[Federal Register Volume 83, Number 73 (Monday, April 16, 2018)]
[Rules and Regulations]
[Pages 16440-16757]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 2018-07179]



[[Page 16439]]

Vol. 83

Monday,

No. 73

April 16, 2018

Part II





 Department of Health and Human Services





-----------------------------------------------------------------------





Centers for Medicare & Medicaid Services





-----------------------------------------------------------------------





42 CFR Parts 405, 417, 422, et al.





Medicare Program; Contract Year 2019 Policy and Technical Changes to 
the Medicare Advantage, Medicare Cost Plan, Medicare Fee-for-Service, 
the Medicare Prescription Drug Benefit Programs, and the PACE Program; 
Final Rule

Federal Register / Vol. 83 , No. 73 / Monday, April 16, 2018 / Rules 
and Regulations

[[Page 16440]]


-----------------------------------------------------------------------

DEPARTMENT OF HEALTH AND HUMAN SERVICES

Centers for Medicare & Medicaid Services

42 CFR Parts 405, 417, 422, 423, 460, and 498

[CMS-4182-F]
RIN 0938-AT08


Medicare Program; Contract Year 2019 Policy and Technical Changes 
to the Medicare Advantage, Medicare Cost Plan, Medicare Fee-for-
Service, the Medicare Prescription Drug Benefit Programs, and the PACE 
Program

AGENCY: Centers for Medicare & Medicaid Services (CMS), HHS.

ACTION: Final rule.

-----------------------------------------------------------------------

SUMMARY: This final rule will revise the Medicare Advantage (MA) 
program (Part C) regulations and Prescription Drug Benefit program 
(Part D) regulations to implement certain provisions of the 
Comprehensive Addiction and Recovery Act (CARA) to further reduce the 
number of beneficiaries who may potentially misuse or overdose on 
opioids while still having access to important treatment options; 
implement certain provisions of the 21st Century Cures Act; support 
innovative approaches to improve program quality, accessibility, and 
affordability; offer beneficiaries more choices and better care; 
improve the CMS customer experience and maintain high beneficiary 
satisfaction; address program integrity policies related to payments 
based on prescriber, provider and supplier status in MA, Medicare cost 
plan, Medicare Part D and the PACE programs; provide an update to the 
official Medicare Part D electronic prescribing standards; and clarify 
program requirements and certain technical changes regarding treatment 
of Medicare Part A and Part B appeal rights related to premiums 
adjustments.

DATES: 
    Effective Date: This rule is effective June 15, 2018.
    The incorporation by reference of certain publications listed in 
the rule is approved by the Director of the Federal Register as of June 
15, 2018.
    Applicability Dates: The applicability date of the provisions of 
this rule is January 1, 2019 except for the provisions in Sec. Sec.  
422.100(f)(4) and (5) and 422.101(d) (discussed in section II.A.4. of 
this final rule (Maximum Out-of-Pocket Limit for Medicare Parts A and B 
Services)) and Sec.  422.100(f)(6) (discussed in section II.A.5. of 
this final rule (Cost Sharing Limits for Medicare Parts A and B 
Services)). Those provisions are applicable for contract year 2020 
(January 1, 2020). E-Prescribing and the Part D Prescription Drug 
Program; Updating Part D E Prescribing Standards discussed in section 
II.D.8. of this final rule is applicable January 1, 2020 conditioned on 
The Office of the National Coordinator for Health Information 
Technology (ONC) adopting the same standard for use in its Electronic 
Health Record Certification Program by that date.

FOR FURTHER INFORMATION CONTACT: 
    Theresa Wachter, (410) 786-1157, Part C Issues.
    Marie Manteuffel, (410) 786-3447, Part D Issues.
    Kristy Nishimoto, (206) 615-2367, Beneficiary Enrollment and 
Appeals Issues.
    Raghav Aggarwal, (410) 786-0097, Part C and D Payment Issues.
    Vernisha Robinson-Savoy, (443) 826-9925, Compliance Program 
Training Issues.
    Frank Whelan, (410) 786-1302, Preclusion List Issues.
    Shelly Winston, (410) 786-3694, Part D E-Prescribing Program.

SUPPLEMENTARY INFORMATION: 

I. Executive Summary and Background

A. Executive Summary

1. Purpose
    The primary purpose of this final rule is to make revisions to the 
Medicare Advantage (MA) program (Part C) and Prescription Drug Benefit 
Program (Part D) regulations based on our continued experience in the 
administration of the Part C and Part D programs and to implement 
certain provisions of the Comprehensive Addiction and Recovery Act and 
the 21st Century Cures Act. The changes are necessary to--
     Support Innovative Approaches to Improving Quality, 
Accessibility, and Affordability;
     Improve the CMS Customer Experience; and
     Implement Other Changes.
    In addition, this final rule makes technical changes related to 
treatment of Part A and Part B premium adjustments and updates the 
NCPDP SCRIPT standard used for Part D electronic prescribing. While the 
Part C and Part D programs have high satisfaction among enrollees, we 
continually evaluate program policies and regulations to remain 
responsive to current trends and newer technologies, and provide 
increased flexibility to serve patients. Specifically, this regulation 
meets the Administration's priorities to reduce burden and provide the 
regulatory framework to develop MA and Part D products that better meet 
the individual patient's health care needs. These changes being 
finalized will empower MA and Part D plans to meet the needs of 
enrollees at the local level, and should result in more enrollee choice 
and more affordable options. Additionally, this regulation includes a 
number of provisions that will help address the opioid epidemic and 
mitigate the impact of increasing drug prices in the Part D program.
2. Summary of the Major Provisions
a. Implementation of the Comprehensive Addiction and Recovery Act of 
2016 (CARA) Provisions
    In line with the agency's response to the President's call to end 
the scourge of the opioid epidemic, this final rule implements 
statutory provisions of the Comprehensive Addiction and Recovery Act of 
2016 (CARA), which amended the Social Security Act and was enacted into 
law on July 22, 2016. CARA includes new authority for Medicare Part D 
plans to establish drug management programs effective on or after 
January 1, 2019. Through this final rule, CMS has established a 
framework under which Part D plan sponsors may establish a drug 
management program for beneficiaries at risk for prescription drug 
abuse or misuse, or ``at-risk beneficiaries.'' Specifically, under drug 
management programs, Part D plans will engage in case management of 
potential at-risk beneficiaries, through contact with their 
prescribers, when such beneficiary is found to be taking a specific 
dosage of opioids and/or obtaining them from multiple prescribers and 
multiple pharmacies who may not know about each other. Sponsors may 
then limit at-risk beneficiaries' access to coverage of controlled 
substances that CMS determines are ``frequently abused drugs'' to a 
selected prescriber(s) and/or network pharmacy(ies) after case 
management with the prescribers for the safety of the enrollee. CMS 
also limits the use of the special enrollment period (SEP) for dually- 
or other low income subsidy (LIS)-eligible beneficiaries by those LIS-
eligible beneficiaries who are identified as at-risk or potentially at-
risk for prescription drug abuse under such a drug management program. 
Finally, these provisions will codify the current Part D Opioid Drug 
Utilization Review (DUR) Policy and Overutilization Monitoring System 
(OMS) by integrating this current policy with drug management program 
provisions.

[[Page 16441]]

Through the adoption of this policy, from 2011 through 2017, there was 
a 76 percent decrease (almost 22,500 beneficiaries) in the number of 
Part D beneficiaries identified as potential very high risk opioid 
overutilizers. Thus, drug management programs will expand upon an 
existing, innovative, successful approach to reduce opioid 
overutilization in the Part D program by improving quality of care 
through coordination while maintaining access to necessary pain 
medications, and will be an important next step in addressing the 
opioid epidemic and safeguarding the health and safety of our nation's 
seniors.
b. Revisions to Timing and Method of Disclosure Requirements
    Consistent with agency efforts supporting innovative approaches to 
improve quality, accessibility, and affordability and reduce burden, we 
are finalizing changes to align the MA and Part D regulations in 
authorizing CMS to set the manner of delivery for mandatory disclosures 
in both the MA and Part D programs. CMS will use this authority to 
allow MA plans to meet the disclosure and delivery requirements for 
certain documents by relying on notice of electronic posting and 
provision of the documents in hard copy when requested, when previously 
the documents, such as the Evidence of Coverage (EOC), had to be 
provided in hard copy. Additionally, we are changing the timeframe for 
delivery of the MA and Part D EOC to the first day of the Annual 
Election Period (AEP), rather than 15 days prior to that date. Allowing 
Part C and Part D plans to provide the EOC electronically will 
alleviate plan burden related to printing and mailing and reduce the 
number of paper documents that enrollees receive from plans. Changing 
the date by which plans must provide the EOC to enrollees will allow 
plans more time to finalize the formatting and ensure the accuracy of 
the information in the EOC. Changing the date will also separate the 
mailing and receipt of the EOC from the Annual Notice of Change (ANOC), 
which describes the important changes in a patient's plan from one year 
to the next. The ANOC must be delivered 15 days prior to the AEP and 
will be received by enrollees ahead of the EOC, thus allowing enrollees 
to focus on materials that drive decision-making during the AEP. We see 
this final change as an overall reduction of burden that our 
regulations have on plans and enrollees. In aggregate, we estimate a 
savings (to plans for not producing and mailing hardcopy EOCs) of 
approximately $54.7 million each year, 2019 through 2023.
c. Preclusion List Requirements for Prescribers in Part D and 
Individuals and Entities in MA, Cost Plans, and PACE
    This final rule will rescind current regulatory provisions that 
require prescribers of Part D drugs and providers of MA services and 
items to enroll in Medicare in order for the Part D drug or MA service 
or item to be covered. As a replacement, a Part D plan sponsor will be 
required to reject, or require its pharmacy benefit manager to reject, 
a pharmacy claim for a Part D drug if the individual who prescribed the 
drug is included on the ``preclusion list.'' Similarly, an MA service 
or item will not be covered if the provider that furnished the service 
or item is on the preclusion list. The preclusion list will consist of 
certain individuals and entities that are currently revoked from the 
Medicare program under 42 CFR 424.535 and are under an active 
reenrollment bar, or have engaged in behavior for which CMS could have 
revoked the individual or entity to the extent applicable if they had 
been enrolled in Medicare, and CMS determines that the underlying 
conduct that led, or would have led, to the revocation is detrimental 
to the best interests of the Medicare program. We believe that this 
change from an enrollment requirement to a preclusion list requirement 
will reduce the burden on Part D prescribers and MA providers without 
compromising our program integrity efforts.
3. Summary of Costs, Savings and Benefits of the Major Provisions

------------------------------------------------------------------------
          Provision           Savings and benefits          Costs
------------------------------------------------------------------------
Implementation of the         The purpose of this   The creation of lock
 Comprehensive Addiction and   provision is to       in-status is a
 Recovery Act of 2016.         create a lock-in      burden to plans.
                               status for certain    The cost to
                               at-risk               industry is
                               beneficiaries. In     estimated at about
                               addition to the       $2.8 million per
                               benefits of           year. This $2.8
                               preventing opioid     million cost arises
                               and benzodiazepine    from (i) the
                               dependency in         uploading and
                               beneficiaries, we     preparing of
                               estimate, in 2019,    additional notices
                               a reduction of $19    to enrollees
                               million in Trust      ($101,721), (ii)
                               Fund expenditures     the re-negotiation
                               because of reduced    of contracts
                               opioid scripts.       between Part D
                               This $19 million      sponsors and
                               reduction modestly    pharmacies
                               increases to a $20    ($547,415), (iii)
                               million reduction     the programming of
                               in 2023.              edits about lock-
                                                     ins into the
                                                     systems of Part D
                                                     sponsors
                                                     ($2,152,332), and
                                                     (iv) the right of
                                                     enrollees to appeal
                                                     a status of lock-in
                                                     ($35,183).
Revisions to Timing and       We estimate 67% of    ....................
 Method of Disclosure          the current 47.8
 Requirements.                 million
                               beneficiaries will
                               prefer use of the
                               internet versus
                               hard copies. This
                               will result in a
                               savings to the
                               industry of $54.7
                               million each year,
                               2019 through 2023.
                               This is due to a
                               reduction in
                               printing and
                               mailing costs.
Preclusion List Requirements  For 2019, this        For 2019, this
 for Prescribers in Part D     provision saves       provision costs
 and Individuals and           providers $34.4       Part D sponsors or
 Entities in MA, Cost Plans,   million. For 2020     their PBMs $9.3
 and PACE.                     and future years,     million. For 2020
                               there are no          and future years,
                               savings. The $34.4    costs are
                               million in savings    negligible (below
                               to providers arises   $50,000). The $9.3
                               because of removal    million cost arises
                               of the requirement    because of
                               of MA providers and   programming and
                               suppliers and Part    staff resources
                               D prescribers to      needed to produce
                               enroll in Medicare    and send required
                               as a prerequisite     notifications to
                               for furnishing        enrollees and
                               health care items     prescribers.
                               and services. Part
                               C providers and
                               suppliers save
                               $24.1 million in
                               reduced costs while
                               Part D providers
                               save $10.3 million
                               in reduced costs.

[[Page 16442]]

 
Physician Incentive Plans--   For 2019, this        ....................
 Update Stop-Loss Protection   provision reduces
 Requirements.                 required
                               reinsurance
                               resources by $204.6
                               million. The $204.6
                               million savings
                               increases yearly
                               because of expected
                               enrollment
                               increases and
                               medical inflation;
                               the savings is
                               $281.8 million in
                               2023. The savings
                               arise because we
                               are replacing the
                               current insurance
                               schedule in the
                               regulation with
                               updated stop-loss
                               insurance
                               requirements that
                               will allow
                               insurance with
                               higher deductibles.
                               This updated
                               schedule will
                               result in a
                               significant
                               reduction to the
                               cost of obtaining
                               stop-loss
                               insurance. The
                               higher deductibles
                               are consistent with
                               the increase in
                               medical costs due
                               to inflation.
                               Through transfers,
                               the 2019 $204.6
                               million savings
                               results in $71.6
                               savings to the
                               Medicare Trust Fund
                               and $133 million
                               savings (in the
                               form of rebates) to
                               Medicare Advantage
                               (MA) organizations.
                               It is likely that
                               some of the savings
                               to MA organizations
                               will result in
                               increased health
                               care benefits to MA
                               enrollees.
------------------------------------------------------------------------

B. Background

    In the proposed rule titled ``Medicare Program; Contract Year 2019 
Policy and Technical Changes to the Medicare Advantage, Medicare Cost 
Plan, Medicare Fee-for-Service, the Medicare Prescription Drug Benefit 
Programs, and the PACE Program'' which appeared in the November 28, 
2017 Federal Register (82 FR 56336), we proposed to revise the Medicare 
Advantage program (Part C) regulations and Prescription Drug Benefit 
program (Part D) regulations to implement certain provisions of the 
Comprehensive Addiction and Recovery Act (CARA) and the 21st Century 
Cures Act; improve program quality, accessibility, and affordability; 
improve the CMS customer experience; address program integrity policies 
related to payments based on prescriber, provider and supplier status 
in Medicare Advantage, Medicare cost plan, Medicare Part D and the PACE 
programs; provide a proposed update to the official Medicare Part D 
electronic prescribing standards; clarify program requirements; and 
make certain technical changes regarding treatment of Medicare Part A 
and Part B appeal rights related to premium adjustments.
    We received approximately 1,669 timely pieces of correspondence 
containing multiple comments on the CY 2019 proposed rule. While we are 
finalizing several of the provisions from the proposed rule, there are 
a number of provisions from the proposed rule that we intend to address 
later and a few that we do not intend to finalize. We also note that 
some of the public comments were outside of the scope of the proposed 
rule. These out-of-scope public comments are not addressed in this 
final rule. Summaries of the public comments that are within the scope 
of the proposed rule and our responses to those public comments are set 
forth in the various sections of this final rule under the appropriate 
heading. However, we note that in this final rule we are not addressing 
comments received with respect to the provisions of the proposed rule 
that we are not finalizing at this time. Rather, we will address them 
at a later time, in a subsequent rulemaking document, as appropriate.

II. Provisions of the Proposed Rule and Analysis of and Responses to 
Public Comments

A. Supporting Innovative Approaches to Improving Quality, 
Accessibility, and Affordability

1. Implementation of the Comprehensive Addiction and Recovery Act of 
2016 (CARA) Provisions
a. Medicare Part D Drug Management Programs
    The Comprehensive Addiction and Recovery Act of 2016 (CARA), 
enacted into law on July 22, 2016, amended the Social Security Act and 
includes new authority for the establishment of drug management 
programs in Medicare Part D, effective on or after January 1, 2019. In 
accordance with section 704(g)(3) of CARA and revised section 1860D-
4(c) of the Act, CMS must establish through notice and comment 
rulemaking a framework under which Part D plan sponsors may establish a 
drug management program for beneficiaries at-risk for prescription drug 
abuse, or ``at-risk beneficiaries.'' Under such a Part D drug 
management program, sponsors may limit at-risk beneficiaries' access to 
coverage of controlled substances that CMS determines are ``frequently 
abused drugs'' to a selected prescriber(s) and/or pharmacy(ies). While 
such programs, commonly referred to as ``lock-in programs,'' have been 
a feature of many state Medicaid programs for some time, prior to the 
enactment of CARA, there was no statutory authority to allow Part D 
plan sponsors to require beneficiaries to obtain controlled substances 
from a certain pharmacy or prescriber in the Medicare Part D program. 
Thus, although drug management programs are voluntary, this rule 
codifies a framework that will place requirements upon such programs 
when established by Part D sponsors.
    This final rule implements the CARA Part D drug management program 
provisions by integrating them with the current Part D Opioid Drug 
Utilization Review (DUR) Policy and Overutilization Monitoring System 
(OMS) (``current policy'').\1\ This integration will mean that Part D 
plan sponsors implementing a drug management program could limit an at-
risk beneficiary's access to coverage of frequently abused drugs 
beginning 2019 through a beneficiary-specific point-of-sale (POS) claim 
edit and/or by requiring the beneficiary to obtain frequently abused 
drugs from a selected

[[Page 16443]]

pharmacy(ies) and/or prescriber(s) after case management and notice to 
the beneficiary. To do so, the beneficiary will have to meet clinical 
guidelines that factor in that the beneficiary is taking opioids over a 
sustained time period and that the beneficiary is obtaining them from 
multiple prescribers and/or multiple pharmacies. This final rule also 
implements a limitation on the use of the special enrollment period 
(SEP) for low income subsidy (LIS)-eligible beneficiaries who are 
identified as potential at-risk beneficiaries or at-risk beneficiaries.
---------------------------------------------------------------------------

    \1\ In using the term ``current policy'', we refer to the aspect 
of our current Part D opioid overutilization policy that is based on 
retrospective DUR and case management. Please refer to the CMS 
website, ``Improving Drug Utilization Review Controls in Part D'' at 
https://www.cms.gov/Medicare/Prescription-Drug-Coverage/PrescriptionDrugCovContra/RxUtilization.html which contains CMS 
communications regarding the current policy.
---------------------------------------------------------------------------

    We received the following general comments and our responses 
follow:
    Comment: Commenters were overall supportive of our proposal. Some 
commenters found it to be a conservative and uniform approach to 
implementing the CARA drug management program provisions. Other 
commenters included specific suggestions for improvements with their 
overall supportive or neutral comments.
    Response: We thank the commenters for their comments. We summarize 
and respond to specific recommendations later in this preamble.
    Comment: We received a request that we confirm that nothing in the 
final rule impacts PACE organizations' waivers of Part D requirements 
in Sec.  423.153. This commenter also asked that existing waivers of 
Sec.  423.153 be extended to include Sec.  423.153(f) unless such a 
waiver is not needed due to the voluntary nature of drug management 
programs.
    Response: PACE organizations are not excluded from OMS reporting 
under the current policy. Additionally, because of the voluntary nature 
of the provisions under Sec.  423.153(f), a waiver is not necessary for 
PACE organizations. However, to the extent that PACE organizations 
commence drug utilization management activities covered under Sec.  
423.153(f), PACE organizations must comply with the requirements of 
423.153(f).
    Comment: We received comments that expressed concern about the time 
needed for Part D plan sponsors to make the necessary systems changes 
to implement compliant drug management programs.
    Response: Section 704(g)(1) of CARA states that the amendments made 
by this section shall apply to prescription drug plans (and MA-PD 
plans) for plan years beginning on or after January 1, 2019. However, 
given the current national opioid epidemic, we expect that Part D 
sponsors will diligently implement fully-functional drug management 
programs in 2019. Moreover, as the new requirements for drug management 
programs build from and are integrated with existing policy, we expect 
sponsors will be able to implement them expeditiously.
    Comment: We received one suggestion that CMS pilot different 
approaches for implementing the CARA drug management program 
provisions, specifically the ``lock-in'' provisions, as we did before 
implementing our current policy.
    Response: Because the CARA drug management provisions will be 
integrated with our current policy, albeit with some modifications to 
that policy, we are not persuaded that an additional pilot is necessary 
since plan sponsors already have experience with addressing potential 
opioid overutilization.
    Comment: A commenter requested that CMS acknowledge the work it 
will take for Standard Development Organizations (SDOs) to implement 
the finalized CARA provisions. In particular, the commenter noted that 
development of any codes and messaging associated with the new CARA-
related requirements will take time to implement.
    Response: We understand that any modifications to existing 
standards to accurately achieve the desired functionalities to further 
the electronic exchange of information between healthcare stakeholders 
about the final CARA provisions may require time. We rely on SDOs to 
coordinate these efforts, and CMS is committed to working with the SDOs 
during this process, if needed.
    Comment: A commenter requested clarification on how to handle 
concurrent DUR edits, such as formulary-level cumulative opioid MME 
safety edits, and the drug management program. Specifically, the 
comment sought clarification on whether the drug management program 
beneficiary-specific POS claim edits or lock-in limitations would take 
precedence over an approved exception to a cumulative opioid MME safety 
edit.
    Response: A plan sponsor may implement formulary-level coverage 
rules for opioids (that is, prior authorization, quantity limits or 
step therapy) or safety edits, and implement a drug management program. 
The formulary and coverage rules would apply to all enrollees (unless 
they obtain an exception), and the drug management program would apply 
to potential at-risk and at-risk beneficiaries. A Part D sponsor's 
concurrent and retrospective DUR programs should be closely 
coordinated. In certain circumstances, it may be appropriate for a 
sponsor to make an at-risk determination through the drug management 
program for a beneficiary who received an approved exception to a 
cumulative opioid MME safety edit, and as part of the at-risk 
determination, may determine that continuing the approved exception is 
no longer appropriate.
    For example, a plan implemented a hard formulary-level cumulative 
MME opioid edit at 200 MME with 2 or more opioid prescribers. A 
beneficiary received their opioids from 2 prescribers and has a 
cumulative MME that exceeds 200 MME. They trigger the edit and request 
a coverage determination. The prescriber attests to medical necessity 
and the exception request is approved. At a later time, the beneficiary 
seeks opioids from 3 additional prescribers, and meets the CARA/OMS 
criteria. Through case management, the prescriber verifies the 
beneficiary is at-risk and agrees to prescriber lock-in due to care 
coordination issues.
b. Integration of CARA and the Current Part D Opioid DUR Policy and OMS
    Our proposal was to integrate the CARA Part D drug management 
program provisions with our current policy and codify them both. 
Specifically, under this regulatory framework, we proposed that Part D 
plan sponsors may voluntarily adopt drug management programs through 
which they address potential overutilization of frequently abused drugs 
identified retrospectively through the application of clinical 
guidelines/OMS criteria that identify potential at-risk beneficiaries 
and conduct case management which incorporates clinical contact and 
prescriber verification that a beneficiary is an at-risk beneficiary. 
If deemed necessary, a sponsor could limit at-risk beneficiaries' 
access to coverage for such drugs through pharmacy lock-in, prescriber 
lock-in, and/or a beneficiary-specific point-of-sale (POS) claim edit. 
Finally, sponsors would report to CMS the status and results of their 
case management through OMS and any beneficiary coverage limitations 
they have implemented through MARx, CMS' system for payment and 
enrollment transactions. Thus, although drug management programs are 
voluntary, our proposal was to codify a framework that will place 
requirements upon such programs when established by Part D sponsors.
    We stated that we foresee that all plan sponsors will implement 
such drug management programs based on our experience that all plan 
sponsors are complying with the current policy; the fact that our 
proposal largely incorporates the CARA drug management provisions into 
existing

[[Page 16444]]

CMS and sponsor operations; and especially, in light of the national 
opioid epidemic and the declaration that the opioid crisis is a 
nationwide Public Health Emergency.
    Comment: Commenters expressed strong support for integrating the 
drug management program provisions of CARA with the current policy. 
Commenters expressed that our proposal is reasonable, thoughtful, 
thorough, practical, and comprehensive; that it builds on a successful 
existing Medicare Part D program; that it will involve a common set of 
procedures and help ensure a streamlined and efficient process rather 
than creating a separate one that would require additional oversight 
and add administrative burden. We did not receive comments that opposed 
integrating the drug management program provisions of CARA with the 
current policy.
    Response: We thank the commenters for their supportive comments and 
are finalizing this integration approach to our proposal.
(1) Requirements for Part D Drug Management Programs (Sec. Sec.  
423.100 and 423.153)
    We proposed the following definitions in establishing requirements 
for Part D drug management programs.
(i) Definitions (Sec.  423.100)
(A) Definition of ``Potential At-Risk Beneficiary'' and ``At-Risk 
Beneficiary'' (Sec.  423.100)
    Section 1860D-4(c)(5)(C) of the Act contains a definition for ``at-
risk beneficiary'' that we proposed to codify at Sec.  423.100. In 
addition, although the section 1860D-4(c)(5) of the Act does not 
explicitly define a ``potential at-risk beneficiary,'' it refers to a 
beneficiary who is potentially at-risk in several subsections.
    Accordingly, we proposed to define these two terms at Sec.  423.100 
as follows: Potential at-risk beneficiary means a Part D eligible 
individual--(1) Who is identified using clinical guidelines (as defined 
in Sec.  423.100); or (2) With respect to whom a Part D plan sponsor 
receives a notice upon the beneficiary's enrollment in such sponsor's 
plan that the beneficiary was identified as a potential at-risk 
beneficiary (as defined in paragraph (1) of this definition) under the 
prescription drug plan in which the beneficiary was most recently 
enrolled, such identification had not been terminated upon 
disenrollment, and the new plan has adopted the identification.
    At-risk beneficiary means a Part D eligible individual--(1) who 
is--(i) Identified using clinical guidelines (as defined in Sec.  
423.100); (ii) Not an exempted beneficiary; and (iii) Determined to be 
at-risk for misuse or abuse of such frequently abused drugs under a 
Part D plan sponsor's drug management program in accordance with the 
requirements of Sec.  423.153(f); or (2) With respect to whom a Part D 
plan sponsor receives a notice upon the beneficiary's enrollment in 
such sponsor's plan that the beneficiary was identified as an at-risk 
beneficiary (as defined in paragraph (1) of this definition) under the 
prescription drug plan in which the beneficiary was most recently 
enrolled, such identification had not been terminated upon 
disenrollment, and the new plan has adopted the identification. We 
noted that we included the phrase, ``and the new plan has adopted the 
identification'' to both definitions for cases where a beneficiary has 
been identified as a potential at-risk or at-risk beneficiary by the 
immediately prior plan to indicate that the beneficiary's status in the 
subsequent plan is not automatic.
    We received the following comments and our response follows:
    Comment: A commenter did not believe that a definition for a 
``potential at-risk beneficiary'' was needed, nor the additional 
prescriber verification the commenter associated with the definition.
    Response: We disagree. Although as we noted above, section 1860D-
4(c)(5) of the Act does not explicitly define a ``potential at-risk 
beneficiary,'' it refers to a beneficiary who is potentially at-risk in 
section 1860D-4(c)(5)(B)(ii), which addresses initial notices; in 
1860D-4(c)(5)(H)(i) which addresses data disclosures; and in section 
1860D-4(c)(5)(I) which addresses the sharing of information for 
subsequent plan enrollments. Therefore, we proposed to define a 
potential at-risk beneficiary in Sec.  423.100, as the CARA drug 
management program provisions clearly contemplate this status for a 
beneficiary.
    With respect to additional prescriber verification of a potential 
at-risk beneficiary, we believe this comment is based on a 
misunderstanding of our proposal, as we did not propose that a 
beneficiary's status as a potential at-risk beneficiary must be 
verified. Rather, we proposed and are finalizing a requirement, as we 
discuss later in this preamble, that a prescriber must verify that a 
beneficiary is at-risk, which serves as his or her professional opinion 
that a Part D plan sponsor takes into account during case management.
    Comment: We received a question whether an individual who is 
subject to lock-in under his or her Medicaid program and then becomes 
dually-eligible constitutes a potential or at-risk beneficiary under 
our proposed definitions.
    Response: Such a beneficiary would not automatically be considered 
to be a potential at-risk or an at-risk beneficiary under a Part D 
sponsor's drug management program. Rather, whether such a beneficiary 
is a potential at-risk or at-risk beneficiary would depend upon whether 
he or she meets the clinical guidelines and is determined to be an at-
risk beneficiary under the process set forth in this rule. An automatic 
determination based on a beneficiary's inclusion and status in a 
Medicaid drug management program would not be appropriate because each 
Medicaid drug management program has its own criteria and requirements 
for reviewing and addressing recipients who may be at-risk for 
prescription drug abuse or misuse and its own interventions. We also 
note that Medicaid programs are not required to comply with section 
1860D-4(c)(5) as Part D drug management programs are.
    To the extent a Part D sponsor is aware or discovers based on 
reliable information that a beneficiary who meets the clinical 
guidelines was locked-in under a Medicaid drug management program, that 
sponsor may consider that information in deciding whether to determine 
that a beneficiary is an at-risk beneficiary under the requirements of 
this final rule. Also, any beneficiary entering the Part D program will 
be immediately subject to their plan's formulary-level controls to 
address opioid overutilization before they may be identified as 
potentially at-risk, so any opioid overutilization by the beneficiary 
in his or her new Part D plan may be addressed by these controls.
    Comment: We received a comment requesting clarification with regard 
to a person who is locked-in under an employer plan and then becomes 
eligible for a Part D EGWP, if the EGWP can continue the lock-in in the 
Part D plan or at least consider the prior lock-in as part of a new 
determination.
    Response: Beginning with plan year 2019, Part D sponsors, including 
sponsors of EGWPs, may adopt drug management programs that meet the 
requirements we are finalizing in this rule. Under a Part D 
prescription drug management program, sponsors may implement a 
prescriber and/or pharmacy lock-in or beneficiary-specific POS claim 
edit for frequently abused drugs with respect to an at-risk 
beneficiary. Similar to a Medicaid beneficiary who becomes newly 
eligible for Medicare and enrolls in Part D, a person who is locked-in 
under a

[[Page 16445]]

commercial plan does not automatically meet the definition of an at-
risk beneficiary we are finalizing in Sec.  423.100. Rather, such a 
person first must be determined to be an at-risk beneficiary in 
accordance with the requirements we are finalizing at Sec.  423.153(f).
    In other words, in order for a beneficiary to be eligible to be 
immediately locked-in to a prescriber or pharmacy in a Part D plan in 
which they are newly enrolled, the plan from which they most recently 
disenrolled must be a Part D plan in which he or she was determined to 
be an at-risk beneficiary under that plan's drug management program. 
When a new enrollee comes from a non-Part D plan in which the 
beneficiary was subject to lock-in, however, the sponsor can consider 
the prior lock-in if it learns or knows of it based upon reliable 
information which is legally available to the sponsor in conjunction 
with the information it gathers from the case management process, the 
beneficiary, and the sponsor's other relevant internal sources and 
data.
    Comment: A commenter asked if a Part D sponsor may consider opioid 
utilization information from external sources during case management, 
such as a state prescription drug monitoring program (PDMP) in making 
the determination if a beneficiary is at-risk.
    Response: As noted above with respect to beneficiaries who were 
locked-in under an employer or Medicaid plan before enrolling in 
Medicare Part D, we encourage sponsors to use all reliable sources 
legally available to them to obtain an accurate account of a potential 
at-risk or at-risk beneficiary's utilization of frequently abused 
drugs.
    After considering the comments, we are finalizing the definition of 
potential at-risk beneficiary and at-risk beneficiary with minor 
modifications for clarity. First, we are removing the phrase ``and the 
new plan adopted the identification'' from paragraph (2) of both 
definitions. As we noted above, the purpose of this language was to 
indicate that the beneficiary's at-risk status in the subsequent plan 
is not automatic, which we meant for purposes of the limitation on the 
special enrollment period (SEP) for LIS beneficiaries with an at-risk 
status. However, as we discuss later in this preamble, this limitation 
will be triggered or continued by Part D sponsors sending the initial 
and second notices to such beneficiaries, as applicable, so we no 
longer believe this phrase is necessary in these definitions.
    Second, we also are making a minor clarifying change in the 
definition of at-risk beneficiary to explicitly acknowledge that it is 
the Part D sponsor that determines which beneficiaries are at-risk 
beneficiaries under its drug management program.
    The definition of potential at-risk beneficiary will read: A Part D 
eligible individual--(1) Who is identified using clinical guidelines 
(as defined in Sec.  423.100); or (2) With respect to whom a Part D 
plan sponsor receives a notice upon the beneficiary's enrollment in 
such sponsor's plan that the beneficiary was identified as a potential 
at-risk beneficiary (as defined in paragraph (1) of this definition) 
under the prescription drug plan in which the beneficiary was most 
recently enrolled and such identification had not been terminated upon 
disenrollment. The definition of at-risk beneficiary will read: At-risk 
beneficiary means a Part D eligible individual--(1) Who is--(i) 
Identified using clinical guidelines (as defined in Sec.  423.100); 
(ii) Not an exempted beneficiary; and (iii) Determined to be at-risk 
for misuse or abuse of such frequently abused drugs by a Part D plan 
sponsor under its drug management program in accordance with the 
requirements of Sec.  423.153(f); or (2) With respect to whom a Part D 
plan sponsor receives a notice upon the beneficiary's enrollment in 
such sponsor's plan that the beneficiary was identified as an at-risk 
beneficiary (as defined in the paragraph (1) of this definition) under 
the prescription drug plan in which the beneficiary was most recently 
enrolled and such identification had not been terminated upon 
disenrollment.
(B) Definition of ``Frequently Abused Drug'', ``Clinical Guidelines'', 
``Program Size'', and ``Exempted Beneficiary'' (Sec.  423.100)
    Because we use these terms in the proposed definitions of 
``potential at-risk beneficiary'' and ``at-risk beneficiary,'' we 
proposed to define ``frequently abused drug'', ``clinical guidelines'', 
``program size'', and ``exempted beneficiary'' at Sec.  423.100 as 
follows:
 Frequently Abused Drug
    Section 1860D-4(c)(5)(G) of the Act defines ``frequently abused 
drug'' as a drug that is a controlled substance that the Secretary 
determines to be frequently abused or diverted. Consistent with the 
statutory definition, we proposed to define ``Frequently abused drug'' 
at Sec.  423.100 to mean a controlled substance under the Federal 
Controlled Substances Act that the Secretary determines is frequently 
abused or diverted, taking into account the following factors: (1) The 
drug's schedule designation by the Drug Enforcement Administration; (2) 
Government or professional guidelines that address that a drug is 
frequently abused or misused; and (3) An analysis of Medicare or other 
drug utilization or scientific data. This definition is intended to 
provide enough specificity for stakeholders to know how the Secretary 
will determine a frequently abused drug, while preserving flexibility 
to update which drugs CMS considers to be frequently abused drugs based 
on relevant factors, such as actions by the Drug Enforcement 
Administration and/or trends observed in Medicare or scientific data. 
Since we did not receive any specific comments to change this 
definition, we are finalizing it as proposed.
    Comment: A commenter requested that CMS include the criteria, 
resources, and the evidence basis upon which it will rely to determine 
that a drug is a frequently abused drug for purposes of a drug 
management program.
    Response: The definition of frequently abused drug that we are 
finalizing indicates that criteria, resources, and evidence basis will 
be the DEA schedule designation, government, and professional drug 
guidelines, and analyses of drug utilization or scientific data.
    We did not receive any further comment on the definition of 
``frequently abused drug'' and are therefore finalizing it as proposed.
    Consistent with current policy, we proposed that opioids are 
frequently abused drugs, except buprenorphine for medication-assisted 
treatment (MAT) and injectables. As we stated in the preamble to the 
proposed rule, we plan to publish and update a list of frequently 
abused drugs for purposes of Part D drug management programs.
    Comment: All commenters agreed that the Secretary should determine 
that opioids are frequently abused drugs, many referencing the national 
opioid overuse epidemic.
    Response: We appreciate that stakeholders are focused on the opioid 
public health emergency.
    Comment: Some of these commenters agreed with our proposal to 
determine only opioids, except buprenorphine for medication-assisted 
treatment (MAT) and injectables, as frequently abused drugs, at least 
in the initial implementation of Part D drug management programs, in 
order to allow CMS and stakeholders to focus on opioid overuse and gain 
experience with the use of lock-in as a tool to address overutilization 
in the Part D program, before potentially determining other controlled 
substances as

[[Page 16446]]

frequently abused drugs. These commenters urged CMS to wait until drug 
management programs were established, and testing and monitoring 
indicate that the program can be administered in a manner that does not 
limit beneficiary access to needed medications before expanding the 
programs further. Some of these commenters were concerned that an at-
risk beneficiary would have to obtain all frequently abused drugs from 
one pharmacy or one prescriber and that this could disrupt patient care 
if the pharmacy did not carry all frequently abused drugs.
    However, some commenters urged us to determine that all controlled 
substances are frequently abused drugs. These commenters were 
particularly focused on a determination as to benzodiazepines, and to a 
lesser extent, muscle relaxants. Due to this focus, these commenters 
referred to the CDC Guideline that specifically recommends that 
clinicians avoid prescribing opioid pain medication and benzodiazepines 
concurrently whenever possible due to increased risk for overdose. They 
also referred to CMS work in this area: (1) The fact that CMS added a 
concurrent benzodiazepine-opioid flag to OMS in October 2016 in 
response to the CDC Guideline and after our own research on the use of 
benzodiazepines among Medicare beneficiaries \2\ to alert Part D 
sponsors that concurrent use may be an issue that should be addressed 
during case management; \3\ and (2) the fact that we have stated that a 
sponsor may implement a beneficiary-specific claim edit at POS for non-
opioid medications under the current policy.\4\ They further referred 
to a statistic from the National Institute on Drug Abuse that 30 
percent of overdoses involving opioids also involve benzodiazepines.\5\ 
Finally, these commenters pointed out that the FDA has found that the 
growing combined use of opioid medicines with benzodiazepines or other 
drugs that depress the central nervous system has resulted in serious 
side effects, including slowed or difficult breathing and deaths. These 
commenters further noted that in an effort to decrease the use of 
opioids and benzodiazepines, and opioids and other such depressants, 
the FDA added Boxed Warnings--its strongest warnings--to the drug 
labeling of prescription opioid pain and cough medicines, and 
benzodiazepines.\6\ Given these developments, these commenters stressed 
the importance of Part D plan sponsors being able to use the tools that 
will be available to them under drug management programs to address the 
dangers of concurrent opioid and benzodiazepine use.
---------------------------------------------------------------------------

    \2\ https://www.cms.gov/Medicare/Prescription-Drug-Coverage/PrescriptionDrugCovContra/Downloads/Concurrent-Use-of-Opioids-and-Benzodiazepines-in-a-Medicare-Part-D-Population-CY-2015.pdf.
    \3\ Please refer to the memo, ``Medicare Part D Overutilization 
Monitoring System (OMS) Update: Addition of the Concurrent Opioid-
Benzodiazepine Use Flag'' dated October 21, 2016.
    \4\ Supplemental Guidance Related to Improving Drug Utilization 
Review Controls in Part D'' September 6, 2012.
    \5\ https://www.drugabuse.gov/drugs-abuse/opioids/benzodiazepines-opioids.
    \6\ https://www.fda.gov/Drugs/DrugSafety/ucm518473.htm.
---------------------------------------------------------------------------

    Response: In light of these comments, we are persuaded that it is 
appropriate that drug management programs are able to address 
concurrent opioid and benzodiazepine use. Such a determination is 
consistent with the definition of frequently abused drugs that we are 
finalizing. First, the Secretary determines benzodiazepines are 
frequently abused or diverted, taking into account that they are 
controlled substances under the Controlled Substances Act (CSA) and 
that prescription benzodiazepines are on Schedule IV, where the DEA 
places substances that have a potential for abuse. In addition, the 
Secretary takes into account that the FDA has issued a warning about 
the risks associated with using opioids and benzodiazepines 
concurrently. Further, the CDC included in its evidence-based opioid 
prescribing guideline a caution to co-prescribe opioids and 
benzodiazepines. Finally, CMS' own statistics reveal that 51 percent of 
Part D beneficiaries that will be identified as potentially at-risk 
under the 2019 clinical guidelines we are finalizing are using opioids 
and benzodiazepines concurrently compared to 24 percent across all Part 
D opioid users. This statistic is indicative that concurrent use is 
even more of a danger among potential at-risk beneficiaries than 
Medicare Part D beneficiaries generally. Therefore, the Secretary 
determines that benzodiazepines are a frequently abused drug for 
purposes of Part D drug management programs beginning in 2019. However, 
the clinical guidelines will still only consider a beneficiary's opioid 
use, as we explain just below.
    Comment: A commenter agreed with our statement in the proposed rule 
that there is difficulty in establishing overuse guidelines for non-
opioid substances. The commenter stated that this underscores the need 
for a robust evidence base to support determining that additional types 
of drugs are frequently abused drugs.
    Response: We agree with the commenter's concern, and for this 
reason we are not modifying the clinical guidelines for 2019 to include 
benzodiazepine use, even though benzodiazepines will be considered a 
frequently abused drug for 2019. This means that a beneficiary who is 
determined to be at-risk based on clinical guidelines that look at the 
beneficiary's opioid use could have a coverage limitation applied under 
a drug management program to both opioids and benzodiazepines to manage 
current and future concurrent use. For example, a sponsor could require 
an at-risk beneficiary to obtain both opioids and benzodiazepines from 
one selected pharmacy.
    We believe that this is appropriate based on the robust evidence 
that concurrent benzodiazepine use with opioids results in an even 
higher risk of an adverse health event than use of opioids alone. We 
will expect to rarely see a sponsor apply a limitation only to an at-
risk beneficiary's access to coverage for benzodiazepines, since to do 
so, the beneficiary would have to have met the clinical guidelines 
which look at opioid use that is potentially risky. However, we 
acknowledge that prescriber agreement during case management could 
rarely lead to such an outcome. For example, no opioid prescriber 
agrees to a beneficiary-specific POS claim edit for opioids, but 
rather, all but one states they will no longer prescriber opioids to 
coordinate the beneficiary's use. However, the benzodiazepine 
prescriber agrees to such an edit for benzodiazepines. We discuss 
prescriber agreement in more detail later in this preamble.
    Given that we are finalizing two categories of drugs as frequently 
abused drugs for 2019, depending upon what a plan sponsor learns during 
case management, we reiterate that the sponsor may have to permit a 
beneficiary to obtain frequently abused drugs from more than one 
pharmacy and/or more than one prescriber in order to provide reasonable 
access, if the sponsor applies lock-in as a coverage limitation, which 
we discuss later in this preamble.
    Comment: A few commenters suggested that Part D sponsors be able to 
expand their drug management programs to include additional frequently 
abused drugs based on their experience with their enrollees. One 
suggested that a sponsor be required to submit such an expansion to CMS 
for approval.
    Response: We disagree with this comment. Section 1860D-4(c)(5)(G) 
of the Act defines ``frequently abused

[[Page 16447]]

drug'' as a drug that is a controlled substance that the Secretary 
determines to be frequently abused or diverted. Consistent with this 
statutory provision, we believe it is appropriate that the 
determination of frequently abused drugs not be plan-specific, but 
rather be consistent across Part D plans, as this will permit better 
oversight and promote consistency across all Part D drug management 
programs.
    We proposed that future determinations of frequently abused drugs 
by the Secretary primarily be included in the annual Medicare Parts C&D 
Call Letter or in similar guidance, if necessary, to address midyear 
entries to the drug market or evolving government or professional 
guidelines or relevant data analysis, which will be subject to public 
comment. We proposed that this approach would be consistent with our 
approach under the current policy and necessary for Part D drug 
management programs to be responsive to changing public health issues 
over time.
    Comment: We received comments supportive of our proposal to apply 
the standards we are establishing in rulemaking to future 
determinations of frequently abused drugs through the annual Medicare 
Parts C&D Call Letter, or in similar guidance. We did not receive any 
comments that opposed this proposed approach.
    Response: We appreciate the comments.
    Comment: A commenter asked us to confirm that we would use the same 
process to determine that a drug is no longer a frequently abused drug.
    Response: We will apply the same regulatory standards and use the 
same process that we use to determine that a drug is a frequently 
abused drug when determining that a drug no longer is a frequently 
abused drug for purposes of Part D drug management programs.
    Comment: A few commenters urged CMS to exclude abuse-deterrent (AD) 
opioids from this definition of ``frequently abused drug'' as there is 
no evidentiary data to support the thesis that AD opioids are 
frequently abused and existing observation data supports their 
exclusion from this broad standard.
    Response: The FDA requires a boxed warning on opioid abuse-
deterrent formulations (ADFs), because even with these formulations 
there is still potential for addiction, abuse, misuse, and diversion. 
The FDA has also noted \7\ that ``abuse-deterrent technologies have not 
yet proven successful at deterring the most common form of abuse--
swallowing a number of intact capsules or tablets to achieve a feeling 
of euphoria. Moreover, the fact that a product has abuse-deterrent 
properties does not mean that there is no risk of abuse. It means, 
rather, that the risk of abuse is lower than it would be without such 
properties.'' Also, ADFs do not prevent patients who may be using 
opioids for therapeutic reasons from taking higher doses than 
prescribed or diverting the opioid. For these reasons, we disagree that 
abuse-deterrent formulations should be excluded from the determination 
of frequently abused drugs.
---------------------------------------------------------------------------

    \7\ ``Abuse-Deterrent Opioids--Evaluation and Labeling Guidance 
for Industry'', U.S. Department of Health and Human Services, Food 
and Drug Administration, Center for Drug Evaluation and Research 
(CDER), Clinical Medical, April 2015.
---------------------------------------------------------------------------

    Comment: A few commenters asked CMS to clarify whether methadone, a 
Part D drug when indicated for pain, would be included in the 
definition of a frequently abused drug under the drug management 
program. Other commenters agreed with excluding buprenorphine for MAT 
from the definition of frequently abused drug as not to limit patient 
access to treatment and noted that removing buprenorphine as a 
frequently abused drug is consistent with the CDC's approach to exclude 
buprenorphine from the determination of a person's daily opioid MME.
    Response: Yes, methadone for pain is included in the definition of 
a frequently abused drug for purposes of Part D drug management 
programs, consistent with current policy/OMS. Although buprenorphine is 
recognized by the DEA as a drug of abuse, we thank the commenters that 
agreed with excluding buprenorphine for MAT from the definition of 
frequently abused drug so that access to MAT, such as buprenorphine, is 
not impacted. However, the commenters' reference to the CDC's exclusion 
of buprenorphine from the determination of a person's daily opioid MME 
made us believe that commenters may be conflating the definition of a 
frequently abused drug with the clinical guidelines and associated 
opioid dosage thresholds. Therefore, we realize that we need to be more 
specific about what opioid use, opioid prescribers, and opioid 
dispensing pharmacies means in the clinical guidelines, which we also 
discuss later.
    Since the publication of the proposed rule, the CDC removed the 
conversion factors for all formulations of buprenorphine, for pain and 
for MAT, from the most recent CDC MME conversion factor file (https://www.cdc.gov/drugoverdose/data-files/CDC_Oral_Morphine_Milligram_Equivalents_Sept_2017.xlsx). Therefore, CMS 
cannot determine the MME. As such, buprenorphine products are not used 
to determine the beneficiary's average daily MME. However, we will 
still use prescription opioids, including all formulations of 
buprenorphine for pain and MAT, to determine opioid prescribers and 
opioid dispensing pharmacies in the clinical guidelines.
 Clinical Guidelines & Program Size
    Section 1860D-4(c)(5)(C)(i)(I) of the Act requires at-risk 
beneficiaries to be identified using clinical guidelines that indicate 
misuse or abuse of frequently abused drugs and that are developed by 
the Secretary in consultation with stakeholders. We proposed to include 
a definition of ``clinical guidelines'' that cross references standards 
that we proposed at Sec.  423.153(f) for how the guidelines will be 
established and updated. Specifically, we proposed to define clinical 
guidelines for purposes of a Part D drug management program in Sec.  
423.100 as criteria to identify potential at-risk beneficiaries who may 
be determined to be at-risk beneficiaries under such programs, and that 
are developed in accordance with the standards in Sec.  423.153(f)(16) 
and beginning with contract year 2020, will be published in guidance 
annually.
    We also proposed to add Sec.  423.153(f)(16) to state that 
potential at-risk beneficiaries and at-risk beneficiaries are 
identified by CMS or a Part D sponsor using clinical guidelines that: 
(1) Are developed with stakeholder consultation; (2) Are based on the 
acquisition of frequently abused drugs from multiple prescribers, 
multiple pharmacies, the level of frequently abused drugs, or any 
combination of these factors; (3) Are derived from expert opinion and 
an analysis of Medicare data; and (4) Include a program size estimate. 
This proposed approach to developing and updating the clinical 
guidelines is intended to provide enough specificity for stakeholders 
to know how CMS will determine the guidelines by identifying the 
standards we will apply in determining them.
    This proposed approach also indicated that the program size will be 
determined as part of the process to develop the clinical guidelines--a 
process into which stakeholders will provide input. Section 1860D-
4(c)(5)(C)(iii) of the Act states that the Secretary shall establish 
policies, including the guidelines and exemptions, to ensure that the 
population of enrollees in drug management programs could be 
effectively managed by plans. We proposed to define ``program size'' in

[[Page 16448]]

Sec.  423.100 to mean the estimated population of potential at-risk 
beneficiaries in drug management programs (described in Sec.  
423.153(f)) operated by Part D plan sponsors that the Secretary 
determines, as part of the process to develop clinical guidelines, can 
be effectively managed by such sponsors.
    Comment: We did not receive any specific comments about the 
definition we proposed for clinical guidelines in Sec.  423.100, nor 
the standards we proposed in Sec.  423.153(f)(16).
    Response: We are therefore finalizing the definition and standards 
as proposed, with one modification adding language so that the 
guidelines will be published in guidance annually beginning with 
contract year 2020 guidance, since we are publishing the 2019 clinical 
guidelines in this final rule.
    Comment: We received comments supportive of our proposal to apply 
the standards we are establishing in rulemaking for clinical guidelines 
in Sec.  423.153(f)(16) to develop future OMS criteria through the 
annual Medicare Parts C&D Call Letter process beginning with plan year 
2020.
    We did not receive comments that specifically opposed this proposed 
approach.
    Response: We appreciate these comments.
    Because Part D drug management programs will be integrated with the 
current policy/OMS beginning in 2019, there will be no separate OMS 
criteria in 2019 and beyond. For plan year 2019, we proposed the 
clinical guidelines to be the OMS criteria established for plan year 
2018. The clinical guidelines for use in drug management programs we 
proposed for 2019 are: Use of opioids with an average daily MME greater 
than or equal to 90 mg for any duration during the most recent 6 months 
and either: 4 or more opioid prescribers and 4 or more opioid 
dispensing pharmacies OR 6 or more opioid prescribers, regardless of 
the number of opioid dispensing pharmacies.
    We estimated that these criteria would identify approximately 
33,053 potential at-risk beneficiaries in the Part D program based on 
2015 data, whom we believe are at the highest risk of death or overdose 
due to their opioid use. Also, under our proposal, we stated that Part 
D plan sponsors will not be able to vary the criteria of the guidelines 
to include more or fewer beneficiaries in their drug management 
programs, as they may under the current policy, except that we proposed 
to continue to permit plan sponsors to apply the criteria more 
frequently than CMS will apply them through OMS in 2018, which can 
result in sponsors identifying beneficiaries earlier. This is because 
CMS evaluates enrollees quarterly using a 6-month look back period, 
whereas sponsors may evaluate enrollees more frequently (for example, 
monthly).
    We also described other clinical guidelines that we considered in 
the Regulatory Impact Analysis section of the proposed rule. 
Stakeholders were invited to comment on those options and any others 
that would identify more or fewer potential at-risk beneficiaries.
    Comment: We received comments that were overall supportive of the 
clinical guidelines/criteria we proposed for 2019 with the estimated 
program size of 33,053. However we did receive a few comments 
suggesting criteria for the clinical guidelines that were not among the 
alternate options we included in the RIA. Some of these supportive 
comments supported the guidelines without reservation, making 
statements such as noting the guidelines align with the CDC Guideline 
or that they understood or supported CMS' desire to gain experience 
with the use of lock-in as a drug management tool before adopting 
clinical guidelines with flexibility and/or that would identify more 
potential at-risk beneficiaries. These commenters want CMS to adopt a 
clear and universal set of guidelines which minimizes customer and 
provider confusion, as well as administrative burden when submitting 
and receiving OMS quarterly reports. These commenters assert that 
voluntary plan guidelines would increase confusion and fragmentation 
across the Medicare landscape. However, some commenters urged that Part 
D plan sponsors should have complete flexibility to identify potential 
at-risk beneficiaries, or at least some flexibility to identify 
additional ones consistent with our current policy. These commenters 
emphasized that sponsors should be able to establish and update 
targeting criteria and program features based on evolving clinical 
evidence and feedback and the specific needs of their members. Some of 
these commenters referred to the experience Part D sponsors and their 
PBMs have gained in identifying opioid overutilization among their plan 
members over the last several years and the need to be able to do more 
to address the opioid overuse crisis. Some commenters referred in 
particular to beneficiaries who do not have an average daily MME of 
greater or equal to 90 mg but who are filling opioids prescriptions 
from many different prescribers or pharmacies that they may currently 
address but would not be able to under our proposal. These commenters 
pointed out that such beneficiaries benefit from better coordination of 
care, which case management and coverage limitations on frequently 
abused drugs can support. Another commenter referred to beneficiaries 
with high dose utilization regardless of the number of prescribers as 
appropriate for review by drug management programs.
    As to program size, a commenter stated that the proposed clinical 
guidelines would identify a reasonable number of potential at-risk 
beneficiaries. Another commenter proposed alternative criteria 
involving a lower MME level that it stated would identify more than 
300,000 Part D beneficiaries as potentially at-risk, whereas the other 
commenters (including those commenters that requested increased 
flexibility) did not provide a program size estimate. On the other 
hand, we did not receive comments that the clinical guidelines we 
proposed would identify a potential at-risk beneficiary population that 
cannot be effectively managed by Part D plan sponsors, and because the 
proposed guidelines are the same as the OMS criteria for 2018 that were 
established through the 2018 Parts C&D Call Letter process, we did not 
expect such comments.
    We received a few comments that the proposed clinical guidelines 
appear to be aimed at primarily limiting the program size arbitrarily 
rather than permitting scientific evidence and clinical research to 
dictate the most appropriate guidelines.
    Response: We appreciate the commenters that provided a specific 
suggestion for criteria; however, these criteria were not among the 
alternate options we included in the RIA. Therefore, we decline to 
adopt these suggestions, as the clinical guidelines are to be developed 
by the Secretary in consultation with stakeholders.
    We were persuaded by the commenters that Part D sponsors should 
have some flexibility in adopting targeting criteria for potential at-
risk beneficiaries in order to be able to identify more such 
beneficiaries, which in turn enables sponsors to be able to do more to 
address the opioid overuse public health emergency. In addition, 
flexibility in adopting targeting criteria for potential at-risk 
beneficiaries is consistent with the current policy, and we wish to be 
more conservative in varying from that policy for the same reasons. 
However, we still believe it prudent to place certain parameters around 
the beneficiaries who may be identified as potentially at-risk by 
sponsors for their drug management programs, particularly as we gain

[[Page 16449]]

experience with the use of lock-in as a drug management tool.
    Given that no other commenter recommended a specific program size, 
there is no discernible consensus that a population of more than 
300,000 would be manageable for Part D sponsors. We therefore decline 
to adopt these criteria as the clinical guidelines for that reason, and 
also because we want sponsors to focus on the Part D population that is 
at the highest risk. Also, as we noted previously, the statute requires 
us to establish policies to ensure that the populations of enrollees in 
a prescription drug management program can be effectively managed by 
plans. Therefore, we disagree that the clinical guidelines arbitrarily 
limit the size of these programs.
    After publication of the proposed rule, we conducted an analysis of 
the clinical guidelines/OMS criteria for 2019 that we proposed using 
2017 PDE data, as the original estimates were based on 2015 data. We 
were pleased to confirm that the current policy, which will be 
integrated into Part D drug management programs, continues to make 
substantial progress in reducing potential opioid overutilization in 
the Part D program. The reduction in the number of beneficiaries 
meeting the OMS criteria between 2015 and 2017 far outpaced previous 
trends. We thank the Part D sponsors that have executed the current 
policy, the providers who have participated, and the various 
stakeholders who have provided helpful input over the years.
    According to this analysis, the 2019 clinical guidelines/OMS 
criteria we proposed would identify an estimated 11,753 potential at-
risk beneficiaries rather than the 33,053 we originally estimated. 
Given the incremental approach we have taken with the current policy 
over the years since its inception, this revised estimate provides an 
opportunity to adjust the clinical guidelines/OMS criteria downward in 
terms of prescriber and pharmacy thresholds which will incorporate more 
potential at-risk beneficiaries in 2019.
    Therefore, after considering the comments and this updated data, we 
are doing two things with respect to our clinical guidelines proposal, 
which we will identify a similar program size as the one we proposed, 
as well as strike a balance between those commenters wanting complete 
flexibility to adopt criteria to identify potential at-risk 
beneficiaries and those urging no flexibility. First, we are finalizing 
alternative criteria that we considered in the RIA as Option 3 as 
minimum criteria. These minimum criteria are: Use of opioids with an 
average daily MME greater than or equal to 90 mg for any duration 
during the most recent 6 months and either: 3 or more opioid 
prescribers and 3 or more opioid dispensing pharmacies OR 5 or more 
opioid prescribers, regardless of the number of opioid dispensing 
pharmacies.
    This means that beneficiaries meeting these criteria will be 
reported to sponsors by OMS and sponsors with drug management programs 
must review each case and report their findings back to OMS as they do 
today consistent with how they have operated under the current policy. 
In addition, sponsors may not vary these minimum criteria. However, as 
we previously stated, sponsors will be permitted to apply the minimum 
criteria more frequently using their own prescription claims data than 
CMS will apply them through OMS quarterly. According to our analysis of 
2017 PDE data, these minimum criteria would identify 44,332 potential 
at-risk beneficiaries and is the option based on 90 MME in the RIA that 
has a revised program size estimate which is closest to our original 
estimate of 33,053 but that would not identify fewer at-risk 
beneficiaries. Given the scope of the opioid crisis, and current data 
showing significant reduction in the number of beneficiaries meeting 
the OMS criteria, finalizing criteria that would have resulted in a 
smaller program size could undermine the increasing momentum in 
addressing opioid overutilization in the Medicare Part D program.
    Second, we are finalizing supplemental criteria to provide sponsors 
with some flexibility in adopting criteria for their drug management 
programs. This means that sponsors may continue to report additional 
beneficiaries to OMS--as they do today under the current policy. 
However, unlike the current policy, such beneficiaries must meet the 
following supplemental criteria: Use of opioids (regardless of average 
daily MME) during the most recent 6 months with 7 or more opioid 
prescribers OR 7 or more opioid dispensing pharmacies.
    These supplemental criteria were included in the additional 
criteria options that we considered and are included in a options chart 
in the Regulatory Impact Analysis (RIA) of the proposed rule; 
specifically, in Row 2 of option 6. Using 2017 data, we estimate that 
these supplemental criteria would identify an additional 22,841 
potential at-risk beneficiaries. We believe these criteria would be 
responsive to the concern of the commenters who, in urging us to allow 
flexibility for sponsors to adopt targeting criteria, expressed 
concerns about not being able to continue to address plan members who 
are receiving opioids from a large number of prescribers or pharmacies 
but who do not meet a particular MME threshold.
    We note that we do not anticipate that OMS will report 
beneficiaries meeting these supplemental criteria to sponsors; however, 
Part D sponsors may review beneficiaries who meet them--and must report 
them to OMS if they do--at a level that is manageable for their drug 
management programs in conjunction with the potential at-risk 
beneficiaries reported by OMS minimum criteria, whom they must address.
    Thus, the final clinical guidelines for 2019 will result in an 
estimated program size of approximately 67,173 beneficiaries--44,332 of 
whom Part D sponsors with drug management programs must review and 
22,841 of whom such sponsors may review. We believe this program size 
can be effectively managed by plans because we have already received 
feedback from Part D sponsors through the final 2018 Medicare Parts C&D 
Call Letter process that 33,000 beneficiaries are manageable. Thus, we 
conclude that 44,332 beneficiaries are associated with the option 
included in the RIA of the proposed rule that is the closest in number 
without identifying fewer potential at-risk beneficiaries and is 
consistent with historical program size under the current policy. 
Moreover, we received no comments that 33,053 beneficiaries is the 
largest program size Part D sponsors can manage. Finally, as we stated 
above, sponsors may review the additional 22,841 beneficiaries at a 
level that is manageable for their drug management programs.
    These final criteria for 2019 meet the definition of clinical 
guidelines that we are finalizing. They are criteria to identify 
potential at-risk beneficiaries who may be determined to be at-risk 
beneficiaries under drug management programs, and they were developed 
in accordance with the standards we are finalizing in Sec.  
423.153(f)(16) and beginning for 2020, will be published in guidance 
annually. These criteria also adhere to the standards we proposed in 
Sec.  423.153(f)(16) because: (1) They were developed with stakeholder 
consultation in that we solicited comment on them in the proposed rule; 
(2) they are based on the acquisition of frequently abused drugs from 
multiple prescribers, multiple pharmacies, and the level of frequently 
abused drugs in that they identify potential at-risk beneficiaries 
taking opioids and obtaining them from 7 or more prescribers or 7 or 
more pharmacies; (3)

[[Page 16450]]

they are derived from our and commenters' expert opinion that obtaining 
opioids from many prescribers or many pharmacies is a potentially 
dangerous utilization pattern of frequently abused drugs due to an 
apparent lack of coordination of care that warrants further review and 
this opinion is supported by the fact that this pattern is highly 
unusual in the Part D program as it represents 0.11 percent of 
beneficiaries; and (4) they include a program size estimate.
    We have consolidated the clinical guidelines/OMS criteria in Table 
1 for easier reference. We note that we were not persuaded by the 
commenter who urged us to adopt criteria that would address high opioid 
use regardless of the number of prescribers or pharmacies, as one 
purpose of drug management programs, and lock-in tools specifically, is 
to promote better care coordination among multiple providers.
    Comment: Some commenters suggested that if we have concerns with 
allowing Part D sponsors flexibility in adopting targeting criteria for 
potential at-risk beneficiaries, that we establish a process through 
which a sponsor could submit their guidelines to CMS.
    Response: We thank these commenters for their idea, but we prefer 
the approach we have taken as providing consistency across the entire 
Part D program and a program size, as required by CARA.
    Comment: A few commenters urged caution in the use of policies 
determining access to medications based upon thresholds such as MME, 
which the commenters viewed as a potentially problematic type of one-
size-fits all approach. These commenters noted that scientific 
literature does not support the establishment of a recommended maximum 
dose for opioids. These commenters also pointed out that the use of 
such thresholds may result in a false impression of a superior safety 
profile, which we interpreted to mean that referring to a specific MME 
level as potentially dangerous may give the impression that a level 
below that amount is universally safe.
    Response: We agree with the commenter that the CDC Guideline--and 
our clinical guidelines for Part D drug management programs that refer 
to it--are not intended as a maximum threshold for prescribing, as we 
noted in the preamble to the proposed rule. In the absence of dosing 
limits in the FDA-approved labeling for opioids, we are using the CDC 
guideline to establish a threshold to identify potentially high-risk 
beneficiaries who may benefit from closer monitoring and to create 
alignment between Government programs.
    Moreover, our implementation of the CARA drug management program 
provisions focuses on beneficiaries who are receiving opioids from 
multiple prescribers and/or multiple pharmacies, not just at a certain 
MME level. In addition, our finalized requirements for drug management 
programs require Part D sponsors to engage in case management with 
prescribers, obtain their verification that the beneficiary is at-risk 
and their agreement before implementing a prescriber lock-in or 
beneficiary-specific claim edit, as long as the prescribers are 
responsive to case management. This means that decisions about the 
amount of frequently abused drugs an at-risk beneficiary should receive 
are made by the beneficiary's prescriber(s) if they are responsive and 
not based on the targeting threshold for review of the beneficiary's 
utilization. Thus, this approach is aimed at addressing overutilization 
of frequently abused drugs while maintaining access to such drugs when 
medically necessary in the Part D program.
    Comment: A commenter proposed modifying ``for any duration'' in the 
clinical guidelines to permit beneficiaries a reasonable overlap time 
to refill medications and suggested that CMS set a reasonable overlap 
period of no more than 3 days for the purposes of identifying potential 
at-risk beneficiaries.
    Response: CMS performed an extensive analysis of the OMS criteria 
using 2015 data (https://www.cms.gov/Medicare/Prescription-Drug-Coverage/PrescriptionDrugCovContra/Downloads/Revised-OMS-Criteria-Modification-Analysis.pdf). Adjusting the clinical guideline MME 
calculation for each beneficiary to account for overlapping fills would 
be difficult to operationalize from a data analysis perspective since 
it would be dependent on the number of fills and the opioids dispensed, 
including strength each beneficiary received. For this reason, CMS 
chose to calculate the MME daily dose using the average daily dose 
during the opioid usage. We included ``for any duration'' in the 
clinical guidelines since this means that these beneficiaries reached 
or exceeded the MME level in a short period of time, and received their 
opioids from multiple prescribers and pharmacies. This indicates 
potential coordination of care issues or misuse. We found that the 
number of additional overutilizers with an episode length less than 90 
days for any of the MME dose thresholds analyzed ranged from only 57 to 
320 beneficiaries, or 1 to 2 percent of the 90+ day episode opioid 
overutilizer count. Therefore, we included these beneficiaries as 
potential opioid overutilizers under the current policy, and we will 
continue to utilize this methodology for OMS reporting of potential at-
risk beneficiaries for drug management programs.
    If a sponsor performs case management for a potential at-risk 
beneficiary who was reported through OMS and discovers that the high 
use was a result of appropriate prescription overlap and not misuse, we 
would expect the sponsor to stop conducting case management for that 
beneficiary, and to not send the initial notice to the beneficiary.
    Comment: A commenter requested that CMS clarify that the language 
``for any duration during the most recent 6 months'' means that the 
opioid use occurred during the most recent 6 months and not 6 months of 
consistent use.
    Response: We confirm that this language means that the opioid use 
occurred during the most recent 6 months.
    Comment: A commenter suggested that CMS apply path analysis to 
develop clinical guidelines to identify potential at-risk beneficiaries 
using the Integrated Data Repository (IDR), which is a data warehouse 
that integrates multiple data sources and supports analytics across 
CMS.
    Response: We thank the commenter for suggesting an approach in the 
IDR to improve identification of potential at-risk beneficiaries for 
CMS to consider.
    We proposed that under the clinical guidelines, prescribers 
associated with the same single Tax Identification Number (TIN) be 
counted as a single prescriber, because we have found under the current 
policy that such prescribers are typically in the same group practice 
that is coordinating the care of the patients served by it, and failing 
to do so would result in a high volume of false positives reported 
through OMS. Thus, it is appropriate to count such prescribers as one, 
so as not to identify beneficiaries through OMS who are not potentially 
at-risk.
    In this regard, in applying the clinical guidelines criteria, CMS 
proposed to count prescribers with the same TIN as one prescriber, 
unless any of the prescribers are associated with multiple TINs. We 
also proposed that when a pharmacy has multiple locations that share 
real-time electronic data, all locations of the pharmacy collectively 
be treated as one pharmacy under the clinical guidelines. For example, 
under the criteria we are finalizing, a beneficiary who meets the 90 
MME criterion and received opioid

[[Page 16451]]

prescriptions from 3 prescribers in the same group practice and 2 
independent opioid prescribers (1 group practice + 2 prescribers = 3 
prescribers) and filled the prescriptions at 4 opioid dispensing 
pharmacies that do not share real-time electronic data, will still meet 
the criteria, which is appropriate. However, a beneficiary who meets 
that 90 MME criterion and received opioid prescriptions from 3 
prescribers in the same group practice and 1 independent opioid 
prescriber (1 group practice + 1 prescriber = 2 prescribers) and filled 
the prescriptions at 4 opioid dispensing pharmacies that do not share 
real-time electronic data will not meet the criteria.
    Comment: Several commenters supported the proposal conceptually to 
count prescribers associated with the same single TIN as a single 
prescriber, but many of these commenters noted that some Part D plans 
sponsors and PBMs do not have access to prescriber TIN information. A 
few commenters recommended that CMS count prescribers with the same 
National Provider Identifier (NPI) as a single prescriber, and a 
commenter suggested that CMS require prescribers to share real-time 
electronic data through an electronic health record (EHR).
    Response: We appreciate the support for this proposal as well as 
the information on the operational challenges. After considering these 
comments, we are finalizing this aspect of the clinical guidelines for 
2019. Part D plan sponsors without the ability to group prescribers 
using the TIN through data analysis will have to make these 
determinations during case management. If a sponsor finds that the 
multiple opioid prescribers for the beneficiary are from a single group 
practice, and therefore, the beneficiary does not meet the clinical 
guidelines, the sponsor could stop conducting case management for that 
beneficiary, and would not send the initial notice to the beneficiary. 
We will issue guidance and updated OMS technical user guides to plan 
sponsors at a later time, including data sources and standard responses 
used in OMS reporting, which may include providing such feedback to 
CMS.
    In addition, this information may be discovered after the sponsor 
provided the beneficiary the initial notice. In such an event, the 
sponsor would send the beneficiary an alternate second notice that the 
beneficiary is not at-risk. To the comments about grouping by NPI, we 
clarify that under the current policy/OMS we use the NPI to first 
identify single prescribers, and then we further group single 
prescribers with the same single TIN. We will continue this methodology 
for the clinical guidelines under the drug management program. We 
appreciate the comment regarding real-time prescriber data, but we did 
not propose such a system for Part D prescribers.
    Comment: We received several comments supporting our proposal that 
when a pharmacy has multiple locations that share real-time electronic 
data, all locations of the pharmacy collectively be treated as one 
pharmacy under the clinical guidelines. We also received many comments 
that Part D plan sponsors and their PBMs do not have the systems 
capabilities to account for pharmacies that have multiple locations 
that share real-time electronic data, in order to treat all locations 
of the pharmacy collectively as one pharmacy. We received one comment 
that they are able to, but that there are operational challenges to 
synthesizing the data to be useful for drug management programs.
    Response: As we stated in the proposed rule, section 1860D-
4(c)(5)(D) of the Act specifies that for purposes of limiting access to 
coverage of frequently abused drugs to those obtained from a selected 
pharmacy, if the pharmacy has multiple locations that share real-time 
electronic data, all such locations of the pharmacy collectively are 
treated as one pharmacy. Because of this statutory requirement, it 
makes sense to us to consider such multiple locations as one pharmacy 
for purposes of the clinical guidelines, similar to how we account for 
group practices, to reduce false positives, particularly because the 
purpose of the guidelines is to identify when a beneficiary may be at 
risk for overutilization because they use multiple pharmacies. 
Therefore, we are finalizing this aspect of the clinical guidelines for 
2019.
    We understand that we, and apparently most sponsors and their PBMs, 
do not have the systems capability to automatically determine when a 
pharmacy is part of a chain. Therefore, Part D plan sponsors without 
this capability will have to make these determinations during case 
management. If through such case management, a plan sponsor finds that 
multiple locations of a pharmacy used by the beneficiary share real-
time electronic data, the sponsor will be required to treat those 
locations as one pharmacy. This may result in the sponsor not or no 
longer conducting case management for a beneficiary because the 
beneficiary does not meet the clinical guidelines, or in the sponsor 
sending the beneficiary an alternate second notice that the beneficiary 
is not at-risk if the sponsor discovers this information after it 
provided the beneficiary with the initial notice.
    We note that group practices and chain pharmacies are discussed 
later in this preamble in the context of the selection of a 
prescriber(s) and pharmacy(ies) in cases when a Part D plan limits a 
beneficiary's access to coverage of frequently abused drugs to selected 
pharmacy(ies) and/or prescriber(s).
    As noted above, Table 1 shows that in 2017 approximately 44,332 
beneficiaries would have met the minimum criteria of the 2019 clinical 
guidelines that we are finalizing, which is approximately 0.10 percent 
of the 45 million beneficiaries enrolled in Part D in 2017. 
Approximately, 22,841 additional beneficiaries will have met the 
supplemental criteria that we are finalizing, which is approximately 
0.05 percent. To derive this estimated population of potential at-risk 
beneficiaries, we analyzed prescription drug event data (PDE) from 
2017,\8\ using the CDC opioid drug list and MME conversion factors, and 
applying the criteria we are finalizing as the clinical guidelines. 
This estimate is over-inclusive because we did not exclude 
beneficiaries in long-term care (LTC) facilities who will be exempted 
from drug management programs, as we discuss later in this section.
---------------------------------------------------------------------------

    \8\ Unique count of beneficiaries who met the criteria in any 6 
month measurement period (January 2017-June 2017; April 2017-
September 2017; or July 2017-December 2017).
---------------------------------------------------------------------------

    However, based on similar analyses we have conducted, this 
exclusion will not result in a noteworthy reduction to our estimate. 
Also, we were unable to count all locations of a pharmacy that has 
multiple locations that share real-time electronic data as one, which 
is a topic we discussed earlier and will return to later. Thus, there 
likely are beneficiaries counted in our estimate who will not be 
identified as potential at-risk beneficiaries because they are in an 
LTC facility or only use multiple locations of a retail chain pharmacy 
that share real-time electronic data.

[[Page 16452]]

[GRAPHIC] [TIFF OMITTED] TR16AP18.000

    As clarified above, since the CDC removed all formulations of 
buprenorphine, for pain and for MAT, from the most recent CDC MME 
conversion factor file, buprenorphine products are not used to 
determine the beneficiary's average daily MME. However, we will use 
prescription opioids, including all buprenorphine products for pain and 
MAT, to determine opioid prescribers and opioid dispensing pharmacies 
under the minimum criteria. Similarly, sponsors must include all 
prescription opioids, including all buprenorphine products, to 
determine opioid prescribers and opioid dispensing pharmacies under the 
supplemental criteria.
 Exempted Beneficiary
    We proposed that an exempted beneficiary, with respect to a drug 
management program, would mean an enrollee who: (1) Has elected to 
receive hospice care; (2) Is a resident of a long-term care facility, 
of a facility described in section 1905(d) of the Act, or of another 
facility for which frequently abused drugs are dispensed for residents 
through a contract with a single pharmacy; or (3) Has a cancer 
diagnosis. While the first two exceptions are required under CARA, we 
proposed to exercise the authority in section 1860D-4(c)(5)(C)(ii)(III) 
of the Act to treat a beneficiary who has a cancer diagnosis as an 
exempted individual. We did not propose to exempt additional categories 
of beneficiaries.
    We received the following comments and our response follows:
    Comment: Commenters were overall supportive of our proposal to 
exempt beneficiaries who have a cancer diagnosis. A few of the 
commenters noted that the CDC Guideline recommendations do not apply to 
active cancer treatment. Many of these commenters asked for more 
guidance on how this exemption, which is a feature of the current 
policy, would be operationalized. Others felt the exemption is too 
broad and could be applied to beneficiaries who have not been treated 
for cancer in years or who are being treated for non-terminal cancer 
but possibly do have an opioid overuse issue that needs to be 
addressed. A few commenters disagreed with the exemption as an 
inappropriate

[[Page 16453]]

one-size-fits-all approach. Even the commenters who did not support the 
exemption noted that the cancer population is unique and must be 
handled delicately.
    Response: We thank the commenters for their supportive comments as 
to the exemption for cancer. Our intent is to exempt beneficiaries who 
are currently being treated for active cancer-related pain from Part D 
drug management programs and this is the exemption we are finalizing 
based on the comments. While our current policy generally excludes 
beneficiaries with cancer diagnoses from OMS reporting,\9\ we believe 
it is appropriate to be more specific with respect to regulatory 
parameters for Part D prescription drug management programs. Therefore, 
the comments have persuaded us that we need to be more precise with 
this codified exemption.
---------------------------------------------------------------------------

    \9\ Currently, for OMS, the following beneficiaries are excluded 
from OMS reporting: Those with ICD-10-CM codes associated with 
American Medical Association (AMA) Physician Consortium for 
Performance Improvement (PCPI) ICD-10 cancer diagnoses in the Common 
Working File (CWF) data during the 12 months prior to the end of the 
measurement period or cancer RxHCCs in the latest Risk Adjustment 
Processing System (RAPS). Note, this is currently aligned with the 
Pharmacy Quality Alliance opioid overuse measure specifications.
---------------------------------------------------------------------------

    As we noted in the proposed rule, there are some limitations around 
this exemption under the current policy due to our current data sources 
which will remain when implementing the drug management program 
clinical guidelines. For example, there may be a lag in current year 
diagnosis data in CMS systems and the RxHCC codes from the risk 
adjustment processing system are based on diagnosis data from the past 
year. Therefore, Part D plan sponsors will have to identify such 
exempted beneficiaries through the case management process if they are 
inadvertently reported through OMS or when the sponsor is reviewing 
cases pursuant to applying the minimum clinical guidelines more 
frequently than CMS and the supplemental criteria of the clinical 
guidelines. Plan sponsors may have more recent cancer diagnosis 
information or learn this information through clinical contact with 
prescribers. Plan sponsors may currently refer to the CDC Guideline as 
a reference which distinguishes active cancer treatment from cancer 
survivors with chronic pain who have completed cancer treatment, are in 
clinical remission, or are under cancer surveillance only. We will 
monitor health care guidelines that address this topic and issue 
guidance as warranted to further refine the execution of the exemption 
for beneficiaries being treated for active cancer-related pain that we 
are finalizing.
    While we understand the concerns of the commenters who did not 
support this exemption about potential inappropriate opioid use among 
this population, we note that this exemption is a feature of the 
current policy, which has reportedly been working well and we therefore 
believe it is appropriate to extend it to drug management programs. We 
agree that this population deserves heightened protection but we are 
finalizing an exemption that we believe is narrowly tailored to address 
the concerns of commenters who urged us to proceed with caution with 
respect to this exemption.
    Comment: Many commenters supported the exemption for beneficiaries 
in the LTC setting. A few commenters recommended that we not exempt LTC 
beneficiaries from retrospective drug utilization review (DUR) 
processes. A commenter asked if it could still implement a beneficiary-
specific claim edit at POS for frequently abused drugs if it 
independently determined an LTC resident to be at-risk.
    Response: Section 1860D-4(c)(5)(C)(ii) exempts beneficiaries in the 
LTC setting, and we therefore do not have the authority to permit plans 
to include them in Part D drug management programs. We are finalizing 
this exemption as proposed. Because beneficiary-specific POS claim 
edits for frequently abused drugs are included in drug management 
programs through the integration approach we are finalizing, a sponsor 
may not implement such an edit for an exempt beneficiary.
    However, while exempt beneficiaries are exempt from drug management 
programs, they are not exempt from retrospective DUR processes. Part D 
plan sponsors still must comply with its other utilization management 
obligations in Sec.  423.153, and could implement a beneficiary-
specific edit for drugs other than frequently abused drugs, for 
example, if necessary to comply with those obligations. In addition, 
sponsors may also still review the use of drugs that constitute 
frequently abused drugs by beneficiaries in LTC facilities and work 
with such facilities to identify patterns of inappropriate or medically 
unnecessary care among enrollees. However, as just stated, the sponsors 
cannot implement beneficiary-specific edits for drugs that constitute 
frequently abused drugs, nor prescriber or pharmacy lock-in for such 
drugs.
    Comment: A commenter requested that CMS exempt any Part D claim 
submitted by a Network Long-Term Care Pharmacy (NLTCP), as defined in 
Chapter 5 of the Medicare Prescription Drug Benefit Manual, asserting 
that such pharmacies are required to meet minimum performance and 
service criteria, including performing drug utilization reviews and 
identifying inappropriate drug usage. Another asked for clarification 
on whether beneficiaries serviced by long-term care pharmacies are 
exempt or if the exemption is limited to beneficiaries in long-term 
care facilities.
    Response: Section 1860D-4(c)(5)(C)(ii) of the Act exempts residents 
of a long-term care facility rather than pharmacy claims submitted by 
long-term care pharmacies. Therefore, we find it is appropriate to 
finalize an exemption that takes the same approach as the statute. 
However, we note that beneficiaries serviced by long-term care 
pharmacies may meet another exemption, such as the one for 
beneficiaries residing in facilities for which frequently abused drugs 
are dispensed for residents through a contract with a single pharmacy.
    Comment: A few commenters stated that they will need the Long-Term 
Institution (LTI) report to be released on a monthly basis rather than 
the current quarterly basis.
    Response: We thank the commenters for their comment and will 
explore if more frequent reporting is feasible.
    Comment: Many commenters supported the proposed exemption for 
beneficiaries who are residents of a facility for which frequently 
abused drugs are dispensed for residents through a contract with a 
single pharmacy. Others urged us to propose one.
    Response: We clarify for commenters that the proposed rule included 
an exemption for beneficiaries who are residents of a facility for 
which frequently abused drugs are dispensed for residents through a 
contract with a single pharmacy, as required by Section 1860D-
4(c)(5)(C)(ii). Therefore, we are finalizing this exemption as 
proposed.
    Comment: Many commenters urged us to extend an exemption to 
beneficiaries in assisted living facilities, asserting that such 
beneficiaries are at very low risk of substance abuse and that applying 
lock-in to them could be disruptive and undermine their care. Other 
commenters opposed such an exemption and urged us to proceed with 
caution in carving out multiple exemptions that could undermine the 
purpose of drug management programs. Other commenters referred to the 
difficulty in identifying such beneficiaries to exempt them.

[[Page 16454]]

    Response: Based on the comments received, we are not persuaded that 
beneficiaries in assisted living facilities should be exempt from Part 
D drug management programs, because we do not believe that these 
facilities routinely dispense drugs to their residents through a 
contract with a single pharmacy, and therefore these beneficiaries 
could be identified by the clinical guidelines on this or another basis 
and be potentially at-risk. However, if a sponsor learned during case 
management that a beneficiary resides in an assisted living facility 
that does dispense drugs through a contract with a single pharmacy, 
then the sponsor must exempt such resident from its drug management 
program.
    In addition, we are persuaded that many exemptions for certain 
group of beneficiaries or ones that are crafted too broadly would risk 
undermining the purpose of drug management programs. Therefore, we 
decline to establish a separate exemption for assisted living facility 
residents. We note that several required features of Part D drug 
management programs, such as case management, multiple written 
beneficiary notices, the right to appeal and our general oversight, 
will serve as beneficiary safeguards should a Part D sponsor 
inappropriately limit a beneficiary's coverage to frequently abused 
drugs through a drug management program.
    Comment: A commenter questioned how a drug management program 
should handle at-risk beneficiaries who move in and out of an LTC 
facility.
    Response: An at-risk beneficiary who moves into an LTC facility 
becomes an individual exempted from a drug management program and a 
sponsor must remove such beneficiary from such program as soon as it 
reliably learns that the beneficiary has moved into an LTC facility, 
whether that be via the beneficiary, the facility, a pharmacy, a 
prescriber, or an internal or external report. A beneficiary who moves 
out of an LTC facility is no longer exempted unless he or she meets 
another prong of the finalized definition of exempted beneficiary.
    Comment: Several commenters suggested that an exemption for 
beneficiaries who are receiving non-hospice palliative and end-of-life 
care would be appropriate in light of the exemption for beneficiaries 
who have elected hospice care. A few of these commenters asserted that 
without an exemption in the regulation, beneficiaries could be included 
in a drug management program at a plan sponsor's discretion and 
experience restricted access to pain-control medication when they need 
them the most. Some commenters noted that the CDC Guideline exempts 
patients receiving palliative and end-of-life care. Others disagreed, 
asserting that we had put sufficient safeguards in place to protect 
such beneficiaries in drug management programs. Other commenters 
referred to the difficulty in identifying such beneficiaries in order 
to exempt them.
    Response: We are persuaded that beneficiaries who are receiving 
non-hospice palliative and end-of-life care but have not elected 
hospice should be exempted from Part D drug management programs. While 
we wish to exercise caution and thoughtfulness in establishing 
regulatory exemptions versus clinical guidelines/criteria, as we noted 
above, we agree based on the multiple comments that such beneficiaries 
should be treated the same as beneficiaries who have elected hospice 
care for purposes of drug management programs, as they are very similar 
in their health care status, if not their health benefit plan status. 
While we expect that Part D plan sponsors and PBMs would not 
inappropriately place such beneficiaries in their drug management 
programs, an actual regulatory exemption from drug management programs 
would be more definitive. Furthermore, adding these exemptions would 
align the drug management programs with the CDC Guideline, which was 
developed by experts and specifically provides recommendations for 
primary care clinicians who are prescribing opioids for chronic pain 
outside of active cancer treatment, palliative care, and end-of-life 
care. Therefore for consistency with the CDC Guideline, beneficiaries 
who are receiving non-hospice palliative and end-of-life care but who 
have not elected hospice will be exempted from Part D drug management 
programs as well.
    As discussed in the proposed rule, the data challenges to identify 
these Part D beneficiaries will still exist for CMS and we anticipate 
for Part D sponsors also. Therefore, we will explore options for 
refining OMS reporting in this regard, and sponsors will have to 
identify these exempted beneficiaries through the case management 
process.
    We also remind Part D sponsors that drugs and biologicals covered 
under the Medicare Part A per-diem payments to a Medicare hospice 
program are excluded from coverage under Part D. For a prescription 
drug to be covered under Part D for a beneficiary who has elected 
hospice, the drug must be for treatment unrelated to the terminal 
illness or related conditions. This is because drugs and biologicals 
covered under the Medicare Part A per-diem payments to a Medicare 
hospice program are excluded from coverage under Part D. Therefore, in 
2014,\10\ we strongly encouraged sponsors to place beneficiary-level PA 
requirements on only four categories of prescription drugs including 
analgesics. As a result, a small number of beneficiaries who elected 
hospice care have been identified and excluded from the current policy/
OMS.
---------------------------------------------------------------------------

    \10\ Please see the most recent CMS guidance, ``Update on Part D 
Payment Responsibility for Drugs for Beneficiaries Enrolled in 
Medicare Hospice'', issued on November 15, 2016.
---------------------------------------------------------------------------

    Comment: A few commenters requested clarification on the practical 
meaning of an exempted individual. Specifically, they asked if the 
beneficiary is exempted from only coverage limitations or from 
retrospective DUR processes. A commenter opposed our proposal that drug 
management programs would supersede the current policy in that 
beneficiary-specific edits would no longer be permitted on non-opioid 
medications. Another commenter requested clarification on the status of 
existing beneficiary-specific POS claim edits for opioids and 
benzodiazepines beginning January 1, 2019.
    Response: Exempted beneficiaries are exempted from Part D drug 
management programs. Also, because we are integrating the ``lock-in'' 
component of the drug management programs with the current policy, 
going forward, beneficiary-specific POS edits and lock-in for 
frequently abused drugs will be permitted only in compliance with Sec.  
423.153(f). However, as we noted earlier, the prescription drug 
management program requirements that we are finalizing in this rule do 
not affect plan sponsors' obligation to comply with other requirements 
pertaining to coverage or utilization management. Part D plan sponsors 
are still obligated to conduct other drug utilization review and 
management consistent with existing DUR requirements, which includes 
reviewing utilization for any Part D drug and may include implementing 
beneficiary-specific POS claim edits on drugs that are not frequently 
abused drugs, if necessary. However, we do not have specific guidance 
in this area, but we would expect the sponsor to employ the same level 
of diligence and documentation with respect to beneficiary-level POS 
claim edits for non-frequently abused drugs that we

[[Page 16455]]

require for drug management programs, consistent with current 
policy.\11\
---------------------------------------------------------------------------

    \11\ See ``Supplemental Guidance Related to Improving Drug 
Utilization Review Controls in Part D,'' dated September 6, 2012.
---------------------------------------------------------------------------

    In addition, beneficiaries for whom Part D sponsors have 
implemented beneficiary-specific POS claim edits for opioids and/or 
benzodiazepines before January 1, 2019 can continue to be subject to 
those edits under the current policy after December 31, 2018, which 
means that they may remain in place unless removed under the current 
policy. For example, as the result of a coverage determination or 
appeal.\12\ To the extent that such a beneficiary is reported through 
OMS on January 1, 2019 or later to a sponsor with a drug management 
program, that sponsor must comply with the requirements we are 
finalizing in this rule.
---------------------------------------------------------------------------

    \12\ Patient Safety Analysis Overutilization Monitoring System 
User Guide. January 2018.
---------------------------------------------------------------------------

    Comment: A commenter suggested that CMS develop a process by which 
additional categories of exempted individuals could be evaluated and 
added that are evidence-based and involve health care practitioners.
    Response: We will evaluate the implementation of the drug 
management programs. Based on this experience or new or emerging 
relevant health care information, we will consider proposing additional 
exemptions through rulemaking as necessary.
    Comment: A commenter asked how to handle retroactive notifications 
that would qualify a beneficiary for an exemption.
    Response: As we stated in a previous response with regard to 
beneficiaries who move into LTC facilities, a sponsor must remove an 
exempted beneficiary from a drug management program as soon as it 
reliably learns that the beneficiary is exempt, whether that be via the 
beneficiary, the facility, a pharmacy, a prescriber, or an internal or 
external report.
    Based on these comments, we are finalizing with modification the 
following definition for exempted beneficiary: An exempted beneficiary, 
with respect to a drug management program, will mean an enrollee who: 
(1) Has elected to receive hospice care or is receiving palliative or 
end-of-life care; (2) is a resident of a long-term care facility, of a 
facility described in section 1905(d) of the Act, or of another 
facility for which frequently abused drugs are dispensed for residents 
through a contract with a single pharmacy; or (3) is being treated for 
active cancer-related pain. Given this exemption, CMS will report 
potential at-risk beneficiaries who meet the minimum criteria of the 
clinical guidelines to sponsors through the OMS. Currently, we have the 
ability to exempt beneficiaries in LTC facilities, in hospice, and with 
active cancer-related pain. Sponsors may have more current data or 
obtain information through the case management and notification 
processes to further exempt beneficiaries, including those receiving 
palliative or end-of-life care.
(ii) Requirements of Drug Management Programs (Sec. Sec.  423.153, 
423.153(f))
    As noted previously, we proposed to codify a regulatory framework 
under which Part D plan sponsors may adopt drug management programs to 
address overutilization of frequently abused drugs. Therefore, we 
proposed to amend Sec.  423.153(a) by adding this sentence at the end: 
``A Part D plan sponsor may establish a drug management program for at-
risk beneficiaries enrolled in their prescription drug benefit plans to 
address overutilization of frequently abused drugs, as described in 
paragraph (f) of this section,'' in accordance with our authority under 
revised section 1860D-4(c)(5)(A) of the Act.
    We also proposed to revise Sec.  423.153 by adding a new paragraph 
(f) about drug management programs for which the introductory sentence 
will read: ``(f) Drug Management Programs. A drug management program 
must meet all the following requirements.'' Thus, the requirements that 
a Part D plan sponsor must meet to operate a drug management program 
will be codified in various provisions under Sec.  423.153(f).
    We received the following comments and our response follows:
    Comment: While CMS received many comments that were supportive of 
drug management programs as a whole, we did not receive comments 
specific to these provisions.
    Response: We are therefore finalizing as proposed.
(iii) Written Policies & Procedures (Sec.  423.153(f)(1))
    We proposed to require Part D sponsors document their programs in 
written policies and procedures that are approved by the applicable P&T 
committee and reviewed and updated as appropriate, which is consistent 
with the current policy. Also consistent with the current policy, we 
proposed to require that these policies and procedures address the 
appropriate credentials of the personnel conducting case management and 
the necessary and appropriate contents of files for case management. We 
additionally proposed to require sponsors to monitor information about 
incoming enrollees who will meet the definition of a potential at-risk 
and an at-risk beneficiary in proposed Sec.  423.100 and respond to 
requests from other sponsors for information about potential at-risk 
and at-risk beneficiaries who recently disenrolled from the sponsor's 
prescription drug benefit plans.
    To codify these requirements, we proposed the written policies and 
procedures specified at Sec.  423.153(f)(1) (see 82 FR 56510).
    We received the following comments and our response follows:
    Comment: We received a comment strongly supportive of the 
requirements in this provision.
    Response: We thank the commenter for the support.
    Comment: We received a few comments inquiring what credentials are 
needed for clinical staff who conduct case management. The commenters 
were concerned that the clinical staff conducting case management be 
adequately qualified to perform it in terms of education and training. 
These commenters stated that unqualified case managers could 
significantly detract from the benefit of Part D drug management 
programs.
    Response: We agree that the requirement that clinical staff conduct 
case management needs more detail. CMS expects that such clinical staff 
conducting case management as part of a Part D plan sponsor's drug 
management program would be a physician or other appropriate health 
care professional with sufficient expertise to conduct medical 
necessity reviews related to potential opioid overutilization. While we 
are not specifying particular credentials for clinical staff, in 
response to these comments, we are clarifying in the finalized version 
of Sec.  423.153(f)(1)(i) that clinical staff must have a current and 
unrestricted license to practice within the scope of his or her 
profession in a State, Territory, Commonwealth of the United States 
(that is, Puerto Rico), or the District of Columbia.
    Comment: We received several comments that a dentist should be 
required to be included on the case management team when a prescriber 
of frequently abused drugs is a dentist.
    Response: We decline to adopt this recommendation. We do not want 
to be overly prescriptive as to the specific background of licensed 
clinical staff conducting case management. We believe the plan should 
have some flexibility, beyond what is discussed in the preceding 
response and described in Sec.  423.153(f)(1)(i), to determine 
appropriate credentials of the clinical

[[Page 16456]]

staff conducting case management based on the facts and circumstances 
of the case.
    Comment: We received a question asking how prescriber agreement 
should be documented and shared with appropriate parties. We also 
received a few comments that a Part D sponsor must ensure that any 
records of contacts between the sponsors and prescribers under drug 
management programs must be easily accessible to at-risk beneficiaries 
who wish to appeal and that these records are easily able to be auto-
forwarded to the Independent Review Entity (IRE).
    Response: We agree that such information must be documented and 
available to appropriate parties including at-risk beneficiaries and 
the IRE, when applicable. To comply with Sec.  423.153(f)(1)(ii), 
sponsors must document contact with prescribers during case management, 
for example, if a prescriber agreed with the plan sponsor to implement 
a limit on the beneficiary's access to coverage for frequently abused 
drugs pursuant to Sec.  423.153(f)(4). Also, the sponsor must document 
if the beneficiary calls the sponsor to provide his or her pharmacy or 
prescriber preferences for lock-in. To make this clearer, we are adding 
language to Sec.  423.153(f)(1)(ii) such that the necessary and 
appropriate contents of files for case management must include 
documentation of the substance of prescriber and beneficiary contacts.
    Comment: We received a comment that we should require Part D plan 
sponsors' policies and procedures for clinical contact to include 
secure identity verification safeguards to protect prescribers from 
``phishing'' communications that attempt to trick prescribers into 
disclosing patient information.
    Response: We decline to make this a requirement specific to Part D 
drug management programs. We note that health care providers' offices 
and Part D sponsors are both covered entities under Health Insurance 
Portability and Accountability Act of 1996. We also encourage Part D 
sponsors to have written policies and procedures for their staff who 
contact providers to proactively identify themselves in a manner that 
should reasonably satisfy the providers of their identity and for 
providers to likewise have written practice policies and procedures to 
reasonably establish the identity of the staff of health benefit plans 
who contact them and do not proactively establish their identity.
    Given these comments and our responses, we are finalizing Sec.  
423.153(f)(1) with modification to include the changes regarding the 
licensure of the clinical staff conducting case management and the 
required documentation of the substance of prescriber and beneficiary 
contacts.
(iv) Case Management/Clinical Contact/Prescriber Verification (Sec.  
423.153(f)(2))
    To meet the requirements of section 1860D-4(c)(5)(C) and section 
1860D-4(c)(5)(B)(i)(II) of the Act, we proposed in a new Sec.  
423.153(f)(2) to require Part D sponsors' clinical staff to engage in 
case management for each potential at-risk beneficiary for the purpose 
of engaging in clinical contact with the prescribers of frequently 
abused drugs and verifying whether a potential at-risk beneficiary is 
an at-risk beneficiary. Specifically, we proposed that a new Sec.  
423.153(f)(2) would state that the sponsor's clinical staff must 
conduct case management for each potential at-risk beneficiary for the 
purpose of engaging in clinical contact with the prescribers of 
frequently abused drugs and verifying whether a potential at-risk 
beneficiary is an at-risk beneficiary. Proposed Sec.  423.153(f)(2)(i) 
would further state that, except as provided in paragraph (f)(2)(ii) of 
this section, the sponsor must do all of the following:
     Send written information to the beneficiary's prescribers 
that the beneficiary meets the clinical guidelines and is a potential 
at-risk beneficiary;
     Elicit information from the prescribers about any factors 
in the beneficiary's treatment that are relevant to a determination 
that the beneficiary is an at-risk beneficiary, including whether 
prescribed medications are appropriate for the beneficiary's medical 
conditions or the beneficiary is an exempted beneficiary; and
     In cases where the prescribers have not responded to the 
inquiry described in (f)(2)(i)(B), make reasonable attempts to 
communicate telephonically with the prescribers within a reasonable 
period after sending the written information.
    We proposed to add paragraph (ii) to Sec.  423.153(f)(2) that would 
specify that the exception would be for identification by prior plan. 
If a beneficiary was identified as a potential at-risk or an at-risk 
beneficiary by his or her most recent prior plan, and such 
identification has not been terminated in accordance with paragraph 
(f)(14) of this section, the sponsor meets the requirements in 
paragraph (f)(2)(i) of this section, so long as the sponsor obtains 
case management information from the previous sponsor and such 
information is still clinically adequate and up to date. This proposal 
is to avoid unnecessary burden on health care providers when additional 
case management outreach is not necessary because it has already been 
performed by a prior Part D sponsors for the beneficiary. We discuss 
potential at-risk and at-risk beneficiaries who change plans again 
later in this preamble.
    The information that the plan sends to the prescribers and elicits 
from them is intended to assist a Part D sponsor to understand why the 
beneficiary meets the clinical guidelines and if a limitation on access 
to coverage for frequently abuse drugs is warranted for the safety of 
the beneficiary. Also, sponsors will use this information to choose 
standardized responses in OMS and provide information to MARx about any 
plan coverage limitations that the sponsors implement. We will address 
required reporting to OMS and MARx by sponsors again later.
    Our proposed Sec.  423.153(f)(2) used the terms ``reasonable 
attempts'' and ``reasonable period'' rather than specify a required 
number of attempts or a specific timeframe for plan sponsor to call 
prescribers. We explained that this was due to the competing priorities 
of sponsors' diligently addressing opioid overutilization in the Part D 
program through case management, which may necessitate telephone calls 
to the prescribers, while being cognizant of the need to be judicious 
in contacting prescribers telephonically in order to not unnecessarily 
disrupt their practices. We further stated that we wished to leave 
flexibility in the regulation text for sponsors to balance these 
priorities on a case-by-case basis in their drug management programs. 
However, we note that we proposed a 3 attempts/10 business days 
requirement for sponsors to conclude that a prescriber is unresponsive 
to case management in Sec.  423.153(f)(4) discussed later in this 
section.
    We received the following comments and our response follows:
    Comment: We received a comment requesting that a plan sponsor be 
able to communicate to CMS if no prescriber will verify that the 
beneficiary is at-risk.
    Response: We plan to expand and modify OMS and the MARx system to 
accommodate the CARA drug management program provisions we are 
finalizing here. We will issue additional guidance and technical 
instructions as needed.
    Comment: We received a comment asking that we recommend that Part D 
sponsors encourage prescribers during case management to discuss drug 
management programs with their patients. We also received a request 
that we issue guidance to plan sponsors directing them to encourage 
prescribers, as part of the required clinical contact,

