[Federal Register Volume 84, Number 153 (Thursday, August 8, 2019)]
[Rules and Regulations]
[Pages 39054-39173]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 2019-16603]



[[Page 39053]]

Vol. 84

Thursday,

No. 153

August 8, 2019

Part II





Department of Health and Human Services





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





Centers for Medicare & Medicaid Services





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





42 CFR Part 412





Medicare Program; Inpatient Rehabilitation Facility (IRF) Prospective 
Payment System for Federal Fiscal Year 2020 and Updates to the IRF 
Quality Reporting Program; Final Rule

  Federal Register / Vol. 84 , No. 153 / Thursday, August 8, 2019 / 
Rules and Regulations  

[[Page 39054]]


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

DEPARTMENT OF HEALTH AND HUMAN SERVICES

Centers for Medicare & Medicaid Services

42 CFR Part 412

[CMS-1710-F]
RIN 0938-AT67


Medicare Program; Inpatient Rehabilitation Facility (IRF) 
Prospective Payment System for Federal Fiscal Year 2020 and Updates to 
the IRF Quality Reporting Program

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

ACTION: Final rule.

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

SUMMARY: This final rule updates the prospective payment rates for 
inpatient rehabilitation facilities (IRFs) for federal fiscal year (FY) 
2020. As required by the statute, this final rule includes the 
classification and weighting factors for the IRF prospective payment 
system's (PPS) case-mix groups (CMGs) and a description of the 
methodologies and data used in computing the prospective payment rates 
for FY 2020. This final rule rebases and revises the IRF market basket 
to reflect a 2016 base year rather than the current 2012 base year. 
Additionally, this final rule revises the CMGs and updates the CMG 
relative weights and average length of stay (LOS) values beginning with 
FY 2020, based on analysis of 2 years of data (FYs 2017 and 2018). 
Although we proposed to use a weighted motor score to assign patients 
to CMGs, we are finalizing based on public comments the use of an 
unweighted motor score to assign patients to CMGs beginning with FY 
2020. Additionally, we are finalizing the removal of one item from the 
motor score. We are updating the IRF wage index to use the concurrent 
fiscal year inpatient prospective payment system (IPPS) wage index 
beginning with FY 2020. We are amending the regulations to clarify that 
the determination as to whether a physician qualifies as a 
rehabilitation physician (that is, a licensed physician with 
specialized training and experience in inpatient rehabilitation) is 
made by the IRF. For the IRF Quality Reporting Program (QRP), we are 
adopting two new measures, modifying an existing measure, and adopting 
new standardized patient assessment data elements. We are also making 
updates to reflect our migration to a new data submission system.

DATES: 
    Effective date: These regulations are effective on October 1, 2019.
    Applicability dates: The updated IRF prospective payment rates are 
applicable for IRF discharges occurring on or after October 1, 2019, 
and on or before September 30, 2020 (FY 2020). The new and updated 
quality measures and reporting requirements under the IRF QRP are 
applicable for IRF discharges occurring on or after October 1, 2020.

FOR FURTHER INFORMATION CONTACT:
    Gwendolyn Johnson, (410) 786-6954, for general information.
    Catie Kraemer, (410) 786-0179, for information about the IRF 
payment policies and payment rates.
    Kadie Derby, (410) 786-0468, for information about the IRF coverage 
policies.
    Kate Brooks, (410) 786-7877, for information about the IRF quality 
reporting program.

SUPPLEMENTARY INFORMATION: 
    Inspection of Public Comments: All comments received before the 
close of the comment period are available for viewing by the public, 
including any personally identifiable or confidential business 
information that is included in a comment. We post all comments 
received before the close of the comment period as soon as possible 
after they have been received at http://www.regulations.gov. Follow the 
search instructions on that website to view public comments.
    The IRF PPS Addenda along with other supporting documents and 
tables referenced in this final rule are available through the internet 
on the CMS website at http://www.cms.hhs.gov/Medicare/Medicare-Fee-for-Service-Payment/InpatientRehabFacPPS/.