[[Page 16457]]

to perform a comprehensive substance abuse disorder screening and/or 
assessment of the patient deemed to be a potential at-risk beneficiary, 
and if indicated, refer him or her for follow-up treatment with a pain 
specialist or addiction treatment provider.
    Response: We encourage Part D plan sponsors to undertake both of 
these suggestions, but decline to require it at this time, as we 
believe prescribers, in their professional discretion by and large will 
undertake appropriate adjusted treatment plans with their patients and/
or MA-PDs will negotiate such issues with their network providers. We 
also remind commenters that not all Part D prescription drug plans have 
network providers.
    Comment: We received some comments that Part D sponsors should not 
be permitted to telephone prescribers in order to avoid disrupting 
their practices.
    Response: We decline to adopt this suggestion. The clinical 
guidelines identify beneficiaries who are potentially at-risk for a 
serious adverse health event, including death, due to their opioid use 
and apparent lack of coordinated care. The requirements we are 
finalizing permit sponsors to escalate the steps they take during case 
management to engage in clinical contact with the beneficiary's 
prescribers of frequently abused drugs. We would expect such 
prescribers to understand such sponsors' attempts to make them aware of 
important information in this regard that they likely do not know.
    Comment: We received a comment that integrated delivery systems use 
communication tools other than telephone calls to escalate matters to 
prescribers and that CMS should allow such systems to use such tools 
instead.
    Response: Our intent is for Part D sponsors to use the most 
effective means designed to elicit a prescriber response to case 
management. Therefore, based on this comment, we are modifying the 
regulatory language in Sec.  423.153(f)(2)(i)(C).
    Comment: We received a question whether a gaining sponsor must 
immediately lock-in a new enrollee if the sponsor receives notice from 
the losing sponsor that the enrollee was locked-in by the losing 
sponsor.
    Response: No. Part D sponsors are responsible for their own drug 
management programs. As such, a gaining sponsor is not required to but 
may do so under certain circumstances as we discuss later in this 
preamble. Also, we note that with respect to at-risk beneficiaries that 
are new to a plan, sponsors that do not take any action should be aware 
that such beneficiaries may later be reported through OMS if they meet 
the clinical guidelines. Also, we note that pursuant to Sec.  
423.153(f)(2)(i), the sponsor must conduct case management for every 
potential at-risk beneficiary, unless an exception applies.
    After considering these comments, we are finalizing the proposed 
language in Sec.  423.153(f)(2) with the modification described.
(v) Limitations on Access to Coverage for Frequently Abused Drugs 
(Sec.  423.153(f)(3))
    We proposed to describe all the tools that will be available to 
sponsors to limit an at-risk beneficiary's access to coverage for 
frequently abused drugs under a drug management program in Sec.  
423.153(f)(3). Our proposal specified that subject to the requirements 
of paragraph (f)(4) of this section, a Part D plan sponsor may do all 
of the following:
     Implement a point-of-sale claim edit for frequently abused 
drugs that is specific to an at-risk beneficiary.
     In accordance with paragraphs (f)(10) and (f)(11) of this 
section, limit an at-risk beneficiary's access to coverage for 
frequently abused drugs to those that are--
    ++ Prescribed for the beneficiary by one or more prescribers;
    ++ Dispensed to the beneficiary by one or more network pharmacies; 
or
    ++ Specified in both paragraphs (f)(3)(ii)(B)(1) and (2) of this 
section.
    Paragraph (iii)(A) will state that if the sponsor implements an 
edit as specified in paragraph (f)(3)(i) of this section, the sponsor 
must not cover frequently abused drugs for the beneficiary in excess of 
the edit, unless the edit is terminated or revised based on a 
subsequent determination, including a successful appeal. Paragraph 
(iii)(B) will state that if the sponsor limits the at-risk 
beneficiary's access to coverage as specified in paragraph (f)(3)(ii) 
of this section, the sponsor must cover frequently abused drugs for the 
beneficiary only when they are obtained from the selected pharmacy(ies) 
and/or prescriber(s), or both, as applicable, (1) in accordance with 
all other coverage requirements of the beneficiary's prescription drug 
benefit plan, unless the limit is terminated or revised based on a 
subsequent determination, including a successful appeal, and (2) except 
as necessary to provide reasonable access in accordance with paragraph 
(f)(12) of this section.
    We received the following comments and our response follows:
    Comment: We received a question whether a Part D sponsor, under a 
drug management program, may implement a combination of a beneficiary-
specific POS claim edit, prescriber and/or pharmacy lock-in for 
frequently abused drugs, and whether these limitations may be 
implemented at different times. Another comment recommended that plan 
sponsors be permitted to establish a prescriber lock-in concurrently 
with a beneficiary-specific POS claim edit and not require the plan to 
contact the prescribers separately for each limitation.
    Response: We acknowledge that there may be cases where a plan may 
impose one or more coverage limitations for frequently abused drugs 
simultaneously on an at-risk beneficiary, and at a later time, add new 
limitations and/or terminate existing ones. Thus, a plan sponsor may 
choose to implement multiple limitations on access to coverage for 
frequently abused drugs for an at-risk beneficiary at one time.
    For instance, after case management, a plan sponsor may decide to 
pursue implementation of a POS claim edit, prescriber lock-in, and 
pharmacy lock-in for an at-risk beneficiary simultaneously because of 
the circumstances of the particular case. In this instance, prescriber 
agreement would be necessary to implement the POS edit and the 
prescriber lock-in.
    A plan sponsor may also implement additional coverage limitations 
over time (for example, start with a beneficiary-level POS edit, 
subsequently add a prescriber lock-in, and subsequently add a pharmacy 
lock-in) because the case has not resolved itself as expected after 
initial case management. We remind plan sponsors that when implementing 
additional coverage limitations, the plan sponsor must repeat the case 
management process including prescriber verification, prescriber 
agreement, if applicable, and notice requirements for each additional 
limitation, and that such actions would also confer a new 60 day appeal 
timeframe. We discuss this scenario further in the appeal section of 
this preamble.
    Furthermore, a plan sponsor might also terminate existing 
limitations on access to coverage over time (for example, an at-risk 
beneficiary may have a POS edit and pharmacy lock-in and the plan 
sponsor terminates the pharmacy lock-in and leaves in place the POS 
edit).
    While we are allowing plan sponsors to make such additions/
terminations to limitations to access to coverage for frequently abused 
drugs for an at-risk beneficiary, we recognize that such

[[Page 16458]]

changes might be disruptive and/or confusing for the beneficiary, and 
thus strongly discourage plans from making frequent changes to such 
limitations for a particular at-risk beneficiary. To minimize such 
disruption and ensure such actions are taken in the manner contemplated 
by the statute, we have added a provision at Sec.  423.153(f)(5)(iv) to 
the regulation text which specifies that, if a plan intends to make 
changes to the limitations imposed on a beneficiary under their drug 
management program after the beneficiary has been identified as at-
risk, the plan sponsor is required to provide the beneficiary notices 
under the rules established at Sec.  423.153(f)(5) through (f)(8) and 
discussed later in this preamble. Additionally, we will closely monitor 
information submitted by sponsors to CMS in OMS and MARx and complaint 
data to make sure plans are not inappropriately disrupting beneficiary 
access to coverage for frequently abused drugs by making frequent 
changes to the limitations on access to coverage. While we are not 
currently imposing limitations on how many times the plan can make such 
changes, we will re-evaluate this policy in the future if it becomes 
problematic.
    In response to this comment, we are finalizing this provision as 
proposed, except we are modifying Sec.  423.153(f)(3) to state a Part D 
plan sponsor may do ``any or all of the following,'' and Sec.  
423.153(f)(3)(ii)(C) to simply state ``both.'' This will make clearer 
that read as a whole, Sec.  423.153(f)(3) means that a Part D sponsor 
may use the tool of a beneficiary-specific point-of-sale edit, or 
prescriber or pharmacy lock-in, or any combination of these three tools 
to limit an at-risk beneficiary's access to coverage of frequently 
abused drugs under its drug management program.
(vi) Requirements for Limiting Access to Coverage for Frequently Abused 
Drugs (Sec.  423.153(f)(4))
    We proposed in Sec.  423.153(f)(4) that before a Part D plan 
sponsor could limit the access of at-risk beneficiary to coverage for 
frequently abused drugs, the sponsor would first be required to take 
certain actions. We proposed in paragraph Sec.  423.153(f)(4)(i)(A) 
that a sponsor would be required to conduct the case management 
discussed earlier, which includes clinical contact to determine whether 
prescribed medications are appropriate for the potential at-risk 
beneficiary's medical conditions that is required by section 1860D-
4(c)(5)(C)(iv) of the Act and prescriber verification that the 
beneficiary is an at-risk beneficiary in accordance with Section 1860D-
4(c)(5)(B)(i)(II).
    We also proposed in paragraph Sec.  423.153(f)(4)(i)(B) that the 
sponsor would be required to obtain the agreement of the prescribers of 
frequently abused drugs with the limitation, unless the prescribers 
were not responsive to the required case management. We invited 
stakeholders to comment on not requiring prescriber agreement to 
implement pharmacy lock-in.
    We further proposed in paragraph Sec.  423.153(f)(4)(i)(C) that the 
sponsor must first provide notices that complied with Sec.  
423.153(f)(5) and (f)(6) to the beneficiary in accordance with section 
1860D-4(c)(5)(B)(i)(I) of the Act. We additionally proposed in 
paragraph Sec.  423.153(f)(4)(ii) that a sponsor has complied with the 
requirement in Sec.  423.153(f)(2)(i)(C) to make reasonable attempts to 
communicate telephonically with prescribers with a reasonable period if 
the prescribers were not responsive after 3 attempts to contact them 
within 10 business days. Finally, we proposed language in Sec.  
423.153(f)(4)(ii) that would provide an exception to the case 
management requirement in Sec.  423.153(f)(2) in cases when a potential 
or an at-risk beneficiary was identified as such by the beneficiary's 
most recent prior prescription drug benefit plan and the sponsor had 
obtained the case management information from the sponsor and updated 
it as appropriate. We discussed such cases elsewhere in this section. 
We also discuss proposed Sec.  423.153(f)(4)(iv) that would have 
imposed a 6-month delay before a sponsor could implement prescriber 
lock-in later in this preamble.
    We received the following comments and our responses follow:
    Comment: A commenter suggested that we allow a coverage limitation 
to be put in place through a drug management program if a prescriber 
requests one to assist in coordinating the care for his or her patient.
    Response: If the beneficiary meets the clinical guidelines/OMS 
criteria we are finalizing, and a prescriber requests during case 
management that a coverage limitation be implemented for the 
beneficiary, the sponsor may implement it in accordance with the 
requirements we are finalizing for drug management programs in this 
rule.
    Comment: Many commenters stated that Part D sponsors should not 
have to seek prescriber agreement to limit at-risk beneficiaries to a 
pharmacy(ies) for access to coverage for frequently abused drugs. These 
commenters argued that requiring prescriber agreement for pharmacy 
lock-in would create additional administrative burden and 
inefficiencies and thus prevent drug management programs from 
responding in a timely fashion to potentially dangerous overutilization 
of frequently abused drugs. These commenters also argued that sponsors 
of stand-alone Part D plans do not have contracts with most of the 
prescribers and, therefore, have limited opportunity to have clinical 
contact with these prescribers. Moreover, many commenters felt it was 
not appropriate to require that the prescriber agree to pharmacy lock-
in when the pharmacy is not required to agree when a sponsor applies 
prescriber lock-in to an at-risk beneficiary.
    Other commenters supported our proposal to require prescriber 
agreement for pharmacy lock-in. These commenters argued that provider 
discretion and clinical judgment is appropriate to prevent pharmacy 
lock-in from being implemented by Part D sponsors inappropriately and 
impeding legitimate patient access.
    Response: CMS was persuaded by commenters' rationale that requiring 
prescriber agreement for pharmacy lock-in could undermine one purpose 
of drug management programs, which is to promptly address potentially 
dangerous overutilization of frequently abused drugs. While we 
recognize that prescriber agreement is an essential component of 
prescriber lock-in, and prescriber agreement is preferred in the case 
of a beneficiary-specific claim edit for frequently abused drugs, we 
are now persuaded that prescriber agreement to pharmacy lock-in is not 
essential, as pharmacy lock-in is primarily about where the drugs are 
dispensed and not who wrote the prescription or its dosage. Therefore, 
we are finalizing this provision with this modification. Plan sponsors 
will not be required to obtain the agreement of the prescribers of 
frequently abused drugs to implement a pharmacy lock-in. However, we do 
note that should a prescriber proactively alert the plan sponsor that 
they do not believe that pharmacy lock-in is appropriate for a 
particular at-risk beneficiary, we expect the plan sponsor to take such 
information into consideration.
    On the point of prescriber agreement, we also wish to note that it 
was unclear in some of the statements if the commenters understood that 
section 1860D-4(c)(5)(C)(iv) and Section 1860D-4(c)(5)(B)(i)(II) of the 
Act require, respectively, that a Part D sponsor engage in clinical 
contact with prescribers regarding whether medications are appropriate 
for a beneficiary's medical condition and to

[[Page 16459]]

verify that a beneficiary is at-risk before limiting access to coverage 
for frequently abused drugs. Thus, eliminating the need to obtain 
prescriber agreement to a pharmacy lock-in does not eliminate the 
requirement to comply with Sec.  423.153(f)(2) and (f)(4)(i)(A) with 
respect to pharmacy lock-in.
    Comment: Several commenters asked CMS to provide additional details 
about what options Part D plan sponsors would have if a prescriber does 
not agree to a pharmacy lock-in.
    Response: As mentioned above, we are not finalizing the proposal 
that sponsors must receive prescriber agreement before placing an at-
risk beneficiary in pharmacy lock-in.
    Comment: In general, commenters supported our proposal that a Part 
D sponsor would have to obtain prescriber agreement before implementing 
prescriber lock-in or a beneficiary-specific claim edit at POS for 
frequently abused drugs to limit an at-risk beneficiary's access to 
coverage for frequently abused drugs, in cases when a prescriber is 
responsive to case management. These commenters maintained that the 
prescribers are in the best position to understand the beneficiary's 
background and know additional relevant considerations.
    However, many commenters voiced their recommendation that the Part 
D sponsor be able to implement prescriber lock-in without obtaining 
agreement from all prescribers. Several commenters expressed that it 
would be difficult to get all prescribers to agree to any limitation, 
and suggested that as long as at least one prescriber of frequently 
abused drugs agreed to the limitation, sponsors should be able to 
proceed with a prescriber lock-in. Commenters suggested that plan 
sponsors will have already coordinated with the prescribers during case 
management, at which time the sponsor will have confirmed the 
appropriateness of the medication and verified with a prescriber that 
the beneficiary is at risk. Thus, these commenters further suggested 
that obtaining formal approval of the lock-in will only serve to delay 
initiating the lock-in.
    Commenters also raised the point that a given prescriber may be 
contributing to the overutilization, in which case his or her approval 
may not be obtained and requested clarification how a sponsor should 
act in a beneficiary's best interest if prescribers disagree with each 
other about the implementation of a claim edit or lock-in. Some 
commenters recommended that CMS require approval only from the primary 
prescriber of frequently abused drugs, as determined by case 
management.
    Response: We agree that in order for drug management programs to 
operate effectively, and prevent the resource-intensive process of 
obtaining agreement from multiple prescribers, a Part D sponsor should 
not have to obtain the agreement to prescriber lock-in of all the at-
risk beneficiary's prescribers of frequently abused drugs. Therefore, 
we are changing the language of Sec.  423.153(f)(4)(i)(B) to refer to 
at least one prescriber, which means that only one prescriber has to 
agree to prescriber lock-in or a beneficiary-specific POS edit.
    In addition, we believe the language of Sec.  423.153(f)(4)(ii)(B) 
needs to be clearer that prescribers must be responsive in the case of 
a prescriber lock-in, meaning that non-responsive prescribers cannot 
constitute agreement as they can in the case of a beneficiary-specific 
POS edit. Therefore, we are finalizing the Sec.  423.153(f)(4) with 
this modification in paragraph (ii)(A) and a new (B).
    Comment: We received a comment suggesting that a better approach to 
prescriber agreement would be for at-risk beneficiaries to identify a 
primary prescriber to help drug management and increase beneficiary 
safety.
    Response: As noted above, we have modified our proposal and are 
finalizing that all prescribers do not have to agree to prescriber 
lock-in in order for a plan to implement prescriber lock-in for an at-
risk beneficiary; rather, at least one prescriber has to agree. 
However, we believe that the prescriber who agrees to prescriber lock-
in for a beneficiary should be identified through the plan sponsor as a 
result of case management, and not the at-risk beneficiary. There may 
be a conflict of interest in having an at-risk beneficiary select whom 
they consider to be their ``primary'' prescriber for purposes of 
prescriber agreement, given they might be motivated to select a 
``primary'' prescriber that they feel would not agree to prescriber 
lock-in, such that they can continue receiving inappropriate amounts of 
frequently abused drugs. We reiterate that the requirement that at 
least one prescriber agree is for agreement to lock-in is different 
from the beneficiary's preferences for the prescriber to which they 
will be locked into, which we discuss later in this preamble.
    Comment: We received comments that a prescriber should be able to 
agree, disagree or neither agree nor disagree with a limitation on a 
beneficiary's access to coverage for frequently abused drugs.
    Response: A prescriber is of course free to have any of these 
reactions to case management. A plan sponsor cannot implement 
prescriber lock-in for the beneficiary, unless at least one prescriber 
agrees to prescriber lock-in, as discussed earlier. Typically, we would 
expect the one prescriber to agree to prescriber lock-in and agree to 
serve as the prescriber. A sponsor cannot lock-in a beneficiary to a 
prescriber who disagrees, unless the prescriber changes their mind, 
which must be documented in the case file.
    We foresee a situation when a prescriber initially disagrees with 
prescriber lock-in and asserts that he or she must be able to continue 
to prescribe frequently abused drugs for the beneficiary. In such a 
case, if another prescriber has agreed to serve as the prescriber to 
which the beneficiary is locked into, a plan sponsor may need to again 
ask the first prescriber if he or she would agree to be a prescriber 
the beneficiary is locked into, and the beneficiary is ultimately 
locked into two prescribers to ensure reasonable access pursuant to 
Sec.  423.153(f)(12), which we discuss further below. This could 
happen, for example, when a beneficiary has been obtaining opioids from 
multiple prescribers and benzodiazepines from one psychiatrist. A 
sponsor may have to permit an at-risk beneficiary to obtain opioids 
from the prescriber who agreed to the lock-in limitation and 
benzodiazepines from the psychiatrist, who initially did not agree to 
prescriber lock-in, but ultimately does agree to serve that beneficiary 
in a lock-in capacity.
    With respect to a beneficiary-specific POS claim edit for 
frequently abused drugs, however, a plan sponsor may not implement one 
at a dosage that is lower than the highest dosage a prescriber asserts 
is medically necessary, which is consistent with our current 
policy.\13\
---------------------------------------------------------------------------

    \13\ Supplemental Guidance Related to Improving Drug Utilization 
Review Controls in Part D, September 6, 2012.
---------------------------------------------------------------------------

    If a prescriber neither agrees nor disagrees with a limitation on 
access to coverage for frequently abused drugs, such a prescriber may 
be considered by the sponsor to be non-responsive, and an at-risk 
beneficiary could not be locked into that prescriber.
    Comment: We received a comment suggesting that 30 days be the time 
period during which a Part D sponsors must attempt to reach an 
unresponsive prescriber.
    Response: We believe 30 days is too long considering that drug 
management programs involve frequently abused drugs and multiple 
prescribers and

[[Page 16460]]

pharmacies; that the clinical guidelines identify beneficiaries who are 
at potentially at high risk for an adverse health event due to the 
amount of such drugs they are taking; and that there is an apparent 
lack of coordinated care.
    Comment: We received a comment that a sponsor should only be 
required to attempt to reach a prescriber twice in 10 business days 
rather than 3 times in order to establish that the prescriber is 
unresponsive.
    Response: We decline to make this change as this is our current 
policy and we received minimal comment on this proposed requirement. 
The purpose of the policy is to ensure that sponsors have diligently 
tried to involve prescribers in the case management process.
    We wish to note that we believe the language we proposed in Sec.  
423.153(f)(4)(iii) which provides an exception to case management is 
duplicative of the language we discussed above that we are finalizing 
in Sec.  423.153(f)(2)(ii). Therefore, we are deleting the language in 
Sec.  423.153(f)(4)(iii).
    Given the foregoing, we are finalizing Sec.  423.153(f)(4) with 
modification, including ones to assist the reader in more easily 
understanding the cross-references.
    We will also state in paragraph (ii)(A) that, except as provided in 
paragraph (ii)(B) which regards a prescriber limitation, if the sponsor 
complied with the requirement of paragraph (f)(2)(i)(C) of this section 
about attempts to reach prescribers, and the prescribers were not 
responsive after 3 attempts by the sponsor to contact them within 10 
business days, then the sponsor has met the requirement of paragraph 
(f)(4)(i)(B) of this section which regards eliciting information from 
the prescribers. Paragraph (i)(B) will state that the sponsor may not 
implement a prescriber limitation pursuant to Sec.  
423.153(f)(3)(ii)(A) if no prescriber was responsive.
(vii) Beneficiary Notices and Limitation of Special Enrollment Period 
(Sec. Sec.  423.153(f)(5), 423.153(f)(6), 423.153(f)(7), 423.153(f)(8), 
423.38)
(A) Initial Notice to Beneficiary and Sponsor Intent To Implement 
Limitation on Access to Coverage for Frequently Abused Drugs (Sec.  
423.153(f)(5))
    The notices referred to in proposed Sec.  423.153(f)(4)(i)(C) are 
the initial and second notice that section 1860D-4(c)(5)(B)(i)(I) of 
the Act requires Part D sponsors to send to potential at-risk and at-
risk beneficiaries regarding their drug management programs.
    We proposed in Sec.  423.153(f)(5) that if a Part D plan sponsor 
intends to limit the access of a potential at-risk beneficiary to 
coverage for frequently abused drugs, the sponsor will be required to 
provide an initial written notice to the potential at-risk beneficiary. 
We also proposed that the language be approved by the Secretary and be 
in a readable and understandable form that contains the language 
required by section 1860D-4(c)(5)(B)(ii) of the Act, as well as 
additional detail specified in the proposed regulation text.
    In proposed paragraph (f)(5)(ii)(C)(2)--which will require a 
description of public health resources that are designed to address 
prescription drug abuse--we proposed to require that the notice contain 
information on how to access such services. We also included a 
reference in proposed paragraph (ii)(C)(4) to the fact that a 
beneficiary will have 30 days to provide information to the sponsor, 
which is a timeframe we discuss later in this preamble. We proposed an 
additional requirement in paragraph (ii)(C)(5) that the sponsor include 
the limitation the sponsor intends to place on the beneficiary's access 
to coverage for frequently abused drugs, the timeframe for the 
sponsor's decision, and, if applicable, any limitation on the 
availability of the SEP. Finally, we proposed a requirement in 
paragraph (ii)(C)(8) that the notice contain other content that CMS 
determines is necessary for the beneficiary to understand the 
information required in the initial notice.
    We noted that our proposed implementation of the statutory 
requirements for the initial notice will permit the notice also to be 
used when the sponsor intends to implement a beneficiary-specific POS 
claim edit for frequently abused drugs.
    Although section 1860D-4(c)(5) is silent as to the sequence of the 
steps of clinical contact, prescriber verification, and the initial 
notice, we proposed to implement these requirements such that they will 
occur in the following order: first, the plan sponsor will conduct the 
case management which encompasses clinical contact and prescriber 
verification required by Sec.  423.153(f)(2) and obtain prescriber 
agreement if required by Sec.  423.153(f)(4), and subsequently, if 
applicable, the plan sponsor will provide the initial notice indicating 
the sponsor's intent to limit the beneficiary's access to frequently 
abused drugs. Further, under our proposal, although the proposed 
regulatory text of (f)(4)(i) states that the sponsor must verify with 
the prescriber(s) that the beneficiary is an at-risk beneficiary in 
accordance with the applicable statutory language, the beneficiary will 
still be a potential at-risk beneficiary from the sponsor's perspective 
when the sponsor provides the beneficiary the initial notice. This is 
because the sponsor has yet to solicit information from the beneficiary 
about his or her use of frequently abused drugs, and such information 
may have a bearing on whether a sponsor identifies a potential at-risk 
beneficiary as an at-risk beneficiary.
    Moreover, we proposed that a sponsor should not send a potential 
at-risk beneficiary an initial notice until after the sponsor has been 
in contact with the beneficiary's prescribers of frequently abused 
drugs as part of case management, so as to avoid unnecessarily alarming 
the beneficiary. This is because the result of case management may be 
that the sponsors takes a ``wait and see'' approach to observe if the 
prescribers adjust their management of, and opioid prescriptions they 
are writing for, the beneficiary. We noted that while this approach is 
acceptable, we still expect sponsors to address the most egregious 
cases of apparent opioid overutilization without unreasonable delay.
    Under our proposed approach, a sponsor will provide an initial 
notice to a potential at-risk beneficiary if the sponsor intends to 
limit the beneficiary's access to coverage for frequently abused drugs, 
and the sponsor will provide a second notice to an at-risk beneficiary 
when it actually imposes a limit on the beneficiary's access to 
coverage for frequently abused drugs. Alternatively, the sponsor will 
provide an alternate second notice if it decides not to limit the 
beneficiary's access to coverage for frequently abused drugs. The 
second notice and alternate second notice are discussed later in this 
final rule.
    Finally, we proposed to require at Sec.  423.153(f)(5)(iii) that 
the Part D plan sponsor make reasonable efforts to provide the 
beneficiary's prescriber(s) of frequently abused drugs with a copy of 
the notice required under paragraph (f)(5)(i).
    We received the following comments related to the initial notice, 
and general comments applicable to all the proposed notices, and our 
responses follow:
    Comment: We received many comments related to our proposal to 
require written beneficiary notice both when a plan identifies the 
beneficiary as potentially at risk for prescription drug abuse, and 
again when the plan determines the beneficiary is at risk and 
implements a beneficiary-level POS edit

[[Page 16461]]

and/or a pharmacy or prescriber lock-in for frequently abused drugs. 
Some commenters disagreed with our proposal to require two notices, 
stating that a second notice would be unnecessary, confusing, or overly 
burdensome.
    Several other commenters strongly supported our proposal to require 
the two notifications, including the proposed change to the existing 
OMS process that would require the initial and second notices before a 
plan imposes a beneficiary-specific edit at POS. Commenters stated that 
requiring multiple notices will increase the likelihood that affected 
beneficiaries will be notified of their status and aware of how they 
could dispute it. A commenter wanted CMS to require more than two 
notices, because CMS did not propose to require acknowledgement of 
receipt from the beneficiary.
    Response: We thank those commenters who agreed with our proposals 
to require two notices and to integrate existing OMS process into a 
uniform process for all drug management program restrictions. While we 
appreciate the concerns expressed by commenters who do not agree with 
our proposal, as we noted in the proposed rule, the statute at Sec.  
1860D-4(c)(5)(B) clearly requires written beneficiary notification both 
upon identification as a potential at-risk beneficiary and again when 
the plan determines the beneficiary is at risk. We do not agree that 
additional notices beyond what we proposed should be required, as it 
would be overly burdensome on plans and provide little value to 
beneficiaries.
    Comment: Several commenters asked if stakeholders will have an 
opportunity to comment on the beneficiary notices and for more 
information on whether they can be modified by plans and when they will 
be released. A commenter requested that CMS conduct focus-group testing 
with beneficiaries to ensure the notice is understandable.
    Response: As discussed in section III.B.14 of this final rule, 
these notices are subject to approval by the Office of Management and 
Budget (OMB) under the Paperwork Reduction Act of 1995 (44 U.S.C. 3501 
et seq.). The notices will be posted in the Federal Register to give 
stakeholders an opportunity to review and comment before final versions 
of the notices are posted. CMS will consider testing through 
beneficiary focus groups, time permitting. The notices and accompanying 
instructions will contain detailed information about permissible 
modifications by plans. CMS intends to release the notices with 
sufficient time for plan sponsors to implement them into their drug 
management programs.
    Comment: We received some comments related to requirements to 
translate these beneficiary notices. Some of the commenters stated that 
these notices should be designated to be among materials subject to 
translation requirements in proposed Sec. Sec.  422.2268 and 423.2268. 
A commenter asked for clarification on whether plans are required to 
include section 1557 taglines with these notices.
    Response: While CMS is still developing instructions related to 
translation requirements to provide guidance on the requirements at 
Sec. Sec.  422.2268 and 423.2268, we note that, 423.128(d)(1)(iii) 
requires Part D plan sponsors' call centers to have interpreter 
services available to call center personnel to answer questions from 
limited-English proficient beneficiaries. These obligations are based 
on Medicare regulations and other civil rights laws, such as Title VI 
of the Civil Rights Act of 1964, that apply to Medicare health and drug 
plans. Applicability of Section 1557, and the scope of requirements for 
access for limited English proficient beneficiaries, and what is a 
significant communication are determined by the Office for Civil Rights 
(OCR).
    Comment: A commenter urged CMS to consider implementing additional 
requirements for beneficiary notification, including establishing 
requirements stipulating information that must be written on envelopes 
containing written notices, adding requirements for telephonic or email 
notification in addition to written notices, and requirements for 
prescribers to contact beneficiaries to confirm receipt of the required 
notices.
    Response: We agree with the commenter that detailed beneficiary 
notification is important, both upon identification as a potential at-
risk beneficiary and again either confirming the at-risk identification 
or that the plan has determined the beneficiary is not at-risk. 
However, we disagree with this commenter that additional notice 
requirements are necessary or advisable. We believe it would be overly 
burdensome to require plans to include specific information on the 
outside of mailing envelopes and there is no such precedent for similar 
beneficiary notices in the Part D program, such as notices of coverage 
denials or transition letters. While CMS expects that prescribers of 
frequently abused drugs will communicate regularly with their patients, 
we do not believe it is necessary to require prescribers to confirm 
that beneficiaries received the required plan notices. Finally, we note 
that, while CMS does not require telephonic or email notification in 
addition to the required written notices, plans are not precluded from 
doing so.
    Comment: A commenter asked why CMS proposed to require that the 
initial notice contain contact information for other organizations that 
can provide assistance to beneficiaries regarding the sponsor's drug 
management program.
    Response: Such information is statutorily required under Sec.  
1860D-4(c)(5)(B)(ii)(VII) to be included in the initial notice. As 
specified in the statute, it should be similar to the information 
provided in other standardized Part D beneficiary notices. We expect 
the notice may include, for example, contact information for the 
enrollee's State Health Insurance Program (SHIP), 1-800-MEDICARE, the 
Medicare Rights Center, and/or other organizations as appropriate.
    Comment: We received some comments that supported our proposal to 
require plan sponsors to make reasonable efforts to provide copies of 
notices to the potentially at-risk and at-risk beneficiary's 
prescriber(s).
    Response: We thank these commenters for their support.
    Comment: A few commenters opined that Part D plan sponsors and 
third party administrators do not have access to a list of all State 
and Federal public health resources designed to address prescription 
drug abuse. These commenters stated that requiring plans operating in 
multiple states to compile such a list would be overly burdensome, and 
requested that CMS provide templates containing such information as 
required under proposed Sec.  423.153(f)(5)(ii)(C)(2). Another 
commenter asked if MA-PD plans will be allowed to include information 
about plan-specific mental health benefits in addition to State and 
Federal resources.
    Response: CMS appreciates the input provided by these commenters. 
While the notice templates and instructions are still under 
development, CMS expects to provide information on Federal and State 
public health resources to assist plans in meeting the statutory 
requirement at Sec.  1860D-4(c)(5)(B)(ii)(II) to include such 
information in the initial notice. Under the existing regulations at 
Sec.  423.505(i), Part D plan sponsors are ultimately responsible for 
adhering to all terms and conditions of their contract with CMS, 
including compliance with all Federal laws, regulations and CMS 
instructions related to activities or responsibilities delegated to a 
third party. Pursuant to the regulation at Sec.  
423.153(f)(5)(ii)(C)(2), which we are finalizing as proposed, plans 
will be also required to include

[[Page 16462]]

information about relevant benefits and services covered by the plan, 
such as medical, mental health and MAT benefits.
    Comment: Some commenters stated that CMS should specify in 
regulation text that initial notices must not be sent to potential at-
risk beneficiaries until the plan has communicated with and received 
clinical information from the beneficiary's prescribers. These 
commenters noted that failure to conduct case management prior to 
sending the initial notice would interfere with doctor-patient 
relationships and unnecessarily alarm beneficiaries who may be 
determined not to be at-risk.
    Response: We agree with these commenters that initial notices 
should not be sent to beneficiaries before the plan has engaged in case 
management and attempted to communicate with the beneficiary's 
prescriber(s), and this is specified in the regulation text at Sec.  
423.153(f)(2)(i). However, we know from experience with the OMS process 
that prescribers are not always responsive to the plan's attempts to 
make clinical contact; therefore, we proposed at Sec.  
423.153(f)(2)(i)(C) that plans must make additional attempts to contact 
such prescribers. Additionally, we proposed at Sec.  423.153(f)(4) that 
plans cannot limit access to frequently abused drugs unless the plan 
has conducted case management and obtained agreement from prescribers 
(or made certain attempts to contact prescribers). We believe this 
approach strikes an appropriate balance between ensuring sufficient 
access to frequently abused drugs and protecting at-risk beneficiaries 
from potential harm in the absence of improved care coordination.
    After consideration of the comments received on this section, we 
are finalizing our proposal with modification to clearly codify the 
policy that a sponsor should not provide the initial notice to the 
beneficiary until after the sponsor has engaged in the required case 
management by adding the phrase ``after conducting the case management 
required by Sec.  423.153(f)(2)'' at the beginning of Sec.  
423.153(f)(5)(i).
(B) Limitation on the Special Enrollment Period for LIS Beneficiaries 
With an At-Risk Status (Sec.  423.38)
    Section 704(a)(3) of CARA gave the Secretary the discretion to 
limit the SEP for full benefit dually eligible (FBDE) beneficiaries 
outlined in section 1860D-1(b)(3)(D) of the Act. In addition to 
providing relevant information to a potential at-risk beneficiary, we 
proposed that the initial notice will notify dually- and other low 
income subsidy (LIS)-eligible beneficiaries that they would be unable 
to use the special enrollment period (SEP) for LIS beneficiaries due to 
their potential at-risk status. (Hereafter, this SEP is referred to as 
the ``duals' SEP''). This limitation is related to, but distinct from, 
other changes to the duals' SEP discussed in the proposed rule.
    We proposed that once a dually- or other LIS-eligible individual is 
identified as a potential at-risk beneficiary, and the sponsor intends 
to limit the beneficiary's access to coverage for frequently abused 
drugs, the sponsor will provide an initial notice to the beneficiary 
and the duals' SEP would no longer be available to the otherwise 
eligible individual. This means that he or she would be unable to use 
the duals' SEP to enroll in a different plan or disenroll from the 
current Part D plan. The limitation would be effective as of the date 
the Part D plan sponsor identifies an individual to be potentially at-
risk.
    We proposed that, consistent with the timeframes discussed in 
proposed paragraph Sec.  423.153(f)(7), if the Part D plan sponsor 
takes no additional action to identify the individual as an at-risk 
beneficiary within 90 days from the initial notice, the ``potentially 
at-risk'' designation and the duals' SEP limitation would expire. If 
the sponsor determines that the potential at-risk beneficiary is an at-
risk beneficiary, the duals' SEP would not be available to that 
beneficiary until the date the beneficiary's at-risk status is 
terminated based on a subsequent determination, including a successful 
appeal, or at the end of a 12-month period calculated from the 
effective date of the limitation, as specified in the second notice 
provided under Sec.  423.153(f)(6), whichever is sooner.
    We noted that auto- and facilitated enrollment of LIS eligible 
individuals and plan annual reassignment processes would still apply to 
dual- and other LIS-eligible individuals who were identified as an at-
risk beneficiary in their previous plan. Furthermore, we noted that the 
proposed enrollment limitations for Medicaid or other LIS-eligible 
individuals designated as at-risk beneficiaries would not apply to 
other Part D enrollment periods, including the AEP or other SEPs, 
including when an individual has a gain, loss, or change in Medicaid or 
LIS eligibility. We proposed that the ability to use the duals' SEP 
would not be permissible once the individual is enrolled in a plan that 
has identified him or her as a potential at-risk beneficiary or at-risk 
beneficiary under Sec.  423.100 of this final rule. (See section 
II.A.10 for a more detailed discussion of Part D SEP changes.)
    We received the following comments and our response follows:
    Comment: We received many comments supporting the limitation of the 
duals' SEP for those individuals identified as potential at-risk or at-
risk for overutilizing frequently abused drugs. Commenters noted that 
this limitation would support care coordination for this population, 
ensure that these beneficiaries are effectively managed, and prevent 
those that do abuse drugs from frequent plan switching, and either 
changing to a Part D plan without a drug management program, or 
accessing opioids because of a gap in information sharing across plans. 
Several commenters stated that this move would support their state's 
efforts in curbing the opioid epidemic.
    Response: We appreciate the support for our proposal to limit the 
SEP for individuals identified as potential at-risk or at-risk for 
overutilizing frequently abused drugs.
    Comment: A commenter requested that CMS confirm that any 
limitations on Part D LIS-eligible individuals would not impact the 
ability of such individuals to make an enrollment or disenrollment 
during other enrollment periods for which he or she is eligible. 
Commenters specifically asked about the AEP and the SEPs available for 
individuals to enroll in or disenroll from Program for All-inclusive 
Care (PACE) or enroll in a 5-Star plan.
    Response: We note that the enrollment limitation for a potential 
at-risk or an at-risk individual will not apply to other Part D 
enrollment periods, including the AEP or other SEPs, including new SEPs 
that will be established at Sec.  423.38(c)(9) and (c)(10) and are 
discussed in more detail in section II.A.10. of this final rule. In the 
event that an individual is subject to this limitation, but is eligible 
for another enrollment period, he or she may use that enrollment period 
to make a change. For example, a potential at-risk or at-risk dually- 
or other LIS-eligible individual who is subject to the duals' SEP 
limitation may use the PACE SEP to enroll in or disenroll from PACE, or 
they may use the 5-Star Rating SEP to enroll in an MA plan, PDP, or 
cost plan with a Star Rating of 5 stars during the year in which that 
plan has the 5-star overall rating, provided the enrollee meets the 
other requirements to enroll in that plan.
    Comment: A commenter asked for clarification as to whether the SEP 
limitation for potential at-risk or at-risk individuals would apply 
when a

[[Page 16463]]

beneficiary loses Medicaid eligibility and goes through the deeming 
process permitted in capitated models under Financial Alignment 
Initiative demonstrations. The commenter stated that, in their state, a 
beneficiary is allowed to remain in the demonstration Medicare-Medicaid 
Plan (MMP) for up to 3 months while he or she tries to regain Medicaid 
eligibility. If the beneficiary regains Medicaid eligibility within 
this 3 month window, would the state be required to allow the 
beneficiary to change his or her enrollment? The commenter stated, 
that, now, they automatically re-enroll the beneficiary back into the 
MMP.
    Response: The period of deemed continued eligibility provides an 
opportunity for individuals in Dual Special Needs Plans (D-SNPs) or 
MMPs who lose Medicaid eligibility to stay enrolled in their plan for a 
short time,\14\ while they try to regain Medicaid eligibility. However, 
should an individual be eligible to leave the plan, and takes an action 
to leave the plan, using any valid SEP, the plan must honor the 
disenrollment request. It is our view that a change in Medicaid status, 
especially loss of Medicaid eligibility, is an important event with 
potentially significant financial impacts to the beneficiary. As a 
result, the SEP outlined in Sec.  423.38(c)(9) will remain available to 
a potential at-risk or at-risk individual, even if the person is 
provided a deeming period by an MMP or D-SNP. This will permit 
individuals in a capitated model under the Financial Alignment 
Initiative demonstrations to change plans using the duals' SEP, within 
3 months of a gain, loss, or change to Medicaid or LIS eligibility, or 
notification of such.
---------------------------------------------------------------------------

    \14\ Under the capitated model of the Financial Alignment 
Initiative demonstration, MMPs may provide up to 3 months of deemed 
continued eligibility for individuals who lose MMP eligibility due 
to short-term loss of Medicaid. As outlined in Chapter 2 of the 
Medicare Managed Care Manual, D-SNPs must provide at least 1 month 
and up to 6 months of deemed continued eligibility for individuals 
who lose eligibility due to loss of Medicaid, but are reasonably 
expected to regain Medicaid within that timeframe.
---------------------------------------------------------------------------

    Comment: We received several comments relating to the operational 
aspects of implementing this limitation on the duals' SEP. Commenters 
requested clarification on how a plan sponsor would know if a potential 
at-risk or at-risk beneficiary was not eligible to use the duals' SEP, 
and how the MARx system would be operationalized to effectuate this 
change. A commenter requested clarification on how these individuals 
would be prevented from utilizing the duals' SEP.
    Response: Information related to an individual's at-risk status, 
including the beginning and end dates for any limitation imposed, will 
be stored in MARx and available to plans for enrollment processing via 
the User Interface (UI) and the beneficiary eligibility query (BEQ). 
CMS will reject a submitted enrollment for a beneficiary who is subject 
to the SEP limitation and the plan will be notified with a unique 
transaction reply code (TRC). We will also notify plans via a TRC if a 
member has a change in their at-risk status period. We will provide 
further subregulatory guidance on system and operational changes that 
will occur to effectuate this limitation, as well as the larger drug 
management program.
    Comment: To further assist in these efforts to curb opioid misuse, 
a commenter requested that CMS share data about any members in Part D 
plans who are subject to this SEP limitation to target Medicaid wrap 
services, including supplemental behavioral health and substance use 
treatment services.
    Response: We thank the commenter for their suggestion and we will 
explore data sharing for states to provide additional services to these 
individuals.
    Comment: A commenter recommended that CMS allow potential at-risk 
or at-risk individuals to use the duals' SEP to change to another plan 
if that plan has an established drug management program in place.
    Response: We appreciate the comment; however, we disagree with 
allowing individuals identified as potentially at risk or at risk to 
use the duals' SEP. Even if an at-risk individual joined another plan 
that had a drug management program in place, there would be challenges 
in terms of preventing a gap managing their potential or actual 
overutilization of frequently abused drugs due to the timing of 
information sharing between the plans and possible difference in 
provider networks.
    Comment: A commenter stated that because the ``at-risk'' status is 
transferable from one plan to another, an individual will not avoid the 
implications of the lock-in by utilizing the SEP. As such, the 
commenter believed that the dual SEP should not be limited.
    Response: We disagree. First, for general clarification purposes, 
the at-risk determination will not automatically transfer and be 
applied by a new Part D plan in the event a potentially at-risk or at-
risk beneficiary changes plans. Even though a gaining plan will be able 
to see if a new member had an at-risk determination with their prior 
plan, the new plan will still have to make their own determination 
regarding the individual's status and send the individual the 
appropriate notice, which will trigger the SEP limitation, as we have 
explained elsewhere in this preamble. Although the beneficiary's prior 
at-risk designation is an indicator that the new plan will have to 
initiate case management and may even allow them to bypass the first 
notice and go straight to issuing the second notice, the at-risk 
determination is not directly transferable.
    In addition, while we assume that all Part D sponsors will have 
drug management programs in place, it is not a requirement.
    With respect to the need for the SEP limitation, this policy is 
still needed to prevent potential and at-risk beneficiaries from making 
frequent plan changes after they receive the initial and second 
notices, as applicable, and thus, avoid the care coordination that drug 
management plans are intended to provide.
    We note that the SEP limitation--whether it is a first time 
designation or one that is being applied after enrollment into a new 
plan--will be effective as of the date on the initial notice that the 
Part D plan sponsor provides to an individual identified to be 
potentially at-risk. We are revising that language in Sec.  
423.38(c)(4) to state that beneficiaries that have been notified that 
they are potentially at-risk or at-risk, and such identification has 
not been terminated in accordance with Sec.  423.153(f)), will not be 
able to use the duals' SEP.
    Comment: A commenter encouraged CMS to offer increased resources to 
SHIPs to provide targeted outreach to the dual eligible and LIS 
populations who will be impacted by these changes. The commenter stated 
that CMS should also conduct outreach and education to providers and 
pharmacies, including mental health and substance use providers, as 
well as community based organizations (such as recovery learning 
communities), as these changes have a specific impact on beneficiaries 
with substance use disorders. The commenter stated that these efforts 
will help ensure that beneficiaries most likely to be impacted by these 
changes, and their providers, are made aware well in advance of 
implementation. Also, the commenter encouraged CMS and the 
Administration for Community Living (ACL) to provide continued funding 
for state Ombudsman programs that serve dual eligible populations 
enrolled in demonstration products, and to allow states to use this 
funding to serve dual

[[Page 16464]]

eligible beneficiaries enrolled in any integrated care product, 
including, for example FIDE SNPs.
    Response: CMS appreciates the comment, and we will continue to 
explore avenues for beneficiary and provider outreach and education; 
however, provisions for addressing cost and funding resources is 
outside of the scope of this rule.
    Comment: Several commenters opposed the limitation of the duals' 
SEP for at-risk beneficiaries. Commenters cited issues, such as limited 
access to prescription drugs and the possible risks of medical 
complications and increased costs resulting from such access barriers. 
They also noted the vulnerability and special needs of this population. 
A commenter stated that this limitation is unnecessary, as the current 
OMS program in Part D typically resolves cases of potential misuse 
without resorting to any beneficiary-specific tactic and would result 
in beneficiaries losing access to an important patient protection.
    Response: We appreciate the comments. As we stated in the proposed 
rule, based on the 2015 data in CMS' OMS, more than 76 percent of all 
beneficiaries estimated to be potential at-risk beneficiaries are LIS-
eligible individuals. It is our view that the SEP limitation will be an 
important tool to reduce the opportunities for dual and LIS-eligible 
beneficiaries designated as at-risk to switch plans, and circumvent the 
care coordination that a drug management program is designed to provide 
for this vulnerable population, especially as our nation faces an 
opioid epidemic. As stated previously, the enrollment limitation for a 
potential at-risk or an at-risk individual would not apply to other 
Part D enrollment periods, including the AEP or other SEPs. In the 
event that a potential at-risk or at-risk dually- or other LIS-eligible 
individual is subject to this limitation, but that individual is 
eligible to make an enrollment change using a different and valid 
election period, he or she may do so.
    In the case where an individual is prescribed a specific drug that 
is not on the sponsor's formulary, the individual always has the right 
to request a coverage determination for the drug. Each Part D sponsor 
that provides prescription drug benefits for Part D drugs and manages 
this benefit through the use of a formulary must establish and maintain 
exceptions procedures for receipt of an off formulary drug. A Part D 
sponsor must grant an exception whenever it determines that the drug is 
medically necessary, consistent with the physician's or other 
prescriber's statement, and that the drug would be covered but for the 
fact that it is an off formulary drug. Since these protections apply to 
all beneficiaries, they also protect dually-eligible and other LIS-
eligible beneficiaries.
    Comment: A couple of commenters stated that maintaining maximum 
flexibility regarding enrollment in Medicare Part D and the ability to 
change PDPs best serves the interests of low-income beneficiaries, 
especially American Indian and Alaskan Native (A/I and A/N) 
beneficiaries. The commenters further stated that a decision to change 
plans is often made in order to access a specific prescription drug. 
The commenters further requested that, if the proposed regulation is 
retained, CMS specify an exemption for Indian Health Service (IHS)-
eligible individuals as inserting the Medicare Part D drug plans into 
the relationship between Medicare/IHS beneficiaries and their IHS/
Tribal providers would not be helpful. We discuss IHS beneficiaries 
again further below.
    Response: CMS disagrees with establishing population-based 
exceptions to the duals' SEP limitation. In our view, all potential at-
risk and at-risk beneficiaries should be afforded the opportunity to 
benefit from the care coordination that the drug management program is 
designed to provide. We do not believe it is prudent at this time to 
carve out a subset of at-risk beneficiaries to which special rules 
apply. As previously mentioned, there are opportunities for potential 
at-risk and at-risk individuals to make enrollment choices during other 
election periods. Also, an individual always has the right to request a 
coverage determination, including an exception request for an off-
formulary drug.
    Comment: A couple of commenters expressed concern about this SEP 
limitation not being appealable. A commenter urged CMS to make the loss 
of the duals' SEP for potential at-risk beneficiaries appealable, as an 
at-risk beneficiary's other non-opioid-related conditions may justify 
the using of an SEP. A commenter noted that the proposal stipulated an 
appeals process for beneficiaries wishing to appeal their at-risk 
status, but encouraged CMS in its final rule to clarify whether the 
loss of a duals' SEP would be appealable in any way, and urge CMS to 
make a provision for beneficiaries who may need access to this SEP 
despite their at-risk status.
    Response: Similar to all other enrollment decisions, the limitation 
on the duals' SEP for potential at-risk or at-risk individuals is not 
appealable. However, after an individual is determined to be at-risk, 
he or she may appeal that determination. We intend to provide maximum 
transparency to the beneficiary by ensuring, consistent with the 
statutory requirements, that the beneficiary has information about 
appeal rights during the at-risk determination process.
    Comment: A commenter stated that nothing in the law would make a 
dual-eligible at- risk or potentially at-risk beneficiary ineligible 
for an SEP.
    Response: We disagree with the commenter. Section 704(a)(3) of CARA 
gives the Secretary the discretion to limit the SEP for FBDE 
beneficiaries outlined in section 1860D-1(b)(3)(D) of the Social 
Security Act (the Act). As discussed previously, the duals' SEP was 
extended to all other subsidy-eligible beneficiaries by regulation so 
that all LIS-eligible beneficiaries are treated uniformly.
    Comment: A commenter is concerned that dually- and other LIS-
eligible individuals inappropriately identified as potentially at-risk 
may not understand the process for correcting a determination that was 
made in error or may otherwise be inappropriate. The commenter further 
stated that some beneficiaries will be erroneously identified and not 
confirmed as at-risk and they should not be subject to the SEP 
limitation as a result of poor data, plan error, or some other reason 
unrelated to the beneficiary's action.
    Response: We appreciate the comments. We believe that there will be 
sufficient safeguards in the design and implementation of prescription 
drug management programs to prevent errors and provide beneficiaries 
with an opportunity to make corrections. CMS expects that exempt 
individuals will be identified through OMS. For those that are not 
excluded based on this data, they should be excluded by their plans 
during case management, as clinical contact and prescriber verification 
and agreement should occur before an initial notice of potential at-
risk status is sent to the individual and the SEP limitation is 
imposed. Thereafter, if a beneficiary believes he or she has been 
identified in error, the beneficiary has a chance to submit relevant 
information in response to the initial notice. If a determination is 
made that a beneficiary is an at-risk beneficiary, a Part D sponsor 
must also provide a second written notice to the beneficiary which is 
required to provide clear instruction on how a beneficiary may submit 
further applicable information to the sponsor. A beneficiary is also 
provided a right to redetermination of the at-risk status. CMS expects 
these measures will provide adequate protections for all beneficiaries.

[[Page 16465]]

    Comment: Another commenter requested clarification that the SEP is 
only removed for LIS beneficiaries once the plan sponsor has completed 
case management activities, including prescriber agreement.
    Response: We appreciate the question regarding when the duals' SEP 
limitation goes into effect. The duals' SEP limitation can go into 
effect without prescriber agreement; however, before the initial notice 
is sent, which informs the beneficiary of the limitation, the sponsor 
is required to engage in case management and attempt to communicate 
with the beneficiary's prescriber(s).
    Comment: A commenter urged CMS to make a provision for LIS 
beneficiaries who lose access to their SEP, but need access to non-
opioid drugs. For example, if an LIS beneficiary is determined to be 
at-risk and loses an SEP, and is later diagnosed with a different 
chronic condition that requires medication not on the beneficiary's 
current formulary. The commenter requested that CMS specify in the 
final rule that such a beneficiary would be given special consideration 
when submitting an appeal to their current plan to gain coverage of 
necessary non-opioid drugs.
    Response: We do not believe any ``special consideration'' is 
necessary. An enrollee--regardless of LIS eligibility--always has the 
right to request a coverage determination for a drug. In all cases, the 
standard is that the plan must notify the enrollee of its coverage 
determination decision as expeditiously as the enrollee's health 
condition requires, but no later than the applicable adjudication 
timeframe (24 hours for an expedited coverage determination, 72 hours 
for a standard coverage determination).
    Comment: A commenter noted that, while they agree with the proposal 
to implement the SEP provision, there may be an increase in complaints 
and grievances against the sponsor. The commenter encourages CMS to 
exclude beneficiaries identified as potentially at-risk and at-risk 
from Consumer Assessment of Healthcare Providers and Systems (CAHPS) 
surveys and not count complaints related to the duals' SEP limitation 
in the Complaint Tracking Module (CTM) numbers for star-rating 
purposes.
    Response: Thank you for the comment. Our Star Ratings proposal did 
not address this topic, and we plan to take this comment under 
advisement.
    After consideration of these comments, we are finalizing the 
provision on the CARA duals' SEP limitation at Sec.  423.38(c)(4) with 
a modification to specify that beneficiaries that have been notified 
that they are potentially at-risk or at-risk as defined in Sec.  
423.100, and such identification has not been terminated in accordance 
with Sec.  423.153(f)), will not be able to use the duals' SEP.
    The duals' SEP limitation will align with the revised timeframes 
for the potential-at-risk and at-risk status as addressed in section 
423.153(f) of this final rule. That is, if the Part D plan sponsor 
takes no additional action to identify the individual as an at-risk 
beneficiary within 60 days from the date on the initial notice, the 
``potentially at-risk'' designation and the duals' SEP limitation will 
expire. At-risk determinations will be for an initial 12 month period, 
with the option to extend for a maximum of 24 months in total (that is, 
an additional 12 month period) upon reassessment of the beneficiary's 
at-risk status at the completion of the initial 12 month period.
(C) Second Notice to Beneficiary and Sponsor Implementation of 
Limitation on Access to Coverage for Frequently Abused Drugs (Sec.  
423.153(f)(6))
    Section 1860D-4(c)(5)(B)(i)(I) of the Act requires Part D sponsors 
to provide a second written notice to at-risk beneficiaries when they 
limit their access to coverage for frequently abused drugs. We proposed 
to codify this requirement in Sec.  423.153(f)(6)(i). As with the 
initial notice, our proposed implementation of the statutory 
requirement for the second notice will also permit it to be used when 
the sponsor implements a beneficiary-specific POS claim edit for 
frequently abused drugs. Specifically, we proposed to require the 
sponsor to provide the second notice when it determines that the 
beneficiary is an at-risk beneficiary and to limit the beneficiary's 
access to coverage for frequently abused drugs. We further proposed to 
require the second notice to include the effective and end date of the 
limitation. Thus, this second notice will function as a written 
confirmation of the limitation the sponsor is implementing with respect 
to the beneficiary, and the timeframe of that limitation.
    We also proposed that the second notice, like the initial notice, 
contain language required by section 1860D-4(c)(5)(B)(iii) of the Act 
to which we proposed to add detail in the regulation text. The second 
notice must also be approved by the Secretary and be in a readable and 
understandable form, as well as contain other content that CMS 
determines is necessary for the beneficiary to understand the 
information required in the notice. In paragraph (2), we proposed 
language that will require a sponsor to include the limitation the 
sponsor is placing on the beneficiary's access to coverage for 
frequently abused drugs, the effective and end date of the limitation, 
and if applicable, any limitation on the availability of the SEP. We 
proposed an additional requirement in paragraph (6) that the sponsor 
include instructions how the beneficiary may submit information to the 
sponsor in response to the request described in paragraph (4). In Sec.  
423.153(f)(6)(iii), we proposed that the sponsor be required to make 
reasonable efforts to provide the beneficiary's prescriber(s) of 
frequently abused drugs with a copy of the notice, as we proposed with 
the initial notice. Finally, we proposed a requirement in paragraph (7) 
that the notice contain other content that CMS determines is necessary 
for the beneficiary to understand the information required in the 
initial notice.
    Also, the sponsor will generally be required to send two notices--
the first signaling the sponsor's intent to implement a POS claim edit 
or limitation (both referred to generally as a ``limitation''), and the 
second upon implementation of such limitation. Under our proposal, the 
requirement to send two notices will not apply in certain cases 
involving at-risk beneficiaries who are identified as such and provided 
a second notice by their immediately prior plan's drug management 
program.
    We received the following comments and our responses follow:
    Comment: We received many comments related to our proposal 
requiring plans to provide a second written notice to beneficiaries 
before implementing a restriction under the plan's drug management 
program, most of which supported the proposal. Other commenters opposed 
it, expressing a belief that only one notice would be sufficient. Some 
of these commenters offered ideas for various alternative approaches 
for CMS to consider, such as including information in the plan's 
Evidence of Coverage that would replace the notices described in the 
proposed rule, or using a single notice similar to the current OMS 
requirement. Other commenters stated that the two notices required for 
lock-in should be limited to lock-in and plans should continue to be 
permitted to send a single notice when implementing a beneficiary-level 
POS edit.
    Response: We disagree with the comments recommending requiring a 
single beneficiary notice or replacing one or both notices with general 
information in other documents. Section

[[Page 16466]]

1860D-4(c)(5)(B) requires two written notices before a beneficiary can 
be locked-in to a prescriber or pharmacy, and includes a high level of 
specificity about the content of the notices. Moreover, the required 
initial and second notices contain important information about access 
restrictions that may be or will be placed on potentially at-risk and 
at-risk beneficiaries, resources such beneficiaries may need to treat 
potential drug dependency issues, and notification of important 
beneficiary rights.
    We also disagree with comments stating that the proposed notice 
requirements for the lock-in program should be limited to lock-in, and 
that CMS should retain existing beneficiary notice policies, including 
sending only one notice, when implementing beneficiary-level POS edits. 
Currently, the application of a beneficiary-level POS claim edit is not 
considered a coverage determination and does not trigger appeal rights 
under Subpart M. As we explained in the proposed rule, the 
implementation of a beneficiary-specific POS claim edit or a limitation 
on the at-risk beneficiary's coverage for frequently abused drugs to a 
selected pharmacy(ies) or prescriber(s) will be an aspect of an at-risk 
determination (a type of initial determination that will confer appeal 
rights on the beneficiary, consistent with section 1860D-4(c)(5)(E) of 
the Act) under our proposal establishing the Part D drug management 
program. As discussed in subsection (c) of this preamble, we are 
finalizing the proposal to integrate the current OMS process with lock-
in to create a uniform drug management program for Part D. Under this 
final rule, since the application of a beneficiary-level POS edit for 
frequently abused drugs can only be applied upon the plan's at-risk 
determination and is subject to appeal, it is necessary to treat those 
edits the same as limitations on selected pharmacy(ies) or 
prescriber(s). Furthermore, we believe that establishing an 
inconsistency with respect to notice requirements would be confusing 
for beneficiaries and plans. For these reasons, and because we believe 
the second notice, which identifies the action taken by the plan and 
instructs the beneficiary how to exercise their statutory appeal 
rights, is an important beneficiary protection, the notice is required 
both for lock-in and for POS edits for frequently abused drugs.
    Comment: A commenter suggested that CMS require that the second 
notice, in addition to the initial notice, include a description of all 
State and Federal public health resources addressing prescription drug 
abuse that are available to the beneficiary.
    Response: While we agree that this information is important to 
communicate to affected beneficiaries, we recognize the potential 
burden that multiple notices may place on plan sponsors as well as 
beneficiaries. We note that such information is required in the initial 
notice, and the statute does not require it in the second notice. While 
CMS will not preclude plans from providing this information again, for 
example, if requested by the enrollee, we do not believe it is 
necessary to require that it be included in both notices.
    After consideration of comments received, we are finalizing our 
proposal without modification to require plans to send both the initial 
and second notice before implementing a beneficiary-level POS edit or a 
pharmacy or prescriber lock-in under a drug management program.
(D) Alternate Second Notice When Limit on Access Coverage for 
Frequently Abused Drugs by Sponsor Will Not Occur (Sec.  423.153(f)(7))
    Although not explicitly required by the statute, we proposed at 
Sec.  423.153(f)(7) that if a sponsor determines that a potential at-
risk beneficiary is not an at-risk beneficiary and does not implement 
the limitation on the potential at-risk beneficiary's access to 
coverage of frequently abused drugs it described in the initial notice, 
then the sponsor will be required to provide the beneficiary with an 
alternate second notice. Specifically, we proposed that such alternate 
second notice use language approved by the Secretary in a readable and 
understandable form, and contain the following information: The sponsor 
has determined that the beneficiary is not an at-risk beneficiary; the 
sponsor will not limit the beneficiary's access to coverage for 
frequently abused drugs; if applicable, the SEP limitation no longer 
applies; clear instructions that explain how the beneficiary may 
contact the sponsor; and other content that CMS determines is necessary 
for the beneficiary to understand the information required in this 
notice.
    As with the other notices, we proposed that the Part D sponsor be 
required to make reasonable efforts to provide the beneficiary's 
prescriber(s) of frequently abused drugs with a copy of this notice.
    We received the following comments and our response follows:
    Comment: We received a few comments on this proposal. Some of these 
commenters supported the proposal and agreed that such notice is 
necessary to minimize beneficiary confusion and limit unneeded appeals 
when a plan decides not to implement any restrictions on frequently 
abused drugs. A commenter disagreed with our proposal to require an 
alternate second notice, stating such notice is not necessary.
    Response: As we stated in the proposed rule, we believe that this 
alternate notice is necessary to ensure beneficiaries who received the 
initial notice of an intended limitation on access to frequently abused 
drugs under the plan's drug management program are informed of the 
outcome of the plan's decision not to take such action. We are 
finalizing Sec.  423.153(f)(7) without modification.
(E) Timing of Notices and Exceptions to Timing (Sec.  423.153(f)(8))
    Section 1860D-4(c)(5)(B)(iv) of the Act requires a Part D sponsor 
to provide the second notice to the beneficiary on a date that is not 
less than 30 days after the sponsor provided the initial notice to the 
beneficiary. Although not specifically required by CARA, we believe it 
is also important to establish a maximum timeframe by which the plan 
must send the second notice or the alternate second notice, to ensure 
that plans do not leave a case open indefinitely. We proposed to 
specify at Sec.  423.153(f)(8)(i) that a Part D sponsor must provide 
the second notice described in paragraph (f)(6) or the alternate second 
notice described in paragraph (f)(7), as applicable, on a date that is 
not less than 30 days and not more than the earlier of the date the 
sponsor makes the relevant determination or 90 days after the date of 
the initial notice described in paragraph (f)(5).
    Section 1860D-4(c)(5)(B)(iv)(II) of the Act explicitly provides for 
an exception to the required 30 day minimum timeframe for issuing a 
second notice. Specifically, the statute permits the Secretary to 
identify through rulemaking concerns regarding the health or safety of 
a beneficiary or significant drug diversion activities that will 
necessitate that a Part D sponsor provide the second written notice to 
the beneficiary before the minimum 30 day time period normally required 
has elapsed.
    As we explained in the proposed rule, because this provision also 
allows an at-risk identification to carry forward to the next plan, we 
believe it is appropriate to permit a gaining plan to provide the 
second notice to an at-risk beneficiary so identified by the most

[[Page 16467]]

recent prior plan without having to wait the minimum 30 days, if 
certain conditions are met. This is consistent with our current policy 
under which a gaining sponsor may immediately implement a beneficiary-
specific POS claim edit, if the gaining sponsor is notified that the 
beneficiary was subject to such an edit in the immediately prior plan 
and such edit had not been terminated.\15\
---------------------------------------------------------------------------

    \15\ See ``Beneficiary-Level Point-of-Sale Claim Edits and Other 
Overutilization Issues,'' August 25, 2014.
---------------------------------------------------------------------------

    As such, at Sec.  423.153(f)(8)(ii), we proposed one exception to 
the timing of the notices, applicable to at-risk beneficiaries who 
switch plans. The exception allows a gaining plan sponsor to 
immediately provide the second notice described in paragraph (f)(6) to 
a beneficiary for whom the gaining sponsor received notice that the 
beneficiary was identified as an at-risk beneficiary by the prior plan 
and such identification had not been terminated. The exception is only 
permissible if the gaining sponsor is implementing either a 
beneficiary-specific POS edit as described in paragraph (f)(3)(i) under 
the same terms as the prior plan, or a limitation on access to coverage 
as described in paragraph (f)(3)(ii), if such limitation will require 
the beneficiary to obtain frequently abused drugs from the same 
pharmacy location and/or the same prescriber, as applicable, that was 
selected under the immediately prior plan under (f)(9).
    We received the following comments and our responses follow:
    Comment: Some commenters recommended that the timeframe between the 
first and second notices be shortened to within 15 days, which the 
commenters believe would provide sufficient time for beneficiaries to 
submit preferences. A commenter noted that there is no added value in 
waiting 30 days after the initial notice to provide the second notice 
because it contains similar information.
    Response: We disagree with these commenters. Outside of 
circumstances identified by the Secretary through rulemaking, section 
1860D-4(c)(5)(B)(iv) requires that the second notice be provided ``on a 
date that is not less than 30 days'' after the initial notice. 
Moreover, because the statute gives significant deference to 
beneficiary preferences, CMS does not believe that 15 days is 
sufficient for beneficiaries to receive the initial notice, identify 
their preferences for prescribers and/or pharmacies, potentially confer 
with the preferred prescribers and/or pharmacies, communicate 
preferences to their plan, and give the plan sufficient time to 
implement the limitation in their systems, including situations where 
the plan determines that an exception to preferences under Sec.  
423.153(f)(10) is warranted.
    Comment: We received several comments supporting our proposal to 
establish a maximum timeframe by which sponsors must send the second or 
alternate second notice. However, most of these commenters expressed 
concerns that 90 days is too long because potentially at-risk 
beneficiaries would be subject to a limitation on their SEP without 
appeal rights during that 90 day timeframe. Commenters stated that, if 
those beneficiaries identified as potentially at-risk did not lose 
access to the SEP, 90 days would be acceptable. Other commenters 
expressed a belief that plans would not need 90 days to obtain 
beneficiary preferences and implement relevant access limitations upon 
receipt of those preferences.
    Response: We appreciate the commenters' feedback about the proposed 
90 day maximum timeframe. As we noted in the preamble to the proposed 
rule, while section 1860D-4(c)(5)(B)(iv) of the Act requires plans to 
wait a minimum of 30 days from the initial notice before providing the 
second notice, Congress did not establish a maximum timeframe. Because 
case management, clinical contact and prescriber verification 
requirements would be met before the plan sends the initial notice, we 
agree with the commenters that our proposed 90 day maximum timeframe 
between notices could be shortened. Therefore, we are modifying Sec.  
423.153(f)(8)(i) to require the notice required under (f)(6) or 
alternate notice required under (f)(7) to be provided to the 
beneficiary no more than the earlier of the date the sponsor makes the 
relevant determination or 60 days after the date of the initial notice 
required under (f)(5).
    Given the comments received, many of which stated that the 90 day 
maximum timeframe we proposed is too long, we believe 60 days strikes 
the right balance. We do not believe the maximum timeframe should be 
shorter than 60 days, because sponsors may need this time to process 
information from beneficiaries that is received at the end of the 
minimum 30 day timeframe, or to communicate with prescribers who may 
have been unresponsive prior to receiving a copy of the initial notice 
the plan provided to the beneficiary. This revised timeframe is still 
sufficient to limit any potential compliance issues for sponsors 
related to timeliness and unnecessary appeals where such information is 
still being processed. However, we do not expect sponsors to routinely 
take the maximum amount of time to issue the second notice, and note 
that they must send it sooner if they make the relevant determination 
sooner. We note that the SEP is addressed in an earlier section of this 
preamble.
    Comment: We received several comments related to our proposal at 
Sec.  423.153(f)(8)(ii) to, under certain circumstances, permit a 
gaining plan to immediately send a second notice without waiting 30 
days to a beneficiary who is already subject to a drug management 
program coverage limitation (a beneficiary-specific POS claim edit or 
pharmacy or prescriber lock-in) in their immediately prior plan. Most 
commenters supported our proposal to establish an exception to the 30-
day notice for at-risk beneficiaries, as identified by the losing plan, 
when such beneficiaries switch plans and the gaining plan decides to 
continue the same limitation(s). Some of these commenters agreed that 
exceptions to the 30 day notice should be limited to circumstances 
where the beneficiary was already given notice by the previous plan. 
Some commenters noted that because a beneficiary may be changing plans 
due to dissatisfaction with their current providers, these 
beneficiaries must also have an opportunity to change their preferences 
with respect to pharmacies and prescribers when they change plans. 
Other commenters supported the exception that we proposed but stated 
that the statute allows exceptions under additional circumstances based 
on the health and safety of the beneficiary or significant drug 
diversion activity. A commenter recommended that CMS should specify 
that when a beneficiary who moves to a new plan offered by the same 
parent organization as their prior plan, the plan is not required to 
send any notice to the beneficiary to continue the restriction because 
such notice would only serve to confuse the beneficiary.
    Response: As we explained in the proposed rule, we believe that 
exceptions to the statutory requirement to wait at least 30 days before 
sending the second notice and implementing a coverage limitation under 
a drug management program should be very limited. Since the drug 
management program is focused on improved care coordination for 
beneficiaries who are utilizing high doses of frequently abused drugs 
and/or have multiple providers, and the statute specifies that such 
exceptions be identified through rulemaking regarding the health or

[[Page 16468]]

safety of the beneficiary or regarding significant drug diversion 
activities, we do not believe that it is appropriate to permit such an 
exception based on a sponsor's concerns about the health and safety of 
a particular beneficiary because that is too subjective and could 
adversely impact such beneficiaries, who could be subject to a coverage 
limitation without notice. Rather, we are finalizing the exception we 
proposed related to at-risk beneficiaries who switch plans and the 
gaining plan decides to continue a limitation(s) under the same terms 
as the losing plan, because we believe, in this instance, the coverage 
limitation(s) can safely be immediately implemented--namely, when the 
beneficiary already has been identified as at-risk by his or her prior 
plan, and the coverage limitations would continue in the same manner 
under his or her new plan. We have not at this time identified 
additional circumstances under which an exception to the 30-day minimum 
between the first and second notices is warranted. We note that this 
final rule does not change existing requirements that Part D plan 
sponsors cannot pay fraudulent claims. With respect to a beneficiary 
who changes plans within the same parent organization, we are 
clarifying that the gaining plan must still meet the requirements set 
forth at Sec.  423.153(f)(8)(ii). We do not believe it is advisable to 
apply a different standard to a gaining plan just because it has the 
same parent organization as the losing plan.
    While we are finalizing our proposed exception to the timing of the 
notices, we agree with the commenters who stated that beneficiaries who 
change plans should still have an opportunity to change their 
preferences for prescribers and pharmacies. Therefore, we are 
clarifying that an at-risk beneficiary's right to submit new 
preferences we are finalizing at (f)(9) also applies to beneficiaries 
who switch plans. While a gaining plan could still implement the 
restriction without providing 30 day advance notice, they must comply 
with the statutory and regulatory requirements to accept beneficiary 
preferences. Under the exception to the notice requirements that we are 
finalizing in this rule, a gaining plan choosing to immediately impose 
the restriction(s) of the prior plan is not required to resend the 
initial notice described at (f)(5) that was sent by the prior plan, but 
must issue a new version of the second notice described at (f)(6). This 
notice, which is being developed by CMS, will allow the gaining plan to 
include updated information from the initial notice that changes with 
the change to the new plan (for example, plan contact information or 
relevant medical benefits available to such beneficiary under the new 
plan).
    After consideration of all comments received on Sec.  
423.153(f)(8), we are finalizing our proposal at paragraph (f)(8)(i) to 
retain the minimum 30 day timeframe between the initial and second or 
alternate second beneficiary notices (except as provided in 
subparagraph (ii)), with a modification establishing a maximum 
timeframe of 60 days between the notices.
    Additionally, we are finalizing the proposed exception to the 
minimum 30 day timeframe at Sec.  423.153(f)(8)(ii), which permits a 
gaining plan to immediately issue the second beneficiary notice 
required by (f)(6) and implement a continuation of the same claim edit 
and/or pharmacy or prescriber lock-in for an at-risk beneficiary who 
was already provided the initial and second notice for such 
limitation(s) from the losing plan. As discussed above, we believe the 
circumstances under which a limitation can be safely implemented 
without advance beneficiary notice and are consistent with the 
requirements for such exceptions at section 1860D-4(c)(5)(iv)(II) are 
limited in scope. While, at this time, we have not identified 
additional circumstances under which we believe an exception to the 30 
day beneficiary notice is warranted under section 1860D-
4(c)(5)(B)(iv)(II), we will continue to evaluate this issue, and may 
establish additional exceptions through future rulemaking.
(viii) Provisions Specific to Limitations on Access to Coverage of 
Frequently Abused Drugs to Selected Pharmacies and Prescribers 
(Sec. Sec.  423.153(f)(4) and 423.153(f)(9) Through (13))
    Some of the drug management program provisions in CARA are only 
relevant to ``lock-in.'' We proposed several regulatory provisions to 
implement these provisions, as follows:
(A) Special Requirement To Limit Access to Coverage of Frequently 
Abused Drugs to Selected Prescriber(s) (Sec.  423.153(f)(4))
    In the proposed rule, we noted that, at that time, we viewed 
prescriber lock-in as a tool of last resort to manage at-risk 
beneficiaries' use of frequently abused drugs, meaning when a different 
approach has not been successful, whether that was a ``wait and see'' 
approach after case management or the implementation of a beneficiary 
specific POS claim edit or a pharmacy lock-in. We also were concerned 
about impacting an at-risk beneficiary's relationship with their 
provider, and we sought comment on whether a 6-month delay before a 
sponsor could implement prescriber lock-in would lessen burden on 
prescribers.
    As a result, we proposed in Sec.  423.153(f)(4)(iv) that a sponsor 
may not limit an at-risk beneficiary's access to coverage of frequently 
abused drugs to a selected prescriber(s) until at least 6 months has 
passed from the date the beneficiary is first identified as a potential 
at-risk beneficiary. We specifically sought comment on whether this 6-
month waiting period would reduce provider burden sufficiently to 
outweigh the additional case management, clinical contact and 
prescriber verification that providers may experience if a sponsor 
later believed a beneficiary's access to coverage of frequently abused 
drugs should be limited to a selected prescriber(s).
    We received the following comments and our response follows:
    Comment: Many commenters expressed significant concerns with the 
proposal to require a Part D plan sponsor to wait at least six months 
from the date the beneficiary is first identified as a potential at-
risk beneficiary before limiting that beneficiary to a prescriber for 
frequently abused drugs, noting that it works against the goal of CARA 
and defeats the purpose of the lock-in program. Moreover, many 
commenters also expressed that a 6 month delay to prescriber lock-in 
was not in the spirit of a national public health emergency, and may 
actually place at-risk beneficiaries at even greater risk for adverse 
health outcomes. A commenter expressed support for the 6 month delay, 
noting that it would allow time for alternative interventions to be 
implemented so as to not burden the prescriber unnecessarily. A 
commenter offered a lengthy legal argument against the 6-month delay 
for prescriber lock-in.
    Response: In light of these comments, we have been persuaded not to 
finalize require a 6 month waiting period before a plan may limit an 
at-risk beneficiary to a prescriber for frequently abused drugs. We 
agree with the majority of commenters that CMS should not impose a 
waiting period for plan sponsors to implement a prescriber lock-in for 
at-risk beneficiaries, and that once a beneficiary is deemed at-risk, a 
plan sponsor should have the full range of limitations on access to 
coverage for frequently abused drugs to employ for such beneficiaries. 
We are persuaded

[[Page 16469]]

that our initial concern about the beneficiary's relationship with a 
provider is significantly outweighed by the more immediate concerns for 
the beneficiary's safety.
    In addition, we are unpersuaded that our proposal would reduce 
burden on providers. This is because a sponsor, in conducting the case 
management is required under Sec.  423.153(f)(2), to contact 
prescribers and the sponsor may seek a prescriber's agreement to a 
beneficiary-specific POS claim edit pursuant to Sec.  423.153(f)(4). 
Thus, we now believe that requiring a sponsor to wait 6 months to 
contact the prescriber again to assist with additional case management 
for the prescriber lock-in, and to possibly obtain the prescriber's 
agreement to such lock-in, will actually increase provider burden.
    For these reasons, we are not finalizing the proposal that a 
sponsor may not limit an at-risk beneficiary's access to coverage of 
frequently abused drugs to a selected prescriber(s) until at least 6 
months has passed from the date the beneficiary is first identified as 
a potential at-risk beneficiary. Therefore, we have removed the 
language from Sec.  423.153(f)(4) relevant to this 6-month waiting 
period for prescriber lock-in.
(B) Selection of Pharmacies and Prescribers (Sec. Sec.  423.153(f)(9) 
Through (13))
(1) Beneficiary Preferences (Sec.  423.153(f)(9))
    Section 1860D-4(c)(5)(D)(iii) of the Act provides that, if a 
sponsor intends to impose, or imposes, a limit on a beneficiary's 
access to coverage of frequently abused drugs to selected pharmacy(ies) 
or prescriber(s), and the potential at-risk beneficiary or at-risk 
beneficiary submits preferences for a network pharmacy(ies) or 
prescriber(s), the sponsor must select the pharmacy(ies) and 
prescriber(s) for the beneficiary based on such preferences, unless an 
exception applies, for example, the beneficiary's preferred provider 
would contribute to the beneficiary's abuse of prescription drugs. We 
address exceptions to beneficiary's preferences later in the preamble.
    In light of this language, we proposed a Part D plan sponsor must 
accept an at-risk beneficiary's preferences for in-network prescribers 
and pharmacies from which to obtain frequently abused drugs unless an 
exception applies. In cases that involve stand-alone PDPs, we proposed 
that a sponsor must accept the beneficiary's selection of prescriber, 
unless an exception applies, because such PDPs do not have provider 
networks. We further proposed that a stand-alone PDP or MA-PD does not 
have to accept a beneficiary's selection of a non-network pharmacy, 
except as necessary to provide reasonable access, which we discuss 
later in this section. Our rationale for this proposal was that the 
selection of network prescribers and pharmacies puts the plan sponsor 
in the best possible position to coordinate the beneficiary's care 
going forward in light of the demonstrated concerns with the 
beneficiary's utilization of frequently abused drugs.
    Also, we did not propose to place a limit on how many times 
beneficiaries can submit their preferences, but we did solicit 
additional comments on this topic. Finally, under our proposal, the 
sponsor would be required to confirm the selection of pharmacy and/or 
prescriber in writing to the beneficiary either in the second notice, 
if feasible, or within 14 days of receipt of the beneficiary's 
submission.
    We received the following comments and our response follows:
    Comment: Commenters widely supported CMS's proposal that the 
pharmacy or prescriber in which an at-risk beneficiary is locked-into 
must be in-network for a plan, except to provide reasonable access or 
when the plan does not have a relevant network. Specifically, 
commenters noted that allowing selection of out of network pharmacies 
or prescribers would undermine keeping beneficiary costs low, and 
efforts to combat pharmacy-based fraud and abuse.
    Response: We thank commenters for their support.
    Comment: CMS received a handful of comments that disagreed that a 
prescriber should have to be in-network, given some Medicare Advantage 
beneficiaries may receive out-of-network treatment from providers due 
to their relationships with the prescriber and the high quality of care 
that they provide. These commenters requested that CMS eliminate the 
requirement that a prescriber generally must be in-network if the plan 
sponsor imposes a limit on a beneficiary's access to coverage for 
frequently abused drugs to a selected prescriber or prescribers.
    Response: We were not persuaded that sponsors should have to accept 
a beneficiary's selection of an out-of-network prescriber or pharmacy, 
unless needed to maintain reasonable access or if the plan does not 
have a relevant network. Our rationale for this is that Section 1860D-
4(c)(5)(D)(iii) refers specifically to the beneficiary selecting a 
network prescriber(s) and/or pharmacy(ies) and the plan sponsor 
accepting such selections based on the beneficiary's preference. We 
therefore believe that the statute does not contemplate requiring Part 
D plan sponsors to select a beneficiary's preference of an out-of-
network prescriber or pharmacy in all instances.
    However, because our requirements for drug management programs--as 
proposed and finalized--permit stand-alone PDPs to use prescriber lock-
in, the requirement for a sponsor to accept the beneficiary's selection 
of a network prescriber is inapplicable, and the sponsor must accept 
the beneficiary's selection of a prescriber, unless an exception 
applies, such as if the selection would contribute to the beneficiary's 
abuse of prescription drugs. With regard to this exception, we note 
that when there is a prescriber or pharmacy network, and the plan 
sponsor asserts it would accept a beneficiary's in-network pharmacy or 
prescriber preference(s) but such selection would contribute to 
prescription drug abuse or drug diversion by the beneficiary, we would 
question why such pharmacy or prescriber is in the sponsor's network.
    We realize that in the case of at-risk beneficiaries enrolled in MA 
plans that provide out-of-network coverage of services and are designed 
and specifically authorized for that purpose (that is, PPO, PFFS, and 
cost plans), these beneficiaries have access to supplemental services 
out of network. However, as we stated above, Section 1860D-
4(c)(5)(D)(iii) states that if an at-risk beneficiary submits 
preferences for which in-network prescribers and pharmacies the 
beneficiary would prefer, the PDP sponsor shall select them. The 
requirement, discussed later, that Part D prescription drug management 
programs ensure reasonable access addresses the sponsor's selection 
out-of-network prescribers and pharmacies when necessary and therefore 
accommodate our regulations at Sec.  422.105; Sec.  422.112 that permit 
out-of-network coverage.
    We note that by requiring a plan sponsor to accept an at-risk 
beneficiary's selection of an out-of-network prescriber, we would in 
effect have a blanket requirement that a coordinated health plan to 
manage an at-risk beneficiary out-of-network, which would be difficult 
to achieve. For those at-risk beneficiaries locked into a particular 
prescriber(s) and/or pharmacy(ies), prescriptions for frequently abused 
drugs would need to be obtained from an in-network prescriber (when 
such a network exists), even in the case of at-risk beneficiaries who 
are enrolled in MA plan that provide for out-of-network coverage.

[[Page 16470]]

Therefore, we are finalizing our provision as proposed.
    We wish to make a point of clarification regarding at-risk 
beneficiaries who are entitled to fill prescriptions or receive 
services from IHS, Tribal, and Urban Indian (ITU) organization 
pharmacies and providers. An IHS I/T/U pharmacy or provider may be the 
selected pharmacy or prescriber for such beneficiaries and they may go 
to such a pharmacy or prescriber pursuant to our reasonable access 
requirement, even if they are not in-network which we discuss again 
later.
    Comment: Regarding a limitation on how many times beneficiaries can 
submit their preferences, many commenters suggested that we allow an 
at-risk beneficiary to submit his or her preferences anywhere from 1 to 
3 times per year, noting that it was important to cap the number of 
times preferences can be submitted. A commenter noted that the 
beneficiary's unlimited opportunity to change preferences for 
prescribers and pharmacies will be problematic and burdensome, and 
recommended that CMS place a limit on the number of times a beneficiary 
may change preferences on an annual basis, unless they can provide good 
cause for requesting the change. Suggested examples of good cause would 
include moving beyond easy access to the prescriber or pharmacy; the 
prescriber has discharged the beneficiary from his/her practice; or the 
pharmacy is unable to provide the requested drugs.
    Response: While commenters raised concerns that at-risk 
beneficiaries should have some parameters around changing their 
preferences for a selected pharmacy or prescriber, CMS must balance 
curbing opioid overuse and misuse with ensuring reasonable access to 
selected pharmacies and prescribers. Therefore, we will allow at-risk 
beneficiaries to submit their preferences to plan sponsors without a 
numerical restriction during the plan year. We note that the sponsor 
does not have to make changes to the selection of pharmacy(ies) and 
prescriber(s) based on the at-risk beneficiaries preferences if the 
plan sponsor believes such changes are contributing to abuse or 
diversion of frequently abused drugs, pursuant to Sec.  423.153(f)(10), 
discussed above. Also, CMS will monitor for these issues and act 
accordingly to ensure efficient operation of the program and prevention 
of excessive administrative burden.
    Comment: A commenter stated that an at-risk beneficiary should not 
be locked-into pharmacies in which the plan sponsor or PBM overseeing 
the drug management program has a financial interest.
    Response: Since the selection of the pharmacy in which an at-risk 
beneficiary is locked into is largely a beneficiary choice, and one 
they are provided specifically in the statute with little exception, 
CMS does not find this comment persuasive, and will finalize this 
provision as proposed.
    Comment: A commenter stated that plan sponsors should be able to 
implement the change in a beneficiary's preference within 14 days after 
the beneficiary has submitted the preference.
    Response: We note that our proposal, which we are finalizing, 
requires the sponsor to inform the beneficiary of the selection in the 
second notice or if not feasible due to the timing of the beneficiary's 
submission of preference, in a subsequent written notice issue no later 
than 14 days after receipt of the submission.
    Accordingly, we are finalizing Sec.  423.153(f)(9), as proposed. We 
note that we added the words ``or change'' in paragraph (iii) for 
consistency with the rest of the regulation text in this section.
(2) Exception to Beneficiary Preferences (Sec.  423.153(f)(10))
    Section 1860D-4(c)(5)(D)(iv) of the Act provides for an exception 
to an at-risk beneficiary's preference of prescriber or pharmacy from 
which the beneficiary must obtain frequently abused drugs, if the 
beneficiary's allowable preference of prescriber or pharmacy will 
contribute to prescription drug abuse or drug diversion by the at-risk 
beneficiary. Section 1860-D-4(c)(5)(D)(iv) of the Act requires the 
sponsor to provide the at-risk beneficiary with at least 30 days 
written notice and a rationale for not accepting his or her allowable 
preference for pharmacy or prescriber from which the beneficiary must 
obtain frequently abused drugs under the plan.
    We received the following comments and our response follows:
    Comment: Commenters generally agreed with our proposal that plan 
sponsors may disallow a beneficiary's selection of a prescriber or 
pharmacy that may contribute to prescription drug abuse or drug 
diversion.
    Response: We appreciate the commenters support.
    Comment: A commenter suggested that CMS require plans/PBMs to 
report the percentage of times when beneficiary preference is/is not 
considered and to track which pharmacy the plan/PBM utilizes to 
override patient preference.
    Response: While we are not currently requiring that plans or PBMs 
report to CMS the percentage of times when beneficiary preference is/is 
not considered and to track which pharmacy the plan/PBM utilizes to 
override patient preference, we will re-evaluate this policy in the 
future if it becomes problematic. Therefore, we will closely monitor to 
make sure plans are not inappropriately choosing to not accept 
beneficiary preferences, in order to ensure efficient operation of the 
program and prevention of excessive administrative burden.
    While we received no comments specific to beneficiary appeal rights 
when the plan's selection of pharmacies or prescribers for lock-in are 
not aligned with the beneficiary's submitted preferences, we remind 
plans that the statute at Sec.  1860D-2(c)(5)(E) specifically states 
that the selection of pharmacy or prescriber for lock-in is subject to 
appeal. If a beneficiary complains about being locked into a pharmacy 
or prescriber that is not the one they selected, such complaint must be 
treated as an appeal. We address beneficiary appeals rights later in 
this preamble.
    We are finalizing the following at Sec.  423.153(f)(10) Exception 
to Beneficiary Preferences, as proposed.
(3) Reasonable Access (Sec. Sec.  423.100, 423.153(f)(11) 
423.153(f)(12))
    If a potential at-risk beneficiary or at-risk beneficiary does not 
submit pharmacy or prescriber preferences, section 1860-D-4(c)(5)(D)(i) 
of the Act provides that the Part D sponsor shall make the selection. 
Section 1860-D-4(c)(5)(D)(ii) of the Act further provides that, in 
making the selection, the sponsor shall ensure that the beneficiary 
continues to have reasonable access to frequently abused drugs, taking 
into account geographic location, beneficiary preference, the 
beneficiary's predominant usage of prescriber or pharmacy or both, 
impact on cost-sharing, and reasonable travel time. We proposed Sec.  
423.153(f)(11) to codify these statutory provisions.
    Since the statute explicitly allows the beneficiary to submit 
preferences, we interpreted the additional reference to beneficiary 
preference in the context of reasonable access to mean that a 
beneficiary allowable preference should prevail over a sponsor's 
evaluation of geographic location, the beneficiary's predominant usage 
of a prescriber and/or pharmacy impact on cost-sharing, and reasonable 
travel time. In the absence of a beneficiary preference for pharmacy 
and/or prescriber, however, a Part D plan sponsor must take into

[[Page 16471]]

account geographic location, the beneficiary's predominant usage of a 
prescriber and/or pharmacy, impact on cost-sharing, and reasonable time 
travel in selecting a pharmacy and/or prescriber, as applicable, from 
which the at-risk beneficiary will have to obtain frequently abused 
drugs under the plan. Thus, absent a beneficiary's allowable preference 
or plan recognition that the beneficiary's selection will contribute to 
prescription drug abuse or drug diversion, we proposed that the sponsor 
must ensure reasonable access by choosing the network pharmacy or 
prescriber that the beneficiary uses most frequently unless the plan is 
a stand-alone PDP and the selection involves a prescriber(s). In the 
latter case, the prescriber will not be a network provider, because 
such plans do not have provider networks. In urgent circumstances, we 
proposed that reasonable access means the sponsor must have reasonable 
policies and procedures in place to ensure beneficiary access to 
coverage of frequently abused drugs without a delay that may seriously 
jeopardize the life or health of the beneficiary or the beneficiary's 
ability to regain maximum function. We stated that determining 
reasonable access may be complicated when an enrollee has multiple 
addresses or his or her health care necessitates obtaining frequently 
abused drugs from more than one prescriber and/or more than one 
pharmacy. Sections 1860D-4(c)(5)(D)(ii)(I) and (II) address this issue 
by requiring the Part D plan sponsor to select more than one prescriber 
to prescribe frequently abused drugs and more than one pharmacy to 
dispense them, as applicable, when it reasonably determines it is 
necessary to do so to provide the at-risk beneficiary with reasonable 
access, which we proposed to codify at Sec.  423.153(f)(12). To address 
chain pharmacies and group practices, we proposed that in the case of a 
group practice, all prescribers of the group practice shall be treated 
as one prescriber and all locations of a pharmacy that share real-time 
electronic data should be treated as one pharmacy.
    We proposed to interpret these provisions to mean that a sponsor 
will be required to select more than one prescriber of frequently 
abused drugs, if more than one prescriber has asserted during case 
management that multiple prescribers of frequently abused drugs are 
medically necessary for the at-risk beneficiary.
    We received the following comments and our response follows:
    Comment: A commenter noted that the reasonable access provisions 
did not allow for situations where a patient who is locked-in is 
hospitalized or develops a new medical condition that requires they see 
a new physician, and that CMS should consider providing additional 
flexibility in such unexpected or unplanned situations.
    Response: We note that drugs dispensed during a hospitalization are 
covered under the Medicare Part A benefit. Aside from that, plans are 
required to provide reasonable access to at-risk beneficiaries in their 
drug management programs under proposed Sec.  423.153(f)(11). Proposed 
Sec.  423.153(f)(12) requires a Part D plan sponsor to select more than 
one prescriber to prescribe frequently abused drugs when it reasonably 
determines it is necessary to do so to provide the at-risk beneficiary 
with reasonable access. To the extent that a new health condition 
necessitates an at-risk beneficiary to change providers who prescribe 
frequently abused drugs rather than see more than one, the beneficiary 
can submit a new prescriber preference, as discussed earlier.
    With respect to a hospital emergency room visit, for example, we 
stated that in urgent circumstances, proposed Sec.  423.153(f)(11) 
requires a Part D sponsor to ensure an at-risk beneficiary has 
reasonable access in the case of emergency services, which we stated 
means that the sponsor must have reasonable policies and procedures in 
place to ensure beneficiary access to coverage of frequently abused 
drugs without a delay that may seriously jeopardize the life or health 
of the beneficiary or the beneficiary's ability to regain maximum 
function. Thus, we believe Sec.  423.153(f)(11) and (12) address the 
commenter's concerns.
    Comment: We received a comment requesting that group practices be 
permitted to designate one or more prescribers when a plan sponsor 
intends to limit a beneficiary's access to coverage of frequently 
abused drugs to a selected prescriber or prescribers at a group 
practice, and permit the group practice to modify such designation from 
time to time. The commenter stated that this requirement should apply 
whether or not the prescribers at the group practice are all associated 
with the same single Tax Identification Number (TIN).
    Response: Under the provision we proposed and are finalizing, all 
prescribers of a group practice are treated as one prescriber. A TIN is 
a mechanism that can assist Part D sponsors in identifying group 
practices, but as discussed earlier in the preamble, case management 
can also reveal the existence of a group practice that is prescribing 
frequently abused drugs to a beneficiary.
    Comment: We received several comments that recommended that CMS re-
evaluate its policy for determining chain pharmacies, as identification 
of which pharmacies share real-time data may be difficult in many 
situations, noting that sponsors do not have an effective way to manage 
such arrangements, and PBMs do not have the systems capabilities to 
discern if their systems are integrated and interchangeable. A 
commenter stated support for CMS' proposal as it relates to chain 
pharmacies, but noted that managing this option will be challenging 
absent additional instructions from CMS.
    Response: Section 1860D-4(c)(5)(D)(ii) of the Act states that with 
respect to a pharmacy that has multiple locations that share real-time 
electronic data, all such locations of the pharmacy shall collectively 
be treated as one pharmacy for purposes of an at-risk beneficiary's 
selection of pharmacies. Until such pharmacies can be determined 
through data, sponsors with drug management programs will have to 
ascertain such pharmacies through the case management and beneficiary 
notification processes. We therefore are finalizing this provision as 
proposed.
    Earlier in the preamble in responding to comments about prescriber 
agreement, we stated that in the case of prescriber lock-in, if a 
prescriber who has not agreed to this limitation insists that he or she 
must be able to continue to prescribe frequently abused drugs for the 
beneficiary, a plan sponsor may need to offer to lock-in the at-risk 
beneficiary to more than one prescriber to ensure reasonable access 
pursuant to Sec.  423.153(f)(12), for example, if the beneficiary has 
been obtaining opioids from one prescriber and benzodiazepines from 
another. Thus, we point out that in finalizing the drug management 
program regulations, we are not interpreting the reasonable access 
provisions to require a sponsor to select more than one prescriber, if 
more than one prescriber has asserted during case management that 
multiple prescribers of frequently abused drugs are medically necessary 
for the at-risk beneficiary but only to consider it in the context of 
the requirement to provide reasonable access. This should also be the 
sponsor's approach when a beneficiary submits a preference for more 
than one prescriber and/or more than one pharmacy as his or her 
preference.
    Also earlier in this preamble, we stated that an IHS pharmacy or 
provider may be the selected pharmacy or

[[Page 16472]]

prescriber for at-risk beneficiaries who are entitled to fill 
prescriptions from IHS, tribal, or Urban Indian (I/T/U) organization 
pharmacies and receive services through the IHS health system, and that 
they may go to such a pharmacy or prescriber pursuant to our reasonable 
access requirement, even if they are not in-network. Therefore, we are 
adding language to Sec.  423.153(f)(12) to address situations when the 
sponsor reasonably determines that the selection of an out-of-network 
prescriber or pharmacy is necessary to provide the beneficiary with 
reasonable access. This language also addresses our earlier comment 
that a stand-alone PDP or MA-PD does not have to accept a beneficiary's 
selection of a non-network pharmacy or prescriber, except as necessary 
to provide reasonable access.
    Given the foregoing, we therefore finalize as proposed the 
following at Sec.  423.153(f)(11), with a modification to include 
language that the sponsor must ensure reasonable access by taking into 
account ``all relevant factors, including but not limited to'' and to 
renumber for better clarity: Reasonable access. In making the 
selections under paragraph (f)(12) of this section, a Part D plan 
sponsor must ensure that the beneficiary continues to have reasonable 
access to frequently abused drugs, taking into account all relevant 
factors, including but not limited to: (i) Geographic location; (ii) 
Beneficiary preference; (iii) The beneficiary's predominant usage of a 
prescriber or pharmacy or both; (iv) The impact on cost-sharing; (v) 
Reasonable travel time; (vi) Whether the beneficiary has multiple 
residences; (vii) Natural disasters and similar situations; and (viii) 
The provision of emergency services.
    We are also finalizing with modification for the addition of 
language requiring the selection of an out-of-network prescriber or 
pharmacy if necessary at Sec.  423.153(f)(12). Paragraphs (f)(12)(i) 
and (ii) will specify the following:
     A Part D plan sponsor must select, as applicable--
    ++ One, or, if the sponsor reasonably determines it necessary to 
provide the beneficiary with reasonable access, more than one, network 
prescriber who is authorized to prescribe frequently abused drugs for 
the beneficiary, unless the plan is a stand-alone PDP, or the selection 
of an out-of-network provider is necessary; and
    ++ One, or, if the sponsor reasonably determines it necessary to 
provide the beneficiary with reasonable access, more than one, network 
pharmacy that may dispense such drugs to such beneficiary, unless the 
selection of an out-of-network pharmacy is necessary.
     For purposes of paragraph (f)(12) of Sec.  423.153, in the 
case of a--
    ++ Pharmacy that has multiple locations that share real-time 
electronic data, all such locations of the pharmacy shall collectively 
be treated as one pharmacy; and
    ++ Group practice, all prescribers of the group practice shall be 
treated as one prescriber.
(4) Confirmation of Pharmacy and Prescriber Selection (Sec.  
423.153(f)(13))
    Section 1860D-4(c)(5)(D)(v) of the Act requires that, before 
selecting a prescriber or pharmacy, a Part D plan sponsor must notify 
the prescriber and/or pharmacy that the at-risk beneficiary has been 
identified for inclusion in the drug management program, which will 
limit the beneficiary's access to coverage of frequently abused drugs 
to selected pharmacy(ies) and/or prescriber(s) and that the prescriber 
and/or pharmacy has been selected as a designated prescriber and/or 
pharmacy for the at-risk beneficiary. We proposed Sec.  423.153(f)(13) 
to codify this statutory requirement.
    We also proposed that plan sponsors must obtain the network 
prescriber's or pharmacy's confirmation that the selection is accepted 
before conveying this information to the at-risk beneficiary, unless 
the prescriber or pharmacy agreed in advance in its network agreement 
to accept all such selections and the agreement specifies how the 
prescriber and pharmacy will be notified of its selection. In these 
cases, the network provider would agree to forgo specific notification 
if selected under a drug management program to serve an at-risk 
beneficiary.
    We received the following comments and our responses follow:
    Comment: We received a comment that CMS should prohibit plan 
sponsors from including in their provider agreements any requirement 
that would require a prescriber to confirm in advance and forego 
specific confirmation, if selected under a drug management program to 
serve an at-risk beneficiary.
    Response: In light of this comment, and given the fact that we are 
finalizing a requirement for prescriber agreement for prescriber lock-
in, as discussed earlier in the preamble, we believe the appropriate 
approach is that the required prescriber agreement during case 
management satisfies the requirement that the plan sponsor notify the 
prescriber that the at-risk beneficiary has been identified for 
inclusion in a drug management program and the prescriber has been 
selected as a prescriber that the beneficiary will be locked into for 
purposes of frequently abused drugs. In our view, the process of 
obtaining the prescriber agreement to prescriber lock-in also serves as 
the receipt of confirmation from the prescriber, not to mention our 
requirement that the sponsor make reasonable efforts to provide the 
prescriber with a copy of the beneficiary notices that the sponsor must 
provide, discussed earlier. Such an approach reduces unnecessary 
repetition of communication with prescribers.
    For network pharmacies, this approach means that the notification 
that the at-risk beneficiary has been identified for inclusion in a 
drug management program and the pharmacy has been selected as a 
pharmacy that the beneficiary will be locked into for purposes of 
frequently abused drugs and the pharmacy's confirmation can be 
negotiated between the plan sponsor and the pharmacy, and if not, the 
plan sponsor must do so on a case-by-case basis, which is also the case 
for out-of-network prescribers and pharmacies.
    Comment: A commenter proposed an additional exception to the 
confirmation requirement for plan sponsors that own or operate their 
own pharmacies, arguing that such confirmation would be unnecessary 
given that the pharmacy would already be confirmed, as part of their 
integrated system.
    Response: We are not persuaded that an exception is needed in these 
situations. If the pharmacy is a separate legal entity from the plan 
sponsor, then the contract could contain a blanket agreement stating 
that the pharmacy agrees to accept at-risk beneficiaries that the plan 
sponsors locks into that pharmacy, as we mentioned in the proposed 
rule. If the pharmacy is the same legal entity as the plan sponsor, 
then notification is automatic, and no further notification or contract 
language would be necessary.
    Based on the comments and our responses, we are finalizing this 
provision with modifications to state the following regarding 
confirmation of selections(s):
     Before selecting a prescriber or pharmacy under this 
paragraph, a Part D plan sponsor must notify the prescriber or 
pharmacy, as applicable, that the beneficiary has been identified for 
inclusion in the drug management program for at-risk beneficiaries and 
that the prescriber or pharmacy or both is(are) being selected as the 
beneficiary's designated prescriber or pharmacy or both for frequently 
abused drugs. For prescribers, this notification occurs during case 
management as described in