Executive Summary

A. Purpose

    This final rule updates the prospective payment rates for IRFs for 
FY 2020 (that is, for discharges occurring on or after October 1, 2019, 
and on or before September 30, 2020) as required under section 
1886(j)(3)(C) of the Social Security Act (the Act). As required by 
section 1886(j)(5) of the Act, this final rule includes the 
classification and weighting factors for the IRF PPS's case-mix groups 
(CMGs) and a description of the methodologies and data used in 
computing the prospective payment rates for FY 2020. This final rule 
also rebases and revises the IRF market basket to reflect a 2016 base 
year, rather than the current 2012 base year. Additionally, this final 
rule revises the CMGs and updates the CMG relative weights and average 
LOS values beginning with FY 2020, based on analysis of 2 years of data 
(FYs 2017 and 2018). Although we proposed to use a weighted motor score 
to assign patients to CMGs, we are finalizing based on public comments 
the use of an unweighted motor score to assign patients to CMGs 
beginning with FY 2020. Additionally, we are finalizing the removal of 
one item from the motor score. We are also updating the IRF wage index 
to use the concurrent FY IPPS wage index for the IRF PPS beginning with 
FY 2020. We are also amending the regulations at 42 CFR 412.622 to 
clarify that the determination as to whether a physician qualifies as a 
rehabilitation physician (that is, a licensed physician with 
specialized training and experience in inpatient rehabilitation) is 
made by the IRF. For the IRF QRP, we are adopting two new measures, 
modifying an existing measure, and adopting new standardized patient 
assessment data elements. We also include updates related to the system 
used for the submission of data and related regulation text. We are not 
finalizing our proposal requiring that IRFs submit data on measures and 
standardized patient assessment data for which the source of the data 
is the IRF-PAI to all patients, regardless of payer, but plan to 
propose this policy in future rulemaking.

B. Summary of Major Provisions

    In this final rule, we use the methods described in the FY 2019 IRF 
PPS final rule (83 FR 38514) to update the prospective payment rates 
for FY 2020 using updated FY 2018 IRF claims and the most recent 
available IRF cost report data, which is FY 2017 IRF cost report data. 
This final rule also rebases and revises the IRF market basket to 
reflect a 2016 base year rather than the current 2012 base year. 
Additionally, this final rule revises the CMGs and updates the CMG 
relative weights and average LOS values beginning with FY 2020, based 
on analysis of 2 years of data (FYs 2017 and 2018). Although we 
proposed to use a weighted motor score to assign patients to CMGs, we 
are finalizing based on public comments the use of an unweighted motor 
score to assign patients to CMGs beginning with FY 2020. Additionally, 
we are finalizing the removal of one item from the motor score. We are 
also updating the IRF wage index to use the concurrent FY IPPS wage 
index for the IRF PPS beginning in FY 2020. We are also amending the 
regulations at Sec.  412.622 to clarify that the determination as to 
whether a physician qualifies as a rehabilitation physician (that is, a 
licensed physician with specialized

[[Page 39055]]

training and experience in inpatient rehabilitation) is made by the 
IRF. We also update requirements for the IRF QRP.