[[Page 16473]]

paragraph (f)(2) or when the prescriber provides agreement pursuant to 
paragraph (f)(4)(i)(B).
     The sponsor must receive confirmation from the 
prescriber(s) or pharmacy(ies) or both, as applicable, that the 
selection is accepted before conveying this information to the at-risk 
beneficiary, unless the pharmacy has agreed in advance in a network 
agreement with the sponsor to accept all such selections and the 
agreement specifies how the pharmacy will be notified by the sponsor of 
its selection.
     A sponsor complies with paragraphs (i) and (ii) as it 
pertains to a prescriber by obtaining the prescriber's agreement 
pursuant to Sec.  423.153(f)(4)(i)(B).
(ix) Drug Management Program Appeals (Sec. Sec.  423.558, 423.560, 
423.562, 423.564, 423.580, 423.582, 423.584, 423.590, 423.602, 423.636, 
423.638, 423.1970, 423.2018, 423.2020, 423.2022, 423.2032, 423.2036, 
423.2038, 423.2046, 423.2056, 423.2062, 423.2122, and 423.2126)
    Section 1860D-4(c)(5)(E) of the Act specifies that the 
identification of an individual as an at-risk beneficiary for 
prescription drug abuse under a Part D drug management program, a 
coverage determination made under such a program, the selection of a 
prescriber or pharmacy, and information sharing for subsequent plan 
enrollments shall be subject to reconsideration and appeal under 
section 1860D-4(h) of the Act. This provision also permits the option 
of an automatic escalation to external review to the extent provided by 
the Secretary.
    As discussed earlier in this preamble, we proposed to integrate the 
lock-in provisions with existing Part D Opioid DUR Policy/OMS. 
Determinations made in accordance with any of those processes, at Sec.  
423.153(f), and discussed previously, are interrelated issues that we 
collectively refer to as an ``at-risk determination.'' In this final 
rule, we are adding a definition of at-risk determination at Sec.  
423.560 to describe a decision made under a plan sponsor's drug 
management program in accordance with Sec.  423.153(f) that involves 
the identification of an individual as an at-risk beneficiary for 
prescription drug abuse; a limitation, or the continuation of a 
limitation, on an at-risk beneficiary's access to coverage of 
frequently abused drugs (that is, a beneficiary specific point-of-sale 
edit the selection of a prescriber and/or pharmacy and implementation 
of lock-in); and information sharing for subsequent plan enrollments.
    We proposed that at-risk determinations made under the processes at 
Sec.  423.153(f) be adjudicated under the existing Part D benefit 
appeals process and timeframes set forth in Subpart M. Consistent with 
the existing Part D benefit appeals process, we proposed that at-risk 
beneficiaries (or an at-risk beneficiary's prescriber, on behalf of the 
at-risk beneficiary) must affirmatively request IRE review of adverse 
plan level appeal decisions made under a plan sponsor's drug management 
program. We also proposed to amend the existing Subpart M rules at 
Sec.  423.584 and Sec.  423.600 related to obtaining an expedited 
redetermination and IRE reconsideration, respectively, to apply them to 
appeals of an at-risk determination made under a drug management 
program. While we did not propose to adopt auto-escalation, the 
proposed approach ensures that an at-risk beneficiary has the right to 
obtain IRE review and higher levels of appeal (ALJ/attorney 
adjudicator, Council, and judicial review). Accordingly, we also 
proposed to add the reference to an ``at-risk determination'' to the 
following regulatory provisions that govern ALJ and Council processes: 
Sec. Sec.  423.2018, 423.2020, 423.2022, 423.2032, 423.2036, 423.2038, 
423.2046, 423.2056, 423.2062, 423.2122, and 423.2126.
    Finally, we also proposed a change to Sec.  423.1970(b) to address 
the calculation of the amount in controversy (AIC) for an ALJ hearing 
in cases involving at-risk determinations made under a drug management 
program in accordance with Sec.  423.153(f).
    In addition to the changes related to the implementation of drug 
management program appeals, we also proposed to make technical changes 
to Sec.  423.562(a)(1)(ii) to remove the comma after ``includes'' and 
replace the reference to ``Sec. Sec.  423.128(b)(7) and (d)(1)(iii)'' 
with a reference to ``Sec. Sec.  423.128(b)(7) and (d)(1)(iv).''
    We received the following comments and our responses follow:
    Comment: A few commenters strongly objected to beneficiaries not 
having appeal rights during their designation as ``potential'' at-risk 
beneficiaries at the time the initial notice is received from the plan 
sponsor.
    Response: As we noted in the proposed rule, when a beneficiary is 
identified as being potentially at-risk, but has not yet been 
definitively identified as at-risk, the plan is not taking any action 
to limit such beneficiary's access to frequently abused drugs. Because 
the plan sponsor has not taken any action to limit a beneficiary's 
access at this point in the process, the situation is not ripe for 
appeal. We proposed that a beneficiary will have the right to appeal a 
determination made under a plan sponsor's drug management program when 
the beneficiary receives the second notice explaining that access to 
coverage for frequently abused drugs will be limited. We believe the 
intent of the statute is to confer appeal rights to beneficiaries at 
the point in the process at which a beneficiary is notified that access 
will be limited and provide an explanation of the restrictions that 
will be applied under the drug management program.
    As discussed earlier in this preamble, the proposed 90 day maximum 
timeframe for the plan sponsor to send the second or alternate second 
notice is being reduced to 60 days under this final rule. Specifically, 
the second or alternate second notice is to be provided to the 
beneficiary no more than the earlier of the date the sponsor makes the 
relevant determination or 60 days after the date of the initial notice. 
This 60 day period may be used by a plan sponsor to process information 
received from beneficiaries or communicate with prescribers who may 
have been unresponsive prior to receiving a copy of the initial notice 
the plan provided to the beneficiary. As we also previously noted in 
this preamble, we do not expect plans to routinely take the maximum 
amount of time to issue the second notice, and note that the plan must 
send it sooner if they make the relevant determination sooner. Reducing 
this period between the initial notice and the second or alternate 
second notice to a maximum of 60 days balances plan sponsors' need for 
time to process information from beneficiaries and prescribers, if 
applicable, with providing timely notice to beneficiaries.
    Comment: Several commenters encouraged CMS to make the appeals 
process regarding lock-in as simple as possible for beneficiaries to 
ensure that those who need particular drugs are able to access them. 
These commenters suggested that CMS implement all of the protections of 
CARA, including automatic escalation to independent review. Several 
commenters do not agree with CMS' interpretation of the CARA language 
on appealing lock-in and believe automatic escalation to the IRE would 
ensure beneficiary due process and access to needed prescription drugs. 
These commenters strongly oppose the use of the existing Part D appeals 
process for appeals of at-risk status or other consequences of drug 
management, and view the process as a significant barrier that will 
increase the timeframe for the lock-in appeals process. Commenters 
expressed concerns regarding case management and physician agreement as 
additional hurdles for beneficiaries who are not at-

[[Page 16474]]

risk, in addition to plan compliance with the current requirements for 
timely appeals. A few commenters stated that CARA contemplates a more 
streamlined process that is easier for beneficiaries to navigate and 
that automatic escalation would allow for improved tracking and 
monitoring of the scope and impact of the lock-in program, in addition 
to providing more uniform decision making across various plan programs. 
A commenter suggested that CMS conduct analysis to determine which 
option would prevent or reduce bias against beneficiaries, as well as 
minimize the timeframe by which the review process occurs, and upon 
implementation closely monitor the decisions of at-risk status to 
ensure decisions are made in the best interest of the beneficiary. A 
commenter recommended a separate appeals process that is similar to the 
grievance process.
    Response: We agree with commenters that the appeals process for 
enrollees identified as at-risk should be as easy to navigate as 
possible. As we noted in the proposed rule, Part D enrollees, plan 
sponsors, and other stakeholders are already familiar with the Part D 
benefit appeals process. Resolving disputes that arise under a plan 
sponsor's drug management program within the existing Part D benefit 
appeals process is not only required by statute, but will allow at-risk 
beneficiaries to be more familiar with, and more easily access, the 
appeals process as opposed to creating a new process specific to 
appeals related to a drug management program. Since the statute 
specifically refers to section 1860D-4(h) of the Act and the process we 
proposed is consistent with the existing appeals process, we disagree 
with the comment that further analysis of options is necessary to 
``prevent or reduce bias against beneficiaries.'' As we noted in the 
proposed rule, affording a plan sponsor the opportunity to review its 
initial determination may result in resolution of the disputed issues 
at a lower level of review and obviate the need for further appeal of 
the issues to the Part D IRE which, in turn, will minimize the time for 
reviewing and resolving disputes. With respect to the monitoring of 
plan sponsors' at-risk decisions, appeal decisions involving at-risk 
status will be subject to review under existing plan sponsor audit 
processes. We do not believe that a process similar to the existing 
grievance process, as recommended by a commenter, would comport with 
the statute, which requires the use of the existing appeals process. 
However, potential at-risk and at-risk beneficiaries retain their 
existing right to file a grievance with the plan if they have 
complaints about the prescription drug management program.
    With respect to the comment on case management and physician 
involvement, these are key components to drug management programs and 
we disagree that these components create additional hurdles for 
beneficiaries within the appeals process. In fact, we believe that the 
extensive case management we expect to be performed under plan 
sponsors' drug management programs, including ongoing communications 
among the plan sponsor, enrollee, prescriber(s) and pharmacy, will 
result in a relatively low volume of appeals under these programs. In 
addition, the appeals that are processed will be informed by the case 
management conducted by the plan sponsor and the involvement of the 
physician.
    Comment: Many commenters agreed with the proposal to utilize the 
existing Part D appeals process for at-risk beneficiaries, including 
not requiring automatic escalation for external review. These 
commenters believed that use of the existing process is the simplest 
and most administratively efficient approach, as it is familiar to 
beneficiaries, plan sponsors, and other stakeholders. These commenters 
also believed that plan sponsors should have the opportunity to review 
additional information and potentially adjust their initial decision 
before the case is reviewed by the IRE.
    Response: We thank the commenters for expressing support for use of 
the existing Part D benefit appeals process for beneficiaries 
identified as at-risk under a plan sponsor's drug management program. 
In addition to comporting with the statutory requirement, we agree with 
the commenters that use of the existing appeals process is the most 
administratively efficient approach and will result in better outcomes 
for at-risk beneficiaries. Not only is the existing appeals process 
familiar to enrollees, plans, and the IRE, but it allows a plan sponsor 
the opportunity to review information it used to make an at-risk 
determination under its drug management program (and any additional 
relevant information submitted as part of the appeal), promotes the 
resolution of issues at a lower level of administrative review and 
potentially reduces the need for the beneficiary to further appeal the 
issues in dispute. However, if the matter is not resolved by the plan 
sponsor at the redetermination level, an at-risk beneficiary will have 
the right to seek review by the Part D IRE.
    Comment: With respect to the calculation of the amount in 
controversy (AIC) for an ALJ hearing or judicial review, a commenter 
expressed support for using a formula based on the value of any refills 
for frequently abused drugs to calculate the AIC, noting that it will 
provide a greater probability for higher review, benefiting both the 
plan and the beneficiary.
    Response: We thank the commenter for expressing support for the 
proposal related to calculation of the AIC at Sec.  423.1970(b)(2) for 
disputes related to identification as an at-risk beneficiary under a 
plan sponsor's drug management program.
    Comment: A few commenters requested clarification as to whether the 
beneficiary Notice of Appeal Rights (reject code 569), which triggers a 
pharmacy to provide the beneficiary with the standardized pharmacy 
notice, Prescription Drug Coverage and Your Rights (CMS-10147), should 
accompany any POS claim rejections regarding prescriber or pharmacy 
lock-in or beneficiary-specific POS edits. Commenters recommended that 
the CMS-10147 not be provided to beneficiaries when a claim rejects at 
POS due to issues under a plan sponsor's drug management program.
    Response: We agree with the commenters that a POS claim rejection 
as a result of a restriction imposed under a plan sponsor's drug 
management program should not trigger delivery of the standardized 
pharmacy notice (CMS-10147). The pharmacy notice informs a beneficiary 
to contact his or her Part D plan to request a coverage determination. 
As discussed above in this final rule, a determination under a plan 
sponsor's drug management program is not a coverage determination as 
defined at Sec.  423.566. Instead, a determination made under a drug 
management program is governed by the provisions proposed at Sec.  
423.153(f) related to at-risk determinations. If a beneficiary 
disagrees with a decision made under Sec.  423.153(f), the beneficiary 
has the right to appeal such decision. The at-risk beneficiary will be 
notified of this appeal right pursuant to the notice described at Sec.  
423.153(f)(6).
    Comment: Several commenters requested clarification that when a 
beneficiary appeals their coverage limitation under the drug management 
program, that the request should be processed as a redetermination and 
not as a coverage determination. A few commenters requested 
clarification as to whether or not the POS edit or a lock-

[[Page 16475]]

in would be a coverage determination. Commenters asked if Chapter 18 of 
the Prescription Drug Benefit Manual would apply, and if so, noted that 
CMS should release proposed changes to the guidance for comment. 
Commenters inquired about how the CARA provisions would impact the 
coverage determination and redetermination processes, including 
approval and denial language used by plan sponsors. A commenter stated 
that they do not believe that these are coverage determinations because 
they involve access issues and being treated as such would pose system, 
policy, and process challenges. This commenter also asked for 
clarification on how this process would impact the appeals auto-forward 
star measure if treated as a coverage determination.
    Response: We did not propose to change the current definition of a 
coverage determination at Sec.  423.566. As we stated in the proposed 
rule, the types of decisions made under a drug management program align 
more closely with the regulatory provisions in Subpart D than with the 
provisions in Subpart M. We believe it is clearer to set forth the 
rules for at-risk determinations as part of Sec.  423.153 and cross 
reference Sec.  423.153(f) in relevant appeals provisions in Subpart M 
and Subpart U. The types of initial determinations made under a drug 
management program (for example, a restriction on the at-risk 
beneficiary's access to coverage of frequently abused drugs to those 
that are prescribed for the beneficiary by one or more prescribers) 
will be subject to the processes proposed at Sec.  423.153(f).
    What we did propose is that at-risk determinations made under the 
processes at Sec.  423.153(f) be adjudicated under the existing Part D 
benefit appeals process and timeframes set forth in Subpart M. Thus, we 
agree with these commenters that a determination made under a drug 
sponsor's drug management program should not be considered a coverage 
determination as defined at Sec.  423.566. If a beneficiary has a 
dispute related to a determination under the processes set forth at 
Sec.  423.153(f), the beneficiary has the right to request a 
redetermination and potentially higher levels of appeal. Therefore, 
drug management program disputes are subject to the appeals provisions 
in Subpart M and Subpart U of the regulations and the guidance in 
Chapter 18 of the Prescription Drug Benefit Manual also applies. 
Disputes under a plan sponsor's drug management program will be 
adjudicated under the existing appeals process and the regulatory 
timeframes will apply. The manual guidance will be updated, as 
necessary, to reflect any changes relevant to drug management program 
disputes. With respect to the redetermination notice, plan sponsors may 
use CMS' model redetermination notice (with modifications) or develop 
their own notice for informing an enrollee of the outcome of the 
appeal.
    Comment: A few commenters suggested that these appeals be limited 
to the beneficiary-level edit, the selected pharmacy or the prescriber, 
and not the underlying criteria for identification and guidance. 
Commenters noted that the appeal should be limited to the issue of 
whether the beneficiary is an appropriate candidate for lock-in, and 
not have any other scope. A commenter stated that the appeal should not 
relate to whether the plan may impose prior authorization or other 
utilization management restrictions on certain prescriptions. Rather, 
according to the commenter, beneficiary appeals should be limited to 
compliance with internal program criteria and CMS guidance, rather than 
allowing beneficiaries to challenge the underlying criteria. A 
commenter asked that CMS clarify how to effectuate a redetermination 
that requires the reversal of one limit, but other limits remain (for 
example, a formulary restriction and lock-in), and which limit takes 
priority. This commenter stated that beneficiaries would have to 
receive decision notices explaining that because of the remaining 
limits, their drug access will continue to be limited. Another 
commenter requested guidance on whether to handle a dispute involving 
beneficiary-specific POS claim edit and a dispute about a pharmacy or 
prescriber selection under the same appeal, or the POS edit as a 
coverage determination and the lock-in as an appeal.
    Response: As explained above, the statute explicitly states that 
one of the issues that can be appealed is the identification as an at-
risk beneficiary for prescription drug abuse under a Part D drug 
management program. With respect to the comment that an enrollee not be 
permitted to challenge the ``underlying criteria,'' we interpret this 
to mean a plan sponsor's clinical guidelines used to identify potential 
at-risk beneficiaries. We believe that a beneficiary disputing his or 
her at-risk determination will inherently be arguing that the plan's 
criteria for identifying at-risk beneficiaries do not apply to his or 
her particular circumstances. In addition to the at-risk determination, 
an enrollee has the right under the statute to appeal the selection of 
a prescriber or pharmacy as well as a coverage determination made under 
a plan sponsor's drug management program. As previously noted, 
determinations made under the processes at Sec.  423.153(f) will be 
adjudicated under the existing Part D benefit appeals process. Such 
determinations include limitation on access to coverage for frequently 
abused drugs, including a POS claim edit for frequently abused drugs 
that is specific to an at-risk beneficiary and a limit on an at-risk 
beneficiary's access to coverage for frequently abused drugs to those 
that are prescribed by one or more prescribers or dispensed to the 
beneficiary by one or more network pharmacies. As also previously 
noted, we did not propose to revise the existing definition of a 
coverage determination. In addition to a determination made under the 
processes at Sec.  423.153(f), a coverage determination, including an 
exception, is also subject to appeal. For example, if an enrollee does 
not dispute a POS edit for a quantity limit on a drug within 60 days of 
the date of the second notice pursuant to Sec.  423.153(f)(6) but later 
requests an exception to the quantity limit and that request is denied 
by the plan sponsor, the enrollee has the right to appeal the denial of 
the exception request. While the enrollee always has the right to 
request a coverage determination, changes to previously imposed 
limitations can also be implemented through ongoing case management and 
a new determination under the processes at Sec.  423.153(f).
    As noted earlier, a commenter asked whether a dispute regarding 
pharmacy or prescriber selection for purposes of lock-in and a dispute 
related to a beneficiary specific POS claim edit should be processed as 
the same appeal. If a beneficiary's request for an appeal raises 
multiple issues related to the limitations imposed on the beneficiary 
under a drug management program, the plan sponsor must address each 
issue as part of the appeal. For example, if the beneficiary's appeal 
request includes a dispute related to pharmacy selection and a POS 
edit, the adjudication and disposition of the appeal would involve both 
issues. All disputes raised in the enrollee's appeal request that arise 
under a plan's drug management program will be adjudicated as a single 
case. Assuming the request is filed timely, an enrollee could later 
appeal another limitation imposed under the drug management program, 
such as the selection of a prescriber, and the adjudication and 
disposition of that appeal would relate to prescriber selection for 
purposes of lock-in and be considered separate and distinct from any 
previous or pending appeal

[[Page 16476]]

requests. An appeal request must be filed within 60 calendar days from 
the date of the notice that explains the limitations imposed under the 
drug management program (unless there is good cause for late filing of 
the appeal). In addition to appealing determinations made under the 
processes at Sec.  423.153(f) that limit a beneficiary's access, a 
beneficiary who is subject to a Part D plan sponsor's drug management 
program always retains the right to request a coverage determination 
under existing Sec.  423.566 for any Part D drug that the beneficiary 
believes may be covered by their plan.
    With respect to effectuation of a redetermination of an at-risk 
determination, we agree with the commenter that the redetermination 
notice should clearly explain which aspect of the program is changing 
(for example, change in pharmacy lock-in) and which restrictions remain 
unchanged and will continue to apply to the beneficiary. We would like 
to clarify that all changes must be effectuated pursuant to the 
effectuation rules at Sec.  423.636 and Sec.  423.638; in other words, 
one change does not take ``priority'' over another applicable change 
with respect to effectuation. For example, if the outcome of a standard 
redetermination related to pharmacy and prescriber lock-in is a change 
to the pharmacy and the prescriber(s) an at-risk enrollee must use, the 
plan sponsor must implement both of those changes concurrently and as 
expeditiously as the enrollee's health condition requires, but no later 
than 7 calendar days from the date the plan sponsor receives the 
redetermination request.
    Comment: A few commenters suggested that CMS confirm that a 
beneficiary should not continue to receive inappropriate fills of 
opioids during the appeals process.
    Response: We thank the commenters for their request for 
confirmation that a beneficiary who has been identified as at-risk, has 
received the second notice, and has requested an appeal should not 
continue to receive ``inappropriate fills'' of opioids during the 
appeals process. We are interpreting ``inappropriate fills'' to mean a 
fill that does not comport with the specific restrictions placed on the 
at-risk beneficiary (for example, pharmacy lock-in). Once the 
beneficiary has been notified via the second notice of applicable 
restrictions, there should be no additional fills of any of the drug(s) 
subject to the drug management program that do not satisfy the 
parameters of the program established for the at-risk beneficiary, 
unless those restrictions are later modified through the appeals 
process.
    Comment: A commenter asked that CMS clarify whether these appeals 
are required to be handled based on the timeframes for a request for 
benefit or a request for payment, and whether or not these are subject 
to the expedited timeframes.
    Response: As noted in the proposed rule, at-risk determinations 
made under the processes at Sec.  423.153(f) would be adjudicated under 
the existing Part D benefit appeals process and timeframes set forth in 
Subpart M and Subpart U. As such, at-risk determinations will be 
subject to the benefit request timeframes set forth at Sec.  
423.590(a). We also proposed to amend the existing Subpart M rules at 
Sec.  423.584 and Sec.  423.600 related to obtaining an expedited 
redetermination and IRE reconsideration, respectively, to apply them to 
appeals of a determination made under a drug management program. 
Consistent with existing rules, the beneficiary must meet the 
requirements set forth in regulation in order to obtain an expedited 
review of their at-risk determination.
    Comment: In the case of a beneficiary appealing the Part D plan 
sponsor's initial selection of a prescriber or pharmacy, a commenter 
requested clarification whether the plan sponsor must obtain 
confirmation of acceptance from the new prescriber and/or pharmacy the 
beneficiary has selected as part of the appeal and whether this 
confirmation needs to be made within the appeals timeframes. This 
commenter expressed concern with obtaining such confirmation within the 
short window for adjudicating the case.
    Response: While we appreciate the commenter's concern regarding the 
timeframe for making a decision, we believe that the current timeframes 
afford the plan sponsor sufficient time to obtain confirmation from a 
prescriber and/or pharmacy that they have accepted the beneficiary's 
selection for lock-in. Under the current Part D benefit appeals 
process, plan sponsors are required to obtain similar information from 
prescribers and we believe that appeals of at-risk determinations 
should not be materially different from the outreach plans conduct as 
part of the coverage determination, exceptions, and benefits appeals 
process. Please refer to the discussion regarding confirmation of 
pharmacy and prescriber selection earlier in this preamble.
    Comment: A few commenters requested clarification as to whether or 
not plans would be permitted to terminate exceptions or implement 
temporary exceptions, in consultation with the prescriber, prior to the 
end of a plan year due to opioid case management and, if so, what prior 
notice requirements will apply.
    Response: Consistent with existing rules for the exceptions process 
at Sec.  423.578(c), if a drug is found to no longer be safe for the 
enrollee, then a previously approved exception request could be 
terminated prior to the end of the plan year. This would include if the 
plan determines that the previously approved exception is no longer 
safe as part of an at-risk determination or ongoing case management 
under its drug management program. A determination made by a plan 
sponsor under the processes at Sec.  423.153(f) is subject to appeal. 
For example, if a determination is made under a plan sponsor's drug 
management program to implement a beneficiary-specific POS claim edit 
for a drug, the beneficiary will be notified of that decision per the 
provisions at Sec.  423.153(f)(6) and the decision may be appealed. If 
the beneficiary does not appeal the decision within 60 calendar days 
from the date of the notice that explains the limitations the plan 
sponsor is placing on the beneficiary's access to coverage for 
frequently abused drugs, the beneficiary retains the right to request a 
coverage determination related to a beneficiary-specific POS edit at 
any time. And, as stated above, changes to previously imposed 
limitations can also be implemented through ongoing case management and 
a new determination under the processes at Sec.  423.153(f).
    Comment: A few commenters expressed concern regarding the lack of 
any proposed review criteria that would be used by plans to evaluate 
these appeals based on the at-risk determination. Commenters stated 
that appeal requests for opioid restrictions do not fit in any existing 
utilization management criteria (for example formulary and tiering 
exceptions criteria) and request additional guidance from CMS. These 
commenters are concerned that if the beneficiary appeals the limitation 
beyond the plan, the IRE or ALJ/attorney adjudicator will likely review 
these restrictions similar to a formulary or tiering exception and not 
based on the at-risk determination. A commenter indicated that this 
type of review may have an adverse impact on plans' D03 STARS Ratings, 
and if approved, an exception must be effectuated through the end of 
the plan year, which could remove the enrollee from case management for 
the rest of the year even if they meet the criteria for such.
    Response: We appreciate the commenters' concerns. If the case goes 
to the IRE, or higher levels of appeal, the administrative case file 
assembled by

[[Page 16477]]

the plan sponsor will contain the relevant information needed by the 
adjudicator to make an informed decision, such as information used by 
the plan sponsor to determine at-risk status, a description of the case 
management the plan has performed and the beneficiary's preference with 
respect to prescriber or pharmacy lock-in. We believe the regulations, 
applicable manual guidance, the plan sponsor's review criteria and case 
management notes on the access limitations that apply to the enrollee 
(which would be included in the administrative case file) will be 
sufficient for an adjudicator to review an appeal. With respect to the 
comment on an approved exception, please refer to the introductory 
section on drug management programs earlier in this preamble for a 
discussion of determinations where continuing an approved exception is 
no longer appropriate.
    Comment: With respect to the handling and reporting of appeals, a 
few commenters expressed concerns regarding the negative impact 
choosing to implement the lock-in procedures could potentially have on 
a plan. A commenter noted that opioid restriction reviews are not 
represented in their reporting and there are no allowable values in the 
audit universes that would designate a case as an opioid restriction. 
As a result, the commenter believes that if an approved exception is 
terminated prior to the end of the plan year, this could be detected on 
audit and the plan sponsor may be found to be non-compliant with 
exception processing requirements.
    Response: If a plan sponsor makes a determination under its drug 
management program per the processes at Sec.  423.153(f) that results 
in a finding that a drug previously approved through the exception 
process is found to no longer be safe for treating the beneficiary's 
disease or medical condition, the previously approved exception can be 
terminated prior to the end of the plan year. With respect to the 
commenter's concern about such a case being reviewed on audit, the plan 
sponsor would not be subject to a finding of non-compliance for having 
terminated a previously authorized exception if such termination is 
consistent with a clinically appropriate determination made under the 
plan sponsor's drug management program.
    Comment: A few commenters encourage CMS to communicate appeal-
related information and requirements in a clear, concise, and 
consistent manner to beneficiaries, the IRE, and plan sponsors to 
support a uniform understanding of the agency's rules and related 
expectations. A commenter stated that beneficiaries are not always 
aware of their exceptions and appeal rights and many do not understand 
how the process works. This commenter expressed concern that there may 
be a lack of transparency in the appeals process or excessive 
administrative burden for the beneficiary and provider, which may 
extend to those who may be inappropriately identified as at-risk and 
subject to unnecessary access restrictions to needed medications.
    Response: We agree with the commenters that appeals-related 
information and requirements should be communicated in a clear, 
concise, and consistent manner to beneficiaries, Part D plan sponsors, 
and the IRE. We will continue to update existing materials and develop 
new CARA related communications, such as the first and second notices 
described elsewhere in this final rule, with these goals in mind.
    After consideration of these comments, we are finalizing with 
modifications the provisions on CARA appeals with two clarifying 
changes. First, in this final rule, we are including a definition of 
at-risk determination to Sec.  423.560 to clarify the types of actions 
made under the processes at Sec.  423.153(f) that are subject to 
appeal. In addition to coverage determinations made under a drug 
management program, an enrollee has the right to appeal the 
identification as an at-risk beneficiary for prescription drug abuse; a 
beneficiary specific point-of-sale (POS) edit; the selection of a 
prescriber or pharmacy for purposes of lock-in; and information sharing 
for subsequent plan enrollments. Second, proposed new paragraph 
(a)(1)(v) at Sec.  423.562 has been revised to clarify that 
determinations made in accordance with the processes at Sec.  
423.153(f) are collectively referred to as an at-risk determination as 
defined at Sec.  423.560.
    Finally, we did not receive comments on the technical changes to 
Sec.  423.562(a)(1)(ii) and we are finalizing those changes as 
proposed.
(x) Termination of a Beneficiary's Potential At-Risk or At-Risk Status 
(Sec.  423.153(f)(14))
    Section 1860-D-4(c)(5)(F) of the Act provides that the Secretary 
shall develop standards for the termination of the identification of an 
individual as an at-risk beneficiary, which shall be the earlier of the 
date the individual demonstrates that he or she is no longer likely to 
be an at-risk beneficiary in the absence of limitations, or the end of 
such maximum period as the Secretary may specify.
    We proposed a maximum 12-month period for both a lock-in period, 
and also for the duration of a beneficiary-specific POS claim edit for 
frequently abused drugs. However, we also noted that if the sponsor 
implements an additional, overlapping limitation on the at-risk 
beneficiary's access to coverage for frequently abused drugs, the 
beneficiary may experience a coverage limitation beyond 12-months. The 
same is true for at-risk beneficiaries who were identified as such in 
the most recent prescription drug plan in which they were enrolled and 
the sponsor of their subsequent plan immediately implements a 
limitation on coverage of frequently abused drugs.
    Section 1860-D-4(c)(5)(F)(ii) of the Act states that nothing in 
CARA shall be construed as preventing a plan from identifying an 
individual as an at-risk beneficiary after such termination on the 
basis of additional information on drug use occurring after the date of 
notice of such termination. Accordingly, termination of an at-risk 
determination will not prevent an at-risk beneficiary from being 
subsequently identified as a potential at-risk beneficiary and an at-
risk beneficiary on the basis of new information on drug use occurring 
after the date of such termination that causes the beneficiary to once 
again meet the clinical guidelines.
    We received the following comments and our response follows:
    Comment: We received widespread comments that suggested that a 
maximum 12-month lock-in period was arbitrary, and that automatic 
termination of a beneficiary's at-risk status after 12 months threatens 
beneficiary safety. Commenters suggested that termination of such 
programs should be based on the needs of the beneficiary following a 
clinical assessment, and that an arbitrary time limit assumes without 
any clinical justification that he or she is no longer at-risk for drug 
abuse after 12 months. Following this period, many commenters also 
recommended plan sponsors should be permitted to conduct a review of 
the beneficiary's at-risk status at the expiration of the first 12 
months whether a beneficiary is determined at-risk, and if so, 
implement a termination after an additional 12 months, for 24 months 
total. While very few commenters supported the 12-month limitation 
timeframe, they did not provide rationale for their support.
    Response: We disagree with commenters that the 12-month period 
lock-in period we proposed is arbitrary. As we noted in the proposed 
rule, during the Stakeholder Listening Session on CARA held in November

[[Page 16478]]

2016, most commenters recommended a maximum 12-month period for lock-
in. We also noted that a 12-month lock-in period is common in Medicaid 
lock-in programs.\16\ Additionally, Section 1860D-4(c)(5)(F) grants the 
Secretary the authority to establish a maximum limitation period, and 
we choose to exercise said authority.
---------------------------------------------------------------------------

    \16\ Medicaid Drug Utilization Review State Comparison/Summary 
Report FFY 2015 Annual Report: Prescription Drug Fee-For Service 
Program (December 2016).
---------------------------------------------------------------------------

    CMS was, however, persuaded that a 12-month limitation maximum 
might be too short to ensure for beneficiary safety in some instances, 
and a longer limitation on access to coverage for frequently abuse 
drugs might be needed in such cases. We also re-reviewed limitation 
periods in Medicaid lock-in programs, and found that another very 
common lock-in period is 24 months. An additional prevalent trend for 
Medicaid lock-in periods is the ability to extend the lock-in period 
based on a review of appropriateness of continuance of lock-in.\17\ 
This trend aligned very closely with the many commenters who suggested 
a 24-month limitation period, and/or the ability of the plan sponsor to 
extend the limitation as a result of a clinical assessment. As a 
compromise between these two options, CMS is finalizing an initial 12-
month limitation period as proposed, but with ability modification 
allowing for the sponsor to extend the limitation for up to an 
additional 12 months. This extension will be dependent upon a clinical 
assessment whether the beneficiary demonstrates that they are no longer 
likely, in the absence of the limitation(s) the plan sponsor has placed 
on their access to coverage for frequently abused drugs, to be an at-
risk beneficiary for prescription drug abuse at the conclusion of the 
initial 12 months of the limitation. Thus, the maximum limitation 
period will be 24 months.
---------------------------------------------------------------------------

    \17\ Medicaid Drug Utilization Review State Comparison/Summary 
Report FFY 2016 Annual Report: Prescription Drug Fee-For Service 
Program (October 2017).
---------------------------------------------------------------------------

    Based on the provisions discussed earlier regarding when prescriber 
agreement is required, we believe the plan sponsor must, as part of the 
required clinical assessment, obtain prescriber agreement to extend a 
prescriber lock-in beyond the initial 12 months. Prescriber agreement 
will also be required with respect to extending beneficiary-specific 
POS edits. However, as with the initial POS edit, one can be extended 
without prescriber agreement if no prescriber is responsive. Also, the 
plan sponsor will be required to send the at-risk beneficiary another 
second notice, indicating that the limitation is being extended, and 
that they continue to be considered as an at-risk beneficiary. Aside 
from the required prescriber agreement just described, a plan sponsor 
will have discretion as to how they clinically assess whether an at-
risk beneficiary's demonstrates whether they are no longer likely to be 
an at-risk beneficiary for prescription drug abuse in the absence of 
limitation at the conclusion of the initial 12 months of the 
limitation. This assessment might include a review of medical records 
or prescription drug monitoring program data, if available to the 
sponsor. Given that the plan sponsor will not be required to obtain 
prescriber agreement to extend pharmacy lock-in past the initial 12 
month period, we expect the plan sponsor to have a clinical basis to 
extend the limitation, such as, the plan sponsor has recently rejected 
claims for frequently abused drugs from non-selected pharmacies to an 
extent that indicates the beneficiary may abuse frequently abused drugs 
without the limitation.
    Comment: A handful of commenters suggested that a limitation to 
coverage for frequently abused drugs only be terminated as a result of 
a clinical assessment by the at-risk beneficiary's prescriber with no 
maximum limitation period.
    Response: CMS believes it advisable to place a time limit on the 
duration of a limitation on access to coverage for frequently abused 
drugs that a plan sponsor can place on an at-risk beneficiary in order 
to balance the beneficiary's right to utilize their Part D benefit 
without encumbrance against with the sponsor's responsibility to manage 
the Part D benefit and promote the safety of its enrollees.
    Comment: A commenter suggested that CMS could consider requiring 
Part D sponsors to send annual notifications to beneficiaries who are 
subjected to a lock-in and their approving prescribers to let them know 
the lock-in will be extended another 12 months. This would afford 
beneficiaries and prescribers an annual opportunity to request that the 
lock-in be reconsidered or raise any concerns.
    Response: We decline to adopt this suggestion, as it does not 
suggest a basis upon which the limitation would be extended. Under the 
provision we are finalizing, a clinical assessment is required and, if 
the limitation on access to coverage is extended beyond the initial 12 
month period, the plan sponsor would be required to send the at-risk 
beneficiary an additional second notice pursuant to Sec.  423.153(f)(6) 
explaining that the limitation is being extended and for how long.
    Also, a beneficiary, their representative, or their prescriber on 
behalf of the beneficiary, is not precluded from requesting that the 
plan revisit its determination that the beneficiary is an at-risk 
beneficiary as defined at Sec.  423.100, or the terms of any limitation 
imposed on the beneficiary under the plan's drug management program.
    Based on these comments and our responses, we are therefore 
finalizing additional language at Sec.  423.153(f)(14). The revised 
language will specify that the identification of an at-risk beneficiary 
as such must terminate as of the earlier of the following:
     The date the beneficiary demonstrates through a subsequent 
determination, including but not limited to, a successful appeal, that 
the beneficiary is no longer likely, in the absence of the limitation 
under this paragraph, to be an at-risk beneficiary; or
     The end of a--
    ++ One year period calculated from the effective date of the 
limitation, as specified in the notice provided under paragraph (f)(6) 
of this section, unless the limitation was extended pursuant to 
paragraph (f)(14)(ii)(B) of this section.
    ++ Two year period calculated from the effective date of the 
limitation, as specified in a notice provided under paragraph (f)(6) of 
this section, subject to the following requirements:

--The plan sponsor determines at the end of the one year period that 
there is a clinical basis to extend the limitation.
--Except in the case of a pharmacy limitation imposed pursuant to 
paragraph (f)(3)(ii)(B) of this section, the plan sponsor has obtained 
the agreement of a prescriber of frequently abused drugs for the 
beneficiary that the limitation should be extended.
--The plan sponsor has provided another notice to the beneficiary in 
compliance with paragraph (f)(6) of this section.
--If the prescribers were not responsive after 3 attempts by the 
sponsor to contact them within 10 business days, then the sponsor has 
met the requirement of paragraph (f)(14)((ii)(B)(2) of this section.
--The sponsor may not extend a prescriber limitation implemented 
pursuant to paragraph (f)(3)(ii)(A) of this section if no prescriber 
was responsive.

[[Page 16479]]

(xi) Data Disclosure and Sharing of Information for Subsequent Sponsor 
Enrollments (Sec.  423.153(f)(15))
    In order for Part D sponsors to conduct the case management/
clinical contact/prescriber verification pursuant to Sec.  
423.153(f)(2), certain data disclosure and sharing of information must 
happen. First, CMS must identify potential at-risk beneficiaries to 
sponsors who are in the sponsors' Part D prescription drug benefit 
plans. In addition, a new sponsor must have information about potential 
at-risk beneficiaries and at-risk beneficiaries who were so identified 
by their immediately prior plan and enroll in the new sponsor's plan 
and such identification had not terminated before the beneficiary 
disenrolled from the immediately prior plan. Finally, as discussed 
earlier, sponsors may identify potential at-risk beneficiaries by their 
own application of the clinical guidelines (that is, applying the 
minimum clinical guidelines more frequently or in applying the 
supplemental clinical guidelines). It is important that CMS be aware of 
which Part D beneficiaries sponsors identify on their own, as well as 
which ones have been subjected to limitations on their access to 
coverage for frequently abused drugs under sponsors' drug management 
programs for Part D program administration and other purposes.
    Regarding data disclosures, section 1860D-4(c)(5)(H) of the Act 
provides that, in the case of potential at-risk beneficiaries and at-
risk beneficiaries, the Secretary shall establish rules and procedures 
to require the Part D plan sponsor to disclose data, including any 
necessary individually identifiable health information, in a form and 
manner specified by the Secretary, about the decision to impose such 
limitations and the limitations imposed by the sponsor under this part. 
We plan to expand and modify the scope of OMS and the MARx system as 
appropriate to accommodate the data disclosures necessary to oversee 
and facilitate Part D drug management programs.
    Section 1860-D-4(c)(5)(I) of the Act requires that the Secretary 
establish procedures under which Part D sponsors must share information 
when at-risk beneficiaries or potential at-risk beneficiaries enrolled 
in one prescription drug plan subsequently disenroll and enroll in 
another prescription drug plan offered by the next sponsor (gaining 
sponsor). We plan to expand the scope of the reporting to MARx under 
the current policy to include the ability for sponsors to report 
similar information to MARx about all pending, implemented, and 
terminated limitations on access to coverage of frequently abused drugs 
associated with their plans' drug management programs.
    We proposed to codify the data disclosure and information sharing 
process under the current policy, with the expansion just described, by 
adding data disclosure requirements in Sec.  423.153.
    We received the following comments and our response follows:
    Comment: We received comments supportive of our proposal regarding 
data disclosures and sharing of information. We did not receive 
comments opposed to our proposal.
    Response: We thank the commenters for their support.
    Comment: A commenter recommended that we clarify sponsors must 
conduct case management with respect to potential at-risk beneficiaries 
who are current utilizers under the Part D sponsor and not such 
beneficiaries who are identified by the prior sponsor. This commenter 
stated further that if sponsors are required to conduct case management 
on potential at-risk beneficiaries identified by the prior sponsor, 
then the response due date should be extended for such cases (that is, 
to next OMS quarter), as sponsors may need to contact the prior sponsor 
for case details to conduct case management for the prior claims data. 
In extending the outlier response due date, this commenter urged us to 
consider that the volume of such cases may differ based on the size of 
the prior sponsor.
    Response: Pursuant to Sec.  423.153(f)(2)(i), sponsors are required 
to conduct case management with respect to all potential at-risk 
beneficiaries who are identified by CMS or the sponsor applying the 
clinical guidelines, regardless of whether the beneficiary meets the 
clinical guidelines based on PDE data from the beneficiary's current 
Part D contract alone or across multiple contracts (including contracts 
the beneficiary was previously enrolled in during the measurement 
period).
    Sec.  423.153(f)(2)(ii) does provide an exception to the case 
management requirements with respect to potential at-risk beneficiaries 
identified as such by their most recent prior plan, if the 
identification has not been terminated and the sponsor obtains case 
management information from the previous sponsor, which is clinically 
adequate and up to date. Under the current policy, a sponsor may report 
in OMS that a beneficiary's case is under review. We plan to keep this 
response. However, because of this comment, we realize that there may 
be some instances in which a sponsor receives notice about a potential 
at-risk beneficiary who has just enrolled in its plan, but the deadline 
to provide information to CMS within 30 days from the date of the most 
recent prior CMS report identifying potential at-risk beneficiaries 
pursuant to proposed Sec.  423.153(f)(15) might be very short. 
Therefore, we are modifying Sec.  423.153(f)(15) such that the sponsor 
would have to provide the information within 30 days from the date of 
the most recent CMS report received after receiving such a notice.
    Comment: We received a comment requesting clarity on the issue of 
patient consent in the sharing of the patient personal health 
information related to implementation of these finalized provisions.
    Response: While the commenter's concerns about sharing personal 
health information are not entirely clear, we note that Part D plan 
sponsors are required under Sec.  423.136 to establish procedures for 
maintenance and sharing of medical records and other health information 
about enrollees in accordance with all applicable Federal and State 
confidentiality laws.
    Comment: We received a question asking what data sources we will 
use to identify LIS beneficiaries who are potentially at-risk.
    Response: We plan to use OMS to identify all potential at-risk 
beneficiaries who meet the minimum criteria of the clinical guidelines, 
discussed earlier, to report to Part D plan sponsors. We will modify 
the OMS as appropriate to implement the Part drug management program 
requirements. We will issue guidance and updated OMS technical user 
guides to plan sponsors at a later time, including data sources used in 
OMS reporting.
    Comment: We received a question whether the original plan that 
identified the beneficiary's at-risk status has a duty to inform the 
new plan of individual's status.
    Response: Plan sponsors will be required to communicate 
beneficiaries' potential and at-risk statuses to each other through the 
data disclosures and information sharing we are finalizing in this 
section.
    Comment: We received a question whether we will be providing new 
response codes for pharmacy and prescriber lock-in in OMS, specifically 
whether we will eliminate the response code ``BSC'' which stands for 
``Beneficiary did not meet sponsor's internal criteria.'' We also 
received some specific suggestions to: (1) Include responses to OMS 
that differentiate between lock-in and a claim edit at POS;

[[Page 16480]]

(2) add a sponsor summary page to OMS; (3) make enhancements to MARx to 
recognize internal and external contract changes; and (4) allow for 
more complete case management information to be shared to obviate the 
needs for sponsors to contact each other.
    Response: We appreciate these suggestions. We plan to expand and 
modify the scope of OMS and MARx as appropriate and technically 
possible in light of the final requirements in this rule to accommodate 
the data disclosures necessary to oversee and facilitate Part D drug 
management programs. We plan to issue guidance about this expansion and 
details on the modifications. Based on these comments, we are 
finalizing Sec.  423.153(f)(15) with modifications to specify the 
following regarding data disclosure:
     CMS identifies potential at-risk beneficiaries to the 
sponsor of the prescription drug plan in which the beneficiary is 
enrolled.
     A Part D sponsor that operates a drug management program 
must disclose any data and information to CMS and other Part D sponsors 
that CMS deems necessary to oversee Part D drug management programs at 
a time, and in a form and manner, specified by CMS. The data and 
information disclosures must do all of the following:
    ++ Provide information to CMS within 30 days of receiving a report 
about a potential at-risk beneficiary from CMS.
    ++ Provide information to CMS about any potential at-risk 
beneficiary that meets paragraph (1) of the definition in Sec.  423.100 
that a sponsor identifies within 30 days from the date of the most 
recent CMS report identifying potential at-risk beneficiaries.
    ++ Provide information to CMS about any potential at-risk 
beneficiary that meets paragraph (2) of the definition in Sec.  423.100 
within 30 days of the date after which the sponsor referred to in 
paragraph (2).
    ++ Provide information to CMS as soon as possible but no later than 
7 days of the date of the initial notice or second notice that the 
sponsor provided to a beneficiary, or as soon as possible but no later 
than 7 days of a termination date, as applicable, about a beneficiary-
specific opioid claim edit or a limitation on access to coverage for 
frequently abused drugs.
    ++ Transfer case management information upon request of a gaining 
sponsor as soon as possible but no later than 2 weeks from the gaining 
sponsor's request when--

--An at-risk beneficiary or potential at-risk beneficiary disenrolls 
from the sponsor's plan and enrolls in another prescription drug plan 
offered by the gaining sponsor; and
--The edit or limitation that the sponsor had implemented for the 
beneficiary had not terminated before disenrollment.

    We note that this final provision contains a technical correction 
to refer to 7 days instead of 7 business days the first instance this 
timeframe is used for consistency and added ``as soon as possible'' in 
Sec.  423.153(f)(15(D). It also substitutes ``provide information'' for 
``respond'' in one place for consistent terminology in this section.
(xii) Out of Scope Comments and Summary
    We received comments on the following topics which were out of 
scope of our proposal and to which we are therefore not responding: (1) 
CMS oversight of Part D drug management programs; (2) Education of Part 
D enrollees and providers regarding prescription drug management 
programs; (3) A seven day limit on opioids for acute pain; (4) 
Additional ideas about how to address the national opioid overuse 
crisis; (5) Opioid use standards in Medicare Set Aside arrangement 
(MSAs).
2. Flexibility in the Medicare Advantage Uniformity Requirements
    We have determined that providing access to services (or specific 
cost sharing for services or items) that are tied to health status or 
disease state in a manner that ensures that similarly situated 
individuals are treated uniformly is consistent with the uniformity 
requirement in the Medicare Advantage (MA) regulations at Sec.  
422.100(d). We solicited comments on this reinterpretation in the 
proposed rule. In response to those comments and our further 
consideration of this issue, we are providing guidance here to MA 
organizations. As discussed in more detail below, the Bipartisan Budget 
Act of 2018 (Pub. L. 115-123) amends section 1853 of the Act to 
authorize waiver of the uniformity requirement beginning in 2020 for MA 
plans that provide additional supplemental benefits (which are not 
required to be health care benefits) to chronically ill enrollees. It 
also amends section 1859 of the Act to require a nationwide revision of 
the Medicare Advantage Value-Based Insurance Design test model 
currently administered by the Center for Medicare and Medicaid 
Innovation, which provides similar flexibility to participating MA 
plans to offer targeted supplemental benefits. Our reinterpretation of 
the uniformity requirements is not identical to these statutory 
changes, but does provide a comparable flexibility for MA plans that is 
consistent with the requirement that MA plans offer uniform benefits, 
with uniform premium and uniform cost-sharing to all enrollees.
    This regulatory requirement that MA plans provide uniform benefits 
implements both section 1852(d) of the Act, which requires that 
benefits under the MA plan are available and accessible to each 
enrollee in the plan, and section 1854(c) of the Act, which requires 
uniform premiums for each enrollee in the plan. Previously, we required 
MA plans to offer all enrollees access to the same benefits at the same 
level of cost sharing. We have determined that these statutory 
provisions and the regulation at Sec.  422.100(d) mean that we have the 
authority to permit MA organizations the ability to reduce cost sharing 
for certain covered benefits, offer specific tailored supplemental 
benefits, and offer lower deductibles for enrollees that meet specific 
medical criteria, provided that similarly situated enrollees (that is, 
all enrollees who meet the medical criteria identified by the MA plan 
for the benefits) are treated the same. In addition, there must be some 
nexus between the health status or disease state and the specific 
benefit package designed for enrollees meeting that health status or 
disease state. As examples, uniformity flexibility will allow an MA 
plan to offer an enrollee with diabetes any or all of the following:
     Reduced cost sharing for endocrinologist visits;
     More frequent foot exams as a tailored, supplemental 
benefit;
     A lower deductible.
    In these examples, non-diabetic enrollees will not have access to 
these tailored cost sharing or supplemental benefits; however, any 
enrollee that develops diabetes will then have access to these 
benefits.
    We believe that our reinterpretation of the uniformity requirement 
is consistent with the underlying Part C statutory requirements because 
targeted supplemental benefits and cost sharing reductions must be 
offered uniformly to all enrollees with a specified health status or 
disease state. By tying specific supplemental benefits to specific 
medical conditions, MA plans would be building upon the concept of 
medical necessity and developing targeted benefits designed to treat 
the illnesses of enrollees who meet specific medical criteria. Further, 
treating similarly situated enrollees equally preserves the uniformity 
of the benefits package. This

[[Page 16481]]

flexibility is similar to our policy over the past several years of 
permitting MA plans to adopt tiered cost-sharing, that is, allowing 
plans to have different cost sharing for contracted providers of the 
same type (for example, hospitals) provided that enrollees are equally 
able to access the lower cost-sharing providers.
    Such flexibility under our new interpretation of the uniformity 
requirement is not without limits, however, as section 1852(b)(1)(A) of 
the Act prohibits an MA plan from denying, limiting, or conditioning 
the coverage or provision of a service or benefit based on health-
status related factors. MA regulations (for example, Sec. Sec.  
422.100(f)(2) and 422.110(a)) reiterate and implement this non-
discrimination requirement. In interpreting these obligations to 
protect against discrimination, we have historically indicated that the 
purpose of the requirements is to protect high-acuity enrollees from 
adverse treatment on the basis of their higher cost health conditions 
(79 FR 29843; 76 FR 21432; and 74 FR 54634). As MA plans consider this 
new flexibility in meeting the uniformity requirement, they must be 
mindful of ensuring compliance with non-discrimination responsibilities 
and obligations.\18\ MA plans that exercise this flexibility must 
ensure that the cost sharing reductions and targeted supplemental 
benefits are for health care services that are medically related to 
each disease condition. CMS will be concerned about potential 
discrimination if an MA plan is targeting cost sharing reductions and 
additional supplemental benefits for a large number of disease 
conditions, while excluding other, potentially higher-cost conditions. 
We will review benefit designs to make sure that the overall impact is 
non-discriminatory and that higher acuity, higher cost enrollees are 
not being excluded in favor of healthier populations.
---------------------------------------------------------------------------

    \18\ Among these responsibilities and obligations are compliance 
with Title VI of the Civil Rights Act, section 504 of the 
Rehabilitation Act, the Age Discrimination Act, section 1557 of the 
Affordable Care Act, and conscience and religious freedom laws.
---------------------------------------------------------------------------

    In identifying eligible enrollees, the MA plan must use medical 
criteria that are objective and measurable, and the enrollee must be 
diagnosed by a plan provider or have their existing diagnosis certified 
or affirmed by a plan provider to assure equal application of the 
criteria. Objective criteria that are contained in written policies and 
that are clearly and adequately communicated to enrollees (such as in 
the EOC and other plan documents) are necessary to ensure that these 
tailored benefits are not provided in a discriminatory fashion and that 
the overall package of benefits is uniform among similarly situated 
individuals. We view this flexibility as an extension of the concept 
that as an enrollee in good health without cardiac problems would not 
receive cardiac rehabilitation services, an enrollee who does not meet 
the medical criteria would not receive the targeted benefits offered by 
an MA plan.
    CMS is currently testing value based insurance design (VBID) 
through the use of our demonstration authority under section 1115A of 
the Act (42 U.S.C. 1315a, added by section 3021 of the Affordable Care 
Act), and we note that Bipartisan Budget Act of 2018 expands the 
testing of the model under section 1115A(b) to all 50 states by 2020. 
This demonstration includes some of the elements that are a part of our 
reinterpretation of the uniformity requirements. However, there are 
also features of the VBID demonstration that are unique to the 
demonstration test, such as the ability for participating plans to 
target Part D benefits, the restriction to certain medical conditions, 
and the requirement that plans apply to participate. We expect the VBID 
demonstration to provide CMS with insights into future VBID innovations 
for the MA program.
    After the publication of the proposed rule, Congress passed the 
Bipartisan Budget Act of 2018 (Pub. L. 115-123). Section 50322 of the 
law expanded supplemental benefits in Section 1852(a)(3) of the Act and 
also authorized waiver of the uniformity requirements to permit MA 
plans to offer targeted supplemental benefits for the chronically ill 
through new provisions, effective in plan year 2020.
    Specifically, the Bipartisan Budget Act of 2018 expands 
supplemental benefits available to chronically ill enrollees by adding 
a new subparagraph (D) to Section 1852(a)(3). This subparagraph expands 
supplemental benefits for the chronically ill to include benefits that 
``have a reasonable expectation of improving or maintaining the health 
or overall function of the chronically ill enrollee and may not be 
limited to being primarily health related benefits.'' These additional 
supplemental benefits will be qualitatively different than the 
supplemental health care benefits that MA plans may currently offer and 
may continue to offer to enrollees who are not chronically ill. In 
addition, it provides authority for the waiver of uniformity 
requirements ``only with respect to supplemental benefits provided to a 
chronically ill enrollee.''
    We have evaluated how this new authority for the Secretary to waive 
uniformity requirements relates to our concurrent reinterpretation of 
uniformity requirements. We believe that a waiver of uniformity 
requirements was authorized in this new provision to allow for the 
delivery of different, non-uniform benefits to a subset of enrollees 
that meet a specific definition: Chronically ill enrollee.\19\ We do 
not believe that our reinterpretation, which also allows for targeted 
benefits based on the disease state or health status, can only be 
accomplished through a waiver of uniformity requirements.
---------------------------------------------------------------------------

    \19\ The Bipartisan Budget Act specifically identifies the 
chronically ill as individuals with (1) one or more morbidities that 
is life threatening and limits overall function (2) has a high risk 
of hospitalization and adverse outcomes, and (3) requires intensive 
care coordination.
---------------------------------------------------------------------------

    We believe that the waiver authorized under the Bipartisan Budget 
Act is necessary in order to allow MA plans the flexibility to offer 
chronically ill enrollees supplemental benefits that are not uniform 
across the entire population of the chronically ill. The Bipartisan 
Budget Act states that supplemental benefits must ``have a reasonable 
expectation of improving or maintaining the health or overall function 
of the chronically ill enrollee.'' This means that MA plans do not have 
to offer uniform supplemental benefits to all chronically ill 
enrollees, and instead, may vary supplemental benefits offered to the 
chronically ill as it relates to the individual enrollee's specific 
medical condition and needs. In other words, a supplemental benefit 
adopted under the new statutory provision may not be provided to a 
chronically ill enrollee if that benefit does not have a reasonable 
likelihood of improving that enrollee's health condition. Therefore, we 
have determined that the waiver of uniformity requirements and the 
enactment of section 1852(a)(3)(D) of the Act does not limit our 
authority to interpret sections 1851(d) and 1854(c) of the Act as 
permitting uniform benefits to include specific services targeted for 
groups of similarly situated specific enrollees based on medical 
criteria.
    Our reinterpretation of uniformity requirements maintains the 
spirit of the MA regulations at Sec.  422.100(d), which aims for equal 
treatment across all similarly situated enrollees. A specific health 
status or disease state--or meeting a specific group of medical 
criteria--is merely a means of ``grouping'' similarly situated 
enrollees for equal access to and treatment in connection with coverage 
of benefits.

[[Page 16482]]

All enrollees in that group must have access to the same targeted 
benefits. The new expansion of supplemental benefits for the 
chronically ill breaks that construct because the needs of one 
chronically enrollee may be very different from those of another within 
the same health status or disease state. As such, a waiver was 
authorized to provide for differences in supplemental benefits across 
chronically ill enrollees in order for MA organization to craft 
specific supplemental benefit offerings for each vulnerable plan member 
so that individual needs are met.
    Further, our reinterpretation of uniformity requirements is 
compatible with the new legislation in Bipartisan Budget Act. Beginning 
in 2020, MA plans may offer three forms of supplemental benefits: 
``standard'' supplemental benefits offered to all enrollees; 
``targeted'' supplemental benefits offered to qualifying enrollees by 
health status or disease state; and ``chronic'' supplemental benefits 
offered to the chronically ill. The first two (standard and targeted) 
will be allowable in 2019. Only ``chronic'' supplemental benefits will 
be evaluated under the new expansive definition in the Bipartisan 
Budget Act and be eligible for a waiver of the uniformity requirements. 
Standard and targeted supplemental benefits will be evaluated under our 
existing interpretation of whether the benefit is ``primarily health 
related.'' It is possible that an enrollee qualifies for a ``targeted'' 
supplemental benefits as well as ``chronic'' supplemental benefits. In 
that circumstance, the MA plan must provide the targeted supplemental 
benefits as long as the enrollee establishes the required health status 
or disease state and the benefits are medically appropriate. However, 
the MA plan must only provide ``chronic'' supplemental benefits if the 
benefit has a reasonable expectation of improving or maintaining the 
health or overall function of the chronically ill enrollee.
    Based on these differences, it will be important for MA plans to 
identify in their bids and in their Evidence of Coverage documents 
which supplemental benefits are offered as ``standard'', ``targeted'', 
or ``chronic'' benefits. CMS will evaluate the acceptability of the 
supplemental benefit offering based on this designation and the 
standards identified in section 1852(a)(3) of the Act. We believe that 
both the new uniformity interpretation and the new statutory provision 
will succeed in increasing MA plans' flexibility and plan options and 
ultimately allow for better health outcomes.
    We received the following comments, and our response follows:
    Comment: A number of commenters supported CMS' implementation of 
this reinterpretation. These commenters stated that their ability to 
lower cost sharing will help beneficiaries seek high value and 
effective care.
    Response: We thank commentators for their support of this 
reinterpretation.
    Comment: Commenters suggested that CMS include regulatory text in 
the final rule that confirms that the flexibility that will be allowed 
in the MA uniformity requirements.
    Response: In this final rule, we are reinterpreting existing 
statutory and regulatory authority to allow MA organizations the 
ability to reduce cost sharing for certain covered benefits, offer 
specific tailored supplemental benefits, and offer different lower 
deductibles for enrollees that meet specific medical criteria. Thus, it 
is unnecessary to provide additional regulation language.
    Comment: A number of commenters requested that CMS provide 
additional sub-regulatory guidance surrounding this policy.
    Response: We will provide additional guidance and update all 
corresponding guidance documents (that is, bid guidance and operational 
guidance) to reflect the new interpretation. This guidance will be 
available before contract year 2019 bids are due.
    Comment: We received a number of comments asking that CMS issue 
sub-regulatory guidance with examples for permissible and impermissible 
actions, as well as examples of what would be considered 
discriminatory. In addition, others suggested that CMS specify the 
medical criteria that MA plans should use to determine enrollee 
eligibility as well as clear guidelines for eligible tailored 
supplemental benefits and/or reduced cost sharing.
    Response: CMS will provide additional operational guidance before 
CY 2019 bids are due.
    Comment: A commenter recommended that CMS open its implementing 
guidance to public comment prior to issuance.
    Response: We appreciate this comment. We will not be able to 
solicit industry comment in time for CY 2019 bids. However, we will 
take this suggestion under consideration as we develop future guidance 
and will reach out for input as needed.
    Comment: A number of commenters requested that CMS to provide 
certain technical clarifications. For instance, commenters questioned 
whether the plan-level deductible could be eliminated, or just reduced, 
and if lower cost sharing means a zero-dollar copay.
    Response: Yes, under this reinterpretation, a plan may reduce or 
eliminate a deductible, co-pay, or cost sharing for Part C services. We 
remind all organizations that this is reinterpretation is about MA 
benefits only and does not permit changes in Part D cost sharing or 
Part D benefits, which must be consistent with Part D applicable law 
and CMS policy. In addition, additional operational guidance will be 
provided before CY 2019 bids are due.
    Comment: We also received comments asking CMS to clarify whether a 
plan may reduce or eliminate certain cost sharing based on 
participation in a disease management program.
    Response: Yes, under this reinterpretation, a plan may restrict 
cost sharing reductions based on participation in a disease management 
program so long as there is equal access to the disease management 
program based on objective criteria related to a health status or 
disease state.
    Comment: We received comments asking CMS to clarify whether a plan 
may offer different co-pays to a subset of the population for some 
visits, but not all.
    Response: We appreciate the comment and are still considering how 
our new interpretation of the uniformity requirement would apply to 
such situations. We intend to provide clarifying guidance on this issue 
through HPMS memoranda and updates to the Medicare Managed Care Manual.
    Comment: A commenter requested that CMS clarify whether reduced 
cost sharing can be extended to premiums.
    Response: No, this flexibility does not extend to premiums; 
beneficiaries in the same plan must have the same premium. Allowing 
different premiums would violate section 1854(c) of the Act, which 
explicitly requires uniform premiums. Our reinterpretation of section 
1854(c), section 1852(d) regarding access to benefits for all 
enrollees, and the regulations implementing those statutes permits only 
reductions in Part C cost sharing and deductibles, and in targeting 
Part C supplemental benefits. As noted elsewhere, these specific 
benefits must be tied to health status or disease state and must be 
applied to health care services that are medically related to each 
disease condition. Additionally, targeted benefits and reduced cost 
sharing must be offered in a manner that ensures that similarly 
situated individuals are treated uniformly is consistent with the 
uniformity

[[Page 16483]]

requirement in the Medicare Advantage (MA) regulations at Sec.  
422.100(d).
    Comment: We received a comment asking CMS to confirm if MA plans 
may choose to apply these flexibilities to out-of-network benefits.
    Response: CMS will provide additional guidance and update all 
corresponding guidance documents to reflect the new interpretation. 
This guidance will be available before CY 2019 bids are due.
    Comment: We received comments requesting that CMS encourage plans 
to offer such flexibilities to beneficiaries with specific conditions 
(for example, dementia), stating that such flexibilities could help the 
ongoing treatment.
    Response: In the proposed rule, we stated that an MA plan may offer 
reduced cost sharing, deductibles, and or targeted supplemental 
benefits to enrollees diagnosed with specific diseases. In identifying 
eligible enrollees, the MA plan must use medical criteria that are 
objective and measurable, and the enrollee must be diagnosed by a plan 
provider or have their existing diagnosis certified or affirmed by a 
plan provider to assure equal application of the objective criteria 
necessary to provide equal treatment of similarly situated individuals. 
We do not have the authority to restrict or mandate which diagnoses or 
health conditions a plan chooses for this flexibility. Plans may 
determine which diagnoses or health conditions they choose to offer 
these flexibilities. CMS encourages plans to consider the population of 
their plan when making these decisions.
    Comment: We received a number of comments requesting that CMS allow 
reduced cost sharing and targeting supplemental benefits based on 
conditions unrelated to medical conditions, such as living situation 
and income. A commenter suggested CMS allow plans to reduced premiums 
for beneficiaries who sign up for automated premium payments.
    Response: The revised uniformity interpretation does not allow 
plans to reduce cost sharing and offer targeted supplemental benefits 
based on criteria unrelated to a diagnosis or health condition. We have 
determined that a plan may only provide access to targeted supplemental 
benefits (or specific cost sharing for certain services or items) based 
on health status or disease state. In identifying eligible enrollees, 
the MA plan must use medical criteria that are objective and 
measurable. In addition, MA plans that exercise this flexibility must 
ensure that the cost sharing reductions and targeted supplemental 
benefits are for health care services that are medically related to 
each diagnosis or health condition. Note that, effective CY 2020, the 
Bipartisan Budget Act of 2018 calls for a new category of supplemental 
benefits to be made available to chronically ill enrollees that are not 
limited to being primarily health related. Because the new benefits 
will not be limited to the primarily health related standard, it is 
possible for certain offerings to address issues beyond a specific 
medical condition, such as social supports. However, the basis for 
offering the new benefits will be based solely on an enrollees' 
qualification as ``chronically ill'' and may not be based on conditions 
unrelated to medical conditions, such as living situation and income.
    Comment: We received a comment urging CMS to include an affirmation 
that C-SNPs would automatically be permitted to adjust benefits and 
cost sharing based on the eligibility groupings that CMS has approved 
for each C-SNP.
    Response: CMS will update sub-regulatory guidance to clarify the 
impact of both this reinterpretation and the Bipartisan Budget Act on 
SNP policy.
    Comment: A commenter suggested that CMS should also provide 
clarification on how the additional benefit flexibility for highly 
integrated dual eligible special needs plans (D-SNPs), as outlined in 
Chapter 16b of the Medicare Managed Care Manual, is retained and/or 
modified under these provisions.
    Response: Chapter 16b and any corresponding guidance will be 
updated to clarify any impact this reinterpretation has on D-SNP 
policy.
    Comment: A commenter asked CMS to allow plans to provide certain 
supplemental benefits only to fully integrated D-SNP (FIDE SNP) 
enrollees who do not meet nursing home level of care requirements that 
would otherwise make them eligible for home and community-based 
services under an Elderly Waiver.
    Response: CMS will update sub-regulatory guidance to clarify the 
impact of both this reinterpretation and the Bipartisan Budget Act on 
D-SNP policy.
    Comment: We received some comments suggesting that CMS allow plans 
to reduce cost sharing and offer targeting supplemental benefits based 
on functional status, in addition to a medical condition.
    Response: There must be an underlying disease condition that is 
diagnosed, such as Alzheimer's disease or Parkinson's disease, in order 
for the plan to reduce cost sharing and offer targeted supplemental 
benefits. As stated in the proposed rule, in identifying eligible 
enrollees, the MA plan must use medical criteria that are objective and 
measurable, and the enrollee must be diagnosed by a plan provider or 
have their existing diagnosis certified or affirmed by a plan provider 
to assure equal application of the objective criteria necessary to 
provide equal treatment of similarly situated individuals. 
Specifically, MA plans offering targeted benefits will be responsible 
for developing the criteria to identify enrollees who fall within each 
of the clinical categories selected by an organization. Furthermore, 
cost sharing reductions and targeted supplemental benefits must be for 
health care services that are medically related to each disease 
condition.
    Note that, effective CY 2020, the Bipartisan Budget Act of 2018 
calls for a new category of supplemental benefits to be made available 
to chronically ill enrollees that are not limited to being primarily 
health related. Because the new benefits will not be limited to the 
primarily health related standard, it is possible for certain offerings 
to address issues beyond a specific medical condition, such as social 
supports. However, the basis for offering the new benefits will be 
based solely on an enrollees' qualification as ``chronically ill'' and 
may not be based on conditions unrelated to medical conditions, such as 
living situation and income.
    Comment: We received a comment asking CMS to expand our definition 
of health status or disease state to include ``medically complex 
patients.''
    Response: We have determined that a plan may only provide access to 
targeted supplemental benefits (or specific cost sharing for certain 
services or items) based on health status or disease state. In 
identifying eligible enrollees, the MA plan must use medical criteria 
that are objective and measurable. MA plans offering targeted benefits 
are responsible for developing the criteria to identify enrollees who 
fall within each of the clinical categories selected by an 
organization.
    Comment: We received comments requesting that CMS clarify whether a 
plan may reduce cost sharing only for a subset of high-quality network 
providers as long as all members with the same health status or disease 
state receive the same lower cost sharing for using these providers.
    Response: Yes, under this flexibility, a plan may reduce cost 
sharing for certain high-quality providers to members with a specified 
health status or disease state. MA plans may identify high-value 
providers across all Medicare provider types. This can include 
physicians and practices, hospitals,

[[Page 16484]]

skilled-nursing facilities, home health agencies, ambulatory surgical 
centers, etc.
    Comment: Some commenters suggested CMS delay implementation, 
stating that plans need time to enhance their existing internal tools 
and systems to accommodate varying benefit structures for different 
sub-populations within a single plan. Some commented that this may be 
administratively burdensome to implement, and therefore, may not be 
equal adoption across all MA organizations.
    Response: CMS will permit this flexibility beginning in CY 2019. MA 
organizations that need additional time to consider whether and how to 
take advantage of this new flexibility are not required to offer 
targeted supplemental benefits or reductions in cost sharing or 
deductibles. We believe it is important to allow plans the flexibility 
to target and better provide for the needs of their enrollees. Our 
reinterpretation of the uniformity requirements offers flexibility to 
MA organizations in designing their coverage and is not a mandate.
    Comment: Some commenters recommended that only high-performing 
plans be permitted to provide flexibility in the MA Uniformity 
Requirements.
    Response: CMS appreciates these comments and believes this 
flexibility will help enrollees seek higher value care. Therefore, CMS 
will permit all plans to use this flexibility beginning in CY 2019. CMS 
appreciates these comments and believes this flexibility will help 
enrollees seek higher value care. This flexibility is not a change to 
the regulation; it is a reinterpretation of an existing regulation. 
Therefore, all MAOs must comply with uniformity requirements regardless 
of individual plan performance. CMS will permit all plans to use this 
flexibility beginning in CY 2019.
    Comment: We received a number of comments suggesting that this 
reinterpretation is premature. Some commenters suggested that CMS wait 
until the VBID demonstration has concluded.
    Response: The existing VBID demonstration will continue. 
Information regarding this demonstration can be found at https://innovation.cms.gov/initiatives/vbid/. While we have adopted features of 
the VBID demonstration, the VBID demonstration and the new uniformity 
flexibilities are distinct. CMS will permit this flexibility beginning 
in CY 2019, as we believe it is important to allow plans the 
flexibility to target and better provide for the needs of their 
enrollees. We hope that the VBID demonstration will provide CMS with 
insights into future innovations for the MA program.
    Comment: Some commenters suggested that CMS take a measured 
approach by setting initial limits on the number of targeted conditions 
and tailored benefit packages that an MA plan can offer.
    Response: The existing uniformity flexibility regulatory authority 
does not allow CMS to limit the number of targeted conditions without 
additional rulemaking.
    Comment: Some suggested that CMS adopt the oversight requirements 
in the VBID demonstration in allowing plans to use this flexibility 
under the new reinterpretation.
    Response: Currently, the VBID demonstration has a number of 
oversight requirements, including some marketing restrictions, 
monitoring to ensure compliance with demonstration rules, data 
reporting to help CMS evaluate outcomes, and restricting low performing 
plans from participation. CMS has no plans to adopt these additional 
demonstration requirements. First, CMS has a robust compliance and 
auditing program to oversee MA plans and all benefit packages are 
reviewed by CMS. Therefore, we do not believe any additional monitoring 
or compliance is needed. Second, MA rules require that this benefit be 
available in marketing materials and transparent to enrollees. 
Therefore, we cannot restrict marketing this benefit. Third, we believe 
we do not need to introduce any additional uniformity reporting as the 
VBID reporting is designed to aide demonstration evaluation. However, 
CMS will monitor the implementation of this flexibility and make 
appropriate adjustments as needed.
    Comment: Commenters asked that CMS clarify how this flexibility 
impacts the VBID demonstration.
    Response: The existing VBID demonstration will continue. We note 
that Bipartisan Budget Act of 2018 expands the testing authority under 
section 1115A(b) to all 50 states. This flexibility will not impact the 
VBID demonstration, which is separate from this rulemaking. The new 
flexibilities discussed here will have no impact on current VBID 
operations. Information regarding this demonstration can be found at 
https://innovation.cms.gov/initiatives/vbid/. The VBID demonstration 
will provide CMS with insights into future innovations for the MA 
program.
    Comment: A commenter asked if CMS planned to implement reporting 
requirements related to this flexibility, noting that such requirements 
are in the VBID demonstration.
    Response: CMS has no plans to add any reporting requirements 
related to uniformity flexibility at this time. We do note that MA 
plans must explain the targeted supplemental benefits and reductions in 
cost sharing and deductibles in their bids (OMB 0938-0763), including 
information necessary for CMS to evaluate if there is any 
discrimination involved. In addition, MA plans must include 
descriptions of these benefits in benefit disclosures required under 
Sec.  422.111.
    Comment: We received a number of comments expressing concern that 
this policy could increase beneficiary confusion, particularly as it 
relates to marketing materials provided during the annual election 
process.
    Response: To mitigate beneficiary confusion, CMS will require MA 
plans that take advantage of this flexibility to include benefit 
flexibility information in their CY 2019 EOC. Also, indication of 
additional benefits and/or reduced cost sharing for enrollees with 
certain health conditions will be displayed in Medicare Plan Finder.
    Comment: We received several comments asking CMS to clarify whether 
plans will be permitted to market this flexibility to potential 
enrollees. Some suggested CMS permit marketing. Others suggested CMS 
prohibit marketing.
    Response: Plans will be allowed to market the additional benefits 
and/or reduced cost sharing to potential enrollees to give 
beneficiaries the information necessary to choose the best plan for 
their health care needs. Plans will be required to follow the same CMS 
marketing rules for this benefit, as they are required to follow when 
marketing any other benefit. This includes ensuring that materials are 
not materially inaccurate or misleading or otherwise make material 
misrepresentations. Specifically, CMS will require that plans include 
comprehensive benefit flexibility information in their CY 2019 EOC and 
indicate the additional benefits and/or reduced cost sharing in 
Medicare Plan Finder.
    Comment: A number of commenters expressed concern that this policy 
may lead to discrimination. For example, some commenters expressed 
concern that a plan may balance the reduction of cost sharing for one 
group by increasing cost sharing for others. Further, some commenters 
expressed concern that this could lead to lead to ``cherry-picking'' by 
plans for beneficiaries with low-cost conditions while discriminating 
against

[[Page 16485]]

those with higher-cost chronic conditions.
    Response: As noted in the preamble language, the implementation of 
this flexibility must not violate existing anti-discrimination rules 
(for example, service category cost sharing and per member per month 
actuarial equivalence standards communicated by CMS annually in the 
Call Letter). Organizations that exercise this flexibility must ensure 
that the cost sharing reductions and targeted supplemental benefits 
only apply to healthcare services that are medically related to each 
health status or disease state. CMS will not permit cost sharing 
reductions across all benefits for an enrollee; cost sharing reductions 
must be for specific benefits related to a specific health status or 
disease state. Specifically, plans must not target cost sharing 
reductions and additional supplemental benefits for a large number of 
disease conditions, while excluding other higher-cost conditions. CMS 
will review benefit designs to make sure that targeted disease state(s) 
and/or clinical condition(s) included in the benefit design are non-
discriminatory and that higher acuity, higher cost enrollees are not 
being excluded in favor of healthier populations.
    Comment: A commenter recommended that plan members should have full 
appeal rights with respect to denial of access to supplemental 
benefits.
    Response: All negative coverage decisions are subject to appeal 
rights. CMS is reinterpreting existing statutory language at section 
1854(c) and 1852(d) of the Act, and the implementing regulation at 
Sec.  422.100(d), to allow MA organizations the ability to reduce cost 
sharing for certain covered benefits, offer specific tailored 
supplemental benefits, and offer lower deductibles for enrollees that 
meet specific medical criteria. We have reviewed and considered all 
comments on this clarification and will begin implementing this 
additional flexibility in CY 2019. In addition, we will provide 
additional operational guidance before CY 2019 bids are due.
3. Segment Benefits Flexibility
    In reviewing section 1854(h) of the Act and Medicare Advantage (MA) 
regulations governing plan segments, we have determined that the 
statute and existing regulations may be interpreted to allow MA plans 
to vary supplemental benefits, in addition to premium and cost sharing, 
by segment so long as the supplemental benefits, premium, and cost 
sharing are uniform within each segment of an MA plan's service area. 
Plans segments are county-level portions of a plan's overall service 
area which, under current CMS policy, are permitted to have different 
premiums and cost sharing amounts as long as these premiums and cost 
sharing amounts are uniform throughout the segment. As county-level 
areas, these are separate rating setting areas within the plan's 
service area; no further subdivision is permitted. We are proposed to 
revise our interpretation of the existing statute and regulations to 
allow MA plan segments to vary by supplemental benefits in addition to 
premium and cost sharing, consistent with the MA regulatory 
requirements defining segments at Sec.  422.262(c)(2).
    We received the following comments, and our response follows:
    Comment: We received a number of comments supporting the 
implementation of this reinterpretation.
    Response: We thank commentators for their support of this 
reinterpretation.
    Comment: Many commenters requested that CMS clarify if this 
segmentation can be offered to a sub-set of the network providers.
    Response: The MA regulations at Sec.  422.2 define a provider 
network as occurring at the MA plan level: ``. . . the providers with 
which an MA organization contracts or makes arrangements to furnish to 
furnish covered health care services to Medicare enrollees under a MA 
coordinated care plan or network PFFS plan''. In implementing its 
network adequacy standard CMS allows for networks at the MA plan level 
(a provider specific plan) or at the contract level. In addition to 
being inconsistent with the regulations we believe that allowing 
networks to be established at the MA plan segment level would introduce 
an unnecessary level of complexity to the MA program.
    Comment: A commenter asked if there are any restrictions to the 
benefits that may vary and if all supplemental benefits and services 
are eligible, or is this specific to a set of supplemental benefits?
    Response: Plans may vary supplemental benefits by plan segment 
consistent with the bid submitted for the segment. All basic benefits 
(that is, Part A and B benefits) must be offered by all MA plans in all 
segments.
    Comment: A commenter asked if the maximum out-of-pocket (MOOP) 
amount was one of the elements that may vary.
    Response: Yes, because the MOOP is an element of the cost-sharing 
structure of the plan, each segment may have its own MOOP. This 
flexibility already exists in MA.
    Comment: Commenters asked CMS to clarify if in sub-regulatory 
guidance that plans are allowed to display multiple segments in the 
Evidence of Coverage (EOC), Summary of Benefits, and other coverage 
documents.
    Response: Plans will be required to follow the same CMS 
communication, disclosure and marketing guidelines for each segment In 
addition, as noted in section II.B, CMS will require plans to include 
comprehensive benefit flexibility information in their CY 2019 (EOC).
    Comment: A commenter noted that CMS uses both ``supplemental 
benefits'' and ``benefits'' in the preamble language and asked CMS 
explicitly clarify if this new segment benefit flexibility applies only 
to supplemental benefits and not to the core MA benefit package to 
which beneficiaries are entitled.
    Response: Thank you for the comment. All MA plans must provide 
basic benefits--meaning Part A and Part B benefits consistent with the 
cost-sharing limits identified in section 1854(e)(4)(A) \20\ and Sec.  
422.100(j) and (k)--in all segments. We have determined that the 
statute and existing regulations may be interpreted to allow MA plans 
to vary supplemental benefits, in addition to premium and cost sharing, 
by segment, as long as the benefits, premium, and cost sharing are 
uniform within each segment of an MA plan's service area. Supplemental 
benefits include cost-sharing reductions from the actuarial equivalent 
on average of original Medicare for basic benefits and coverage of 
additional services and items not covered by original Medicare.
---------------------------------------------------------------------------

    \20\ Beginning in 2006, an MA plan may reduce cost sharing below 
the actuarial value specified in section 1854(e)(4)(A) of the Act 
only as a mandatory supplemental benefit. The actuarial value of the 
deductibles, coinsurance, and copayments applicable to the basic 
benefits on average to enrollees in an MA plan must be equal to the 
actuarial value of the deductibles, coinsurance, and copayments that 
would be applicable with respect to such benefits on average to 
individuals enrolled in original Medicare.
---------------------------------------------------------------------------

    Comment: Some commenters expressed concern that CMS is moving too 
quickly in implementing this reinterpretation and that such flexibility 
should be tested on a small scale first.
    Response: We believe this flexibility will allow plans to better 
target and provide for the needs of their populations. CMS will monitor 
the implementation of this flexibility and make appropriate adjustments 
as needed. In addition, we note that MA organizations are not required 
to use this flexibility to vary benefits, cost-sharing and premium at 
the segment level.

[[Page 16486]]

    Comment: We received many comments related to concern about benefit 
transparency and that this flexibility to offer segments with varied 
benefits, cost-sharing, or premiums, may lead to beneficiary confusion. 
Commenters expressed concern that this flexibility will result in 
beneficiary confusion regarding the differences between plans, which 
may create a confusing environment for Medicare beneficiaries trying to 
make informed decisions when choosing plans.
    Response: Plans will be required to follow existing rules governing 
mandatory disclosures (for example, Sec.  422.111), communications and 
marketing. In addition, CMS will require plans to include comprehensive 
benefit flexibility information in their CY 2019 EOC.
    In this final rule, CMS is adopting a reinterpretation of section 
1854(h) of the Act and Sec. Sec.  422.100(d)(2) and 422.262 to allow MA 
organizations the ability to vary supplemental benefits, in addition to 
premium and cost sharing, by segment, as long as the benefits, premium, 
and cost sharing are uniform within each segment of an MA plan's 
service area. We have reviewed comments on our proposal and have 
considered these comments as we finalize the policy. Plans will be 
permitted to begin implementing this flexibility in CY 2019.
4. Maximum Out-of-Pocket Limit for Medicare Parts A and B Services 
(Sec. Sec.  422.100(f)(4) and (5) and 422.101(d))
    As provided at Sec. Sec.  422.100(f)(4) and (5) and 422.101(d)(2) 
and (3), all Medicare Advantage (MA) plans (including employer group 
waiver plans (EGWPs) and special needs plans (SNPs)), must establish 
limits on enrollee out-of-pocket cost sharing for basic benefits 
(meaning Parts A and B services) that do not exceed the annual limits 
established by CMS. CMS added Sec.  422.100(f)(4) and (5), effective 
for coverage in 2011, under the authority of sections 1852(b)(1)(A), 
1856(b)(1), and 1857(e)(1) of the Act in order not to discourage 
enrollment by individuals who utilize higher than average levels of 
health care services (that is, in order for a plan not to be 
discriminatory) (75 FR 19709-11). Section 1858(b)(2) of the Act 
requires a limit on in-network out-of-pocket expenses for enrollees in 
regional MA plans. In addition, local preferred provider organization 
(LPPO) plans, under Sec.  422.100(f)(5), and regional PPO (RPPO) plans, 
under section 1858(b)(2) of the Act and Sec.  422.101(d)(3), are 
required to have a ``catastrophic'' limit inclusive of both in- and 
out-of-network cost sharing for all Parts A and B services, the annual 
limit which is also established by CMS; all cost sharing (that is, 
deductibles, coinsurance, and copayments) for Parts A and B services, 
excluding plan premium, must be included in each plan's maximum out-of-
pocket (MOOP) amount subject to these limits. As stated in the CY 2018 
final Call Letter \21\ and in the 2010 final rule (75 FR 19710), CMS 
currently sets MOOP limits based on a beneficiary-level distribution of 
Parts A and B cost sharing for individuals enrolled in Medicare Fee-
for-Service (FFS) for local and regional MA plans.
---------------------------------------------------------------------------

    \21\ The CY 2018 final Call Letter may be accessed at https://www.cms.gov/Medicare/Health-Plans/MedicareAdvtgSpecRateStats/Announcements-and-Documents.html.
---------------------------------------------------------------------------

    CMS proposed to amend Sec. Sec.  422.100(f)(4) and (5) and 
422.101(d)(2) and (3) to clarify that CMS may use Medicare FFS data to 
establish the annual MOOP limits, which have historically been linked 
to values that approximate the 85th and 95th percentile of out-of-
pocket expenditures for beneficiaries in original Medicare. The 
proposal included that CMS have authority to increase the voluntary 
MOOP limit to another percentile level of Medicare FFS, increase the 
number of service categories that have higher cost sharing in return 
for offering a lower MOOP amount, and implement more than two levels of 
MOOP and cost sharing limits to encourage plan offerings with lower 
MOOP limits. CMS also proposed that it have authority to increase the 
number of service categories that have higher cost sharing in return 
for offering a lower (voluntary) MOOP amount. To codify these various 
authorities, CMS proposed regulation text permitting CMS to set the 
annual MOOP limits to strike a balance between limiting maximum 
beneficiary out-of-pocket costs and potential changes in premium, 
benefits, and cost sharing, with the goal of ensuring beneficiary 
access to affordable and sustainable benefit packages. CMS intends to 
use the annual Call Letter process to communicate its application of 
the regulation and to transition changes to MOOP limits over time, 
beginning no earlier than in CY 2020, to avoid disruption to benefit 
designs and minimize potential beneficiary confusion.
    As noted in the proposed rule, CMS discussed in the 2010 rulemaking 
(75 FR 19709) that it provides greater flexibility in establishing cost 
sharing for basic benefits to MA plans that adopt a lower, voluntary 
MOOP limit than is available to plans that adopt the higher, mandatory 
MOOP limit. The number of beneficiaries with access to a voluntary MOOP 
limit plan and the proportion of total enrollees in a voluntary MOOP 
limit plan has decreased significantly from CY 2011 to CY 2017.
    Currently, CMS sets the mandatory MOOP amount at approximately the 
95th percentile of projected beneficiary out-of-pocket spending. Stated 
differently, 5 percent of Medicare FFS beneficiaries are expected to 
incur approximately $6,700 or more in Parts A and B deductibles, 
copayments, and coinsurance. CMS sets the voluntary MOOP amount of 
$3,400 to represent approximately the 85th percentile of projected 
Medicare FFS out-of-pocket costs. The Office of the Actuary conducts an 
annual analysis to help CMS determine these MOOP limits. Since the MOOP 
requirements for local and regional MA plans were finalized in 
regulation, a strict application of the 95th and 85th percentiles would 
have resulted in MOOP limits for local and regional MA plans 
fluctuating from year-to-year. To avoid enrollee confusion, allow plans 
to provide stable benefit packages year over year, and minimize 
disincentives to the adoption of the lower voluntary MOOP amount 
because of fluctuations in the amount, CMS has exercised discretion in 
order to maintain stable MOOP limits from year-to-year that approximate 
but are not exactly at the 85th and 95th percentile of, beneficiary 
cost sharing in Medicare FFS.
    In the proposed rule, CMS explained that it would want to change 
the MOOP limits if a consistent pattern of increasing or decreasing 
costs emerges over time. CMS also summarized how stakeholders have 
suggested changes to how CMS establishes MOOP limits, including 
suggestions to use the most appropriate data to inform its decision-
making, increase the MOOP limits and the number of service categories 
that have higher cost sharing in return for a plan offering a lower 
MOOP limit, and implement different levels of MOOP and service category 
cost sharing standards to encourage plan offerings with lower MOOP 
limits.
    CMS explained in the proposed rule its goal to establish future 
MOOP limits based on the most relevant and available data, or 
combination of data, that reflects beneficiary health care costs in the 
MA program and maintains MA benefit stability over time. Medicare FFS 
data currently represents the most relevant and available data at this 
time so the proposal included codifying use of Medicare FFS data in 
Sec. Sec.  422.100(f)(4) and (5) and 422.101(d)(2) and (3).
    CMS also explained in the proposed rule that it wished to have 
flexibility to

[[Page 16487]]

change its existing methodology (of using the 85th and 95th percentiles 
of projected beneficiary out-of-pocket Medicare FFS spending) in the 
future. The proposed rule was explicitly based on a policy objective of 
striking the appropriate balance between limiting MOOP costs and 
potential changes in premium, benefits, and cost sharing with the goal 
of making sure beneficiaries can access affordable and sustainable 
benefit packages. While CMS intends to continue using the 85th and 95th 
percentiles of projected beneficiary out-of-pocket spending for the 
immediate future to set MA MOOP limits, the proposed amendments to 
Sec. Sec.  422.100(f)(4) and (5) and 422.101(d)(2) and (3) were to 
incorporate authority to balance these factors to set the MOOPs. The 
flexibility contemplated by the proposed rule would permit CMS to 
annually adjust mandatory and voluntary MOOP limits based on changes in 
market conditions and to ensure the sustainability of the MA program 
and benefit options.
    The proposed rule also explained how CMS would, in advance of each 
plan year, use the annual Call Letter and other guidance documents to 
explain its application of the regulations and the data used to 
identify MOOP limits. In addition, CMS committed to transitioning any 
significant changes adopted using the new proposed authority over time 
to avoid disruption to benefit designs and minimize potential 
beneficiary confusion.
    In conclusion, CMS proposed to amend Sec. Sec.  422.100(f)(4) and 
(5) and 422.101(d)(2) and (3) to clarify that CMS may use Medicare FFS 
data to establish annual MOOP limits and to adopt a flexible standard 
for setting the MOOPs. This flexible standard would authorize CMS to 
increase the voluntary MOOP limit to another percentile level of 
Medicare FFS beneficiary spending; increase the number of service 
categories that have higher cost sharing in return for offering a lower 
MOOP amount; and implement more than two levels of MOOP and cost 
sharing limits (as a means to encourage plan offerings with lower MOOP 
limits).
    We received the following comments on this proposal, and our 
response follows,
    Nearly all commenters who provided feedback on this provision 
(Maximum Out-of-Pocket Limit for Medicare Parts A and B Services 
(Sec. Sec.  422.100(f)(4) and (5) and 422.101(d))) also provided 
feedback on the proposal at section II.B.5 (Cost Sharing Limits for 
Medicare Parts A and B Services (Sec.  422.100(f)(6))). In this 
section, we address comments that focus on either this section or both 
sections, while we address comments that focus on cost sharing limits 
in section II.B.5.
    Comment: The majority of commenters supported this proposal, 
stating that CMS should primarily use Medicare FFS and MA encounter 
data to inform its decision-making, and that CMS should consider 
authorizing more than two levels of MOOP and associated cost sharing 
standards to encourage plan offerings with lower MOOP limits. Some 
commenters also made suggestions for levels of MOOP limits and cost 
sharing service category adjustments that could be especially 
beneficial.
    Response: We thank commenters for their support. CMS's goal is to 
establish future MOOP limits based on the most relevant and available 
data, or combination of data, that reflects beneficiary health care 
costs in the MA program and maintains benefit stability over time. This 
final rule limits that data to the FFS Medicare data, but as other data 
sources become accessible, relevant, and of the quality necessary to 
make these determinations, we will engage in rulemaking to change the 
rule.
    Comment: Many commenters expressed concern with MA encounter data 
being used at this time to establish MOOP levels based on data quality 
issues. Commenters also encouraged CMS to continue working with MA 
organizations to improve the validity and reliability of MA encounter 
data. A commenter suggested CMS consider other data such as of 
Marketplace Qualified Health Plan review data.
    Response: Medicare FFS data is the most relevant and available data 
at this time. CMS will consider future rulemaking to use MA encounter 
cost data as well as Medicare FFS data to establish MOOP limits. In 
determining completeness and accuracy of MA encounter data CMS does 
consider the various managed care payment arrangements and payment 
policies that may exist between organizations, as compared to Medicare 
FFS data (which are based on relatively consistent payment schedules 
and payment policies). At this time we cannot commit to a timeline for 
use of MA encounter data or other data sources to establish MOOP 
limits. As we learn more and are able to establish standards for the 
completeness and sufficiency of alternate data sources, we will revisit 
this issue.
    Comment: Some commenters noted concern with the specific 
methodology that CMS would use other than the 85th or 95th percentile 
of Medicare FFS beneficiary costs to establish MOOP limits and how 
abrupt changes may impact cost sharing and the levels of MOOP limits. A 
commenter also stated concern about what level of change to MOOP limits 
would be considered ``significant'' and necessitate a multi-year 
transition. Some commenters suggested CMS maintain the current 
voluntary and mandatory MOOP limits (that is, $3,400 and $6,700) and 
establish additional MOOP limits between these levels with prorated 
cost sharing standards to minimize any impact to benefit design and 
beneficiaries. Some commenters suggested CMS further change the 
regulatory cost sharing standards for inpatient, skilled nursing 
facility, emergency care, and other professional services as an 
incentive for plans to adopt lower MOOP limits, while other commenters 
cautioned CMS to limit changes to these categories to prevent 
discrimination.
    Response: We appreciate the feedback and will take these 
suggestions and concerns under consideration. CMS plans to transition 
changes under the finalized regulations over time, beginning no earlier 
than CY 2020, to avoid disruption to benefit designs and minimize 
potential beneficiary confusion. The regulation standard adopted in 
this final rule for Sec. Sec.  422.100(f)(4) and (5) and 422.101(d)(2) 
and (3) (that the MOOP be set to strike a balance between limiting 
maximum beneficiary out of pocket costs and potential changes in 
premium, benefits, and cost sharing, with the goal of ensuring 
beneficiary access to affordable and sustainable benefit packages) will 
apply to determinations regarding a transition period from one 
particular MOOP to another MOOP. We anticipate that sudden and 
significant shifts in the MOOP would cause sudden changes in premiums, 
benefits and cost sharing, which are identified under the new 
regulation text as something to be minimized. Consistent with past 
practice, CMS will continue to publish the expected changes for the 
next year and a description of how the regulation standard is applied 
(that is, the methodology used) in the annual Call Letter prior to bid 
submission so that MA plans can submit bids consistent with MA 
standards. CMS has historically provided prior notice and an 
opportunity to comment on the Call Letter guidance document and does 
not expect that to change. This will provide MA organizations adequate 
time to comment and prepare for changes. We anticipate potential 
changes in MOOP limits or cost sharing based on MA benefit design 
strategies will be

[[Page 16488]]

conveyed through existing enrollee communication materials.
    Comment: Several commenters were concerned about CMS's strategy to 
promote plan adoption of lower MOOP limits by increasing the cost 
sharing flexibility for those plans. They suggested that allowing this 
flexibility may result in discriminatory benefit designs as plans may 
raise cost sharing limits for certain service categories more likely to 
be utilized by vulnerable beneficiaries, and that such beneficiaries 
would be especially disadvantaged if they do not reach the lower, 
voluntary MOOP limit. Some commenters identified concern for specific 
service categories if their cost sharing limits were raised (for 
example, inpatient and professional services) and requested CMS be 
especially thoughtful when considering changes to these categories. A 
few commenters proposed that CMS consider lowering cost sharing limits 
for mandatory MOOP plans as another method to encourage adoption of a 
lower MOOP limit.
    Response: CMS agrees that while increasing flexibility for MA plans 
that voluntarily offer lower MOOP limits can allow for improved plan 
design, it will be important to make sure that vulnerable patient 
populations are not discriminated against and that plan designs are not 
confusing to beneficiaries. Other existing regulations governing cost 
sharing designs of MA plans--such as the prohibition on discrimination 
(Sec.  422.100(f)(2)), requirement that certain services have cost 
sharing that is no higher than FFS Medicare limits (Sec.  422.100(j)), 
and requirement that overall plan cost-sharing for coverage of basic 
benefits must be actuarially equivalent to the level of cost sharing 
(deductible, copayments, or coinsurance) charged to beneficiaries under 
the original Medicare program option (Sec.  422.254(b)(4))--remain in 
place and are unchanged by this final rule. CMS will manage the 
flexibility plans have in setting cost sharing limits to make sure that 
plan designs are not discriminatory. For example, CMS does not intend 
to significantly increase cost sharing limits as a percentage of 
Medicare FFS above current levels for inpatient, primary, and specialty 
care based on cost sharing standards that CMS publishes in its annual 
Call Letter. CMS intends to continue the practice of furnishing 
information to MA organizations about the methodology used to establish 
cost sharing limits and the thresholds CMS identifies as non-
discriminatory through the annual Call Letter process or Health Plan 
Management System (HPMS) memoranda and solicit comments, as 
appropriate.
    Comment: Some commenters reported concern with the proposal to 
amend Sec.  422.100(f)(6) and implement it as described in the proposed 
rule strategy because of unintended consequences, such as beneficiaries 
having to choose between plans offering different levels of MOOP limits 
and variability in cost sharing across services. A commenter suggested 
that CMS update plan selection resources such as Medicare Plan Finder 
(MPF) to simplify the plan selection process and assist beneficiaries 
choose the plan that best fits their unique health care needs.
    Response: We agree that cost sharing must not be discriminatory and 
that it is important to make sure that beneficiaries have adequate 
information to support their plan enrollment decision-making. 
Beneficiaries typically make decisions based on plan characteristics 
that are important to their needs (for example, benefits, cost sharing, 
MOOP limit, plan premium, and providers) and are not familiar with the 
complexities associated with bidding guidance and cost sharing 
standards that plans use to prepare bids. To minimize beneficiary 
confusion, CMS will continue evaluations and enforcement of the current 
authority prohibiting plans from misleading beneficiaries in their 
communication materials. In addition, we will disapprove a plan bid if 
its proposed benefit design substantially discourages enrollment in 
that plan by certain Medicare-eligible individuals. In addition, CMS 
will continue efforts to improve plan offerings and plan comparison 
tools and resources (for example, MPF and 1-800-MEDICARE).
    Comment: We received a comment that noted the importance of MOOP 
limits as part of a benefit offering for beneficiary protection and 
that there are MA plans being marketed that do not have a MOOP for out-
of-network services.
    Response: CMS notes that all Medicare LPPOs and RPPOs are required 
to have a combined in- and out-of-network MOOP limit. HMO-POS plans may 
offer out-of-network benefits as supplemental benefits, but are not 
required to have these services contribute to the in-network MOOP limit 
or a combined in- and out-of-network MOOP limit.
    We received over 40 comments pertaining to the proposal, with the 
majority reflecting support to amend Sec. Sec.  422.100(f)(4) and (5) 
and 422.101(d)(2) and (3) to clarify that CMS may use Medicare FFS data 
to establish annual MOOP limits. The majority of comments also 
supported the regulation amendment to add a standard governing CMS 
establishment of MOOP limits (to strike a balance between limiting 
maximum beneficiary out of pocket costs and potential changes in 
premium, benefits, and cost sharing, with the goal of ensuring 
beneficiary access to affordable and sustainable benefit packages). As 
noted in the proposed rule, CMS will interpret and implement these 
amendment to give CMS the authority to change MOOP limits; increase the 
number of service categories that have higher cost sharing in return 
for offering lower MOOP limits; and implement more than two levels of 
MOOP limits. Consistent with past practice, CMS will continue to 
publish the expected changes for the next year and a description of how 
the regulation standard is applied in the annual Call Letter prior to 
bid submission so that MA plans can submit bids consistent with MA 
standards. CMS plans to transition changes under the finalized 
regulations over time, beginning no earlier than CY 2020, to avoid 
disruption to benefit designs and minimize potential beneficiary 
confusion. After careful consideration of all of the comments we 
received, we are finalizing the proposal to amend Sec. Sec.  
422.100(f)(4) and (5) and Sec.  422.101(d)(2) and (3) as described with 
an applicability date of January 1, 2020; this applicability date is 
consistent with our intent that these new standards apply to cost 
sharing limits set for plans years after 2019. We are also finalizing 
minor revisions as follows:
    (1) In Sec.  422.100(f)(5), we are finalizing the regulation text 
without the phrase ``annually determined by CMS using Medicare Fee for 
Service and to establish appropriate'' in the introductory text; we 
believe that the regulation text finalized in the paragraph (f)(5)(ii) 
is sufficiently clear on this point.
    (2) In Sec.  422.100(f)(5)(ii), we will finalize the text with 
``CMS sets'' in place of ``CMS will set'' for clarity.
5. Cost Sharing Limits for Medicare Parts A and B Services (Sec.  
422.100(f)(6))
    In addition to MOOP Limits, MA plan cost sharing for Parts A and B 
services is subject to additional regulatory requirements and limits in 
Sec. Sec.  417.454(e), 422.100(f)(6), and 422.100(j). Section 
422.100(f)(6) provides that cost sharing must not be discriminatory and 
CMS determines annually the level at which certain cost sharing becomes 
discriminatory. Sections 417.454(e) and 422.100(j) are based on how 
section 1852(a)(1)(B)(iii) and (iv) of the Act directs that cost

[[Page 16489]]

sharing for certain services may not exceed the cost sharing levels in 
Medicare Fee-for-Service (FFS); under the statute and the regulations, 
CMS may add to that list of services. CMS identifies Parts A and B 
services that are more likely to be used by enrollees in establishing 
its cost sharing parameters for review and evaluation. The review 
parameters are currently based on Medicare FFS data and reflect a 
combination of patient utilization scenarios and length of stays or 
services used by average to sicker patients. CMS uses multiple 
utilization scenarios for some services (for example, inpatient care) 
to guard against MA organizations distributing or designing cost 
sharing amounts in a manner that is discriminatory. Review parameters 
are also established for frequently used professional services, such as 
primary and specialty care services.
    CMS proposed to amend Sec.  422.100(f)(6) to clarify that it may 
use Medicare FFS data to establish appropriate cost sharing limits for 
certain services that are not discriminatory. In addition, CMS proposed 
to amend the regulation to reflect that CMS would use FFS data and MA 
encounter data to inform patient utilization scenarios to help identify 
MA plan cost sharing standards and thresholds that are not 
discriminatory. We specifically solicited comment on whether to codify 
that use of MA encounter data for this purpose in Sec.  422.100(f)(6). 
In this final rule, we reiterate our intent to use the annual Call 
Letter process to communicate its application of the regulation and 
announce our intent to transition changes to cost sharing standards 
over time, beginning no earlier than in CY 2020, to avoid disruption to 
benefit designs and minimize potential beneficiary confusion. This 
proposal is not related to a statutory change.
    In the proposed rule, CMS explained that it sought to codify 
authorization to allow CMS to use the most relevant and appropriate 
information in determining whether specific cost sharing is 
discriminatory and to set standards and thresholds above which CMS 
believes cost sharing is discriminatory. In addition, CMS stated its 
intent to continue the practice of furnishing information to MA 
organizations about the methodology used to establish cost sharing 
limits and the thresholds CMS identifies as non-discriminatory through 
the annual Call Letter process. We referenced soliciting comments 
before finalizing guidance as necessary and appropriate. We expect this 
process will allow MA organizations to prepare plan bids consistent 
with parameters that CMS have determined to be non-discriminatory. In 
addition, and as appropriate, CMS noted that we may also issue guidance 
using Health Plan Management System (HPMS) memoranda.
    CMS noted in the proposed rule that while it has not established a 
specific service category cost sharing limit for all possible services, 
CMS has issued guidance that MA plans must pay at least 50 percent of 
the contracted (or Medicare allowable) rate and that cost sharing for 
services cannot exceed 50 percent of the total MA plan financial 
liability for the benefit in order for the cost sharing for such 
services to be considered non-discriminatory (Medicare Managed Care 
Manual, Chapter 4, Section 50.1 at https://www.cms.gov/Regulations-and-Guidance/Guidance/Manuals/internet-Only-Manuals-IOMs-Items/CMS019326.html). We stated our belief that cost sharing (service 
category deductibles, copayments, or co-insurance) that fails to cover 
at least half the cost of a particular service or item acts to 
discriminate against those for whom those services and items are 
medically necessary and discourages enrollment by beneficiaries who 
need those services and items. If an MA plan uses a copayment method of 
cost sharing, then the copayment for an in-network Medicare FFS service 
category cannot exceed 50 percent of the average contracted rate of 
that service without CMS seriously questioning and reviewing the cost-
sharing as discriminatory. CMS does not believe that cost sharing at 
such high levels can legitimately serve any purpose other than 
discriminating against the enrollees who need and frequently use those 
services. Some service categories may identify specific benefits for 
which a unique copayment will apply, while others are grouped, such as 
durable medical equipment or outpatient diagnostic and radiological 
services, which contain a variety of services with different levels of 
cost which may reasonably have a range of copayments.
    As discussed in section II.A/B.4 in the proposed rule and this 
final rule, CMS uses (and will continue to use under revisions 
finalized for Sec. Sec.  422.100 and 422.101) Medicare FFS data in 
setting limits and thresholds for MA cost sharing for the basic 
benefits (that is, the Part A and Part B services that MA plans must 
cover). Medicare FFS data currently represents the most relevant and 
available data at this time. CMS uses it as well to evaluate the cost 
sharing for specific services, apply the anti-discrimination standard 
currently at Sec.  422.100(f)(6), and consider whether to exercise 
CMS's authority to add (by regulation) categories of services for which 
cost sharing may not exceed levels in Medicare FFS.
    As noted with regard to setting MOOP limits under Sec. Sec.  
422.100 and 422.101, CMS may consider future rulemaking regarding the 
use of MA encounter data to understand program health care costs and 
compare to Medicare FFS data in establishing cost sharing limits. 
Therefore, in addition to proposing to codify use of the FFS data, CMS 
proposed to include in Sec.  422.100(f)(6) that CMS would use MA 
encounter data to inform utilization scenarios used to identify 
discriminatory cost sharing.
    CMS explained that its proposal to amend Sec.  422.100(f) would 
allow use of the most relevant and appropriate information in 
determining cost sharing standards and thresholds. For example, 
analyses of MA utilization encounter data can be used with Medicare FFS 
data to establish the appropriate utilization scenarios to determine MA 
plan cost sharing standards and thresholds. CMS solicited comments and 
suggestions on this proposal, particularly whether additional 
regulation text is needed to achieve CMS's goal of setting and 
announcing each year presumptively discriminatory levels of cost 
sharing.
    We received the following comments on this proposal, and our 
response follows,
    Nearly all commenters who provided feedback on this provision (Cost 
Sharing Limits for Medicare Parts A and B Services (Sec.  
422.100(f)(6))) also provided feedback on section II.B. 4 (Maximum Out-
of-Pocket Limit for Medicare Parts A and B Services (Sec. Sec.  
422.100(f)(4) and (5) and 422.101(d))). In this section, we address 
commenters that primarily focus on cost sharing limits, while section 
II.B.4 addresses commenters that focus on MOOP limits or both of these 
provisions.
    Comment: The majority of commenters supported the proposal, stating 
that CMS should use Medicare FFS data to establish non-discriminatory 
cost sharing limits as it is currently the most relevant and 
appropriate information in determining cost sharing standards and 
thresholds. Commenters also supported providing guidance through the 
annual Call Letter to achieve CMS's goal of setting and announcing each 
year presumptively discriminatory levels of cost sharing that will not 
be considered discriminatory or in violation of other applicable 
standards.
    Response: We thank the commenters for their support. CMS intends to 
continue the practice of furnishing

[[Page 16490]]

information to MA organizations about the methodology used to establish 
cost sharing limits and the thresholds CMS identifies as non-
discriminatory through the annual Call Letter process. We will also 
continue to solicit comments before finalizing guidance as necessary 
and appropriate. Addressing changes in these vehicles that solicit 
comments provides for more timely and effective changes to protect 
beneficiaries. We expect this process will allow MA organizations to 
prepare plan bids consistent with parameters that CMS have determined 
to be non-discriminatory. In addition, and as appropriate, CMS will 
announce and issue guidance using HPMS memoranda.
    Comment: Many commenters were concerned about the quality of MA 
encounter data and questioned whether such data should be used to 
establish cost sharing limits. A few commenters were concerned about 
using MA encounter data to inform utilization scenarios, as proposed, 
based on data quality issues. A commenter proposed that CMS consider 
using a phased in approach over multiple years by blending Medicare FFS 
and MA encounter data for utilization analyses to address data quality 
concerns.
    Response: We understand the concerns expressed by commenters about 
using MA encounter data to estimate costs associated with specific 
health care services. However, we believe MA encounter data can be used 
to understand utilization trends in establishing the utilization 
scenarios selected for cost sharing standards (for example, 6-day and 
10-day inpatient cost sharing standards). Medicare FFS data currently 
represents the most relevant and available data at this time but we 
believe adding MA encounter data to FFS data will improve our 
utilization scenarios for the MA population. CMS may consider future 
rulemaking to incorporate MA encounter data with Medicare FFS data to 
establish cost sharing limits as well. Under this final rule, CMS will 
use Medicare FFS data along with MA encounter data to help inform 
utilization scenarios (for example, inpatient lengths of stay) in 
establishing cost sharing standards as we continue to rely on Medicare 
FFS data to determine cost sharing dollar limits. We believe the use of 
MA encounter data to inform utilization scenarios is reasonable as we 
are using it in conjunction with Medicare FFS data, which mitigates 
concerns about the completeness and quality of the MA encounter data.
    Comment: Several commenters were concerned about CMS's strategy to 
promote plan adoption of lower MOOP limits by increasing the cost 
sharing flexibility for those plans. Commenters expressed concern that 
allowing this flexibility may result in discriminatory benefit designs 
as plans may raise cost sharing limits for certain service categories 
more likely to be utilized by vulnerable beneficiaries. Some commenters 
referenced specific service categories of concern if cost sharing 
limits were raised (for example, inpatient and professional services) 
and requested CMS be especially thoughtful when considering changes to 
these categories.
    Response: CMS agrees that while increasing flexibility in cost 
sharing standards for plans that voluntarily offer lower MOOP limits 
can allow for improved plan design, it will be important to make sure 
that vulnerable patient populations are not discriminated against and 
that plan designs are not confusing to beneficiaries. CMS will manage 
the flexibility plans have in setting cost sharing limits to make sure 
that plan designs are not discriminatory.
    Comment: Some commenters noted concern with the specific 
methodology that CMS would use to establish cost sharing limits and how 
abrupt any changes may be from one contract year to the next. A few 
commenters requested CMS provide additional guidance on its 
implementation of the proposed changes to Sec.  422.100(f)(6).
    Response: CMS intends to use the annual Call Letter process to 
communicate its application of the regulation and to transition changes 
to cost sharing standards over time, beginning no earlier than CY 2020, 
to avoid disruption to benefit designs and minimize potential 
beneficiary confusion. Consistent with past practice, CMS will continue 
to publish annual limits, expected changes for the next year, and a 
description of how the regulation standard is applied (that is, the 
methodology used) in the annual Call Letter prior to bid submission so 
that MA plans can submit bids consistent with CMS standards. This will 
provide MA organizations adequate time to comment and prepare for 
changes.
    We received over 40 comments pertaining to the proposal, with the 
majority reflecting support to amend Sec.  422.100(f)(6) to permit use 
of Medicare FFS data to establish cost sharing limits that will not be 
considered discriminatory for Part A and B services in MA plans. 
Commenters also generally supported continued use of the annual Call 
Letter process for explaining our application and implementation of the 
revised Sec.  422.100(f)(6). After careful consideration of all the 
comments, we are finalizing our proposal to use Medicare FFS data along 
with MA encounter data to inform utilization scenarios (for example, 
inpatient lengths of stay) and rely on Medicare FFS data to determine 
cost sharing standards and thresholds. We are finalizing these 
amendments with an applicability date of January 1, 2020; this 
applicability date is consistent with our intent that these new 
standards apply to cost sharing limits set for plans years after 2019. 
As MA encounter cost data quality improves, CMS will consider future 
rulemaking to incorporate with Medicare FFS data to establish cost 
sharing limits. CMS intends to use the annual Call Letter process to 
communicate its application of the regulation and plans to transition 
changes under the finalized regulations over time, beginning no earlier 
than CY 2020, to avoid disruption to benefit designs and minimize 
potential beneficiary confusion. We are also finalizing a minor 
revision to paragraph (f)(6) to improve the flow of the text. 
Specifically, we are separating the last sentence into two sentences 
divided by a semicolon with minor grammatical edits.
6. Meaningful Differences in Medicare Advantage Bid Submissions and Bid 
Review (Sec. Sec.  422.254 and 422.256)
    As provided at Sec. Sec.  422.254(a)(4) and 422.256(b)(4), CMS will 
only approve a bid submitted by a Medicare Advantage (MA) organization 
if its plan benefit package (PBP) is substantially different from those 
of other plans offered by the organization in the same area with 
respect to key plan characteristics such as premiums, cost sharing, or 
benefits offered. MA organizations may submit bids for multiple plans 
in the same area under the same contract only if those plans are 
substantially different from one another based on CMS's annual 
meaningful difference evaluation. CMS proposed to eliminate the 
meaningful difference requirement beginning with MA bid submissions for 
contract year (CY) 2019. Separate meaningful difference rules were 
concurrently adopted for MA and stand-alone prescription drug plans 
(PDPs), but this specific proposal was limited to the meaningful 
difference provision related to the MA program. A proposal related to 
the Part D meaningful difference regulation is addressed at section 
III. II.A.16. of this final rule.
    In the proposed rule, CMS explained the goal of eliminating the 
meaningful difference requirement: To improve competition, innovation, 
available

[[Page 16491]]

benefit offerings, and provide beneficiaries with affordable plans that 
are tailored for their unique health care needs and financial 
situation. Other regulations prohibit plans from misleading 
beneficiaries in their communication materials, provide CMS the 
authority to disapprove a bid if a plan's proposed benefit design 
substantially discourages enrollment in that plan by certain Medicare-
eligible individuals, and allow CMS to non-renew a plan that fails to 
attract a sufficient number of enrollees over a sustained period of 
time (Sec. Sec.  422.100(f)(2), 422.510(a)(4)(xiv), 422.2264, and 
422.2260(e)). Therefore, CMS explained in the proposed rule, MA 
organizations could be expected to continue designing PBPs that, within 
a service area, are different from one another with respect to key 
benefit design characteristics. CMS stated its belief that any 
potential beneficiary confusion would be minimized when comparing 
multiple plans offered by the MA organization. For example, 
beneficiaries may consider the following factors when they make their 
health care decisions: Plan type, Part D coverage, differences in 
provider network, Part B and plan premiums, and unique populations 
served (for example, special needs plans). In addition, CMS stated its 
intent to continue the practice of furnishing information to MA 
organizations about the bid evaluation methodology through the annual 
Call Letter process and/or Health Plan Management System (HPMS) 
memoranda and solicit comments, as appropriate. This process allows CMS 
to articulate bid requirements and MA organizations to prepare bids 
that satisfy CMS requirements and standards prior to bid submission in 
June each year.
    As stated in the proposed rule, although challenged by choices, 
beneficiaries do not want their plan choices to be limited and 
understand key decision factors such as premiums, out-of-pocket cost 
sharing, Part D coverage, familiar providers, and company offering the 
plan.\22\ CMS noted that more sophisticated approaches to consumer 
engagement and decision-making should help beneficiaries, caregivers, 
and family members make informed plan choices. CMS cited supporting 1-
800-MEDICARE and enhancements to MPF that have improved the customer 
experience, such as including MA and Part D benefits and a new consumer 
friendly tool for the CY 2018 Medicare open enrollment period. This new 
tool assists beneficiaries in choosing a plan that meets their unique 
health and financial needs based on a set of 10 quick questions.
---------------------------------------------------------------------------

    \22\ Jacobson, G. Swoope, C., Perry, M. Slosar, M. How are 
seniors choosing and changing health insurance plans? Kaiser Family 
Foundation. 2014
---------------------------------------------------------------------------

    As stated in the October 22, 2009, proposed rule (74 FR 54670 
through 73) and April 15, 2010, final rule (75 FR 19736 through 40), 
CMS's goal for the meaningful difference evaluation was to ensure a 
proper balance between affording beneficiaries a wide range of plan 
choices and avoiding undue beneficiary confusion in making coverage 
selections. The meaningful difference evaluation was initiated when 
cost sharing and benefits were relatively consistent within each plan, 
and similar plans within the same contract could be readily compared by 
measuring estimated out-of-pocket costs (OOPC) and other factors 
currently integrated in the evaluation's methodology. Detailed 
information about the meaningful difference evaluation is available in 
the CY 2018 Final Call Letter issued April 3, 2017, (pages 115-118) and 
information about the CMS OOPC model is available at: https://www.cms.gov/Medicare/Prescription-Drug-Coverage/PrescriptionDrugCovGenIn/OOPCResources.html. As discussed in the CY 
2018 Final Call Letter, the differences between similar plans must have 
at least a $20 per member per month estimated beneficiary out-of-pocket 
cost difference. Differences in plan type (for example, HMO, LPPO), SNP 
sub-type, and inclusion of Part D coverage are considered meaningful 
differences, which align with beneficiary decision-making. As noted in 
the proposed rule, premiums, risk scores, actual plan utilization, and 
enrollment are not included in the evaluation because these factors 
will introduce risk selection, costs, and margin into the evaluation, 
resulting in a negation of the evaluation's objectivity. CMS clarified 
that the OOPC model uses the lowest cost sharing value for each service 
category to estimate out-of-pocket costs, which may or may not be a 
relevant comparison between different plans for purposes of evaluating 
meaningful difference when variable cost sharing of this type is 
involved.
    Based on CMS's efforts to revisit MA standards and the 
implementation of the governing law to find flexibility for MA 
beneficiaries and plans, MA organizations are able to: (1) Tier the 
cost sharing for contracted providers as an incentive to encourage 
enrollees to seek care from providers the plan identifies based on 
efficiency and quality data which was communicated in CY 2011 guidance; 
(2) establish Provider-Specific Plans (PSPs) designed to offer 
enrollees benefits through a subset of the overall contracted network 
in a given service area, which are sometimes referred to as narrower 
networks, and which was collected in the PBP beginning in CY 2011; and 
(3) beginning in CY 2019, provide different cost sharing and/or 
additional supplemental benefits for enrollees based on defined health 
status or disease state within the same plan (Flexibility in the 
Medicare Advantage Uniformity Requirements). These flexibilities allow 
MA organizations to provide beneficiaries with access to health care 
benefits that are tailored to individual needs, but make it difficult 
for CMS to objectively measure meaningful differences between plans. 
Items 1 and 3 provide greater cost sharing flexibility to address 
individual beneficiary needs but result in a much broader range of cost 
sharing values being entered into the PBP.
    CMS restated its commitment to ensuring transparency in plan 
offerings so that beneficiaries can make informed decisions about their 
health care plan choices while also noting the importance of 
encouraging competition, innovation, and providing access to affordable 
health care approaches that address individual needs. CMS recognized 
that the current meaningful difference methodology evaluates the entire 
plan and does not capture differences in benefits that are tied to 
specific health conditions. As a result, CMS noted the meaningful 
difference evaluation will not fully represent benefit and cost sharing 
differences experienced by enrollees and could lead to MA organizations 
to focus on CMS standards, rather than beneficiary needs, when 
designing benefit packages. CMS noted the challenges with trying to 
capture differences in provider network, more tailored benefit and cost 
sharing designs, or other innovations. In addition, we are concerned 
that plans may be forced to potentially develop more complicated and 
confusing benefit designs to achieve differences between plans.
    CMS recognized to satisfy current CMS meaningful difference 
standards, MA organizations may have to change benefit coverage or cost 
sharing in certain plans to establish the necessary benefit value 
difference, even if substantial difference exists based on factors CMS 
is currently unable to incorporate into the evaluation (such as tiered 
cost sharing, and unique benefit packages based on enrollee health 
conditions). Although these changes in benefits coverage may be 
positive or negative, CMS stated concern that the meaningful difference 
requirement

[[Page 16492]]

results in organizations potentially reducing the value of benefit 
offerings. These are unintended consequences of the existing meaningful 
difference evaluation and may restrict innovative benefit designs that 
address individual beneficiary needs and affordability.
    As discussed in the proposed rule, CMS continually evaluates 
consumer engagement tools and outreach materials (including marketing, 
educational, and member materials) to ensure information is formatted 
consistently so beneficiaries can easily compare multiple plans. Annual 
guidance and model materials are provided to MA organizations to assist 
them in providing resources, such as the plan's Annual Notice of Change 
(ANOC) and Evidence of Coverage (EOC), which contain valuable 
information for the enrollee to evaluate and select the best plan for 
their needs. CMS invests substantial resources in engagement strategies 
such as 1-800-MEDICARE, MPF, standard and electronic mail, and social 
media to continuously communicate with beneficiaries, caregivers, 
family members, providers, community resources, and other stakeholders.
    CMS noted that MA organizations may be able to offer a portfolio of 
plan options with clear differences between benefits, providers, and 
premiums which will allow beneficiaries to make more effective 
decisions if the MA organizations are not required to change benefit 
and cost sharing designs in order to satisfy Sec. Sec.  422.254 and 
422.256. Currently, MA organizations must satisfy CMS meaningful 
difference standards (and other requirements), rather than solely 
focusing on beneficiary purchasing needs when establishing a range of 
plan options. CMS also noted additional beneficiary protections 
including: Plans are required to not mislead beneficiaries in 
communication materials; CMS may disapprove a bid if CMS finds that a 
plan's proposed benefit design substantially discourages enrollment in 
that plan by certain Medicare-eligible individuals; and CMS may 
terminate plans that fail to attract a sufficient number of enrollees 
over a sustained period of time (Sec. Sec.  422.100(f)(2), 
422.510(a)(4)(xiv), 422.2264, and 422.2260(e)). For these reasons, CMS 
proposed to remove Sec. Sec.  422.254(a)(4) and 422.256(b)(4) to 
eliminate the meaningful difference requirement for MA bid submissions. 
CMS also solicited comments and suggestions on making sure 
beneficiaries have access to innovative plans that meet their unique 
needs.
    We received the following comments on this proposal, and our 
response follows:
    Comment: Some commenters fully supported the proposal, stating that 
eliminating the meaningful difference requirement will support plan 
innovation and provide Medicare beneficiaries access to plans that meet 
their unique needs. Several commenters noted that eliminating the 
current meaningful difference requirement that established arbitrary 
differences between plans will allow MA organizations to put the 
beneficiary at the center of benefit design. This will result in MA 
organizations being able to offer a portfolio of plan options with 
clear differences between benefits, providers, and premiums that are 
easily understood by beneficiaries. Commenters also noted that CMS's 
efforts to support beneficiaries make informed choices by maintaining 
existing requirements for marketing materials and nondiscriminatory 
benefit designs will sufficiently safeguard beneficiaries if the 
meaningful difference requirement is eliminated.
    Response: We thank the commenters for supporting the proposal. We 
believe this proposed change could result in more innovative products 
that are more competitive and market-driven within a less restrictive 
regulatory framework.
    Comment: A commenter supported the proposal and questioned how the 
agency will ensure potential savings from eliminating the meaningful 
difference requirement will be passed on to beneficiaries in the form 
of lower premiums, while also maintaining coverage of essential and 
appropriate benefits.
    Response: CMS expects that the elimination of the meaningful 
difference evaluation, in conjunction with the expansion of benefit 
flexibilities, will allow organizations to provide benefit offerings 
that satisfy the unique needs of beneficiaries, increase enrollee 
satisfaction, reduce overall plan expenditures, and result in more 
affordable plans. All MA plans must provide enrollees in that plan with 
all Parts A and B services so beneficiaries are assured a minimum 
package of covered services; many plans also provide supplemental 
benefits, at the MA organization's option. While CMS reviews and 
approves MA PBPs and premiums for actuarial soundness and satisfying 
CMS standards, we do not have the legal authority to dictate MA 
organizations' business decisions to establish premiums at a specific 
level. MA organizations can adjust their plan offerings to reflect 
annual changes in medical costs and payment rates and may do so in a 
variety of ways, such as adjustments to cost sharing amounts, adding or 
subtracting supplemental benefits, or making changes to the monthly 
premium(s). Plans face competition in their defined market areas and 
must also comply with Part C standards related to changes in benefits, 
cost sharing, and premium. In addition, all beneficiaries are made 
aware of plan changes including premium for the upcoming year and can 
choose to switch plans during the annual election period.
    Comment: Several commenters disagreed with the proposal to 
eliminate the meaningful difference requirement because they believe it 
is a beneficiary protection. Reasons for maintaining the meaningful 
difference requirement included: Concerns about the ability of Medicare 
beneficiaries to make the nuanced comparisons among various plan types 
and benefit packages, limited resources to assist beneficiaries with 
complicated decisions, expectation that older people and people with 
disabilities do not use technology to the same extent as non-Medicare 
beneficiary populations (thereby limiting the usefulness of MPF, a 
primary means of CMS assistance to beneficiaries in comparing plans), 
and unknown resource availability to support call centers to assist 
beneficiaries who do not have access to or use the internet. Several 
comments were concerned that narrower networks could be potentially 
discriminatory or a means of limiting benefit access for enrollees. 
Another commenter had concerns that eliminating the meaningful 
difference requirement may encourage plan risk segmentation based on 
benefit design but did not include any rationale for their concern. 
Some commenters referenced plan selection research, such as National 
Institutes of Health, and Brookings studies,\23\ noting Consumers Union 
findings that indicate beneficiaries face challenges in navigating the 
Medicare market due to not using available tools (such as MPF),

[[Page 16493]]

confusion when using MPF, and high rates of individuals not making an 
active health plan selection because of choice anxiety. Several 
commenters also noted their general concern that the net effect of 
eliminating the meaningful difference requirement and other proposals 
pursued in the proposed rule may have unintended consequences regarding 
beneficiary confusion that will negate the value of market innovation, 
especially for people with lower income and educational levels.
---------------------------------------------------------------------------

    \23\ Bertko J, Ginsburg PB, Lieberman S, Trish E, Antos J. 
Medicare Advantage: Better information tools, better beneficiary 
choices, better competition. U.S.C.-Brookings Schaeffer Initiative 
for Health Policy. Nov. 2017. Retrieved from https://www.brookings.edu/wp-content/uploads/2017/11/ma-consumer-reforms.pdf.
    Cognitive Functioning and Choice between Traditional Medicare 
and Medicare Advantage; J. Michael McWilliams, Christopher C. 
Afendulis, Thomas G. McGuire, and Bruce E. Landon; Health Affairs, 
September 2011 (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3513347/
).
    The Evidence is Clear: Too Many Health Insurance Choices Can 
Impair, Not Help Consumer Decision Making; Lynn Quincy and Julie 
Silas; Consumers Union, November 2012 (http://consumersunion.org/pdf/Too_Much_Choice_Nov_2012.pdf).
---------------------------------------------------------------------------

    Response: We acknowledge the commenters' concerns about beneficiary 
confusion. We believe that the tools CMS provides for beneficiaries to 
make decisions and our enforcement of communication and marketing 
requirements (such as the prohibition on misleading beneficiaries) 
mitigate and address these concerns. Under our existing authority at 
Sec.  422.110, CMS will monitor to ensure organizations are not 
engaging in activities that are discriminatory or potentially 
misleading or confusing to Medicare beneficiaries. We note that CMS has 
authority, clarified in this final rule, to review marketing (review in 
advance of use) and communication (review after use) materials to 
ensure compliance with MA program requirements. CMS will conduct 
outreach with organizations that appear to offer a large number of 
similar plans in the same county following bid submissions and 
communicate any general concerns through the annual Call Letter process 
and/or HPMS memoranda. CMS network adequacy requirements apply to all 
Part C provider networks to ensure adequate network provider access for 
enrollees. With regard to concerns about risk segmentation, CMS 
believes risk segmentation is not beneficial to MA organizations or 
enrollees who want to maintain stable benefits and premiums, but if an 
organization wanted to purposely create risk segmentation within its 
plan offerings, it could do so with or without the meaningful 
difference evaluation. The agency will continue to monitor and address 
potential concerns as part of our existing authority to review and 
approve bids. We expect eliminating the meaningful difference 
requirement will improve plan choices for beneficiaries by driving 
provider network and benefit package innovation and affordable health 
care coverage. MA organizations also consider beneficiary choice 
anxiety when developing their own portfolio of plan offerings, so that 
sales and broker personnel and marketing materials can highlight key 
differences between plan offerings and support informed choice. 
Beneficiaries also rely on established health plan characteristics to 
guide their decision making, such as preferences for plan type (for 
example, HMO or PPO), providers (for example, established primary care 
physician being in network), presence of Part D benefits, cost sharing, 
plan premium, and brand.\24\ In addition, dually eligible beneficiaries 
may choose D-SNPs that provide more standardized plan options with 
little or no cost sharing responsibilities instead of a non-D-SNP plan 
without these benefits. This allows beneficiaries to reduce the number 
of health plan options of interest (for example, focus on MA 
organizations offering SNP options) and simplify the process to choose 
their health plan. After taking into account specific preferences, such 
as plan type, beneficiaries may choose from a limited subset of 
available plan options with the assistance of plan communication 
materials and existing CMS resources such as MPF and 1-800-MEDICARE. In 
addition, CMS will continue to prohibit plans from misleading 
beneficiaries in their communication materials, disapprove a plan's bid 
if its proposed benefit design substantially discourages enrollment in 
that plan by certain Medicare-eligible individuals, and allow CMS to 
terminate a plan that fails to attract a sufficient number of enrollees 
over a sustained period of time so that any potential beneficiary 
confusion is minimized when comparing multiple plans offered by the 
organization (Sec. Sec.  422.100(f)(2), 422.510(a)(4)(xiv), 422.2264, 
and 422.2260(e)).
---------------------------------------------------------------------------

    \24\ Jacobson, G., Swoope, C., Perry, M., Slosar, M. How are 
seniors choosing and changing health insurance plans? Kaiser Family 
Foundation. 2014.
    Atherly, A., Dowd, B., Feldman, R. The Effect of Benefits, 
Premiums, and Health Risk on Health Plan Choice in the Medicare 
Program. Health Services Research. 2004. Retrieved from https://onlinelibrary.wiley.com/doi/full/10.1111/j.1475-6773.2004.00261.x.
    McCormack LA, Garfinkel SA, Hibbard JH, Norton EC, Bayen UJ. 
Health plan decision making with new medicare information materials. 
Health Services Research. 2001;36(3):531-554.
    Abaluck, Jason, and Jonathan Gruber. 2011. ``Choice 
Inconsistencies among the Elderly: Evidence from Plan Choice in the 
Medicare Part D Program.'' American Economic Review, 101(4): 1180-
1210.
    Uhrig, J., Harris-Kojetin, L., Bann, C., Kuo, T. Do Content and 
Format Affect Older Consumers' Use of Comparative Information in a 
Medicare Health Plan Choice? Results from a Controlled Experiment. 
2006. Retrieved from http://journals.sagepub.com/doi/pdf/10.1177/1077558706293636.
---------------------------------------------------------------------------

    Comment: Several commenters had concern that eliminating the 
meaningful difference requirement would promote ``gaming'' among plan 
sponsors (for example, offering a large number of plan options in a 
service area) which may challenge or complicate beneficiary decision-
making because of the potential increase in plan options; these 
commenters questioned if elimination of the requirement provides enough 
benefits to outweigh the risks. A few commenters questioned whether 
there is evidence that innovation is or will be inhibited by the 
meaningful difference evaluation. A commenter recommended CMS formally 
survey MA organizations about the impact of meaningful difference 
standards as well as survey beneficiaries regarding their satisfaction 
with MA plan offerings. Some commenters suggested CMS first pursue 
adjusting the meaningful difference requirement before eliminating it 
by either waiving the requirement if MA organizations can provide 
alternative evidence to CMS that their plan offerings are substantively 
different, significantly reducing the current $20 meaningful difference 
threshold between similar plans to provide more flexibility, accounting 
for differences in premiums, and providing broader consideration of 
provider network differences in the evaluation. A commenter requested 
that instead of eliminating the meaningful difference requirement, CMS 
revise the evaluation and require plan actuaries to attest to actuarial 
value differences among plans using a utilization profile that is 
representative of the plan population. A few comments stated that if 
CMS was to place a limit on the number of plans an organization could 
offer that CMS take into consideration the appropriate level within an 
organizational structure to establish the limit (for example, parent, 
legal entity, or contract organization), mergers and acquisitions, and 
that CMS treat full-provider networks separately from more limited 
provider networks.
    Response: As discussed in the proposed rule, CMS is concerned the 
meaningful difference requirement may force MA organizations to design 
benefit packages to meet CMS standards rather than address beneficiary 
needs. CMS has been made aware of these concerns through comments 
submitted in response to recent Call Letters and the Request for 
Information (April 2017), that highlighted how MA organizations may be 
forced to meet arbitrary limits between their plans to comply with CMS 
meaningful difference standards. Based on this information CMS does not 
believe formal surveys are necessary to determine the unintended 
consequences of the meaningful difference evaluation. Our proposal to 
eliminate the meaningful difference requirement aimed to improve 
competition, innovation, available benefit offerings, and provide 
beneficiaries with

[[Page 16494]]

affordable plans that are tailored for their unique health care needs 
and financial situation. The number of MA plan bids may increase 
because of a variety of factors, that are not related to the 
elimination of the meaningful difference requirement, such as payments, 
bidding and service area strategies, serving unique populations, and in 
response to other program constraints or flexibilities. CMS expects 
that eliminating the meaningful difference requirement will improve 
plan choice for beneficiaries by driving provider network and benefit 
package innovation and affordable health care coverage. CMS believes 
that eliminating the current meaningful difference requirement will 
allow MA organizations to put the beneficiary at the center of benefit 
design as MA organizations will not be pressured to make benefit 
changes to comply with an arbitrary requirement that may ultimately 
result in higher premiums and/or cost sharing for beneficiaries. This 
will result in MA organizations being able to offer a portfolio of plan 
options with clear differences between benefits, providers, and 
premiums that are more easily understood by beneficiaries. In order to 
capture differences in provider networks, more tailored benefit and 
cost sharing designs, or other innovations, the evaluation process 
would have to use more varied and complex assumptions to identify plans 
that are not meaningfully different from one another. CMS believes that 
such an evaluation could result in more complicated and potentially 
confusing benefit designs and would require investment of greater 
administrative resources for MA organizations and CMS, while not 
producing results that are useful to beneficiaries. CMS expects that 
eliminating the meaningful difference requirement will improve the plan 
options available for beneficiaries. As it is unknown how many 
organizations will choose to add plan options as a result of this 
provision, we are unable to estimate the impact to beneficiaries should 
this lead to more competition. CMS expects increased competition will 
lead to potentially lower premiums and/or cost-sharing for Medicare 
beneficiaries. CMS does not anticipate beneficiaries will need 
additional time to compare differences between plans related to the 
elimination of the meaningful difference requirement. This particular 
change is expected to help MA organizations differentiate plan 
offerings more effectively so that beneficiaries can make decisions 
more efficiently. We believe that the tools and information CMS 
provides for beneficiaries to make decisions (for example, Medicare 
Plan Finder, Medicare and You Handbook, 1-800-MEDICARE), in addition to 
our enforcement of communication and marketing requirements, aim to 
mitigate any potential choice overload. We are not pursuing adjustments 
to the meaningful difference requirement (for example, waivers) because 
the use of a waiver or justification process introduces subjectivity 
into the benefit review and we believe the goal of increasing 
flexibility is better served by eliminating the requirement. With this 
final rule, organizations will have more flexibility to design MA plans 
in a manner that is more focused on beneficiary needs. Finally, we do 
not intend to establish a specific number of plans that any one 
organization could offer. The MA program has a different market 
structure than standalone PDPs, that is, PDPs serve entire regions 
while MA organizations may serve different service areas based on 
county. The same MA organization may have multiple plans but those 
plans may only overlap in a limited number of counties. Depending on 
the market structure (for example, makeup of providers and consumers) 
it may be helpful for MA organizations to provide offerings from 
multiple plan types so that beneficiaries have valuable options. In 
addition, it may be helpful for MA organizations to offer SNP plans to 
meet the needs of different beneficiary populations. CMS will monitor 
and address potential concerns as part of our existing authority to 
review and approve bids.
    Comment: A few commenters requested that CMS conduct an evaluation 
to estimate whether eliminating the meaningful difference requirement 
would create choice anxiety among beneficiaries and its potential 
effect on future enrollment. A few commenters also questioned if CMS 
had presented sufficient reasons to justify eliminating the meaningful 
difference requirement.
    Response: In the proposed rule (82 FR 56363 through 56365) and in 
the responses in this section, we have discussed our supporting 
rationale to eliminate the meaningful difference requirement. After 
carefully considering the commenters' concerns, we believe our proposal 
will result in improved options--both in terms of innovative plans and 
affordability--for beneficiaries and that existing safeguards, along 
with beneficiary decision making education and tools, will be 
successful in managing beneficiary choice anxiety concerns.
    Comment: A commenter requested clarification on how this proposal, 
in conjunction with others, affects expectations for state Medicaid 
agencies and SNPs.
    Response: CMS does not anticipate that eliminating the meaningful 
difference requirement, in conjunction with other proposals, would 
affect state Medicaid agencies. To the extent that clarification of 
state Medicaid or SNP issues is required as a result of the regulation 
changes in this final rule, CMS would communicate this guidance through 
the annual Call Letter process, HPMS memoranda, and Medicare Managed 
Care Manual updates. In addition, the CMS Medicare-Medicaid 
Coordination Office (MMCO) may provide assistance for states and D-
SNPs. The Center for Medicare is working collaboratively with MMCO in 
the regulations drafting process and implementation steps related to 
this rule. Separately, MMCO is re-examining the potential need for 
resources related to implementing the provisions of section 50311 of 
the Bipartisan Budget Act of 2018.
    Comment: Several commenters requested that CMS issue guidance 
regarding the distinctions in plan options that would be permissible 
and operational guidance on the implementation of this proposal in the 
annual Call Letter to support CY 2019 bid development and submission.
    Response: MA organizations can use the information contained in 
this final rule about the elimination of the meaningful difference 
requirements and CMS expectations to prepare CY 2019 bid submissions. 
CMS intends to continue using the annual Call Letter process in future 
years for releasing draft versions of bid-related guidance for comment 
and to provide additional guidance regarding general concerns we may 
have with organizations' portfolio of plan offerings. In addition, we 
will provide information about potential concerns regarding activities 
that are potentially discriminatory or potentially misleading or 
confusing to Medicare beneficiaries.
    Comment: Several commenters noted concern about resources to 
support beneficiaries choose a health plan and navigate their benefits 
(for example, 1-800-MEDICARE, MPF, SHIP counselors, and the Medicare 
Ombudsman program) and supported improvements to MPF that allow 
beneficiaries to more easily narrow down their choices based on 
personalized information (for example, more filters and pre-selection 
criteria to identify important plan characteristics that limit plan 
options to evaluate). Several commenters offered to provide

[[Page 16495]]

input to MPF changes, while others encouraged CMS to establish a group 
of representatives (for example, MA organizations, advocacy 
organizations, provider groups, and other stakeholders) to help develop 
MPF improvements, health plan decision-making education materials, and 
other information to improve the health plan selection process and 
overall experience for beneficiaries. Some comments indicated that 
changes to the MPF should occur prior to eliminating the meaningful 
difference evaluation. Commenters also had an interest in CMS 
establishing communications and marketing guidance so that MA 
organizations can describe how an organization's plan offerings are 
different in situations where multiple plan options are compared (for 
example, providing additional information in the Summary of Benefits). 
In addition, other comments noted the need for CMS to solicit input 
from multiple stakeholders to improve communication materials (for 
example, ANOC and EOC).
    Response: These recommendations are not strictly within the scope 
of this final rule provision. We do however appreciate the many 
comments and suggestions related to improving the health plan decision 
making process and overall experience for beneficiaries. We agree with 
the need for clear and complete information and intend to continue 
improving the MPF to make it as user friendly as possible. We are 
sharing these comments and suggestions with the CMS Office of 
Communications. Additionally, we would encourage third party 
organizations that support beneficiaries in their decision-making to 
take advantage of existing resources 1-800-MEDICARE, MPF, SHIP 
counselors, and the Medicare Ombudsman program. CMS will take commenter 
suggestions under careful consideration and will continue to include 
stakeholders and beneficiaries in the planning, preparation, testing, 
and execution process for MPF; CMS subjects some model enrollee 
communication materials to periodic consumer testing and also considers 
comments submitted from MA organizations and stakeholders on an ongoing 
basis. In addition, CMS will look for ways to incorporate the 
suggestions from commenters about how the health plan selection process 
can be simplified for beneficiaries through existing and possibly new 
Medicare materials. MA organizations have and are encouraged to use 
existing flexibilities to highlight differences between their own plan 
offerings for beneficiaries in marketing and communications materials 
(for example, summary of benefits).
    We received over 65 comments pertaining to the proposal; the great 
majority reflected mixed support for eliminating the meaningful 
difference requirement. After careful consideration of all of the 
comments we received, we are finalizing the elimination of the 
meaningful difference requirement from Sec. Sec.  422.254 and 422.256 
as proposed. Under our existing authority at Sec.  422.2268, CMS will 
monitor to ensure organizations are not engaging in activities that are 
discriminatory or potentially misleading or confusing to Medicare 
beneficiaries. CMS will communicate and work with organizations that 
appear to offer a large number of similar plans in the same county, 
raising and discussing with such MA organizations any concerns. CMS 
plan checks would include plans offered under each contract, unique 
plan type, and county. Plan types currently include: (1) HMO and HMO-
POS not offering all Parts A and B services out-of-network, (2) HMO POS 
offering all Parts A and B services out-of-network, (3) LPPO, (4) RPPO, 
(5) PFFS, and (6) unique SNP types (that is, different chronic 
diseases, institutional categories, and dual-eligible sub-types). From 
a beneficiary's perspective, CMS would expect plans within the same 
contract, plan type, and county to be distinguishable by beneficiaries 
using such factors as the inclusion or exclusion of Part D coverage, 
provider network, plan premium, Part B premium buy-down, estimated out-
of-pocket costs, and benefit design so that MA organizations can market 
their plans clearly. CMS intends to issue guidance through the annual 
Call Letter process and HPMS memoranda to help organizations design 
plan options that avoid potential beneficiary confusion prior to bid 
submission.
7. Coordination of Enrollment and Disenrollment Through MA 
Organizations and Effective Dates of Coverage and Change of Coverage 
(Sec. Sec.  422.66 and 422.68)
    In addition to general authority for the Secretary to establish the 
process through which MA plan election is made by Medicare 
beneficiaries, section 1851(c)(3)(A)(ii) of the Act authorizes the 
Secretary to implement default enrollment rules for the Medicare 
Advantage (MA) program. This default enrollment is in addition to the 
statutory direction that beneficiaries who do not elect an MA plan are 
defaulted to original (fee-for-service) Medicare. Section 
1851(c)(3)(A)(ii) states that the Secretary may establish procedures 
whereby an individual currently enrolled in a non-MA health plan 
offered by an MA organization at the time of his or her Initial 
Coverage Election Period is deemed to have elected an MA plan offered 
by the organization if he or she does not elect to receive Medicare 
coverage in another way. We proposed new regulation text to establish 
limits and requirements for these types of default enrollments to 
address our administrative experience with and concerns raised about 
these types of default enrollments under our existing practice. Based 
on our experience with the seamless conversion process thus far, we 
proposed to codify at Sec.  422.66(c)(2) requirements for seamless 
default enrollments upon initial eligibility for Medicare. As proposed, 
such default enrollments would be into dual eligible special needs 
plans (D-SNPs) and would be subject to five substantive conditions: (1) 
The state has approved use of this default enrollment process and 
provided Medicare eligibility information to the MA organization; (2) 
CMS has approved the MA organization to use the default enrollment 
process before any enrollments are processed; (3) the individual is 
enrolled in an affiliated Medicaid managed care plan and is dually 
eligible for Medicare and Medicaid; (4) the MA organization provides a 
notice that meets CMS requirements to the individual; and (5) the 
individual does not opt out of the default enrollment. We proposed that 
coverage under these types of default enrollments begin on the first of 
the month that the individual's Part A and Part B eligibility is 
effective. We also proposed changes to Sec. Sec.  422.66(d)(1) and 
(d)(5) and 422.68 that coordinate with the proposal for Sec.  422.66.
    As noted in the proposed rule, we initially addressed default 
enrollment upon conversion to Medicare in a 2005 rulemaking (70 FR 4606 
through 4607) and released subregulatory guidance \25\ to provide an 
optional enrollment mechanism in 2006. This mechanism permitted MA 
organizations to develop processes and, with CMS approval, provide 
seamless continuation of coverage by way of enrollment in an MA plan 
for newly MA eligible individuals who are currently enrolled in other 
health plans offered by the MA organization (such as commercial or 
Medicaid plans) at the time of the individuals' initial eligibility for 
Medicare. The guidance emphasized

[[Page 16496]]

that approved MA organizations not limit seamless continuation of 
coverage to situations in which an enrollee becomes eligible for 
Medicare by virtue of age, and directed MA organizations to implement 
seamless conversions to include all newly eligible Medicare 
beneficiaries, including those whose Medicare eligibility is based on 
disability. From its inception, the guidance required that individuals 
receive advance notice of the proposed MA enrollment and have the 
ability to ``opt out'' of such an enrollment prior to the effective 
date of coverage. This guidance has been in practice for the past 
decade, but we encountered complaints and heard concerns about the 
practice.
---------------------------------------------------------------------------

    \25\ https://www.cms.gov/Medicare/Eligibility-and-Enrollment/MedicareMangCareEligEnrol/Downloads/CY_2018_MA_Enrollment_and_Disenrollment_Guidance_6-15-17.pdf.
---------------------------------------------------------------------------

    The Advance Notice of Methodological Changes for Calendar Year (CY) 
2016 for Medicare Advantage (MA) Capitation Rates, Part C and Part D 
Payment Policies and 2016 Call Letter discussed the opportunity to 
integrate Medicare and Medicaid benefits via seamless continuation of 
coverage into D-SNPs, and we received positive comments from state 
Medicaid agencies supporting this enrollment mechanism and requesting 
clarification of the approval process. We also received comments from 
beneficiary advocates asking for additional consumer protections (for 
example, requiring written beneficiary confirmation and a special 
enrollment period for those enrolled using this optional mechanism).
    On October 21, 2016,\26\ in response to inquiries regarding this 
enrollment mechanism, its use by MA organizations, and the beneficiary 
protections currently in place, we announced a temporary suspension of 
acceptance of new proposals for seamless continuation of coverage. We 
discovered, based on our subsequent discussions with beneficiary 
advocates and MA organizations approved for this enrollment mechanism, 
that MA organizations find it difficult to comply with our current 
guidance and approval parameters, especially the requirement to 
identify commercial members who are approaching Medicare eligibility 
based on disability when the other plan offered by the MA organization 
is a commercial insurance plan. MA organizations also outlined 
challenges in confirming entitlement to Medicare Parts A and B within 
necessary timeframes and obtaining the individual's Medicare number--
which in 2018 will become a random and unique number instead of the 
Social Security Number-based identifier used today. As discussed in 
more detail below, we anticipate that the switch from the SSN-based 
identifier will exacerbate this difficulty.
---------------------------------------------------------------------------

    \26\ https://www.cms.gov/Medicare/Eligibility-and-Enrollment/MedicareMangCareEligEnrol/Downloads/HPMS_Memo_Seamless_Moratorium.pdf.
---------------------------------------------------------------------------

    We noted in the proposed rule how organizations operating Medicaid 
managed care plans are better able to meet these requirements when 
states provide data, including the individual's Medicare number, to 
identify individuals about to become Medicare eligible; MA 
organizations with state contracts to offer D-SNPs will be able to 
obtain (under their agreements with state Medicaid agencies) the data 
necessary to process and submit default enrollments to CMS without 
needing to collect information from the Medicare beneficiaries. 
Therefore, we proposed to revise Sec.  422.66 to permit default 
enrollment only for Medicaid managed care enrollees who are newly 
eligible for Medicare and who are enrolled into a D-SNP administered by 
an MA organization with the same parent organization as the 
organization that operates the Medicaid managed care plan in which the 
individual remains enrolled. At Sec.  422.66(c)(2)(i)(B), we also 
proposed to limit these default enrollments to situations where the 
state has actively facilitated and approved the MA organization's use 
of this enrollment process and articulates this in the agreement with 
the MA organization offering the D-SNP and by providing necessary 
identifying information to the MA organization.
    The proposal was designed to support state efforts to increase 
enrollment of dually eligible individuals into fully integrated systems 
of care There is evidence \27\ that such systems improve health 
outcomes so supporting efforts to increase use those systems is 
consistent with overall CMS policy. Further, we believe then, and now, 
that the proposal provided states with additional flexibility and 
control.
---------------------------------------------------------------------------

    \27\ There is a growing evidence that integrated care and 
financing models can improve beneficiary experience and quality of 
care, including:
     Health Management Associates, Value Assessment of the 
Senior Care Options (SCO) Program, July 21, 2015, available at: 
http://www.mahp.com/unify-files/HMAFinalSCOWhitePaper_2015_07_21.pdf;
     MedPAC chapter ``Care coordination programs for dual-
eligible beneficiaries,'' June 2012, available at: http://www.medpac.gov/docs/default-source/reports/chapter-3-appendixes-care-coordination-programs-for-dual-eligible-beneficiaries-june-2012-report-.pdf?sfvrsn=0.
     Anderson, Wayne L., Zhanlian Fen, and Sharon K. Long, 
RTI International and Urban Institute, Minnesota Managed Care 
Longitudinal Data Analysis, prepared for the U.S. Department of 
Health and Human Services Assistant Secretary for Planning and 
Evaluation (ASPE), March 2016, available at: https://aspe.hhs.gov/report/minnesota-managed-care-longitudinal-data-analysis.
---------------------------------------------------------------------------

    To ensure individuals are aware of the default MA enrollment and of 
the changes to their Medicare and Medicaid coverage, we also proposed, 
at Sec.  422.66(c)(2)(i)(C) and (c)(2)(iv), a requirement for MA 
organizations to issue a notice no fewer than 60 days before the 
default enrollment effective date to the enrollee. The notice \28\ must 
include clear information on the D-SNP, as well as instructions to the 
individual on how to opt out (or decline) the default enrollment and 
how to enroll in Original Medicare or a different MA plan.
---------------------------------------------------------------------------

    \28\ Enrollment requirements and burden are currently approved 
by OMB under control number 0938-0753 (CMS-R-267). Since this rule 
will not impose any new or revised requirements/burden, we are not 
making any changes to that control number.
---------------------------------------------------------------------------

    We also proposed, in paragraph (c)(2)(i)(E) and (2)(ii), that MA 
organizations must obtain approval from CMS before implementing default 
enrollment. We explained that under our proposal in paragraph 
(c)(2)(i)(B), CMS approval would be granted only if the applicable 
state approves the default enrollment through its agreement with the MA 
organization. We also noted that MA organizations would be required to 
implement default enrollment in a non-discriminatory manner, consistent 
with their obligations under Sec.  422.110; that is, MA organizations 
could not select for default enrollment only certain members of the 
affiliated Medicaid plan who were identified as eligible for default 
enrollment. Lastly, we proposed authority for CMS to suspend or rescind 
approval at any time it determined that the MA organization is not in 
compliance with the requirements. We requested comment on whether this 
authority to rescind approval should be broader. We also explained that 
we continued to consider whether a time limit on the approval (such as 
2 to 5 years) would be appropriate so that CMS would have to revisit 
the processes and procedures used by an MA organization in order to 
assure that the regulation requirements are still being followed. We 
were particularly interested in comment on this point in conjunction 
with our alternative proposal (discussed later in this section) to 
codify the existing parameters for this type of seamless conversion 
default enrollment such that all MA organizations would be able to use 
this default enrollment process for newly eligible and newly enrolled 
Medicare beneficiaries in the MA organization's non-Medicare coverage.

[[Page 16497]]

    Under our proposal, default enrollment of individuals at the time 
of their conversion to Medicare would be more limited than the default 
enrollments Congress authorized the Secretary to permit in section 
1851(c)(3)(A)(ii) of the Act. However, we also proposed some 
flexibility for MA organizations that wish to offer seamless 
continuation of coverage to their non-Medicare members (commercial, 
Medicaid or otherwise) who are gaining Medicare eligibility. We further 
proposed to amend Sec.  422.66(d)(5) and to establish, through 
subregulatory guidance, a new and simplified positive (that is, ``opt 
in'') election process that would be available to all MA organizations 
for their commercial, Medicaid or other non-Medicare plan members. To 
reflect this proposal for a simplified election process, we proposed to 
add text in Sec.  422.66(d)(5) authorizing a simplified election for 
purposes of converting existing non-Medicare coverage to MA coverage 
offered by the same organization. This new simplified enrollment 
process aimed to lessen burden for MA organizations, make enrollment 
easier for the newly-eligible beneficiary to complete, and provide 
opportunity for beneficiary choice, so that beneficiaries could remain 
with the organization that offers their non-Medicare coverage or select 
another MA plan that meets their individual needs with respect to 
provider network, prescription drug formularies, and cost and benefit 
structures. We explained that our new election process would provide a 
longer period of time for MA organizations to accept enrollment 
requests than the time period in which MA organizations would be 
required to effectuate default enrollments, as organizations would be 
able to accept simplified enrollments throughout the individual's 
Initial Coverage Election Period (ICEP), provided he or she enrolled in 
both Medicare Parts A and B when first eligible. We proposed to use 
existing authority to create this new enrollment mechanism, which would 
be available to MA organizations in the 2019 contract year. We 
solicited comments on the proposed changes to Sec.  422.66(d)(5) and 
the form and manner of the simplified enrollments.
    In addition to these proposals and solicitations for comment 
related to default and seamless enrollments for newly eligible Medicare 
beneficiaries, we proposed amendments to Sec. Sec.  422.66(d)(1) and 
422.68 that are also related to MA enrollment. Currently, as described 
in the 2005 final rule (70 FR 4606 through 4607), Sec.  422.66(d)(1) 
requires MA organizations to accept enrollment requests from an 
individual who is enrolled in a non-Medicare health plan offered by the 
MA organization during the month immediately preceding the month in 
which he or she is entitled to both Part A and Part B and who meets MA 
eligibility requirements. We are concerned that in some instances, this 
regulation has been interpreted as meaning that the enrollment request 
must be filed during the month before Medicare entitlement occurs. To 
clarify the requirement and be more consistent with section 
1851(c)(3)(A)(ii), we proposed to amend Sec.  422.66(d)(1) to add text 
clarifying that seamless continuation of coverage is available to an 
individual who requests enrollment during his or her Initial Coverage 
Election Period. We also proposed a revision to Sec.  422.68(a) to 
ensure that ICEP elections made during or after the month of 
entitlement to both Part A and Part B are effective the first day of 
the calendar month following the month in which the election is made. 
This proposed revision would codify subregulatory guidance that MA 
organizations have been following since 2006. This proposal is also 
consistent with the proposal at Sec.  422.66(c)(2)(iii) regarding the 
effective date of coverage for default enrollments into D-SNPs. We also 
solicited comment on these related proposals.
    In conclusion, we proposed to add regulation text at Sec.  
422.66(c)(2)(i) through (iv) to set limits and requirements for a 
default enrollment of the type authorized under section 
1851(c)(3)(A)(ii). We proposed a clarifying amendment to Sec.  
422.66(d)(1) regarding when seamless continuation coverage can be 
elected and revisions to Sec.  422.66(d)(5) to reflect our proposal for 
a new and simplified positive election process that will be available 
to all MA organizations and their members who enroll in an MA plan 
offered by the same entity that offers the individual's pre-Medicare 
coverage. Lastly, we proposed revisions to Sec.  422.68(a) to ensure 
that ICEP elections made during or after the month of entitlement to 
both Part A and Part B are effective the first day of the calendar 
month following the month in which the election is made. We solicited 
comments on all these proposals.
    In addition, we presented an alternative for consideration and 
comment. Because we recognized that our proposal narrowed the scope of 
default enrollments compared to what CMS approved under section 
1851(c)(3)(A) of the Act in the past, we discussed in the proposed rule 
that we continued to consider retaining processes similar to the pre-
moratorium seamless conversion process. That seamless conversion 
mechanism is outlined currently in section 40.1.4 of Chapter 2 of the 
Medicare Managed Care Manual and had been in practice through October 
2016. As an alternative we considered proposing regulations to codify 
that guidance as follows--
     Articulating the requirements for an MA organization's 
proposal to use the seamless conversion mechanism, including 
identifying eligible individuals in advance of Medicare eligibility;
     Establishing timeframes for processing and the effective 
date of the enrollment; and
     Requiring notification to individuals at least 60 days 
prior to the conversion of their right to opt-out or decline the 
enrollment.
    In considering this alternative, we contemplated additional 
beneficiary protections, including the issuance of an additional notice 
to ensure that individuals understood the implication of taking no 
action when notified of the default enrollment. While this alternative 
would lead to increased use of the seamless conversion enrollment 
mechanism than what had been used in the past, we expressed concern 
that the operational challenges, particularly in relation to the new 
Medicare Beneficiary Identification number, could be significant for MA 
organizations to overcome at this time.
    We also explained how we considered proposing regulations to limit 
the use of default enrollment to only beneficiaries who are eligible 
for Medicare based on age. While this alternative would simplify an MA 
organization's ability to identify eligible individuals, we noted 
concerns about disparate treatment among newly eligible beneficiaries 
based on their reason for obtaining Medicare entitlement.
    We invited comments on our proposal and the alternate approaches we 
identified, including the following:
     Codify the existing parameters for this type of seamless 
conversion default enrollment such that all MA organizations would be 
able to use this default enrollment process for newly eligible and 
newly enrolled Medicare beneficiaries already covered by the MA 
organization's non-Medicare coverage.
     Codify the existing parameters for this type of seamless 
conversion default enrollment, as described previously, but allow that 
use of default enrollment to be limited to only the aged population.
    We also asked for solutions to address the concerns we identified 
in the proposed rule, particularly related to

[[Page 16498]]

how MA organizations could identify commercial members who are 
approaching Medicare eligibility based on disability, as well as how 
plans could confirm MA eligibility and process enrollments without 
access to the individual's Medicare number.
    We received the following comments and our responses follow:
    Comment: We received significant support for our proposal to permit 
default MA enrollments, especially for dually-eligible beneficiaries 
who are newly eligible for Medicare. Most commenters supported the 
proposal to permit only D-SNPs to receive defaulted enrollments for 
dually-eligible beneficiaries. Some commenters who supported our 
proposal also supported the alternative we noted for consideration that 
would permit default enrollment of newly Medicare-eligible individuals 
enrolled in a non-Medicare health plan offered by the same 
organization.
    Response: We appreciate the widespread support we received for the 
proposal. In our view, this proposal and our final rule support state 
efforts to increase enrollment of dually eligible individuals in fully 
integrated systems of care.
    We appreciate the responses to our solicitation of feedback on 
expanding default enrollment to include individuals enrolled in 
commercial health plans offered by an MA organization. As noted in the 
proposed rule (82 FR 56366) and above, our experience with the current 
seamless conversion enrollment mechanism makes it clear that 
organizations attempting to seamlessly convert individuals from 
commercial coverage (that is, private coverage and Marketplace 
coverage) are, for the most part, unable to comply with our current 
guidance and approval parameters, especially the expectation that 
organizations have the means to identify their commercial members who 
are approaching Medicare eligibility based on disability. Given these 
challenges, we did not specifically propose to codify default 
enrollment from commercial coverage. We also solicited feedback on how 
MA organizations might overcome the challenges in confirming 
entitlement to Medicare Parts A and B within necessary timeframes and 
obtaining the individual's Medicare number, given that in 2018 this 
will become a random and unique number instead of a Social Security 
Number-based identifier. We received only a few responses to our 
solicitation of ideas on how to resolve these issues; commenters 
generally deferred to CMS to find a way to identify non-MA members when 
those members approach Medicare eligibility and for CMS to convey this 
information to plans well in advance of the Medicare eligibility date. 
In light of these comments, CMS may consider expanding default 
enrollment to occur from commercial or other coverage arrangements in 
future rulemaking. We are not finalizing the alternate proposal on 
which we solicited comment.
    Comment: A commenter asked that we expand default enrollment to 
those enrolled in other ``state innovated models'' and delivery systems 
other than Medicaid managed care, such as ACOs. The same commenter 
asked that we allow the default enrollment provisions to be applied to 
individuals enrolled in coverage other than comprehensive Medicaid 
managed care, including prepaid inpatient health plans, prepaid 
ambulatory health plans, and primary care case management. Another 
commenter asked that we consider expanding our proposal for default 
enrollment and/or changing the current parameters for passive 
enrollment to allow a State to enroll any dually-eligible individual 
(whether in a Medicaid managed care plan or in a Medicaid Fee-for-
Service program) into a D-SNP at any time.
    Response: We appreciate the comments. As proposed, default 
enrollment would be subject to several substantive conditions, one of 
which required that anyone being considered for default enrollment be 
enrolled in a Medicaid managed care plan affiliated with the MA 
organization. Our proposal was specific to allowing default enrollment 
of individuals enrolled in comprehensive Medicaid managed care plans--
rather than limited-benefit plans or case management arrangements--into 
D-SNPs when these Medicaid managed care plan enrollees first become 
eligible for Medicare. We believe that our overall goals of encouraging 
integrated care are best met by limiting the default enrollment to the 
context of comprehensive Medicaid managed care plans at this point and 
may revisit an expansion of this regulation in future rulemaking. We 
plan to further clarify allowable scenarios in subsequent guidance. 
However, given the parameters of section 1851(c)(3)(A)(ii) of the Act, 
we are unable to finalize a regulation that so substantially expands 
the population of beneficiaries subject to this default enrollment to 
include Medicaid beneficiaries who are not enrolled in a health plan 
offered by an MA organization.
    Comment: Several commenters who support our proposal for default 
enrollment recommend that, if finalized, we ensure that beneficiaries 
who do not speak English as a primary language receive outreach in 
their language, preferably by both mail and telephone.
    Response: We appreciate these comments and agree that clear 
communication with individuals identified for default enrollment is an 
important protection, especially with regard to the potential impact of 
MA plan enrollment on an individual's access to care. We note that 
existing law, such as Title VI of the Civil Rights Act of 1964 
(applicable to MA organizations in connection with Medicare coverage) 
and 42 CFR 438.10 (applicable to Medicaid managed care plans) address 
requirements for providing access to enrollees who have limited English 
proficiency (LEP). Guidance on the Civil Rights Act of 1964 and 
authorities that are not limited to Medicare or Medicare is issued by 
the HHS Office for Civil Rights (OCR). We refer the commenter to 
section II.B.5 of this final rule on marketing and communications 
requirements. We believe, therefore, that revisions to our proposed 
rule are not necessary.
    Comment: Several commenters stated that the network for the MA plan 
should be substantially identical and should not be substantially 
narrower than the network of the Medicaid plan from which default 
enrollment would occur.
    Response: Although we did not include specific provider network 
criteria in our proposal for default MA enrollment, we note that CMS 
currently has in place network adequacy requirements that would apply 
to any MA plan into which default enrollment occurs. States also have 
the opportunity to use their State Medicaid agency contracts with D-
SNPs to create additional provider network continuity requirements. 
Therefore, we do not believe that additional criteria are warranted.
    Comment: Several of the commenters who opposed our proposal for 
default enrollment asked that in the event that our proposal for 
default enrollment is finalized, we consider additional beneficiary 
protections, such as a minimum star rating for the MA plan into which 
default enrollment would occur and the exclusion of MA plans that have 
been assessed a civil monetary penalty or have been sanctioned within 
the previous 18 months. Another commenter expressed concern about the 
potential for individuals to be default enrolled into an MA plan with a 
low star rating when there are MA plans with higher star ratings 
offered by other organizations in the same area. These commenters note 
that organizations with high star ratings that do not offer

[[Page 16499]]

a Medicaid plan would not be permitted to conduct default enrollment.
    Response: We appreciate the comments we received regarding the 
significance of the compliance history of an MA organization that 
wishes to conduct default MA enrollment and the suggestion of a minimum 
star rating. We agree with these commenters that standards governing 
the quality of the MA D-SNP are appropriate to adopt as well. We 
believe that default enrollment should not be permitted into an MA plan 
offered by an MA organization with a low star rating and/or recent 
issues of significant noncompliance with our regulatory requirements 
such that CMS has imposed a suspension on new enrollments. Since 
default MA enrollment is based on an opt-out, rather than opt-in, 
approach, we believe it is important to ensure that individuals are not 
enrolled by default into MA plans offered by poor performing 
organizations. Therefore, we are finalizing the regulation with 
additional paragraphs ((c)(2)(i)(F) and (G)) that limit default 
enrollment authority to MA plans that have an overall rating of 3 Stars 
(or are low enrollment or new contracts) and that are not under a 
prohibition on new enrollments.
    Comment: Most commenters expressed support for limiting CMS 
approval of an organization's request to conduct default enrollment to 
a specific time frame. Those who mentioned a specific time frame 
suggested a period of 2 to 5 years. A commenter suggested that CMS 
conduct a review after initial approval only if there is an indication 
of disruption in care.
    Response: CMS oversight of plans' implementation of the default 
enrollment process is an important beneficiary protection. We agree 
with the suggestions of a 5 year timeframe, as it provides a reasonable 
amount of time for MA organizations to implement and then assess the 
approved process, limits administrative burden for MA organizations to 
request continued approval, and provides them the opportunity to update 
their processes as operational enhancement or new technologies emerge. 
However, in our view, should beneficiary complaints or allegations of 
noncompliance come to our attention, we need to be able to conduct a 
review of an organization's default enrollment process prior to the 
expiration of the five year period. Therefore, we will include in the 
final rule an approval time period of 5 years with a provision that 
permits CMS to suspend or rescind approval if CMS determines that the 
MA organization is not in compliance with the requirements or Sec.  
422.66(c)(2) or other MA program standards.
    Comment: A commenter suggested that we share with states the 
criteria we will use to review plan proposals to offer default 
enrollment, adding that this may promote uniformity with implementation 
across the various states.
    Response: The requirements for default enrollment are outlined in 
this regulation. In addition, we will consider additional guidance, 
which is available to states, industry, advocates, and the general 
public, as necessary.
    Comment: Most commenters expressed support for our proposal to 
permit simplified elections for seamless continuation of commercial 
coverage into a MA plan offered by the same organization. A commenter 
expressed opposition to the offering of a simplified (opt-in) 
enrollment mechanism to anyone enrolled in a Medicaid managed care 
plan. Another commenter asked that we consider making the simplified 
(opt-in) enrollment mechanism available to all beneficiaries, including 
those who are not in their ICEP and those who are not enrolled in a 
non-Medicare plan offered by the same organization.
    Response: We appreciate the support for our proposal to promote 
beneficiary choice and simplify the enrollment process for all MA 
organizations that offer non-Medicare coverage. However, we disagree 
with the suggestion to prohibit use of the simplified enrollment 
mechanism by those enrolled in Medicaid managed care plans. In our 
view, an eligible individual always has the option to make an active 
choice into an MA plan that meets their needs when in an election 
period. Further, as not all individuals in Medicaid managed care plans 
will be automatically enrolled into a D-SNP (such as those individuals 
enrolled in Medicaid managed care plans whose parent organizations have 
opted not to use the default enrollment mechanism or those individuals 
whose Medicaid managed care enrollment is in a Medicaid prepaid health 
plan that covers a limited scope of benefits), the simplified 
enrollment mechanism will lessen burdens on the enrollee and MA 
organizations that offer such plans. We believe that a simplified 
election process for beneficiaries who wish to convert from their non-
Medicare coverage to MA coverage offered by the same entity will 
facilitate a more efficient enrollment process overall.
    As described in the proposed rule, this mechanism will be available 
to any MA organization that chooses to offer it. It will be potentially 
available to any beneficiary who wishes to join an MA plan offered by 
the same MA organization that offers his or her non-Medicare coverage 
at the time of his or her initial Medicare eligibility. The simplified 
enrollment mechanism aims to lessen the amount of information that an 
MA organization needs to collect from the beneficiary and to use 
information the MA organization already has. MA organizations that do 
not already have an existing relationship with an individual must 
collect all the necessary information in which to determine eligibility 
and process the enrollment request under Sec.  422.60.
    We appreciate the feedback to finalize use of a simplified 
enrollment mechanism authorized under Sec.  422.66(d)(5) as amended in 
this final rule. We will permit individuals who are in their ICEP and 
enrolled in any type of non-Medicare plan to use the simplified (opt-
in) enrollment mechanism to request enrollment in any type of MA plan 
offered by the same MA organization that offers the non-Medicare 
coverage.
    Comment: A few commenters responded to our solicitation of feedback 
on limiting default enrollment to only the aged. Most of these 
commenters opposed this limitation; a commenter supported it. Those who 
oppose limiting default enrollment to only the aged believe that 
allowing default enrollment to be offered only to those whose Medicare 
eligibility is based on age, instead of to all beneficiaries, would be 
discriminatory on its face because the exclusion is based on having a 
disability or ESRD. Another commenter believes that states and plans 
should be allowed to determine whether including all individuals 
approaching Medicare eligibility is feasible and, if not feasible, 
include only those whose Medicare eligibility is based on age.
    Response: We thank the commenters and agree that it would be 
inappropriate to exclude individuals whose Medicare eligibility is 
based on disability from default enrollment. We believe that an 
individual's eligibility to be included in default enrollment should be 
based on his or her projected Medicare eligibility in general and not 
on the specific reason for Medicare eligibility. We are, therefore, 
finalizing this aspect of our proposal as described in the notice of 
proposed rulemaking and are not including any authority to limit 
default enrollment (under paragraph (c)) or seamless conversions (under 
paragraph (d)) to beneficiaries whose eligibility is based on age.

[[Page 16500]]

    Comment: In the event that our proposal for default enrollment is 
finalized, several commenters who opposed our proposal for default 
enrollment ask that default-enrolled beneficiaries be provided 
transition coverage, allowing use of an off-formulary drug, and 
allowing a beneficiary to maintain an out-of-network provider for 12 
months, similar to the Medicare-Medicaid financial alignment 
demonstration.
    Response: We appreciate these comments and note that several of the 
concerns expressed are addressed in other areas of current regulation 
and guidance. With regard to formulary concerns, we note that all plans 
offering Part D coverage must meet CMS' formulary adequacy requirements 
and, in addition, must offer a transition period upon a member's 
enrollment in a new plan. Specifically, under Sec.  423.120(b)(3), new 
enrollees must be provided a temporary supply of non-formulary Part D 
drugs, as well as Part D drugs with utilization management 
restrictions, and can work with their new plan and provider to switch 
to a different formulary drug or request an exception during their 
first 90 days of enrollment in the new MA plan. States may also use 
their State Medicaid Agency contracts with D-SNPs to create additional 
continuity requirements. With regard to the commenters' suggestion that 
we require MA organizations to allow new members to receive care from 
out-of-network providers for 12 months, similar to the Medicare-
Medicaid financial alignment demonstration, we note that a 6 month 
continuity of care period is more common for demonstration plans. In 
addition, we note that this period can be offered by demonstration 
plans due to the demonstration authority itself; we do not have similar 
authority to impose a similar requirement on MA organizations that 
choose to implement the default enrollment process.
    Comment: The few commenters who opposed default enrollment cite as 
the basis for their position the lack of beneficiary choice and the 
potential for disruption in care resulting from default enrollment into 
a plan with different benefits, cost-sharing, provider network and 
formulary.
    Response: In response to these comments, we note that an important 
feature of this enrollment process is clear and timely advance notice 
to the individual regarding default MA enrollment and the opportunity 
to decline the enrollment up to and including the day prior to the 
enrollment effective date. We, therefore, disagree with these 
commenters that the default MA enrollment process, as proposed and as 
finalized in this rule, does not involve beneficiary choice. The notice 
requirements in the final rule will provide the beneficiary a least a 2 
month period in which to review his or her Medicare options and make an 
informed choice. Further, the new MA Open Enrollment Period, discussed 
at section II.B.1 of this final rule, would be available to any 
beneficiary who was default enrolled in an MA plan pursuant to Sec.  
422.66(c)(2). Upon an individual's new enrollment in an MA plan during 
the individual's ICEP, he or she would have 3 months, under the MA Open 
Enrollment Period discussed in Sec.  422.62(a)(5), to make a change to 
another MA plan or select Original Medicare for health coverage. 
Additionally, as individuals eligible for default enrollment would only 
be those dually-eligible, they would also be eligible to use their 
quarterly opportunity under the duals SEP, as outlined in II.A.10 of 
this final rule, to make a Part D election, as well as any other 
election periods for which they may qualify, to make a change. In this 
context, a Part D election would include enrollment into an MA plan 
that includes a Part D benefit. We believe that there are adequate 
protections in place, as finalized with these amendments to Sec.  
422.66(c)(2) and elsewhere in this final rule, for beneficiary choice 
in connection with the initial election period when someone is first 
entitled to or eligible for Medicare.
    The regulation we proposed requires the MA organization conducting 
default enrollment to provide notice that describes the costs and 
benefits of the MA plan into which the default enrollment would occur, 
as well as the process for accessing care under the plan. We agree with 
the commenters that information on the differences between an 
individual's current non-Medicare coverage and the new MA plan, 
including a statement as to whether the individual's current primary 
care provider will continue to be available to the individual upon 
enrollment in the MA plan, should be included in the advance 
notification of default enrollment. We also agree that information on 
other types of Medicare plans should be included in the notice to 
ensure an individual who is notified of default enrollment has 
sufficient information and can make an informed choice with regard to 
the coverage option that best meets his or her needs. Therefore, we are 
finalizing additional paragraphs, at (c)(2)(iv), that specific 
information be included in the notice describing the default enrollment 
and the ability to opt-out:
    (A) Information on the differences in premium, benefits and cost 
sharing between the individual's current Medicaid managed care plan and 
the dual eligible MA special needs plan and the process for accessing 
care under the MA plan;
    (B) The individual's ability to decline the enrollment, up to and 
including the day prior to the enrollment effective date, and either 
enroll in Original Medicare or choose another MA plan; and
    (C) A general description of alternative Medicare health and drug 
coverage options available to an individual in his or her Initial 
Coverage Election Period.
    In addition, we are including in the regulation that this 
information and the notice about the default enrollment is in addition 
to any mandatory disclosures required under Sec.  422.111.
    Comment: Several commenters who opposed our proposal for default 
enrollment expressed support for our proposal to develop a simplified 
(opt-in) enrollment mechanism, as long as differences between an 
individual's current and new plan are clearly communicated and that he 
or she is made aware of all options available to newly Medicare-
eligible individuals. These commenters note that an individual's 
initial eligibility for Medicare is a critical decision point and that 
information on the full range of Medicare coverage options is important 
to help ensure that those approaching Medicare eligibility are aware of 
the resources available to them and of any time-limited enrollment 
opportunities, such as the option to obtain Medigap on a guaranteed 
issue basis.
    Response: With respect to the new simplified (opt-in) election 
mechanism that would be available to all MA organizations for MA 
enrollments of their commercial, Medicaid or other non-Medicare 
members, we note that MA organizations that choose to implement this 
optional election mechanism will be required to follow existing rules 
governing mandatory disclosures (for example, Sec.  422.111), 
communications and marketing that are applicable to other beneficiary-
initiated enrollment requests. Required disclosures include a 
description of the MA plan benefits, including applicable conditions 
and limitations, premiums and cost-sharing (such as copayments, 
deductibles, and coinsurance), any other conditions associated with 
accessing benefits and for purposes of comparison, a description of the 
benefits offered under original

[[Page 16501]]

Medicare. Also included under Sec.  422.111 is the requirement to 
disclose the number, mix, and distribution (addresses) of providers 
from whom enrollees may reasonably be expected to obtain services. We 
will provide additional information on this optional enrollment 
mechanism in subregulatory guidance.
    Given these substantial existing disclosure requirements that will 
be applicable to the new simplified (opt-in) election mechanism, as 
well as our ongoing public outreach and education activities for 
individuals new to Medicare, we do not believe that additional notice 
or disclosure requirements are warranted.
    Comment: A few commenters asked that we reduce the requirement to 
identify newly-eligible Medicare beneficiaries from 90 to 60 days.
    Response: We believe the commenters' reference to a 90 day 
requirement for advance notification of newly-eligible Medicare 
beneficiaries is based on the current subregulatory guidance applicable 
to the seamless conversion enrollment mechanism. This guidance will be 
revised as a result of this final rule to account for default 
enrollment and the new simplified (opt-in) enrollment mechanism. The 
rule we are finalizing requires notice to the affected beneficiary at 
least 60 days in advance of the enrollment effective date (the month in 
which the individual is first entitled to both Part A and Part B). This 
reflects a change from the current seamless conversion process, which 
requires identification of beneficiaries that will be seamlessly 
enrolled 90 days in advance. While we believe that timely 
identification of individuals approaching Medicare eligibility is an 
important beneficiary protection that helps to ensure that plans are 
able to provide timely advance notification and submission of 
enrollment transactions to CMS, we also believe that for default 
enrollment this shorter timeframe does not have an adverse beneficiary 
impact. MA plans that are authorized to use this default enrollment 
process must identify all eligible enrollees in time to provide the 
required advance notification to individuals eligible for default 
enrollment no fewer than 60 days before the default enrollment 
effective date.
    Comment: Several commenters suggested that CMS consider allowing 
default enrollment from Medicaid managed care plans into fully 
integrated dual eligible special needs plans (FIDE SNPs), which are a 
type of special needs plan designed to promote the full integration and 
coordination of Medicaid and Medicare benefits for dual eligible 
beneficiaries by a single managed care organization.
    Response: We thank the commenters for their feedback and agree that 
allowing default enrollment from Medicaid managed care plans into FIDE 
SNPs is consistent with the proposed rule. FIDE SNPs are a specific 
type of approved MA-PD dual eligible special needs plan. We will 
finalize revised text to clarify that FIDE SNPs are permitted to use 
the default enrollment mechanism, subject to the other requirements in 
the rule.
    Comment: A commenter stated that Congress should revisit default 
enrollment in traditional Medicare. This commenter believes that to the 
extent that MA quality is superior, enrollment should default to the 
highest quality option, rather than to traditional Medicare.
    Response: As acknowledged by the commenter, this comment is outside 
of the scope of this regulation and our authority under section 1851. 
CMS's authority is circumscribed by the Medicare statute, particularly 
section 1851(c)(3)(A)(ii) of the Act with regard to default 
enrollments.
    Comment: A commenter suggests that plans conducting default 
enrollment be allowed to send the notification of default enrollment up 
to 90 days after an individual's initial Medicare eligibility, adding 
that this would increase enrollment into integrated plans.
    Response: We appreciate the suggestion; however we disagree with 
permitting notification of default enrollment after enrollment or, as 
implied by the commenter, effectuating the default enrollment up to 90 
days after the initial date of Medicare eligibility. As described in 
our proposal, states have the information to identify newly eligible 
Medicare beneficiaries before the actual first date of Medicare 
eligibility; therefore, they have the information necessary to provide 
to their contracted MA organizations so that the integrated coverage 
can begin at the earliest possible date--the date the individual first 
has both Medicare Parts A and B. As such, the effective date for 
default enrollment will always coincide with the date of an 
individual's entitlement to and eligibility for Medicare Parts A and B, 
which would not allow the commenter's suggested change. We note as well 
that the commenter's suggestion would result in notification of the 
default enrollment well after the enrollment effective date, resulting 
in a period of time during which the individual is not aware of his or 
her enrollment in an MA plan, does not have the information necessary 
to access benefits and would be financially liable for healthcare 
services received from providers not contracted with the MA plan. To 
ensure that individuals receive timely advance notification of the 
default enrollment, we are declining the commenter's suggestion. We 
note that individuals who are enrolled into a MA plan through default 
enrollment continue to have a three-month opportunity to change their 
enrollment using the MA Open Enrollment Period, as outlined in Sec.  
422.62(a)(5). Further, an individual who chooses to opt out of default 
enrollment into an MA plan is still able to make an election during his 
or her Initial Coverage Election Period, which begins 3 months before 
and lasts 3 months after the month of initial Medicare eligibility.
    Comment: A commenter suggested that default enrollment not be 
allowed where Medicare-Medicaid financial alignment demonstration plans 
are available.
    Response: We are committed to partnership with state Medicaid 
agencies to pursue integrated care approaches that work for each state. 
We believe that the proposed regulatory language requiring state 
approval for default enrollment into D-SNPs provides an appropriate 
safeguard that ensures any default enrollments are consistent with the 
state's Medicare-Medicaid integration goals.
    Comment: A commenter who opposes default enrollment into D-SNPs 
stated that it will lead to reduced competition and fewer D-SNP 
offerings for beneficiaries, resulting in higher costs and fewer 
benefits over time.
    Response: We appreciate the comment but disagree with the 
commenter's assessment and conclusion regarding the impact of default 
MA enrollment on competition in the market and the number of D-SNP 
offerings. As default enrollment accounts only for those newly eligible 
for Medicare, it is our view that D-SNPs provide a valuable service to 
all beneficiaries--those currently and newly in the Medicare program.
    After review of the comments, and as discussed earlier, we are 
finalizing the proposed changes to Sec. Sec.  422.66(c) and 
422.68(d)(1) and (5) with the following modifications:
     Paragraph 422.66(c)(2)(i) will be revised to clarify that 
we will allow default enrollment into a FIDE-SNP administered by an MA 
organization under the same parent organization as the organization 
that operates the Medicaid managed care plan in which the individual 
remains enrolled.
     Paragraph 422.66(c)(2)(i) will be revised to require a 
minimum star rating

[[Page 16502]]

on the contract receiving default enrollments for an MA organization to 
be approved for default enrollment. We are revising the paragraph to 
require that, for an organization to be approved for default 
enrollment, it must have an overall quality rating, from the most 
recently issued ratings, under the rating system described in 
Sec. Sec.  422.160 through 422.166, of at least 3 stars or is a low 
enrollment contract or new MA plan as defined in Sec.  422.252. In 
addition, the MA organization must not be under an enrollment 
suspension.
     Paragraph 422.66(c)(2)(ii) will be revised to include an 
approval period not to exceed 5 years, subject to CMS authority to 
rescind or suspend approval if the plan is non-compliant.
     Paragraph 422.66(c)(2)(iv) will be revised to require that 
the notice issued by the MA organization include information on the 
differences in premium, benefits and cost sharing between the 
individual's current Medicaid managed care plan and the dual eligible 
MA special needs plan and the process for accessing care under the MA 
plan; an explanation of the individual's ability to decline the 
enrollment, up to and including the day prior to the enrollment 
effective date, and either enroll in Original Medicare or choose 
another MA plan; and a general description of alternative Medicare 
health and drug coverage options available to an individual in his or 
her Initial Coverage Election Period.
     Paragraph 422.66(c)(2)(iv) will be revised to clarify that 
the mandatory notice is in addition to the information and documents 
required to be provided to new enrollees under Sec.  422.111.
8. Passive Enrollment Flexibilities To Protect Continuity of Integrated 
Care for Dually Eligible Beneficiaries (Sec.  422.60(g))
    Beneficiaries who are dually eligible for both Medicare and 
Medicaid typically face significant challenges in navigating the two 
programs, which include separate or overlapping benefits and 
administrative processes. Fragmentation between the two programs can 
result in a lack of coordination for care delivery, potentially 
resulting in unnecessary, duplicative, or missed services. One method 
for overcoming this challenge is through integrated care, which 
provides dually eligible beneficiaries with the full array of Medicaid 
and Medicare benefits for which they are eligible through a single 
delivery system, thereby improving quality of care, beneficiary 
satisfaction, and care coordination, and reducing administrative 
burden.
    In the proposed rule, we proposed a limited expansion of CMS' 
regulatory authority to initiate passive enrollment for certain dually 
eligible beneficiaries who are currently enrolled in an integrated D-
SNP into another integrated D-SNP in instances where integrated care 
coverage would otherwise be disrupted, such as during a state re-
procurement of Medicaid managed care contracts that results in current 
Medicaid managed care plans not being renewed, or when beneficiaries 
are enrolled in an integrated D-SNP that non-renews its MA contract at 
the end of the contract year. The intent of CMS' proposal was to 
improve care coordination and minimize disruption in care by promoting 
enrollment in integrated care arrangements for dually eligible 
beneficiaries currently enrolled in an integrated D-SNP.
    Specifically, we proposed authorizing CMS to passively enroll 
certain dually eligible individuals currently enrolled in an integrated 
D-SNP into another integrated D-SNP, after consulting with the state 
Medicaid agency that contracts with the D-SNP or other integrated 
managed care plan, when CMS determines that the passive enrollment will 
promote continuity of care and integrated care under Sec.  
422.60(g)(1)(iii). We also proposed, under Sec.  422.60(g)(2), a number 
of requirements an MA plan would have to meet in order to qualify to 
receive passive enrollments under paragraph (g)(1)(iii). These proposed 
requirements are detailed below.
     MA plans receiving the passive enrollments must be highly 
integrated D-SNPs, thereby restricting passive enrollment to those MA 
plans that operate as a FIDE SNP or meet the integration standard for a 
highly-integrated D-SNP, as defined in Sec.  422.2 and described in 
Sec.  422.102(e), respectively.
     In an effort to promote continuity of care, receiving MA 
plans must have substantially similar provider and facility networks 
and Medicare- and Medicaid-covered benefits as the integrated MA plan 
(or plans) from which beneficiaries are passively enrolled.
     D-SNP contracts must have a minimum overall MA Star Rating 
of at least 3 stars for the year prior to receipt of passive enrollment 
or be a low enrollment or new MA contract (which do not have a Star 
Rating because of the insufficient data available).
     Receiving MA plans must not have any prohibition on new 
enrollment imposed by CMS.
     Receiving MA plans must have appropriate limits on premium 
and cost-sharing for beneficiaries.
    We solicited comments on our proposal to identify plans for 
receiving passive enrollments, particularly on the minimum quality 
standards relevant to dually eligible beneficiaries. We also solicited 
comments on whether to limit passive enrollment authority to 
circumstances that would not raise total cost to the Medicare and 
Medicaid programs. Additionally, we requested feedback on how to 
calculate the projected impact on Medicare and Medicaid costs from 
exercise of this authority.
    In the proposed rule, we noted that we had also considered 
proposing new (or additional) beneficiary notification requirements for 
passive enrollments that occur under proposed paragraph (g)(1)(iii), 
including the provision of two notifications to enrollees prior to the 
effective date. Citing the existing beneficiary notifications that are 
currently required under Medicare regulations and concerns regarding 
the quantity of notifications sent to beneficiaries, we did not propose 
to modify the existing notification requirements under paragraph (g)(4) 
of the proposed rule. However, we solicited comment on alternatives 
regarding beneficiary notices, including comments about the content and 
timing of such notices.
    We received the following comments and our responses follow.
    Comment: Many commenters expressed support for CMS' proposal for a 
limited expansion of the current passive enrollment authority in order 
to promote continued enrollment of dually eligible beneficiaries in 
integrated D-SNPs, preserve and promote care integration, and limit 
disruptions in care under certain circumstances. Several commenters 
supported CMS' goal of care continuity while expressing their belief 
that the best way to empower beneficiaries is through mechanisms where 
beneficiaries opt in to integrated care. A commenter requested that CMS 
consider how passive enrollment of beneficiaries from an existing 
integrated D-SNP into another integrated D-SNP could create disruptions 
in care. A few commenters opposed our passive enrollment proposal due 
to concerns that passive enrollment limits beneficiary choice and 
erodes the role of competition in the marketplace. A commenter 
suggested that a better alternative for beneficiaries in integrated D-
SNPs that are non-renewing is for them to revert to FFS Medicare. 
Another commenter noted that passive enrollment in other

[[Page 16503]]

circumstances has proven to be too confusing for dually eligible 
beneficiaries.
    Response: We appreciate the support by most commenters of our goals 
of promoting continuity and quality of care for dually eligible 
beneficiaries currently enrolled in integrated D-SNPs in situations 
where they would otherwise experience an involuntary disruption in 
either Medicare or Medicaid coverage. As we stated in the proposed rule 
(82 FR 56369-56370), we anticipate using this new authority exclusively 
in limited situations related to market disruptions related to D-SNP 
non-renewal or changes in state Medicaid managed care organization 
procurements; therefore, we anticipate that this authority, as 
finalized, will have no significant impact on competition in the 
Medicare Advantage marketplace. We also proposed that D-SNPs meet 
certain requirements related to integration, quality, performance, and 
provider network and benefits comparability relative to the enrollees' 
previous coverage. We believe these safeguards will ensure continuity 
of care and limit any disruption associated with a plan change for 
affected enrollees. In addition, we believe the beneficiary notice 
requirements for passively enrolled individuals described in Sec.  
422.60(g)(4) ensure that beneficiaries will receive appropriate advance 
notice regarding the costs and benefits of their new coverage, the 
process for accessing care under the new plan, and an explanation of 
the beneficiary's ability to decline the enrollment or choose another 
plan. As described elsewhere in this final rule, we are strengthening 
the notice requirements associated with passive enrollment under this 
new limited expansion of CMS' passive enrollment authority. Finally, we 
note that all individuals enrolled into an integrated D-SNP under CMS' 
passive enrollment authority will have a special election period (SEP) 
under Sec.  422.60(g)(5), which as finalized in this rule refers to the 
new SEP established in this final rule at Sec.  423.38(c)(10). This SEP 
will allow individuals to opt out of the passive enrollment within 3 
months of notification of a CMS or state-initiated enrollment action or 
that enrollment action's effective date (whichever is later). This SEP 
is in addition to any other election periods for which they qualify. 
During the SEP, a beneficiary would be able choose FFS Medicare or 
other coverage based on their personal preferences. Therefore, we are 
finalizing the proposed limited expansion of CMS' passive enrollment 
authority at Sec.  422.60(g)(1)(iii). However, we note that we are 
making a technical revision to paragraph (g)(1)(iii) to clarify that a 
plan must meet all the requirements under paragraph (g)(2) to be 
eligible to receive passive enrollment.
    Comment: A commenter stated that any beneficiary who has chosen FFS 
Medicare should not be passively enrolled. Several commenters suggested 
that passive enrollment be extended to existing and new dually eligible 
beneficiaries in FFS Medicare and stand-alone Part D plans. A few 
commenters recommended passively enrolling dually eligible 
beneficiaries into a D-SNP when states enroll beneficiaries into a 
mandatory Medicaid long-term services and supports (LTSS) program.
    Response: While we appreciate commenters' support for coordinated 
care options for individuals who are not currently enrolled in an MA 
plan, we note that our intent in proposing an expansion of CMS' passive 
enrollment authority was to promote continuity of integrated care for 
those beneficiaries enrolled in an integrated D-SNP but who would 
experience an involuntary disruption in their Medicare or Medicaid 
coverage in the absence of passive enrollment into a comparable 
integrated D-SNP. This authority could not be used to transition 
enrollees currently in FFS Medicare to an MA plan.
    Comment: Some commenters agreed that passive enrollment eligibility 
should be limited to highly integrated D-SNPs. A commenter recommended 
limiting eligibility for passive enrollment to integrated D-SNPs with 
the experience and size to meet the unique needs of the dual eligible 
population. A few commenters expressed concern that the scope of our 
proposal was too limited because only Fully Integrated Dual Eligible 
(FIDE) SNPs and other MA plans that meet the integration standard for a 
highly-integrated D-SNP, as defined in Sec.  422.2 and described in 
Sec.  422.102(e), respectively, would be qualified to receive the 
passive enrollments. These commenters noted the limited number of 
highly integrated D-SNPs and FIDE SNPs currently in the market. A few 
commenters recommended extending eligibility to include all D-SNPs that 
meet minimum quality standards and can demonstrate appropriate levels 
of integrated benefits. Another commenter recommended that CMS allow 
states the flexibility to determine which D-SNPs are eligible to 
participate in passive enrollment.
    Response: We appreciate the commenters' perspectives on this issue. 
We may re-examine this issue as we gain experience, but we have 
concluded that it is more prudent to focus this form of passive 
enrollment on a narrow set of circumstances that offer the highest 
levels of integration between Medicare and Medicaid. This will allow us 
to better monitor implementation and will promote integration, which 
has been associated with better outcomes.\29\ We also note that our 
proposed criteria are minimum standards only; states can establish 
additional criteria to determine which D-SNPs may be eligible for 
passive enrollment. As such, we are finalizing the scope of the 
proposed passive enrollment authority for dually eligible beneficiaries 
enrolled in an integrated D-SNP, without modification.
---------------------------------------------------------------------------

    \29\ There is a growing evidence that integrated care and 
financing models can improve beneficiary experience and quality of 
care, including:
     Health Management Associates, Value Assessment of the 
Senior Care Options (SCO) Program, July 21, 2015, available at: 
http://www.mahp.com/unify-files/HMAFinalSCOWhitePaper_2015_07_21.pdf.
     MedPAC chapter ``Care coordination programs for dual-
eligible beneficiaries,'' June 2012, available at: http://www.medpac.gov/docs/default-source/reports/chapter-3-appendixes-care-coordination-programs-for-dual-eligible-beneficiaries-june-2012-report-.pdf?sfvrsn=0.
     Anderson, Wayne L., Zhanlian Fen, and Sharon K. Long, 
RTI International and Urban Institute, Minnesota Managed Care 
Longitudinal Data Analysis, prepared for the U.S. Department of 
Health and Human Services Assistant Secretary for Planning and 
Evaluation (ASPE), March 2016, available at: https://aspe.hhs.gov/report/minnesota-managed-care-longitudinal-data-analysis.
---------------------------------------------------------------------------

    Comment: Several commenters encouraged CMS to consider further 
expanding our proposed passive enrollment authority to transition 
enrollees of non-renewing Medicare-Medicaid Plans (MMPs) into an 
integrated D-SNP.
    Response: We clarify that under the Financial Alignment Initiative 
capitated model demonstrations, MA regulations--including those 
governing passive enrollments--apply to MMPs unless waived. As has been 
the case to date under the demonstrations, we will continue to use our 
demonstration authority to waive applicable MA regulatory requirements 
in three-way contracts as necessary, and in partnership with each 
state, to achieve each individual demonstration's objectives.
    Comment: Several commenters supported the requirement for 
consultation with the state Medicaid agency that contracts with an 
eligible D-SNP, as proposed in Sec.  422.60(g)(1)(iii). Some commenters 
noted that this consultation would ensure both the proper utilization 
of CMS' passive enrollment authority and consistency

[[Page 16504]]

with states' integration goals and priorities. A commenter noted that 
this consultation would result in a more seamless process for states, 
integrated D-SNPs, and dually eligible beneficiaries. A few commenters 
noted that passive enrollment should occur at state discretion and 
pursuant to the State Medicaid Agency Contract with the D-SNP required 
under Sec.  422.107.
    Response: We appreciate the support for the proposed requirement 
that CMS consult with state Medicaid agencies to make a determination 
that D-SNPs meet the passive enrollment eligibility criteria and that 
the use of passive enrollment will promote integrated care and 
continuity of care for full-benefit dual eligible beneficiaries 
currently enrolled in an integrated D-SNP. We are committed to working 
with states to ensure that any passive enrollments under this authority 
meet CMS requirements as well as state priorities.
    Comment: A commenter requested that CMS clearly communicate the 
criteria for an integrated D-SNP to be eligible to accept passive 
enrollees in subregulatory guidance.
    Response: We anticipate issuing subregulatory guidance about the 
criteria for the passive enrollment authority finalized in this rule. 
We believe that the amendments to Sec.  422.60(g) as finalized here are 
sufficiently clear, particularly in light of the detailed discussion in 
the proposed rule and these various responses to comment, that 
implementation in CY2019 will not be confusing for D-SNPs that are 
qualified to receive enrollments.
    Comment: A commenter expressed concern that passive enrollment 
authority would be delegated to states. Another commenter recommended 
that CMS provide more clarification on whether CMS or state Medicaid 
agencies would be managing passive enrollment into integrated D-SNPs 
under our proposal, as well as on the implementation process for such 
passive enrollments.
    Response: When circumstances arise in which passive enrollment into 
an integrated D-SNP could potentially be applied, CMS will consult with 
the applicable state Medicaid agency, consistent with Sec.  
422.60(g)(1)(iii) as finalized. We anticipate that such consultation 
would include collaboration between CMS and the state Medicaid agency 
on issues such as identifying plans that meet the requirements in Sec.  
422.60(g)(2), decisions about enrollee assignment, and communications 
with impacted plans. We clarify that, as is the case today with respect 
to other passive enrollments into MA plans, affected D-SNPs will submit 
enrollment transactions to CMS' MARx system.
    Comment: Several commenters supported our proposed requirement in 
Sec.  422.60(g)(2)(ii) that a receiving integrated D-SNP have 
substantially similar provider and facility networks to the other MA 
integrated D-SNP plan (or plans) from which the passively enrolled 
beneficiaries are enrolled. A few commenters suggested that CMS limit 
the application of provider network and benefit similarity in order not 
to further narrow the scope of permissible passive enrollments into D-
SNPs.
    Response: We appreciate the support of our proposed requirement for 
provider network comparability as a minimum requirement for an 
integrated D-SNP's eligibility for passive enrollment. We disagree with 
the commenters' suggestion that we limit our eligibility analysis on 
provider network comparability given our emphasis on continuity of care 
in the application of this limited expansion of CMS' passive enrollment 
authority. We believe that this comparability analysis will minimize 
the number of enrollees whose provider relationships are disrupted as a 
result of passive enrollment and will encourage retention following 
enrollees' transition to a new integrated D-SNP. We are therefore 
finalizing the requirements for assessing network comparability as a 
condition for eligibility for passive enrollment under Sec.  
422.60(g)(1)(iii) as proposed.
    Comment: Several commenters requested clarification on how CMS will 
determine that the receiving integrated D-SNP has substantially similar 
provider and facility networks and Medicare- and Medicaid-covered 
benefits as the D-SNP from which the beneficiaries were passively 
enrolled.
    Response: We appreciate the commenters' request for clarification 
and anticipate issuing clarifications through subregulatory guidance. 
The subregulatory guidance will articulate the process and timing for 
the losing and receiving D-SNPs to submit networks through the CMS 
Health Plan Management System. CMS will also review plan benefit 
packages submitted by the impacted D-SNPs as well as engage the State 
Medicaid agency to ensure covered services are similar to services 
currently being received by impacted dual eligible beneficiaries.
    Comment: In addition to our proposed network comparability 
requirement, several commenters recommended the use of an ``intelligent 
assignment'' process for passively enrolling beneficiaries into a D-SNP 
based on the providers and prescription drugs associated with each 
individual beneficiary. Several commenters also recommended that, in 
our analysis of benefits comparability, CMS consider the comparability 
of the receiving D-SNP's formulary.
    Response: We agree that intelligent assignment processes would be 
helpful for ensuring care continuity and minimizing enrollee 
disruption. We will consider the availability of intelligent assignment 
processes when effectuating passive enrollments under this authority 
and will also consider intelligent assignment options in the future. 
However, we note that all plans offering Part D coverage must meet CMS' 
formulary adequacy requirements and, in addition, must offer a 
transition period upon a member's enrollment in a new plan. 
Specifically, under Sec.  423.120(b)(3), new enrollees must be provided 
a temporary supply of non-formulary Part D drugs, as well as Part D 
drugs with utilization management restrictions, and can work with their 
new plan and provider to switch to a different formulary drug or 
request an exception during their first 90 days of enrollment in their 
new plan.
    Comment: A commenter expressed concern that passive enrollment 
could further limit enrollee choice in states in which biologic 
medications are reimbursed at low rates under Medicaid.
    Response: We appreciate the commenter's concern about access to 
medically necessary drugs. We note that Medicare covers nearly all 
prescription drugs for dually eligible individuals under Parts A, B, 
and D. Medicaid coverage of drugs for dually eligible individuals is 
generally limited to over-the-counter drugs and products and 
prescription drugs that are otherwise excluded from the definition of a 
Part D drug. For dually eligible beneficiaries, the drugs referenced by 
this commenter would be covered under Medicare Part B rather than 
Medicaid.
    Comment: Several commenters recommended a transition period during 
which passively enrolled beneficiaries can see current providers that 
are not in their new plan's network. A few commenters also suggested 
that care plans and authorized services be continued for a period of 
time following passive enrollment.
    Response: We appreciate the commenters' suggestion that we 
incorporate continuity of care requirements into our proposed passive 
enrollment processes. We believe our finalization of the requirement 
for substantially similar provider and facility networks under Sec.  
422.60(g)(2)(ii) will facilitate continuity of care in most

[[Page 16505]]

cases. In addition, as previously discussed, the Part D transition 
requirements provide continuity of prescription drug benefits during a 
beneficiary's first 90 days of coverage in a new plan, including in 
cases where passive enrollment has been effectuated. We encourage 
states to consider using their State Medicaid Agency Contracts with D-
SNPs as a vehicle for requiring that any passive enrollments into 
integrated D-SNPs apply transition rules that align with those 
applicable to Medicaid managed care organizations under Sec.  
438.62(b). As previously noted, we are finalizing our provider and 
benefits comparability requirements at Sec.  422.60(g)(2)(ii) without 
further modification.
    Comment: Several commenters responded to our request for comment on 
CMS' proposal that an integrated D-SNP meet certain quality criteria to 
qualify for passive enrollment, particularly with respect to the 
proposed requirement that a D-SNP have an overall quality rating of at 
least 3 stars based on the MA Star Ratings system. Several commenters 
expressed support for our proposed application of a minimum overall MA 
Star Rating of at least 3 stars. A commenter noted that CMS' 
consultation with the state Medicaid agency would ensure that an 
integrated D-SNP's Medicaid performance is considered in addition to 
the Medicare performance captured by the MA Star Ratings. Several 
commenters recommended raising the minimum required MA Star Rating 
level. A commenter noted concerns with the MA Star Ratings as a basis 
for our proposed quality requirement because star ratings may be 
affected more by the percentage of dually eligible members enrolled in 
an MA plan than other factors and suggested requiring state approval 
instead of a minimum MA Star Rating. Some commenters expressed concern 
that use of MA Star Ratings does not capture plans' performance related 
to services covered under Medicaid or other factors affecting plan 
capacity to ensure access to care for passively enrolled individuals.
    Response: We appreciate commenters' support for establishing 
minimum quality criteria as part of our assessment of an integrated D-
SNP's eligibility for passive enrollment under this provision. We call 
attention to our revision to Sec.  422.60(g)(2)(iii), clarifying that 
the minimum star rating of at least 3 stars for a D-SNP to be eligible 
to receive passive enrollment from the most recently issued MA Star 
Rating for the D-SNP under the rating system described in Sec. Sec.  
422.160 through 422.166. While we acknowledge the limitations 
commenters identified with the MA Star Ratings, especially with respect 
to assessing the quality of Medicaid services provided under an 
integrated D-SNP, we believe the MA Star Ratings system is CMS' most 
effective and methodologically sound tool for measuring plan 
performance and quality and ensuring that passive enrollments are 
limited to MA plans that have demonstrated a commitment to quality. 
With regard to the methodological concerns related to the impact of 
enrollees' socioeconomic status on MA contract performance, we direct 
the commenter's attention to the discussion in this final rule about 
the MA and Part D Quality Rating System about adjustments to the 
ratings to address those and similar concerns in section II.A.11.t. We 
note that the additional required consultation with states in Sec.  
422.60(g)(1)(iii) as part of the process of determining that an 
integrated D-SNP meets the criteria for receipt of passive enrollment 
will provide valuable information regarding the performance and quality 
of the organization's Medicaid product. We are therefore finalizing the 
quality requirements under Sec.  422.60(g)(2)(iii) with a clarification 
that the most recently issued overall MA Star Rating is the applicable 
rating for determining eligibility to receive passive enrollment. We 
note as well that new and low enrollment plans are generally not 
assigned an overall Star Rating because of the lack of data from a 
prior performance period (new plans) or insufficient number of 
enrollees for reliable sampling (low enrollment); therefore, the 
regulation text as proposed and as finalized, permits new and low 
enrollment plans that meet the other requirements to also receive these 
passive enrollments. However, we will consider revisiting the minimum 
MA Star Rating level in future rulemaking once we gain additional 
experience with implementing passive enrollments into integrated D-
SNPs.
    Comment: Several commenters made additional recommendations for 
specific minimum quality measures and other criteria relevant to dually 
eligible beneficiaries that CMS should consider as part of our 
determination of integrated D-SNPs' eligibility for passive enrollment 
under proposed Sec.  422.60(g)(1)(iii). A few commenters recommended 
that CMS require integrated D-SNPs to have additional accreditation, 
such as the National Committee for Quality Assurance (NCQA) Medicaid 
plan accreditation and long-term services and supports (LTSS) 
accreditation. A commenter recommended using measures developed by the 
multi-stakeholder Core Quality Measures Collaborative. Another 
commenter suggested evaluating an integrated D-SNP's behavioral health 
services by number of days on waiting list and availability of a 
behavioral health expert. This commenter also suggested several methods 
for assessing LTSS.
    Response: We appreciate the additional information these commenters 
provided regarding accreditation and measures relevant to dually 
eligible beneficiaries. Since the number of plans eligible to receive 
passive enrollment under our proposed limited expansion of passive 
enrollment authority is projected to be small, we believe it is 
important to consider minimizing burden to eligible plans and ensuring 
that there are an adequate number of plans to receive enrollments. MA 
Star Ratings are based on currently reported plan data and do not 
impose additional reporting or specific accreditation requirements on 
integrated D-SNPs. As stated previously, we are finalizing the quality 
requirements for receipt of passive enrollment under Sec.  
422.60(g)(1)(iii) as proposed.
    Comment: We received no comments supporting a limitation of our 
proposed expansion of CMS' passive enrollment authority to 
circumstances that would not raise total cost to the Medicare and 
Medicaid programs. A few commenters stated they would not support a 
cost-effectiveness test as a standalone requirement for determining a 
D-SNP's eligibility to receive passive enrollments under our proposed 
rule. In addition, several commenters expressed concerns about 
establishing such a limitation for a variety of reasons. A commenter 
stated that a cost-effectiveness test would limit CMS' ability to align 
enrollment and preserve continuity of care. Another commenter believed 
that this approach did not consider long-term savings resulting from 
better integration. A few commenters also noted that the added cost and 
administrative burden involved in identifying these circumstances and 
measuring the cost-effectiveness of passive enrollment would 
potentially offset any cost-savings. Another commenter believed that 
choosing integrated D-SNPs for passive enrollment based on an 
artificial cost estimate would be inconsistent with the MA bid process 
and good faith contracting efforts.
    Response: We thank commenters for their comments on this issue. We 
are not adding a cost-effectiveness test for passive enrollments under 
paragraph (g)(1)(iii) in this final rule.