C. Summary of Impacts
[GRAPHIC] [TIFF OMITTED] TR08AU19.000

I. Background

A. Historical Overview of the IRF PPS

    Section 1886(j) of the Act provides for the implementation of a 
per-discharge PPS for inpatient rehabilitation hospitals and inpatient 
rehabilitation units of a hospital (collectively, hereinafter referred 
to as IRFs). Payments under the IRF PPS encompass inpatient operating 
and capital costs of furnishing covered rehabilitation services (that 
is, routine, ancillary, and capital costs), but not direct graduate 
medical education costs, costs of approved nursing and allied health 
education activities, bad debts, and other services or items outside 
the scope of the IRF PPS. Although a complete discussion of the IRF PPS 
provisions appears in the original FY 2002 IRF PPS final rule (66 FR 
41316) and the FY 2006 IRF PPS final rule (70 FR 47880), we are 
providing a general description of the IRF PPS for FYs 2002 through 
2019.
    Under the IRF PPS from FY 2002 through FY 2005, the prospective 
payment rates were computed across 100 distinct CMGs, as described in 
the FY 2002 IRF PPS final rule (66 FR 41316). We constructed 95 CMGs 
using rehabilitation impairment categories (RICs), functional status 
(both motor and cognitive), and age (in some cases, cognitive status 
and age may not be a factor in defining a CMG). In addition, we 
constructed five special CMGs to account for very short stays and for 
patients who expire in the IRF.
    For each of the CMGs, we developed relative weighting factors to 
account for a patient's clinical characteristics and expected resource 
needs. Thus, the weighting factors accounted for the relative 
difference in resource use across all CMGs. Within each CMG, we created 
tiers based on the estimated effects that certain comorbidities would 
have on resource use.
    We established the federal PPS rates using a standardized payment 
conversion factor (formerly referred to as the budget-neutral 
conversion factor). For a detailed discussion of the budget-neutral 
conversion factor, please refer to our FY 2004 IRF PPS final rule (68 
FR 45684 through 45685). In the FY 2006 IRF PPS final rule (70 FR 
47880), we discussed in detail the methodology for determining the 
standard payment conversion factor.
    We applied the relative weighting factors to the standard payment 
conversion factor to compute the unadjusted prospective payment rates 
under the IRF PPS from FYs 2002 through 2005. Within the structure of 
the payment system, we then made adjustments to account for interrupted 
stays, transfers, short stays, and deaths. Finally, we applied the 
applicable adjustments to account for geographic variations in wages 
(wage index), the percentage of low-income patients, location in a 
rural area (if applicable), and outlier payments (if applicable) to the 
IRFs' unadjusted prospective payment rates.
    For cost reporting periods that began on or after January 1, 2002, 
and before October 1, 2002, we determined the final prospective payment 
amounts using the transition methodology prescribed in section 
1886(j)(1) of the Act. Under this provision, IRFs transitioning into 
the PPS were paid a blend of the federal IRF PPS rate and the payment 
that the IRFs would have received had the IRF PPS not been implemented. 
This provision also allowed IRFs to elect to bypass this blended 
payment and immediately be paid 100 percent of the federal IRF PPS 
rate. The transition methodology expired as of cost reporting periods 
beginning on or after October 1, 2002 (FY 2003), and payments for all 
IRFs now consist of 100 percent of the federal IRF PPS rate.
    Section 1886(j) of the Act confers broad statutory authority upon 
the Secretary to propose refinements to the IRF PPS. In the FY 2006 IRF 
PPS final rule (70 FR 47880) and in correcting amendments to the FY 
2006 IRF PPS final rule (70 FR 57166), we finalized a number of 
refinements to the IRF PPS case-mix classification system (the CMGs and 
the corresponding relative weights) and the case-level and facility-
level adjustments. These refinements included the adoption of the 
Office of Management and Budget's (OMB) Core-Based Statistical Area 
(CBSA) market definitions; modifications to the CMGs, tier 
comorbidities; and CMG relative weights, implementation of a new 
teaching status adjustment for IRFs; rebasing and revising the market 
basket index used to update IRF payments, and updates to the rural, 
low-income percentage (LIP), and high-cost outlier adjustments. 
Beginning with the FY 2006 IRF PPS final rule (70 FR 47908 through 
47917), the market basket index used to update IRF payments was a 
market basket reflecting the operating and capital cost structures for 
freestanding IRFs, freestanding inpatient psychiatric facilities 
(IPFs), and long-term care hospitals (LTCHs) (hereinafter referred to 
as the rehabilitation, psychiatric, and long-term care (RPL) market 
basket). Any reference to the FY 2006 IRF PPS final rule in this final 
rule also includes the provisions effective in the correcting 
amendments. For a detailed discussion of the final key policy changes 
for FY 2006, please refer to the FY 2006 IRF PPS final rule.
    In the FY 2007 IRF PPS final rule (71 FR 48354), we further refined 
the IRF PPS case-mix classification system (the CMG relative weights) 
and the case-level adjustments, to ensure that IRF PPS payments would 
continue to reflect as accurately as possible the costs of care. For a 
detailed discussion of the FY 2007 policy revisions, please refer to 
the FY 2007 IRF PPS final rule.
    In the FY 2008 IRF PPS final rule (72 FR 44284), we updated the 
prospective payment rates and the outlier threshold, revised the IRF 
wage index policy, and clarified how we determine high-cost outlier 
payments for transfer cases. For more information on the policy changes