[[Page 16506]]

    Comment: In response to our request for comments on beneficiary 
notices for passive enrollments that would occur under proposed 
paragraph (g)(1)(iii), a few commenters supported maintaining the 
current requirement that receiving plans send one enrollee notice 
requirement when passive enrollment is applied, arguing that states or 
receiving plans could voluntarily choose to add more notifications as 
necessary, and that additional notices added to plan burden. A 
commenter noted that, because the Medicaid Managed Care Rule under 
Sec.  438.54(c)(3) requires the State to notice beneficiaries regarding 
passive enrollment into a Medicaid managed care plan but does not 
specify the number of notices required, a requirement of one notice 
under our proposed passive authority resulted in better alignment 
between Medicare and Medicaid requirements. However, many commenters 
recommended a more robust noticing process, including increasing the 
number of required notices to two for these passive enrollments. Some 
commenters also recommended that impacted plans provide the notices in 
beneficiaries' primary language and identify for each enrollee any 
providers or prescription drugs not included under their new plan. A 
few commenters recommended additional telephonic outreach for 
beneficiaries whose notices are returned by the postal service as 
undeliverable and for those whose primary language is not English.
    Response: We agree with most commenters on this issue that, on 
balance, two notices may be more beneficial than one notice when 
enrollees are being passively enrolled from one integrated D-SNP into 
another under paragraph (g)(1)(iii). A second notice provides an 
additional opportunity for the receiving D-SNP to connect with new 
members and to ensure they receive information about their benefits, 
rights, and options. We believe the benefits from an additional notice 
outweigh the additional burden. In contrast, passive enrollments 
effectuated under paragraphs (g)(1)(i) and (ii)--in other words, when 
an immediate termination as provided in Sec.  422.510(b)(2)(i)(B) 
occurs or when CMS determines a plan poses a potential risk of harm to 
enrollees--are typically performed under time constraints which may 
make the provision of two notices impracticable.
    We are therefore finalizing the notice requirements associated with 
passive enrollments under paragraph (g)(1)(iii) to require two notices 
and to establish parameters around the timing of such notices. 
Accordingly, we are adding new paragraph (g)(4)(ii) to require that 
plans receiving passive enrollments under paragraph (g)(1)(iii) send 
two notices to enrollees that describe the costs and benefits of the 
plan and the process for accessing care under the plan and clearly 
explain the beneficiary's ability to decline the enrollment or choose 
another plan. In addition, we are adding new paragraph (ii)(A) to 
specify that the first notice provided under paragraph (ii) must be 
provided, in a form and manner determined by CMS, no fewer than 60 days 
prior to the enrollment effective date. We are also adding a new 
paragraph (ii)(B) to specify that the second notice must be provided--
again, in a form and manner determined by CMS--no fewer than 30 days 
prior to the enrollment effective date.
    We clarify that for passive enrollments under paragraphs (g)(1)(i) 
and (ii), only one notice will be required. This requirement is now 
reflected in new paragraph (4)(i), which also specifies that the notice 
must describe the costs and benefits of the plan and the process for 
accessing care under the plan, as well as the beneficiary's ability to 
decline enrollment or choose another plan, and be provided prior to the 
enrollment effective date (or as soon as possible after the effective 
date if prior notice is not practical).
    We appreciate commenters' suggestions about the importance of 
telephonic outreach and will encourage affected plans to conduct this 
additional telephonic outreach. We will also encourage the D-SNPs 
losing members to passive enrollment into another plan to share 
information about their enrollees' language preferences to facilitate 
the provision of information in non-English languages and alternate 
formats as applicable. As we gain additional experience using this 
passive enrollment authority, we will consider the development of 
additional guidance or further rulemaking about beneficiary notice 
requirements as necessary.
    Comment: We received a number of comments about the content of 
beneficiary notices sent to passively enrolled individuals. Some 
commenters recommended that notices used as part of this process be 
consumer tested. Several commenters recommended that notices include 
alternative options for Medicare coverage, such as available PACE 
organizations. A few commenters suggested that the notices include 
information on the Special Election Period (SEP) and opt-out process. A 
few commenters also recommended that beneficiaries have access to 
individual counseling regarding their benefit options. A commenter 
recommended that notices be designed to ensure informed consent by 
affected enrollees.
    Response: We appreciate the suggestions commenters provided about 
the content of beneficiary notices for passive enrollment under 
paragraph (g)(1)(iii). We note that CMS currently requires notices sent 
to passively enrolled individuals to clearly explain the beneficiary's 
ability to decline the enrollment or choose another plan. We are 
therefore finalizing the requirements related to notice content without 
modification at Sec.  422.60(g)(4)(i) and (ii), as described elsewhere 
in this preamble. We agree with commenters who emphasized the 
importance of providing additional information and counseling to inform 
beneficiary choice. As we move forward with implementation of this 
limited expansion of CMS' passive enrollment authority, we will 
consider developing a notice template that includes information about 
the availability of resources for additional information and choice 
counseling in the impacted service area, including SHIP programs, as 
well as 1-800-Medicare and Medicare Plan Finder. We will consider 
opportunities for consumer testing notice language, though we note that 
each instance of passive enrollment under this authority will be unique 
and require tailoring to the specific circumstances. As noted 
previously, we believe that the addition of a second notice will help 
increase beneficiaries' awareness of the change to their coverage and 
ensure individuals have the information to make decisions about whether 
to remain in the new integrated D-SNP or select other coverage that 
better serves their needs.
    Comment: A few commenters recommended any beneficiary who is unable 
to be contacted should not be passively enrolled and should instead be 
defaulted into FFS Medicare.
    Response: We do not agree with these commenters. The individuals 
impacted by our proposal are those already enrolled in an integrated D-
SNP and who, absent our application of CMS' passive enrollment 
authority, would lose access to their current integrated care. Dually 
eligible individuals will have various SEPs available, including the 
Part D SEP for dual and other LIS-eligible beneficiaries discussed in 
section II.A.10 of this final rule and the new SEP at Sec.  
423.38(c)(10) discussed in section II.A.10 of this final rule that 
allows individuals who have been auto-enrolled, facilitated enrolled, 
passively enrolled, or reassigned into a plan by CMS an opportunity to 
change plans. These SEPs will allow any individual who does not wish to 
retain coverage

[[Page 16507]]

under his or her new integrated D-SNP to make a different election, 
including opting for coverage in FFS Medicare. We also note that the 
addition of the SEP at Sec.  423.38(c)(10) to this final rule renders 
the SEP described in current Sec.  422.60(g)(5) duplicative because it 
applies to all individuals who have been enrolled in a plan as a result 
of a CMS- or state-initiated enrollment action, including passive 
enrollment under Sec.  422.60(g). To avoid operational complexity, we 
are therefore finalizing this provision by replacing the language 
describing the SEP for passively enrolled individuals at Sec.  
422.60(g)(5) with a cross-reference to the new SEP described at Sec.  
423.38(c)(10).
    Comment: A commenter suggested that CMS provide additional 
opportunities for states to fully integrate Medicaid and Medicare 
noticing and beneficiary communications materials for integrated 
products.
    Response: We appreciate the support for further integration of 
Medicare and Medicaid benefits information for integrated D-SNPs and 
note that CMS has made progress toward this goal in collaboration with 
some state partners. However, this comment is outside the scope of this 
regulation.
    Comment: Several commenters requested clarification on how the SEP 
related to our proposed passive enrollment provision would be impacted 
by, or would interact with, the proposal to limit the Part D SEP for 
dual and other LIS-eligible beneficiaries.
    Response: As previously discussed, dually eligible beneficiaries 
will have access to other SEPs, including the Part D SEP for dual and 
other LIS-eligible beneficiaries and the new SEP finalized in this rule 
at Sec.  423.38(c)(10) that allows individuals who have been auto-
enrolled, facilitated enrolled, passively enrolled, or reassigned into 
a plan by CMS or a state an opportunity to change plans.
    Comment: A couple of commenters noted a lack of alignment between 
the length of the SEP for passive enrollees under Sec.  422.62(b)(4)--
that is, 60 days--and the 90-day disenrollment period afforded to 
enrollees passively enrolled into a Medicaid managed care organization 
under Sec.  438.56.
    Response: The commenters are correct that the length of the SEP for 
passive enrollees, as described in the proposal, and that of the 
Medicaid managed care disenrollment period are not the same. In certain 
integrated care programs, the combination of changes to the SEP for 
dual eligible beneficiaries (discussed in section II.A.10.of this final 
rule) and the 2-month period for the SEP in proposed Sec.  422.60(g)(5) 
could lead to beneficiary confusion and unintended misalignments 
between Medicare and Medicaid. As noted previously in this preamble, we 
are finalizing Sec.  422.60(g)(5) with modifications to replace the 
language describing the SEP for passively enrolled individuals with a 
cross-reference to the new SEP described at Sec.  423.38(c)(10). This 
SEP will allow individuals to opt out of the passive enrollment within 
3 months of notification of a CMS or state-initiated enrollment action 
or that enrollment action's effective date (whichever is later). We 
believe this change will better align the length of the SEP for 
individuals who are passively enrolled under Sec.  422.60(g) with the 
Medicaid managed care disenrollment period under Sec.  438.56.
    Comment: A commenter encouraged CMS to monitor any negative and 
unintended consequences of our use of passive enrollment after 
implementation of our proposed expanded authority.
    Response: We appreciate the commenters' concerns and clarify that 
we intend to use all currently available mechanisms to monitor any 
passive enrollments into integrated D-SNPs, including grievances and 
complaints reported to impacted plans and to 1-800-Medicare. We are 
committed to making all necessary adjustments as we gain experience 
with the application of passive enrollment in the circumstances 
provided for in this final rule, including future rulemaking as 
necessary.
    After consideration of the comments we received, we are finalizing 
our proposal regarding the expansion of CMS' regulatory authority to 
initiate passive enrollment for certain dually eligible beneficiaries 
who are currently enrolled in an integrated D-SNP into another 
integrated D-SNP at Sec.  422.60(g) with some modifications. 
Specifically, we are making the following modifications:
     We are making a technical revision to paragraph 
(g)(1)(iii) to clarify that a plan must meet all the requirements 
established in paragraph (g)(2) to be eligible to receive passive 
enrollment.
     We are revising paragraph (g)(2)(iii) to require a minimum 
Star Rating that applies for a plan to be eligible to receive passive 
enrollment. For a plan to be eligible to receive passive enrollment, it 
must have an overall quality rating, from the most recently issued 
ratings, under the rating system described in Sec. Sec.  422.160 
through 422.166, of at least 3 stars or is a low enrollment contract or 
new MA plan as defined in Sec.  422.252.
     We are adding new paragraph (g)(4)(ii) to require that 
plans receiving passive enrollments under paragraph (g)(1)(iii) send 
two notices to enrollees that describe the costs and benefits of the 
plan and the process for accessing care under the plan and clearly 
explain the beneficiary's ability to decline the enrollment or choose 
another plan. In addition, we are adding new paragraph (ii)(A) to 
specify that the first notice provided under paragraph (ii) must be 
provided, in a form and manner determined by CMS, no fewer than 60 days 
prior to the enrollment effective date. We are also adding a new 
paragraph (ii)(B) to specify that the second notice must be provided, 
in a form and manner determined by CMS, no fewer than 30 days prior to 
the enrollment effective date. New paragraph (g)(4)(i) will retain the 
original requirement that one notice be provided to passively enrolled 
individuals under paragraphs (g)(1)(i) and (ii).
     We are modifying Sec.  422.60(g)(5) by replacing the 
current language describing the SEP for passively enrolled individuals 
at Sec.  422.60(g)(5) with a cross-reference to the new SEP described 
at Sec.  423.38(c)(10), which provides a 3-month SEP when an enrollee 
has been auto-enrolled, facilitated enrolled, passively enrolled, or 
reassigned into a Part D plan as a result of a CMS or state-initiated 
enrollment action. We note that all D-SNPs are also Part D plans as 
they are required to provide the Part D prescription drug benefit 
pursuant to Sec.  422.2 (definition of specialized MA plans for special 
needs individuals).
9. Part D Tiering Exceptions (Sec. Sec.  423.560, 423.578(a) and (c))
a. Background
    Section 1860D-4(g)(2) of the Act specifies that a beneficiary 
enrolled in a Part D plan offering prescription drug benefits for Part 
D drugs through the use of a tiered formulary may request an exception 
to the plan sponsor's tiered cost-sharing structure. The statute 
requires such plan sponsors to have a process in place for making 
determinations on such requests, consistent with guidelines established 
by the Secretary. The requirements for tiering exceptions, set forth at 
Sec.  423.578(a), require plan sponsors to establish and maintain 
reasonable and complete exceptions procedures that permit enrollees, 
under certain circumstances, to obtain a drug in a higher cost-sharing 
tier at the more favorable cost-sharing applicable to alternative drugs 
on a lower cost-sharing tier of the plan sponsor's formulary. Such an 
exception is granted when the plan sponsor determines that the non-

[[Page 16508]]

preferred drug is medically necessary based on the prescriber's 
supporting statement.
    As we stated in the proposed rule, we believe that changes in the 
prescription drug marketplace necessitate revisions to existing 
regulations to ensure that tiering exceptions are adjudicated by plan 
sponsors in the manner the statute contemplates, and are understood by 
beneficiaries. Therefore, we proposed various changes to Sec. Sec.  
423.560, 423.578(a) and 423.578(c) to revise and clarify requirements 
for how tiering exceptions are to be adjudicated and effectuated (82 FR 
56371).
    We received the following general comments on this proposal and our 
responses follow:
    Comment: We received many comments on the proposal. While most 
comments received were generally supportive of our efforts to update 
and improve tiering exceptions policy, there was mixed support for and 
opposition to specific aspects of what we proposed. Many commenters who 
supported our overall proposal noted that beneficiaries have difficulty 
understanding the existing policy, and stated that there is a need for 
a more simplified process. A commenter who opposed revising our 
existing policy for tiering exceptions stated that plans and enrollees 
already understand the current policy and there will be little positive 
outcome. Another commenter agreed that tiering exceptions are an 
important beneficiary protection, but stated a belief that they 
undermine plan sponsors' ability to manage their formularies, which are 
already reviewed by CMS for clinical accuracy. This commenter also 
stated that tiering exceptions provide no incentive for an enrollee to 
try a less expensive drug found on a lower tier if they are able to get 
a more expensive drug at a lower cost.
    Response: We thank the commenters who supported our proposal for 
their support. We agree that this policy area has been confusing for 
beneficiaries and one of our goals in making changes is to make it more 
understandable. We believe that the proposed revisions will streamline 
and clarify the requirements for tiering exceptions, as well as help 
ensure that enrollees have appropriate access to medically necessary 
drugs.
    We disagree with the comment that tiering exceptions provide no 
incentive for enrollees to try lower-cost drugs. On the contrary, Sec.  
1860D-4(g)(2) stipulates that, in order for a tiering exception to be 
approved, the enrollee's prescriber must determine that the preferred 
drug for treatment of the same condition has been or would be less 
effective or have adverse effects for that individual. If the enrollee 
cannot demonstrate that the requested drug is medically necessary, a 
tiering exception cannot be obtained.
    We address comments about specific aspects of the tiering 
exceptions proposal in relevant sections below.
    Comment: Several commenters requested that CMS ensure beneficiaries 
are educated about the availability of tiering exceptions. Some 
commenters expressed a belief that there is little information 
available to beneficiaries about tiering exceptions, and that it is 
difficult to apply to individual situations. Comments offered several 
suggestions, including improving existing educational publications and 
information provided through 1-800-MEDICARE, providing information in 
plain language, and developing notices that provide information at the 
pharmacy counter. Some commenters stated that CMS should require plan 
sponsors to improve information provided in their member materials, and 
noted that plans and pharmacies have a responsibility for educating 
beneficiaries about the availability of tiering exceptions.
    Response: We agree that information about the availability of 
tiering exceptions must be provided to beneficiaries by CMS and their 
Part D plan sponsor. We note that such information is already contained 
in several CMS publications, including Medicare & You (CMS pub. 10050), 
Medicare Appeals (CMS pub. 11525), Your Guide to Medicare Prescription 
Drug Coverage (CMS pub. 11109) and Medicare Rights and Protections (CMS 
pub. 11534), as well as documents that plans are required to provide to 
enrollees, including the Evidence of Coverage, Part D formulary, and 
Annual Notice of Change. Information about the availability of tiering 
exceptions is also included in the standardized pharmacy notice (CMS-
10147) provided to affected enrollees at the point of sale when a claim 
is rejected by their Part D plan sponsor, and in the standardized Part 
D denial notice (CMS-10146), which is provided to enrollees when their 
plan makes an adverse coverage determination. Such information is also 
found on Medicare.gov. CMS will continue to review plan documents and 
beneficiary publications to identify potential areas for improvement, 
and update the documents mentioned above as needed based on this final 
rule, including consideration of how to clarify when a tiering 
exception may be available.
    Comment: Several commenters recommended that CMS ensure consistent 
understanding of tiering exceptions policy by providing specific 
guidance to plan sponsors related to the review of tiering exception 
requests, including examples using various formulary structures that 
illustrate the steps of the process, and guidance to determine the 
lowest applicable tier and appropriate alternative drugs. A commenter 
expressed concern that the proposed rule conflicts with current 
guidance in Chapter 18 of the Medicare Prescription Drug Benefit 
Manual.
    Response: We appreciate the commenters' suggestions for additional 
guidance to ensure that plan sponsors understand the revised policy and 
properly process tiering exception requests. CMS manual guidance will 
be updated to reflect the changes made through this final rule. With 
respect to the comment about the existing version of Chapter 18, we 
note that existing guidance reflects existing regulations and policy.
    Comment: A commenter asserted that utilization management tools, 
such as the use of tiered cost-sharing to encourage use of lower-cost 
drugs, put unnecessary burden on prescribers and cause access delays 
for beneficiaries. The commenter stated that exception requests usually 
require prescribers to submit a written statement supporting the 
exception request, and noted that prescribers are not compensated for 
time spent preparing these statements or obtaining utilization 
management information for the specific plans used by their patients. 
This commenter also suggested that if there was greater transparency on 
which medications are subject to utilization management tools, it would 
reduce the administrative burden placed on physicians.
    Response: We thank the commenter for sharing their concerns. 
Because section 1860D-4(g)(2) of the Act specifies that a tiering 
exception could be granted ``if the prescribing physician determines 
that the preferred drug for treatment of the same condition either 
would not be as effective for the individual or would have adverse 
effects for the individual or both,'' we do not believe CMS has 
authority to require plans to provide tiering exceptions in the absence 
of such a statement from the prescriber. Under existing Sec.  
423.568(a), plans are required to accept oral requests for benefits at 
the coverage determination level, including exception requests, and CMS 
encourages plans to accept oral prescriber supporting statements for 
exception requests when appropriate.
    Comment: A commenter recommended that SNPs, MMPs, and defined 
standard benefit plans be exempt from the tiering exceptions process. 
This commenter also asked that

[[Page 16509]]

CMS explain how tiering exceptions are applied to Low Income Subsidy 
(LIS) beneficiaries.
    Response: We appreciate the commenter's recommendation. In 
accordance with Sec.  423.578(a), the exceptions process applies to 
Part D plans that provide prescription drug benefits through the use of 
a tiered formulary. Given the fixed copays for LIS beneficiaries, that 
are based on whether the drug is a brand or generic product pursuant to 
Sec.  423.782(a)(2)(iii)(A), tiering exceptions do not apply. 
Regardless of whether the beneficiary meets the medical necessity 
criteria for the drug in the higher tier, it would not change the brand 
vs. generic nature of the requested drug, so the cost-sharing would 
remain fixed.
b. Limitations on Tiering Exceptions
    We proposed to revise Sec.  423.578(a)(2) to read as follows: 
``Part D plan sponsors must establish criteria that provide for a 
tiering exception consistent with paragraphs Sec.  423.578(a)(3) 
through (a)(6) of this section.'' This adds a cross-reference to 
revised paragraph (a)(6), which revises allowable limitations plan 
sponsors are permitted to establish in their tiering exceptions 
procedures.
    At Sec.  423.578(a)(6), we proposed to revise the regulations to 
specify how a Part D plan sponsor may limit tiering exceptions. The 
proposed revision strikes the existing regulation text which permits 
plans to disallow tiering exceptions for any non-preferred drug to 
cost-sharing associated with a dedicated generic tier. We proposed to 
replace it with new regulation text at Sec.  423.578(a)(6) specifying 
that a Part D plan sponsor will not be required to offer a tiering 
exception for a brand name drug or biological product to a preferred 
cost-sharing level that applies only to generic alternatives. Under our 
proposal, plans would be required to approve tiering exceptions for 
non-preferred generic drugs when the plan determines that the enrollee 
cannot take the preferred generic alternative(s), including when the 
preferred generic alternative(s) are on dedicated generic tier(s) and 
when the lower tier(s) contain a mix of brand and generic alternatives. 
In other words, plans would no longer be permitted to exclude a tier 
containing alternative drug(s) with more favorable cost-sharing from 
their tiering exceptions procedures altogether just because that lower-
cost tier includes only generic drugs.
    We proposed to revise existing tiering exceptions policy for brand 
name and generic drugs, and proposed a new policy for requests 
involving biological products. First, we proposed to revise Sec.  
423.578(a)(6) by adding new paragraphs (i) and (ii), which would permit 
plans to limit the availability of tiering exceptions for the following 
drug types to a preferred tier that contains the same type of 
alternative drug(s) for treating the enrollee's condition:
     Brand name drugs for which an application is approved 
under section 505(c) of the Federal Food, Drug, and Cosmetic Act (21 
U.S.C. 355(c)), including an application referred to in section 
505(b)(2) of the Federal Food, Drug, and Cosmetic Act (21 U.S.C. 
355(b)(2)); and
     Biological products, including biosimilar and 
interchangeable biological products, licensed under section 351 the 
Public Health Service Act.
    With the proposed revisions, approved tiering exceptions for brand 
name drugs would generally be assigned to the lowest applicable cost-
sharing associated with brand name alternatives, and approved tiering 
exceptions for biological products would generally be assigned to the 
lowest applicable cost-sharing associated with biological alternatives. 
As discussed above, cost sharing for approved tiering exceptions for 
non-preferred generic drugs would be assigned to the lowest applicable 
cost-sharing associated with alternative drug(s) that could be either 
brand name or generic drugs.
    We proposed at Sec.  423.578(a)(6)(i) to codify that plans are not 
required to offer tiering exceptions for brand name drugs or biological 
products at a cost-sharing level of alternative drug(s) for treating 
the enrollee's condition where the alternatives include only the 
following drug types:
     Generic drugs for which an application is approved under 
section 505(j) of the Federal Food, Drug, and Cosmetic Act (21 U.S.C. 
355(j)), or
     Authorized generic drugs as defined in section 505(t)(3) 
of the Federal Food, Drug, and Cosmetic Act (21 U.S.C. 355(t)(3)).
    We proposed to codify existing CMS policy treating authorized 
generics as generics for purposes of tiering exceptions because the 
process used by CMS to collect Part D plan formulary data does not 
allow us to clearly identify whether a plan sponsor includes coverage 
of authorized generic National Drug Codes (NDCs). Under this regulatory 
proposal, a plan sponsor could not completely exclude a lower tier 
containing only generic and authorized generic drugs from its tiering 
exception procedures; rather, the plan sponsor would be permitted to 
limit tiering exceptions for a particular brand drug or biological 
product to the lowest cost sharing tier containing alternatives of the 
same drug type. Plans will be required to grant a tiering exception for 
a higher cost generic or authorized generic drug to the cost sharing 
associated with the lowest tier containing generic and/or authorized 
generic alternatives when the medical necessity criteria are met.
    Finally, we proposed to revise and redesignate existing Sec.  
423.578(a)(7) as new Sec.  423.578(a)(6)(iii), to specify that, ``If a 
Part D plan sponsor maintains a specialty tier, as defined in Sec.  
423.560, the sponsor may design its exception process so that Part D 
drugs and biological products on the specialty tier are not eligible 
for a tiering exception.'' We also proposed to add the following 
definition to Subpart M at Sec.  423.560:
    Specialty tier means a formulary cost-sharing tier dedicated to 
very high cost Part D drugs and biological products that exceed a cost 
threshold established by the Secretary.
    The proposed changes retain the existing regulatory policy that 
permits Part D plan sponsors to disallow tiering exceptions for any 
drug that is on the plan's specialty tier. While we did not propose to 
specify it in regulation text, we stated in the preamble to the 
proposed rule (82 FR 56372) that, if the specialty tier has cost 
sharing more preferable than another tier, then a drug placed on such 
other non-preferred tier is eligible for a tiering exception to the 
cost sharing applicable to the specialty tier if an applicable 
alternative drug is on the specialty tier and the other requirements of 
Sec.  423.578(a) are met. In other words, while plans are not required 
to allow tiering exceptions for drugs on the specialty tier to a more 
preferable cost-sharing tier, the specialty tier is not exempt from 
being considered a preferred tier for purposes of tiering exceptions.
    We received the following comments and our responses follow:
    Comment: We received many comments on this aspect of our proposal. 
Most commenters were supportive of the proposal to remove the generic 
tier exclusion and replace it with limitations that apply to brand name 
drugs and biological products. Some commenters opposed our proposal to 
remove the generic tier exclusion, stating that this would discourage 
plans from offering $0 copayment tiers and increase costs for 
enrollees. Others opposed the proposal to allow plans to limit tiering 
exceptions for brand name drugs only when brand alternatives are on a 
lower tier, noting that allowing plans to limit tiering exceptions for 
brand drugs to the lowest

[[Page 16510]]

cost-sharing associated with brand alternatives does not provide 
sufficient relief for enrollees with a medical need for a brand drug 
because they cannot take a lower cost generic. Commenters expressed 
concern that this would eliminate beneficiaries' ability to seek 
tiering exceptions in many cases, and also stated that nothing in the 
statute permits these limitations.
    Response: We thank commenters who supported the proposed changes 
for their support. As we stated in the proposed rule, we believe a 
policy that allows beneficiaries with a medical need for a non-
preferred product to seek and obtain more favorable cost-sharing 
through the tiering exceptions process must be balanced by reasonable 
limitations to ensure that all enrollees have access to medically 
necessary drugs at the most favorable cost-sharing terms possible.
    We disagree with the commenters opposed to our proposal to require 
plans to include dedicated generic tiers in their tiering exceptions 
procedures. As we discussed in the preamble to the proposed rule (82 FR 
56371), most Part D formularies now include multiple generic tiers, as 
well as multiple higher-cost tiers that contain a mix of brand and 
generic drugs. To encourage the use of generic drugs, we proposed to 
revise the existing regulatory policy to permit tiering exceptions into 
dedicated generic tiers, but allow plans to limit those exceptions to 
requests involving non-preferred generic drugs. Because approval of a 
tiering exception continues to require that the enrollee demonstrate a 
medical need for the non-preferred drug, and because plans will not be 
required to permit exceptions for brand name drugs or biological 
products to the cost-sharing associated with dedicated generic tiers, 
we do not believe this change will result in changes to plan benefit 
design.
    We disagree with the comments asserting that the statute does not 
permit tiering exceptions for non-preferred brand name drugs to be 
limited to the cost sharing associated with preferred brand name drugs. 
Section 1860D-4(g)(2) of the Act specifies that Part D plan sponsors 
offering a tiered drug benefit must have a process for tiering 
exceptions, consistent with guidelines established by the Secretary for 
making such determinations, where ``a nonpreferred drug could be 
covered under the terms applicable for preferred drugs'' (emphasis 
added). While we agree that the statutory language does not 
specifically refer to brand name and generic drugs, it clearly gives 
CMS authority to establish guidelines for plan procedures, and does not 
require that such exceptions be available in all circumstances.
    Comment: Several commenters supported our proposal to treat 
authorized generic drugs in the same manner as generic drugs for 
tiering exceptions.
    Response: We thank the commenters for their support.
    Comment: We received some comments requesting that CMS specify that 
multi-source drugs and other drugs that do not meet the definition of a 
generic or authorized generic drug, but that a plan may place on a 
generic-labeled tier, also be treated as generic drugs for purposes of 
tiering exceptions.
    Response: We disagree with these comments. As discussed above, we 
are revising the tiering exceptions regulations to specify that 
authorized generic drugs be treated as generic drugs. We recognize that 
other drugs may be treated in a similar manner to generic drugs, 
including being placed on generic-labeled drug tiers; however, we 
believe further expansion of what drugs are treated as generics would 
introduce additional complexity to a process that beneficiaries and 
plans already have difficulty understanding. For example, whether a 
brand drug is a ``multi-source'' drug is dependent on multiple factors 
and may change over time. An authorized generic is determined at the 
time of FDA approval and does not change as long as the drug is 
marketed under that approval, regardless of how many other 
interchangeable drugs may be introduced to or leave the market. Because 
tier placement of the same drug can vary widely across Part D plans, we 
believe that applying rules based on FDA approval type is the best way 
to limit confusion and create a consistent policy. Additionally, we 
believe that an enrollee who cannot take a brand drug on a lower-cost 
tier, regardless of the tier label, should be able to obtain the brand 
drug on a higher-cost tier at the more favorable cost-sharing of the 
brand drug on the lower-cost tier.
    Comment: We received many comments related to our proposal to 
retain the current regulatory policy allowing plans to exclude 
specialty tier drugs from their tiering exceptions process. Commenters 
were divided on whether they supported or opposed this proposal. Some 
commenters asked CMS to confirm that drugs on the specialty tier will 
continue to be exempt from tiering exceptions.
    Commenters who supported our proposal stated that tiering 
exceptions should not be allowed for specialty tier drugs because 
alternative drugs on lower tiers are not typically appropriate or 
therapeutically equivalent, even though they may treat the same 
condition.
    Commenters who opposed this limitation on tiering exceptions noted 
that vulnerable beneficiaries who need to access specialty tier drugs 
often do not have alternative options on more preferred tiers and can 
accrue very high out of pocket costs. A few noted that cost-prohibitive 
out of pocket expenses can lead to decreased adherence to drug 
therapies and put patients at risk. Some commenters questioned CMS' 
authority to allow plans to exclude specialty tier drugs from the 
tiering exceptions process because the statute gives beneficiaries the 
right to request a tiering exception for any non-preferred drug when 
the formulary contains a preferred drug for the same condition that has 
lower cost sharing. A commenter stated that prohibiting tiering 
exceptions for specialty tier drugs discriminates against beneficiaries 
who need them.
    Response: We appreciate the comments expressing concern about 
beneficiary access to very high cost drugs. While CMS is aware that 
access to needed drug therapies can be impacted by the out of pocket 
expenses associated with these drugs, we do not believe that requiring 
plans to offer tiering exceptions for specialty tier drugs will result 
in the desired effect. In order for a drug to be placed on the 
specialty tier, the plan's negotiated price for the drug must exceed a 
monthly threshold established by the Secretary ($670 for 2018). Along 
with the protection against tiering exceptions for specialty tier drugs 
that is afforded to plans, CMS also requires plans to limit enrollee 
cost sharing for the specialty tier to 25 percent coinsurance (up to 33 
percent if the plan waives all or part of the Part D deductible), which 
aligns with the statutorily defined maximum cost sharing for the 
defined standard benefit at section 1860D-2(b)(2)(A). When high cost 
drugs are placed on the specialty tier instead of a Non-Preferred Brand 
or Non-Preferred Drug tier, which can have up to 50 percent 
coinsurance, the cost to enrollees who would not qualify for a tiering 
exception is often considerably lower than if the same drug were placed 
on one of these other non-preferred tiers. Additionally, many specialty 
tier drugs, particularly biological products, often do not have viable 
alternatives on lower-cost tiers. The statutory basis for approval of a 
tiering exception request is the presence of an alternative drug(s) on 
a lower cost-sharing tier of the plan's formulary; therefore, even if a 
plan sponsor permitted tiering exceptions for

[[Page 16511]]

specialty tier drugs, such requests would not be approvable if the 
plan's formulary did not include any alternative drugs on a lower tier.
    We disagree with the comments positing that allowing plans to 
exclude the specialty tier from their tiering exceptions procedures is 
inconsistent with the statute. As discussed above in this section, 
section 1860D-4(g)(2) of the Act gives CMS authority to establish 
guidelines for Part D plan sponsors' tiering exceptions procedures, and 
does not require such exceptions to be available in all circumstances. 
For the reasons stated earlier, we believe that our current policy of 
allowing plans to exclude specialty tier drugs from their tiering 
exceptions procedures, coupled with the maximum allowable coinsurance 
of 25 percent to 33 percent for the specialty tier, affords the most 
beneficiaries the most protection from high out-of-pocket expenses 
associated with very high cost drugs.
    Comment: A few commenters suggested that CMS permit plan sponsors 
to designate two specialty tiers on their formularies--a non-preferred 
specialty tier, as well as a preferred specialty tier that would have 
lower cost sharing. These commenters expressed a belief that permitting 
plans to have two specialty tiers would encourage increased competition 
among specialty drugs, giving plans greater leverage in price 
negotiations, resulting in more affordable access for Part D enrollees 
and lower costs for the program. The commenters also noted that 
permitting two specialty tiers could encourage enrollees to try 
preferred specialty products and could reduce the need for enrollees to 
seek coverage through the non-formulary exceptions process.
    Response: While we appreciate these comments, we disagree with the 
suggestion to permit Part D plans to have a preferred and a non-
preferred specialty tier. As discussed above, CMS limits specialty tier 
cost sharing to the statutorily mandated amount for the defined 
standard Part D benefit. While we did not propose to allow plans to 
establish multiple specialty tiers, we are making significant changes 
to existing tiering exceptions policy through this final rule, 
including removal of the generic tier exclusion and addition of the 
brand-to-brand limitation discussed above in subsection b. 
Additionally, while the plan's cost for a drug must exceed a CMS-
specified monthly cost threshold in order to be placed on the specialty 
tier, CMS does not require all drugs exceeding that threshold be placed 
on the specialty tier. In other words, if plans wish to encourage the 
use of certain specialty drugs over others, they can do so within 
existing formulary benefit designs. As such, we are not making 
additional changes in this policy area before having an opportunity to 
consider the effects of the changes in this rule. CMS will continue to 
disallow plan benefit packages with more than one specialty tier.
    Comment: We received some comments requesting that CMS clarify 
whether select care/select diabetic or other $0 copayment tiers can be 
excluded from a plan's tiering exceptions procedures. These commenters 
supported a policy that would permit such an exclusion, stating that 
requiring tiering exceptions to $0 or very low cost tiers would 
discourage plans from offering them and increase overall beneficiary 
out of pocket costs.
    Response: We appreciate the commenter's requests for clarification. 
As discussed above, we proposed to revise the existing regulatory text 
that permits plans to exclude generic tiers from their tiering 
exceptions procedures. We did not propose to permit plans to exclude 
any formulary tiers other than the specialty tier, and do not agree 
that such an exclusion is advisable. As we stated in the proposed rule, 
we believe that tiering exceptions are an important enrollee protection 
and must not be restricted to such a degree. Under the proposed rule, 
which we are finalizing without modification, plans can establish 
tiering exceptions procedures where they do not have to offer such 
exceptions for brand name drugs or biological products to more 
preferred cost-sharing tiers that do not contain an alternative brand 
name or biological product, respectively. We believe that permitting 
additional restrictions that make certain low-cost tiers wholly 
inaccessible to beneficiaries with a medical need for a non-preferred 
drug would be inappropriate.
    Comment: A commenter urged CMS to monitor Part D plan formularies 
to ensure that plans do not change their formularies in an effort to 
decrease opportunities for tiering exceptions. Another commenter 
suggested that CMS consider requiring plan sponsors to establish 
evidence-based formularies that tie enrollee cost-sharing to the 
appropriateness of medications based on safety and efficacy.
    Response: All Part D plan formularies must be approved by CMS as 
part of the bid review process described at Sec.  423.272. Under Sec.  
423.120(b)(1), formularies must be developed and reviewed by a pharmacy 
and therapeutic committee that makes clinical decisions based on 
scientific evidence and standards of practice and considers safety and 
efficacy when determining inclusion of a drug on a formulary, including 
tier placement.
    Comment: We received a comment requesting that CMS clarify non-
formulary drugs approved for a formulary exception continue to be 
ineligible for tiering exceptions. Another commenter suggested that CMS 
consider ways to make it easier for individuals applying for a 
formulary exception to also apply for a tiering exception, if 
applicable.
    Response: We appreciate the commenter's request for clarification. 
We did not propose to revise the existing requirement set forth at 
Sec.  423.578(c)(4)(iii) which establishes that an enrollee may not 
request a tiering exception for a non-formulary drug approved under the 
formulary exceptions rules at Sec.  423.578(b). Under the proposed 
changes to tiering exceptions rules, which we are finalizing as 
proposed, an enrollee may not obtain a tiering exception for an 
approved non-formulary drug. We note that, if an enrollee obtains an 
exception to a utilization management requirement such as step therapy 
or a quantity limit, such enrollee may also request a tiering 
exception, pursuant to Sec.  423.578(a) and (c). The model Part D 
coverage determination request form, developed by CMS with stakeholder 
feedback, permits an enrollee or their prescriber, on the enrollee's 
behalf, to request a tiering exception along with, for example, prior 
authorization. The form includes check boxes for various types of 
requests, including an exception to cost-sharing.
    Comment: We received some comments opposed to requiring plans to 
consider tiering exceptions for non-preferred drugs to specialty tier 
cost-sharing when the specialty tier cost-sharing is more favorable for 
the enrollee. Some of these commenters stated that such a policy would 
be confusing for enrollees because the specialty tier is often a 
higher-numbered tier (for example, tier 5 on a 5-tier formulary). 
Commenters also stated that it would be overly burdensome for plans to 
administer such a policy, particularly if the exception request is for 
a drug on a copayment tier to a coinsurance tier (for example, tier 4--
Non Preferred Drug has a $100 copayment and tier 5--Specialty has a 25 
percent coinsurance). These commenters opined that allowing a drug with 
a copayment to be approved to a coinsurance tier would bypass formulary 
design and require extensive price review and calculation to determine 
which tier is more favorable. A commenter asked CMS to clarify whether 
plans would be permitted to

[[Page 16512]]

retain specialty tier supply limits such as a 30 day supply, even if 
the enrollee wishes to obtain a 90 day supply and a tiering exception 
is approved.
    Response: We appreciate the comments received on this aspect of the 
proposal. We are persuaded by the comments received that requiring 
plans to consider tiering exceptions into the specialty tier would be 
confusing and difficult for plans to implement, and are not finalizing 
this aspect of the proposal. While we believe many of the concerns 
expressed by commenters would be addressed by clarifying that such a 
policy would only apply if the requested drug meets the specialty tier 
cost threshold, we recognize it would still be difficult to explain to 
enrollees, who probably would have no knowledge as to whether any given 
drug would meet the specialty tier cost threshold and would be very 
unlikely to request such an exception. As noted above, we did not 
propose regulation text for such a requirement, and therefore, while we 
are not finalizing it, we are also not making any changes to the 
proposed regulation text.
    Comment: A few commenters stated that CMS should conduct an 
analysis of Part D plan formularies to ensure plans are not 
discriminating against beneficiaries by always placing certain classes 
of drugs on specialty tiers. A commenter asserted that, without 
standardized tiering in Part D, nothing prevents plans from putting 
high cost brand name drugs on specialty tiers to avoid having to offer 
tiering exceptions. The commenter stated that CMS should establish 
additional requirements for tiered formularies, such as requiring that 
all generic drugs be placed on tier 1 or tier 2. Another commenter 
recommended that CMS continue to explore improvements to benefit design 
and meaningful exceptions to high cost-sharing.
    Response: Pursuant to existing Part D policy and the proposed 
definition of specialty tier, it is a tier dedicated to very high cost 
drugs, which are often brand name drugs or biological products. As 
noted in a previous response, pursuant to Sec.  423.120(b)(1), 
formularies must be developed and reviewed by a pharmacy and 
therapeutic committee that makes clinical decisions based on scientific 
evidence and standards of practice, and considers safety and efficacy 
when determining inclusion of a drug on a formulary, including that 
drug's tier placement. While CMS does not prohibit plan sponsors from 
having a mix of both brand and generic drugs on each tier, it is our 
expectation that a tier label be representative of the drugs that make 
up that tier. Additionally, consistent with Sec.  30.2.7 of Chapter 6 
of the Medicare Prescription Drug Benefit Manual, CMS reviews 
formularies for the placement of drugs in non-preferred tiers in the 
absence of therapeutically similar drugs in preferred tiers.
    Comment: A few commenters stated that CMS should increase the $670 
specialty tier cost threshold to reduce the number of drugs that 
qualify and, therefore, reduce out of pocket spending for 
beneficiaries.
    Response: As we did not propose to change the specialty tier 
threshold in this rule, we decline to adopt this recommendation.
    After consideration of the comments received, we believe our 
proposed revisions to Sec.  423.578(a)(6) regarding the limitations 
plans are permitted to establish for tiering exceptions strike an 
appropriate balance between allowing plans to manage their formularies 
and ensuring enrollee access to this statutory protection. These 
revisions prohibit plans from excluding generic drug tiers from their 
tiering exceptions procedures, and permit plans to limit tiering 
exceptions for brand name drugs to the lowest applicable cost sharing 
associated with preferred brand name alternatives, and tiering 
exceptions for biological products to the lowest applicable cost 
sharing associated with preferred biological product alternatives. We 
are finalizing the proposed revisions to Sec.  423.578(a)(6) and the 
proposed definition of specialty tier at Sec.  423.560 without 
modification, noting the clarification discussed above that plans are 
not required to treat the specialty tier as a preferred cost-sharing 
tier for purposes of tiering exceptions. CMS continues to explore ways 
to ensure Part D enrollees are able to access very high cost, medically 
necessary prescription drugs.
d. Alternative Drugs for Treatment of the Enrollee's Condition
    We noted in the proposed rule that we have received comments from 
plan sponsors and PBMs requesting that CMS provide additional guidance 
on how to determine what constitutes an alternative drug for purposes 
of tiering exceptions, including establishment of additional 
limitations on when such exceptions are approvable. The statutory 
language for tiering and formulary exceptions at sections 1860D-4(g)(2) 
and 1860D-4(h)(2) of the Act, respectively, specifically refers to a 
preferred or formulary drug ``for treatment of the same condition.'' 
While our proposal did not include regulation text specific to the 
meaning of an alternative drug, we clarified in the preamble that we 
interpret this language to refer to the condition as it affects the 
enrollee--that is, taking into consideration the individual's overall 
clinical condition, including the presence of comorbidities and known 
relevant characteristics of the enrollee and/or the drug regimen, which 
can factor into which drugs are appropriate alternative therapies for 
that enrollee.
    We received the following comments on this section and our 
responses follow:
    Comment: We received several comments related to how to determine 
which drugs should be considered alternatives for treating the 
enrollee's health condition. Some of these commenters were supportive 
of the additional information we provided in the preamble to the 
proposed rule about how to determine alternative drugs. Most of the 
commenters stated that a more specific regulatory definition of 
alternative drug is needed. Some commenters recommended that the 
definition specify that alternative drugs must be one or more of the 
following: supported in drug compendia or treatment guidelines for use 
in the same place in therapy, FDA-approved for the same indication as 
the requested drug, in the same therapeutic class and/or category as 
the requested drug, use the same route of administration as the 
requested drug, and/or have the same mechanism of action as the 
requested drug.
    Several commenters provided various hypothetical scenarios using 
specific diagnoses and drugs and asked that CMS clarify whether a 
tiering exception would be allowed under our interpretation. A 
commenter asked CMS to provide examples that include how to determine 
what an appropriate alternative drug is. Another commenter stated that 
plan sponsors will continue to inaccurately apply rules for tiering 
exceptions because CMS does not define what a preferred alternative 
drug is. A few commenters stated that CMS' proposed interpretation of 
``same condition'' will limit exception requests and negatively impact 
beneficiaries. A few commenters stated that this interpretation has no 
statutory basis, and one of the commenters asserted that our 
clarification basing what constitutes an alternative drug on the 
individual characteristics and condition of the enrollee would make it 
easy for plans to claim there are no alternatives for treating that 
enrollee and therefore no tiering exception would be allowed.
    Response: The statutory language noted above related to approval of 
a tiering exception request broadly refers to preferred drugs ``for 
treatment of the

[[Page 16513]]

same condition.'' We believe that most of the criteria suggested by 
commenters would be more restrictive than the statute allows if plans 
were required to apply such criteria to all tiering exception 
situations, and we therefore disagree that such criteria should be 
specified in regulation. For example, if the mechanism of action or 
route of administration of a plan's preferred alternative drug would 
cause adverse effects for a particular enrollee versus the non-
preferred drug for treating the same condition, this could be the basis 
for that enrollee to seek a tiering exception for the non-preferred 
drug. Also, CMS does not specify the classification system that must be 
used on Part D plan formularies; therefore, establishing a requirement 
that alternative drugs must be in the same therapeutic class would 
introduce inconsistency because what one plan considers the same drug 
class may be different than another plan for the same drugs. The 
changes to the tiering exception regulations that we are finalizing in 
this rule do not require plans to consider a drug for which the 
enrollee's condition is not a medically accepted indication to be an 
alternative drug for purposes of a tiering exception request. Because 
payment under Part D cannot be made for any drug that does not meet the 
definition of a Part D drug for the prescribed indication, such drug 
could not reasonably be considered an alternative drug for treatment of 
the enrollee's condition.
    In response to comments suggesting that our interpretation of ``for 
treatment of the same condition'' is inconsistent with the statute, we 
disagree. As we noted in the proposed rule, we interpret this language 
to refer to the condition as it affects the enrollee. Given the 
language in section 1860D-4(g)(2) of the Act states that an exception 
could be covered if the prescribing physician determines that the 
preferred drug would not be as effective ``for the individual'' or 
would have adverse effects ``for the individual,'' we believe it is 
appropriate to interpret the standard for the ``same condition'' to be 
referring to the individual.
    While we are not making any changes to the regulations with respect 
to defining alternative drugs, we wish to note that plan medical 
directors are required to be involved in the development and oversight 
of policies and procedures for processing exception requests, including 
criteria for determining alternative drugs, as part of their 
responsibility under Sec.  423.562(a)(5) to ensure the clinical 
accuracy of all coverage determinations and redeterminations involving 
medical necessity. Additionally, Sec.  423.566(d) requires that, before 
issuing an adverse coverage determination based on lack of medical 
necessity, including exception requests, it must be reviewed by a 
physician or appropriate health care professional. These policies 
requiring clinician involvement in the establishment and application of 
plan coverage rules contemplate that those individuals apply reasonable 
clinical judgment, based on sound medical and scientific evidence and 
acceptable standards of practice, in adjudicating exception requests, 
including consideration of alternative drugs on the plan's formulary.
    While we agree that in certain situations and with certain medical 
conditions, what is reasonably considered an alternative drug may be 
limited in ways suggested by commenters, we disagree that such 
designations should be codified in regulation to apply to all tiering 
exceptions for the reasons previously stated, and because we do not see 
a good reason to codify these types of clinical considerations only for 
tiering exceptions, when we have not proposed to do so for other types 
of coverage determinations. We also believe these clarifications 
provide sufficient guidance for plans to determine what drugs should be 
considered alternatives for treating the enrollee's condition, and will 
ensure that plans do not apply unreasonable clinical or policy 
standards to their interpretation of the meaning of alternative drug so 
as to inappropriately refuse to allow tiering exceptions. Therefore, we 
are not adding a definition of alternative drug in this final rule.
    As discussed earlier in this preamble, CMS will update any existing 
agency guidance related to tiering exceptions as needed to ensure that 
it comports with the requirements of this final rule.
    Comment: A commenter asked CMS to clarify whether a tiering 
exception should be approved when the requested drug is not being 
prescribed for a medically accepted indication, or does not otherwise 
meet the definition of a Part D drug.
    Response: Pursuant to the existing regulation at Sec.  423.578(e), 
which we did not propose to revise, enrollees are not permitted to use 
the exceptions process to obtain coverage for a drug that is not being 
prescribed to treat a medically accepted indication as defined in 
section 1860D-2(e)(4) of the Act, or does not otherwise meet the 
definition of a Part D drug at Sec.  423.100. Thus, a plan cannot 
approve a tiering exception request if the requested drug is not being 
used to treat a medically accepted indication or does not meet the 
definition of a Part D drug.
    After consideration of the comments received on this section, we 
are finalizing our proposal without modification, and have chosen not 
to further specify how to determine what an alternative drug for 
treating the enrollee's condition is.
e. Approval of Tiering Exception Requests
    We proposed to revise Sec.  423.578(c)(3) by renumbering the 
provision and adding a new paragraph (ii) to codify our current policy 
that cost sharing for an approved tiering exception request is assigned 
at the lowest applicable tier when preferred alternatives sit on 
multiple lower tiers. Under our proposal, assignment of cost sharing 
for an approved tiering exception must be at the most favorable cost-
sharing tier containing alternative drugs, unless such alternative 
drugs are not applicable pursuant to limitations set forth under 
proposed Sec.  423.578(a)(6).
    We received the following comments and our responses follow:
    Comment: We received several comments related to this aspect of our 
proposal. Commenters were divided, with some supporting our proposal 
and others opposed. Commenters in support of the proposal to require 
approval at the lowest applicable tier stated that this policy allows 
beneficiaries who cannot take less expensive drugs to obtain needed 
drugs at an affordable price. Some commenters noted that they supported 
this aspect of the proposal because we also proposed to allow plans to 
limit tiering exceptions for brand name drugs to the lowest tier 
containing alternative brand name drugs. A few commenters expressed a 
belief that this policy would be easy for beneficiaries to understand.
    Commenters who opposed our proposal stated that requiring approval 
to the lowest applicable tier interferes with plans' ability to manage 
their formularies. A few commenters expressed a belief that our 
proposal is not consistent with the statute, which states that the 
requested drug could be covered at terms applicable to preferred drugs 
but does not specify that it be the terms applicable to the most 
preferred alternatives. A commenter stated that Sec.  1860D-4(g)(2) 
does not specifically refer to a right to obtain a drug at the lowest 
cost-sharing tier. Another commenter stated that requiring plans to 
provide high cost drugs at the lowest tier instead of the next lower 
tier increases premiums for all beneficiaries and provides only 
slightly lower cost-sharing for a few individuals.

[[Page 16514]]

    Response: We thank commenters who were supportive of our proposal 
for their support. We agree that our policy of approval to the lowest 
applicable tier containing alternatives provides the most relief for 
beneficiaries with a medical need for a non-preferred drug.
    We disagree that our proposal is inconsistent with the statute. 
Section 1860D-4(g)(2) provides that if a plan sponsor uses formulary 
tiers and offers lower cost sharing for ``preferred drugs'' (plural) 
included in the formulary, an enrollee may request an exception to the 
tiered cost-sharing structure, and under such an exception, a non-
preferred drug could be covered ``under the terms applicable for 
preferred drugs'' (plural) if the prescriber determines that ``the 
preferred drug'' (singular) for the same condition would not be as 
effective or would have adverse effects, or both. The statute clearly 
contemplates that while there can be multiple drugs that are preferred 
drugs relative to the requested drug, and the prescribing physician can 
determine that ``the'' preferred drug would not be as effective or 
would have adverse effects. We believe it is reasonable to interpret 
this provision to permit an enrollee to seek a tiering exception under 
which he or she would pay the cost sharing applicable to the most 
preferred drug among one or more preferred drugs.
    After consideration of the comments received, we are finalizing 
without modification our proposal at Sec.  423.578(c)(3), which 
specifies that cost-sharing for approved tiering exceptions is assigned 
at the lowest applicable tier when preferred alternatives sit on 
multiple lower tiers.
f. Additional Technical Changes and Corrections
    Finally, we proposed various technical changes and corrections to 
improve the clarity of the tiering exceptions regulations and 
consistency with the regulations for formulary exceptions. 
Specifically, we proposed the following:
     Revise the introductory text of Sec.  423.578(a) to 
clarify that a ``requested'' non-preferred drug for treatment of an 
enrollee's health condition may be eligible for an exception.
     Revise Sec.  423.578(a)(1) to include ``tiering'' when 
referring to the exceptions procedures described in this subparagraph.
     Revise Sec.  423.578(a)(4) by making ``conditions'' 
singular and by adding ``(s)'' to ``drug'' to account for situations 
when there are multiple alternative drugs.
     Revise Sec.  423.578(a)(5) by removing the text specifying 
that the prescriber's supporting statement ``demonstrate the medical 
necessity of the drug'' to align with the existing language for 
formulary exceptions at Sec.  423.578(b)(6). The requirement that the 
supporting statement address the enrollee's medical need for the 
requested drug is already explained in the introductory text of Sec.  
423.578(a).
     Redesignate paragraphs Sec.  423.578(c)(3)(i) through 
(iii) as paragraphs Sec.  423.578(c)(3)(i)(A) through (C), 
respectively. This proposed change will improve consistency between the 
regulation text for tiering and formulary exceptions.
    We received no comments on the proposed technical changes and 
corrections and are finalizing them without modification.
    After consideration of all comments received on the tiering 
exceptions proposal, we are finalizing the proposed regulation text 
without modification. As discussed above, CMS will review agency 
guidance and beneficiary communications and revise as needed to be 
consistent with this final rule.
10. Establishing Limitations for the Part D Special Election Period 
(SEP) for Dually Eligible Beneficiaries (Sec.  423.38)
    As discussed in section II.A.1 of this final rule, the MMA added 
section 1860D-1(b)(3)(D) to the Act to establish a special election 
period (SEP) for full-benefit dual eligible (FBDE) beneficiaries under 
Part D. This SEP, codified at Sec.  423.38(c)(4), was later extended to 
all other subsidy-eligible beneficiaries by regulation (75 FR 19720). 
The SEP allows eligible beneficiaries to make Part D enrollment changes 
(that is, enroll in, disenroll from, or change Part D plans, including 
Medicare Advantage Prescription Drug (MA-PD) plans) once a month 
throughout the year, unlike other Part D enrollees who generally may 
switch plans only during the annual enrollment period (AEP) each fall.
    With over 10 years of programmatic experience, we have observed 
certain enrollment trends in terms of FBDE and other LIS beneficiaries:
     Most LIS beneficiaries do not make an active choice to 
join a PDP.
     Once in a plan, whether it was a CMS-initiated enrollment 
or a choice they made on their own, most LIS beneficiaries do not make 
changes during the year.
     A small subset (0.8 percent) of LIS beneficiaries use the 
SEP to actively enroll in a plan of their choice and then disenroll 
within 2 months.
    In addition, the application of the continuous SEP carries 
different service delivery implications for enrollees of MA-PD plans 
and related products than for standalone enrollees of PDPs. At the 
outset of the Part D program, when drug coverage for dually eligible 
beneficiaries was transitioned from Medicaid to Medicare, there were 
concerns about how CMS would effectively identify, educate, and enroll 
dually eligible beneficiaries. While processes (for example, auto-
enrollment, reassignment) were established to facilitate coverage, the 
continuous SEP served as a fail-safe to ensure that the beneficiary was 
always in a position to make a choice that best served their healthcare 
needs. Unintended consequences have resulted from this flexibility, 
including, as noted by the Medicare Payment Advisory Commission (MedPAC 
\30\), opportunities for marketing abuses.
---------------------------------------------------------------------------

    \30\ Medicare Payment Advisory Commission, ``Report to Congress: 
Medicare Payment Policy,'' March 2008.
---------------------------------------------------------------------------

    Among the key obstacles the continuous SEP (and resulting plan 
movement) can present are--
     Interfering with the coordination of care among the 
providers, health plans, and states;
     Hindering the ability for beneficiaries to benefit from 
case management and disease management;
     Inefficient use of the effort and resources needed to 
conduct enrollee needs assessments and developing plans of care for 
services covered by Medicare and Medicaid;
     Limiting a plan's opportunity for continuous coordinated 
treatment of chronic conditions; and
     Diminishing incentives for plans to innovate and invest in 
serving potentially high-cost members.
    To support plan sponsors' efforts to administer benefits to 
beneficiaries, including coordination of Medicare and Medicaid 
benefits, and maximize care management and positive health outcomes, we 
proposed to amend Sec.  423.38(c)(4) to make the SEP for FBDE and other 
subsidy-eligible individuals available only in certain circumstances. 
Specifically, we proposed to revise to Sec.  423.38(c) to specify that 
the SEP is available only as follows:
     In new paragraph (c)(4)(i), eligible beneficiaries (that 
is, those who are dual or other LIS-eligible and do not meet the 
definition of at-risk beneficiary or potential at-risk beneficiary 
under proposed Sec.  423.100) would be able to use the SEP once per 
calendar year.
     In new paragraph (c)(4)(iii), eligible beneficiaries who 
have been assigned to a plan by CMS or a State would be able to use the 
SEP before that election becomes effective (that is, opt out and

[[Page 16515]]

enroll in a different plan) or within 2 months of their enrollment in 
that plan.
     In new paragraph (c)(9), dual and other LIS-eligible 
beneficiaries who have a change in their Medicaid or LIS-eligible 
status would have an SEP to make an election within 2 months of the 
change, or of being notified of such change, whichever is later. This 
SEP would be available to beneficiaries who experience a change in 
Medicaid or LIS status regardless of whether they have been identified 
as potential at-risk beneficiaries or at-risk beneficiaries under 
proposed Sec.  423.100.
     In addition, we also proposed to remove the phrase ``at 
any time'' in the introductory language of Sec.  423.38(c) for the sake 
of clarity.
    We considered multiple alternatives related to the SEP proposal. In 
the proposed rule, we described and asked for comments on two 
alternatives:
    Limit of two or three uses of the SEP per year. We considered 
applying a simple numerical limit to the number of times the LIS SEP 
could be used by any beneficiary within each calendar year. We 
specifically considered limits of either two or three uses of the SEP 
per year.
    Limits on midyear MA-PD plan switching. We also considered an 
option that would prohibit SEP use into non-integrated MA-PD plans, but 
allow continuous use of the dual SEP to allow eligible beneficiaries to 
enroll into FIDE SNPs or comparably integrated products for dually 
eligible beneficiaries or standalone PDPs.
    We received the following comments and our responses follow:
    Comment: Some commenters supported the proposal and agreed that 
continuity of enrollment could maximize coordination of care and 
positive health outcomes. However, the majority of commenters opposed 
the proposal based on a variety of factors. Most of these commenters 
expressed concerns about the impact on the dual-eligible population 
which, they noted, not only has limited financial resources, but also 
higher rates of disability, higher rates of cognitive impairment, and 
lower health literacy. These circumstances, commenters noted, often 
contribute to more complex and changing health needs and difficulties 
with medication adherence. Citing these circumstances, many commenters 
believed these beneficiaries needed the flexibility to change their 
healthcare coverage at any time during the year.
    Commenters also believed that the proposal was too complex and 
would be difficult for beneficiaries to understand and for plans to 
administer. They noted that limited and, in some cases, multi-layered 
SEPs were unnecessary when the existing ongoing SEP has worked well and 
has proved to be simpler to communicate and understand.
    Many commenters also said that the proposal would have an even 
greater impact given the proposed changes related to midyear formulary 
changes. Commenters noted that since plans have the ability to change 
formularies or provider networks during the year, the ongoing dual SEP 
is a vital beneficiary protection.
    Lastly, commenters said that the proposed dual SEP limitation 
could, in actuality, hamper CMS' stated goal of bringing Medicare and 
Medicaid into better alignment because it could inadvertently 
discourage dual eligible beneficiaries from enrolling in integrated 
products. Commenters noted that because beneficiaries are often 
hesitant to change plans, they may opt to stay in their current plan 
instead of trying an integrated option. In other cases, commenters 
expressed concern that beneficiaries who are assigned into a plan by 
CMS or a State may panic and disenroll immediately if they believe 
pressured to make an immediate decision. Commenters said that the 
ongoing SEP gives beneficiaries the comfort and time to make a 
deliberate and educated choice.
    Response: We thank the commenters for their thoughtful feedback. We 
are mindful of the unique health care challenges that dual and other 
LIS-eligible beneficiaries may face. The goals of the proposal were to 
improve administration of benefits and coordination of care and we 
believed that this could best be accomplished through continuity of 
enrollment. While we acknowledge that many commenters prefer the 
ongoing nature of the existing dual SEP, we still believe that adopting 
some limitations is an appropriate step toward encouraging care 
coordination, achieving positive health outcomes, and discouraging 
extraneous beneficiary movement during the plan year.
    In response to comments, we are modifying our approach. In lieu of 
the proposed dual SEP limitation that would only allow a onetime use 
per year with certain exceptions, we are instead revising the dual SEP 
so that it is similar to the ``two or three uses per year'' alternative 
discussed in the proposed rule. Specifically, the dual SEP is being 
amended so that it can be used once per calendar quarter during the 
first nine months of the year (that is, one election during each of the 
following time periods: January-March, April-June, July-September). 
During the last quarter of the year, a beneficiary can use the AEP to 
make an election that would be effective on January 1. In addition to 
this change, the exception outlined at Sec.  423.38(c)(4)(ii) related 
to CMS and State-initiated elections will not be finalized as proposed. 
(Instead, as discussed below, CMS will be using its authority under 
Sec.  423.38(c)(8)(ii) to establish a coordinating SEP for those who 
are enrolled into a plan by CMS or a State at new Sec.  423.38(c)(10).
    We believe that limiting use of the dual SEP, but in a less 
restrictive manner, strikes the appropriate balance of our stated goals 
and the concerns raised by commenters, for the reasons that follow. We 
consider this approach to be less confusing for both plan sponsors and 
beneficiaries than our proposal because it provides a date-based 
parameter that is easier to comprehend without the additional layers of 
exceptions. By still allowing multiple changes throughout the year, 
dual and other LIS-eligible beneficiaries will maintain additional 
flexibilities not afforded to other Part D-eligible beneficiaries, but 
there may be times when these individuals cannot change plans and have 
that choice effective the next month either because they already made 
an election during that calendar quarter (during the first nine months 
of the year) or because they are making an election during the AEP. We 
believe that having certain periods when individuals must maintain 
enrollment in a particular plan will increase opportunities for 
coordination of care and case management. Even though these periods of 
required continuity of enrollment will be shorter than what was 
proposed, we believe it still matches our stated goals and addresses 
the concerns expressed by commenters.
    While we believe this limitation is an appropriate control to put 
in place, we also believe that it will not impact the vast majority of 
individuals eligible for the dual SEP. As discussed in the proposed 
rule, 2016 data demonstrated that most beneficiaries do not use the 
dual SEP and, of those who do use it, the majority (74.5 percent) only 
used it once. Analysis of 2017 data continues to show that 
beneficiaries who use the SEP use it only one time (85.5 percent). Of 
those who use it two times, the average time between elections is 3.4 
months, which is roughly the duration of a calendar quarter.
    Given this flexibility, we believe that dual and other LIS-eligible 
beneficiaries will have the freedom to choose a plan that works for 
their evolving health care needs during the year. For those that have 
an opportunity to enroll in an integrated product, they will be able to 
do so and know that if it does not suit their needs, they can choose 
another

[[Page 16516]]

plan in the near future. The same logic can be applied to those who 
want to explore other plan options during the year due to formulary, 
provider network, or health status changes. We note, though, that as 
discussed earlier, individuals who have been identified as an at-risk 
beneficiary or potential at-risk beneficiary under Sec.  423.100 will 
not be able to use the dual SEP. As discussed in section II.A.1, we are 
specifying at Sec.  423.38(c)(4) that this particular limitation 
applies once the beneficiary has been notified that he or she has been 
identified as a potential at-risk beneficiary or at-risk beneficiary, 
and the limitation will continue until such identification has been 
terminated consistent with Sec.  423.153(f).
    Comment: Many commenters recommended a wide range of modifications 
or alternatives to the dual SEP limitation outlined in the proposed 
rule. Suggestions included the following:
     Allow beneficiaries to disenroll to FFS at any time.
     Instead of limiting the use of the dual SEP, require a 
minimum enrollment duration in a plan.
     Limit to onetime use per year, without exceptions, to 
mitigate administrative burden.
     Delay any sort of SEP limitation and, instead, contemplate 
for future rulemaking.
    Some commenters--both those who supported and opposed the concept 
of a limitation to the dual SEP--expressed a preference for one of the 
two alternatives discussed in the proposed rule. There were some who 
supported the concept of expanding the onetime annual election to 2-3 
uses per year because it provided more flexibility. Some commenters 
expressed support for the more complex approach that would have allowed 
limited use of the dual SEP for enrollment in integrated products, 
standalone PDPs, and FFS, but not any non-integrated MA plans.
    Along these lines, there was varied feedback for dual SEP use for 
enrollment into integrated products. Some said that it should be 
allowed as a onetime exception, some said that it should be an ongoing 
opportunity, while others said that it should be the only allowable use 
of the dual SEP. A commenter encouraged CMS to work with States to 
define which plans would be considered ``integrated'' and another 
commenter suggested that CMS maintain and publicize a list of 
integrated plans.
    Response: We believe that the wide array of feedback that 
commenters provided on the proposal represents the complexity and 
varying interests of those who would be impacted by a change to the 
dual SEP. Given that the majority of commenters preferred more 
flexibility than what we proposed, we are opting to finalize a 
limitation that is along the lines of the ``two or three uses per 
year'' alternative described in the proposed rule.
    We contemplated allowing multiple uses per year at any time, but 
thought that an approach that allowed for quarterly elections (that is, 
the dual SEP in coordination with the AEP) was preferable because it 
would be easier to keep track of and for beneficiaries to understand. 
With a multiple-use-per-year-at-any-time policy, if a beneficiary makes 
several elections in the beginning of the year, as they approach the 
end of the year it may be hard to remember how many elections they have 
made or whether any more are available. With an approach that allows 
for quarterly elections, however, they only need to remember if they 
made an election in the last few months. If they have not, it is likely 
that they are eligible for a quarterly dual SEP use or the AEP. A 
quarterly approach also mitigates scenarios where a beneficiary makes 
multiple elections in the first half of the year and is then locked 
into a plan for the latter half of the year.
    Comment: In addition to the modifications/alternatives discussed 
above, a number of commenters believed that if limitations were 
established for the dual SEP, CMS should consider additional exceptions 
for certain beneficiary groups or conditions. Specifically, commenters 
believed exceptions would serve as important beneficiary protections 
for the following individuals/circumstances:
     Those who have a new or existing disability.
     Those with a new or altered disease state or diagnosis.
     American Indians and Alaska Natives who also receive 
services through the Indian Health Service.
     Enrollees whose prescription drugs are not covered under 
their plan's formulary or whose providers change during the year.
     Individuals whose caregiver arrangements change during the 
year.
     Individuals who must comply with Medicaid open enrollment 
periods or those who meet the ``for cause'' standards established for 
enrollees in Medicaid managed care plans.
     Those whose providers request an SEP on their behalf.
    Response: We believe that by allowing the dual SEP to be used 
quarterly during the first nine months of the year in conjunction with 
the AEP at the end of the year, we are mitigating the need for the 
exceptions suggested by the commenters. Dual or other LIS-eligible 
beneficiaries who fall into any of these categories would still be able 
to use the dual SEP. The only way that they may be limited is if they 
had already made a recent election into a plan. If that were the case, 
they may have to wait several months to make another change. (A more 
detailed discussion of different election periods and when they are 
considered ``used'' and effective can be found below.) Again, we do not 
see the frequency of movement that would lead us to believe that this 
will be an issue for the vast majority of LIS-eligible beneficiaries.
    We would note that in addition to the dual SEP, there are already a 
number of protections in place for all beneficiaries who have Part D 
coverage and are unable to change plans. For example, beneficiaries can 
request transition fills--prescription drugs that are not on a plan's 
formulary or that are on a plan's formulary but require prior 
authorization or step therapy under a plan's utilization management 
rules--during the first 90 days of enrollment in a new plan as provided 
under Sec.  423.120(b)(3). In addition, beneficiaries can request a 
formulary or tiering exception to obtain a drug that is not on their 
plan's formulary or to obtain a drug at a lower cost-sharing tier.
    While we understand that commenters believe that the ability to 
change plans at any time is an important beneficiary protection, we 
believe it is worth re-stating that the changes finalized at Sec.  
423.38(c)(4) will still provide for multiple uses of the dual SEP 
throughout the year and this is a flexibility that is not afforded to 
all Part D enrollees. During other parts of the year, dual and other 
LIS-eligible individuals will still have access to the AEP in the fall 
or, if applicable, the initial enrollment period (IEP) or the new MA 
open enrollment period (OEP) discussed in section II.B.1. Beneficiaries 
may also continue to be eligible for other SEPs outlined in Sec.  
422.62(b) and Sec.  423.38(c), which includes circumstances like a 
change or residence or other exceptional circumstances as determined by 
CMS.
    In addition, we will be finalizing the SEP opportunity that was 
contemplated in the proposed rule for beneficiaries assigned to a plan 
by CMS or a State. While this was proposed at new Sec.  
423.38(c)(4)(iii) as an additional use of the dual SEP, and would have 
been available before that election became effective or within 2 months 
of enrollment in the plan, we will be finalizing this as a new and 
separate

[[Page 16517]]

SEP at Sec.  423.38(c)(10). We believe that establishing this as a 
separate SEP is more straightforward because it makes clear that this 
opportunity is separate and in addition to the elections allowable 
under the revised dual SEP.
    This new SEP will allow individuals who have been auto-enrolled, 
facilitated enrolled, or reassigned into a plan by CMS, as well as 
those who have been subject to passive enrollment processes discussed 
in section II.A.8, an opportunity to change plans. Unlike the proposed 
SEP, this new SEP will be available even if a beneficiary meets the 
definition of an at-risk beneficiary or potential at-risk beneficiary. 
Beneficiaries would be able to use this new CMS/State assignment SEP 
before that enrollment becomes effective (that is, opt out and enroll 
in a different plan) or within 3 months of the assignment effective 
date, whichever is later. (Note that this SEP will not apply to 
individuals who have been subject to default enrollment processes 
discussed in section II.A.7, as they will be able to use the new Open 
Enrollment Period (OEP) to make an election.)
    Comment: A commenter requested a mechanism for plan sponsors to 
determine if the enrollment prior to the enrollee's SEP request was 
assigned by the CMS or the State. Another commenter requested 
clarification that States may make passive enrollment decisions where 
otherwise permitted, such as in Medicare-Medicaid Plans (MMPs), 
regardless of whether an individual has exhausted his or her SEP 
options for the year.
    Response: CMS is exploring possible mechanisms that would allow 
plan sponsors to determine if the enrollee's most recent enrollment 
transaction was one that was initiated by CMS or the State. In the 
interim, plan sponsors should ask the enrollee if they received a 
notice that indicates that they have been assigned to a plan and have 
certain SEP opportunities.
    If a beneficiary is assigned to a plan by CMS or a State, the 
enrollment change does not count against any of their SEP 
opportunities. That is, if a State passively enrolls a dual-eligible 
beneficiary in April, the beneficiary would still have their second 
quarter dual SEP, as well as the SEP associated specifically with the 
passive enrollment.
    Comment: Several commenters sought clarification on how the dual 
SEP limitation would affect and interact with other election periods. 
Commenters stated that it was unclear how the SEP changes in Sec.  
423.38 would relate to the AEP and OEP. A few commenters sought 
verification that the SEPs for Program of All-inclusive Care for the 
Elderly (PACE) eligible beneficiaries, institutionalized individuals, 
and enrollments into 5-star plans would be unaffected. A commenter 
requested clarification whether the once-per-year SEP falls outside of 
the AEP, or whether the SEP also applies during this same AEP 
timeframe.
    Response: As noted in the proposed rule and above, other election 
periods, including the AEP and the new OEP, are still available to 
eligible individuals. The established SEPs that allow beneficiaries to 
enroll in 5-star plans and PACE, as well as the SEP that allows 
elections for those who move into, reside in, or move out of an 
institution, are unaffected. If used, they would not count as use of 
the dual SEP. If the beneficiary is eligible for multiple election 
periods, plan sponsors (or other enrollment facilitators) may need to 
determine which election period the beneficiary would like to use, 
especially if the election periods would result in different enrollment 
effective dates. This is consistent with subregulatory guidance in 
Chapter 2 of the Medicare Managed Care Manual (section 30.6), Chapter 3 
of the Medicare Prescription Drug Manual (section 30.4), and current 
enrollment processing procedures for any enrollment request received 
when the individual is eligible for more than one election period.
    The dual SEP will be considered ``used'' based on the application 
date. If, for example, an election is made in March and effective in 
April, we would consider the beneficiary as having used their first 
quarter (Q1) dual SEP, even though coverage would not be effective 
until the second quarter of the calendar year. If a dual or other LIS-
eligible beneficiary makes an election during the AEP (October 15th 
through December 7th), coverage would be effective January 1.
    If, for example, a beneficiary is reassigned into a new plan in the 
fall for coverage effective January 1, they would be able to make an 
election under the AEP or the new CMS/State assignment SEP. If they opt 
out of the reassignment before it becomes effective and choose to stay 
in their current plan, this would be considered a cancellation and no 
election period is required.
    We recognize that when looking at all of the election periods and 
associated timeframes in whole, there are multiple opportunities both 
within this SEP and other election periods for an individual to make a 
choice that best meets their needs. We believe that enrollment is an 
individual-based exercise, and 1-800-MEDICARE, SHIPs, advocacy 
helplines, plans, and enrollment brokers, already have processes in 
place to work with individual beneficiaries and determine the election 
periods for which they may be eligible. Ultimately, as already outlined 
in Chapter 3 of the Prescription Drug Benefit Manual (section 30), it 
is the plan sponsor's responsibility to determine the enrollment period 
for each enrollment/disenrollment request. In some cases, plan sponsors 
may need to contact the beneficiary directly to confirm the election 
period.
    Table 2 summarizes the election periods discussed above and the 
suggested hierarchy of election periods (highest to lowest). Readers 
should note that it is not a comprehensive list of all election periods 
and does not negate a plan sponsor's responsibility to contact a 
beneficiary if they believe that multiple election periods may be 
available. More detailed information will be provided in subregulatory 
guidance.

                        Table 2--Election Periods
------------------------------------------------------------------------
                                                          Considered
         Election period               Available           ``Used''
------------------------------------------------------------------------
Part D IEP......................  Based on when       Upon effective
                                   first eligible      date.
                                   for Part D.
MA OEP (must meet OEP             Annually..........  Upon application
 requirements).                                        date.
SEP--5-Star plans...............  Ongoing...........  Available as long
                                                       as election is in
                                                       5-Star plan.
SEP--PACE.......................  Ongoing for         Available as long
                                   enrollment into     as election is in
                                   PACE; two month     PACE plan; upon
                                   window after        application date
                                   disenrollment       for election
                                   from PACE.          subsequent to
                                                       PACE
                                                       disenrollment.
SEP--Institutionalized..........  Ongoing if moving   Available while in
                                   into/residing in    facility; upon
                                   facility; two       application date
                                   month window        for election
                                   after moving out    subsequent to
                                   of facility.        moving out of
                                                       facility.

[[Page 16518]]

 
SEP--CMS/State Assignment.......  Within 3 months *   Upon application
                                   of assignment or    date.
                                   notification of
                                   assignment,
                                   whichever is
                                   later.
SEP--Change in Dual/LIS Status..  Within 3 months *   Upon application
                                   of status change    date.
                                   or notification
                                   of change,
                                   whichever is
                                   later.
Dual SEP........................  Ongoing--One use    Upon application
                                   per calendar        date.
                                   quarter during
                                   the first nine
                                   months of the
                                   year.
AEP.............................  Annually..........  Multiple elections
                                                       can be submitted
                                                       during AEP, last
                                                       rec'd will be
                                                       considered the
                                                       choice.
------------------------------------------------------------------------
* As discussed below, the finalized SEPs will allow for a 3-month
  opportunity to change plans, not the 2-month window noted in the
  proposed rule.

    Comment: A few commenters requested clarification on how plan 
sponsors would be able to determine if a beneficiary has used their 
allowable dual SEP election. Commenters asked whether this information 
would be available in MARx or as a batch enrollment query (BEQ). 
Commenters also asked who is responsible for validating the SEP and 
noted that beneficiaries may be frustrated if they are unaware that 
they have exhausted their allowable use of the dual SEP and their 
enrollment is denied. A commenter asked that plans not be penalized for 
rejections related to the dual SEP.
    Response: Plan sponsors continue to be responsible for determining 
the eligibility and enrollment period for enrollment/disenrollment 
requests. As noted earlier, plan sponsors and other enrollment 
facilitators may need to ask questions of the beneficiary to determine 
if they are eligible for the dual SEP or another election period. As a 
part of this process, we assume that beneficiaries are informed about 
the enrollment process and told that a submitted enrollment form does 
not always guarantee enrollment in a plan. Further, the enrollment 
module in MARx will be updated to no longer allow use of the dual SEP 
more than once per calendar quarter during the first nine months of the 
year. Enrollment transactions submitted for an individual who has 
already used their quarterly opportunity will be rejected, and sponsors 
would notify the individual of the denial, as they do today. While the 
commenter did not specify which penalties they wanted waived, as stated 
earlier, the vast majority of beneficiaries do not use the dual SEP 
multiple times, let alone within a 3-month period, so any rejected 
transactions should be minimal.
    Comment: A commenter asked that we confirm that the dual SEP 
applies to individuals considered full-benefit dual eligible 
beneficiaries under Sec.  423.773(c)(1).
    Response: The dual SEP, with the parameters established in this 
rule, is available for full benefit dual eligible individuals and other 
subsidy-eligible beneficiaries as defined at Sec.  423.772.
    Comment: A few commenters recommended that we modify the proposed 
SEP at Sec.  423.38(c)(9) to allow for a three-month or unlimited 
window post LIS-change, not a 2-month window. These commenters said 
that the outreach and education time can be lengthy and two months does 
not provide the beneficiary with enough time to make a fully-informed 
choice. In addition, a commenter requested that we clarify whether a 
change in co-pay level only is considered a change in LIS-eligible 
status and would prompt eligibility for the dual SEP. Another commenter 
asked how the change in status SEP would affect those going through the 
deeming process.
    Response: We appreciate this insight from commenters and believe 
that a three-month window should give the beneficiary adequate time to 
understand their coverage changes and determine if it is in their best 
interest to change plans. Accordingly, we are revising Sec.  
423.38(c)(9) to allow individuals to make an election within 3 months 
of a gain, loss, or change to Medicaid or LIS eligibility, or 
notification of such a change, whichever is later. A change in co-pay 
level, or any change, resulting from the deeming process, would be 
considered a change in LIS eligibility.
    As discussed previously, the SEP for dual/LIS status change is 
separate from the dual SEP. If, for example, a Medicare beneficiary 
becomes eligible for Medicaid during the year, they would be able to 
use the dual/LIS status change SEP to change plans. In addition, 
because they are now a dually-eligible beneficiary, they would also be 
able to make their allowable quarterly dual SEP election during the 
first nine months of the year.
    Comment: A commenter noted that the Medicaid managed care rule at 
42 CFR 438.56(c)(2)(i) includes a 90-day period for plan changes 
following enrollment, and that dual/LIS SEPs should align so as to 
avoid conflicts between Medicare and Medicaid rules.
    Response: We appreciate the identification of the potential 
conflict. We believe that because of the various election periods that 
are available, including the new SEPs that are being finalized in this 
rule, there should not be a coordination issue with Medicaid managed 
care rules. Specifically, a beneficiary can still use the dual SEP 
quarterly during the first nine months of the year, the new three-month 
SEP for change in Medicaid status, the new three-month CMS/State 
assignment SEP, and the AEP.
    Comments: A commenter recommended that if the proposal was 
finalized, CMS should allow beneficiaries the right to file an appeal 
to switch plans in instances where their Part D plan has made a 
material change (such as to its formulary or to its pharmacy network) 
during the plan year.
    Response: Enrollment decisions are not appealable and we do not 
believe it would be prudent to set up an enrollment appeals process at 
this time. Given that dual and other LIS-eligible beneficiaries will 
still be able to use the dual SEP on a quarterly basis during the first 
nine months of the year, we believe that there is a readily accessible 
remedy for this enrollment issue. The beneficiary will still be able to 
change plans, but in the event that they have already used up their 
dual SEP election, they may have to wait to make another change, unless 
they are eligible for one of the many other SEPs. Again, we expect this 
circumstance to be extremely rare.
    Comment: A few commenters recommended that in addition to MA and 
Part D plans, CMS apply the SEP limitations to Medicare-Medicaid Plans

[[Page 16519]]

(MMPs) as part of the Financial Alignment Initiative demonstration.
    Response: We clarify that under the Financial Alignment Initiative 
capitated model demonstrations, MA regulations--including those 
governing SEPs--apply to MMPs unless waived. As has been the case to 
date under the demonstrations, we will continue to use our 
demonstration authority to waive applicable MA regulatory requirements 
in three-way contracts as necessary, and in partnership with each 
state, to achieve each individual demonstration's objectives.
    Comment: A commenter requested clarification regarding the federal 
vs. state authority over the dual SEP.
    Response: Other than state laws relating to state licensure and 
plan solvency the standards established under Part D supersede any 
state law or regulation with respect to Part D plans.
    Comment: Many commenters provided valuable feedback related to our 
request for suggestions on how to educate the affected population and 
other stakeholders of changes to the dual SEP. Suggestions included the 
following:
     Development of more outreach materials, including non-
English materials.
     Direct notification to affected individuals.
     Increased resources for SHIPs.
     Coordination with the Administration for Community Living 
and State ombudsmen.
     Television advertisements.
     Educational opportunities sales agents, providers and 
community partners.
     Broader education about the dual SEP in general.
    Response: We appreciate the feedback provided by commenters and 
will keep these suggestions in mind as we proceed with implementation 
of the dual SEP limitation beginning in plan year 2019.
    Comment: A commenter recommended changes to Medicaid managed care 
disenrollment rules outlined at 42 CFR 438.56.
    Response: Medicaid disenrollment rules are outside the scope of 
proposals set forth in the proposed rule and, as such, will not be 
considered for this rulemaking.
    After review of the comments, and as discussed above, we are 
finalizing the proposed changes to Sec.  423.38 with the following 
modifications:
     Paragraph (c)(4) is revised to allow eligible 
beneficiaries (that is, those who are dual or other LIS-eligible) use 
of the dual SEP once per calendar quarter during the first nine months 
of the year. We are further specifying that the limitation applicable 
to at-risk beneficiaries and potential at-risk beneficiaries (as 
defined under Sec.  423.100 and discussed in section II.A.1) is 
effective upon notification of that status and ends upon termination of 
that status consistent with Sec.  423.153(f).
     New paragraph (c)(9), which provides dual and other LIS-
eligible beneficiaries who have a change in their Medicaid or LIS-
eligible status an SEP, is modified to allow a 3-month window to make a 
change.
     Proposed paragraph (c)(4)(iii) allowing eligible 
beneficiaries who have been assigned to a plan by CMS or a State use of 
the dual SEP before that election becomes effective or within 2 months 
of their enrollment in that plan will not be finalized. Instead, a new 
CMS/State assignment SEP is established at Sec.  423.38(c)(10) to allow 
individuals in a similar circumstance (that is, auto- or facilitated 
enrolled, reassigned, default or passively enrolled by CMS or a state) 
an opportunity to change plans upon notification or within 3 months of 
the assignment effective date, whichever is later.
    Further detail on the SEP changes will be provided in subregulatory 
guidance. As suggested by a commenter, we will monitor the impact of 
this change and consider future modifications if there is evidence that 
beneficiaries are being harmed.
11. Medicare Advantage and Part D Prescription Drug Plan Quality Rating 
System
a. Introduction
    We are committed to transforming the health care delivery system--
and the Medicare program--by putting a strong focus on person-centered 
care, in accordance with the CMS Quality Strategy, so each provider can 
direct their time and resources to each beneficiary and improve their 
outcomes. As part of this commitment, one of our most important 
strategic goals is to improve the quality of care for Medicare 
beneficiaries. The Part C and D Star Ratings support the efforts of CMS 
to improve the level of accountability for the care provided by health 
and drug plans, physicians, hospitals, and other Medicare providers. We 
currently publicly report the quality and performance of health and 
drug plans on the Medicare Plan Finder tool on www.medicare.gov in the 
form of summary and overall ratings for the contracts under which each 
MA plan (including MA-PD plans) and Part D plan is offered, with drill 
downs to ratings for domains, ratings for individual measures, and 
underlying performance data. We also post additional measures on the 
display page \31\ at www.cms.gov for informational purposes. The goals 
of the Star Ratings are to display quality information on Medicare Plan 
Finder to help beneficiaries, families, and caregivers make informed 
choices by being able to consider a plan's quality, cost, and coverage; 
to provide information for public accountability; to incentivize 
quality improvement; to provide information to oversee and monitor 
quality; and to accurately measure and calculate scores and stars to 
reflect true performance. In addition, CMS has made strides in 
recognizing the challenges of serving high risk, high needs populations 
while continuing the focus on improving health care for these important 
groups.
---------------------------------------------------------------------------

    \31\ http://go.cms.gov/partcanddstarratings (under the 
downloads).
---------------------------------------------------------------------------

    In this final rule, as part of the Administration's efforts to 
improve transparency, we are codifying the existing Star Ratings system 
for the MA and Part D programs with some changes. As noted later in 
this section in more detail, the changes we proposed and are finalizing 
include more clearly delineating the rules for adding, updating, and 
removing measures and modifying how we calculate Star Ratings for 
contracts that consolidate. As we explained in the proposed rule, 
codifying the Star Ratings methodology will provide plans with more 
stability to plan multi-year initiatives, because the rulemaking 
process will create a longer lead time for changes and MA organizations 
and Part D sponsors will know the measures several years in advance. We 
have received comments for the past several years from MA organizations 
and other stakeholders asking that CMS use Federal Register rulemaking 
for the Star Ratings system; we discuss in section II.A.11.c. of this 
final rule (regarding plans for the transition period before the 
codified rules are used) how section 1853(b) authorizes CMS to 
establish and annually modify the Star Ratings system using the Advance 
Notice and Rate Announcement process because the system is an integral 
part of the policies governing Part C payment. We believe this is an 
appropriate time to codify the methodology, because the rating system 
has been used for several years now and is relatively mature so there 
is less need for extensive changes every year; the smaller degree of 
flexibility in having codified regulations rather than using the 
process for adopting payment methodology changes may be appropriate. 
Further, by adopting and

[[Page 16520]]

codifying the rules that govern the Star Ratings system, we are 
demonstrating a commitment to transparency and predictability for the 
rules in the system so as to foster investment.
b. Background
    We originally acted upon our authority to disseminate information 
to beneficiaries as the basis for developing and publicly posting the 
5-star ratings system (sections 1851(d) and 1852(e) of the Act). The MA 
statute explicitly requires that information about plan quality and 
performance indicators be provided to beneficiaries to help them make 
informed plan choices. These data are to include disenrollment rates, 
enrollee satisfaction, health outcomes, and plan compliance with 
requirements.
    The Part D statute (at section 1860D-1(c)) imposes a parallel 
information dissemination requirement with respect to Part D plans, and 
refers specifically to comparative information on consumer satisfaction 
survey results as well as quality and plan performance indicators. Part 
D plans are also required by regulation (Sec.  423.156) to make 
Consumer Assessment of Healthcare Providers and Systems (CAHPS) survey 
data available to CMS and are required to submit pricing and 
prescription drug event data under statutes and regulations specific to 
those data. Regulations require plans to report on quality improvement 
and quality assurance and to provide data which CMS can use to help 
beneficiaries compare plans (Sec. Sec.  422.152 (b)(3) and 
423.153(c)(5)). In addition we may require plans to report statistics 
and other information in specific categories (Sec. Sec.  422.516 and 
423.514).
    Currently, for similar reasons of providing information to 
beneficiaries to assist them in plan enrollment decisions, we also 
review and rate section 1876 cost plans on many of the same measures 
and publish the results. We also proposed to continue to include 1876 
cost contracts in the MA and Part D Star Rating system to provide 
comparative information to Medicare beneficiaries making plan choices. 
We proposed specific text, to be codified at Sec.  417.472(k), 
requiring that 1876 cost contracts to agree to be rated under the 
quality rating system specified at subpart D of part 422. Cost 
contracts are also required by regulation (Sec.  417.472(j)) to make 
CAHPS survey data available to CMS. As is the case today, no Quality 
Bonus Payments (QBP) will be associated with the ratings for 1876 cost 
contracts.
    In line with Sec. Sec.  422.152 and 423.153, CMS uses the 
Healthcare Effectiveness Data and Information Set (HEDIS), Health 
Outcomes Survey (HOS), CAHPS data, Part C and D Reporting requirements 
and administrative data, and data from CMS contractors and oversight 
activities to measure quality and performance of contracts. We have 
been displaying plan quality information based on that and other data 
since 1998.
    Since 2007, we have published annual performance ratings for stand-
alone Medicare PDPs. In 2008, we introduced and displayed the Star 
Ratings for Medicare Advantage Organizations (MAOs) for both Part C 
only contracts (MA-only contracts) and Part C and D contracts (MA-PDs). 
Each year since 2008, we have released the MA Star Ratings. An overall 
rating combining health and drug plan measures was added in 2011, and 
differential weighting of measures (for example, outcomes being 
weighted 3 times the value of process measures) began in 2012. The 
measurement of year to year improvement began in 2013, and an 
adjustment (Categorical Adjustment Index) was introduced in 2017 to 
address the within-contract disparity in performance revealed in our 
research among beneficiaries that are dual eligible, receive a low 
income subsidy, and/or are disabled.
    The MA and Part D Star Ratings measure the quality of care and 
experiences of beneficiaries enrolled in MA and Part D contracts, with 
5 stars as the highest rating and 1 star as the lowest rating. The Star 
Ratings provide ratings at various levels of a hierarchical structure 
based on contract type, and all ratings are determined using the 
measure-level Star Ratings. Contingent on the contract type, ratings 
may be provided and include overall, summary (Part C and D), and domain 
Star Ratings. Information about the measures, the hierarchical 
structure of the ratings, and the methodology to generate the Star 
Ratings is detailed in the annually updated Medicare Part C and D Star 
Ratings Technical Notes, referred to as Technical Notes, available at 
http://go.cms.gov/partcanddstarratings.
    The MA and Part D Star Ratings system is designed to provide 
information to the beneficiary that is a true reflection of the plan's 
quality and encompasses multiple dimensions of high quality care. The 
information included in the ratings is selected based on its relevance 
and importance such that the ratings can meet the needs of 
beneficiaries using them to inform plan choice. While encouraging 
improved health outcomes of beneficiaries in an efficient, person 
centered, equitable, and high quality manner is one of the primary 
goals of the ratings, they also provide feedback on specific aspects of 
care and performance that directly impact outcomes, such as process 
measures and the beneficiary's perspective. The ratings focus on 
aspects of care and performance that are within the control of the 
health plan and can spur quality improvement. The data used in the 
ratings must be complete, accurate, reliable, and valid. A delicate 
balance exists between measuring numerous aspects of quality and the 
need for a small data set that minimizes reporting burden for the 
industry. Also, the beneficiary (or his or her representative) must 
have enough information to make an informed decision without feeling 
overwhelmed by the volume of data.
    The Patient Protection and Affordable Care Act (Pub. L. 111-148), 
as amended by the Healthcare and Education Reconciliation Act (Pub. L. 
111-152), provides for quality ratings, based on a 5-star rating system 
and the information collected under section 1852(e) of the Act, to be 
used in calculating payment to MA organizations beginning in 2012. 
Specifically, sections 1853(o) and 1854(b)(1)(C) of the Act were added 
and amended to provide, respectively, for an increase in the benchmark 
against which MA organizations bid and in the portion of the savings 
between the bid and benchmark available to the MA organization to use 
as a rebate. Under the Act, Part D plan sponsors are not eligible for 
quality based payments or rebates. We finalized a rule on April 15, 
2011 to implement these provisions and to use the existing Star Ratings 
system that had been in place since 2007 and 2008. (76 FR 21485-
21490).\32\ In addition, the Star Ratings measures are tied in many 
ways to responsibilities and obligations of MA organizations and Part D 
sponsors under their contracts with CMS. We believe that continued poor 
performance on the measures and overall and summary ratings indicates 
systemic and wide-spread problems in an MA plan or Part D plan. In 
April 2012, we finalized regulations to use consistently low summary 
Star Ratings--meaning 3 years of summary Star Ratings below 3 stars--as 
the basis for a contract termination for Part C and Part D plans. 
(Sec. Sec.  422.510(a)(14) and 423.509(a)(13)). Those regulations 
further reflect the role the Star Ratings have had in CMS' oversight, 
evaluation, and monitoring of MA and Part D plans to ensure compliance 
with the

[[Page 16521]]

respective program requirements and the provision of quality care and 
health coverage to Medicare beneficiaries.
---------------------------------------------------------------------------

    \32\ The ratings were first used as part of the QBP 
Demonstration for 2012 through 2014 and then used for payment 
purposes as specified in sections 1853(o) and 1854(b)(1)(C) of the 
Act and the regulation at 42 CFR 422.258(d)(7).
---------------------------------------------------------------------------

    The true potential of the use of the MA and Part D Star Ratings 
system to reach our goals and to serve as a catalyst for change can 
only be realized by working in tandem with our many stakeholders, 
including beneficiaries, plans, and advocates. The following guiding 
principles have been used historically in making enhancements and 
updates to the MA and Part D Star Ratings:
     Ratings align with the current CMS Quality Strategy.
     Measures developed by consensus-based organizations are 
used as much as possible.
     Ratings are a true reflection of plan quality and enrollee 
experience; the methodology minimizes risk of misclassification.
     Ratings are stable over time.
     Ratings treat contracts fairly and equally.
     Measures are selected to reflect the prevalence of 
conditions and the importance of health outcomes in the Medicare 
population.
     Data are complete, accurate, and reliable.
     Improvement on measures is under the control of the health 
or drug plan.
     Utility of ratings is considered for a wide range of 
purposes and goals.
    ++ Accountability to the public.
    ++ Enrollment choice for beneficiaries.
    ++ Driving quality improvement for plans and providers.
     Ratings minimize unintended consequences.
     Process of developing methodology is transparent and 
allows for multi-stakeholder input.
    We used these goals to guide our proposal and intend to use them to 
guide how we interpret and apply the final regulations. For each 
provision we proposed, we solicited comment on whether our specific 
proposed regulation text best serves these guiding principles. We also 
solicited comment on whether additional or other principles are better 
suited for these roles in measuring and communicating quality in the MA 
and Part D programs in a comparative manner.
    As we continue to consider making changes to the MA and Part D 
programs in order to increase plan participation and improve benefit 
offerings to enrollees, we also solicited feedback from stakeholders on 
how well the existing stars measures create meaningful quality 
improvement incentives and differentiate plans based on quality. We 
solicited comments on those topics, and have considered them in 
adopting this final rule, as noted in the responses below, and will 
consider them for future rulemaking. We specifically asked for feedback 
on the following topics:
     Additional opportunities to improve measures so that they 
further reflect the quality of health outcomes under the rated plans.
     Whether CMS' current process for establishing the cut 
points for Star Rating can be simplified, and if the relative 
performance as reflected by the existing methodology to establish cut 
points accurately reflects plan quality.
     How CMS should measure overall improvement across the Star 
Ratings measures. In the proposed rule, we specifically requested input 
on additional improvement adjustments that could be implemented, and 
the effect that these adjustment could have on new entrants (here 
meaning new MA organizations and/or new plans offered by existing MA 
organizations).
     Additional adjustments to the Star Ratings measures or 
methodology that could further account for unique geographic and 
provider market characteristics that affect performance (for example, 
rural geographies or monopolistic provider geographies), and the 
operational difficulties that plans could experience if such 
adjustments were adopted.
     In order to further encourage plan participation and new 
market entrants, whether CMS should consider implementing a 
demonstration to test alternative approaches for putting new entrants 
(that is, new MA organizations) on a level playing field with renewing 
plans from a Star Ratings perspective for a pre-determined period of 
time.
     Adding measures that evaluate quality from the perspective 
of adopting new technology (for example, the percent of beneficiaries 
enrolled through online brokers or increasing implementation of the use 
of telemedicine) or improving the ease, simplicity, and satisfaction of 
the beneficiary experience in a plan.
     Including survey measures of physicians' experiences. 
(Currently, we measure beneficiaries' experiences with their health and 
drug plans through the CAHPS survey.) Physicians also interact with 
health and drug plans on a daily basis on behalf of their patients. We 
noted in the proposed rule that we are considering developing a survey 
tool for collecting standardized information on physicians' experiences 
with health and drug plans and their services.
    CMS appreciates the feedback we received on our proposals and on 
the solicitations for comment on the various topics. In the sections 
that follow, which are arranged by topic area, we summarize the 
comments we received on the background section and policies, proposals 
and solicitations summarized there and provide our responses to the 
comments. (In each section in II.B.11.c through w, we summarize the 
proposals from the corresponding section of the proposed rule, the 
applicable comments, and our responses.)
    Comment: Most commenters supported both the principles and the 
decision to codify the methodology for the Part C and D Star Ratings. 
Of the commenters who supported those aspects of our overall proposal, 
a few suggested adding principles, such as the measure data should be 
timely and that distinctions between measure-level Star Ratings (cut 
points) should be meaningful.
    Response: CMS appreciates the support to codify the methodology for 
the Part C and D Star Ratings. We will codify the methodology in this 
final rule as outlined in this preamble, and will consider the 
additional principles raised by the commenters for adoption in the 
future as we continue to refine the principles in consultation with 
experts and stakeholders through the regulatory process.
    Comment: Several commenters requested CMS to continue updating the 
methodology though the Call Letter instead though regulation. 
Commenters were concerned that the regulatory process would lead to CMS 
not being able to act quickly when there are public health or patient 
safety concerns or when treatment guidelines are changed. Commenters 
also cited other concerns, including introducing a burdensome 
regulatory process that delays the implementation of essential measures 
which can improve the quality of care for patients with chronic 
illness, as reasons to not to finalize this proposal but to continue 
using the Call Letter process to modify the Star Ratings methodology. 
They also noted that there are already multiple opportunities for 
comment on new measures; thus, the regulatory process does not create 
additional transparency. A few commenters supported the general effort 
to put the Star Ratings principles and process into regulation, but 
encouraged CMS to adopt a few exceptions (such as allowing new measures 
(but not measures with substantive changes) to enter Star Ratings 
through the Call Letter process).
    Response: CMS understands the commenters' concerns about how the 
regulatory process may, in some cases, prevent CMS from quickly 
changing or adopting measures. However, given the

[[Page 16522]]

level of support for the proposal and the need to provide the industry 
with longer lead times for new measures, we will finalize the proposal 
to implement substantive changes through regulation and use the Call 
Letter to make non-substantive changes, suggest and solicit feedback on 
new measures that will be proposed in regulation, and address emergent 
public health or patient safety concerns by retiring existing measures 
as needed or introducing new measures for the display page that will be 
proposed for Star Ratings as appropriate. We also address comments on 
our proposals related to the type of updates and changes that we 
proposed to adopt without rulemaking, pursuant to specific rules 
proposed for Sec. Sec.  422.164 and 423.184, in section II.A.11.h.
    Comment: A commenter requested that measure changes take 3 years to 
implement in the Star Ratings and that five years should elapse before 
those changes could impact payment.
    Response: We thank the commenter for the suggestion, but are 
finalizing the timeframes proposed in the proposed rule because the 
majority of commenters supported the proposed timeframes. Some of the 
commenters did raise concerns about extending the timeframes for 
implementing and updating measures. Changing the timeframes for 
measures updates to at least 3 years will significantly slow the 
implementation of substantive and non-substantive changes, in 
particular, when the changes are non-substantive.
    Comment: A commenter encouraged CMS to adopt financial incentives 
for stand-alone prescription drug plans based on Part D Star Ratings.
    Response: CMS thanks the commenter for the suggestion, but CMS 
cannot adopt such financial incentives without statutory authority. The 
Quality Bonus Payment (QBP) program for MA plans is statutory and the 
statute does not allow CMS to pay QBPs to stand-alone prescription drug 
plans.
    Comment: We solicited comments on potentially adding measures in 
the future that evaluate quality from the perspective of adopting new 
technology. Many commenters supported adding a measure related to the 
use of technology, but multiple commenters cautioned that CMS rely on 
and use evidence that technology impacts health outcomes or improves 
the experiences of beneficiaries in order to adopt specific measures of 
that type. A number of commenters cautioned CMS to move carefully and 
slowly on promoting technology due to the potential for unintended 
consequences. A few commenters did not support measuring the adoption 
of technology, because such adoption may not always be in the best 
interest of the patient or enrollee. A few commenters did not support 
such measurement because adoption of technology is hard to measure well 
and may not lead to greater member satisfaction or correlate with other 
measures of plan performance. Those commenters discouraged such a 
focus, believing that beneficiaries will vary in their interest in 
whether plans and providers adopt new technologies, so measures of such 
adoption many not inform plan choice. A few commenters also feared that 
measures of adoption of technology may end up reflecting geographic 
differences and the socioeconomic status of members enrolled in the 
plan rather than the quality or performance of the plan itself. With 
respect to CMS' proposal to possibly add new measures that address the 
issue of new technology in the future, such as telemedicine, a 
commenter pointed out that ``Use of new technologies'' is not clearly 
defined and can span a number of technologies implemented across plans 
but not in a uniform manner or across all service areas. A commenter 
recommended that CMS continue to look at the incorporation of new 
technologies into Star Ratings measures but withhold any proposals for 
CY 2019 and CY 2020 until more formal proposals can be put forth for 
notice and comment prior to adoption. A commenter specifically urged 
measures of e-prescribing and e-prior authorization in Star Ratings. 
Another commenter urged CMS to explicitly capture in CAHPS composites 
(that is, the combination of two or more survey items into a measure) 
the use of telemedicine, as current survey wording may not do so.
    Response: CMS appreciates comments received on adding measures that 
evaluate quality from the perspective of adopting new technology and 
will continue to monitor developments in this area for future 
consideration. Although we are not finalizing the adoption of such a 
measure in this rule, we will continue to investigate how best to 
address incorporating new technologies into the Star Ratings measures. 
We note that for HEDIS 2019, NCQA is examining the addition of 
telehealth services in existing HEDIS measures where appropriate. 
NCQA's proposed method would use specific codes and code modifiers to 
clearly define which telehealth services would be allowed for each 
specific measure. Proposed changes to incorporate telehealth services 
will be posted for the HEDIS 2019 public comment period in February. We 
appreciate receiving the comment about telemedicine and CAHPS; we 
recognize telemedicine is an evolving area and may propose changes to 
CAHPS survey questions in the future after discussions with the Agency 
for Healthcare Research and Quality.
    Comment: A commenter specifically requested CMS provide certified 
software for measures not developed by external stewards, such as the 
Medication Therapy Management (MTM) and SNP Care Management measures.
    Response: These measures are based on data reported to CMS through 
the Part C and D Reporting Requirements. CMS is not clear how providing 
certified software for these measures will facilitate the submission of 
these measures. CMS also notes that the MTM measure is developed by an 
external steward (PQA).
    Comments: Many commenters indicated the need for greater alignment 
with providers (physicians, hospitals, medical groups, accountable care 
organizations, and plans) to make the quality measures more consistent, 
both to reduce burden and duplication and to more effectively 
incentivize behavior. For example, a few commenters urged use of 
measures aligned with the Merit-based Incentive Payment System (MIPS) 
program.
    Response: CMS thanks the multiple commenters for these suggestions 
and appreciate the concern about burden and duplication, as well as the 
potential value of consistently reinforcing the same message. CMS is 
continuing to work with measure developers to increase consistency in 
measurement across settings.
    Comment: Several commenters encouraged CMS to develop measures 
related to how well the care that is received by beneficiaries reflects 
the beneficiaries' concerns, values, and goals.
    Response: CMS is tracking work by measure developers in this area 
and thanks the commenters for the suggestion.
    Comment: Many commenters supported CMS continuing to develop and 
implement new measure concepts beyond those in current or currently 
anticipated measure sets. Among the most common suggestions were 
outcome measures, especially new patient-reported outcome measures, 
quality of life, and functional status measures (including Healthy Days 
at Home). Several commenters also encouraged measuring care for cancer, 
prevention of diabetes and other chronic conditions, long-term 
management of chronic obstructive pulmonary disease (COPD), as well as 
advanced care

[[Page 16523]]

planning, advanced directives and palliative care. A few other 
commenters highlighted concerns about measure gaps, such as for pain 
management, autoimmune disorders, mental illness, dementia/cognitive 
impairment, anticoagulation drug safety, and measures specific to 
patients with multiple co-morbidities, especially co-morbid diabetes 
and cardiovascular disease. A few commenters referred to NQF-endorsed 
measures used in other programs, such as change in functional status 
after spine or hip replacement surgery. A commenter encouraged CMS to 
utilize a comprehensive measure of adult vaccination, while another 
encouraged adoption of a vaccine cost-sharing measure. A commenter 
urged CMS to develop more medication adherence and appropriate use 
measures and to assign a high weight in the Star Ratings program. 
Another commenter suggested that any future transition of care measures 
include detailed information on all drug therapies prescribed and 
broader sharing of discharge information.
    In addition, a few commenters urged CMS to provide quality and 
performance information about physicians within plans or to measure 
plans on the engagement of their network of physicians in value-based 
purchasing designs (that is, payment designs that reward or increase 
payments based on quality or capitated payments to physicians/
practitioners, medical groups and ACOs).
    Several comments highlighted promoting and measuring network 
adequacy and potential delays in care or medication related to this, 
and a few encouraged CMS to reward plans that maintain adequate 
networks with increased Star Ratings. A number of commenters urged CMS 
to measure access to medical specialists and subspecialists, such as 
Mohs surgeons, cataract surgeons, and ophthalmologists, while a couple 
of commenters supported the assessment of pharmacy networks broken down 
by specialty drug access. The two comments about networks of physician 
and surgeon specialists urged CMS to leverage extant measurement with 
the MIPS and Quality Payment Program (QPP) to also help measure plan 
network adequacy. A commenter urged CMS to look beyond simple numbers 
of physicians and specialists, since contracting and affiliation in 
medical groups and ACOs may effectively limit the access patients have 
to the full network.
    Response: CMS appreciates the breadth of suggestions for new 
measures and will take these under consideration, including internal 
discussion and sharing them with the measure developers. We will also 
study the value and feasibility of deriving additional metrics (such as 
additional patient-reported outcome measures) from existing data 
collection efforts, like HOS.
    Comment: Several commenters urged the development of geographic 
and/or provider market characteristic adjusters in order to normalize 
variations outside plans' control. Some stated such adjustments would 
specifically prevent measure bias against state-contracted SNPs.
    Response: CMS appreciates this comment and will take it into 
consideration. As we consider adjustments to the Star Ratings measures, 
we need to ensure that the adjustments do not mask true differences in 
the quality of care across the country.
    Comment: A few commenters requested information about a Star 
Ratings policy for natural disasters.
    Response: CMS provided a detailed proposal concerning treatment of 
Star Ratings measures for contracts affected by disasters in the 2019 
draft Call Letter that would apply to the 2019 and 2020 Star Ratings. 
We plan to propose codifying this policy through future rulemaking for 
performance periods after 2019 and ratings after the 2021 Star Ratings.
    Comment: Several commenters questioned whether the Star Ratings 
regulations apply to PACE organizations.
    Response: The MA Star Ratings regulations do not apply to PACE 
organizations but to the extent that a PACE organization offers a plan 
including qualified prescription drug coverage, it is a Part D sponsor 
and therefore subject to the Part D regulations. This would include the 
Part D Star Ratings regulations adopted in this final rule as 42 CFR 
423.182-423.186. We have not produced Star Ratings for PACE 
organizations to date and are exploring the PACE waiver authority to 
continue to exclude PACE organizations from this requirement.
    Comment: Several commenters made suggestions for possible Medicare 
Plan Finder enhancements, including adding the capability to compare 
plans by population type as well as mobile enhancements. A commenter 
suggested including the overall Star Ratings in the Medicare & You 
handbook.
    Response: We appreciate these comments, but believe they are 
outside the scope of the proposed rule. However, we note that CMS is 
currently exploring options for improving the Plan Finder experience 
for Medicare beneficiaries, and that, although the timelines for 
publishing the Medicare & You handbook do not allow for including the 
overall Star Rating in the initial release that occurs in the fall, the 
overall Star Ratings are included in updated versions of the handbook 
that are released after the initial release and publication.
    Comment: We received one comment that PBMs and Part D plan sponsors 
have delegated their responsibilities for the Star Ratings program to 
network pharmacies without providing the pharmacies with additional 
compensation.
    Response: CMS appreciates these comments, but due to the non-
interference clause, CMS is prevented from interfering in contract 
arrangements between sponsors, pharmacies and other providers. CMS has 
indicated to measure stewards and other stakeholders that if such 
pharmacy performance metrics are used as a condition of pharmacy 
network status, measure specifications should be appropriately scaled, 
for example, ensure adequate sample size, and that incentives to 
achieve performance should be appropriately allocated.
    Comment: We received several comments recommending beneficiaries 
designated for lock-in be excluded from certain Star Ratings measures.
    Response: Thank you for the comment. Our Star Ratings proposal did 
not address this topic, and we plan to take these comments under 
advisement. For more information about CARA, please see section B.
    Comment: CMS had solicited feedback on the potential development of 
a physician survey to gather information for Star Ratings measures. The 
majority of commenters opposed the development of a physician survey 
due to the increased financial and administrative burden it would 
entail for both plans and health care providers/physicians who would be 
surveyed. Other commenters raised concerns about the ability of 
physicians to differentiate across plans when physicians interact with 
multiple plans. Multiple commenters were concerned that a physician 
could not accurately complete a survey on this topic since physicians 
often do not personally know the plan in which a beneficiary is 
enrolled. Some commenters noted that it may be difficult to determine 
who within a provider's practice should complete the survey. Other 
concerns raised include small sample sizes, subjectivity of responses, 
and potential for incomplete data.
    Response: CMS appreciates the input provided by commenters 
regarding the

[[Page 16524]]

burden and multiple challenges in developing a survey to evaluate 
physician experience interacting with both Medicare health and drug 
plans. We are not finalizing any aspect of the physician survey in this 
rule, but will take these comments into consideration as we continue to 
explore the feasibility and the value to the Star Ratings program in 
collecting feedback through a physician survey.
    Comment: A handful of commenters were concerned about the 
administration of a physician survey in integrated plans where the 
physician is employed by the plan which may bias the survey results.
    Response: We acknowledge that responses may not be unbiased in 
situations when the physician is employed by the plan. CMS will take 
this into account as we consider whether to develop a physician/
clinician survey in the future.
    Comment: Among the commenters supporting the development of a 
physician survey, commenters noted that the physician is in close 
contact with plans on behalf of their patients so this would complement 
the existing CAHPS survey for enrollees. A couple of commenters noted 
that a physician survey would be a way to measure network adequacy, 
appeals, benefit limit exceptions, and grievances. A few commenters 
recommended that CMS consider a broader survey of clinician 
experiences, including nurses, therapists, care coordinators, and 
pharmacists from a variety of settings. A commenter requested that a 
physician survey be voluntary.
    Response: CMS appreciates the support for the development of a 
physician survey and will solicit feedback from the industry on 
additional topics to be included on the survey if we move forward with 
the development in the future. We believe obtaining feedback from 
physicians is important; however, we will consider all of the comments 
provided before we make a determination about proceeding with 
developmental work.
    Comment: A commenter suggested the development of a general 
physician survey regarding experiences with managed care compared to 
fee-for-service to understand the larger healthcare landscape, while 
another commenter suggested obtaining feedback through other avenues 
outside of the Star Ratings program.
    Response: CMS appreciates these suggestions but they are out of 
scope for the potential development of surveys for the purpose of Star 
Ratings.
    We specifically address adoption of the Star Ratings System 
regulations for the MA and Part D programs in sections II.B.11.c 
through w.
c. Basis, Purpose and Applicability of the Medicare Advantage and 
Prescription Drug Plan Quality Rating System
    We proposed to codify regulation text, at Sec. Sec.  422.160 and 
423.180, that identifies the statutory authority, purpose, and 
applicability of the Star Ratings system regulations that we proposed 
to add under part 422 subpart D and part 423 subpart D. Under our 
proposal, we are continuing to apply the existing purposes of the 
quality rating system, which are to provide comparative information to 
Medicare beneficiaries pursuant to sections 1851(d) and 1860D-1(c) of 
the Act, identify and apply the payment consequences for MA plans under 
sections 1853(o) and 1854(b)(1)(C) of the Act, and evaluate and oversee 
overall and specific performance by MA and Part D plans. To reflect how 
the Part D ratings are used for MA-PD plan QBP status and rebate 
retention allowances, we also proposed specific text, to be codified at 
Sec.  423.180(b)(2), noting that the Part D Star Rating will be used 
for those purposes.
    We proposed, broadly stated, to codify the current quality Star 
Ratings system uses, methodology, measures, and data collection 
beginning with the measurement periods in calendar year 2019. We 
proposed some changes, such as how we handle consolidations from the 
current Star Ratings program, but overall the proposal was to continue 
the Star Ratings system as it has been developed and has stabilized. 
Under the proposal, data would be collected and performance measured 
using these proposed rules and regulations for the 2019 measurement 
period; the associated quality Star Ratings will be used to assign QBP 
ratings for the 2022 payment year and released prior to the annual 
election period held in late 2020 for the 2021 contract year. Because 
of the timing of the release and use in conjunction with the annual 
coordinated election period, these would be the ``2021 Star Ratings.''
    We proposed that the current quality Star Ratings system and 
procedures for revising it remain in place for the 2019 and 2020 Star 
Ratings. Section 1853(b) of the Act authorizes an advance notice and 
rate announcement to announce and solicit comment for proposed changes 
to the MA payment methodology, which CMS has interpreted to include the 
Part C and D Star Ratings program because of the payment consequences 
of Star Ratings under section 1853(o) of the Act. The statute 
identifies specific notice and comment timeframes, but that process 
does not require publication in the Federal Register. We have used the 
draft and final Call Letter, which are attachments to the Advance 
Notice and final Rate Announcement respectively,\33\ to propose for 
comment and finalize changes to the quality Star Ratings system since 
the ratings became a component of the payment methodology for MA and 
MA-PD plans. (76 FR 21487 through 89). Because the Star Ratings system 
has been integrated into the payment methodology since the 2012 
contract year (as a mechanism used to determine how much a plan is 
paid, and not the mechanism by which [or a rule about when] a plan is 
paid), the Star Ratings are part of the process for setting benchmarks 
and capitation rates under section 1853 of the Act, and the process for 
announcing changes to the Star Ratings system falls within the scope of 
section 1853(b) of the Act. Although not expressly required by section 
1853(b) of the Act, CMS has historically solicited comment on 
significant changes to the ratings system using a Request for Comment 
process before the Advance Notice and draft Call Letter are released; 
this Request for Comment \34\ provides MAOs, Part D sponsors, and other 
stakeholders an opportunity to request changes to and raise concerns 
about the Star Ratings methodology and measures before CMS finalizes 
its proposal for the Advance Notice. We intend to continue the current 
process at least until the 2019 measurement period that we proposed as 
the first measurement period under these new regulations, but we may 
discontinue that process at a later date as the Advance Notice/Call 
Letter process and rulemaking process may provide sufficient 
opportunity for public input. In addition, CMS issues annually the 
Technical Notes \35\ that describe in detail how the methodology is 
applied from the changes in policy adopted through the Advance Notice 
and Rate Announcement process. We intend to continue the practice of 
publishing the Technical Notes during the preview periods. Our proposed 
rule included continued use of the draft and final Call Letters as a 
means to provide subregulatory application),

[[Page 16525]]

interpretation, and guidance of the final version of these proposed 
regulations where necessary. Our proposed regulation text does not 
detail these plans for the RFC and Technical Notes because we believe 
such regulation text will be unnecessary. We proposed to codify the 
first performance period (2019) and first payment year (2022) to which 
our proposed regulations will apply at Sec.  422.160(c) and Sec.  
423.180(c).
---------------------------------------------------------------------------

    \33\ Advance Notices and Rate Announcements are posted each year 
on the CMS website at: https://www.cms.gov/Medicare/Health-Plans/MedicareAdvtgSpecRateStats/Announcements-and-Documents.html.
    \34\ Requests for Comment are posted at http://go.cms.gov/partcanddstarratings under the downloads.
    \35\ http://go.cms.gov/partcanddstarratings (under the 
downloads) for the Technical Notes.
---------------------------------------------------------------------------

    We received no comments on our proposed basis, purpose, and 
applicability regulations. For the reasons outlined in the proposed 
rule and summarized here, we are finalizing the regulation text 
proposed at Sec. Sec.  422.160 and 423.180 with one significant 
modification regarding the applicability of the regulations governing 
the Star Ratings of a surviving contract in a contract consolidation. 
In light of the passage of section 53112 of the Bipartisan Budget Act 
of 2018 (Pub. L. 115-123), the consolidation policy described at 
Sec. Sec.  422.162(b)(3) and 423.182(b)(3) will be implemented for the 
2020 QBP ratings and 2020 Star Ratings. We will finalize additional 
text at Sec. Sec.  422.160(c), 422.162(b)(3)(v), 423.180(c) and 
423.182(b)(3)(iii) to apply the regulations that govern the calculation 
of Star Ratings for surviving contracts when the contract consolidation 
is approved on or after January 1, 2019, consistent with the ACCESS Act 
provision.
d. Definitions
    We proposed the following definitions for the respective subparts 
in part 422 and part 423 in paragraph (a) of Sec. Sec.  422.162 and 
423.182 respectively.
     CAHPS refers to a comprehensive and evolving family of 
surveys that ask consumers and patients to evaluate the interpersonal 
aspects of health care. CAHPS surveys probe those aspects of care for 
which consumers and patients are the best or only source of 
information, as well as those that consumers and patients have 
identified as being important. CAHPS initially stood for the Consumer 
Assessment of Health Plans Study, but as the products have evolved 
beyond health plans the acronym now stands for Consumer Assessment of 
Healthcare Providers and Systems.
     Case-mix adjustment means an adjustment to the measure 
score made prior to the score being converted into a Star Rating to 
take into account certain enrollee characteristics that are not under 
the control of the plan. For example age, education, chronic medical 
conditions, and functional health status that may be related to the 
enrollee's survey responses.
     Categorical Adjustment Index (CAI) means the factor that 
is added to or subtracted from an overall or summary Star Rating (or 
both) to adjust for the average within-contract (or within-plan as 
applicable) disparity in performance associated with the percentages of 
beneficiaries who are dually eligible for Medicare and enrolled in 
Medicaid, beneficiaries who receive a Low Income Subsidy or have 
disability status in that contract (or plan as applicable).
     Clustering refers to a variety of techniques used to 
partition data into distinct groups such that the observations within a 
group are as similar as possible to each other, and as dissimilar as 
possible to observations in any other group. Clustering of the measure-
specific scores means that gaps that exist within the distribution of 
the scores are identified to create groups (clusters) that are then 
used to identify the four cut points resulting in the creation of five 
levels (one for each Star Rating), such that the scores in the same 
Star Rating level are as similar as possible and the scores in 
different Star Rating levels are as different as possible. Technically, 
the variance in measure scores is separated into within-cluster and 
between-cluster sum of squares components. The clusters reflect the 
groupings of numeric value scores that minimize the variance of scores 
within the clusters. The Star Ratings levels are assigned to the 
clusters that minimize the within-cluster sum of squares. The cut 
points for star assignments are derived from the range of measure 
scores per cluster, and the star levels associated with each cluster 
are determined by ordering the means of the clusters.
     Consolidation means when an MA organization/Part D sponsor 
that has at least two contracts for health and/or drug services of the 
same plan type under the same parent organization in a year combines 
multiple contracts into a single contract for the start of the 
subsequent contract year.
     Consumed contract means a contract that will no longer 
exist after a contract year's end as a result of a consolidation.
     Display page means the CMS website on which certain 
measures and scores are publicly available for informational purposes; 
the measures that are presented on the display page are not used in 
assigning Part C and D Star Ratings.
     Domain rating means the rating that groups measures 
together by dimensions of care.
     Dual Eligible (DE) means a beneficiary who is enrolled in 
both Medicare and Medicaid.
     HEDIS is the Healthcare Effectiveness Data and Information 
Set which is a widely used set of performance measures in the managed 
care industry, developed and maintained by the National Committee for 
Quality Assurance (NCQA). HEDIS data include clinical measures 
assessing the effectiveness of care, access/availability measures, and 
service use measures.
     Highest rating means the overall rating for MA-PDs, the 
Part C summary rating for MA-only contracts, and the Part D summary 
rating for PDPs.
     Highly-rated contract means a contract that has 4 or more 
stars for their highest rating when calculated without the improvement 
measures and with all applicable adjustments (CAI and the reward 
factor).
     HOS means the Medicare Health Outcomes Survey which is the 
first patient reported outcomes measure that was used in Medicare 
managed care. The goal of the Medicare HOS program is to gather valid, 
reliable, and clinically meaningful health status data in the Medicare 
Advantage (MA) program for use in quality improvement activities, pay 
for performance, program oversight, public reporting, and improving 
health. All managed care organizations with MA contracts must 
participate.
     Low Income Subsidy (LIS) means the subsidy that a 
beneficiary receives to help pay for prescription drug coverage (see 
Sec.  423.34 for definition of a low-income subsidy eligible 
individual).
     Measurement period means the period for which data are 
collected for a measure or the performance period that a measures 
covers.
     Measure score means the numeric value of the measure or an 
assigned `missing data' message.
     Measure star means the measure's numeric value is 
converted to a Star Rating. It is displayed to the nearest whole star, 
using a 1-5 star scale.
     Overall Rating means a global rating that summarizes the 
quality and performance for the types of services offered across all 
unique Part C and Part D measures.
     Part C Summary Rating means a global rating that 
summarizes the health plan quality and performance on Part C measures.
     Part D Summary Rating means a global rating of the 
prescription drug plan quality and performance on Part D measures.
     Plan Benefit Package (PBP) means a set of benefits for a 
defined MA or PDP service area. The PBP is submitted by PDP sponsors 
and MA organizations to CMS for benefit analysis, bidding,

[[Page 16526]]

marketing, and beneficiary communication purposes.
     Reliability means a measure of the fraction of the 
variation among the observed measure values that is due to real 
differences in quality (``signal'') rather than random variation 
(``noise''); it is reflected on a scale from 0 (all differences in plan 
performance measure scores are due to measurement error) to 1 (the 
difference in plan performance scores is attributable to real 
differences in performance).
     Reward factor means a rating-specific factor added to the 
contract's summary or overall (or both) rating if a contract has both 
high and stable relative performance.
     Statistical significance assesses how likely differences 
observed in performance are due to random chance alone under the 
assumption that plans are actually performing the same. Although not 
part of the proposed regulatory definition, we clarify that CMS uses 
statistical tests (for example, t-test) to determine if a contract's 
measure value is statistically different (greater than or less than 
depending on the test) from the national mean for that measure, or 
whether conversely, the observed differences from the national mean 
could have arisen by chance.
     Surviving contract means the contact that will still exist 
under a consolidation, and all of the beneficiaries enrolled in the 
consumed contract(s) are moved to the surviving contracts.
     Traditional rounding rules mean that the last digit in a 
value will be rounded. If rounding to a whole number, look at the digit 
in the first decimal place. If the digit in the first decimal place is 
0, 1, 2, 3 or 4, then the value should be rounded down by deleting the 
digit in the first decimal place. If the digit in the first decimal 
place is 5 or greater, then the value should be rounded up by 1 and the 
digit in the first decimal place deleted.
    We received no comments on the proposed definitions in paragraph 
(a) of Sec. Sec.  422.162 and 423.182 and are therefore finalizing 
without modification.
e. Contract Ratings
    Star Ratings and data reporting are at the contract level for most 
measures. Currently, data for measures are collected at the contract 
level including data from all plan benefit packages (PBPs) under the 
contract, except for the following Special Needs Plan (SNP)-specific 
measures which are collected at the PBP level: Care for Older Adults--
Medication Review, Care for Older Adults--Functional Status Assessment, 
and Care for Older Adults--Pain Assessment. The SNP-specific measures 
are rolled up to the contract level by using an enrollment-weighted 
mean of the SNP PBP scores. Although we discussed and solicited comment 
on the feasibility and burden of collecting data at the PBP (plan) 
level and the reliability of ratings at the plan level, we proposed to 
continue the practice of calculating the Star Ratings at the contract 
level and that all PBPs under the contract would have the same overall 
and/or summary ratings at paragraph (b)(1) of Sec. Sec.  422.162 and 
423.182.
    However, beneficiaries select a plan, rather than a contract, so we 
discussed in the proposed rule how we considered whether data should be 
collected and measures scored at the plan level. We have explored the 
feasibility of separately reporting quality data for individual D-SNP 
PBPs, instead of the current reporting level. For example, in order for 
CAHPS measures to be reliably scored, the number of respondents must be 
at least 11 people and reliability must be at least 0.60. In the 
proposed rule, we summarized our findings. Our current analyses show 
that, at the PBP level, CAHPS measures could be reliably reported for 
only about one-third of D-SNP PBPs due to sample size issues, and HEDIS 
measures could be reliably reported for only about one-quarter of D-SNP 
PBPs. If reporting were done at the plan level, a significant number of 
D-SNP plans will not be rated and in lieu of a Star Rating, Medicare 
Plan Finder will display that the plan is ``too small to be rated.'' 
However, when enough data are available, plan level quality reporting 
will reflect the quality of care provided to enrollees in that plan. 
Plan-level quality reporting will also give states that contract with 
D-SNPs plan-specific information on their performance and provide the 
public with data specific to the quality of care for dual eligible (DE) 
beneficiaries enrolled in these plans. For all plans as well as D-SNPs, 
reporting at the plan level will significantly increase plan burden for 
data reporting and will have to be balanced against the availability of 
additional clinical information available at the plan level. Plan-level 
ratings will also potentially increase the ratings of higher-performing 
plans when they are in contracts that have a mix of high and low 
performing plans. Similarly, plan-level ratings will also potentially 
decrease the ratings of lower-performing plans that are currently in 
contracts with a mix of high and low performing plans. Measurement 
reliability issues due to small sample sizes will also decrease our 
ability to measure true performance at the plan level and add 
complexities to the rating system. We solicited comments on balancing 
the improved precision associated with plan level reporting (relative 
to contract level reporting) with the negative consequences associated 
with an increase in the number of plans without adequate sample sizes 
for at least some measures; we asked for comments about this for D-SNPs 
and for all plans as we continue to consider whether rating at the plan 
level is feasible or appropriate. In particular, we solicited feedback 
on the best balance and whether changing the level at which ratings are 
calculated and reported better serves beneficiaries and our goals for 
the Star Ratings system.
    We also indicated that we were exploring whether some measure data 
could be reported at a higher level (parent organization versus 
contract) to ease and simplify reporting while continuing to remain 
useful (for example, call center measures as we anticipate that parent 
organizations use a consolidated call center to serve all contracts and 
plans) for the Star Ratings. Further, we said we are exploring if 
contract market area reporting is feasible when a contract covers a 
large geographic area. For example, when HEDIS reporting began in 1997, 
there were contract-specific market areas that evolved into reporting 
by market area for five states with large Medicare populations.\36\ We 
are planning to continue work in this area to determine the best 
reporting level for each measure that most accurately reflects 
performance and minimizes to the extent possible plan reporting burden. 
As we consider alternative reporting units, we solicited comments and 
suggestions about requiring reporting at different levels (for example, 
parent organization, contract, plan, or geographic area) by measure. In 
addition, section 50311(d) of the Bipartisan Budget Act of 2018 after 
publication of the proposed rule, amended section 1853 to require the 
Secretary to determine the feasibility of quality measurement at the 
plan level for all MA plans. CMS will use the feedback received from 
the proposed rule as we consider reporting options in the future and 
continue to evaluate this issue consistent with the Bipartisan Budget 
Act provision.
---------------------------------------------------------------------------

    \36\ The following states were divided into multiple market 
areas: CA, FL, NY, OH, and TX.
---------------------------------------------------------------------------

    We proposed to continue calculating the same overall and/or summary 
Star Ratings for all PBPs offered under an MA-only, MA-PD, or PDP 
contract and

[[Page 16527]]

to codify this policy in regulation text at Sec. Sec.  422.162(b) and 
423.182(b). We also proposed a cost plan regulation at Sec.  417.472(k) 
to require cost contracts to be subject to the part 422 and part 423 
Medicare Advantage and Part D Prescription Drug Program Quality Rating 
System. Specifically, we proposed, at paragraph (b)(1) that CMS will 
calculate overall and summary ratings at the contract level and 
proposed regulation text that cross-references other proposed 
regulations regarding the calculation of measure scoring and rating, 
and domain, summary and overall ratings. Further, we proposed to 
codify, at (b)(2) of each section, that data from all PBPs offered 
under a contract will continue to be used to calculate the ratings for 
the contract. For SNP specific measures collected at the PBP level, we 
proposed that the contract level score will be an enrollment-weighted 
mean of the PBP scores using enrollment in each PBP as reported as part 
of the measure specification, which is consistent with current 
practice. The proposed text is explicit that domain and measure 
ratings, other than the SNP-specific measures, are based on data from 
all PBPs under the contract.
    We received the following comments related to our proposals, and 
our responses follow:
    Comment: Most commenters opposed moving to plan-level reporting and 
expressed overwhelming support for retaining the current contact-level 
measurement. Commenters raised concerns about the additional 
complexity, administrative burden and reporting requirements of plan-
level reporting. Additionally, commenters reiterated our concerns 
regarding the reliability of the scores at the plan level, as well as 
the inability to report some measure due to inadequate sample sizes. A 
commenter urged CMS to continue reporting Star Ratings at the contract 
level for PDPs.
    Response: CMS appreciates commenters' support for contract-level 
reporting and acknowledge the complexities of moving to plan-level 
reporting given the challenges of accurately measuring quality with 
smaller groups and sample sizes and the additional administrative 
burden that would be placed on contracts.
    Comment: A handful of commenters supported plan-level reporting 
also recognized it may not be practical for all quality measures. Some 
of the commenters noted the utility for beneficiaries who choose among 
plans. A commenter suggested CMS require Part D plans to report certain 
medication-related measures at the Formulary ID level.
    Response: We agree that ideally for consumer choice, plan-level 
reporting or Formulary ID level reporting for Part D plans would be 
preferable, because it provides more detailed and targeted data. 
However, we need to consider the operational and methodological 
challenges of reporting at the plan level, including the ability to 
accurately measure performance at that level.
    Comment: A commenter stated that plan-level reporting would be open 
to potential gaming by contracts constructing the plan-level geographic 
areas to maximize Star Ratings for the greatest number of enrollees. 
The commenter suggested that contracts would consider how well each 
plan was performing in the Star Ratings program to determine the 
geographic area of each plan.
    Response: We appreciate this comment and will take it into account 
as we consider this issue in the future.
    Comment: A commenter noted that plan-level reporting would stifle 
innovation and discourage plans from serving difficult areas. This 
would limit the ability of contracts to implement innovative models in 
one plan prior to expanding.
    Response: We appreciate this comment since we clearly do not want 
our Star Ratings policies to stifle innovation. We will take this 
comment into consideration as we continue to consider options for 
different levels of reporting.
    Comment: A handful of commenters expressed interest in measurement 
at the local health services area, including by state. Many of these 
commenters noted that it will be challenging to move to reporting at 
the local geographic area. Issues to be considered include how to 
handle contracts that serve major metropolitan areas that cross state 
lines. A couple of commenters suggested that CMS consider creating 
additional contract numbers or market-level designations for a 
contract. A commenter recommended that CMS discontinue the moratorium 
that does not allow for existing H numbers to be split to allow more 
meaningful measurement.
    Response: CMS is committed to examining the feasibility of 
alternative levels of reporting, including by geographic area. The 
suggestions provided by commenters through the proposed rule will be 
taken into consideration as alternatives are explored. Additionally, 
section 50311(d) of the Bipartisan Budget Act of 2018 (P.L. 115-123) 
enacted after publication of the proposed rule, amended section 1853 to 
require the Secretary to determine the feasibility of quality 
measurement at the plan level for all MA plans. CMS plans to obtain 
additional feedback from stakeholders on this issue given the 
challenges of developing options that would be feasible for both the 
industry and CMS. CMS' contractor for the Star Ratings program is 
planning to convene a Technical Expert Panel following the publication 
of the final rule and this is one of the issues the panel will address. 
This panel will periodically meet to provide feedback on different 
critical Star Ratings issues. Information from the Technical Expert 
Panel will be publicly shared.
    Comment: A commenter expressed concern about pursuing market area 
reporting as such reporting could result in limiting the health care 
options for higher-need populations.
    Response: CMS appreciates this comment and does not want to limit 
options for higher-need populations. We will take the comment into 
consideration as we continue to consider options for different levels 
of reporting.
    Comment: A handful of commenters recommended adjusting the Star 
Ratings to account for variables that contribute to underperformance in 
certain geographic areas, network characteristics and patient 
characteristics by applying, for example, the case-mix adjustment 
process currently used for the CAHPS measures.
    Response: CMS appreciates this comment and will take it into 
account as we continue to consider options for different levels of 
reporting. As we contemplate case-mix adjustment, we need to ensure 
that we are not adjusting away true differences in the quality of care 
across contracts in different geographic areas or with different 
network structures.
    Comment: A commenter raised concerns of the possibility for gaming 
in connection with separate ratings for new contracts. If CMS is to 
proceed, the commenter would like to see simulations of the ratings.
    Response: CMS appreciates this comment and clearly does not want to 
implement changes that would encourage gaming of the Star Ratings 
system. We will take this comment into consideration as we continue to 
analyze different ways to rate contracts.
    Comment: A commenter raised a question about a potential error on 
page 82 FR 56380 in the sentence that reads ``For SNP specific measures 
collected at the PBP level, we propose that the contract level score 
would be an enrollment-weighted mean of the PBP

[[Page 16528]]

scores using enrollment in each PBP as reported as part of the measure 
specification, which is consistent with current practice.'' The 
commenter noted that the current practice weights the PBP scores by 
eligible population.
    Response: The text from the proposed rule is correct. The eligible 
population and the enrollment reported as part of the measure 
specification are the same.
    Comment: A handful of commenters from sponsoring organizations 
suggested separate reporting by Dual SNPs and non-Dual SNPs, and 
rolling up all Dual SNP PBPs and non-Dual SNP PBPs separately within a 
contract. A couple of commenters noted that moving to plan level 
reporting for all SNPs is complex with many pros and cons so they 
recommended that CMS continue contract-level reporting until all of the 
consequences can be fully evaluated.
    Response: CMS appreciates these comments, including the issues 
raised by commenters regarding the complexities of moving to plan/PBP-
level reporting by SNPs and non-SNPs. Given that some contracts just 
have SNP PBPs and other contracts offer both SNP and non-SNP plans, CMS 
needs to evaluate how this would impact reporting of measures and 
calculations. We agree that all of the benefits and disadvantages need 
to be weighed before a final decision is made about how to proceed and 
CMS is committed to continuing to obtain feedback from the industry on 
changes to the level of reporting. CMS continues to evaluate this 
issue. Additionally, in light of the passage of the Bipartisan Budget 
Act of 2018, CMS is required to examine the feasibility of plan-level 
reporting for both SNP and non-SNP plans. Any related changes would be 
proposed through future rulemaking.
    Comment: A couple of commenters supported the idea of reporting the 
call center and appeals measures at the parent organization level since 
in most cases these functions are organized at the parent organization 
level, while a couple of commenters did not like having different 
levels of reporting for different measures, arguing that it would 
create more complexity in the Star Rating program.
    Response: CMS appreciates the suggestions received from commenters 
and will continue to look at the advantages of moving to a different 
level of reporting for these and other measures. Any related changes 
would be proposed through future rulemaking.
    Comment: A commenter supports CMS' current process for rolling up 
SNP plan-benefit package level information to the contract level.
    Response: CMS thanks this commenter for their support for our 
current policy of calculating SNP measures.
    Comment: A handful of commenters recommended that CMS not make any 
changes in the unit for reporting until additional analyses are 
completed that ensures that any changes are fair and equitable to all 
sponsors. A commenter suggests an industry-wide workgroup to discuss 
potential changes to reporting levels and operational challenges.
    Response: We acknowledge these comments and agree that we need to 
do more analysis and obtain additional feedback from sponsors before we 
make any changes in the level of reporting. We support the desire to 
make sure that any changes are fair and equitable to all sponsoring 
organizations. As noted in a previous response, CMS' contractor for the 
Star Ratings program is planning to convene a Technical Expert Panel 
following the publication of the final rule and this is one of the 
issues the panel will address.
    For the reasons indicated in the proposed rule and our responses to 
the related comments, we are finalizing the provisions as proposed in 
paragraphs (b)(1) and (2) of Sec. Sec.  422.162 and 423.182 and Sec.  
417.472(k) without substantive modification. However, we realized that 
paragraphs (b)(1) as proposed did not specify that summary ratings also 
include the reward factor and the Categorical Adjustment Index as 
described in Sec. Sec.  422.166(f) and 423.186(f); we are finalizing 
additional text to clarify that in paragraphs (b)(1). In addition, we 
are slightly revising the last two sentences of paragraphs (b)(2) of 
the same regulation sections to clarify that the rule for including 
plan-level only measures is applicable to the SNP-specific measures 
that are reported only at the plan level.
f. Contract Consolidations
    We proposed a change in how contract-level Star Ratings are 
assigned in the case of contract consolidations. We noted in the 
proposed rule how we have historically permitted MAOs and Part D 
sponsors to consolidate contracts when a contract novation occurs to 
better align business practices. As noted in MedPAC's March 2016 Report 
to Congress (http://www.medpac.gov/docs/default-source/reports/march-2016-report-to-the-congress-medicare-payment-policy.pdf), there has 
been a continued increase in the number of enrollees being moved from 
lower Star Rating contracts that do not receive a QBP to higher Star 
Rating contracts that do receive a QBP as part of contract 
consolidations, which increases the size of the QBPs that are made to 
MAOs due to the large enrollment increase in the higher rated, 
surviving contract. We are worried that this practice results in 
masking low quality plans under higher rated surviving contracts. This 
does not provide beneficiaries with accurate and reliable information 
for enrollment decisions, and it does not truly reward higher quality 
contracts. We proposed to modify the calculation of Star Ratings for 
surviving contracts that have consolidated to address these concerns. 
Instead of assigning the surviving contract the Star Rating that the 
contract would have earned without regard to whether a consolidation 
took place, we proposed to assign and display on MPF Star Ratings based 
on the enrollment-weighted mean of the measure scores of the surviving 
and consumed contract(s) so that the ratings reflect the performance of 
all contracts (surviving and consumed) involved in the consolidation. 
Under our proposal, the calculation of the measure, domain, summary, 
and overall ratings will be based on these enrollment-weighted mean 
scores. We estimated that the number of contracts impacted by the 
proposal would be small relative to all contracts that qualify for 
QBPs. During the period from 1/1/2015 through 1/1/2017 annual 
consolidations for MA contracts ranged from a low of 7 in 2015 to a 
high of 19 in 2016 out of approximately 500 MA contracts. As proposed 
in Sec. Sec.  422.162(b)(3)(i)-(iii) and 423.182(b)(3)(i)-(iii), CMS 
will use enrollment-weighted means of the measure scores of the 
consumed and surviving contracts to calculate ratings for the first and 
second plan years following the contract consolidations. We believe 
that use of enrollment-weighted means will provide a more accurate 
snapshot of the performance of the underlying plans in the new 
consolidated contract, such that both information to beneficiaries and 
QBPs are not somehow inaccurate or misleading. We also proposed, 
however, that the process of weighting the enrollment of each contract 
and applying this general rule will vary depending on the specific 
types of measures involved in order to take into account the 
measurement period and data collection processes of certain measures. 
Our proposal was to treat ratings for determining Quality Bonus Payment 
(QBP) status for MA contracts differently than displayed Star Ratings 
for the first year following the consolidation for consolidations that 
involve the same parent organization and plans of the same plan type.

[[Page 16529]]

    We proposed to codify our new policy at Sec. Sec.  422.162(b)(3) 
and 423.182(b)(3). First, we proposed generally, at paragraph (b)(3)(i) 
of each regulation, that CMS will assign Star Ratings for consolidated 
contracts using the provisions of paragraph (b)(3). We proposed in 
Sec.  422.162(b)(3) both a specific rule to address the QBP rating for 
the first year after the consolidation and a rule for subsequent years. 
As Part D plan sponsors are not eligible for QBPs, Sec.  423.182(b)(3) 
was proposed without the QBP aspect. We proposed in Sec.  
422.162(b)(3)(iv) and Sec.  423.182(b)(3)(ii) the process for assigning 
Star Ratings for posting on the Medicare Plan Finder for the first 2 
years following the consolidation.
    For the first contract year following a consolidation, we proposed 
to use the enrollment-weighted means as calculated below to set Star 
Ratings for MPF publication:
     The Star Ratings measure scores for the consolidated 
entity's first plan year will be based on enrollment-weighted measure 
scores using the July enrollment of the measurement period of the 
consumed and surviving contracts for all measures, except the survey-
based and call center measures.
     The survey-based measures (that is, CAHPS, HOS, and HEDIS 
measures collected through CAHPS or HOS surveys) will use enrollment of 
the surviving and consumed contracts at the time the sample is pulled 
for the rating year. For example, for a contract consolidation that is 
effective January 1, 2021 the CAHPS sample for the 2021 Star Ratings 
will be pulled in January 2020 so enrollment in January 2020 will be 
used. The call center measures will use mean enrollment during the 
study period. We stated that we believed that these proposals for 
survey-based measures are more nuanced and account for how the data 
underlying those measures are gathered and that the enrollment-weighted 
means better reflect the true underlying performance of both the 
surviving and consumed contracts.
    For the second year following the consolidation, for all MA and 
Part D Sponsors, we proposed to calculate the Star Ratings will be 
calculated as follows:
     The enrollment-weighted measure scores using the July 
enrollment of the measurement period of the consumed and surviving 
contracts will be used for all measures except HEDIS, CAHPS, and HOS.
     We proposed that HEDIS and HOS measure data will be used 
as reported in the second year after consolidation. The current 
reporting requirements for HEDIS and HOS already combine data from the 
surviving and consumed contract(s) following the consolidation, so we 
did not propose any modification or averaging of these measure scores. 
For example, for HEDIS if an organization consolidates one or more 
contracts during the change over from measurement to reporting year, 
then only the surviving contract is required to report audited summary 
contract-level data but it must include data on all members from all 
contracts involved.
     We proposed to require that the CAHPS survey sample (that 
would be selected following the consolidation) would include enrollees 
in the sample universe from which the sample is drawn from both the 
surviving and consumed contracts. If there are two contracts (that is, 
Contract A is the surviving contract and Contract B is the consumed 
contract) that consolidate, and Contract A has 5,000 enrollees eligible 
for the survey and Contract B has 1,000 eligible for the survey, the 
universe from which the sample will be selected will be 6,000.
    CMS proposed that these rules would be used to calculate the 
measure scores in the first and second year after consolidation; 
following those two years, CMS proposed to use the other rules proposed 
in Sec. Sec.  422.166 and 423.186 to calculate the measure, domain, 
summary, and overall Star Ratings for the consolidated contract. In the 
third year after consolidation and subsequent years, the performance 
period for all the measures will be after the consolidation, so our 
proposal limited the special rules for calculating post-consolidation 
the Star Ratings to the Ratings issued the first 2 years after 
consolidation.
    When consolidations involve two or more contracts for health and/or 
drug services of the same plan type under the same parent organization 
combining into a single contract at the start of a contract year, we 
proposed to calculate the QBP rating for that first year following the 
consolidation using the enrollment-weighted mean, using traditional 
rounding rules, of what would have been the QBP ratings of the 
surviving and consumed contracts using the contract enrollment in 
November of the year the Star Ratings were released. In November of 
each year following the release of the ratings on Medicare Plan Finder, 
the preliminary QBP ratings are displayed in the Health Plan Management 
System (HPMS) for the year following the Star Ratings year. For 
example, if the first year the consolidated entity is in operation is 
plan year 2020, the 2020 QBP rating displayed in HPMS in November 2018 
will be based on the 2019 Star Ratings (which are released in October 
2018) and calculated using the weighted mean of the November 2018 
enrollment of the surviving and consumed contracts. Because the same 
parent organization is involved in these situations, we believe that 
many administrative processes and procedures are identical in the 
Medicare health plans offered by the sponsoring organization, and using 
a weighted mean of what will have been their QBP ratings accurately 
reflects their performance for payment purposes. In subsequent years 
after the first year following the consolidation, QBPs status will be 
determined based on the consolidated entity's Star Rating posted on 
MPF. Under our proposal, the measure, domain, summary, and (in the case 
of MA-PD plans) the overall Star Ratings posted on Medicare Plan Finder 
for the second year following consolidation would be based on the 
enrollment-weighted measure scores so would include data from all 
contracts involved. Consequently, we stated that we believed the 
ratings used for QBP status determinations would reflect the care 
provided by both the surviving and consumed contracts.
    In conclusion, we proposed a new set of rules regarding the 
calculation of Star Ratings for consolidated contracts to be codified 
at paragraphs (b)(3) of Sec. Sec.  422.162 and 423.182. We solicited 
comment on this proposal and whether our separate treatment of 
different measure types during the first and second year adequately 
addresses the differences in how data are collected (and submitted) for 
those measures during the different periods. We also solicited feedback 
on whether sponsoring organizations believe that the special rule for 
consolidations involving the same parent organization and same plan 
types adequately addresses how those situations are different from 
cases where an MA organization buys or sells a plan or contract from or 
to a different entity and whether these rules should be extended to 
situations where there are different parent organizations involved. For 
commenters that support the latter, we also requested comment on how 
CMS should determine that the same administrative processes are used 
and whether attestations from sponsoring organizations or evidence from 
prior audits should be required to support such determinations.
    Following publication of our proposed rule, Congress enacted the 
Bipartisan Budget Act of 2018. Section 53112 of the Act amended section 
1853(o) to require an adjustment to the Star Ratings, quality bonus 
under

[[Page 16530]]

section 1853(o) and rebate allocation under section 1854 based on the 
quality rating to ``prevent the artificial inflation'' of Star Ratings 
after consolidation. That required adjustment applies for 
consolidations approved on or after January 1, 2019. The statutory 
change requires the adjustment be applied when a single MA organization 
consolidates contracts and reflect an enrollment-weighted average of 
scores or ratings for the underlying contracts. We believe that our 
proposal is generally consistent with the new statutory requirement, 
with minor exceptions. The proposal would not have applied until a 
later period, but, as noted in section II.A.11.c of this final rule, we 
will finalize these provisions to be applicable beginning with the 2020 
QBPs and 2020 Star Ratings produced in fall 2019 to be consistent with 
the statute. Our proposal was for consolidations involving a single 
parent organization while the statute focused on consolidations 
involving a single MA organization; applying the proposed policy to 
consolidations at the level of the parent organization instead of the 
specific MA organization captures more consolidations. We read the 
Bipartisan Budget Act as setting a floor rather than a ceiling on our 
authority to establish and set the rules governing the Stars Rating 
system. In addition, our proposal also was more specific as to how 
enrollment-weighted ratings at the measure and contract level would be 
used following the consolidation. We believe the additional detail in 
our proposal is explicitly authorized as the statutory change leaves it 
to the Secretary to identify the specific appropriate adjustments.
    We received the following comments on our proposals and 
solicitations for feedback, and our responses follow:
    Comment: Commenters expressed overwhelming support for our rules 
outlined at Sec. Sec.  422.162(b)(3) and 423.182(b)(3) to calculate 
contract-level Star Ratings in the case of contract consolidations. 
Commenters stated that this would be a more accurate picture of the 
performance of the underlying contracts. Commenters noted that this 
would help eliminate the gaming that can occur when consolidations of 
multiple contracts in distinct geographic areas result in artificial 
increases the Star Ratings and Quality Bonus Payment (QBP) ratings. A 
number of commenters suggested that this approach was fair and 
equitable to all stakeholders. Some commenters supported this change as 
a short-term solution, but they wanted CMS to consider how in the 
future the ratings could more accurately reflect the care provided at 
the local market area. Commenters recognized that quality reporting at 
the local market area is a sizeable change and would not be feasible 
for a number of years.
    Response: CMS appreciates the commenters' support for revising how 
Star Ratings and QBP ratings are calculated when two or more contracts 
consolidate. We believe that the Bipartisan Budget Act indicates that 
Congress is similarly concerned about these issues and our proposal to 
address them. We also agree with commenters that local market area 
reporting would be preferable in cases when the contracts are 
geographically dispersed. Although moving to local market area 
reporting has many challenges, CMS is committed to work with 
stakeholders to examine the feasibility of local market area reporting. 
Any potential changes that would change the consolidation policy in the 
direction of local market area reporting would be proposed in future 
rulemaking.
    Comment: A commenter recommended that CMS issue contract numbers at 
the state level and then base Star Ratings at the state level to avoid 
consolidations across disperse geographic areas.
    Response: State-based contract numbers would be administratively 
burdensome for both contracts and CMS, would significantly increase 
reporting burden of contracts, and would create measurement challenges 
since many contracts at the state level will not have a sufficient 
number of enrollees by state to calculate reliably the quality measures 
that are part of the Part C and D Star Ratings program. Contracts that 
serve disperse geographic areas often have the majority of their 
enrollees in one or two states with smaller enrollment in other states.
    Comment: A commenter suggested using the unrounded final summary 
mean rather than the rounded final Star Rating.
    Response: CMS is assuming this commenter is referring to the QBP 
rating for the first year of the consolidation. For all other years, 
the QBP rating of the contract would be based on the Star Ratings 
posted on Medicare Plan Finder; therefore for the second year following 
a consolidation, the same rules for calculating the Star Ratings for 
QBP and for MPF posting would apply (that is, Sec. Sec.  
422.162(b)(3)(iii)). The preliminary QBP rating is produced and posted 
in HPMS in November of each year for the bids that will be submitted 
the following year. The QBP appeals process is based on these ratings 
posted in November. In April prior to the bids being due, CMS would 
update the QBP rating using an enrollment-weighted QBP ratings of all 
contracts involved in the consolidation which are already rounded.
    Comment: A commenter asked CMS to consider a grace period that 
would neither reward nor disadvantage the surviving contract as a 
result of acquiring a poor performing contract.
    Response: Under our current policy, a sponsor can gain financially 
by consolidating enrollees from a poor-performing contract into a 
contract that receives a QBP and thereby receive bonus payments that it 
would not have been entitled to receive had the consolidation not 
occurred. The revised methodology for calculating Star Ratings and QBPs 
for the surviving contract takes into consideration the performance of 
all contracts involved; thus, it is a more accurate measure of 
performance. We do not believe that a more accurate reflection of 
performance can be fairly termed a ``reward'' or a ``disadvantage'' of 
contract consolidation.
    Comment: A handful of commenters expressed concern regarding the 
consolidation policy stating that they thought the calculations were 
too complex. A commenter stated it would limit the beneficiary options 
to enroll in plans with richer benefits since there would not be the 
same incentives to consolidate lower performing contracts into higher 
performing ones receiving QBPs.
    Response: Most of the calculations for the revised consolidation 
policy will be handled by CMS, though contracts will have an 
opportunity to review the calculations as part of the normal Star 
Rating review process. The consolidation policy should not make it more 
difficult for contracts to produce the data that are needed for the 
Star Ratings program. The premise behind all of the calculations is to 
combine the already gathered (or currently gathered) data from all 
contracts involved in the consolidation using an enrollment-weighted 
average. This policy should not create a situation which limits options 
for beneficiaries to enroll in plans with richer benefits. As always, 
beneficiaries may is able to choose in their service area any plan that 
best meets their needs. If a beneficiary decides to remain in a 
contract that consolidates, the ratings for that contract will now more 
accurately reflect performance of that contract.
    Comment: A commenter suggested that CMS post by year end in HPMS a 
worksheet with the exact enrollment and overall Star Rating values 
which CMS intends to use for determining QBP ratings for consolidated 
contracts.

[[Page 16531]]

    Response: In November of each year, CMS posts in HPMS the 
preliminary QBP ratings for the bids that will be submitted the 
following year. This starts the QBP appeals process. In April of each 
year prior to bids being submitted, CMS posts in HPMS the final QBP 
rating following the appeals process. In November of each year or at 
year end, CMS would not be aware of future consolidations that would be 
announced near the time of the bid so would be unable to post a 
combined rating for the consolidated contracts at this time. As long as 
CMS is aware of the consolidation by April at the time of the HPMS 
posting, the combined rating for the consolidated contracts would be 
posted at that time. A parent organization would have sufficient 
information to calculate the enrollment weighted QBP rating of a 
consolidated contract using the preliminary QBP ratings posted in HPMS 
in November of each year.
    Comment: A handful of commenters requested that CMS clarify the 
timing of this provision. These commenters expressed a preference for 
it not to begin until the 2021 Star Ratings and 2022 QBPs.
    Response: The proposed rule stated that all of the changes related 
to Star Ratings would go into effect for performance periods in 2019 
(thus, for the 2021 Star Ratings and 2022 QBPs). However, in light of 
the passage of the Bipartisan Budget Act provision which requires 
enrollment-weighted adjustments to the Star Ratings for contract 
consolidations approved on or after 1/1/19, we are finalizing the 
regulation text on this policy to be applicable to consolidations that 
occur on or after the same date. The final regulations at Sec. Sec.  
422.162(b)(3) and 423.182(b)(3) will apply to the star ratings of 
surviving contracts from contract consolidations that are approved on 
or after January 1, 2019. Thus, the policy will be implemented for the 
2020 Star Ratings and the 2020 QBPs. We note that while the statute is 
specific to MA ratings, we are finalizing the same policy for Part D 
Ratings on the same timeframe to have consistent methodology across 
Part C and D for beneficiaries choosing a contract.
    Comment: A few commenters were interested in a similar policy for 
consolidations between different parent organizations.
    Response: We treat the purchase of a contract, multiple contracts 
or all of the contracts offered by a parent organization by different 
parent organization is known as a novation, not a consolidation, even 
though the consolidation will generally also require similar contract 
documents and approvals from CMS. Where one entity is buying all or 
part of the business of another entity, we did not propose and do not 
intend to apply the consolidation policy finalized in this rule. In 
novations, the structure of each of the individual contracts being 
purchased does not change and the contracts still provide the same 
services within the same service area before and after the novation is 
completed, only the company that owns the business and is the MA 
organization under the contract has changed. The Star Rating for each 
individual contract transfers with the contract and remains intact 
until the next rating cycle. Novations can occur at any point during 
the calendar year.
    A consolidation by contrast is when two or more contracts owned by 
the same parent organization are combined into a single contract. The 
overall service area of the two contracts are combined, the contract 
number of the consumed contract(s) is retired and the contract number 
of the surviving contract now provides all of the services in the 
combined service area. To consolidate contracts, all of the contracts 
must be owned by the same parent organization. Consolidations can only 
occur at the change from one year to another year and must be submitted 
and approved by CMS by a specific deadline in the annual contracting 
process. If one parent organization buys another contract owned by a 
different parent organization, the sponsor could consolidate multiple 
contracts using the rules outlined in this rule the year after the 
novation takes place. With a consolidation, the rule finalized here for 
the calculation of the Star Rating of the surviving contract would 
apply.
    Comment: A commenter wanted CMS to propose other alternatives and 
offer additional opportunities to comment, but no additional detail was 
provided on suggested alternatives.
    Response: CMS appreciates the request for other alternatives. 
Commenters to the proposed rule did not suggest other ways to handle 
contracts that consolidate and expressed overwhelming support for this 
policy. CMS will continue to consider if there is a better way to 
account for differences in performance across geographic areas and will 
provide opportunities to engage stakeholders and obtain additional 
input.
    For the reasons set forth in the proposed rule and our responses to 
the related comments summarized earlier, we are finalizing the 
provisions as proposed at Sec. Sec.  422.162(b)(3) and 423.182(b)(3), 
except for modifying the timeframe applying these new rules. The 
revised consolidation policy would be applicable for the Rating for any 
surviving contract after a consolidation that is approved on or after 
January 1, 2019. Although the statute related to consolidations is 
specific to MA ratings, we are finalizing the same policy for Part D 
ratings on the same timeframe to have consistent methodology across 
Part C and D for beneficiaries choosing a contract.
g. Data Sources
    Under 1852(e) of the Act, MA organizations are required to collect, 
analyze, and report data that permit measurement of health outcomes and 
other indices of quality. The Star Ratings system is based on 
information collected consistent with section 1852(e) of the Act. 
Section 1852(e)(3)(B) of the Act prohibits the collection of data on 
quality, outcomes, and beneficiary satisfaction other than the types of 
data that were collected by the Secretary as of November 1, 2003; there 
is a limited exception for SNPs to collect, analyze, and report data 
that permit the measurement of health outcomes and other indicia of 
quality. The statute does not require that only the same data be 
collected, but that we do not change or expand the type of data 
collected until after submission of a Report to Congress (prepared in 
consultation with MA organizations and accrediting bodies) that 
explains the reason for the change(s). We clarify here that the types 
of data included under the Star Ratings system are consistent with the 
types of data collected as of November 1, 2003. Since 1997, Medicare 
managed care organizations have been required to annually report 
quality of care performance measures through HEDIS. We have also been 
conducting the CAHPS survey since 1997 to measure beneficiaries' 
experiences with their health plans. HOS began in 1998 to capture 
changes in the physical and mental health of MA enrollees. To some 
extent, these surveys have been revised and updated over time, but the 
same types of data--clinical measures, beneficiary experiences, and 
changes in physical and mental health, respectively--have remained the 
focus of these surveys. In addition, there are several measures in the 
Stars Ratings System that are based on performance that address 
telephone customer service, members' complaints, disenrollment rates, 
and appeals; however these additional measures are not collected 
directly from the sponsoring organizations for the primary purpose of 
quality measurement so they are not information collections governed by

[[Page 16532]]

section 1852(e). These additional measures are calculated from 
information that CMS has gathered as part of the administration of the 
Medicare program, such as information on appeals forwarded to the 
Independent Review Entity under subparts M, enrollment, and compliance 
and enforcement actions.
    The Part D program was implemented in 2006, and while there is no 
parallel provision regarding applicable Part D sources of data, we have 
used similar datasets, for example CAHPS survey data, for 
beneficiaries' experiences with prescription drug plans. Section 1860D-
4(d) of the Act specifically directs the administration and collection 
of data from consumer surveys in a manner similar to those conducted in 
the MA program. All of these measures reflect structure, process, and 
outcome indices of quality that form the measurement set under Star 
Ratings. Since 2007, we have publicly reported a number of measures 
related to the drug benefit as part of the Star Ratings. For MA 
organizations that offer prescription drug coverage, we use the same 
Part D measures focusing on administration of the drug benefit as is 
used for stand-alone PDPs. Similar to MA measures of quality relative 
to health services, the Part D measures focus on customer service and 
beneficiary experiences, effectiveness, and access to care relative to 
the drug benefit. We believe that the Part D Star Ratings are 
consistent with the limitation expressed in section 1852(e) of the Act 
even though the limitation does not apply to our collection of Part D 
quality data from Part D sponsors.
    We intend to continue to base the types of information collected in 
the Part C Star Ratings on section 1852(e) of the Act, and we proposed 
at Sec.  422.162(c)(1) that the type of data used for Star Ratings will 
be data consistent with the section 1852(e) limits and data gathered 
from CMS administration of the MA program. In addition, we proposed in 
Sec.  422.162(c)(1) and in Sec.  423.182(c)(1) to include measures that 
reflect structure, process, and outcome indices of quality, including 
Part C measures that reflect the clinical care provided, beneficiary 
experience, changes in physical and mental health, and benefit 
administration, and Part D measures that reflect beneficiary 
experiences and benefit administration. The measures encompass data 
submitted directly by MA organizations (MAOs) and Part D sponsors to 
CMS, surveys of MA and Part D enrollees, data collected by CMS 
contractors, and CMS administrative data. We also proposed, primarily 
so that the regulation text is complete on this point, a regulatory 
provision at Sec. Sec.  422.162(c)(2) and 423.182(c)(2) that requires 
MA organizations and Part D plan sponsors to submit unbiased, accurate, 
and complete quality data as described in paragraph (c)(1) of each 
section. Our authority to collect quality data is clear under the 
statute and existing regulations, such as section 1852(e)(3)(A) and 
1860D-4(d) and Sec. Sec.  422.12(b)(2) and 423.156. We proposed the 
paragraph (c)(2) regulation text to ensure that the quality ratings 
system regulations include a regulation on this point for readers and 
to avoid confusion in the future about the authority to collect this 
data. In addition, it is important that the data underlying the ratings 
are unbiased, accurate, and complete so that the ratings themselves are 
reliable. This regulation text will clearly establish the sponsoring 
organization's responsibility to submit data that can be reliably used 
to calculate ratings and measure plan performance.
    We received the following comments on this proposal, and our 
responses follow:
    Comment: A few commenters supported codifying language to clearly 
establish the sponsoring organization's responsibility to submit data 
that can be reliably used to calculate ratings and measure plan 
performance.
    Response: CMS appreciates stakeholders' support of our effort to 
codify language to ensure that that the data submitted are accurate and 
reliable. We are finalizing the language as proposed.
    Comment: Responses were mixed on whether audit data should be used 
in the Star Ratings. A couple of commenters opposed including measures 
in the Star Ratings program that rely on audit findings as a data 
source. Other commenters stated given the Beneficiary Access and 
Performance Problems measure that previously included enforcement 
actions was moved out of the 2019 Star Ratings and to the display page, 
they strongly urged CMS to re-incorporate audit information, including 
information about enforcement actions, in Star Ratings. Those in favor 
of using audit information noted that the key purposes of Quality 
Rating System are to provide comparative information to Medicare 
beneficiaries, to base payment on quality, and to oversee the overall 
performance of plans. These commenters opposed CMS removal of audit 
findings and enforcement actions from the Star Ratings since 
deficiencies, in particular repeat deficiencies, may impact beneficiary 
access to drugs and services and the Star Ratings will not reflect 
these issues.
    Response: We appreciate the commenters' feedback and concerns 
received on the use of audits, compliance actions, and enforcement 
actions in the Star Ratings. In the proposed rule, the Beneficiary 
Access and Performance Problems measure was not proposed for the 2021 
Star Ratings even though some stakeholders strongly support including 
some recognition in the Star Ratings program when serious or repeat 
deficiencies are uncovered in audits or other means. These stakeholders 
argue that such deficiencies directly impact beneficiary access to 
needed services and drugs and therefore should be part of the Star 
Ratings program. We will continue to consider the comments as we 
continue our dialogue with stakeholders on this issue and any future 
changes will be proposed in future rulemaking.
    For the reasons set forth in the proposed rule and our responses to 
the related comments, we are finalizing the provisions regarding the 
data sources for measures and ratings as proposed in Sec. Sec.  
422.162(c) and 423.182(c) with two modifications. In Sec.  
422.162(c)(1), we are finalizing additional text to clarify that CMS 
administrative data will be used in the scoring for measures; the new 
text aligns the Part C regulation with the parallel Part D regulation. 
As noted in the proposed rule (82 FR 56382), some measures are based on 
data that CMS (or a contractor) has related to performance by 
sponsoring organizations and we are including a reference to CMS 
administrative data consistent with that longstanding policy. In 
addition, in Sec.  423.182(c)(2), we are finalizing additional text to 
clarify that the reported data permit measurement of health outcomes 
and other indices of quality, consistent with the scope of the measures 
in the Star Ratings program.
h. Adding, Updating, and Removing Measures
    We are committed to continuing to improve the Part C and D Star 
Ratings system by focusing on improving clinical and other outcomes. We 
anticipate that new measures will be developed and that existing 
measures will be updated over time. NCQA and the Pharmacy Quality 
Alliance (PQA) continually work to update measures as clinical 
guidelines change and develop new measures focused on health and drug 
plans. To address these anticipated changes, we proposed in Sec. Sec.  
422.164 and 423.184 specific rules to govern the addition, update, and 
removal of measures. We also proposed to apply these rules to the 
measure set proposed

[[Page 16533]]

in this rulemaking, to the extent that there are changes to the measure 
set between the effective date of this final rule and the Star Ratings 
based on this final rule (that is the ratings based on the performance 
periods beginning on or after January 1, 2019).
    As discussed in more detail in the following paragraphs, we 
proposed the following general rules to govern adding, updating, and 
removing measures:
     For data quality issues identified during the calculation 
of the Star Ratings for a given year, we proposed to continue our 
current practice of removing the measure from the Star Ratings.
     That new measures and substantive updates to existing 
measures would be added to the Star Ratings system based on future 
rulemaking but that prior to such a rulemaking, CMS would announce new 
measures and substantive updates to existing measures and solicit 
feedback using the process described for changes in and adoption of 
payment and risk adjustment policies in section 1853(b) of the Act 
(that is the Call Letter attachment to the Advance Notice and Rate 
Announcement).
     That existing measures (currently existing or existing 
after a future rulemaking) used for Star Ratings would be updated 
(without rulemaking) with regular updates from the measure stewards 
through the process described for changes in and adoption of payment 
and risk adjustment policies in section 1853(b) of the Act when the 
changes are not substantive.
     That existing measures (currently existing or existing 
after a future rulemaking) used for Star Ratings would be removed from 
use in the Star Ratings when there has been a change in clinical 
guidelines associated with the measure or reliability issues identified 
in advance of the measurement period; CMS would announce the removal 
using the process described for changes in and adoption of payment and 
risk adjustment policies in section 1853(b) of the Act. Removal might 
be permanent or temporary, depending on the basis for the removal.
    We proposed specific rules for updating and removal that would be 
implemented through subregulatory action, so that rulemaking would not 
be necessary for certain updates or removals. CMS proposed to announce 
application of the regulation standards in the Call Letter attachment 
to the Advance Notice and Rate Announcement process issued under 
section 1853(b) of the Act.
    First, we proposed to codify, at Sec. Sec.  422.164(a) and 
423.184(a), regulation text stating the general rule that CMS would 
add, update, and remove measures used to calculate Star Ratings as 
provided in Sec. Sec.  422.164 and 423.184. In each paragraph regarding 
addition, updating, and removal of measures and the use of improvement 
measures, we also proposed to make certain of these changes without 
future rulemaking by applying the standards and authority in the 
regulation text. CMS proposed to solicit feedback of its application of 
such rules using the draft and final Call Letter each year. In 
addition, CMS proposed in paragraph (a) of each section to issue a 
complete list of the measure set for each year in the Technical Notes 
or similar guidance document.
    Second, we proposed, in paragraph (b) of these sections, that CMS 
would review the quality of the data on which performance, scoring, and 
rating of measures is done each year. We proposed to continue our 
current practice of reviewing data quality across all measures, 
variation among organizations and sponsors, and measures' accuracy, 
reliability, and validity before making a final determination about 
inclusion of measures in the Star Ratings. We explained that this rule 
was designed to ensure that Star Ratings measures accurately measure 
true plan performance. If a systemic data quality issue is identified 
during the calculation of the Star Ratings, paragraph (b) would 
authorize CMS to remove the measure from that year's rating.
    Third, we proposed to address the addition of new measures in 
paragraph (c).
    In the proposed rule, we explained that our proposal regarding the 
addition of measures was guided by the principles we reiterated in this 
final rule in section II.A.11.b. Measures should be aligned with best 
practices among payers and the needs of the end users, including 
beneficiaries. Our strategy is to continue to adopt measures when they 
are available, that are nationally endorsed, and in alignment with the 
private sector, as we do today through the use of measures developed by 
NCQA and the PQA, and the use of measures that are endorsed by the 
National Quality Forum (NQF). We proposed to codify that CMS would 
continue to review measures of this type for adoption at Sec. Sec.  
422.164(c)(1) and 423.184(c)(1). We do not intend this standard to 
require that a measure be adopted by an independent measure steward or 
endorsed by NQF in order for us to propose its use for the Star 
Ratings, but that these are considerations that will guide us as we 
develop such proposals. We also proposed that CMS would develop its own 
measures as well when appropriate to measure and reflect performance in 
the Medicare program. For the 2021 Star Ratings, we proposed to have 
measures that encompass outcome, intermediate outcome, patient/consumer 
experience, access, process, and improvement measures. It is important 
to have a mix of different types of measures in the Star Ratings 
program to understand how all of the different facets of the provision 
of health and drug services interact. For example, process measures are 
evidence-based best practices that lead to clinical outcomes of 
interest. Process measures are generally easier to collect, while 
outcome measures are sometimes more challenging requiring in some cases 
medical record review and more sophisticated risk-adjustment 
methodologies.
    Over time new measures would be added and measures would be removed 
from the Star Ratings program to meet our policy goals. As new measures 
are added, we noted in the proposed rule that our general guidelines 
for deciding whether to propose new measures through future rulemaking 
would use the following criteria:
     Importance: The extent to which the measure is important 
to making significant gains in health care processes and experiences, 
access to services and prescription medications, and improving health 
outcomes for MA and Part D enrollees.
     Performance Gap: The extent to which the measure 
demonstrates opportunities for performance improvement based on 
variation in current health and drug plan performance.
     Reliability and Validity: The extent to which the measure 
produces consistent (reliable) and credible (valid) results.
     Feasibility: The extent to which the data related to the 
measure are readily available or could be captured without undue burden 
and could be implemented by the majority of MA and Part D contracts.
     Alignment: The extent to which the measure or measure 
concept is included in one or more existing federal, State, and/or 
private sector quality reporting programs.
    As explained in the proposed rule, CMS would balance these criteria 
as part of our decision-making process so that each new measure 
proposed for addition to the Star Ratings meets each criteria in some 
fashion or to some extent. We intend to apply these criteria

[[Page 16534]]

to identify and adopt new measures for the Star Ratings, which would be 
done through future rulemaking and include explanations for how and why 
we propose to add new measures. We also proposed to follow the process 
in our proposed paragraphs (c)(2) through (4) of Sec. Sec.  422.164 and 
423.184 when a new measure has been identified for inclusion in the 
Star Ratings. We proposed to initially solicit feedback on any 
potential new measures through the Call Letter and to codify that as a 
requirement at paragraph (c)(2) of each section.
    As new performance measures are developed and adopted, we proposed, 
at Sec. Sec.  422.164(c)(3) and (4) and 423.184(c)(3) and (4), that 
they would initially be incorporated into the display page for at least 
2 years but that we would keep a new measure on the display page for a 
longer period if CMS finds there are reliability or validity issues 
with the measure. As noted in the Introduction, the rulemaking process 
creates a longer lead time for changes, in particular to add a new 
measure to the Star Ratings or to make substantive changes to measures 
as discussed later in this section. Here is an example timeline for 
adding a new measure to the Star Ratings. In this scenario, the new 
measure has already been developed by the NCQA and the PQA, and 
endorsed by the NQF. Otherwise, that process may add an extra 3 to 5 
years to the timeline.
     January 2019: Solicit feedback in the draft 2020 Call 
Letter on whether to add the new measure.
     April 2019: Summarize feedback in the 2020 Call Letter on 
adding the new measure.
     2020/2021: Propose adding the new measure to the 2024 Star 
Ratings (2022 measurement period) in a proposed rule; finalize through 
rulemaking (for 1/1/2022 effective date).
     2020: Performance period and collection of data for the 
new measure and collection of data for posting on the 2022 display 
page.
     2021: Performance period and collection of data for the 
new measure and collection of data for posting on the 2023 display 
page.
     Fall 2021: Publish new measure on the 2022 display page 
(2020 measurement period).
     January 1, 2022: Applicability date of new measure for 
Star Ratings.
     2022: Performance period and collection of data for the 
new measure and collection of data for inclusion in the 2024 Star 
Ratings.
     Fall 2022: Publish new measure on the 2023 display page 
(2021 measurement period).
     Fall 2023: Publish new measure in the 2024 Star Ratings 
(2022 measurement period).
     2025: QBP status and rebate retention allowances are 
determined for the 2025 payment year.
    Fourth, at Sec. Sec.  422.164(d) and 423.184(d) we proposed to 
address updates to measures based on whether an update is substantive 
or non-substantive. Since quality measures are routinely updated (for 
example, when clinical codes are updated), we proposed to adopt rules 
for the incorporation of non-substantive updates to measures that are 
part of the Star Ratings system without going through new rulemaking. 
As proposed in paragraphs (d)(1) of Sec. Sec.  422.164 and 423.184, we 
would only incorporate updates without rulemaking for measure 
specification changes that do not substantively change the nature of 
the measure.
    Substantive changes (for example, major changes to methodology or 
specifications) to existing measures would be proposed and finalized 
through rulemaking. In paragraphs (d)(2) of Sec. Sec.  422.164 and 
423.184, we proposed to initially solicit feedback on whether to make 
the substantive measure update through the Call Letter prior to the 
measurement period for which the update would be initially applicable. 
For example, if the change announced significantly expands the 
denominator or population covered by the measure (for example, the age 
group included in the measures is expanded), the measure would be moved 
to the display page for at least 2 years and proposed through 
rulemaking for inclusion in Star Ratings. We noted in our proposal that 
this process for substantive updates would be similar to the process 
proposed for adopting new measures under proposed paragraph (c). As 
appropriate, the legacy measure may remain in the Star Ratings while 
the updated measure is on the display page if, for example, the updated 
measure expands the population covered in the measure and the legacy 
measure remains relevant and measures a critical topic for the Star 
Ratings. Adding the substantively updated measure to the Star Ratings 
would be proposed through rulemaking.
    We proposed to adopt rules to incorporate specification updates 
that are non-substantive in paragraph (d)(1). Non-substantive updates 
that occur (or are announced by the measure steward) during or in 
advance of the measurement period would be incorporated into the 
measure and announced using the Call Letter. We proposed to use such 
updated measures to calculate and assign Star Ratings without the 
updated measure being placed on the display page. Our proposal was 
explained as consistent with current practice.
    In paragraphs (d)(1)(i)-(v) of Sec. Sec.  422.164 and (d)(1)(i)-(v) 
of 423.184, we proposed to codify a non-exhaustive list of non-
substantive updates announced during or prior to the measurement period 
and how we will treat them under our proposal. The list includes 
updates in the following circumstances:
     If the change narrows the denominator or population 
covered by the measure with no other changes, the updated measure would 
be used in the Star Ratings program without interruption. For example, 
if an additional exclusion--such as excluding nursing home residents 
from the denominator--is added, the change will be considered non-
substantive and will be incorporated automatically. In our view, 
changes to narrow the denominator generally benefit Star Ratings of 
sponsoring organizations and should be treated as non-substantive for 
that reason.
     If the change does not meaningfully impact the numerator 
or denominator of the measure, the measure would continue to be 
included in the Star Ratings. For example, if additional codes are 
added that increase the number of numerator hits for a measure during 
or before the measurement period, such a change is not considered 
substantive because the sponsoring organization generally benefits from 
that change. This type of administrative change has no impact on the 
current clinical practices of the plan or its providers, and thus will 
not necessitate exclusion from the Star Ratings system of any measures 
updated in this way.
     The clinical codes for quality measures (such as HEDIS 
measures) are routinely revised as the code sets are updated. For 
updates to address revisions to the clinical codes without change in 
the intent of the measure and the target population, the measure would 
remain in the Star Ratings program and would not move to the display 
page. Examples of clinical codes that might be updated or revised 
without substantively changing the measure include:
    ++ ICD-10-CM (``ICD-10'') code sets. Annually, there are new ICD 10 
coding updates, which are effective from October 1 through September 
30th of any given year.
    ++ Current Procedural Terminology (CPT) codes. These codes are 
published and maintained by the American Medical Association (AMA) to 
describe

[[Page 16535]]

tests, surgeries, evaluations, and any other medical procedure 
performed by a healthcare provider on a patient.
    ++ Healthcare Common Procedure Coding System (HCPCS) codes. These 
codes cover items, supplies, and non-physician services not covered by 
CPT codes.
    ++ National Drug Code (NDC). The PQA updates NDC lists biannually, 
usually in January and July.
     If the measure specification change is providing 
additional clarifications such as the following, the measure would also 
not move to the display page since it does not change the intent of the 
measure but provides more information about how to meet the measure 
specifications:
    ++ Adding additional tests that will meet the numerator 
requirements.
    ++ Clarifying documentation requirements (for example, medical 
record documentation).
    ++ Adding additional instructions to identify services or 
procedures that meet (or do not meet) the specifications of the 
measure.
     If the measure specification change is adding additional 
data sources, the measure would also not move to the display page 
because we believe such changes are merely to add alternative ways to 
collect the data to meet the measure specifications without changing 
the intent of the measure.
    We solicited comment on our proposal to add non-substantive updates 
to measures and using the updated measure (replacing the legacy 
measure) to calculate Star Ratings. In particular, we noted our 
interest in stakeholders' views whether only non-substantive updates 
that have been adopted by a measure steward after a consensus-based or 
notice and comment process should be added to the Star Ratings under 
this proposed authority. Further, we solicited comment on whether there 
are other examples or situations involving non-substantive updates that 
should be explicitly addressed in the regulation text or if our 
proposal is sufficiently extensive.
    In addition to updates and additions of measures, we proposed rules 
to address the removal of measures from the Star Ratings to be codified 
in Sec. Sec.  422.164(e) and 423.184(e). In paragraph (e)(1) of each 
section, we proposed the two circumstances under which a measure will 
be removed entirely from the calculation of the Star Ratings. The first 
circumstance we identified was a change or changes in clinical 
guidelines that mean that the measure specifications are no longer 
believed to align with or promote positive health outcomes. We 
explained that as clinical guidelines change, we would need the 
flexibility to remove measures from the Star Ratings that are not 
consistent with current guidelines. We proposed to announce such 
subregulatory removals through the Call Letter so that removals for 
this reason are accomplished quickly and as soon as the disconnect with 
positive clinical outcomes is definitively identified. We noted that 
this proposal is consistent with our current practice. For example, 
previously we retired the Glaucoma Screening measure for HEDIS 2015 
after the U.S. Preventive Services Task Force concluded that the 
clinical evidence is insufficient to assess the balance of benefits and 
harms of screening for glaucoma in adults.
    In the proposed rule, we also explained how we currently review 
measures continually to ensure that the measure remains sufficiently 
reliable such that it is appropriate to continue use of the measure in 
the Star Ratings. We proposed, at paragraph (e)(1)(ii), authority to 
subregulatorily remove measures that show low statistical reliability 
so as to move swiftly to ensure the validity and reliability of the 
Star Ratings, even at the measure level. We explained that we would 
continue to analyze measures to determine if measure scores are 
``topped out'' (that is, showing high performance across all contracts 
decreasing the variability across contracts and making the measure 
unreliable) so as to inform our decision that the measure has low 
reliability. Although some measures may show uniform high performance 
across contracts and little variation between them, we noted we seek 
evidence of the stability of such high performance, and we noted we 
want to balance how critical the measures are to improving care, the 
importance of not creating incentives for a decline in performance 
after the measures transition out of the Star Ratings, and the 
availability of alternative related measures. If, for example, 
performance in a given measure has just improved across all contracts, 
or if no other measures capture a key focus in Star Ratings, a ``topped 
out'' measure with lower reliability may be retained in Star Ratings. 
Under our proposal to be codified at paragraph (e)(2), we would 
announce application of this rule through the Call Letter in advance of 
the measurement period. Below, we summarize the comments we received on 
adding, updating, and removing measures, and provide our responses and 
final decisions.
    Comment: Commenters agreed with the criteria CMS proposed to select 
new measures for the Star Ratings program. Commenters also agreed with 
the proposed measure categories (the measure categories used to assign 
weights to measures as noted in Sec. Sec.  422.166(e) and 423.186(e)), 
though a few commenters asked CMS to include more outcome measures. A 
few commenters also requested that measures be claims-based and not 
based on medical chart review.
    Response: CMS appreciates the support for our criteria for 
selecting new measures. CMS agrees with the desire to add more outcome 
measures to the Star Ratings program and welcomes all suggestions 
(submitted through the annual Call Letter process) for outcome measures 
to include in the Star Ratings program. We realize that medical chart 
review is burdensome and we are continuing to look at ways to minimize 
chart review measures. For example, CMS is exploring whether using 
encounter data for quality measurement would minimize burden for plans 
while resulting in equally accurate and appropriate reflections of 
performance and quality.
    Comment: The majority of commenters agreed with CMS' proposal for 
selecting new measures, announcing and soliciting feedback on new 
measures, finalizing new measures through rulemaking, reporting new 
measures on the display page for a minimum of 2 years prior to becoming 
a Star Rating measure, and keeping new measures on the display page if 
CMS finds reliability or validity issues with the measure 
specifications. Supporters of these proposals noted that the 
introduction of new measures through rulemaking allows greater lead 
time for plans to incorporate new measures, supports stability in the 
Star Rating program, maximizes stakeholder input, and provides 
additional transparency in the Star Ratings selection process. 
Commenters mentioned that increased lead time for the introduction of 
new measures is important especially in any payment program. Commenters 
noted the need for plans to have sufficient time to allocate resources, 
make changes to operations, adjust supporting information systems, and 
plan any specialized educational materials and events. A commenter 
suggested that new measures remain on the display page for 3 years 
which would allow plans to develop internal processes for quality 
measurement and improvement, which the commenter suggests would lead to 
improved health outcomes for beneficiaries; another commenter expressed 
the opinion that reporting a new measure on the display page for 2 
years is too long. Commenters who expressed concern that the time on

[[Page 16536]]

display was too long or suggested exceptions to allow for shorter times 
on display both referred to the need to reflect changes in clinical 
standards and to respond to public health urgencies.
    Response: CMS appreciates receiving feedback on the proposed policy 
to introduce new measures into the Star Ratings program through 
rulemaking. We acknowledge that there is some desire and policy 
rationale to keep measures on the display page for longer than 2 years, 
but CMS is trying to balance the need to introduce new measures in a 
timely manner with giving sponsors sufficient lead time for the 
introduction of new measures. We believe that a 2 year period provides 
the appropriate balance.
    Comment: Some commenters opposed the requirement to propose new 
measures through rulemaking rather than continuing to announce new 
measures through the Call Letter process. The commenters cited the long 
lag between the time measures are developed/approved and the time they 
are included in the Star Ratings, and requested a more expedited 
approach for the inclusion of new measures. Commenters noted that 
adding more lead time would stifle the adoption of new quality measures 
aligned with the latest innovative advances in medicine and technology 
and, thus, prevent Star Rating measures from reflecting the latest 
treatment guidelines and current standards of care. Further, commenters 
mentioned introducing new measures through rulemaking could 
unnecessarily delay implementation of measures needed to address 
clinical area gaps, preventable safety issues, emerging public health 
concerns, and the adoption of evidence-based measures. As a result, 
commenters believed CMS' ability to incentivize improvements in the 
quality of care for Medicare beneficiaries would decrease. A few 
commenters suggested that, if CMS does implement the rulemaking process 
for the introduction of new measures, CMS should consider granting 
exceptions in circumstances in which there are urgent public health and 
patient safety issues to be addressed through quality measures.
    Response: CMS recognizes that introducing new measures through 
rulemaking will make the process longer than CMS' former process of 
introducing new measures through the Call Letter, but we believe doing 
so balances the need for expediency with the need for greater 
transparency and stability for the ratings program. CMS also believes 
the rulemaking process adds an additional opportunity to fine tune 
measures and thus ensure greater measurement accuracy and enhanced 
stability in the Star Ratings program. We note that using rulemaking to 
adopt measures will bring the MA and Part D quality ratings system in 
line with other quality ratings systems and quality data collection 
programs that are used for Medicare payment. We understand the desire 
to have measures that address public health concerns adopted quickly in 
the Star Ratings program. CMS is committed to implementing these types 
of measures as quickly as possible so they can at least be publicly 
reported on the display page prior to being a Star Ratings measure.
    Comment: A few commenters requested that new measures be fully 
defined, tested, and validated by measure stewards prior to being 
considered for Star Ratings, even for CMS developed measures. A 
commenter requested that CMS adopt only measures which have been NQF 
endorsed, publicly reported by NCQA (or the measure steward) for at 
least one measurement period, and reported on the CMS display page for 
at least one measurement period. The commenter also recommended that 
CMS not report new (first year) measures on the display page.
    Response: CMS agrees that measures need to be fully defined, tested 
and validated by measure stewards before used as the basis for Medicare 
payment. Placing new measures on the display page provides transparency 
about CMS' intention to use the measure in the future as part of Star 
Ratings and an opportunity for sponsors to see their scores and 
performance before the measure is used in the Star Ratings. The display 
measures are not assigned Star Ratings or used in the development of 
measure, domain, summary, or overall Star Ratings, so there are no 
payment consequences. Retaining new measures on the display for two 
years gives CMS additional opportunities to identify any data issues 
prior to the measures being included in the Star Ratings program. CMS 
will use endorsed measures as they are available. For some areas which 
CMS judges to be important for the Star Ratings program, endorsed 
measures may not be available. CMS emphasizes that if reliability 
issues with a display measure are identified, the regulations proposed 
and finalized in this rule at Sec. Sec.  422.164(c)(4) and 
423.184(c)(4) prevent the measure from moving to a Star Ratings 
measure. Although a number of commenters to the proposed rule were 
concerned about the rulemaking process preventing CMS from quickly 
responding to public health and patient safety issues, CMS believes 
that reporting new measures as soon as possible on the display page 
will addresses these concerns.
    Comment: The majority of commenters agreed with the process for 
updating existing measures.
    Response: We appreciate the support for the process for updating 
existing measures.
    Comment: Some commenters objected to the proposal for updating 
measures through rulemaking because of the delay between the time 
measures are updated/approved and the time they are re-introduced into 
the Star Ratings program. These commenters requested a more expedited 
approach for updating measures. Most commenters supported CMS in its 
proposal to codify a non-exhaustive list for identifying non-
substantive measure updates. Some commenters requested additional 
information on how the determination is made as to whether a change is 
substantive versus non-substantive. A few commenters wanted a more 
exhaustive list of what are considered non-substantive changes.
    Some commenters expressed the opinion that all measure updates, 
even non-substantive changes, should be announced in advance of the 
measurement period. In addition, a few commenters expressed the opinion 
that all measure updates, whether substantive or non-substantive, 
should be subject to rulemaking. These commenters noted some of the 
same concerns expressed for supporting the addition of new measures 
through rulemaking rather than through the Call Letter process. These 
concerns included allowing plans greater lead time to incorporate 
updates, have sufficient time to allocate resources to incorporate 
updates, make changes to operations, adjust supporting information 
systems, and plan any specialized educational materials and events. A 
commenter, however, expressed the opinion that no measure updates, 
substantive or non-substantive, should be required to go through 
rulemaking, because this would lead to unnecessary gaps in measurement 
for critically important issues.
    Response: CMS appreciates the comments we received on our proposal 
for updating measures. Although there is some disagreement among 
commenters on whether and which updates should go through rulemaking, 
we believe our proposal balances the commenters' concerns by only 
requiring substantive measure updates to go through the rulemaking 
process. Non-substantive updates, such as coding updates, which are not 
significant changes to the measure specifications would continue to be 
announced

[[Page 16537]]

through the Call Letter process. CMS does not have authority to 
determine or direct when measure stewards update measure 
specifications. If non-substantive measure specifications are made 
during the measurement period, CMS believes it is of value to 
incorporate those measure specification updates in that year's Star 
Ratings measures. Non-substantive updates are most often minor code 
updates and are not significant changes to the measure specifications. 
CMS proposed and is finalizing in this rule a comprehensive list of 
measure changes it considers non-substantive in Sec. Sec.  
422.164(d)(1) and 423.184(d)(1); we explained (above and in the 
proposed rule) the basis for our determination that these changes and 
others like them should be implemented without delay or additional 
rulemaking. The list is not exhaustive because additional situations or 
types of changes may also result in little or no change to the results 
of measurement (or generally benefit sponsoring organizations) in a 
similar way. We believe that the standard adopted here--that of non-
substantive changes--is adequately clear to provide notice to 
stakeholders and balance the competing policies identified by 
commenters. CMS encourages plans and other stakeholders to provide 
suggestions for additional non-substantive measure updates to add to 
the current list through future rulemaking.
    Comment: A few commenters expressed disagreement with the proposal 
to continue collecting a legacy measure until an updated measure has 
been on display for 2 years.
    Response: CMS appreciates comments on its proposal to keep legacy 
measures in the Star Ratings during the period when the related updated 
measure goes through rulemaking and is placed on the display page for 2 
years. We intend that a legacy measure may remain in the Star Ratings 
until the updated measure is ready to move into Star Ratings only when 
the area covered by the measure is critical to reflecting whether plans 
are providing appropriate care or for a similar reason that the 
information provided by the legacy measure is important to the Star 
Ratings.
    Comment: There was general agreement among commenters with CMS' 
proposed process for removing measures from the Star Ratings program 
and for announcing the removal in advance of the measurement period. 
However, some commenters did question the criteria for how CMS judges 
measures to be `topped out' or have low statistical reliability.
    Response: CMS appreciates the overall support for its proposal for 
removing measures from the Star Ratings program. Measure scores are 
determined to be `topped out' when they show high performance and 
little variability across contracts, making the measure statistically 
unreliable. However, although some measures may show uniform high 
performance across contracts and little variation between them, CMS 
needs to balance these concerns with how critical the measures are to 
improving care, the importance of not creating incentives for a decline 
in performance after the measures transition out of the Star Ratings, 
and the availability of alternative related measures which address the 
specific clinical concerns.
    For the reasons set forth in the proposed rule and our responses to 
the related comments summarized earlier, we are finalizing the 
provisions related to the adoption, update, and removal of measures as 
proposed at paragraphs (c), (d), and (e) of Sec. Sec.  422.164 and 
423.184 with a minor modification to add the phrase ``nationally 
endorsed'' to Sec.  422.164(c)(1) so that the regulation text is 
identical to the parallel Part D provision at Sec.  423.184(c)(1).
i. Measure Set for Performance Periods Beginning on or After January 1, 
2019
    We proposed the measures included in Table 2 to be collected for 
performance periods beginning on or after January 1, 2019 for the 2021 
Part C and D Star Ratings. The CAHPS measure specification, including 
case-mix adjustment, is described in the Technical Notes and at ma-pdpcahps.org. The HOS measure specification, including case-mix 
adjustment, is described at (http://hosonline.org/globalassets/hos-online/survey-results/hos_casemix_coefficient_tables_c17.pdf). These 
specifications are part of our proposal.
    As indicated in the proposed rule, CMS will not codify a list of 
measures and specifications in regulation text in light of the regular 
updates and revisions contemplated by the rules we have finalized at 
paragraphs (c), (d) and (e) of Sec. Sec.  422.164 and 423.184. We 
would, as finalized in Sec. Sec.  422.164(a) and 423.184(a), issue 
annually the full list of measures in the Technical Notes for each 
year's Star Ratings.
BILLING CODE 4120-01-P

[[Page 16538]]

[GRAPHIC] [TIFF OMITTED] TR16AP18.001


[[Page 16539]]


[GRAPHIC] [TIFF OMITTED] TR16AP18.002


[[Page 16540]]


[GRAPHIC] [TIFF OMITTED] TR16AP18.003


[[Page 16541]]


[GRAPHIC] [TIFF OMITTED] TR16AP18.004


[[Page 16542]]


[GRAPHIC] [TIFF OMITTED] TR16AP18.005


[[Page 16543]]


[GRAPHIC] [TIFF OMITTED] TR16AP18.006


[[Page 16544]]


[GRAPHIC] [TIFF OMITTED] TR16AP18.007


[[Page 16545]]


[GRAPHIC] [TIFF OMITTED] TR16AP18.008


[[Page 16546]]


[GRAPHIC] [TIFF OMITTED] TR16AP18.009

BILLING CODE 4120-01-C
    We summarize the comments received on the proposed measures and 
respond to them by measure in Table 3C for the Part C measures, for 
performance

[[Page 16547]]

periods beginning on or after January 1, 2019.

                        Table 3C--Part C Measures
------------------------------------------------------------------------
              Measure
------------------------------------------------------------------------
Breast Cancer Screening (BCS).....  Comment: A commenter expressed
                                     concerns that due to physical and
                                     mental limitations, all permanently
                                     institutionalized beneficiaries,
                                     including those under age 65,
                                     should be excluded from the Breast
                                     Cancer Screening measure. This
                                     commenter suggested that rather
                                     than undergo a mammogram, an
                                     alternative screening option would
                                     be an Automated Breast Ultrasound
                                     (ABUS).
                                    Response: CMS appreciates this
                                     feedback. CMS has shared comments
                                     received on this measure with NCQA,
                                     the measure steward, for
                                     consideration when their advisory
                                     panels re-evaluate this measures,
                                     as part of the standard HEDIS
                                     process.
Colorectal Cancer Screening (COL).  Comment: CMS received no comments on
                                     this measure.
Annual Flu Vaccine................  Comment: CMS received a number of
                                     general comments on CAHPS measures.
                                    Response: CMS appreciates the
                                     feedback on the CAHPS measures.
                                     Since the comments on CAHPS
                                     measures were mostly not measure
                                     specific, please see the CAHPS
                                     summary of comments received as
                                     well as CMS responses following the
                                     Parts C and D Measure Tables.
                                    Comment: CMS received one comment
                                     that the annual flu vaccine measure
                                     should use claims data as they are
                                     more reliable. Another commenter
                                     stated that beneficiaries in Puerto
                                     Rico are reluctant to vaccinate
                                     against the flu which unfairly
                                     impacts plans in Puerto Rico, and
                                     that asking beneficiaries to
                                     remember when they received a flu
                                     shot is a burden on them.
                                    Response: The flu item is a HEDIS
                                     measure collected through the CAHPS
                                     survey. Flu shot information is
                                     collected through a survey since
                                     there are a variety of places where
                                     people can get flu shots and the
                                     plan may not have a record of a flu
                                     shot in their administrative data
                                     depending on where the flu shot was
                                     received. We note that CMS applies
                                     standards of reliability to CAHPS
                                     results, directly and through
                                     significance testing. The item asks
                                     whether respondents received a flu
                                     shot since July in order to reflect
                                     the timeframe when beneficiaries
                                     typically receive flu shots. This
                                     is a process measure, and CMS does
                                     not adjust process measures for
                                     beneficiary refusals to avoid
                                     biasing the data.
HOS Measures:
    Improving or Maintaining        Comment: CMS received a number of
     Physical Health.                general comments on HOS measures.
                                    Response: CMS appreciates the
                                     feedback on the HOS measures. Since
                                     the comments on HOS measures were
                                     mostly not measure specific, please
                                     see the HOS summary of comments
                                     received as well as CMS responses
                                     following the Parts C and D Measure
                                     Tables.
Improving or Maintaining Mental     Comment: Several commenters
 Health.                             suggested the HOS measures
                                     Improving or Maintaining Physical
                                     Health and Improving or Maintaining
                                     Mental Health fail to consider the
                                     natural aging process or
                                     accommodate vulnerable
                                     beneficiaries and those with
                                     degenerative or progressive
                                     diseases. They pointed out that as
                                     time passes, patients are more
                                     prone to experience certain health
                                     deterioration and argued that
                                     changes in status--positive or
                                     negative--should not be attributed
                                     to the actions of the health plan.
                                     They again suggested that CMS drop
                                     the two year look-back design of
                                     the survey.
                                    Response: HOS yields two patient-
                                     reported outcome measures of change
                                     in global functioning, by using 2-
                                     year change in scores on the
                                     Physical Component Score (PCS) and
                                     the Mental Component Score (MCS),
                                     both of which come from the
                                     Veterans RAND 12-Item Health Survey
                                     (VR-12) portion of the larger
                                     survey. HOS assesses health
                                     outcomes for randomly selected
                                     beneficiaries from each health plan
                                     over a two-year period by using
                                     baseline measurement and a two-year
                                     follow up. In general, functional
                                     health status is expected to
                                     decline over time in older age
                                     groups, mental health status is
                                     not, and the presence of chronic
                                     conditions is associated with
                                     declines in both *.\37\
                                     Longitudinal HOS outcomes
                                     (including death) are adjusted for
                                     baseline age and other well studied
                                     risk factors, including chronic
                                     conditions, baseline health status,
                                     and socio-demographic
                                     characteristics that include
                                     gender, race, ethnicity, income,
                                     education, marital status, Medicaid
                                     status, SSI eligibility, and
                                     homeowner status. Because each
                                     beneficiary's follow up score is
                                     compared to their baseline score
                                     and adjusted for these risk
                                     factors, each beneficiary serves as
                                     his/her own control. CMS recognizes
                                     that Physical Component Summary
                                     (PCS) and Mental Component Summary
                                     (MCS) may decline over time and
                                     that health maintenance, rather
                                     than improvement, is a more
                                     realistic clinical goal for many
                                     older adults. Therefore, MA
                                     Organizations are asked to improve
                                     or maintain the physical and mental
                                     health of their members. Change
                                     scores are constructed and the
                                     results compare actual to expected
                                     changes in physical and mental
                                     health.
Monitoring Physical Activity (PAO)  Comment: CMS received no comments on
                                     this measure.
Adult BMI Assessment (ABA)........  Comment: CMS received one comment
                                     suggesting the BMI measure be
                                     removed from the Star Ratings
                                     program due to the commenter
                                     believing the measure to be `topped
                                     out.' A measure is considered
                                     `topped out' when it shows high
                                     performance across all contracts
                                     decreasing the variability across
                                     contracts and making the measure
                                     unreliable.
                                    Response: CMS appreciates the
                                     feedback; however, from a review of
                                     the Star Ratings data for this
                                     measure, there are many contracts
                                     rated below 4 stars. There have
                                     been significant increases in
                                     ratings for this measure in recent
                                     years so CMS is carefully
                                     monitoring this measure to see if
                                     it should be proposed for
                                     retirement from the Star Ratings in
                                     the future.
Special Needs Plan (SNP) Care       Comment: A commenter recommended
 Management.                         that the SNP Care Management
                                     measure be retired until clear
                                     technical guidance on the measure
                                     specifications can be issued by the
                                     agency and if the measure is
                                     reintroduced, the cut points should
                                     be stratified based on SNP type
                                     (for example, C-SNP, D-SNP), since
                                     the commenter believes various SNP
                                     types have different outcomes on
                                     this measure.