[[Page 39056]]

implemented for FY 2008, please refer to the FY 2008 IRF PPS final 
rule.
    After publication of the FY 2008 IRF PPS final rule (72 FR 44284), 
section 115 of the Medicare, Medicaid, and SCHIP Extension Act of 2007 
(Pub. L. 110-173, enacted December 29, 2007) (MMSEA) amended section 
1886(j)(3)(C) of the Act to apply a zero percent increase factor for 
FYs 2008 and 2009, effective for IRF discharges occurring on or after 
April 1, 2008. Section 1886(j)(3)(C) of the Act required the Secretary 
to develop an increase factor to update the IRF prospective payment 
rates for each FY. Based on the legislative change to the increase 
factor, we revised the FY 2008 prospective payment rates for IRF 
discharges occurring on or after April 1, 2008. Thus, the final FY 2008 
IRF prospective payment rates that were published in the FY 2008 IRF 
PPS final rule (72 FR 44284) were effective for discharges occurring on 
or after October 1, 2007, and on or before March 31, 2008, and the 
revised FY 2008 IRF prospective payment rates were effective for 
discharges occurring on or after April 1, 2008, and on or before 
September 30, 2008. The revised FY 2008 prospective payment rates are 
available on the CMS website at http://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/InpatientRehabFacPPS/Data-Files.html.
    In the FY 2009 IRF PPS final rule (73 FR 46370), we updated the CMG 
relative weights, the average LOS values, and the outlier threshold; 
clarified IRF wage index policies regarding the treatment of ``New 
England deemed'' counties and multi-campus hospitals; and revised the 
regulation text in response to section 115 of the MMSEA to set the IRF 
compliance percentage at 60 percent (the ``60 percent rule'') and 
continue the practice of including comorbidities in the calculation of 
compliance percentages. We also applied a zero percent market basket 
increase factor for FY 2009 in accordance with section 115 of the 
MMSEA. For more information on the policy changes implemented for FY 
2009, please refer to the FY 2009 IRF PPS final rule.
    In the FY 2010 IRF PPS final rule (74 FR 39762) and in correcting 
amendments to the FY 2010 IRF PPS final rule (74 FR 50712), we updated 
the prospective payment rates, the CMG relative weights, the average 
LOS values, the rural, LIP, teaching status adjustment factors, and the 
outlier threshold; implemented new IRF coverage requirements for 
determining whether an IRF claim is reasonable and necessary; and 
revised the regulation text to require IRFs to submit patient 
assessments on Medicare Advantage (MA) (formerly called Medicare Part 
C) patients for use in the 60 percent rule calculations. Any reference 
to the FY 2010 IRF PPS final rule in this final rule also includes the 
provisions effective in the correcting amendments. For more information 
on the policy changes implemented for FY 2010, please refer to the FY 
2010 IRF PPS final rule.
    After publication of the FY 2010 IRF PPS final rule (74 FR 39762), 
section 3401(d) of the Patient Protection and Affordable Care Act (Pub. 
L. 111-148, enacted March 23, 2010), as amended by section 10319 of the 
same Act and by section 1105 of the Health Care and Education 
Reconciliation Act of 2010 (Pub. L. 111-152, enacted March 30, 2010) 
(collectively, hereinafter referred to as ``PPACA''), amended section 
1886(j)(3)(C) of the Act and added section 1886(j)(3)(D) of the Act. 
Section 1886(j)(3)(C) of the Act requires the Secretary to estimate a 