[[Page 16548]]

 
                                    Response: There are no upcoming
                                     clarifications or changes to this
                                     measure specifications for the 2021
                                     Star Ratings. Note that the SNP
                                     care management measure is
                                     collected at the PBP level and the
                                     requirement to complete a timely
                                     HRA for every plan member (which is
                                     the performance metric measured)
                                     applies to all SNP types. Sponsors
                                     are reminded that as part of the
                                     data validation process of plan-
                                     reported data, a reviewer must
                                     submit and review draft findings to
                                     the sponsor prior to submission via
                                     HPMS. Once data validation findings
                                     are submitted to HPMS, sponsors may
                                     formally submit their disagreement
                                     to CMS if necessary.
                                    Comment: A commenter suggested that
                                     some Star Rating measures are
                                     driven primarily by member
                                     outreach. As such, some plans with
                                     large dual-eligible populations are
                                     disproportionately negatively
                                     impacted by members who are more
                                     transient and with frequent address
                                     and phone number changes that
                                     directly result in fewer successful
                                     contacts and lower engagement. For
                                     outreach-driven measures, the
                                     commenter urges CMS to exclude
                                     members who were unreachable after
                                     a justifiable number of documented
                                     good faith attempts.
                                    Response: The requirement to
                                     complete a timely HRA for every
                                     plan member (which is the
                                     performance metric measured)
                                     applies to all SNP types and is
                                     regulatory. There are no upcoming
                                     specification changes that will
                                     affect this measure for the 2021
                                     Star Ratings. Note that plans may
                                     report when members are unreachable
                                     after documented attempts and when
                                     members refuse to complete the HRA,
                                     but those data are not used in
                                     calculating this measure.
SNP measures:
    Care for Older Adults (COA)--   Comment: A commenter expressed
     Medication Review, Care for     concerns about the varying
     Older Adults (COA)--            performance on SNP measures based
     Functional Status Assessment,   on the SNP type stating that the
     Care for Older Adults (COA)--   performance on these measures is
     Pain Assessment.                heavily biased related to type of
                                     SNP plan, rather than indicative of
                                     plan quality.
                                    Response: These measures are
                                     indicators of high quality care for
                                     all plans that focus on special
                                     needs populations. However, for
                                     HEDIS 2019, NCQA is considering
                                     modifications to these measures, to
                                     broaden the denominators to all
                                     patients with multiple chronic
                                     conditions. CMS will keep
                                     considerations in mind that
                                     measures not be primarily driven by
                                     plan type, rather than differences
                                     in quality of care.
Osteoporosis Management in Women    Comment: CMS received comments that
 who had a Fracture (OMW).           there should be different
                                     exclusions for some health
                                     conditions including osteoporosis
                                     because, for some patients, the
                                     treatments identified in the
                                     measure specification (that is for
                                     compliance) are not medically
                                     appropriate. Commenters noted that
                                     many challenges exist in treating
                                     and screening certain health
                                     conditions for patients with
                                     advanced illness. A commenter
                                     suggested that the Star Ratings
                                     clinical metrics may not be sound
                                     for frail patients with advanced
                                     illness.
                                    Response: CMS appreciates receiving
                                     feedback on this measure. For HEDIS
                                     2019, NCQA is examining potential
                                     cross-cutting exclusions for those
                                     with advanced illness from selected
                                     HEDIS[supreg] measures, including
                                     the Osteoporosis Management in
                                     Women Who Had a Fracture measure.
                                     Proposed changes to implement
                                     advanced illness exclusions will be
                                     posted for the HEDIS 2019 public
                                     comment period in February 2018.
                                     Please see additional comments
                                     related to Patients with Advanced
                                     Illness below.
Diabetes Care (CDC)--Eye Exam.....  Comment: CMS received no comments on
                                     this measure.
Diabetes Care (CDC)-- Kidney        Comment: CMS received a few comments
 Disease Monitoring.                 suggesting the Diabetes Care--
                                     Kidney Disease Monitoring measure
                                     be removed from the Star Ratings
                                     program due to the commenters
                                     belief the measure is `topped out.'
                                     A measure is considered `topped
                                     out' when it shows high performance
                                     across all contracts decreasing the
                                     variability across contracts and
                                     making the measure unreliable.
                                    Response: CMS appreciates the
                                     feedback, however, from a review of
                                     the Star Ratings for this measure,
                                     there are many plans rated below 4
                                     stars. A. As noted above in this
                                     preamble, among other
                                     considerations, CMS wants to
                                     balance how critical measures are
                                     to improving care and the
                                     availability of alternative related
                                     measures. If, for example, no other
                                     measures captures a key focus in
                                     Star Ratings, a `topped out'
                                     measure with lower reliability may
                                     be retained in Star Ratings.
                                     Currently, there are no alternative
                                     kidney disease monitoring measures
                                     appropriate for MA Star Ratings.
Diabetes Care (CDC)--Blood Sugar    Comment: CMS received no comments on
 Controlled.                         this measure.
Controlling Blood Pressure (CBP)..  Comment: CMS received a
                                     recommendation that in alignment
                                     with current clinical practice
                                     guidelines, ambulatory and home
                                     blood pressure readings that are
                                     documented in the treating
                                     provider's medical record be
                                     considered acceptable for the
                                     purposes of assessing the efficacy
                                     and appropriateness of a
                                     clinician's treatment plan.
                                    Response: CMS appreciates feedback
                                     on this measure. NCQA is currently
                                     reevaluating the Controlling High
                                     Blood Pressure measure and
                                     proposing to allow for readings
                                     taken from remote monitoring
                                     devices that transmit results
                                     directly to the provider. Details
                                     on this potential change will be
                                     posted for the HEDIS 2019 public
                                     comment period in February 2018.
Rheumatoid Arthritis Management     Comment: CMS received comments that
 (ART).                              evidence of treatment for
                                     rheumatoid arthritis not limited to
                                     disease-modifying anti-rheumatic
                                     drugs (DMARD) should be considered
                                     for compliance (that is, added to
                                     the numerator for the measure).
                                     Commenters noted that some patients
                                     have limited tolerance for DMARDs
                                     along with a much higher rate of
                                     serious adverse medication effects,
                                     particularly serious infections.

[[Page 16549]]

 
                                    Response: CMS appreciates receiving
                                     feedback on this measure. For HEDIS
                                     2019, NCQA is examining potential
                                     cross-cutting exclusions for those
                                     with advanced illness from selected
                                     HEDIS[supreg] measures, including
                                     the Disease-Modifying Anti-
                                     Rheumatic Drug Therapy for
                                     Rheumatoid Arthritis measures.
                                     Proposed changes to implement
                                     advanced illness exclusions will be
                                     posted for the HEDIS 2019 public
                                     comment period in February 2018.
                                     Please see additional comments
                                     related to Patients with Advanced
                                     Illness below. We understand from
                                     public statements that NCQA plans
                                     to reevaluate the Rheumatoid
                                     Arthritis Management measure and
                                     review the evidence for rheumatoid
                                     arthritis treatment with their
                                     advisory panels.
Reducing the Risk of Falling (FRM)  Comment: CMS received no comments on
                                     this measure.
Improving Bladder Control (MUI)...  Comment: CMS received no comments on
                                     this measure.
Medication Reconciliation Post-     Comment: CMS received no comments on
 Discharge (MRP).                    this measure.
Plan All-Cause Readmissions (PCR).  Comment: A commenter suggested that
                                     in order to provide MA
                                     organizations with greater
                                     visibility into plan performance,
                                     CMS should work with the NCQA to
                                     eliminate the calculation whereby a
                                     national average observed rate is
                                     multiplied by the observed to
                                     expected ratio of readmissions for
                                     Plan All-Cause Readmissions. A
                                     commenter noted that NCQA has
                                     announced in early 2018 substantive
                                     changes in the Plan All-Cause
                                     Readmissions measure.
                                    Response: CMS appreciated feedback
                                     on this measure. The calculation
                                     mentioned that uses the observed
                                     readmission rate divided by the
                                     expected readmission rate for a
                                     contract multiplied by the national
                                     average is the process to calculate
                                     the case-mix adjusted contract
                                     rate. A case-mix adjusted rate is
                                     used to ensure that the comparisons
                                     between contracts is fair and
                                     meaningful. It takes into account
                                     how sick patients were when they
                                     went into the hospital the first
                                     time. CMS will discuss with NCQA
                                     the need to better explain the
                                     calculations involved in the
                                     reporting of the measure.
                                    CMS decision: In that NCQA is
                                     planning to make significant
                                     changes to the Plan All-Cause
                                     Readmissions measure (changes to be
                                     published in 2018 and applied in
                                     measurement year 2019) CMS is not
                                     finalizing this as part of the
                                     measure set for the 2019
                                     performance period and the 2021
                                     Ratings. CMS is finalizing this as
                                     a display measure and consistent
                                     with Sec.   422.164(d)(2) will
                                     include this measure on the display
                                     page for 2 years.
Getting Needed Care...............  Comment: CMS received a number of
                                     general comments on CAHPS measures.
                                    Response: CMS appreciates the
                                     feedback on the CAHPS measures.
                                     Since the comments on CAHPS
                                     measures were mostly not measure
                                     specific, please see the CAHPS
                                     summary of comments received as
                                     well as CMS responses following the
                                     Parts C and D Measure Tables.
Getting Appointments and Care       Comment: CMS received many general
 Quickly.                            and specific comments on CAHPS
                                     measures.
                                    Response: CMS appreciates the
                                     feedback on the CAHPS measures.
                                     Since the comments on CAHPS
                                     measures were not always measure
                                     specific, please see the CAHPS
                                     summary of comments received as
                                     well as CMS responses following the
                                     Parts C and D Measure Tables.
                                    Comment: CMS received one comment
                                     that this composite is unfair to
                                     plans in Puerto Rico because
                                     beneficiaries in Puerto Rico are
                                     not necessarily used to having a
                                     specific appointment time.
                                    Response: We thank the commenter for
                                     this comment. We have conducted
                                     some exploratory work related to
                                     this topic and may propose changes
                                     in the future after consulting with
                                     AHRQ.
Customer Service..................  Comment: CMS received a number of
                                     general comments on CAHPS measures.
                                    Response: CMS appreciates the
                                     feedback on the CAHPS measures.
                                     Since the comments on CAHPS
                                     measures were mostly not measure
                                     specific, please see the CAHPS
                                     summary of comments received as
                                     well as CMS responses following the
                                     Parts C and D Measure Tables.
Rating of Health Care Quality.....  Comment: CMS received a number of
                                     general comments on CAHPS measures.
                                    Response: CMS appreciates the
                                     feedback on the CAHPS measures.
                                     Since the comments on CAHPS
                                     measures were mostly not measure
                                     specific, please see the CAHPS
                                     summary of comments received as
                                     well as CMS responses following the
                                     Parts C and D Measure Tables.
Rating of Health Plan.............  Comment: CMS received a number of
                                     general comments on CAHPS measures.
                                    Response: CMS appreciates the
                                     feedback on the CAHPS measures.
                                     Since the comments on CAHPS
                                     measures were mostly not measure
                                     specific, please see the CAHPS
                                     summary of comments received as
                                     well as CMS responses following the
                                     Parts C and D Measure Tables.
Care Coordination.................  Comment: CMS received many general
                                     and specific comments on CAHPS
                                     measures.
                                    Response: CMS appreciates the
                                     feedback on the CAHPS measures.
                                     Since the comments on CAHPS
                                     measures were not always measure
                                     specific, please see the CAHPS
                                     summary of comments received as
                                     well as CMS responses following the
                                     Parts C and D Measure Tables.
Complaints about the Health Plan..  Comment: A commenter recommended
                                     creating an excluded category/sub-
                                     category for complaints related to
                                     CMS/SSA system/enrollment issues or
                                     limitations which would effectively
                                     remove complaints of that type from
                                     this measure.
                                    Response: Data exchanges between CMS
                                     and SSA occur regularly and mostly
                                     without incident. When issues
                                     occur, CMS often looks to plan
                                     sponsors to communicate accordingly
                                     to their members and utilize CMS
                                     resources, such as the MA-PD help
                                     desk, to help address their matter
                                     without referral to CMS and
                                     generation of complaints. CMS is
                                     not instituting such a category/sub-
                                     category at this time. Plan
                                     Sponsors should continue to work
                                     alongside their CMS caseworker as
                                     appropriate to provide assistance.
                                    Comment: A few commenters requested
                                     updates to the CMS CTM standard
                                     operating procedures (SOP). There
                                     was a request to provide
                                     instructions for plans to return
                                     issues (either as a CMS issue or as
                                     a closed complaint) determined by 1-
                                     800-Medicare to be errors. Another
                                     request was that complaints found
                                     to not be the fault of the plan be
                                     considered CMS issues, or
                                     reassigned to another entity.

[[Page 16550]]

 
                                    Response: CMS regularly utilizes
                                     feedback from plans and other
                                     stakeholders to identify
                                     opportunities for continuous
                                     improvement of CMS resources such
                                     as 1-800-Medicare. Due to the
                                     volume of CTM complaints received
                                     annually, CMS cannot investigate
                                     for individual errors. CMS expects
                                     such matters to be rare, and any
                                     impact on plans to be evenly
                                     distributed. Plan Sponsors should
                                     not seek recategorization of
                                     marketing complaints because, as a
                                     result of plan investigation, they
                                     have determined the allegation is
                                     unfounded. However, if a marketing
                                     complaint has been misclassified,
                                     and the narrative reflects that the
                                     alleged misrepresentation occurred
                                     by a Call Center representative,
                                     SHIP, etc., then a Plan Request to
                                     make the complaint a ``CMS Issue''
                                     is appropriate. CMS appreciates the
                                     feedback and will include
                                     additional language in the next
                                     version of the CTM Plan SOP.
                                    Comment: A commenter suggested that
                                     CMS create an excluded category
                                     intended for cases that are
                                     educational and/or are referrals to
                                     the contract.
                                    Response: It is not CMS' intention
                                     for the CTM to communicate plan
                                     information or simply provide
                                     education.
                                    Comment: A commenter stated concerns
                                     that duplicate complaints count
                                     against plan sponsors.
                                    Response: CMS' CTM SOP includes
                                     procedures for the removal of
                                     duplicate complaints with the same
                                     complaint identification numbers,
                                     so there is no impact on plan
                                     sponsors. CMS has taken numerous
                                     steps over the years to reduce the
                                     instances of this occurring and
                                     expect that plan sponsors have
                                     noticed significant improvement in
                                     this area. If a beneficiary's issue
                                     persists or is not be resolved by a
                                     plan, multiple complaints may be
                                     entered into the CTM. These
                                     complaints are not duplicative, but
                                     reflect unresolved or similar
                                     issues. CMS does not support
                                     removing such complaints. Inclusion
                                     of these complaints effectively
                                     rewards plan sponsors who are
                                     prompt with acknowledging and
                                     resolving complaints, and provide
                                     excellent customer service to
                                     beneficiaries.
                                    Comment: A commenter requested clear
                                     processes for when the assignment/
                                     reassignment date should be reset
                                     by CMS caseworkers, so that plan
                                     sponsors can better strategize
                                     their actions.
                                    Response: Assignment/reassignment
                                     date by CMS caseworkers is a topic
                                     outside the scope of this rule.
Members Choosing to Leave the Plan  Comment: A couple of commenters
                                     suggested that the disenrollment
                                     rate does not reflect the plan's
                                     quality and the beneficiary
                                     experience. They note that the
                                     disenrollment rate is impacted by
                                     the pricing and coverage strategies
                                     of the contract. Among those
                                     commenters dissatisfied with what
                                     the disenrollment rate reflects and
                                     does not reflect, a commenter
                                     suggested that this measure be
                                     moved to the display page.
                                    Response: CMS is statutorily
                                     required to report voluntary
                                     disenrollment rates as part of the
                                     Balanced Budget Act of 1997.
                                     Disenrollment rates are a strong
                                     measure of a beneficiary's
                                     satisfaction with a contract.
                                     Beneficiaries who are interested in
                                     seeing why enrollees voluntarily
                                     leave a contract can obtain this
                                     information as a drill down to the
                                     disenrollment rates on Medicare
                                     Plan Finder. CMS respectfully
                                     disagrees that pricing strategies
                                     and the coverage provided by the
                                     contract should not be considered
                                     in assessing the quality and
                                     performance of contracts since they
                                     have a direct impact on access to
                                     services.
                                    Comment: A commenter suggests that
                                     CMS conduct additional analyses to
                                     see if the disenrollment rates
                                     should be adjusted by the
                                     proportion of SNP members.
                                    Response: CMS appreciates this
                                     comment and will analyze the data
                                     to see if any future changes are
                                     needed. Any potential changes would
                                     be subject to future rulemaking.
                                     The current Star Ratings
                                     adjustments for dual status are
                                     incorporated as part of the CAI.
Health Plan Quality Improvement...  For the summary of comments received
                                     and CMS' responses for this
                                     measure, please see section `j.
                                     Improvement Measures' of the
                                     Preamble.
Plan Makes Timely Decisions about   Comment: CMS received a comment
 Appeals.                            opposing the inclusion of
                                     dismissals in the Plan Makes Timely
                                     Decisions about Appeals measure.
                                     The commenter expressed concern
                                     that if the inclusion of dismissals
                                     is a positive factor in the
                                     measure, it would create incentives
                                     for the MA organization to increase
                                     the opportunities to enter
                                     dismissals.
                                    Response: CMS appreciates the
                                     comment about dismissals. To
                                     clarify, the measure for the 2021
                                     Star Ratings includes cases
                                     dismissed by the IRE because the
                                     plan has subsequently approved
                                     coverage/payment. In prior years,
                                     we excluded all cases dismissed/
                                     withdrawn by the IRE from this
                                     measure. The inclusion of
                                     dismissals would only apply to
                                     cases dismissed by the IRE because
                                     the plan issued an untimely but
                                     favorable decision. In other words,
                                     plans may send late Part C appeals
                                     to the IRE while simultaneously (or
                                     shortly thereafter) approving the
                                     late cases which results in the
                                     case being dismissed by the IRE,
                                     thus masking that the plans'
                                     decisions were untimely. Inclusion
                                     of cases where the plan has
                                     subsequently approved for coverage/
                                     payment that are dismissed or
                                     withdrawn at the IRE level could
                                     provide a more accurate assessment
                                     of plans' timeliness in their Part
                                     C appeals processing. Without
                                     excluding this group of dismissals,
                                     a plans' performance may be
                                     artificially improved as a result,
                                     especially if dismissals were
                                     directly related to the plans'
                                     (untimely) approvals.
                                    If an MA plan fails to provide the
                                     appellant with a reconsidered
                                     determination within the required
                                     timeframes, this failure
                                     constitutes an affirmation of its
                                     adverse organization determination,
                                     and the plan must submit the case
                                     file to the IRE for review. This
                                     new measure would more accurately
                                     reflect that MA plans are not
                                     making timely decisions. CMS does
                                     not believe this would create the
                                     incentive described by the
                                     commenter.
                                    CMS acknowledges these comments and
                                     is actively evaluating these
                                     measures and the use of the IRE
                                     data as their data source for
                                     future enhancements.
                                    Comment: CMS received a comment
                                     recommending that this measure be
                                     weighted by membership by
                                     calculating the measure similarly
                                     to the Part D Auto-Forward measure
                                     to ensure plans of all sizes are
                                     measured equally.
                                    Response: The Part C and Part D
                                     appeals systems are different, they
                                     have different rules for how
                                     appeals are handled. There are no
                                     auto-forwards in Part C and the
                                     number of late appeals examines how
                                     well the contract is processing the
                                     appeals in a timely manner.
                                     Additionally each measure has
                                     different specifications.

[[Page 16551]]

 
Reviewing Appeals Decisions.......  Please see response for Part D
                                     Appeals Upheld measure.
Call Center--Foreign Language       Comment: A few commenters
 Interpreter and TTY Availability.   recommended that CMS revise the
                                     measure's sampling methodology for
                                     volume and for volume by language
                                     (including consideration of plans
                                     with larger enrollment sizes), or
                                     revise the foreign languages and
                                     testing frequency. An additional
                                     commenter recommended that CMS
                                     adjust the foreign languages tested
                                     to the languages actually spoken in
                                     that plan's area, and mentioned
                                     that 99 percent of local residents
                                     speak Spanish in Puerto Rico. The
                                     commenter also suggested using a
                                     single, combined measure (or rate)
                                     for both Part C and D.
                                    Response: The Accuracy and
                                     Accessibility Study is performed to
                                     (1) ascertain the accuracy of
                                     responses to plan benefit questions
                                     provided by customer service
                                     representatives when calling the
                                     call center in addition to (2)
                                     testing the availability of
                                     interpreters for Limited English
                                     Proficient callers and (3) testing
                                     TTY functionality. A simple random
                                     sample method is used. To reduce
                                     the burden on a call center with
                                     multiple phone lines, we select
                                     samples across the call centers
                                     instead of the phone lines. The
                                     precision requirement of the sample
                                     size is calculated at the call
                                     center level and is based on the
                                     question response accuracy rates
                                     obtained from the accuracy survey,
                                     and the rate of completed calls
                                     made through Limited English
                                     Proficiency (LEP) accommodations
                                     and TTY services. This methodology
                                     was chosen by CMS, in part, because
                                     the accuracy of the information
                                     provided to a caller in response to
                                     specific benefits questions should
                                     not be impacted by enrollment size
                                     or physical call center location.
                                     If contract enrollment size is
                                     positively correlated with higher
                                     variability and wider margins of
                                     error in these key metrics of this
                                     study, CMS would expect to see
                                     contracts with higher enrollments
                                     having the key metrics closer to 50
                                     percent than the contracts with
                                     lower enrollments. We have not
                                     observed that in the data and will
                                     therefore continue to use the
                                     methodology as designed. Call
                                     centers using more or fewer
                                     representatives are held to the
                                     same expectation that the
                                     information provided to callers is
                                     accurate.
                                    Foreign language testing was never
                                     intended to be proportionate to the
                                     demographics of any contract. Plan
                                     sponsors are required to provide an
                                     interpreter for any caller speaking
                                     a foreign language, and CMS seeks
                                     to ensure that more vulnerable
                                     populations have equal access to
                                     interpreters. Rather than test all
                                     foreign languages which would be
                                     overly burdensome and costly, CMS
                                     selects 6 foreign languages from
                                     among the top 10 most frequently
                                     spoken languages in the U.S.,
                                     according to the Office for Civil
                                     Rights (which makes its selections
                                     from U.S. Census Data). The number
                                     of calls by foreign language is
                                     equally divided and randomly
                                     assigned to each call center across
                                     the biweekly calling schedule. The
                                     number received by the call center
                                     is dependent upon each call
                                     successfully reaching the call
                                     center (for example, disconnects in
                                     an IVR or other factors will impact
                                     the ability of the call to reach a
                                     representative). Internal analysis
                                     across all plans shows that the
                                     methodology is sound and CMS has
                                     confidence in the data.
                                    When testing in Puerto Rico, Spanish
                                     is the native language and English
                                     is treated as a foreign language.
                                     Because some of the accuracy calls
                                     are placed in the native language
                                     in addition to foreign language
                                     testing, Spanish calls are placed
                                     at a higher volume for plans in
                                     Puerto Rico.
                                    By design, the Accuracy and
                                     Accessibility Study schedules and
                                     places calls to phone numbers that
                                     may or may not be the same for Part
                                     C and Part D. Also, the study is
                                     conducted at the call center level
                                     (not the phone number level), and
                                     not all plans use the same call
                                     center for Part C as for Part D.
                                     Finally, the accuracy questions
                                     used in this study either relate to
                                     Part C benefit questions or to Part
                                     D benefit questions. Because the
                                     questions are different for each,
                                     CMS believes performance should be
                                     measured separately for the Part C
                                     and Part D programs.
Statin Therapy for Patients with    Comment: CMS received two comments
 Cardiovascular Disease (SPC).       seeking clarification regarding the
                                     categorization and weighting
                                     discrepancies between the Part C
                                     and Part D statin measures. Two
                                     organizations recommended
                                     classifying both SPC and SUPD as
                                     process measures with a weight of
                                     one.
                                    Response: CMS appreciates the
                                     feedback. The Part C Statin Therapy
                                     for Patients with Cardiovascular
                                     Disease (SPC) measure is the
                                     percent of plan members (males 21-
                                     75 years of age and females 40-75
                                     years of age) who were identified
                                     as having clinical atherosclerotic
                                     cardiovascular disease (ASCVD) and
                                     were dispensed at least one high or
                                     moderate-intensity statin
                                     medication. This Part C measure
                                     focuses on patients who were
                                     dispensed one prescription and
                                     whether the patient filled the
                                     medication at least once.
                                     Therefore, it is a process measure.
                                     The Part D measure is the percent
                                     of the number of plan members 40-75
                                     years old who were dispensed at
                                     least two diabetes medication fills
                                     and received a statin medication
                                     fill. Receiving multiple fills
                                     indicates the patient continues to
                                     take the medication and therefore
                                     suggests adherence. Continuing to
                                     take the prescribed medication is
                                     necessary to reach clinical/
                                     therapeutic goals. Thus, the Part D
                                     measure is an intermediate outcome
                                     measure. We believe that for these
                                     measures as proposed (and finalized
                                     in this rule) are properly
                                     categorized.
------------------------------------------------------------------------

    We summarize the comments received on the proposed measures and 
respond to them by measure in Table 3D for the Part C measures, for 
performance periods beginning on or after January 1, 2019.
---------------------------------------------------------------------------

    \37\ Ware JE, Kosinski M. SF-36 Physical and Mental Health 
Summary Scales: A Manual for Users of Version 1, Second Edition. 
Lincoln, RI: QualityMetric, Incorporated, 2001.

                        Table 3D--Part D Measures
------------------------------------------------------------------------
              Measure
------------------------------------------------------------------------
Call Center--Foreign Language       Please see comments received and
 Interpreter and TTY Availability.   CMS' responses for this measure in
                                     the above Part C Measures table for
                                     the measure Call Center--Foreign
                                     Language Interpreter and TTY
                                     Availability.

[[Page 16552]]

 
Appeals Auto-Forward..............  Comment: CMS received one comment
                                     suggesting that CMS align the Part
                                     D Appeals Auto-Forward measure with
                                     the Part C Plan Makes Timely
                                     Decisions about Appeals measure.
                                     The commenter also complained that
                                     cases that can be approved, but
                                     because the approvals are untimely,
                                     the cases are forwarded to the IRE;
                                     the commenter said this can cause
                                     delays in patient care as the
                                     member, provider, and plan await
                                     the IRE's decision.
                                    Response: CMS appreciates receiving
                                     comments on this measure. However,
                                     the Part C and Part D appeals
                                     systems are not interchangeable.
                                     Each appeal system has its own set
                                     of rules and procedures which mean
                                     that combining or aligning these
                                     measures is not appropriate. We
                                     direct the commenter to the appeal
                                     regulations at Sec.  Sec.   422.590
                                     and 422.592 as compared to Sec.
                                     Sec.   423.568(h). Further, we note
                                     that the MA and Part D plans have
                                     full control of the appeal prior to
                                     it having been sent to the IRE. In
                                     the example cited, if the plan had
                                     approved the original request from
                                     the member, the appeal would not
                                     have needed to be raised to the IRE
                                     level or incurred the additional
                                     waiting time.
Appeals Upheld....................  Comment: CMS received a comment
                                     requesting that CMS adjust the
                                     Reviewing Appeals Decisions measure
                                     to remove from the measure denials
                                     due to lack of response from
                                     providers from the denominator and
                                     the numerator. The commenter also
                                     requested to align timeframes for
                                     the plan with the IRE stating that
                                     the IRE is generally held to the
                                     same adjudication timeframes as the
                                     plan but if additional information
                                     is needed from a prescriber, the
                                     IRE is allowed to extend the
                                     adjudication timeframe to obtain
                                     this information. The commenter
                                     further said that a plan is not
                                     afforded this time and must deny
                                     based on the information provided
                                     in order to prevent cases from
                                     being auto-forwarded to the IRE.
                                     Therefore, the commenter requested
                                     to measure fairness based on the
                                     information the plan had at the
                                     time of the plan's decision. Plans
                                     should also not be penalized for
                                     appeals that were overturned when
                                     providers provided ``new''
                                     information to the IRE, which was
                                     not originally submitted by the
                                     provider at the time of the plan's
                                     original coverage determination or
                                     redetermination. A commenter from a
                                     plan noted that this measure did
                                     not reflect the commenter's true
                                     plan performance.
                                    Additionally, this commenter noted
                                     several instances where cases were
                                     overturned by the IRE due to
                                     allowing non-Part D supported
                                     indications to be considered and
                                     disregarding the commenter's CMS
                                     approved clinical policies. Due to
                                     these issues, the commenter
                                     proposed an alternative formula to
                                     capture Appeals Upheld data and
                                     measure plan performance in this
                                     area.
                                    Response: CMS appreciates the
                                     comment. Plans and sponsors must
                                     have procedures in place for
                                     requesting and obtaining
                                     information necessary for making
                                     timely and appropriate decisions.
                                     The IRE's decision is based on the
                                     information gathered during its
                                     review process. Adjusting appeal
                                     timeframes is not within the scope
                                     of this proposal, however, we note
                                     that the IRE must issue a decision
                                     within the same appeals timeframe
                                     as the plan. Please refer to 42 CFR
                                     423.600(d). The timeframes for the
                                     plan and the IRE are aligned. At
                                     this time, CMS will continue to
                                     include this measure in the Star
                                     Ratings CMS acknowledges these
                                     comments, and is actively
                                     evaluating these measures, and the
                                     use of the IRE data as their data
                                     source. For future enhancements.
Complaints about the Drug Plan....  Please see comments received and
                                     CMS' responses for this measure in
                                     the above Part C Measures table for
                                     the measure Complaints about the
                                     Health Plan.
Members Choosing to Leave the Plan  Please see comments received and
                                     CMS' responses for this measure in
                                     the above Part C Measures table for
                                     the measure Members Choosing to
                                     Leave the Plan.
Drug Plan Quality Improvement.....  For the summary of comments received
                                     and CMS' responses for this
                                     measure, please see section `j.
                                     Improvement Measures' of the
                                     Preamble.
Rating of Drug Plan...............  Comment: CMS received a number of
                                     general comments on CAHPS measures.
                                    Response: CMS appreciates the
                                     feedback on the CAHPS measures.
                                     Since the comments on CAHPS
                                     measures were mostly not measure
                                     specific, please see the CAHPS
                                     summary of comments received as
                                     well as CMS responses following the
                                     Parts C and D Measure Tables.
                                    Comment: A commenter suggested we
                                     consider this measure `topped out.'
                                    Response: We do not agree this
                                     measure is `topped out'' since many
                                     contracts receive less than 4
                                     stars. Previous analyses of CAHPS
                                     scores have suggested that
                                     seemingly small differences of 1
                                     point on a 0-100 scale are
                                     meaningful; differences of 3 points
                                     can be considered medium, and
                                     differences of 5 points can be
                                     considered large.\38\ For instance,
                                     a 3-point increase in some CAHPS
                                     measures has been associated with a
                                     30 percent reduction in
                                     disenrollment from health plans,
                                     which suggests that even ``medium''
                                     differences in CAHPS scores may
                                     indicate substantially different
                                     care experiences.\39\
Getting Needed Prescription Drugs.  Comment: CMS received a number of
                                     general comments on CAHPS measures.
                                    Response: CMS appreciates the
                                     feedback on the CAHPS measures.
                                     Since the comments on CAHPS
                                     measures were mostly not measure
                                     specific, please see the CAHPS
                                     summary of comments received as
                                     well as CMS responses following the
                                     Parts C and D Measure Tables.
                                    Comment: CMS received one comment
                                     that this composite penalizes Part
                                     D plans where patients do not
                                     prefer to fill prescriptions by
                                     mail.
                                    Response: CMS disagrees that this
                                     composite penalizes plans based on
                                     how beneficiaries choose to fill
                                     prescriptions; rather, the item
                                     focuses on ease of getting
                                     prescriptions filled when using the
                                     plan. The composite covers two
                                     topics: How often it was easy to
                                     use your plan to get the medicines
                                     your doctor prescribed (assessed by
                                     one item) and ease of filling
                                     prescriptions (assessed by
                                     combining two items about how often
                                     it was easy to use your plan to
                                     fill a prescription at your local
                                     pharmacy, and how often it was easy
                                     to use your plan to fill a
                                     prescription by mail). The combined
                                     pharmacy/mail score is averaged
                                     with the first item's score to
                                     produce the composite score. This
                                     averaging weights mail and pharmacy
                                     according to how many respondents
                                     say they use each method, so mail
                                     would not count at all if no one in
                                     the plan uses mail.
                                    Comment: A commenter suggested we
                                     consider this measure `topped out.'

[[Page 16553]]

 
                                    Response: We do not agree this
                                     measure is `topped out' since many
                                     contracts receive less than 4
                                     stars. Previous analyses of CAHPS
                                     scores have suggested that
                                     seemingly small differences of 1
                                     point on a 0-100 scale are
                                     meaningful; differences of 3 points
                                     can be considered medium, and
                                     differences of 5 points can be
                                     considered large.\40\ For instance,
                                     a 3-point increase in some CAHPS
                                     measures has been associated with a
                                     30 percent reduction in
                                     disenrollment from health plans,
                                     which suggests that even ``medium''
                                     differences in CAHPS scores may
                                     indicate substantially different
                                     care experiences.\41\
MPF Price Accuracy................  Comment: A commenter asked CMS to
                                     identify which of the two possible
                                     calculations will be included in
                                     the MPF Accuracy measure. The
                                     commenter noted that CMS had
                                     previously proposed to update the
                                     measure to include frequency and
                                     magnitude of prescription drug
                                     event (PDE) prices that exceed MPF
                                     information beginning with the 2016
                                     data but reverted to the old
                                     measurement (only magnitude) with
                                     the 2018 Star Rating release.
                                    Response: The MPF Accuracy measure
                                     will only measure the magnitude of
                                     difference, as has been done in the
                                     past. CMS will continue to
                                     calculate each contract's accuracy
                                     index which measures the amount
                                     that the PDE price is higher than
                                     the MPF price. CMS will consider
                                     for future rule-making, with
                                     stakeholder input, to include both
                                     frequency and magnitude of PDE
                                     prices that exceed MPF information
                                     in the Accuracy measure.
                                    Comment: A commenter suggested that
                                     this measure is `topped out'. A
                                     measure is considered `topped out'
                                     when it shows high performance
                                     across all contracts decreasing the
                                     variability across contracts and
                                     making the measure unreliable.
                                    Response: As announced through the
                                     2019 Call Letter, CMS is proposing
                                     enhancements to this measure for
                                     the CY2022 Ratings. The enhanced
                                     measure will first be put on
                                     display before being added into the
                                     Star Ratings program pursuant to
                                     the rules in Sec.   423.184.
Adherence Measures:
    Medication Adherence for        Comment: A few commenters requested
     Diabetes Medications,           that CMS consider excluding
     Medication Adherence for        beneficiary prescriptions from
     Hypertension, Medication        these measures or create a
     Adherence for Cholesterol       reporting mechanism that allows
     (Statins).                      plans to identify prescriptions for
                                     removal that are documented as
                                     ``discontinued'' or prescriptions
                                     with therapy changes; the commenter
                                     stated that these changes would
                                     avoid the appearance that
                                     beneficiaries with discontinued
                                     medications are non-adherent. A
                                     commenter expressed concerns about
                                     the thresholds for the medication
                                     adherence for diabetes and
                                     cholesterol measures citing that
                                     they are reaching unsafe levels and
                                     do not reflect individual needs
                                     such as in the aging elderly
                                     population. They described several
                                     circumstances that can adversely
                                     affect adherence measures and
                                     suggest noncompliance, such as
                                     prescription data entry errors and
                                     changes in therapy due to clinical
                                     indicators.
                                    A commenter commended CMS on
                                     including adherence measures in the
                                     Star Ratings. Another commenter
                                     recommended that CMS weight MA-PD
                                     and PDP measures differently based
                                     on the plan's ability to influence
                                     outcomes on a measure. It was
                                     recommended that CMS require
                                     beneficiaries to provide a contact
                                     phone number at the time of
                                     enrollment in order to assist plans
                                     in reaching members to impact
                                     adherence. Another commenter was
                                     concerned about the significant
                                     negative impact by LIS members on
                                     adherence measures.
                                    Response: We appreciate the
                                     feedback. CMS' mission is to
                                     promote quality care for our
                                     beneficiaries. In our May 11, 2012
                                     HPMS memo entitled `Prohibition on
                                     Submitting PDEs for non-Part D
                                     prescriptions', we outlined our
                                     concerns related to beneficiary
                                     privacy protections and data
                                     validation for the submissions of
                                     non-Part D data. If Part D sponsors
                                     were to attempt to collect the data
                                     it is unclear how sponsors could
                                     implement sufficient internal
                                     controls to meet audit standards
                                     necessary to ensure the quality of
                                     the data. In addition, requiring
                                     physicians to attest to therapy
                                     changes or discontinuation of a
                                     prior prescription would be an
                                     added burden and counterproductive
                                     to CMS' Patients over Paperwork
                                     initiative. In the case of changes
                                     in therapy (such as holding or
                                     discontinuing medication), we
                                     believe that the 80 percent
                                     compliance threshold incorporates
                                     these circumstances as the ideal
                                     compliance expectation is 100
                                     percent. We will pass along these
                                     comments to the measure steward
                                     (PQA) but we are unable to use
                                     supplemental data to calculate the
                                     measures.
                                    Data entry error is also a concern
                                     of CMS. We believe that Part D
                                     sponsors have the ability to
                                     identify and correct many data
                                     errors at the point-of-sale and
                                     afterward. Similar to the CMS Part
                                     D Potential Exclusion Warning
                                     Report that identifies PDEs for
                                     adjustment or deletion, plan
                                     sponsors could use their POS edits
                                     systems to screen for data entry
                                     errors. For example, screening
                                     criteria based on a maximum or
                                     minimum daily dose or units per day
                                     could identify outliers. In the
                                     example above, if the term ``3
                                     days'' was accidently entered
                                     instead of ``30 days,'' this could
                                     result in a daily dose that is
                                     significantly higher than the
                                     expected maximum daily dose and
                                     would be an outlier. The claim
                                     could be denied at the POS with a
                                     message of `potential data entry
                                     error' notifying the pharmacist or
                                     technician the need to review and
                                     make a correction. In addition, CMS
                                     provides monthly lists to each plan
                                     sponsor of their members who are
                                     identified as non-compliant
                                     starting in April of each year,
                                     this procedure provides Part D
                                     plans ample time to review their
                                     data and submit corrections.
                                    Also, we disagree that stand-alone
                                     PDPs have very little influence on
                                     beneficiaries' medication
                                     adherence. There are many
                                     strategies that can be used to
                                     improve a beneficiary's medication
                                     adherence in addition to prescriber
                                     interventions, such as refill
                                     reminders, formulary and benefits
                                     design, and medication therapy
                                     management programs. Plan sponsors
                                     can also leverage network pharmacy
                                     relationships to address medication
                                     adherence issues, facilitate
                                     medication synchronization, or
                                     provide education and counseling.
                                     In the absence of a contact phone
                                     number for the beneficiary, it may
                                     be beneficial to use these
                                     interventions to reach the
                                     beneficiary at the place of
                                     dispensing. Furthermore, MA-PDs and
                                     PDPs are rated separately to
                                     account for delivery system
                                     differences. Lastly, as finalized
                                     in the 2019 Call Letter, adherence
                                     measures will now be included in
                                     the CAI to account for LIS
                                     beneficiaries.

[[Page 16554]]

 
MTM Program Completion Rate for     Comment: A commented requested CMS
 CMR.                                move away from MTM process measures
                                     and include outcomes-based MTM
                                     measures in the Star Ratings
                                     program in the future. In the
                                     interim, it was recommended that
                                     CMS evaluate changes to the MTM
                                     Comprehensive Medication Review
                                     Completion Rate (CMR) measure
                                     methodology and that CMS partner
                                     with PQA to develop and understand
                                     the feasibility of implementing
                                     outcome and/or patient-experience
                                     based MTM measures.
                                    Response: The CMR completion rate
                                     measure is an initial measure of
                                     the delivery of MTM services, and
                                     we continue to look forward to the
                                     development and endorsement of
                                     outcomes-based MTM measures as
                                     potential companion measures to the
                                     current MTM Completion Rate CMR
                                     measure. We will consider new MTM
                                     measures when available. Past
                                     analyses did not find a correlation
                                     between a sponsor's rate of MTM
                                     program eligibility and the CMR
                                     completion rate, but we will
                                     continue to monitor and work with
                                     the PQA to consider if any
                                     adjustments are needed to this
                                     measure's specifications.
                                    Comment: A commenter opposed
                                     inclusion of the MTM CMR completion
                                     rate measure in the Star Ratings
                                     due to compliance issues. The
                                     commenter suggested allowing
                                     completion of CMRs with the
                                     beneficiary's prescriber when
                                     unable to contact the beneficiary.
                                    Response: As outlined in 42 CFR
                                     423.153(d)(vii)(B)(2), if a
                                     beneficiary is offered the annual
                                     comprehensive medication review and
                                     is unable to accept the offer to
                                     participate, the pharmacist or
                                     other qualified provider may
                                     perform the comprehensive
                                     medication review with the
                                     beneficiary's prescriber,
                                     caregiver, or other authorized
                                     individual. Current guidance
                                     clarifies that while providers are
                                     required to offer a CMR to all
                                     beneficiaries enrolled in the MTM
                                     program, regardless of setting, in
                                     the event the beneficiary is
                                     cognitively impaired or otherwise
                                     unable to participate, we recommend
                                     that the pharmacist or qualified
                                     provider reach out to the
                                     beneficiary's prescriber,
                                     caregiver, or other authorized
                                     individual, such as the resident's
                                     health care proxy or legal
                                     guardian, to take part in the
                                     beneficiary's CMR. This applies to
                                     beneficiaries in any setting and is
                                     not limited to beneficiaries in
                                     long term care (LTC). This does not
                                     apply to situations where the
                                     sponsor is simply unable to reach
                                     the beneficiary or there is no
                                     evidence of cognitive impairment.
                                     Therefore, we are unable to
                                     consider changes to the measure
                                     absent a change in regulation or
                                     guidance.
Statin Use in Persons With          Comment: A few commenters supported
 Diabetes (SUPD).                    CMS in including this SUPD measure
                                     in the Star Ratings. A commenter
                                     noted support of the addition of a
                                     quality metric monitoring the use
                                     of statins in patients with
                                     diabetes, however, feels that CMS
                                     did not provide a thoughtful
                                     explanation for not selecting the
                                     Part C HEDIS measure of Statin
                                     Therapy in Patients with Diabetes
                                     (NCQA measure), which had also been
                                     under consideration. This measure
                                     includes more robust clinical
                                     considerations for patient
                                     eligibility and thus
                                     appropriateness of statin use.
                                    Response: CMS thanks commenters for
                                     feedback on this measure. Both the
                                     NCQA and PQA measures of statin
                                     therapy were proposed for inclusion
                                     in the Star Ratings, one for Part C
                                     and the other for Part D. As the
                                     Pharmacy Quality Alliance (PQA) is
                                     the developer of the Statin Use in
                                     Persons with Diabetes (SUPD)
                                     measure, CMS will share these
                                     comments with the PQA for their
                                     consideration.
                                    Comment: CMS received two comments
                                     seeking clarification regarding the
                                     categorization and weighting
                                     discrepancies between the Part C
                                     and Part D statin measures. Two
                                     organizations recommended
                                     classifying both SPC and SUPD as
                                     process measures with a weight of
                                     one.
                                    Response: Please refer to the Part C
                                     measure response for Statin Use for
                                     Patients with Cardiovascular
                                     Disease (SPC).
------------------------------------------------------------------------

CAHPS: Summary of Additional Comments Received and CMS's Responses
---------------------------------------------------------------------------

    \38\ Paddison CAM, Elliott MN, Haviland AM, Farley DO, 
Lyratzopoulos G, Hambarsoomian K, Dembosky JW, Roland MO. (2013). 
``Experiences of Care among Medicare Beneficiaries with ESRD: 
Medicare Consumer Assessment of Healthcare Providers and Systems 
(CAHPS) Survey Results.'' American Journal of Kidney Diseases 61(3): 
440-449.
    \39\ Lied, T.R., S.H. Sheingold, B.E. Landon, J.A. Shaul, and 
P.D. Cleary. (2003). ``Beneficiary Reported Experience and Voluntary 
Disenrollment in Medicare Managed Care.'' Health Care Financing 
Review 25(1): 55-66.
    \40\ Paddison CAM, Elliott MN, Haviland AM, Farley DO, 
Lyratzopoulos G, Hambarsoomian K, Dembosky JW, Roland MO. (2013). 
``Experiences of Care among Medicare Beneficiaries with ESRD: 
Medicare Consumer Assessment of Healthcare Providers and Systems 
(CAHPS) Survey Results.'' American Journal of Kidney Diseases 61(3): 
440-449.
    \41\ Lied, T.R., S.H. Sheingold, B.E. Landon, J.A. Shaul, and 
P.D. Cleary. (2003). ``Beneficiary Reported Experience and Voluntary 
Disenrollment in Medicare Managed Care.'' Health Care Financing 
Review 25(1): 55-66.
---------------------------------------------------------------------------

    Comment: CMS received a few comments that CAHPS measures are 
subjective and not reliable. A few commenters stated the CAHPS survey 
responses are not actionable.
    Response: CMS strongly disagrees that patient experience of care 
survey measures are not reliable. CAHPS and other patient experience 
measures have been endorsed as critical aspects of healthcare by the 
Institute of Medicine and the World Health Organization.\42\ \43\ CAHPS 
surveys focus on aspects of healthcare quality that patients themselves 
say are important to them and for which patients are the best and/or 
only source of information. Patient experience surveys such as CAHPS 
focus on how patients experienced key aspects of their care, not merely 
how satisfied they were with their care. Patient experience encompasses 
the range of interactions that patients have with the healthcare 
system, including their care from health plans, and from doctors, 
nurses, and staff in hospitals, physician practices, and other 
healthcare facilities.\44\ While patient experience is an inherently 
important dimension of healthcare quality, it is also the case that the 
preponderance of evidence shows that better patient experience is 
associated with better patient adherence to recommended treatment, 
better clinical processes, better hospital patient safety culture, 
better clinical outcomes, reduced unnecessary healthcare use, and fewer

[[Page 16555]]

inpatient complications.\45\ \46\ Therefore, while we acknowledge that 
the CAHPS survey captures individuals' perspectives on their 
experiences of care, it is anchored in measureable aspects of care and 
so can be measured reliably.
---------------------------------------------------------------------------

    \42\ Institute of Medicine. Crossing The Quality Chasm: A New 
Health System for the 21st Century. Washington DC: National Academy 
Press; 2001.
    \43\ Smith, P.C. (Ed.). (2009). Performance measurement for 
health system improvement: experiences, challenges and prospects. 
Cambridge University Press.
    \44\ Agency for Healthcare Research and Quality. What Is Patient 
Experience?. Content last reviewed March 2017. Agency for Healthcare 
Research and Quality, Rockville, MD. http://www.ahrq.gov/cahps/about-cahps/patient-experience/index.html.
    \45\ Price, R.A., Elliott, M.N., Zaslavsky, A.M., Hays, R.D., 
Lehrman, W.G., Rybowski, L., & Cleary, P.D. (2014). Examining the 
role of patient experience surveys in measuring health care quality. 
Medical Care Research and Review, 71(5), 522-554.
    \46\ Price, R.A., Elliott, M.N., Cleary, P.D., Zaslavsky, A.M., 
& Hays, R.D. (2015). Should health care providers be accountable for 
patients' care experiences?. Journal of general internal medicine, 
30(2), 253-256.
---------------------------------------------------------------------------

    Additionally, CAHPS surveys follow scientific principles in survey 
design and development and have been rigorously developed and tested to 
assess the experiences of Medicare beneficiaries. The surveys are 
designed to reliably assess the experiences of a large sample of 
patients and use standardized questions and data collection protocols 
to ensure that information can be compared across health care settings. 
The contract-level reliability of 2017 MA and PDP CAHPS measures meet 
high standards, with the median reliability of publicly-reported MA 
CAHPS measures exceeding 0.72 for all measures and exceeding 0.90 for a 
majority of measures, with 0.70 being a conventional standard for 
reliability. Finally, there are criteria for sample size eligibility 
that must be met for contracts to be included in data collection, and 
CMS also offers contracts the option of augmenting their CAHPS sample 
sizes if they wish to obtain more precise overall results and/or 
perform subgroup analyses with larger samples.
    Comment: Several commenters stated that CAHPS scores may be 
influenced by factors outside the plan's control, such as cost and 
coverage, provider behavior, cultural differences including language, 
and timing of the survey. A few suggested that beneficiaries who are 
frail, have cognitive impairments, or who have low socio-economic 
status may not be able to respond to survey items accurately. A 
commenter requested allowing proxy methods.
    Response: For MA and PDP CAHPS, CMS uses mixed-mode data collection 
to increase the likelihood of survey participation and 
representativeness.\47\ \48\ Survey responses are also case-mix 
adjusted to account for certain respondent characteristics not under 
the control of the health or drug plan such as age, education, dual 
eligible status and other variables. We note that plans do have some 
control over plan-design features such as cost and coverage as well as 
provider behavior, so it would not be appropriate to adjust for these.
---------------------------------------------------------------------------

    \47\ Fowler Jr, F.J., Gallagher, P.M., Stringfellow, V.L., 
Zaslavsky, A.M., Thompson, J.W., & Cleary, P.D. (2002). Using 
telephone interviews to reduce nonresponse bias to mail surveys of 
health plan members. Medical care, 190-200.
    \48\ Elliott, M.N., Zaslavsky, A.M., Goldstein, E., Lehrman, W., 
Hambarsoomians, K., Beckett, M.K., & Giordano, L. (2009). Effects of 
survey mode, patient mix, and nonresponse on CAHPS[supreg] hospital 
survey scores. Health services research, 44(2p1), 501-518.
---------------------------------------------------------------------------

    CMS currently provides translations of the MA and PDP CAHPS Survey 
in Spanish, Chinese, and Vietnamese, and we are developing a Korean 
translation. All translations are the product of translation and review 
by native speakers of the target languages and have had multiple rounds 
of qualitative testing with Medicare beneficiaries with characteristics 
similar to the MA and PDP CAHPS population. By providing survey 
translations, CMS promotes standardization by assuring that questions 
are presented similarly to beneficiaries across and within languages, 
which also promotes comparability of the results across vendors and 
contracts. The survey administration protocol for MA and PDP CAHPS does 
not permit ``live,'' ``individual,'' or ``real-time'' translation of 
the survey by an interpreter, as such an approach does not promote 
comparability of data and there is no mechanism for assuring the 
accuracy and consistency of the translation. If plans need additional 
translations they should contact us at MP-CAHPS@cms.hhs.gov. The MA and 
PDP CAHPS protocol does allow for the use of proxy respondents in cases 
where a respondent is unable to complete the survey.
    Comment: A commenter stated that the CAHPS survey is long, and a 
couple commenters expressed concern about low response rates.
    Response: CMS shortened the MA CAHPS survey in 2017 by removing 
questions and measures not used in Star Ratings, and we also improved 
phone contact information. As a result of CMS's continuing efforts to 
improve response rates, overall MA and PDP CAHPS response rates 
increased from 2016 to 2017, despite national trends of declining 
response rates for most other surveys. Further, meta-analyses of 
surveys that follow the rigorous probability sampling and survey 
approaches used by MA and PDP CAHPS find little relationship between 
response rates and nonresponse bias.\49\ Moreover, research specific to 
patient experience, CAHPS, and MA and PDP CAHPS surveys finds no 
evidence nonresponse bias affects comparison of case-mix adjusted 
scores between contracts or other similar reporting units.\50\ \51\ 
\52\ \53\ \54\
---------------------------------------------------------------------------

    \49\ Groves, R.M., & Peytcheva, E. (2008). The impact of 
nonresponse rates on nonresponse bias: a meta-analysis. Public 
opinion quarterly, 72(2), 167-189.
    \50\ Klein, D.J., Elliott, M.N., Haviland, A.M., Saliba, D., 
Burkhart, Q., Edwards, C., & Zaslavsky, A.M. (2011). Understanding 
nonresponse to the 2007 Medicare CAHPS survey. The Gerontologist, 
51(6), 843-855.
    \51\ Saunders C.L., Elliott M.N., Lyratzopoulos G., Abel G.A. 
(2016) ``Do differential response rates to patient surveys between 
organisations lead to unfair performance comparisons? Evidence from 
the English Cancer Patient Experience Survey'' Medical Care 54(1): 
45-54.
    \52\ Bone A., McGrath-Lone L., Day S., et al. Inequalities in 
the care experiences of patients with cancer: Analysis of data from 
the National Cancer Patient Experience Survey 2011-2012. BMJ Open. 
2014;4:e004567.
    \53\ El Turabi A., Abel G.A., Roland M., et al. Variation in 
reported experience of involvement in cancer treatment decision 
making: Evidence from the National Cancer Patient Experience Survey. 
Br J Cancer. 2013;109:780-787.
    \54\ Lyratzopoulos G., Neal R.D., Barbiere J.M., et al. 
Variation in number of general practitioner consultations before 
hospital referral for cancer: findings from the 2010 National Cancer 
Patient Experience Survey in England. Lancet Oncol. 2012;13:353-365.
---------------------------------------------------------------------------

    Comment: A commenter requested more insight into statistical 
components such as case-mix adjustment, statistical significance, and 
reliability, and another requested that CMS provide all case-mix 
adjustment flags to the survey vendors to facilitate an additional 
validation.
    Response: CMS provides a detailed explanation of the CAHPS 
methodology including case-mix adjustment in the annual Star Ratings 
Technical Notes, in CAHPS plan reports provided to each contract each 
year, and on the MA and PDP CAHPS web page (https://www.ma-pdpcahps.org). CMS also provides survey vendors all of the necessary 
data to perform case-mix adjustment validation. Plans are welcome to 
contact MP-CAHPS@cms.hhs.gov with specific questions about MA and PDP 
CAHPS.
    Comment: A commenter requested that plans be able to add their own 
questions to the surveys to validate and clarify responses.
    Response: CMS allows plans to add a limited number of items to the 
MA and PDP CAHPS survey that do not affect responses to the survey or 
pose undue burden to the beneficiary. These rules are to ensure the 
highest possible response rate as well as comparability of the data 
across contracts.
HOS: Summary of Additional Comments Received and CMS's Responses
    Comment: CMS received several comments on the HOS measures. Some 
commenters supported patient reported

[[Page 16556]]

outcome measures. Several commenters, however, suggested that the HOS 
has drawbacks in design, methodology, administration, and reporting 
that disproportionately affect SNP populations and fail to accommodate 
diverse populations and the most vulnerable beneficiaries. Some 
commenters stated that the longitudinal two year look-back design of 
the HOS is especially challenging in populations with high rates of 
degenerative or progressive conditions coupled with pervasive low 
socioeconomic status and high social risk factors. Commenters suggested 
that CMS should change sampling methodology to require larger sample 
sizes or allow plans to request oversampling of typically under-
represented groups. In addition, some commenters would like to 
discontinue the use of proxies for self-report as, the commenters 
argue, there is strong evidence indicating proxies' responses are not 
equivalent to beneficiaries' responses.
    Response: CMS is supportive of increasing sample sizes and is not 
opposed to oversampling to ensure a representative sample but to date 
has received no HOS oversampling requests from any plans. We are 
currently reexamining the HOS with a focus on diverse, dual-eligible 
populations and will explore the feasibility of increasing the required 
sample size. CMS already adjusts the HOS data to control for many 
beneficiary characteristics not under the control of the plan, 
including age, gender, race, ethnicity, income, education, marital 
status, Medicaid status, SSI eligibility, homeowner status, chronic 
conditions, and baseline health status. CMS does not plan to 
discontinue the HOS proxy response option. Because the HOS has both 
mail and telephone components, it is likely that some mail 
questionnaires would be completed by proxies whether permitted or not. 
CMS considers it preferable to collect information about whether the 
beneficiary or a proxy answered the survey than to assume the 
beneficiary answered the questions. Every attempt is made to obtain a 
response from the beneficiary before a proxy response is allowed. Also, 
when a proxy was used at baseline and the beneficiary remains unable to 
complete the follow up survey, attempts are made to re-contact the same 
proxy in order to reduce variability in responses. Finally, frailer 
members, including the most vulnerable beneficiaries, who are unable to 
complete the survey independently are excluded from the HOS if a proxy 
response option is not available.
    Comment: Several commenters mentioned the two year look-back period 
is challenging to beneficiaries. A commenter suggested that keeping the 
identity of sample respondents confidential limits opportunities for 
improvement activities, and another suggested the resulting data may be 
too old to be actionable. A few commenters recommended the elimination 
of HOS measures because the measures are too generic for Star Ratings 
and the information from the surveys is not actionable.
    Response: The Health Outcome Survey (HOS) yields two patient-
reported outcome measures of change in global functioning, by using 2-
year change in scores on the Physical Component Score (PCS) and the 
Mental Component Score (MCS), both of which come from the Veterans RAND 
12-Item Health Survey (VR-12) portion of the larger survey. These 
measures are of unique and high value, as demonstrated by their higher 
weight in calculating the Overall Star Ratings. Critics of the HOS 
often point out the 3 years between HOS baseline data collection and 
health plans receiving member-level results, which include the 
identities of respondents. Contributing to the perceived ``lag'' is the 
longitudinal component of the HOS; beneficiaries who complete the 
baseline HOS must be resurveyed two years later to generate data for 
the HOS ``outcome'' measures. HOS data are hardly ``old.'' In fact, HOS 
baseline results are distributed nine months after data collection 
ends, and performance measurement reports and beneficiary-level data 
are distributed about one year after follow-up data collection ends. 
Further, CMS contends that a majority of plans improve or maintain the 
physical and/or mental health of their membership over time. That is, 
the measure requires time to capture change and in fact does capture 
positive change or maintenance of global functioning for the majority 
of plans' members. The Physical Component Score (PCS) and the Mental 
Component Score (MCS), as derived from the VR-12, have been validated 
in multiple studies of VA and elderly populations. The appendix of each 
contract's annual performance measurement report explains how the 
measures are calculated and adjusted to minimize bias in results. CMS 
encourages all plans to familiarize themselves with the methods 
described in the reports and to utilize the background materials 
available on the HOS website that validate the Improving or Maintaining 
Physical Health and Improving or Maintaining Mental Health measures.
    Comment: Commenters also suggested that CMS provide HOS translation 
and instrument adaptation for languages in addition to English, 
Spanish, or Chinese.
    Response: CMS responds to requests for translations of the survey 
into other languages from vendors, who in turn reflect the requests of 
plans. CMS currently provides translations of the HOS in Spanish and 
Chinese, and a Russian translation will be available in 2019. All 
translated versions are the product of translation and review by native 
speakers of these languages and are subject to multiple rounds of 
qualitative testing with Medicare beneficiaries with characteristics 
similar to the HOS population. As a result, the adoption of a 
translated survey tool takes a significant amount of time. By providing 
survey translations, CMS promotes standardization and assures that 
questions are presented similarly to beneficiaries across and within 
languages, which also promotes comparability of the results across 
vendors and contracts. The survey administration protocol for HOS does 
not permit ``live,'' ``individual,'' or ``real-time'' translation of 
the survey by interpreters because such an approach does not promote 
comparability of data and there is no mechanism for assuring the 
accuracy and consistency of the translation. However, the HOS protocol 
does allow for the use of proxy respondents in cases where a 
beneficiary is unable to complete the survey.
    Comment: A commenter reported that they have observed that plans 
with lower membership generally have higher scores on HOS measures than 
plans with higher enrollment.
    Response: We appreciate the comment. CMS is not aware of any formal 
studies that have been done to address the hypothesized link between 
contract size and performance on longitudinal measures.
Patients With Advanced Illness: Comments Received and CMS's Responses
    Comment: CMS received several comments concerning the exclusion 
from measures of patients with advanced illness and in palliative care; 
those who have refused treatment, assessment, or recommended 
screenings; and those who are unable to achieve the desired clinical 
threshold despite having reached the maximum medical therapy and self-
care practices available for the condition. Commenters recommended that 
exclusions or adjustments to measures be made for these patients, or 
that alternate metrics be developed for these patients, since for many 
of them comfort or improving

[[Page 16557]]

quality of life is a greater part of care than curative treatments. In 
particular, some commenters identified specific HEDIS and HOS measures 
which should be excluded or modified for patients with advanced 
illness: Rheumatoid Arthritis, Statin Use, Improving or Maintaining 
Physical Health, and Improving or Maintaining Mental Health. Commenters 
note that there are many challenges treating and screening certain 
health conditions for patients with advanced illness. A commenter 
suggested that the seriously ill population be excluded from preventive 
and HOS measures, as feasible. While commenters agreed that MA plans 
should advance preventive care and maintain or improve physical health 
for the majority of their enrollees, they argued that there will always 
be a subset of enrollees facing serious illness and continued decline. 
Commenters encouraged CMS to work with measure stewards such as NCQA 
and explore other options that can exclude the seriously ill population 
from such measures. Commenters suggested that the exclusion of the 
seriously ill population from these measures will protect against 
discriminatory enrollment, and will not unfairly evaluate plans that 
support this population in making diagnostic and treatment decisions 
based on the patient's preferences. Finally, some commenters suggested 
that patients with advanced illness who have refused services and 
treatments should also be excluded from measure calculations. They 
stated a patient's goal for comfort rather than further treatment 
should be primary. A commenter suggested that the under 65 population 
residing in nursing homes should be excluded from measures for many of 
the same reasons they wanted those with advanced illness excluded--
advanced sickness, nearing the end of life, refusing treatment, and 
sometimes a patient's choice on comfort not care.
    Response: CMS appreciates feedback on the noted measure adjustments 
and exclusions. For HEDIS 2019, NCQA is examining potential cross-
cutting exclusions for those with advanced illness from selected HEDIS 
measures that may not be clinically appropriate for these individuals. 
NCQA is considering various advanced illness conditions and service use 
(for example, indications of frailty, receipt of palliative care or 
nursing care services) for potential exclusion. We anticipate that NCQA 
will consider these comments as their advisory panels re-evaluate 
measures as part of the standard HEDIS process. Proposed changes to 
implement advanced illness exclusions will be posted for the HEDIS 2019 
public comment period in February 2018. CMS currently has no plans to 
exclude members with serious illness from the HOS.
Additional Comments and Responses
    Comment: CMS received one suggestion that CMS create a new, fixed 
identification code for each measure that would be consistent year-
over-year.
    Response: The measure codes are not published publicly for 
beneficiaries. CMS publishes a Star Ratings measure history in the 
Technical Notes each year that cross references the measure codes. 
Plans are welcome to use their own internal coding systems.
    Comment: A commenter suggested that CMS make PDEs available for 
members in drug assistance programs.
    Response: CMS thanks the commenter for this suggestion. However, 
this suggestion is outside the scope of the proposed rule. This comment 
will be shared with others in CMS who will be interested in the 
suggestion.
    Comment: A commenter suggested that CMS exclude beneficiaries' Part 
D trial medication use from the measures.
    Response: CMS believes this request is specific to the adherence 
measures. The adherence measures require at least two fills on 
different dates for any drug within the drug class for inclusion in the 
measure. The two claim requirement essentially excludes many trial 
prescription periods where the beneficiary failed the initial drug and 
is switched to a different drug class. Since the adherence measures are 
for chronic conditions, CMS expects that the beneficiary would continue 
on one drug within the drug class in the measure. Identifying trial 
periods using PDEs outside this definition would be difficult to 
determine and accepting other source data would be prohibited as 
previously stated.
Summary of Changes
    For the reasons set forth in the proposed rule and our responses to 
the related comments summarized earlier, CMS is finalizing the Part C 
and Part D performance measures for the performance periods beginning 
on or after January 1, 2019 with one modification. In that NCQA is 
planning to make substantive changes to the Plan All-Cause Readmissions 
measure that would affect measurement year 2019, CMS is not finalizing 
this as a measure in the 2021 and 2022 Star Ratings but will move this 
measure to the display page for two years. CMS's finalization of the 
proposed measures does include the specifications (metric and 
performance period), domain assignment, measure category, data source 
for the measures, and statistical method for assigning Star Ratings 
(based on Sec. Sec.  422.166(a) and 423.186(a)) as listed in the 
proposed table. However, we note that our finalization of the proposed 
measures does not include the weight of each category as presented in 
the proposed table. For discussion of CMS's final decision to change 
the weight of measures in the Patients' Experience and Complaints 
category and in the Measures Capturing Access category from a weight of 
1.5 to a weight of 2, see section `q. Measure Weights' of this 
preamble. See also Sec. Sec.  422.166(e) and 423.186(e) of this 
regulation for final measure weight assignments. Finally, we note that 
the summary of comments received and CMS's responses for the Health 
Plan Quality Improvement and the Drug Plan Quality Improvement measures 
are presented in the next section (`j. Improvement Measures') of this 
preamble.
j. Improvement Measures
    In the 2013 Part C and D Star Ratings, we implemented the Part C 
and D improvement measures (CY2013 Rate Announcement, https://www.cms.gov/Medicare/Health-Plans/MedicareAdvtgSpecRateStats/Downloads/Announcement2013.pdf). The improvement measures address the overall 
improvement or decline in individual measure scores from the prior to 
the current year. We proposed to continue the current methodology 
detailed in the Technical Notes for calculating the improvement 
measures and to codify it at Sec. Sec.  422.164(f) and 423.184(f). For 
a measure to be included in the improvement calculation, the measure 
must have numeric value scores in both the current and prior year and 
not have had a substantive specification change during those years. In 
addition, the improvement measure would not include any data on 
measures that are already focused on improvement (for example, HOS 
measures focused on improving or maintaining physical or mental 
health). The Part C improvement measure includes only Part C measure 
scores, and the Part D improvement measure includes only Part D measure 
scores. We proposed to codify these criteria at paragraph (f)(1)(i) 
through (iii) of Sec. Sec.  422.164 and 423.184. We proposed to 
annually identify the subset of measures to be included in the 
improvement measures through the Call Letter, similar to our proposal 
for regular updates and removal of measures. Under our proposal, once 
the measures to be used for the improvement

[[Page 16558]]

measures are identified, CMS would determine which contracts have 
sufficient data for purposes of applying and scoring the improvement 
measure(s). We again proposed to follow current practices: The 
improvement measure score would be calculated only for contracts that 
have numeric measure scores for both years for at least half of the 
measures identified for use in the improvement measure. We proposed 
this standard for determining contracts eligible for an improvement 
measure at paragraph (f)(2).
    We proposed at Sec. Sec.  422.164(f)(3) and (4) and 423.184(f)(3) 
and (4) the process for calculating the improvement measure score(s) 
and a special rule for any identified improvement measure for a 
contract that received a measure-level Star Rating of 5 in each of the 
2 years examined, but whose associated measure score indicates a 
statistically significant decline in performance over the time period.
    As proposed, the improvement measure would be calculated in a 
series of distinct steps:
     The improvement change score (the difference in the 
measure scores in the 2-year period) will be determined for each 
measure that has been identified as part of an improvement measure and 
for which a contract has a numeric score for each of the 2 years 
examined.
     Each contract's improvement change score will be 
categorized as a significant change or not by employing a two-tailed t-
test with a level of significance of 0.05.
     The net improvement per measure category (outcome, access, 
patient experience, process) will be calculated by finding the 
difference between the weighted number of significantly improved 
measures and significantly declined measures, using the measure weights 
associated with each measure category.
     The improvement measure score will then be determined by 
calculating the weighted sum of the net improvement per measure 
category divided by the weighted sum of the number of eligible 
measures.
     The improvement measure scores will be converted to 
measure-level Star Ratings by determining the cut points using 
hierarchical clustering algorithms.
    We proposed at Sec. Sec.  422.166(a)(2)(iii) and 422.186(a)(2)(iii) 
that the improvement measure score cut points would be determined using 
two separate clustering algorithms. We explained in the preamble that 
improvement measure scores of zero and above will use the clustering 
algorithm to determine the cut points for the Star Rating levels of 3 
and above. Improvement measure scores below zero will be clustered to 
determine the cut points for 1 and 2 stars. Although the preamble of 
the proposed rule indicated this level of detail, our proposed 
regulation text, at proposed paragraphs (f)(4)(v) and (vi) of 
Sec. Sec.  422.164 and 423.184, did not. In paragraph (4)(v), we 
referred only to ``hierarchical clustering algorithms'' without 
specifying the detailed treatment for scores of greater than, equal to, 
or less than zero; in paragraph (4)(vi), we cross-referenced the text 
proposed at Sec. Sec.  422.166(a)(2) and 423.186(a)(2), which did 
include the specific text specifying the detailed treatment for scores 
of greater than, equal to, or less than zero in connection with the 
ratings for the improvement measures. While our proposed regulation 
text was ultimately consistent, it included cross-references not 
explained in the preamble.
    We also proposed that the Part D improvement measure cut points for 
MA-PDs and PDPs would be determined using separate clustering 
algorithms. The Part D improvement measure cut points for MA-PDs and 
PDPs would be reported separately. Finally, we proposed a special rule 
in paragraph (f)(3) to hold harmless sponsoring organizations that have 
5-star ratings for both years on a measure used for the improvement 
measure calculation. This hold harmless provision was added in 2014 to 
avoid the unintended consequence for contracts that score 5 stars on a 
subset of measures in each of the 2 years. For any identified 
improvement measure for which a contract received a rating of 5 stars 
in each of the years examined, but for which the measure score 
demonstrates a statistically significant decline based on the results 
of the significance testing (at a level of significance of 0.05) on the 
change score, the measure would be categorized as having no significant 
change. The measure would be included in the count of measures used to 
determine eligibility for the improvement measure and in the 
denominator of the improvement measure score. We explained in the 
proposed rule that the intent of the hold harmless provision for a 
contract that receives a measure rating of 5 stars for each year is to 
prevent the measure from lowering a contract's improvement measure when 
the contract still demonstrates high performance. We proposed in 
section II.A.12.r another hold harmless provision to be codified at 
Sec. Sec.  422.166(g)(1) and 423.186(g)(1).
    We requested comment on the methodology for the improvement 
measures, including rules for determining which measures are included, 
the conversion to a Star Rating, and the hold harmless provision for 
individual measures that are used for the determination of the 
improvement measure score.
    We received the following comments on our proposals, and our 
responses follow:
    Comment: The overwhelming majority of commenters supported the 
concept of the improvement measures.
    Response: CMS appreciates the overwhelming support for the 
underlying rationale of the improvement measures.
    Comment: A commenter opposed the codification of the improvement 
measures and urged CMS to discontinue its use in the Star Ratings 
program. The commenter believes that the improvement measures are 
unnecessary, distort the signal provided by the Star Ratings, blur the 
distinction between high performing contracts and other contracts, and 
can lead to misclassification.
    Response: CMS believes that continuous improvement is an important 
component of the Star Ratings program and necessary to achieve the 
ultimate goal of providing the best care to beneficiaries and realizing 
the most positive outcomes. The improvement measures provide a distinct 
aspect of performance and as implemented, provide a true reflection of 
this aspect of performance. CMS is cognizant of the challenges of 
improvement for contracts that have high performance; thus, CMS 
implemented the hold harmless provisions. One hold harmless provision 
addresses high performance at the measure level, and the other 
addresses high performance at the highest rating level. The hold 
harmless provisions coupled with the two-step clustering for converting 
the improvement measure scores to measure-level Star Ratings safeguard 
against possible misclassification. CMS appreciates the comments and 
will continue to look at ways to further enhance the Star Ratings.
    Comment: Some commenters suggested excluding CAHPS and HOS measures 
from the improvement measure because they believe the measures are 
subjective in nature. A commenter further justified the removal of the 
survey measures citing the challenges in sample selection that have 
occurred in recent years that have led some plans to appeal their 
results as not statistically significant.
    Response: CMS reviews and selects the improvement measures annually 
and publishes the list in the draft Call Letter, we proposed to follow 
the same

[[Page 16559]]

process going forward. For a measure to be included in the improvement 
calculation, the measure must have numeric value scores in both the 
current and prior year and not have had a substantive specification 
change during those years. In addition, the improvement measure will 
not include any data on measures that are already focused on 
improvement (for example, HOS measures focused on improving or 
maintaining physical or mental health). CAHPS and HOS measures are 
patient experience not patient satisfaction surveys. The voice of the 
beneficiary is a critical component of the information needed for the 
Star Ratings program to realize its goals. If an issue arises with any 
aspect of the standard protocol regarding sampling in the Star Ratings 
program, CMS carefully reviews any impact of the deviation and assesses 
the risk of unintended consequences on the integrity of the ratings. 
Further, CMS develops and tests analytical adjustments to mitigate and 
address all such concerns. Although there did exist minor deviations in 
the protocol for sampling in the Star Ratings in the past, CMS is 
confident that the ratings were not affected and the measures possessed 
all attributes necessary to preserve and maintain the high standards of 
the Star Ratings program.
    Comment: Many commenters supported an expansion of the measure-
level hold harmless provision for a contract that receives 4 or more 
stars in each of the two-years for a measure. Some commenters noted the 
lack of alignment between the highly-rated contracts' hold harmless 
provision for the application of the improvement measure(s) for the 
identification of a contract's highest rating at Sec.  422.166(g)(1) 
and Sec.  423.186(g)(1) and the measure-level hold harmless provision 
at (Sec.  422.164(f)(3) and Sec.  423.184(f)(3).
    Response: CMS appreciates the thoughtful consideration of the hold 
harmless provisions for the improvement measure methodology. As noted, 
the hold harmless provision at the measure level applies a different 
threshold than the hold harmless provision for a highly-rated 
contract's highest rating. A measure, in general, assesses a single, 
distinct aspect of care while an overall or summary rating provides a 
global indicator of quality of care and performance.
    At the basic building block of the rating system, the measure, a 
measure-level rating of 4 stars allows opportunity for improvement with 
a focus on a singular concept. A measure-level Star Rating of 5 does 
not allow the same degree of possible improvement. The measure-level 
hold harmless provision was designed to protect a contract from being 
adversely impacted by the improvement measure(s) without discouraging 
continuous improvement. CMS believes that changing the hold harmless to 
measures that receive at least 4 stars each year would serve to hamper 
advances and innovation in the care of all populations; in addition, it 
could serve to discourage continuous improvement by suggesting that 4 
stars--rather than 5--is the highest achievement on the measure.
    CMS is cognizant of the additional challenges of improvement for 
highly-rated contracts; improvement is more difficult for a contract 
with high performance as compared to a lower-rated contract that has 
more opportunity for improvement. The hold harmless provision for a 
contract's highest rating provides the safeguard for contracts that 
receive an overall or summary rating of 4 stars or more without the use 
of the improvement measures and with all applicable adjustments (CAI 
and the reward factor). A highly-rated contract will have their final 
highest rating as the higher of either the rating calculated including 
or excluding the improvement measures.
    CMS believes there should be a differentiation in the hold harmless 
provisions to appropriately address the amount of information each 
provides, to incentivize contracts to continuously improve, and to 
provide adequate safeguards for high achieving contracts.
    Comment: A few commenters expressed explicit support for the 
current methodology for determining the improvement rating including 
the use of separate clustering algorithms to convert the improvement 
measure scores to a measure-level Star Rating and the separate 
clustering algorithms for the Part D summary rating for PDPs and MA-
PDs.
    Response: CMS appreciates these comments.
    Comment: A commenter suggested that CMS develop a measure to assess 
a decline in performance.
    Response: The current improvement measures capture both improvement 
and decline. The calculation for the improvement measure score and the 
associated methodology to convert the improvement measures scores to 
measure-level Star Ratings are designed such that a contract that has 
below average improvement, indicated by an improvement measure score 
less than zero, will receive an improvement measure-level Star Rating 
less than 3 stars.
    Comment: A commenter expressed concern with the improvement 
methodology and believes it creates a double-jeopardy situation because 
it includes both significance testing and national performance.
    Response: The Star Ratings are designed to incentivize contracts to 
provide the best quality and care to beneficiaries. The methodology 
employed to determine the improvement measure-level Star Ratings is 
designed to align with the underlying principles of the Star Ratings 
methodology. The use of statistical significance allows the changes of 
each individual measure used for the determination of the improvement 
measure score to be assessed for meaningful differences. The use of the 
clustering algorithm to determine the cut points and ultimately, the 
assignment of the measure-level Star Ratings, allows a contract's 
performance to be assessed relative to all contracts that are required 
to report. The determination of the measure-level Star Ratings is done 
in a manner to minimize misclassification. The clustering for the 
improvement measures is done twice to ensure that a contract with 
average or above average performance, demonstrated by an improvement 
measure score of zero or above, will receive a measure-level Star 
Ratings of at least 3 stars. A contract whose performance declined, 
demonstrated by an improvement measure score of less than zero, will 
receive a measure-level Star Rating less than 3 stars. Further, CMS 
designed the hold harmless provisions as a safeguard for contracts 
maintaining high performance at the measure-level or at the contract's 
highest Star Rating to ensure that the improvement measure-level Star 
Ratings provide a true signal.
    Comment: A commenter suggested reducing the number of improvement 
measures with a focus on newer measures.
    Response: CMS appreciates this comment. For a measure to be 
included in the improvement calculation, the measure must have numeric 
value scores in both the current and prior year and not have had a 
substantive specification change during those years. In addition, the 
improvement measure will not include any data on measures that are 
already focused on improvement (for example, HOS measures focused on 
improving or maintaining physical or mental health). CMS has focused on 
all measures that meet these criteria to create incentives to improve 
care across a broad spectrum of measures. Limiting the set of measures 
used to determine the improvement measure to strictly new measures has 
the potential of limiting

[[Page 16560]]

the focus of improvement activities by a contract. CMS is committed to 
incentivizing contracts to provide the best quality and care to 
beneficiaries. Striving for continuous improvement across all aspects 
of care would be compromised if the focus of improvement was restricted 
to newer measures only.
    Comment: A commenter suggested that CMS ensure that MA contracts 
that are subject to the use of the improvement measures realize a 
benefit from their inclusion.
    Response: CMS has developed a hold harmless provision for a highly-
rated contract's highest rating. All other contracts have the 
improvement measure(s) included in their rating. CMS believes the 
information provided by the ratings must be a true reflection of the 
quality and experience of beneficiaries enrolled in the contract. 
Ensuring that MA contracts that are subject to the use of the 
improvement measures realize a benefit from their inclusion has the 
potential of distorting the signal and does not align with the Star 
Ratings program's guiding principles.
    Comment: A commenter suggested removing the improvement measure in 
the future to streamline and simplify the Star Ratings program.
    Response: CMS disagrees with the commenter. CMS recognizes the 
importance of acknowledging quality improvement in health and drug 
plans. The improvement measures provide an additional dimension to the 
Star Ratings program. At this time, there are no plans to remove the 
measures from the Star Ratings program as we are committed to improving 
the quality of care and experiences for Medicare beneficiaries.
    Comment: A commenter questioned whether the measures Getting Needed 
Care and Customer Service are included in the improvement measure set.
    Response: Annually, CMS reviews the Star Ratings measure set to 
identify the improvement measures. Both Getting Needed Care and 
Customer Service meet the inclusion criteria for an improvement measure 
and will be designated as improvements measures in the 2021 Star 
Ratings program. A specification change prompted a temporary exclusion 
of these measures from the improvement measure in the 2018 Star 
Ratings.
    Comment: A commenter believes that there exists a potential 
disadvantage for SNPs and Medicare/Medicaid plans due to their 
propensity of having lower enrollments which ultimately results in 
fewer of these types of plans from meeting the requirements for the 
calculation of an improvement measure rating. The issue, the commenter 
believes, is attenuated by the sampling requirements for a subset of 
the population, like the HOS measures.
    Response: CMS appreciates these comments. The contract must have a 
minimum number of numeric scores and measures of a certain type to 
reliably determine an improvement measure score. To date, we have not 
seen an issue with smaller contracts obtaining an improvement measure 
score.
    Comment: Some commenters suggested increased transparency in the 
determination of the improvement measure because of the complexity of 
its determination. Other commenters expressed the concern regarding 
their ability to predict the improvement measure-level Star Ratings. 
Further, commenters requested clearer explanations of the methodology.
    Response: The Star Ratings program is designed to incentivize 
contracts to provide the best care to their beneficiaries. The 
improvement measure employs two consecutive years of data. To realize 
the goal of the best care, contracts must continually seek ways to 
improve the care they provide. The improvement measures provides a 
quantification of the improvement made in the two-year period.
    CMS will apply the methodology explained in the preamble and 
adopted in the regulations at Sec. Sec.  422.164(f) and 423.184(f). The 
improvement methodology is detailed in the annual Technical Notes 
available at http://go.cms.gov/partcanddstarratings. CMS is always 
willing to answer questions related to the calculation of the Star 
Ratings including the improvement measure methodology. Further, upon 
request, CMS will provide a detailed calculation worksheet for a 
contract's improvement measures. Contracts should contact the Part C & 
D Star Ratings Team at PartCandDStarRatings@cms.hhs.gov for answers to 
any questions related to the MA Star Ratings.
    Comment: A commenter urged CMS to review the rules guiding the 
selection of the improvement measures to ensure that each measure is 
under the control of the contract and that the measure is not topped 
out.
    Response: CMS supports the request for reviewing the measures 
designated for use in the improvement measures. CMS annually reviews 
the measures used in the Star Ratings and releases the measures that 
will be used to determine the improvement measures in the draft Call 
Letter. Although some measures may show uniform high performance across 
contracts suggesting that they are topped out, CMS needs to balance 
these concerns with how critical the measures are to improving care, 
the importance of not creating incentives for a decline in performance 
after the measures transition out of the Star Ratings, and the 
availability of alternative related measures which address the specific 
clinical concerns. MAOs and Part D sponsors have control over all 
measures included in the Star Ratings' program; thus, the measures 
selected for the improvement measure(s) are all under the control of 
the contract.
    Comment: A commenter suggested several adjustments to address their 
belief that the improvements measure is based on the following 
perceived flawed assumptions: all plans have the same opportunity to 
improve on both mature and new measures year after year; high- and low-
performing plans have equal opportunity for improvement; and the hold 
harmless provision protects plans. The suggested adjustments included: 
The use of a log scale for evaluating performance instead of a linear 
scale; weighting improvement achieved relative to current performance; 
and adjusting the threshold for significant improvement. (The commenter 
suggested changing the level of significance to 0.025 as opposed to 
0.05, or in other words employing the threshold of 1.645 instead of 
1.96 in the testing for significance.)
    Response: CMS appreciates the comments and the suggested 
enhancements for the improvement measure methodology. CMS remains 
cognizant of the additional challenges for improvement for contracts 
with high performance on their highest rating and at the individual 
measure level. CMS does not believe the underlying assumptions for the 
methodology for the determination of the improvement measure-level Star 
Ratings is flawed. There is less room for improvement for contracts 
that are highly-rated, thus there is a hold harmless provision for a 
contract's highest rating. In addition, there is less room for 
improvement for a measure score if a contract is performing at the 
highest rating, 5 stars, for each of the two consecutive years examined 
for the improvement score. CMS implemented a hold harmless provision at 
the measure level to ensure a contract receiving 5 stars for each year 
of the two years examined would not be subject to the possible 
categorization of a significant decline for the measure.
    At this time, CMS employs a level of significance of 0.05 for all 
significance testing across the aspects of the methodology. The use of 
a 0.05 level of significance is typical for statistical analyses. CMS 
will consider the

[[Page 16561]]

suggestions as we enhance the Star Ratings methodology to best address 
the concerns of our stakeholders while maintaining the integrity of the 
Star Ratings system.
    Comment: Some commenters suggested that the improvement measures 
should consider measure-level Star Ratings and the measure score in the 
hold harmless provision. Some commenters provided examples of an 
increase in a measure-level Star Rating for a specific measure used in 
the improvement measure that was accompanied by a significant decrease 
in the measure score. Commenters believe that such scenarios should be 
part of the hold harmless provision or considered counted as an not 
applicable (NA) measure, those not factoring in the determination of 
the improvement measure score.
    Response: CMS will consider a potential enhancement to the hold 
harmless provision that considers both the measure-level Star Rating 
and the measure score. Any changes would be proposed through future 
rulemaking.
    Comment: Some commenters suggested that a measure that receives 5 
stars for each of the two years should be a positive influence on the 
improvement measure score and counted as a significant improvement.
    Response: CMS appreciates this feedback. A measure used for the 
determination of the improvement measure score that receives a measure-
level Star Rating of 5 stars in each of the two years examined would be 
subject to the 5-star measure hold harmless rule and would benefit from 
the 5-star measure-level Star Rating in the calculation of the summary 
or overall rating. In addition, contracts do have the opportunity to 
earn a reward factor for high and stable relative performance across 
measures pursuant to Sec. Sec.  422.166(f)(1) and 423.186(f)(1) 
discussed in section II.A.11.s of this final rule.
    Comment: Some commenters recommended a predictable gold standard be 
established for determining meaningful improvement as a set percentage 
reduction of a sub-optimal measure rate. The commenters believe this 
approach would result in a more tailored approach of meaningful 
improvement per contract and recognize the natural concept of 
diminishing returns. For example, if a 5 percent reduction in the sub-
optimal rate was classified as meaningful, an increase of 1 percent for 
a contract whose rate was 80 percent in year 1 would be a meaningful 
improvement (1/(100 - 80) or 5 percent) while a contract with a rate of 
60 percent in year 1 would need an increase in their rate of 2 percent 
(an increase to 62 percent) for a 5 percent reduction which would be 
classified as a meaningful reduction in their suboptimal rate (2/(100 - 
60) or 5 percent).
    Response: We will consider these comments for the future as we make 
enhancements to this measure.
    Comment: A few commenters recommended that CMS either adjust its 
methodology and assign ``not applicable'' when determining 
``Improvement, Decline, or No Change'' for measures that increased in 
measure-level Star Ratings in the year two of the comparison or add 
these measures to the ``held harmless'' provision for measures. The 
commenters noted that the current methodology for a measure is based on 
measure scores as opposed to measure-level Star Ratings.
    Response: CMS appreciates this feedback. CMS will further consider 
the measure-level hold harmless provisions to examine the influence of 
the measure scores and measure-level Star Ratings on the improvement 
measures.
    Comment: Some commenters supported a revision to the hold harmless 
measure provision for an improvement measure when a contract received 
5-star ratings for each of the 2 years examined. Although the 
commenters believe that the current measure-level hold harmless does 
align with its intent to prevent an adverse impact on a contract's 
rating, a few commenters suggested modifying the provision to allow a 
measure-level Star Rating of 5 stars for each of the 2 years examined 
to be counted as a significant improvement in the measure's associated 
net improvement category. Other commenters suggested a hold harmless 
provision if mathematically it is not possible to have a 5-star measure 
score difference that would be classified as significant improvement. A 
commenter suggested another version of a measure-level hold harmless in 
which an adjustment factor would be employed for contracts that had 
incremental improvement at the measure-level score but who could not 
attain ``Significant Improvement'' due to performance requirements 
above 100 percent (mathematically) and when the current measure-level 
hold harmless provision would not be applied. Further, the commenter 
believes the adjustment factor would acknowledge the increased 
difficulty in moving from 2 to 3 versus 4 to 5.
    Response: CMS appreciates this feedback. CMS will further consider 
these suggestions for a future enhancement to the hold harmless 
provision at the measure-level.
    Comment: A commenter suggested using a logarithmic scale instead of 
a linear scale in the significance testing for classifying significant 
changes to the measure score to address the law of diminishing returns.
    Response: CMS appreciates the careful consideration of the 
improvement measure methodology. CMS is cognizant of the additional 
challenges for both highly-rated contracts and contracts that receive a 
5 star measure-level rating for each of the two years examined used 
determining the improvement measure. Improvement is easier at the 
summary levels for a contract that is not highly-rated. Likewise, 
improvement for an individual measure is easier when there is more room 
for improvement.
    The current hold harmless provisions were designed to address the 
concern related to the concept of diminishing returns. The improvement 
measure safeguards for contracts at the highest-rating level by 
contract-type and at the measure-level determination of the improvement 
scores allow a transparent method of addressing the challenges of 
improvement for high performing contracts.
    The suggested use of a logarithmic scale instead of a linear scale 
will be considered during our ongoing review of the methodology. Any 
enhancements to the methodology must be balanced by the approachability 
of the methodology to our stakeholders including the beneficiaries.
    Comment: A commenter suggested creating an improvement score for 
measures that could potentially be part of the improvement measures, 
but only have one year's worth of data. The commenter noted that 
improvement activities begin during the first year of a measure being 
included in the Star Ratings program. The focus on a first year measure 
coupled with the significant impact of the improvement measure on a 
contract's rating according to the commenter justified first year 
measures being included in the improvement measure.
    Response: CMS has designed the improvement measures to assess the 
level of improvement from one year to the next.
Summary of Regulatory Changes
    For the reasons set forth in the proposed rule and our responses to 
the related comments summarized above, we are finalizing the 
improvement measure provisions as proposed in Sec. Sec.  422.164(f) and 
423.184(f) with minor modifications. First, in the regulation text at 
Sec. Sec.  422.164(f)(4)(vi) and 423.184(f)(4)(vi), we have corrected 
the

[[Page 16562]]

cross reference to Sec. Sec.  422.166(a)(2)(i) through (iii) and 
422.186(a)(2)(i) through (iii) for the clustering of the improvement 
measure to clarify the methodology for converting the improvement 
measure scores to measure-level Star Ratings. Second, we are also 
finalizing Sec.  422.164(f)(4)(vi) without the sentence that provided 
for separate measure thresholds for the Part D improvement score for 
MA-PDs and PDPs in favor of revising the first sentence as follows: 
``The Part D improvement measure cut points for MA-PDs will be 
determined using separate clustering algorithms in accordance with 
Sec. Sec.  422.166(a)(2)(i) through (iii) and 423.186(a)(2)(i) through 
(iii) of this chapter.''
k. Data Integrity
    The data underlying a measure score and rating must be complete, 
accurate, and unbiased for it to be useful for the purposes we have 
proposed at Sec. Sec.  422.160(b) and 423.180(b). As part of the 
current Star Ratings methodology, all measures and the associated data 
have multiple levels of quality assurance checks. Our longstanding 
policy has been to reduce a contract's measure rating if we determine 
that a contract's measure data are incomplete, inaccurate, or biased. 
Data validation is a shared responsibility among CMS, CMS data 
providers, contractors, and Part C and D sponsors. When applicable (for 
example, data from the IRE, PDE, call center), CMS expects sponsoring 
organizations to routinely monitor their data and immediately alert CMS 
if errors or anomalies are identified so CMS can address these errors.
    We proposed to codify at Sec. Sec.  422.164(g) and 423.184(g) 
specific rules for the reduction of measure ratings when CMS identifies 
incomplete, inaccurate, or biased data that have an impact on the 
accuracy, impartiality, or completeness of data used for the impacted 
measures. Data may be determined to be incomplete, inaccurate, or 
biased based on a number of reasons, including mishandling of data, 
inappropriate processing, or implementation of incorrect practices that 
impacted specific measure(s). One example of such situations that give 
rise to such determinations includes a contract's failure to adhere to 
HEDIS, HOS, or CAHPS reporting requirements. Our modifications to 
measure-specific ratings due to data integrity issues are separate from 
any CMS compliance or enforcement actions related to a sponsor's 
deficiencies. This policy and these rating reductions are necessary to 
avoid falsely assigning a high star to a contract, especially when 
deficiencies have been identified that show we cannot objectively 
evaluate a sponsor's performance in an area.
    As a standard practice, we check for flags that indicate bias or 
non-reporting, check for completeness, check for outliers, and compare 
measures to the previous year to identify significant changes which 
could be indicative of data issues. CMS has developed and implemented 
Part C and Part D Reporting Requirements Data Validation standards to 
assure that data reported by sponsoring organizations pursuant to 
Sec. Sec.  422.516 and 423.514 satisfy the regulatory obligation. 
Sponsor organizations should refer to specific guidance and technical 
instructions related to requirements in each of these areas. For 
example, information about HEDIS measures and technical specifications 
is posted on: http://www.ncqa.org/HEDISQualityMeasurement/HEDISMeasures.aspx. Information about Data Validation of Reporting 
Requirements data is posted on: https://www.cms.gov/Medicare/Prescription-Drug-Coverage/PrescriptionDrugCovContra/PartCDDataValidation.html and https://www.cms.gov/Medicare/Prescription-Drug-Coverage/PrescriptionDrugCovContra/RxContracting_ReportingOversight.html.
    We proposed, in paragraphs (g)(1)(i) through (iii), rules for 
specific circumstances where we believe a specific response is 
appropriate. First, we proposed a continuation of a current policy: To 
reduce HEDIS measures to 1 star when audited data are submitted to NCQA 
with an audit designation of ``biased rate'' or BR based on an 
auditor's review of the data if a plan chooses to report; this proposal 
will also apply when a plan chooses not to submit and has an audit 
designation of ``non-report'' or NR. Second, we proposed to continue to 
reduce Part C and D Reporting Requirements data, that is, data required 
pursuant to Sec. Sec.  422.514 and 423.516, to 1 star when a contract 
did not score at least 95 percent on data validation for the applicable 
reporting section or was not compliant with data validation standards/
sub-standards for data directly used to calculate the associated 
measure. In our view, data that do not reach at least 95 percent on the 
data validation standards are not sufficiently accurate, impartial, and 
complete for use in the Star Ratings. We explained in the preamble that 
as the sponsoring organization is responsible for these data and 
submits them to CMS, a negative inference is appropriate, to conclude 
that performance is likely poor. Third, we proposed a new specific rule 
to implement scaled reductions in Star Ratings for appeal measures in 
both Part C and Part D.
    The data downgrade policy was adopted to address instances when the 
data that will be used for specific measures are not reliable for 
measuring performance due to their incompleteness or biased/erroneous 
nature. For instances where the integrity of the data is compromised 
because of the action or inaction of the sponsoring organization (or 
its subcontractors or agents), this policy reflects the underlying 
fault of the sponsoring organization for the lack of data for the 
applicable measure. Without some policy for reduction in the rating for 
these measures, sponsoring organizations could ``game'' the Star 
Ratings and merely fail to submit data that illustrate poor 
performance. As stated in the proposed rule, we believe that removal of 
the measure from the ratings calculation will unintentionally reward 
poor data compilation and submission activities such that our only 
recourse is to reduce the rating to 1 star for affected measures.
    For verification and validation of the Part C and D appeals 
measures, we proposed to use statistical criteria to determine if and 
how a contract's appeals measure-level Star Ratings would be reduced 
for missing IRE data. We explained that the proposed criteria would 
allow us to use scaled reductions for the appeals measures to account 
for the degree to which the data are missing. The completeness of the 
IRE data is critical to allow fair and accurate measurement of the 
appeals measures. All plans are responsible and held accountable for 
ensuring high quality and complete data to maintain the validity and 
reliability of the appeals measures.
    In response to past stakeholder concerns about CMS's prior practice 
of reducing measure ratings to one star based on any finding of data 
inaccuracy, incompleteness, or bias, CMS initiated the Timeliness 
Monitoring Project, TMP, in CY 2017.\55\ The first submission for the 
TMP was for the measurement year 2016 related to Part C organization 
determinations and reconsiderations and Part D coverage determinations 
and redeterminations. The timeframe for the submitted data was 
dependent on the enrollment of the contract, with smaller

[[Page 16563]]

contracts submitting data from a 3-month period, medium-sized contracts 
submitting data from a two-month period, and larger contracts 
submitting data from a one-month period.\56\
---------------------------------------------------------------------------

    \55\ This project was discussed in the November 28, 2016 HPMS 
memo, ``Industry-wide Appeals Timeliness Monitoring.'' https://www.cms.gov/Medicare/Prescription-Drug-Coverage/PrescriptionDrugCovGenIn/Downloads/Industry-wide-Timeliness-Monitoring.pdf. https://www.cms.gov/Medicare/Prescription-Drug-Coverage/PrescriptionDrugCovGenIn/Downloads/Industry-wide-Appeals-Timeliness-Monitoring-Memo-November-28-2016.pdf.
    \56\ Contracts with a mean annual enrollment of less than 50,000 
are required to submit data for a three-month time period. Contracts 
with a mean enrollment of at least 50,000 but at most 250,000 are 
required to submit data for a two-month time period. Contracts with 
a mean enrollment greater than 250,000 are required to submit data 
for a one-month period.
---------------------------------------------------------------------------

    We proposed to use TMP data and other data sources whenever 
possible, such as information from audits, to determine whether the 
data at the Independent Review Entity (IRE) are complete and to 
evaluate the level of missing data. Given the financial and marketing 
incentives associated with higher performance in Star Ratings, 
safeguards are needed to protect the Star Ratings from actions that 
inflate performance or mask deficiencies.
    We proposed to reduce a contract's Part C or Part D appeal measures 
Star Ratings for IRE data that are not complete or otherwise lack 
integrity based on the TMP or audit information. The reduction would be 
applied to the measure-level Star Ratings for the applicable appeals 
measures. There are varying degrees of data issues and as such, we 
proposed a methodology for reductions that reflects the degree of the 
data accuracy issue for a contract instead of a one-size fits all 
approach. The proposed methodology employs scaled reductions, ranging 
from a 1-star reduction to a 4-star reduction; the most severe 
reduction for the degree of missing IRE data would be a 4-star 
reduction which will result in a measure-level Star Rating of 1 star 
for the associated appeals measures (Part C or Part D). The data source 
for the scaled reduction is the TMP or audit data, however the specific 
data used for the determination of a Part C IRE data completeness 
reduction are independent of the data used for the Part D IRE data 
completeness reduction. If a contract receives a reduction due to 
missing Part C IRE data, the reduction would be applied to both of the 
contract's Part C appeals measures. Likewise, if a contract receives a 
reduction due to missing Part D IRE data, the reduction would be 
applied to both of the contract's Part D appeals measures. We solicited 
comment on this proposal and its scope; we were looking in particular 
for comments related to how to use the process in this proposal to 
account for data integrity issues discovered through means other than 
the TMP and audits of sponsoring organizations.
    CMS's proposed scaled reduction methodology is a three-stage 
process using the TMP or audit information to determine: first, whether 
a contract may be subject to a potential reduction for the Part C or 
Part D appeals measures; second, the basis for the estimate of the 
error rate; and finally, whether the estimated error rate is 
significantly greater than the cut points for the scaled reductions of 
1, 2, 3, or 4 stars.
    Once the scaled reduction for a contract is determined using this 
methodology, the reduction is applied to the contract's associated 
appeals measure-level Star Ratings. The minimum measure-level Star 
Rating is 1 star. If the difference between the associated appeals 
measure-level Star Rating (before the application of the reduction) and 
the identified scaled reduction is less than one, the contract receives 
a measure-level Star Rating of 1 star for the appeals measure.
    Under the proposed methodology, the error rate for the Part C and 
Part D appeals measures using the TMP or audit data and the projected 
number of cases not forwarded to the IRE for a 3-month period is used 
to identify contracts that may be subject to an appeals-related IRE 
data completeness reduction. We proposed a minimum error rate to 
establish a threshold for the identification of contracts that may be 
subject to a reduction. The establishment of the threshold focuses the 
possible reductions on contracts with error rates that have the 
greatest potential to distort the signal of the appeals measures. Since 
the timeframe for the TMP data is dependent on the enrollment of the 
contract, (with smaller contracts submitting data from a 3-month 
period, medium-sized contracts submitting data from a 2-month period, 
and larger contracts submitting data from a one-month period), the use 
of a projected number of cases over a 3-month period allows a 
consistent time period for the application of the criteria proposed.
    The calculated error rate formula (Equation 1) for the Part C 
measures is determined by the quotient of the number of cases not 
forwarded to the IRE and the total number of cases that should have 
been forwarded to the IRE. The number of cases that should have been 
forwarded to the IRE is the sum of the number of cases in the IRE 
during TMP or audit data collection period and the number of cases not 
forwarded to the IRE during the same period.
[GRAPHIC] [TIFF OMITTED] TR16AP18.010

    The calculated error rate formula (Equation 2) for the Part D 
measures is determined by the quotient of the number of untimely cases 
not auto-forwarded to the IRE and the total number of untimely cases.
[GRAPHIC] [TIFF OMITTED] TR16AP18.020

    Under the proposed methodology, the projected number of cases not 
forwarded to the IRE in a 3-month period is calculated by multiplying 
the number of cases found not to be forwarded to the IRE based on the 
TMP or audit data by a constant determined by the TMP time period. 
Contracts with mean annual enrollments greater than 250,000 that 
submitted data from a 1-month period would have their number of cases 
found not to be forwarded to the IRE based on the TMP data multiplied 
by the constant 3.0. Contracts with mean enrollments of 50,000 but at 
most 250,000 that submitted data from a 2-month period would have their 
number of cases found not to be forwarded to the IRE based on the TMP 
data multiplied by the constant 1.5. Small contracts with mean 
enrollments less than 50,000 that submitted data for a 3-month period 
would have their number of cases found not to be forwarded to the IRE 
based on the TMP data multiplied by the constant 1.0.

[[Page 16564]]

    We proposed that contract ratings be subject to a possible 
reduction due to lack of IRE data completeness if both following 
conditions are met:
     The calculated error rate is 20 percent or more.
     The projected number of cases not forwarded to the IRE is 
at least 10 in a 3-month period.
    The requirement for a minimum number of cases is needed to address 
statistical concerns with precision and small numbers. If a contract 
meets only one of the conditions, the contract would not be subject to 
reductions for IRE data completeness issues.
    If a contract is subject to a possible reduction based on the 
aforementioned conditions, a confidence interval estimate for the true 
error rate for the contract is calculated using a Score Interval 
(Wilson Score Interval) at a confidence level of 95 percent.
    The midpoint of the score interval will be determined using 
Equation 3.
[GRAPHIC] [TIFF OMITTED] TR16AP18.011

    The z score that corresponds to a level of statistical significance 
of 0.05, commonly denoted as
[GRAPHIC] [TIFF OMITTED] TR16AP18.022

but for ease of presentation represented here as z. (The z value that 
will be used for the purpose of the calculation of the interval is 
1.959964.).
    For the Part C appeals measures, the midpoint of the confidence 
interval is calculated using Equation 3 along with the calculated error 
rate from the TMP, which is determined by Equation 1. The total number 
of cases in Equation 3 is the number of cases that should have been in 
the IRE for the Part C TMP data.
    For the Part D appeals measures, the midpoint of the confidence 
interval is calculated using Equation 3 along with the calculated error 
rate from the TMP, which is determined by Equation 2. The total number 
of cases in Equation 3 is the total number of untimely cases for the 
Part D appeals measures.
    Letting the calculated error rate be represented by p and the total 
number of cases represented as n, Equation 3 can be streamlined as 
Equation 4:
[GRAPHIC] [TIFF OMITTED] TR16AP18.012

    The lower bound of the confidence interval estimate for the error 
rate is calculated using Equation 5 below:
[GRAPHIC] [TIFF OMITTED] TR16AP18.021

    For each contract subject to a possible reduction, the lower bound 
of the interval estimate of the error rate will be compared to each of 
the thresholds in Table 4. If the contract's calculated lower bound is 
higher than the threshold, the contract will receive the reduction that 
corresponds to the highest threshold that is less than the lower bound. 
In other words, the contract's lower bound is being employed to 
determine whether the contract's error rate is significantly greater 
than the thresholds of 20 percent, 40 percent, 60 percent, and 80 
percent. The proposed scaled reductions are in Table 4, and were 
proposed in narrative form at paragraph (g)(1)(iii)(D) of both 
regulations.
    We further proposed that the reductions due to IRE data 
completeness issues be applied after the calculation of the measure-
level Star Rating for the appeals measures. The proposed reduction 
would be applied to the Part C appeals measures and/or the Part D 
appeals measures.
    We noted in the proposed rule that a contract's lower bound could 
be statistically significantly greater than more than one threshold. We 
proposed that the reduction be determined by the highest threshold that 
the contract's lower bound exceeds. For example, if the lower bound for 
a contract is 64.560000 percent, the contract's estimated value is 
significantly greater than the thresholds of 20 percent, 40 percent, 
and 60 percent because the lower bound value 64.560000 percent is 
greater than each of these thresholds. The lower bound for the 
contract's confidence interval is not greater than 80 percent. 
Therefore, in this example, the contract will be subject to the 
reduction that corresponds to the 60 percent threshold, which is three 
stars.

 Table 4--Appeals Measure Star Ratings Reductions by the Incomplete Data
                               Error Rate
------------------------------------------------------------------------
                                                           Reduction for
 Proposed thresholds using the lower bound of confidence  incomplete IRE
         interval estimate of the error rate (%)           data (stars)
------------------------------------------------------------------------
20......................................................               1
40......................................................               2
60......................................................               3
80......................................................               4
------------------------------------------------------------------------

    We proposed regulation text at Sec.  422.164(g)(1)(iii)(A) through 
(N) and Sec.  423.184(g)(1)(iii)(A) through (K) to codify these 
parameters and formulas for the scaled reductions. We noted in the 
proposed rule that the proposed text for the Part C regulation includes 
specific paragraphs related to MA and MA-PD plans that are not included 
in the proposed text for the Part D regulation but that the two are 
otherwise identical.
    In addition, we proposed in Sec. Sec.  422.164(g)(2) and 
423.184(g)(2) to authorize reductions in a Star Rating for

[[Page 16565]]

a measure when there are other data accuracy concerns (that is, those 
not specified in paragraph (g)(1)). We proposed an example in paragraph 
(g)(2) of another circumstance where CMS will be authorized to reduce 
ratings based on a determination that performance data are incomplete, 
inaccurate, or biased: the failure of a contract to adhere to the 
HEDIS, CAHPS, or HOS reporting requirements. We also proposed this 
other situation would result in a reduction of the measure rating to 1 
star.
    We noted in the proposed rule that we had taken several steps in 
past years to protect the integrity of the data we use to calculate 
Star Ratings. We welcomed comments about alternative methods for 
identifying inaccurate or biased data and comments on the proposed 
policies for reducing stars for data accuracy and completeness issues 
and comments on the proposed methodology for scaled reductions for the 
Part C and Part D appeals measures to address the degree of missing IRE 
data.
    We received the following comments on our proposals and our 
responses follow:
    Comment: There was overwhelming support for the use of scaled 
reductions for the completeness of the IRE data for the appeals 
measures. Some commenters explicitly stated that the use of scaled 
reductions avoids the one-size-fits-all approach.
    Response: CMS appreciates the overwhelming support for the proposed 
scaled reduction methodology.
    Comment: Some commenters suggested other potential criteria for 
consideration for the scaled reductions methodology. A commenter 
suggested CMS consider the volume of appeals instead of plan size for 
determining the reductions. Other commenters suggested including 
enrollment as part of the rules for the allowable excluded number of 
cases, using the timely percentages as basis for scaled reduction, or 
using the errors relative to enrollment level as the thresholds.
    Response: CMS appreciates the careful consideration of alternative 
options for the scaled reduction methodology. A thorough examination 
and identification of potential unintended consequences must be done 
for any possible modification to the Star Ratings methodology. 
Additional analysis will be done to further explore relations among 
enrollment, appeals volume, untimely, and timely percentages. CMS 
believes the proposed methodology provides the best foundation for 
scaled reductions and will consider these comments as we contemplate 
future enhancements.
    Comment: Some commenters expressed support for the data integrity 
policies for non-appeals measures. A commenter supported the proposal 
to reduce a contract's measure-level Star Rating to 1-star for measures 
related to Part C and D reporting requirements measures when the 
contract does not meet CMS expectations for data validation. Another 
commenter supported the reduction for HEDIS measures that received an 
audit designation of ``Biased Rate.'' Another commenter supported the 
high standard of 95 percent on validation audits, but believed it is 
important to distinguish between generally well-functioning plans that 
may have an occasional error versus plans that have significant, 
systematic errors.
    Response: CMS appreciates the support of the data integrity 
policies. The data integrity policies align with our commitment to data 
quality and preserves the integrity of the Star Ratings. CMS believes 
the data integrity policies are designed to distinguish between 
occasional errors and systematic issues. For example, both the 
validation audits and scaled reduction methodology allow for the 
occasional error and target only those contracts that exceed a 
specified error rate.
    Comment: A commenter requested clarification on how CMS plans to 
use the Data Validation Audit.
    Response: The Data Validation Audit is one method to ensure the 
data used for Star Ratings are accurate. The two Star Rating measures 
(SNP Care Management (Part C) and Medication Therapy Management (MTM) 
Program Completion Rate for Comprehensive Medication Reviews (CMR) 
(Part D)) are based on Part C and D Reporting Requirements data and 
calculated using data reported by plan sponsors and validated via an 
independent data validation using CMS standards. Per the Star Ratings 
Technical Notes, contracts that did not score at least 95 percent on 
data validation for these reporting sections and/or were not compliant 
with data validation standards/sub-standards for at least one of the 
data elements used to calculate the measures are not rated in this 
measure, and the contract's measure score is reduced to 1 star. CMS has 
relied on the Data Validation Audit to confirm the integrity of these 
plan-reported data since these measures were first added to the Star 
Ratings program. In the 2019 draft Call Letter CMS proposed to define a 
contract as being non-compliant if it either receives a ``No'' or a 1, 
2, or 3 on the 5-point Likert scale in the specific data element's data 
validation in order to align with changes in the Data Validation Audit.
    If further clarification is needed, please feel free to contact the 
Part C&D Data Quality Team at: PARTCDQA@cms.hhs.gov
    Comment: Some commenter expressed concern or opposed using audit 
findings as a data source to validate the appeals measures.
    Response: The Timeliness Monitoring Project (TMP) data will be the 
primary data used to validate the completeness for the Part C and D 
appeals measures. However, CMS may also use audit data to validate the 
appeals measures if additional information is uncovered during the 
audit process that demonstrates that the data for the appeals measures 
are not complete.
    Comment: A commenter requested clarification regarding the use of 
TMP data that are submitted at the parent-organization level. 
Specifically, the commenter was unsure if the reporting level would be 
at contract level or all contracts under the parent organization would 
receive the same scaled reduction.
    Response: Although the data for the TMP are submitted by the parent 
organization, the observations are recorded at the contract level. The 
TMP data for each parent organization are disaggregated to contract-
level data. The scaled reduction would be separately and independently 
determined for each contract under a parent organization. If a contract 
has no untimely cases or no cases that should have been forwarded to 
the IRE in the TMP timeframe, the contract would not be subject to a 
possible IRE data completeness reduction for the associated appeals 
measure. This analysis would be done on a contract-by-contract basis 
using only data for the applicable contract.
    Comment: A commenter expressed concern about the lack of a data-
driven methodology used to determine data integrity issues. Further the 
commenter asked for a data-driven, streamlined approach that does not 
use audit data.
    Response: The Star Ratings program and its associated methodology 
generally employ a comprehensive, scientific, data-driven approach. CMS 
has moved away from relying on audit data for determining the 
completeness of the appeals measures with the introduction of the TMP 
data. However, we are not adopting a rule to prohibit use of audit data 
where such data are reliable and relevant to understanding and 
determining whether the data used for a particular measure (even 
appeals measures) are erroneous, incomplete or biased.
    Comment: Some commenters requested additional information on the 
timeline for contracts to submit

[[Page 16566]]

information on scaled reductions along with simulations to allow 
contracts to better understand the impact of the scaled reduction 
methodology. Another commenter requested that CMS share all simulated 
data related to scaled reductions.
    Response: CMS will issue a memo each year outlining the timeframe 
associated with the TMP data collection. The TMP data used for the Star 
Ratings program will align with the measurement period of the Star 
Ratings year.
    The first submission for the TMP focused on the 2016 measurement 
year for Part C organization determinations and reconsiderations and 
Part D coverage determinations and redeterminations. CMS gained 
valuable insight about the audit universes, and the completeness of the 
IRE data.
    In December 2017, CMS provided each contract with the results of 
its TMP analysis. The Part C and D IRE data completeness percentage 
provided is equivalent to the calculated error rate discussed in the 
scaled reduction methodology section outlined in the NPRM. A contract 
can simulate the scaled reduction for the 2018 Star Ratings appeals 
measures by following the methodology for scaled reductions. First, a 
contract can use the data provided to determine whether it would be 
subject to a possible reduction due to lack of IRE data completeness 
based on the calculated error rate and projected number of cases not 
forwarded to the IRE. (To determine the projected number of cases the 
factor based on the enrollment needs to be multiplied by the number of 
cases detailed on the December report.) Next, if the contract is 
subject to a possible reduction, the lower bound of the Wilson Score 
interval is calculated using the formulas in the NPRM along with the 
calculated error rate. The lower bound can then be compared to the 
thresholds in Table 3 to identify the reduction to the associated 
appeals measure-level Star Ratings.
    Comment: A commenter did not believe the exclusion of a measure 
affected by data integrity issues is sufficient to prevent 
gamesmanship. Instead, the commenter suggested a hybrid approach that 
the commenter believes is less punitive. This method would exclude 
measures that received 4 or 5 stars and would levy an automatic 
reduction to 1 star for data integrity issues for measures that 
received 3 or less stars.
    Response: The accuracy of the measure data is key to the Star 
Ratings methodology. Excluding a measure from the Star Ratings due to 
data integrity issues instead of using a measure-level Star Rating of 1 
distorts the signal of the true quality and performance of a contract 
and does not align with the intent of the data integrity policies. We 
therefore disagree with the commenter.
    Comment: Some commenters supported expanding polices to reduce Star 
Ratings when the data are not reported or do not meet validation 
requirements. A few commenters suggested the use of scaled reductions 
for all measures in the Star Ratings program including HEDIS measures. 
Another commenter supported expanding the scaled reductions to other 
measures with special consideration of organizations demonstrating 
commitment to compliance.
    Response: CMS appreciates the support of the data integrity policy 
and will consider expanding the policies to be as comprehensive as 
feasible. Currently, for most measures, including HEDIS measures, we do 
not have enough information to calculate scaled reductions.
    Comment: Some commenters expressed concern regarding the possible 
use of audit data. The commenters stated that using audit data results 
in artificially inflated ratings for contracts that are not audited 
compared to contracts that are audited. A commenter stated the goals 
and analytic approaches associated with an audit do not align with 
those of the Star Ratings program. In addition, a commenter wanted any 
findings from enforcement activities excluded from the Star Ratings 
since not all contracts are audited each year. A commenter requested 
information about how CMS would ensure equity between audited and non-
audited contracts. In addition, another commenter asked for 
clarification of the `other data' that may be used for assessing data 
completeness. A commenter encouraged CMS immediately remove the impact 
of audit findings on the Star Ratings for the determination of 2019 
QBPs.
    Response: CMS appreciates the comments. All contracts are required 
to submit TMP data on an annual basis. The TMP data are typically the 
same data used for CMS program audits but are collected from all MA and 
Part D sponsoring organizations which shall ensure equity among all 
contracts. As part of the 2019 draft Call Letter, CMS proposed to 
remove the Beneficiary Access and Performance Problems (BAPP) measure 
from the Star Ratings. This proposal was finalized in the 2019 Final 
Call Letter to remove the BAPP measure from the Star Ratings program 
effective for the 2019 Star Ratings.
    Comment: A commenter suggested a hold harmless provision when there 
are data issues. The commenter provided the example of the measure of 
providing translation services that was removed from the Star Ratings 
in the past for contracts that have worked hard to perform well on a 
measure.
    Response: CMS removes measures from the Star Ratings if a 
systematic issues exists with data quality across all (or a majority 
of) contracts as described in Sec. Sec.  422.164(b) and 423.184(b). It 
is the policy of CMS not to assign measure-level Star Ratings if data 
issues are present across the board that suggest that the measure 
results are not reliable. When systemic data issues are present for a 
measure, it is difficult to accurately determine performance across 
contracts. The policy proposed for adding, updating and removing 
measures is presented in Sec. Sec.  422.164 and Sec. Sec.  423.184. The 
removal of measures from the Star Ratings is detailed in in Sec. Sec.  
422.164(b) and Sec. Sec.  423.184(b).
    Comment: A few commenters expressed concern that the CMS approach 
for data integrity issues for HEDIS measures is duplicative of the 
HEDIS audit process.
    Response: The data integrity policy for HEDIS measures uses the 
information provided by the NCQA compliance auditor, and thus aligns 
with their findings.
    Comment: A commenter stated that the reductions in the Star Ratings 
for integrity blurs the distinction between quality measurement and 
compliance and audit activities. Further, the commenter stated that the 
focus of the ratings should be clinical quality and beneficiary 
satisfaction. Another commenter expressed concern of the continuation 
of the downgrade to 1-star for the HEDIS and measures related to the 
Part C&D reporting requirements.
    Response: CMS considers data quality as paramount to accurate and 
reliable measurement. As such, CMS uses multiple sources of information 
to assess the multiple facets of data quality. The Star Ratings were 
designed to provide a true signal of the quality and performance of a 
contract. Star Ratings that are generated from data that lack quality 
or, in other words, flawed data--whether because of bias, 
incompleteness, or inaccuracy--impact the integrity of the ratings. 
Star Ratings that do not provide a true signal of the quality and 
performance of the Medicare health and drugs plans offered under a 
contract threaten the core of the Star Ratings program. CMS is 
committed to maintaining the integrity of the ratings. By taking steps 
to downgrade measure ratings when underlying data

[[Page 16567]]

quality issues exists, CMS is preserving the integrity of the Star 
Ratings and incentivizing sponsoring organizations to take steps to 
improve data integrity and eliminate problems.
    Comment: Some commenters suggested modifications to other facets of 
the data integrity policy. A commenter suggested that if an identified 
data issue did not harm beneficiaries, plans should be able to resubmit 
the data with limited penalty. Other commenters stated that CMS should 
provide contracts the opportunity to correct data errors without 
penalties. A commenter suggested that contracts should be offered a 
preliminary review of their data midway through the reporting year to 
allow identification of any issues and the chance to correct them 
before the end of the year. Another commenter suggested that CMS take 
into account the necessary distinction between a deliberate submission 
of inaccurate data and the unintentional occurrence of minor errors and 
mistakes when addressing data integrity. In addition, the commenter 
outlined an approach to penalize plans based on beneficiary impact, 
nature of issue, health plan activity, history of data integrity 
issues, and timing that would be reviewed by a third party.
    Response: CMS appreciates the careful consideration and suggestions 
for potential revisions to the data integrity policy. The data 
underlying a measure score and rating must be complete, accurate, and 
unbiased to allow the Star Ratings to be a true reflection of a 
contract's quality and performance. CMS's longstanding policy has been 
to reduce a contract's measure rating if a contract's measure data are 
incomplete, inaccurate, or biased but, as the proposal of scaled 
reductions indicates, CMS will consider and implement alternatives and 
improvements. We must, however, remain mindful of the timing and 
resource considerations at play with the annual release of Star 
Ratings.
    Data validation is a shared responsibility among CMS, CMS data 
providers, contractors, and Part C and D sponsors. CMS encourages 
organizations to routinely monitor their data and immediately alert CMS 
if errors or anomalies are identified so CMS can address these errors. 
Contracts are afforded opportunities to review their data before the 
Star Ratings are calculated, during data collection and during the Plan 
Preview periods for the Star Ratings. CMS will continue to review the 
policies and solicit feedback from stakeholders.
    Comment: A few commenters expressed concern about the perceived 
punitive nature of the data integrity policies. A commenter suggested 
that contracts should be rewarded if they have near-perfect 
performance. For example, the commenter suggested that contracts that 
receive a 5-Star Rating on the Part D Appeals Timeliness measure and do 
not qualify for the Part D Appeals Upheld measure because they received 
less than 10 appeals, should automatically receive a 5-Star Rating on 
the Part D Appeals Upheld measure.
    Response: CMS believes the integrity of the data is fundamental to 
the Star Ratings program. CMS maintains high standards for data quality 
to ensure that the Star Ratings are a true reflection of the quality, 
performance and experience of the beneficiaries enrolled in MA and Part 
D contracts. CMS employs a data-driven approach for determining the 
measure-level Star Ratings. The data integrity policies serve to 
preserve the integrity of the Star Ratings and encourage contracts and 
sponsors to strive for the highest data quality; they are not designed 
or intended to be punitive. The measure level reductions for data 
integrity concerns are not made to punish a sponsor but rather to 
reflect that the data available are incomplete and inaccurate.
    In the commenter's example, the contract did not meet the minimum 
number of cases reviewed by the IRE to be measured in the Appeals 
Upheld measure. This specification is necessary to ensure an adequate 
sample of cases for which to evaluate the contract's original 
decisions. The contract's TMP results regarding the completeness of the 
IRE data has no relevance on whether CMS can evaluate the contract in 
this measure. It remains that CMS cannot reliably calculate a percent 
of cases upheld by the IRE if there are too few IRE cases reviewed for 
the contract.
    Comment: A commenter suggested the removal of the Part C and D 
appeals measures until CMS can adequately address the underlying data 
integrity issues that are associated with the IRE and contracts.
    Response: CMS is firmly committed to the integrity of the Star 
Ratings systems. CMS believes that the data integrity policy and the 
rating reductions are necessary to avoid falsely assigning a high star 
to a contract, especially when deficiencies have been identified that 
show CMS cannot objectively evaluate a sponsor's performance in an 
area. To address challenges in validating the appeals measures, CMS 
implemented the collection of the TMP data. Concerns and reviews to 
assure data integrity will remain for as long as necessary to collect 
data in order to provide reliable Star Ratings and comparable 
information about plan quality and performance. CMS believes that our 
rule, as proposed and finalized, strikes the right balance in support 
of the underlying policies.
    For the reasons set forth in the proposed rule and our responses to 
the related comments summarized above, we are finalizing data integrity 
provisions as proposed at Sec. Sec.  422.164(g) and 423.184(g) without 
substantive modification. We are finalizing the following minor 
editorial changes to the regulation text: (1) In Sec.  
422.164(g)(1)(ii) to add a reference to ``substandards'' as well as 
standards that govern data validation; (2) in Sec.  422.164(g)(1)(iii) 
to improve the flow of the last sentence in the introductory paragraph 
and to correct the verb tenses in paragraphs (A), (C) and (K); (3) in 
Sec.  423.184(g)(1)(i) to identify the data that are subject to data 
validation; (4) in Sec.  423.184(g)(1)(ii) to add the sentence proposed 
as paragraph (ii)(A) to the introductory paragraph and redesignate the 
remaining paragraphs; and (5) in redesignated Sec.  
423.184(g)(1)(ii)(A), (C), and (F) to correct the verb tenses and 
capitalization of ``Star Ratings''. Finally, in Sec.  423.184(g)(1)(ii) 
A-L we aligned the regulatory text with Sec.  422.164(g)(1)(ii) A-N 
where appropriate. Sec.  422.164(g)(1)(ii) A-N has more provisions to 
account for the differences in calculations between Part C and D 
appeals measures.
l. Measure-Level Star Ratings
    We proposed in Sec. Sec.  422.166(a) and 423.186(a) the methods for 
calculating Star Ratings at the measure level. As part of the Part C 
and D Star Ratings system, Star Ratings are currently calculated at the 
measure level. To separate a distribution of scores into distinct 
groups or star categories, a set of values must be identified to 
separate one group from another group. The set of values that break the 
distribution of the scores into non-overlapping groups is a set of cut 
points. We proposed to continue to determine cut points by applying 
either clustering or a relative distribution and significance testing 
methodology; we proposed to codify this policy in paragraphs (a)(1) of 
each section. We proposed in paragraphs (a)(2) and (a)(3) of each 
section that for non-CAHPS measures (including the improvement 
measures, which were specifically addressed in paragraphs (a)(2)(iii), 
we would use a clustering methodology and that for CAHPS measures, we 
would use relative distribution and significance testing. Measure 
scores will be converted to a 5-

[[Page 16568]]

star scale ranging from 1 to 5, with whole star increments. A rating of 
5 stars will indicate the highest Star Rating possible, while a rating 
of 1 star will be the lowest rating on the scale. We proposed to use 
the two methodologies described as follows to convert measure scores to 
measure-level Star Ratings.
    We proposed to use the clustering method for all Star Ratings 
measures, except for the CAHPS measures. For each individual measure, 
we would determine the measure cut points using all measure scores for 
all contracts required to report that do not have missing, flagged as 
biased, or erroneous data. For the Part D measures, we proposed to 
determine MA-PD and PDP cut points separately. The scores would be 
grouped such that scores within the same rating (that is 1 star, 2 
stars, etc.) are as similar as possible, and scores in different 
ratings are as different as possible. The hierarchical clustering 
algorithm and the associated tree and cluster assignments using SAS (a 
statistical software package) are currently used to determine the cut 
points for the assignment of the measure-level Star Ratings. We stated 
that we would continue use of this software, but that improvements in 
statistical analysis would not result in rulemaking or changes in these 
eventual rules providing for the use of a clustering methodology. We 
stated our belief that the software used to apply the clustering 
methodology is generally irrelevant.
    Conceptually, the clustering algorithm identifies natural gaps 
within the distribution of the scores and creates groups (clusters) 
that are then used to identify the cut points that result in the 
creation of a pre-specified number of categories. The Euclidean 
distance between each pair of contracts' measure scores serves as the 
input for the clustering algorithm. The hierarchical clustering 
algorithm begins with each contract's measure score being assigned to 
its own cluster. Ward's minimum variance method is used to separate the 
variance of the measure scores into within-cluster and between-cluster 
sum of squares components in order to determine which pairs of clusters 
to merge. For the majority of measures, the final step in the algorithm 
is done a single time with five categories specified for the assignment 
of individual scores to cluster labels. The cluster labels are then 
ordered to create the 1 to 5-star scale. The range of the values for 
each cluster (identified by cluster labels) is examined. We proposed 
that this final range of values and labels would be used to determine 
the set of cut points for the Star Ratings as follows: The measure 
score that corresponds to the lower bound for the measure-level ratings 
of 2 through 5 will be included in the star-specific rating category 
for a measure for which a higher score corresponds to better 
performance; for a measure for which a lower score is better, the 
process will be the same except that the upper bound within each 
cluster label will determine the set of cut points; the measure score 
that corresponds to the cut point for the ratings of 2 through 5 will 
be included in the star-specific rating category; and in cases where 
multiple clusters have the same measure score value range, those 
clusters will be combined, leading to fewer than 5 clusters. Under our 
proposal to use clustering to set cut points, we stated that we would 
require the same number of observations (contracts) within each rating 
and instead will use a data-driven approach.
    As proposed in paragraphs (a)(2)(iii) of each section the 
improvement measures for Part C and Part D would be determined using 
the hierarchical clustering algorithm twice, once for raw scores of 
zero or greater and again for raw scores below for the identification 
of the cut points that will allow the conversion of the improvement 
measure scores to the star scale. The Part D improvement measure score 
clustering for MA-PDs and PDPs will be reported separately. Improvement 
scores of zero or greater would be assigned at least 3 stars for the 
improvement Star Rating, while improvement scores of less than zero 
would be assigned either 1 or 2 stars. For contracts with improvement 
scores greater than or equal to zero, the clustering process will 
result in three clusters with measure-level Star Ratings of 3, 4, or 5 
with the lower bound of each cluster serving as the cut point for the 
associated Star Rating. For those contracts with improvement scores 
less than zero, the clustering algorithm will result in two clusters 
with measure-level Star Ratings of 1 or 2..
    We proposed in paragraphs (a)(3) of each section to use another 
method using percentile standing relative to the distribution of scores 
for other contracts, measurement reliability standards, and statistical 
significance testing to determine star assignments for the CAHPS 
measures. This method will combine evaluating the relative percentile 
distribution of scores with significance testing and measurement 
reliability standards in order to maximize the accuracy of star 
assignments based on scores produced from the CAHPS survey. For CAHPS 
measures, contracts are first classified into base groups by 
comparisons to percentile cut points defined by the current-year 
distribution of case-mix adjusted contract means. Percentile cut points 
are rounded to the nearest integer on the 0-100 reporting scale, and 
each base group includes those contracts whose rounded mean score is at 
or above the lower limit and below the upper limit. Then, the number of 
stars assigned is determined by the base group assignment, the 
statistical significance and direction of the difference of the 
contract mean from the national mean, an indicator of the statistical 
reliability of the contract score on a given measure (based on the 
ratio of sampling variation for each contract mean to between-contract 
variation), and the standard error of the mean score. Table C4, which 
we proposed to codify in narrative form at Sec. Sec.  422.166(a)(3) and 
423.186(a)(3), details the CAHPS star assignment rules for each rating. 
We proposed that all statistical tests, including comparisons involving 
standard error, would be computed using unrounded scores.
    We proposed that if the reliability of a CAHPS measure score is 
very low for a given contract, less than 0.60, the contract would not 
receive a Star Rating for that measure. For purposes of applying the 
criterion for 1 star on Table 4, at item (c), low reliability scores 
are defined as those with at least 11 respondents and reliability 
greater than or equal to 0.60 but less than 0.75 and also in the lowest 
12 percent of contracts ordered by reliability. The standard error is 
considered when the measure score is below the 15th percentile (in base 
group 1), significantly below average, and has low reliability: In this 
case, 1 star will be assigned if and only if the measure score is at 
least 1 standard error below the unrounded cut point between base 
groups 1 and 2. Similarly, when the measure score is at or above the 
80th percentile (in base group 5), significantly above average, and has 
low reliability, 5 stars would be assigned if and only if the measure 
score is at least 1 standard error above the unrounded cut point 
between base groups 4 and 5.

[[Page 16569]]



                  Table 5--CAHPS Star Assignment Rules
------------------------------------------------------------------------
          Star                  Criteria for assigning Star Ratings
------------------------------------------------------------------------
1.......................  A contract is assigned one star if both
                           criteria (a) and (b) are met plus at least
                           one of criteria (c) and (d):
                          (a) its average CAHPS measure score is lower
                           than the 15th percentile; AND
                          (b) its average CAHPS measure score is
                           statistically significantly lower than the
                           national average CAHPS measure score;
                          (c) the reliability is not low; OR
                          (d) its average CAHPS measure score is more
                           than one standard error (SE) below the 15th
                           percentile.
2.......................  A contract is assigned two stars if it does
                           not meet the one[dash]star criteria and meets
                           at least one of these three criteria:
                          (a) its average CAHPS measure score is lower
                           than the 30th percentile and the measure does
                           not have low reliability; OR
                          (b) its average CAHPS measure score is lower
                           than the 15th percentile and the measure has
                           low reliability; OR
                          (c) its average CAHPS measure score is
                           statistically significantly lower than the
                           national average CAHPS measure score and
                           below the 60th percentile.
3.......................  A contract is assigned three stars if it meets
                           at least one of these three criteria:
                          (a) its average CAHPS measure score is at or
                           above the 30th percentile and lower than the
                           60th percentile, AND it is not statistically
                           significantly different from the national
                           average CAHPS measure score; OR
                          (b) its average CAHPS measure score is at or
                           above the 15th percentile and lower than the
                           30th percentile, AND the reliability is low,
                           AND the score is not statistically
                           significantly lower than the national average
                           CAHPS measure score; OR
                          (c) its average CAHPS measure score is at or
                           above the 60th percentile and lower than the
                           80th percentile, AND the reliability is low,
                           AND the score is not statistically
                           significantly higher than the national
                           average CAHPS measure score.
4.......................  A contract is assigned four stars if it does
                           not meet the 5-star criteria and meets at
                           least one of these three criteria:
                          (a) its average CAHPS measure score is at or
                           above the 60th percentile and the measure
                           does not have low reliability; OR
                          (b) its average CAHPS measure score is at or
                           above the 80th percentile and the measure has
                           low reliability; OR
                          (c) its average CAHPS measure score is
                           statistically significantly higher than the
                           national average CAHPS measure score and
                           above the 30th percentile.
5.......................  A contract is assigned five stars if both
                           criteria (a) and (b) are met plus at least
                           one of criteria (c) and (d):
                          (a) its average CAHPS measure score is at or
                           above the 80th percentile; AND
                          (b) its average CAHPS measure score is
                           statistically significantly higher than the
                           national average CAHPS measure score;
                          (c) the reliability is not low; OR
                          (d) its average CAHPS measure score is more
                           than one SE above the 80th percentile.
------------------------------------------------------------------------

We requested comments on our proposed methods to determine cut points.
    In the proposed rule, we also acknowledged our past practice of 
publishing pre-determined 4-star thresholds for certain measures. We 
asked commenters who supported the return of the pre-determined 4-star 
thresholds to provide suggestions on how to minimize the risk of 
``misclassifying'' a contract's performance. For example, 
misclassification occurs when scoring a ``true'' 4-star contract as a 
3-star contract, or vice versa. The potential for misclassification is 
increased if the cut points result in the creation of ``cliffs'' 
between adjacent categories within the Star Ratings that could lead to 
the potential of different ratings between contracts with nearly 
identical Star Ratings that lie on the opposite sides of a fixed 
threshold. In addition, we ask commenters that supported pre-determined 
thresholds ways in which CMS can continue to create incentives for 
quality improvement. We also solicited comments on alternative 
recommendations for revising the cut point methodology. We summarized 
examples of alternatives we were considering: Methodologies that will 
minimize year-to-year changes in the cut points by setting the cut 
points so they are a moving average of the cut points from the 2 or 3 
most recent years; and setting caps on the degree to which a measure 
cut point could change from one year to the next. We solicited comments 
on these particular methodologies and recommendations for other ways to 
provide stability for cut points from year to year.
    We received the following comments on our proposals and our 
responses follow:
    Comment: There was widespread support for the use of the clustering 
algorithm to determine the cut points, although the overwhelming 
majority recommended some changes to how CMS determines the cut points.
    Response: CMS appreciates the support of the use of the clustering 
algorithm for the determination of the cut points. CMS carefully 
reviewed the feedback which reflects very diverse and conflicting 
opinions on the appropriate way to set cut points. CMS is actively 
considering a wide range of options for modifying the approach for 
determining cut points and needs to fully simulate alternative options 
in order to avoid implementing an option that could have unintended 
consequences. Thus, we are finalizing the clustering algorithm for the 
determination of cut points (for non-CAHPS measures) as proposed while 
we continue to simulate alternative options. CMS will use the feedback 
from this NPRM to guide and examine options for an enhanced methodology 
for converting the measure scores to measure-level Star Ratings, which 
would be proposed in a future regulation.
    Comment: The majority of commenters listed or identified several 
desirable attributes for the cut points, including having them be 
predetermined and released before the beginning of the measurement 
period, and increasing the stability and predictability of them. A 
handful of commenters noted that the cut points must represent 
meaningful differences among the star categories.
    Many commenters expressed concern about the influence of outliers 
on the cut points. Some of the suggestions for decreasing the influence 
of outliers included removing them from the clustering algorithm, using 
a trimmed data set, or raising the minimum measure-level denominator 
threshold from 30 to 100 to reduce the number of outliers based on 
small numbers. In addition, many commenters that expressed a preference 
for stability supported a cap, a restriction on the maximum movement 
for a measure's cut points from one year to the next, to achieve the 
desired characteristic. A commenter suggested employing a cap similar 
to NCQA's method which relies on assigning a cap based on the maximum 
change in the relative distribution of the measure scores. The 
commenter believed this would allow

[[Page 16570]]

CMS's clustering methodology to move cut points (for example, moving 
the 4 and 5 star cut points up) without extreme changes based on the 
movement of relatively few MA contracts. Another commenter who 
supported stability stated that the thresholds from one year to the 
next should not be allowed to decrease. The majority of commenters who 
supported caps did not provide a specific value or methodology, but 
rather the advantages that caps would allow.
    Some commenters suggested averaging cut points over multiple years 
for stability. Many commenters referenced CMS's previous policy that 
identified 4-star predetermined thresholds for specific measures and 
supported their return. A few commenters supported a weighted average 
based on several years of data to determine the cut points. A few 
commenters supported using a multiple-year trend to project measure cut 
points in advance of the measurement period.
    Response: We appreciate the careful consideration of possible 
modifications to the methodology used for determining the cut points 
for the conversion of measure scores to the measure-level Star Ratings 
scale. CMS is examining a number of potential options for determining 
cut points that would capture the greatest number of desirable 
attributes that our stakeholder have identified (pre-determined, 
stable, predictable cut points with minimal (if any) influence by 
outliers, restricted movement across years) while maintaining the 
integrity of the Star Ratings in order to propose a new or enhanced 
policy for establishing measure-level ratings in the near future. We 
believe that the number and scope of alternatives require additional 
consideration and testing before we can finalize a different 
methodology for setting cut points for non-CAHPS measures. In the 
meantime, we believe that the clustering methodology presents a valid 
approach to accurately reflect the quality of care for MA and Part D 
sponsors, while creating incentives for continued quality improvement. 
The goal of clustering is to assign stars that maximize the differences 
across star categories and minimize the differences within star 
categories to minimize the risk of misclassification. The clustering 
methodology also accounts for changes in the distribution of scores 
over time. We understand the desire to create more stability in the 
assignment of cut points and in performance expectations, but we want 
to ensure that any potential alternative methodologies do not create 
unintended consequences.
    Comment: Some commenters stated their support for transparency. 
Some commenters believe that increased transparency can be achieved by 
releasing all data for the Star Ratings program. A commenter suggested 
that CMS improve transparency in national performance reflected in 
display measures by calculating and publishing individual measure cut 
points for display measures instead of national averages. Other 
commenters believe transparency would be achieved by the implementation 
of pre-determined thresholds before the start of the measurement 
period.
    Response: CMS appreciates these comments. CMS agrees with the 
commenters that transparency in the ratings system is important. Each 
year CMS releases public use files of the performance data underlying 
the Star Ratings, available at http://go.cms.gov/partcanddstarratings. 
In addition, the cut points for the specific Star Ratings' year are 
available in the annual Technical Notes using the same link used for 
the data sets. A Cut Point Trend document is updated and released 
annually to provide a single source for multiple years of Star Ratings 
cut points. The Cut Point Trend document is organized in a user-
friendly format by measure and is available using the aforementioned 
link.
    Display measures are collected through data sources such as 
Medicare Part C and Part D reporting requirements, Part D Prescription 
Drug Event (PDE) data, Healthcare Effectiveness Data and Information 
Set (HEDIS) information, and Consumer Assessment of Healthcare 
Providers and Systems (CAHPS) surveys. The display measures are not 
included in determination of the Star Ratings on Medicare Plan Finder 
and thus, are not assigned Star Ratings. Display measures provide 
useful information about plan quality that sponsors can take action 
upon in order to improve the quality of care provided to their members. 
To allow comparisons, national averages of the display measure scores 
are available in the annual MA Part C & D Measure Technical Notes. (The 
display measure data set and Technical Notes are posted on the same 
site as the MA Star Ratings information.)
    CMS is examining a number of potential options for determining cut 
points that would capture the greatest number of desirable attributes 
that the commenters have identified (pre-determined, stable, 
predictable cut points with minimal (if any) influence by outliers, 
restricted movement across years) while maintaining the integrity of 
the Star Ratings. CMS is simulating the alternatives to the current cut 
point methodology. Further, CMS is identifying potential unintended 
consequences and examining ways to mitigate any identified risk to the 
integrity of the Star Ratings program. CMS is finalizing the clustering 
algorithm for the determination of cut points as proposed based on the 
positive and useful aspects of that methodology and to allow us the 
time to fully consider the options suggested by our stakeholders for 
enhancements to make it an even stronger methodology for converting the 
measure scores to measure-level Star Ratings. Any changes would be 
proposed in a future regulation.
    Comment: Some commenters suggested alternative methodologies to 
determine cut points. A commenter suggested the use of a forced 
distribution rather than clustering to capture the true distribution of 
plan performance; assigning cut points by applying an adjustment factor 
to the prior year's results based on historical performance; or 
calculating the average change in the median from the prior 3 years and 
apply that to determine the current cut points. A few commenters 
suggested using the industry average. A commenter suggested using the 
industry average as the basis of a 3-star rating.
    Response: CMS appreciates these comments. CMS believes that using a 
data driven approach to determine cut points aligns with our policies 
and guiding principles. As part of our guiding principles, we want to 
develop an enhanced methodology that ensures that the ratings are a 
true reflection of plan quality and minimizes the risk of 
misclassification. A forced distribution carries a high risk of 
misclassification because the cut points would not maximize the 
differences of contracts across star categories and minimize the 
differences of contracts within the same star category. An average as 
the basis of a 3-star rating would not accurately take into account the 
skewed distribution of many measures. CMS is examining a number of 
potential options for determining cut points while maintaining the 
integrity of the Star Ratings, including examining whether we can 
adjust prior performance results to determine current cut points. CMS 
will propose and solicit comment on an enhanced cut point methodology 
in a future regulation.
    Comment: Some commenters stated that the current clustering 
algorithm to identify the cut points for the Star Ratings' measures 
does not always accurately reflect the quality improvement that 
contacts have achieved especially for measures scores

[[Page 16571]]

with a limited range in their distribution. Some commenters explicitly 
stated their opposition to some of the proposed methodologies. A 
commenter was against a moving average approach amid concerns of the 
longevity of such a method. Another commenter did not support caps due 
to the belief that caps would mask true performance. Another commenter 
did not support weighted clustering. A commenter suggested benchmarking 
independent of clustering to determine the cut points; the commenter 
justified the recommendation based on the belief that increases in the 
average measure scores over time leads to decreased variability of plan 
performance and tight clustering of plan performance which results in 
insignificant percentile scores having large impacts on the Star 
Ratings.
    Response: CMS appreciates our stakeholder's feedback and will use 
it to guide the development of an enhanced methodology. So as not to 
implement a methodology that may inordinately increase the risk of 
misclassification, CMS will analyze and simulate the options to assess 
the impact of the methodology on the Star Ratings. The goal of 
clustering and the elimination of pre-determined 4-star thresholds for 
the 2016 Star Ratings was to more accurately measure performance.
    The current methodology for converting measure scores to measure-
level Star Ratings for non-CAHPS measures identifies the gaps that 
exist within the distribution of scores based on the criterion for 
assigning the groups. If the distribution is extremely restricted such 
that 5 unique groups cannot be formed, the output will result in 5 
groups that are not unique. In this rare situation, there would be less 
than 5 star categories and the ordered groups will be assigned the 
higher ratings on the scale.
    Comment: A commenter expressed concern about determining cut points 
using all MA data because such an approach fails to take into account 
the significant underlying differences in enrollment of plans. The 
commenter supported both stratified reporting and the determination of 
cut points after grouping plans into relevant cohorts (stratification 
at the contract level on key population characteristics, such as 
proportion Dual/LIS/Disabled).
    Response: CMS appreciates this feedback. The Star Ratings system 
does not determine cut points for subsets of the population because it 
does not align with its underlying principles. However, CMS has 
developed the Categorical Adjustment Index (CAI), which is a factor 
that is added to or subtracted from a contract's Overall and/or Summary 
Star Ratings to adjust for the average within-contract disparity in 
performance associated with a contract's percentages of beneficiaries 
with Low Income Subsidy/Dual Eligible (LIS/DE) and disability status. 
These adjustments are performed both with and without the improvement 
measures included. The value of the CAI varies by a contract's 
percentages of beneficiaries with Low Income Subsidy/Dual Eligible 
(LIS/DE) and disability status. In addition, CMS displays Part C and D 
performance data stratified by race and ethnicity at: https://www.cms.gov/About-CMS/Agency-Information/OMH/research-and-data/statistics-and-data/stratified-reporting.html.
    Comment: A commenter expressed support of the identification of cut 
points, as they can provide insight into performance throughout the 
year, leading to greater quality improvements.
    Response: CMS appreciates this support of the identification and 
utility of cut points.
    Comment: A commenter requested simulations on the proposed cut 
point methodologies.
    Response: CMS remains committed to transparency. CMS regularly 
solicits and values the feedback from our stakeholders. The feedback 
received guides the development of the policy options. CMS will 
continue to remain transparent in the development process for an 
enhanced cut point methodology as we move forward to propose a 
modified, new, or different policy for the assignment of measure-level 
Star Ratings.
    Comment: A commenter urged CMS to re-evaluate the cut points to 
ensure the Star Ratings accurately reflect plan quality and are based 
on evidence. The commenter expressed concern about the number of 
measures within the MA Star Ratings program that are based on physician 
action and compliance. In order for plans to comply with and earn 
incentives from CMS, the commenter believes that plans must often set 
unrealistic targets within their physician contracts in order for the 
plan to score well due to the Star Ratings cut points. The commenter 
believes that there may be instances when compliance with a measure is 
contrary to appropriate care, and contracts may be penalized.
    Response: CMS appreciates this comment. Plans should always set 
clinically appropriate targets for their physicians. There is no reason 
why the current methodology for setting the cut points to assign 
ratings to raw performance scores would require a physician to provide 
inappropriate care.
    Comment: CMS received a handful of comments related to converting 
CAHPS scores to stars. There was support for the current methodology 
(which was proposed) although several commenters suggested the cut 
points are too narrow and a few would like to re-implement pre-
determined cut points for CAHPS. A commenter stated that the relative 
distribution and significance testing methodology in CAHPS is biased in 
a negative direction and that these adjustments do not appropriately 
address the variability in CAHPS survey results.
    Response: We appreciate comments received on the CAHPS methodology. 
Three factors enter into CAHPS star assignment: The ranking of the 
contract in relation to other contracts, a statistical significance 
test that takes into consideration the degree of certainty that the 
score is above or below the national average, and examination of 
measure reliability. The significance test is applied in the same way 
in the positive and negative directions and is not biased. CAHPS 
measures meet high standards of reliability and thus variability in the 
scores reflects variability in performance. This methodology improves 
the performance of the star system and ensures that 4 and 5 stars are 
reserved for contracts with strong evidence of high performance and 
that 1 and 2 stars are reserved for contracts with strong evidence of 
low performance. We note that the base group is not an entitlement to a 
certain Star Rating.
    Previous analyses of CAHPS scores have suggested that seemingly 
small differences of 1 point on a 0-100 scale are meaningful; 
differences of 3 points can be considered medium, and differences of 5 
points can be considered large.\57\ For instance, a 3-point increase in 
some CAHPS measures has been associated with a 30 percent reduction in 
disenrollment from health plans, which suggests that even ``medium'' 
differences in CAHPS scores may indicate substantially different care 
experiences.\58\
---------------------------------------------------------------------------

    \57\ Paddison CAM, Elliott MN, Haviland AM, Farley DO, 
Lyratzopoulos G, Hambarsoomian K, Dembosky JW, Roland MO. (2013). 
``Experiences of Care among Medicare Beneficiaries with ESRD: 
Medicare Consumer Assessment of Healthcare Providers and Systems 
(CAHPS) Survey Results.'' American Journal of Kidney Diseases 61(3): 
440-449.
    \58\ Lied, T.R., S.H. Sheingold, B.E. Landon, J.A. Shaul, and 
P.D. Cleary. (2003). ``Beneficiary Reported Experience and Voluntary 
Disenrollment in Medicare Managed Care.'' Health Care Financing 
Review 25(1): 55-66.

---------------------------------------------------------------------------

[[Page 16572]]

Summary of Regulatory Changes
    For the reasons set forth in the proposed rule and our responses to 
the related comments summarized above, we are finalizing the 
methodology to determine cut points as proposed in Sec. Sec.  
422.166(a) and 423.186(a). CMS is committed to incorporating the 
feedback received from commenters on the methodology for determining 
cut points for non-CAHPS measures and will thoroughly analyze other 
potential methodologies to ensure that unintended consequences are 
avoided and the cut points resulting from any enhancements are 
consistent with principles and policy goals for the Star Ratings 
system. Changes to the methodology for the determination of cut points 
for non-CAHPS measures will be proposed in a future rule.
    We are finalizing the methodology to determine cut points for CAHPS 
measures in Sec. Sec.  422.166(a)(3) and 423.186(a)(3) substantively as 
proposed. We are finalizing the regulation text with minor technical 
revisions to improve readability.
m. Hierarchical Structure of the Ratings
    We proposed to continue our existing policy to use a hierarchical 
structure for the Star Ratings. Currently, and as proposed, the basic 
building block of the MA Star Ratings system is the measure. Because 
the MA Star Ratings system consists of a large collection of measures 
across numerous quality dimensions, the measures will be organized in a 
hierarchical structure that provides ratings at the measure, domain, 
Part C summary, Part D summary, and overall levels. The proposed 
regulations text at Sec. Sec.  422.166 and 423.186 are built on this 
structure and provides for calculating ratings at each ``level'' of the 
system. The organization of the measures into larger groups increases 
both the utility and efficiency of the rating system. At each 
aggregated level, ratings are based on the measure-level stars. Ratings 
at the higher level are based on the measure-level Star Ratings, with 
whole star increments for domains and half-star increments for summary 
and overall ratings; a rating of 5 stars will indicate the highest Star 
Rating possible, while a rating of 1 star will be the lowest rating on 
the scale. Half-star increments are used in the summary and overall 
ratings to allow for more variation at the higher hierarchical levels 
of the ratings system. We believe this greater variation and the 
broader range of ratings provide more useful information to 
beneficiaries in making enrollment decisions while remaining consistent 
with the statutory direction in sections 1853(o) and 1854(b) of the Act 
to use a 5-star system. These policies for the assignment of stars will 
be codified with other rules for the ratings at the domain, summary, 
and overall level. Domain ratings employ an unweighted mean of the 
measure-level stars, while the Part C and D summary and overall ratings 
employ a weighted mean of the measure-level stars and up to two 
adjustments. We proposed to codify these policies at paragraphs (b)(2), 
(c)(1) and (d)(1) of Sec. Sec.  422.166 and 423.186.
    We received the following overall comments on our proposal and our 
response follows:
    Comment: All commenters supported the existing hierarchical 
structure of the Star Ratings program and its associated policies.
    Response: CMS appreciates the continued support of the existing 
organization of the Star Ratings measures and the policies associated 
with it. CMS firmly believes the structure increases the utility and 
efficiency of the rating system and appreciates the positive response 
to it.
n. Domain Star Ratings
    Groups of measures that together represent a unique and important 
aspect of quality and performance are organized to form a domain. 
Domain ratings summarize a plan's performance on a specific dimension 
of care. Currently the domains are used purely for purposes of 
displaying data on Medicare Plan Finder to organize the measures and 
help consumers interpret the data. We proposed to continue this policy 
at Sec. Sec.  422.166(b)(1)(i) and 423.186(b)(1)(i).
    At present, there are nine domains--five for Part C measures for 
MA-only and MA-PD plans and four for Part D measures for stand-alone 
PDP and MA-PD plans. We proposed to continue to group measures for 
purposes of display on Medicare Plan Finder and to continue use of the 
same domains as in current practice in Sec. Sec.  422.166(b)(1)(i) and 
423.196(b)(1)(i). The current domains are listed in Tables 5 and 6.

                         Table 6--Part C Domains
------------------------------------------------------------------------
                                 Domain
-------------------------------------------------------------------------
Staying Healthy: Screenings, Tests and Vaccines.
Managing Chronic (Long Term) Conditions.
Member Experience with Health Plan.
Member Complaints and Changes in the Health Plan's Performance.
Health Plan Customer Service.
------------------------------------------------------------------------


                         Table 7--Part D Domains
------------------------------------------------------------------------
                                 Domain
-------------------------------------------------------------------------
Drug Plan Customer Service.
Member Complaints and Changes in the Drug Plan's Performance.
Member Experience with the Drug Plan.
Drug Safety and Accuracy of Drug Pricing.
------------------------------------------------------------------------

    Currently, Star Ratings for domains are calculated using the 
unweighted mean of the Star Ratings of the included measures. They are 
displayed to the nearest whole star, using a 1-5 star scale. We 
proposed to continue this policy at paragraph (b)(2)(ii). We also 
proposed that a contract must have stars for at least 50 percent of the 
measures required to be reported for that domain for that contract type 
to have that domain rating calculated; we explained this was necessary 
to have enough data to reflect the contract's performance on the 
specific dimension. For example, if a contract is rated only on one 
measure in Staying Healthy: Screenings, Tests and Vaccines, that one 
measure will not necessarily be representative of how the contract 
performs across the whole domain so we do not believe it is appropriate 
to calculate and display a domain rating. We proposed to continue this 
policy by providing, at paragraph (b)(2)(i), that a minimum number of 
measures must be reported for a domain rating to be calculated.
    We received the following comments on our proposal and our 
responses follow:
    Comment: Commenters supported the use of the current domains and 
the associated policies related to the calculation of the Star Ratings 
for domains.
    Response: CMS appreciates our stakeholders' support of the use of 
the domains and associated policies related to the domains.
    Comment: A commenter noted the usefulness of the domains for 
displaying the data on Medicare Plan Finder (MPF). In addition, the 
commenter believed the domains helped consumers interpret the data on 
MPF.
    Response: The domains were designed to summarize a plan's 
performance on a specific dimension of care. CMS appreciates the 
positive feedback related to domains and the agreement that they serve 
not only to organize data on MPF, but also serve as an aid to 
consumers' interpretation of the data displayed.
Summary of Regulatory Changes
    For the reasons set forth in the proposed rule and our responses to 
the related comments summarized above,

[[Page 16573]]

we are finalizing the provisions identifying the domains and for rating 
at the domain level as proposed at Sec. Sec.  422.166(b) and 423.186(b) 
without modification.
o. Part C and D Summary Ratings
    In the current rating system the Part C summary rating provides a 
rating of the health plan quality and the Part D summary rating 
provides a rating of the prescription drug plan quality. We proposed, 
at Sec. Sec.  422.166(c) and 423.186(c), to codify regulation text 
governing the adoption of Part C summary ratings and Part D summary 
ratings. An MA-only plan and a Part D stand-alone plan will receive a 
summary rating only for, respectively, Part C measures and Part D 
measures.
    First, in paragraphs (c)(1) of each section, we proposed the 
overall formula for calculating the summary ratings for Part C and Part 
D. Under current policy, the summary rating for an MA-only contract is 
calculated using a weighted mean of the Part C measure-level Star 
Ratings with up to two adjustments: The reward factor (if applicable) 
and the Categorical Adjustment Index (CAI). Similarly, the current 
summary rating for a PDP contract is calculated using a weighted mean 
of the Part D measure-level Star Ratings with up to two adjustments: 
The reward factor (if applicable) and the CAI. We proposed in 
Sec. Sec.  422.166(c)(1) and 423.186(c)(1) that the Part C and Part D 
summary ratings would be calculated as the weighted mean of the 
measure-level Star Ratings with an adjustment to reward consistently 
high performance (reward factor) and the application of the CAI, 
pursuant to paragraph (f) (where we proposed the specifics for these 
adjustments) for Parts C and D, respectively.
    Second, and also consistent with current policy, we proposed an MA-
only contract and PDP would have a summary rating calculated only if 
the contract meets the minimum number of rated measures required for 
its respective summary rating: A contract must have scores for at least 
50 percent of the measures required to be reported for the contract 
type to have the summary rating calculated. We proposed to codify the 
necessary text as paragraph (c)(2)(i) of Sec. Sec.  422.166 and 423.186 
the same rules will be applied to both the Part C and Part D summary 
ratings for the minimum number of rated measures. We proposed that 
these regulations would also apply to calculating the summary Part C 
and Part D ratings of MA-PD plan; the MA-PD plan would have to meet the 
minimum number of rated measures for each summary rating type. We also 
proposed (at paragraph (c)(2)(ii)) that the improvement measures 
themselves are not included in the count of minimum number of measures 
for the Part C or Part D summary ratings. Third, we proposed a 
paragraph (c)(3) in both Sec. Sec.  422.166 and 423.186 to provide that 
the summary ratings are on a 1 to 5 star scale in half-star increments. 
Traditional rounding rules would be employed to round the summary 
rating to the nearest half-star. We explained in connection with this 
proposal how the policies proposed in Sec. Sec.  422.166(h) and 
423.186(h) regarding posting summary ratings on MPF would apply. The 
summary rating would be displayed in HPMS and Medicare Plan Finder to 
the nearest half star if a contract had not met the measure requirement 
for calculating a summary rating, the display in HPMS (and on Medicare 
Plan Finder) for the applicable summary rating would be the flag, ``Not 
enough data available'' or if the measurement period is less than 1 
year past the contract's effective date the flag would be, ``Plan too 
new to be measured.''
    We solicited comments on the calculations for the Part C and D 
summary ratings. We received the following comments on our proposal and 
our responses follow:
    Comment: The majority of the commenters supported the policies, 
methodology, and display of the summary ratings as proposed.
    Response: CMS appreciates the ongoing support of the summary 
ratings.
    Commenter: A commenter recommended a revision to the rule that 
requires a contract to have numeric scores for at least 50 percent of 
the required measures for the summary-specific rating to have a summary 
rating calculated. The commenter suggested a change to the rule such 
that a summary rating would be calculated if a contract had at least 
half of the weighted value of the full measure set for the summary-
specific rating.
    Response: CMS appreciates the feedback for a possible revision to 
the rule that determines whether a summary rating would be calculated. 
The Part C summary rating provides a rating of the health plan quality 
and the Part D summary rating provides a rating of the prescription 
drug plan quality. The summary ratings include information from 
multiple dimensions of quality and performance. CMS plans to evaluate 
the suggestion of using 50 percent of the total weighted value of the 
measure set as the threshold for calculating summary ratings to examine 
whether such a change would still allow an accurate reflection of the 
quality of the health plan or prescription drug plan.
Summary of Regulatory Changes
    For the reasons set forth in the proposed rule and our responses to 
the related comments summarized above, we are finalizing the provisions 
governing summary ratings as proposed at Sec. Sec.  422.162(c) and 
423.182(c) without modification.
p. Overall Rating
    The overall Star Rating is a global rating that summarizes the 
plan's quality and performance for the types of services offered by the 
plans under the rated contract. We proposed at Sec. Sec.  422.166(d) 
and 423.186(d) to codify the standards for calculating and assigning 
overall Star Ratings for MA-PD contracts. The overall rating for an MA-
PD contract is proposed to be calculated using a weighted mean of the 
Part C and Part D measure level Star Ratings, respectively, with an 
adjustment to reward consistently high performance described in 
paragraph (f)(1) and the application of the CAI, pursuant to described 
in paragraph (f)(2).
    Consistent with current policy, we proposed at paragraph (d)(2) 
that an MA-PD would have an overall rating calculated only if the 
contract receives both a Part C and Part D summary rating and has 
scores for at least 50 percent of the required measures for the 
contract type. As with the Part C and D summary ratings, the Part C and 
D improvement measures will not be included in the count for the 
minimum number of measures for the overall rating. Any measure that 
shares the same data and is included in both the Part C and Part D 
summary ratings would be included only once in the calculation for the 
overall rating. For example, the measures ``Members Choosing to Leave 
the Plan'' and ``Complaints about the Plan'' use the same data for both 
the Part C and Part D measure for an MA-PD plan and under the proposal, 
would be counted only once for the overall rating. As with summary 
ratings, we proposed that overall MA-PD ratings would use a 1 to 5 star 
scale in half-star increments; traditional rounding rules would be 
employed to round the overall rating to the nearest half-star. These 
policies are proposed as paragraphs (d)(2)(i) through (iv).
    We also explained in the proposed rule how the overall rating would 
be posted in accordance with our general proposed policy at Sec. Sec.  
422.166(h) and 423.186(h), including the specific messages for lack of 
ratings for certain reasons. Applying that rule, if an MA-

[[Page 16574]]

PD contract has only one of the two required summary ratings, the 
overall rating would not be calculated and the display in HPMS would be 
the flag, ``Not enough data available.''
    For QBP purposes, low enrollment contracts and new MA plans are 
defined in Sec.  422.252. Low enrollment contract means a contract that 
could not undertake Healthcare Effectiveness Data and Information Set 
(HEDIS) and Health Outcomes Survey (HOS) data collections because of a 
lack of a sufficient number of enrollees to reliably measure the 
performance of the health plan; new MA plan means an MA contract 
offered by a parent organization that has not had another MA contract 
in the previous 3 years. Low enrollment contracts and new plans do not 
receive an overall or summary rating because of the lack of necessary 
data. However, they are treated as qualifying plans for the purposes of 
QBPs. Section 1853(o)(3)(A)(ii)(II) of the Act, as implemented at Sec.  
422.258(d)(7), provides that for 2013 and subsequent years, CMS shall 
develop a method for determining whether an MA plan with low enrollment 
is a qualifying plan for purposes of receiving an increase in payment 
under section 1853(o). This determination is applied at the contract 
level and thus determines whether a contract (meaning all plans under 
that contract) is a qualifying contract. The statute, at section 
1853(o)(3)(A)(iii) of the Act, provides for treatment of new MA plans 
as qualifying plans eligible for a specific QBP. We therefore proposed, 
at Sec. Sec.  422.166(d)(3) and 423.186(d)(3), that low enrollment 
contracts (as defined in Sec.  422.252 of this chapter) and new MA 
plans (as defined in Sec.  422.252 of this chapter) do not receive an 
overall and/or summary rating; they will be treated as qualifying plans 
for the purposes of QBPs as described in Sec.  422.258(d)(7) of this 
chapter. The QBP levels for each rating area are announced through the 
process described for changes in and adoption of payment and risk 
adjustment policies in section 1853(b) of the Act. We noted that this 
aspect of the proposal would merely codify existing policy and 
practice.
    We received the following comments on our proposal and our 
responses follow:
    Comment: Commenters supported the use of the overall rating as a 
global rating that summarizes a contract's quality and performance, as 
well as the proposal to use the current policies for calculating and 
publishing the overall rating.
    Response: CMS values the support of the overall rating and its 
associated methodology.
    Comment: A commenter suggested a revision to the rule for 
calculating the overall rating for an MA-PD contract. As done currently 
and proposed, an MA-PD contract must have both (Part C and Part D) 
summary ratings and measure scores for at least 50 percent of the 
required measures based on contract-type (exclusive of the improvement 
measures) to have an overall rating. The commenter suggested that an 
overall rating for an MA-PD contract require measure scores that total 
at least half of the weighted value of the full measure set.
    Response: CMS appreciates the suggestion of alternative 
requirements for the calculation of an overall rating. Changing the 
requirement for the calculation of an overall rating to be based on the 
majority of the total weight of the Star Ratings measures has the 
potential of confusing the global nature of the overall rating. There 
are substantially more Part C measures in the Star Ratings and the 
total weight of the Part C measures exceeds that of the Part D 
measures. By requiring a contract to have both a Part C and D summary 
rating coupled with the requirement of at least 50 percent of the 
measures, CMS has minimized the potential for the overall rating being 
determined primarily by dimensions of health plan quality instead of 
both health plan and prescription drug plan quality.
    Comment: A commenter suggested the use of a percentile rank 
threshold for the determination of a 5-star overall rating, thus 
allowing the recognition of top performers along with the ability to 
enroll members year-round.
    Response: While CMS thanks the commenter for its suggestion, CMS 
disagrees with using percentile ranking as a threshold for calculating 
overall ratings. One of the underlying design principles of the MA and 
Part D Star Ratings is to incentivize plans to provide the best health 
care possible to our beneficiaries. This underlying principle is 
reflected in the manner that measure scores are converted to Star 
Ratings, as well as the aggregation of the measure-level Star Ratings 
to an overall rating. (Measure-level Star Ratings are the basic 
building block, of the overall rating.) A percentile rank threshold 
approach for the overall rating does not align with the principles of 
the Star Ratings methodology and would arbitrarily apply a threshold 
that could be perceived as a subjective value that would ultimately 
separate 5-star contracts from all other contracts.
Summary of Regulatory Changes
    For the reasons set forth in the proposed rule and our responses to 
the related comments summarized above, we are finalizing the provisions 
for overall ratings as proposed at Sec. Sec.  422.162(d) and 423.182(d) 
without modification.
q. Measure Weights
    Prior to the 2012 Part C and D Plan Ratings (now known as Star 
Ratings), all individual measures included in the program were weighted 
equally, suggesting equal importance. Based on feedback from 
stakeholders, including health and drug plans and beneficiary advocacy 
groups, we moved to provide greater weight to clinical outcomes and 
lesser weight to process measures. Patient experience and access 
measures were also given greater weight than process measures, but not 
as high as outcome measures. The differential weighting was implemented 
to help create further incentives to drive improvement in clinical 
outcomes, patient experience, and access. These differential weights 
for measures were implemented for the 2012 Ratings following a May 2011 
Request for Comments and adopted in the CY2013 Rate Announcement and 
Final Call Letter.
    In the Contract Year 2012 Final Rule for Changes to the Medicare 
Advantage and the Medicare Prescription Drug Benefit Programs rule (79 
FR 21486), we stated that scoring methodologies should also consider 
improvement as an independent goal. To this end, we implemented in the 
CY 2013 Rate Announcement the Part C and D improvement measures that 
measure the overall improvement or decline in individual measure scores 
from the prior to the current year. Given the importance of recognizing 
quality improvement as an independent goal, for the 2015 Star Ratings, 
we proposed and subsequently finalized through the 2015 Rate 
Announcement and final Call Letter an increase in the weight of the 
improvement measure from 3 times to 5 times that of a process measure, 
which is weighted as 1. This weight aligns the Part C and D Star 
Ratings program with value-based purchasing programs in Medicare fee-
for-service which take into account improvement.
    We proposed in Sec. Sec.  422.166(e) and 423.186(e) to continue the 
current weighting of measures in the Part C and D Star Ratings program 
by assigning the highest weight (5) to improvement measures, followed 
by outcome and intermediate outcome measures (weight of 3), then by 
patient experience/

[[Page 16575]]

complaints and access measures (weight of 1.5), and finally process 
measures (weight of 1). We also solicited feedback about increasing the 
weight of the patient experience/complaints and access measures and 
stated our interest in stakeholder feedback on this potential change in 
order to reflect better the importance of these issues in plan 
performance. If we were to increase the weight, we asked for feedback 
about increasing it from a weight of 1.5 to between 1.5 and 3, similar 
to outcome measures. This increased weight would reflect CMS's 
commitment to serve Medicare beneficiaries by putting patients first, 
including their assessments of the care received by plans. We solicited 
comment on this point, particularly the potential change in the weight 
of the patient experience/complaints and access measures.
    Table C7 includes the proposed measure categories, the definitions 
of the measure categories, and the weights. In calculating the summary 
and overall ratings, a measure given a weight of 3 counts three times 
as much as a measure given a weight of 1. In section II.A.11. of the 
proposed rule, we proposed (as Table C2) the measure set and included 
the category and weight for each measure, consistent with this proposal 
for weighting measure by category. We proposed that as new measures are 
added to the Part C and D Star Ratings, we would assign the measure 
category based on these categories and the regulation text proposed at 
Sec. Sec.  422.166(e) and 423.186(e), subject to two exceptions. For 
the first exception, we proposed to codify current policy in paragraphs 
(e)(2) of each section and to assign new measures to the Star Ratings 
program a weight of 1 for their first year in the Star Ratings. In 
subsequent years the weight associated with the measure weighting 
category would be used. This is consistent with current policy.

                          Table 8--Proposed Measure Categories, Definitions and Weights
----------------------------------------------------------------------------------------------------------------
              Measure category                                    Definition                          Weight
----------------------------------------------------------------------------------------------------------------
Improvement.................................  Part C and Part D improvement measures are derived               5
                                               through comparisons of a contract's current and
                                               prior year measure scores.
Outcome and Intermediate Outcome............  Outcome measures reflect improvements in a                       3
                                               beneficiary's health and are central to assessing
                                               quality of care. Intermediate outcome measures
                                               reflect actions taken which can assist in
                                               improving a beneficiary's health status.
                                               Controlling Blood Pressure is an example of an
                                               intermediate outcome measure where the related
                                               outcome of interest will be better health status
                                               for beneficiaries with hypertension.
Patient Experience/Complaints...............  Patient experience measures reflect beneficiaries'             1.5
                                               perspectives of the care and services they
                                               received.
Access......................................  Access measures reflect processes and issues that              1.5
                                               could create barriers to receiving needed care.
                                               Plan Makes Timely Decisions about Appeals is an
                                               example of an access measure.
Process.....................................  Process measures capture the health care services                1
                                               provided to beneficiaries which can assist in
                                               maintaining, monitoring, or improving their
                                               health status.
----------------------------------------------------------------------------------------------------------------

    For the second exception, we proposed (at Sec. Sec.  422.166(e)(3) 
and 423.186(e)(3)) again to codify current policy and to apply a 
special rule for MA-PD and Part D contracts that have service areas 
that are wholly located in Puerto Rico. We recognize the additional 
challenge unique to Puerto Rico related to the medication adherence 
measures used in the Star Ratings program due to the lack of Low Income 
Subsidy (LIS). For the 2017 Star Ratings, we implemented a different 
weighting scheme for the Part D medication adherence measures in the 
calculation of the overall and summary Star Ratings for contracts that 
solely serve a population of beneficiaries in Puerto Rico. We proposed, 
at Sec. Sec.  422.166(e)(3) and 423.186(e)(3), to continue to reduce 
the weights for the adherence measures to 0 for the summary and overall 
rating calculations and maintain the weight of 3 for the adherence 
measures for the improvement measure calculations for contracts with 
service areas that are wholly located in Puerto Rico. We requested 
comment on our proposed weighting strategy for Measure Weights 
generally and for Puerto Rico, including the weighting values 
themselves.
    We received the following comments on our proposal and our 
responses follow:
    Comment: Multiple commenters requested CMS not to increase the 
weight of patient experience/complaints and access measures from a 
weight of 1.5 up to 3. Many of the commenters requested to maintain the 
current weight; however, others requested that CMS decrease the weight 
of patient experience measures citing survey reliability and sampling 
concerns with patient experience surveys. They stated that patient-
reported data are not as reliable as claims, prescription drug event 
data, medical charts, and other data sources. They believe that these 
measures are unfairly subjective and that more weight should be placed 
on more reliable and objective measures like clinical and outcome 
measures. Many cited concerns with response rates, sample size of 
patient experience surveys, and other factors in which the plan has 
less control, as well as industry concerns around accuracy of survey 
responses and research suggesting a weak relationship between care 
received and survey responses. A commenter supported increasing the 
weight of access and patient experience measures that are not based on 
survey data. A commenter opposed the weight increase until we have 
better measures in these areas.
    Response: We refer commenters to section II.A.11.i and Table C2A 
and the narrative comment and responses that follow, which give 
background and additional justification for CAHPS measures. While we 
acknowledge that the CAHPS survey captures individuals' perspectives on 
their experiences of care, it is anchored in measureable aspects of 
care and so can be measured reliably. In addition, CAHPS surveys were 
developed with broad stakeholder input, including a public solicitation 
of measures and a Technical Expert Panel, and the opportunity for 
anyone to comment on the surveys through multiple public comment 
periods through the Federal Register. CMS encourages all plans to 
familiarize themselves with the survey methodology and to review the 
background materials available on the MA and PDP CAHPS website that 
validate the CAHPS measures.
    CMS has pledged to put patients first and to empower patients to 
work with their doctors to make health care decisions that are best for 
them. An increased weight for patient experience/complaints and access 
measures reflects

[[Page 16576]]

CMS's commitment to serve Medicare beneficiaries by including their 
assessments of the care received by plans. In addition, CAHPS measures 
and positive clinical outcomes have been shown to be related. While 
patient experience is an inherently important dimension of healthcare 
quality, it is also the case that the preponderance of evidence shows 
that better patient experience is associated with better patient 
adherence to recommended treatment, better clinical processes, better 
hospital patient safety culture, better clinical outcomes, reduced 
unnecessary healthcare use, and fewer inpatient 
complications.59 60 A recent study found that higher quality 
for patient experience had a statistically significant association with 
lower rates of many in-hospital complications and unplanned 
readmissions to the hospital within 30 days. In other words, better 
patient experience according to the CMS hospital Star Ratings is 
associated with favorable clinical outcomes.\61\ An increased weight 
also reflects the importance of these beneficiary-centered issues in 
plan performance.
---------------------------------------------------------------------------

    \59\ Price, R.A., Elliott, M.N., Zaslavsky, A.M., Hays, R.D., 
Lehrman, W.G., Rybowski, L., & Cleary, P.D. (2014). Examining the 
role of patient experience surveys in measuring health care quality. 
Medical Care Research and Review, 71(5), 522-554.
    \60\ Price, R.A., Elliott, M.N., Cleary, P.D., Zaslavsky, A.M., 
& Hays, R.D. (2015). Should health care providers be accountable for 
patients' care experiences?. Journal of general internal medicine, 
30(2), 253-256.
    \61\ Trzeciak, Stephen et al. ``Association Between Medicare 
Star Ratings for Patient Experience and Medicare Spending per 
Beneficiary for US Hospitals.'' Journal of patient experience 4.1 
(2017): 17-21. PMC. Web. 2 Feb. 2018.
---------------------------------------------------------------------------

    Further, access to health services is a critical issue in the 
healthcare sector and to Medicare beneficiaries. Lack of access can 
result in unmet health needs, delays in receiving the appropriate care, 
inability to access preventative services, unreasonable financial 
burdens, and preventable hospitalizations.\62\ For these reasons, 
access measures, such as appeals measures and call center measures, are 
crucial in the Star Ratings system. Increasing the weight for these 
measures highlights the importance of capturing access to care within 
MA and Part D plans.
---------------------------------------------------------------------------

    \62\ https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/MMS/Downloads/Access-Measures.pdf.
---------------------------------------------------------------------------

    To best meet the needs of our beneficiaries, CMS believes that we 
must listen to their perceptions of care, as well as ensure they have 
access to needed care. Commenters representing beneficiaries strongly 
supported an increase in the weight of the patient experience of care 
and access measures. Therefore, we will finalize an increase in the 
weight for these two categories of measures from 1.5 to 2. Given the 
importance of hearing the voice of patients when evaluating the quality 
of care provided, CMS intends to further increase the weight of these 
measures in the future, so we welcome stakeholder feedback on how to 
improve the CAHPS survey, including the topics it covers, and 
suggestions for additional access measures or modifications to existing 
ones. We expect this change to increase the highest rating for 
approximately 8 percent of contracts and to have no impact on the 
majority of other contracts, while also demonstrating CMS's commitment 
to evaluate the quality of care provided as experienced by 
beneficiaries. Please send feedback about CAHPS to MP-CAHPS@cms.hhs.gov 
and feedback about access measures to PartCandDStarRatings@cms.hhs.gov.
    Comment: A handful of commenters strongly supported the proposed 
weight increase of patient experience/complaints and access measures. 
They emphasized the importance of the beneficiary and caregiver 
perspectives and noted that the beneficiary's voice is an important 
indicator for plan performance in key areas such as the ease of access 
to needed drugs and treatments as well as plan responsiveness to appeal 
requests. Commenters said that by increasing the weights of these 
measures, CMS ensures that beneficiaries are seeing Star Ratings that 
reflect what they are likely to find important about their plan 
selections. These commenters also believed that assessments of quality 
and value by the patient are currently under-valued in Part C and D. 
Therefore, they believed patient experience/complaints and access 
measures should receive a higher weight than the current 1.5.
    Response: CMS appreciates this feedback and agrees the voice of the 
beneficiary must be heard as part of evaluating the quality of health 
and drug plans.
    Comment: CMS received several comments requesting to decrease and 
reclassify HOS measures on Improving or Maintaining Physical and Mental 
Health to receive a patient experience weight of 1.5 or process measure 
weight of 1, as opposed to their current outcome weight of 3. Some 
commenters believed there are methodological limitations to the HOS, 
and they stated that it does not provide a reliable evaluation of the 
patient experience because it relies on variables such as memory and 
the patient's physical and mental status at the time of survey 
completion. We also received comments that because the HOS measures are 
patient-reported measures (in response to survey questions) they are 
not true measures of health outcomes and should be weighted no higher 
than 1 or 1.5.
    Response: We refer the commenter to Table C2 in section II.A.11., 
which gives background and additional justification for HOS measures. 
The HOS assesses health outcomes for randomly selected beneficiaries 
from each health plan over a two-year period by using baseline 
measurement and a two-year follow up. CMS recognizes that the Physical 
Component Score (PCS) and the Mental Component (MCS) may decline over 
time and that health maintenance, rather than improvement, is a more 
realistic clinical goal for many older adults. MAOs are asked to 
improve or maintain the physical and mental health of their members. 
Change scores are constructed and the results compare actual to 
expected changes in physical and mental health. Therefore, the 
Improving or Maintaining Physical and Mental Health measures are not 
patient experience measures because they measure whether plan member's 
physical and mental health is the same or better than expected after 2 
years. While the data come from the HOS, they measure beneficiary 
outcomes and therefore are appropriately classified as outcome measures 
with a weight of 3.
    Additionally, the HOS was developed and continues to be refined 
under the guidance of a Technical Expert Panel comprised of individuals 
with specific expertise in the health care industry and outcomes 
measurement. HOS analysts apply the most recent advances in summarizing 
physical and mental health outcomes results and appropriate risk 
adjustment techniques. CMS also solicits stakeholder input, including 
public solicitation of measures and the opportunity for anyone to 
comment on the survey through multiple public comment periods through 
the Federal Register.
    Comment: A few commenters sought clarification on differences in 
the weights between the Part C and Part D Statin measures. Two 
organizations recommended classifying both the Part C Statin Therapy 
for Patients with Cardiovascular Disease (SPC) and the Part D Statin 
Use in Persons with Diabetes (SUPD) measures as process measures with a 
weight of 1. A commenter supported the weight for the Statin measure 
developed by PQA.
    Response: CMS appreciates the feedback and clarifies the weighting 
decision for each measure below. The Part C Statin Therapy for Patients 
with Cardiovascular Disease (SPC) measure is

[[Page 16577]]

the percent of plan members (males 21-75 years of age and females 40-75 
years of age) who were identified as having clinical atherosclerotic 
cardiovascular disease (ASCVD) and were dispensed at least one high or 
moderate-intensity statin medication. The Part C measure focuses on 
patients who were dispensed one prescription and whether the patient 
filled the medication at least once. Therefore, it is a process measure 
and will receive a weight of 1. The Part D measure is the percent of 
the number of plan members 40-75 years old who were dispensed at least 
two diabetes medication fills and received a statin medication fill. 
Receiving multiple fills indicates the patient continues to take the 
medication and therefore suggests adherence. The Part D measure is not 
a process measure. Continuing to take the prescribed medication is 
necessary to reach clinical/therapeutic goals. Thus, the Part D measure 
is an intermediate outcome measure and will receive a weight of 3.
    Comment: A couple of commenters requested a decrease in the 
improvement measures from the current weight of 5 to a weight of 3 
(like outcome measures). They stated the measures diminish the 
importance of clinical measures and mislead Medicare beneficiaries 
about which are the highest quality health plans.
    Response: CMS recognizes the importance of acknowledging quality 
improvement in health and drug plans. The decision to assign a weight 
of 5 for the improvement measures was originally made to align the Part 
C and D Star Ratings program with value-based purchasing programs in 
Medicare fee-for-service which heavily weight improvement. As part of 
the Part C and D Star Ratings program, we are committed to improving 
the quality of care and experiences for Medicare beneficiaries. Through 
assigning a weight of 5 to improvement, CMS encourages MA and Part D 
contracts to focus on improving the quality of care provided.
    With regard to overall ratings, improvement measures contribute 
significantly less than outcome measures overall. For example for the 
2018 Star Ratings for an MA-PD that does not include a SNP, the overall 
contribution of the improvement measures to the overall rating is close 
to 14 percent, but the overall contribution of outcome and intermediate 
outcome measures is 33 percent.
    CMS believes that continuous improvement is necessary to reach the 
goal of providing the best care to our beneficiaries. While the 
improvement measures are weighted the most of any category in the Star 
Ratings, the improvement measure is a single measure that encompasses 
care across multiple dimensions.
    Comment: A commenter recommended that CMS weight MA-PD and PDP 
measures differently based on the plan's ability to influence outcomes 
on a measure, for example statin use in persons with diabetes. PDPs 
should have less weight placed on measures that largely depend on 
provider behavior, which they have very little ability to impact.
    Response: Currently the only Part D outcome measures are adherence 
measures. CMS disagrees that stand-alone PDPs have very little 
influence on beneficiaries' medication adherence. There are many 
strategies that can be used to improve a beneficiary's medication 
adherence in addition to prescriber interventions, such as refill 
reminders, formulary and benefits design, and medication therapy 
management programs. Plan sponsors can also leverage network pharmacy 
relationships to address medication adherence issues, facilitate 
medication synchronization, or provide education and counseling. In the 
absence of a contact phone number for the beneficiary, it may be 
beneficial to use these interventions to reach the beneficiary at the 
place of dispensing. Furthermore, MA-PDs and PDPs are rated separately 
to account for delivery system differences. Lastly, adherence measures 
will now be included in the CAI to account for LIS beneficiaries which 
we discuss in more detail in section II.A.11.t.
    Comment: A commenter recommended decreasing the weighting of a 
topped out measure rather than discontinuing the measure.
    Response: Measure scores are determined to be `topped out' when 
they show high performance and little variability across contracts, 
making the measure unreliable. CMS removes measures that show low 
statistical reliability so as to move swiftly to ensure the validity 
and reliability of the Star Ratings, even at the measure level. 
However, CMS will retain measures at the same weight if for example, 
performance in a given measure has just improved across all contracts, 
or if no other measures capture a key focus in Star Ratings. CMS will 
take this comment into consideration as we make future enhancements in 
the Star Ratings program.
    Comment: Multiple commenters supported assigning new measures a 
weight of 1 for the first year.
    Response: CMS appreciates the support of the proposed weighting for 
new measures.
    Comment: A commenter supported the weighting for the adherence 
measures in Puerto Rico.
    Response: CMS appreciates the support of the proposed Puerto Rico 
weights.
    For the reasons set forth in the proposed rule and our responses to 
the related comments summarized above, we are finalizing the provisions 
governing the weight of measures as proposed in Sec. Sec.  422.166(e) 
and 423.186(e) with modification. CMS is finalizing the weight of 
patient experience/complaints and access measures at 2 in paragraphs 
(e)(iii) and (iv) given the importance of hearing the perspectives and 
voice of patients in times of need.
r. Application of the Improvement Measure Scores
    Consistent with current policy, we proposed at Sec. Sec.  
422.166(g) and 423.186(g) a hold harmless provision for the inclusion 
or exclusion of the improvement measure(s) for highly-rated contracts' 
highest ratings. We proposed, in paragraphs (g)(1)(i) through (iii), a 
series of rules that specify when the improvement measure is included 
in calculating overall and summary ratings.
    Under our proposal, MA-PDs would have the hold harmless provisions 
for highly-rated contracts applied for the overall rating. For an MA-PD 
that receives an overall rating of 4 stars or more without the use of 
the improvement measures and with all applicable adjustments (CAI and 
the reward factor), a comparison of the rounded overall rating with and 
without the improvement measures would be done. The overall rating with 
the improvement measures used in the comparison would include up to two 
adjustments, the reward factor (if applicable) and the CAI. The overall 
rating without the improvement measures used in the comparison would 
include up to two adjustments, the reward factor (if applicable) and 
the CAI. The higher overall rating would be used for the MA-PD 
contract's overall rating. For an MA-PD that has an overall rating of 2 
stars or less without the use of the improvement measure and with all 
applicable adjustments (CAI and the reward factor), we proposed the 
overall rating would exclude the improvement measures; for all others, 
the overall rating would include the improvement measure.
    MA-only and PDPs would have the hold harmless provisions for 
highly-

[[Page 16578]]

rated contracts applied for the Part C and D summary ratings, 
respectively. For an MA-only or PDP contract that receives a summary 
rating (with applicable adjustments) of 4 stars or more without the use 
of the improvement measure, a comparison of the rounded summary rating 
with and without the improvement measure would be done. The higher 
summary rating would be used for the summary rating for the contract's 
highest rating. For MA-only and PDPs with a summary rating (with 
applicable adjustments) of 2 stars or less without the use of the 
improvement measure would exclude the improvement measure. For all 
others, the summary rating would include the improvement measure. MA-
PDs would have their summary ratings calculated with the use of the 
improvement measure regardless of the value of the summary rating.
    In addition, at paragraph (g)(2), we also proposed text to clarify 
that summary ratings use only the improvement measure associated with 
the applicable Part C or D performance.
    We solicited comments on the hold harmless improvement provision we 
proposed to continue to use, particularly any clarifications in how and 
when it should be applied.
    We received the following comments on our proposal and our 
responses follow:
    Comment: A commenter recommended the exclusion of the hold harmless 
provision for a highly-rated contract if the contract would realize a 
decrease in their overall rating. In addition, the commenter supported 
a hold harmless provision for plans that would be at risk of receiving 
a low performing icon due to application of the quality improvement 
measures.
    Response: CMS currently and as proposed, has a safeguard for 
highly-rated contracts. CMS applies the hold harmless provision for a 
highly-rated contract's highest rating. As proposed, a contract that 
receives 4 stars or more without the use of the improvement measures 
and with all applicable adjustments (CAI and the reward factor) will 
have their final overall rating as the higher of either the rating 
calculated including or excluding the improvement measure(s). CMS 
believes the hold harmless provision is appropriate to apply for 
highly-rated contracts since they have less room for improvement and, 
consequently, may have lower scores for the improvement measure(s).
    CMS believes that the Star Ratings should signal the true quality 
of the contract. A hold harmless provision for contracts that are in 
jeopardy of a low performing icon does not align with the intent of the 
Star Ratings program and threatens its integrity. Low performing 
contracts, including those at risk of receiving a low performing icon, 
have plenty of room for improvement and should not need a hold harmless 
provision.
    Comment: A commenter expressed support for all rules that guide the 
application of the improvement measure(s) in calculating overall and 
summary ratings.
    Response: CMS appreciates the support of the policies that guide 
the application of the improvement measure(s) in the Star Ratings.
    Comment: Overall, commenters supported the use of the hold harmless 
provision for a highly-rated contract's highest rating. However, 
several commenters advocated a modification to the hold harmless 
provision for highly-rated MA-PDs such that the overall rating would be 
determined by the highest rating among the overall rating calculated 
with including both improvement measures, excluding both improvement 
measures, using only the Part C improvement measure, or using only the 
Part D improvement measure.
    Response: CMS appreciates the support of a hold harmless provision 
for a highly-rated contract's highest rating. CMS is committed to 
providing a true signal of the overall quality to beneficiaries who use 
Medicare Plan Finder to aid in the selection of a plan that is right 
for them. Eliminating the use of one of the improvement measure ratings 
in calculating the overall rating has the potential to distort the 
signal for beneficiaries. The overall rating is designed as a global 
rating of the quality of both the health plan and prescription drug 
plan benefits for an MA-PD. While we do agree there is justification 
for a hold harmless provision for a highly-rated MA-PD, CMS is 
committed to preserving the integrity of the rating system. Removing 
one facet of the rating system (Part C or Part D improvement measure) 
while not the other, has the potential to undermine the primary 
function of the rating system. Therefore, we are not finalizing the 
revisions requested by the commenter(s).
    Comment: Some commenters did not support excluding the improvement 
measure(s) from use in a contract's highest rating (with applicable 
adjustments) if the contract received 2 stars or less without the use 
of the improvement measure. The commenters believed that limiting the 
measure to only plans with at least 2.5 stars goes against the 
objective of the improvement measure in encouraging and rewarding 
improvements in performance, particularly among lower-rated plans.
    Response: CMS appreciates the careful review of the proposed policy 
related to the application of the improvement measure(s) for a 
contract's highest rating. After thoughtful deliberation of the 
recommendation of our commenters, CMS has decided to modify the 
proposed methodology for the application of the improvement measures. 
The methodology will be changed such that if the highest rating for a 
contract is less than 4 stars without the use of the improvement 
measure(s) and with all applicable adjustments (CAI and the reward 
factor), the rating will be calculated with the improvement measure(s). 
The modification of the application of the improvement measure(s) 
preserves the safeguard for a highly-rated contract's highest rating, 
but removes what could be perceived as a safeguard for contracts with a 
highest rating of 2 stars or less. In other words, if an MA-PD has an 
overall rating of less than 4 stars without the use of the improvement 
measures and with all applicable adjustments, the improvement measures 
will be used in the calculation of the overall rating. If an MA-only 
contract has a Part C summary rating of less than 4 stars without the 
use of the Part C improvement measure and with all applicable 
adjustments, the Part C improvement measure will be used in the 
determination of the contract's Part C summary rating. If a PDP has a 
Part D summary rating of less than 4 stars without the use of the Part 
D improvement measure and with all applicable adjustments, the Part D 
improvement measure will be used in the determination of the contract's 
Part D summary rating. (An MA-PD will have the Part C or Part D 
improvement measure included in the calculation of the respective Part 
C and Part D summary ratings, because the summary ratings are not the 
highest rating for this type of contract.) The only modification will 
be for contracts with a highest rating of 2 stars or less. After 
consideration of the comments received, we believe it is reasonable to 
also include any applicable improvement measure(s) for contracts with a 
highest rating of 2 stars or less so that the highest rating reflects 
whether the overall quality is improving, staying the same, or 
declining.
Summary of Regulatory Changes
    For the reasons set forth in the proposed rule and our responses to 
the related comments summarized above, we are finalizing the provisions 
addressing use of the improvement measure in summary and overall 
ratings as proposed at Sec. Sec.  422.162(g) and

[[Page 16579]]

423.182(g) with one substantive modification. We are not finalizing 
what was proposed for contracts with a 2-star summary or overall rating 
(with applicable adjustments). We are also finalizing a revision to the 
rule for summary or overall ratings (with applicable adjustments) of 
less than 4 stars to include as well contracts with overall or summary 
ratings of 2 stars.
s. Reward Factor (Formerly Referred to as Integration Factor)
    In 2011, the integration factor was added to the Star Ratings 
methodology to reward contracts that have consistently high 
performance. The integration factor was later renamed the reward 
factor. (The reference to either reward or integration factor refers to 
the same aspect of the Star Ratings.) This factor is calculated 
separately for the Part C summary rating, Part D summary rating for MA-
PDs, Part D summary rating for PDPs, and the overall rating for MA-PDs. 
It is currently added to the summary (Part C or D) and overall rating 
of contracts that have both high and stable relative performance for 
the associated summary or overall rating. The contract's performance is 
assessed using its weighted mean relative to all rated contracts 
without adjustments.
    We proposed to codify the calculation and use of the reward factor 
in Sec. Sec.  422.166(f)(1) and 423.186(f)(1); our proposal was to 
generally codify the current practice for the reward factor. Under our 
proposal, the contract's stability of performance would be assessed 
using its weighted variance relative to all rated contracts at the same 
rating level (overall, summary Part C, and summary Part D). The Part D 
summary thresholds for MA-PDs would be, like current practice 
determined independently of the thresholds for PDPs.
    We proposed to update annually the performance and variance 
thresholds for the reward factor based upon the data for the Star 
Ratings year, consistent with current policy. A multistep process would 
be used to determine the values that correspond to the thresholds for 
the reward factors for the summary and/or overall Star Ratings for a 
contract. The determination of the reward factors would rely on the 
contract's ranking of its weighted variance and weighted mean of the 
measure-level stars to the summary or overall rating relative to the 
distribution of all contracts' weighted variance and weighted mean to 
the summary and/or overall rating. Under the proposal a contract's 
weighted variance would be calculated using the quotient of the 
following two values: (1) The product of the number of applicable 
measures based on rating-type and the sum of the products of the weight 
of each applicable measure and its squared deviation \63\ and (2) the 
product of one less than the number of applicable measures and the sum 
of the weights of the applicable measures. A contract's weighted mean 
performance would be found by calculating the quotient of the following 
two values: (1) The sum of the products of the weight of a measure and 
its associated measure-level Star Ratings of the applicable measures 
for the rating-type and (2) the sum of the weights of the applicable 
measures for the rating type. The thresholds for the categorization of 
the weighted variance and weighted mean for contracts would be based 
upon the distribution of the calculated values of all rated contracts 
of the same type. Because highly-rated contracts may have the 
improvement measure(s) excluded in the determination of their final 
highest rating, each contract's weighted variance and weighted mean 
would be calculated both with and without the improvement measures.
---------------------------------------------------------------------------

    \63\ A deviation is the difference between the performance 
measure's Star Rating and the weighted mean of all applicable 
measures for the contract.
---------------------------------------------------------------------------

    Under the methodology CMS proposed for this factor, a contract's 
weighted variance would be categorized into one of three mutually 
exclusive categories, identified in Table C8A, based upon the weighted 
variance of its measure-level Star Ratings. Its ranking would be 
relative to all other contracts' weighted variance for the rating type 
(Part C summary for MA-PDs and MA-only, overall for MA-PDs, Part D 
summary for MA-PDs, and Part D summary for PDPs), and the manner in 
which the highest rating for the contract was determined--with or 
without the improvement measure(s). For an MA-PD's Part C and D summary 
ratings, its ranking is relative to all other contracts' weighted 
variance for the rating type (Part C summary, Part D summary) with the 
improvement measure. Similarly, a contract's weighted mean would be 
categorized into one of three mutually exclusive categories, identified 
in Table C8B, based on its weighted mean of all measure-level Star 
Ratings and its ranking relative to all other contracts' weighted means 
for the rating type (Part C summary for MA-PDs and MA-only, overall, 
Part D summary for MA-PDs, and Part D summary for PDPs) and the manner 
in which the highest rating for the contract was determined--with or 
without the improvement measure(s). For an MA-PD's Part C and D summary 
ratings, its ranking would be relative to all other contracts' weighted 
means for the rating type (Part C summary, Part D summary) with the 
improvement measure. Further, the same threshold criterion would be 
employed per category regardless of whether the improvement measure was 
included or excluded in the calculation of the rating. The values that 
correspond to the thresholds would be based on the distribution of all 
rated contracts and determined with and without the improvement 
measure(s) and exclusive of any adjustments. Table C8A details the 
criteria for the categorization of a contract's weighted variance for 
the summary and overall ratings. Table C8B details the criteria for the 
categorization of a contract's weighted mean (performance) for the 
overall and summary ratings. Like current practice, the values that 
correspond to the cutoffs would be provided each year during the plan 
preview and are published in the Technical Notes.

                                      Table 8A--Categorization of a Contract Based on Its Weighted Variance Ranking
--------------------------------------------------------------------------------------------------------------------------------------------------------
                Variance category                                                                 Ranking
--------------------------------------------------------------------------------------------------------------------------------------------------------
Low.............................................  Below the 30th percentile.
Medium..........................................  At or above the 30th percentile to less than the 70th percentile.
High............................................  At or above the 70th percentile.
--------------------------------------------------------------------------------------------------------------------------------------------------------


[[Page 16580]]


                                   Table 8B--Categorization of a Contract Based on Weighted Mean (Performance) Ranking
--------------------------------------------------------------------------------------------------------------------------------------------------------
      Weighted mean (performance) category                                                        Ranking
--------------------------------------------------------------------------------------------------------------------------------------------------------
High............................................  At or above the 85th percentile.
Relatively High.................................  At or above the 65th percentile to less than the 85th percentile.
Other...........................................  Below the 65th percentile.
--------------------------------------------------------------------------------------------------------------------------------------------------------

    These definitions of high, medium, and low weighted variance 
ranking and high, relatively high, and other weighted mean ranking were 
proposed to be codified in narrative form in paragraph (f)(1)(ii).
    A contract's categorization for both weighted mean and weighted 
variance determines the value of the reward factor. Table C9 shows the 
values of the reward factor based on the weighted variance and weighted 
mean categorization; we proposed to codify these values (in a narrative 
description) in paragraph (f)(1)(iii). The weighted variance and 
weighted mean thresholds for the reward factor are available in the 
Technical Notes and updated annually.

       Table 9--Categorization of a Contract for the Reward Factor
------------------------------------------------------------------------
                                        Weighted mean
         Weighted variance              (performance)      Reward factor
------------------------------------------------------------------------
Low...............................  High................             0.4
Medium............................  High................             0.3
Low...............................  Relatively High.....             0.2
Medium............................  Relatively high.....             0.1
High..............................  Other...............             0.0
------------------------------------------------------------------------

    We proposed to continue the use of a reward factor to reward 
contracts with consistently high and stable performance over time. 
Further, we proposed to continue to employ the same methodology to 
categorize and determine the reward factor for contracts. As proposed 
in paragraphs (c)(1) and (d)(1), these reward factor adjustments would 
be applied at the summary and overall rating level.
    We received the following comments on our proposal and our 
responses follow:
    Comment: The majority of commenters were supportive of the 
continued use of the reward factor. A commenter expressed support 
specifically related to the reward methodology and the codification of 
the calculation of the reward factor.
    Response: CMS appreciates our stakeholders' support of the reward 
factor.
    Comment: A commenter expressed support of the use of a reward 
factor for the overall rating, but was concerned that the proposed (and 
current) methodology for calculating the reward factor did not 
consistently award contracts that maintained high performance and 
demonstrated incremental improvement at the measure level. Further, the 
commenter linked the potential for a high performing contract not 
receiving a reward factor to flaws in the assignment of measure cut 
points.
    Response: CMS appreciates the careful consideration of the reward 
factor. Since the reward factor is a rating-specific factor, a contract 
can qualify for the reward based on its summary or overall (or both) 
rating if a contract has both high and stable relative performance. CMS 
believes the reward factor methodology identifies the contracts that 
have both high and stable relative performance and recognizes that such 
performance may exist overall (Part C and D performance) or in one 
particular area (health plan quality and performance domain on Part C 
measures or the prescription drug plan quality and performance domain 
on Part D measures). Since the reward factor is based on a relative 
performance, it serves to incentivize plans and recognize plans that 
provide the highest and consistent level of care as reflected in their 
ratings. Ratings calculated using a consistent methodology allow to the 
identification of top performers based on rankings.
    Comment: A commenter suggested that CMS annually publish the list 
of reward factor recipients. The commenter referenced the publication 
of the Categorical Adjustment Index (CAI) final adjustment categories 
for contracts to support the request. Further, the commenter believed 
that the publication of the reward factor recipients would maintain the 
attributes of fairness and transparency of the Star Ratings system.
    Response: CMS appreciates this feedback. As noted in the comment, 
the CAI final adjustment categories per contract are available in the 
annual public use files available using the following link: http://go.cms.gov/partcanddstarratings. While the thresholds for the reward 
factor are published each year in the Technical Notes, the recipients 
of the reward factor are not part of the public use files. However, we 
are persuaded that this is important information for beneficiaries and 
could assist in providing greater transparency into the development and 
assignment of the Star Ratings. Therefore, CMS will begin incorporating 
information related to the distribution and characteristics of 
contracts receiving the reward factor in the annual Fact Sheet for the 
2021 Star Ratings.
Summary of Regulatory Changes
    For the reasons set forth in the proposed rule and our responses to 
the related comments summarized above, we are finalizing the provisions 
as proposed at Sec. Sec.  422.162(f1) and 423.182(f)(1) without 
modification.
t. Categorical Adjustment Index
    As we discussed in the proposed rule, a growing body of evidence 
links the prevalence of beneficiary-level social risk factors with 
performance on measures included in Medicare value-based purchasing 
programs, including MA and Part D Star Ratings. With support from our 
contractors, we undertook research to provide scientific evidence as to 
whether MA organizations or Part D sponsors that enroll a 
disproportionate number of vulnerable beneficiaries are systematically 
disadvantaged by the current Star Ratings. In 2014, we issued a Request 
for Information to gather information directly from organizations

[[Page 16581]]

to supplement the data that CMS collects, as we believe that plans and 
sponsors are uniquely positioned to provide both qualitative and 
quantitative information that is not available from other sources. In 
February and September 2015, we released details on the findings of our 
rese