multifactor productivity (MFP) adjustment to the market basket increase 
factor, and to apply other adjustments as defined by the Act. The 
productivity adjustment applies to FYs from 2012 forward. The other 
adjustments apply to FYs 2010 to 2019.
    Sections 1886(j)(3)(C)(ii)(II) and 1886(j)(3)(D)(i) of the Act 
defined the adjustments that were to be applied to the market basket 
increase factors in FYs 2010 and 2011. Under these provisions, the 
Secretary was required to reduce the market basket increase factor in 
FY 2010 by a 0.25 percentage point adjustment. Notwithstanding this 
provision, in accordance with section 3401(p) of the PPACA, the 
adjusted FY 2010 rate was only to be applied to discharges occurring on 
or after April 1, 2010. Based on the self-implementing legislative 
changes to section 1886(j)(3) of the Act, we adjusted the FY 2010 
prospective payment rates as required, and applied these rates to IRF 
discharges occurring on or after April 1, 2010, and on or before 
September 30, 2010. Thus, the final FY 2010 IRF prospective payment 
rates that were published in the FY 2010 IRF PPS final rule (74 FR 
39762) were used for discharges occurring on or after October 1, 2009, 
and on or before March 31, 2010, and the adjusted FY 2010 IRF 
prospective payment rates applied to discharges occurring on or after 
April 1, 2010, and on or before September 30, 2010. The adjusted FY 
2010 prospective payment rates are available on the CMS website at 
https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/InpatientRehabFacPPS/IRF-Rules-and-Related-Files.html.
    In addition, sections 1886(j)(3)(C) and (D) of the Act also 
affected the FY 2010 IRF outlier threshold amount because they required 
an adjustment to the FY 2010 RPL market basket increase factor, which 
changed the standard payment conversion factor for FY 2010. 
Specifically, the original FY 2010 IRF outlier threshold amount was 
determined based on the original estimated FY 2010 RPL market basket 
increase factor of 2.5 percent and the standard payment conversion 
factor of $13,661. However, as adjusted, the IRF prospective payments 
were based on the adjusted RPL market basket increase factor of 2.25 
percent and the revised standard payment conversion factor of $13,627. 
To maintain estimated outlier payments for FY 2010 equal to the 
established standard of 3 percent of total estimated IRF PPS payments 
for FY 2010, we revised the IRF outlier threshold amount for FY 2010 
for discharges occurring on or after April 1, 2010, and on or before 
September 30, 2010. The revised IRF outlier threshold amount for FY 
2010 was $10,721.
    Sections 1886(j)(3)(C)(ii)(II) and 1886(j)(3)(D)(i) of the Act also 
required the Secretary to reduce the market basket increase factor in 
FY 2011 by a 0.25 percentage point adjustment. The FY 2011 IRF PPS 
notice (75 FR 42836) and the correcting amendments to the FY 2011 IRF 
PPS notice (75 FR 70013) described the required adjustments to the FY 
2010 and FY 2011 IRF PPS prospective payment rates and outlier 
threshold amount for IRF discharges occurring on or after April 1, 
2010, and on or before September 30, 2011. It also updated the FY 2011 
prospective payment rates, the CMG relative weights, and the average 
LOS values. Any reference to the FY 2011 IRF PPS notice in this final 
rule also includes the provisions effective in the correcting 
amendments. For more information on the FY 2010 and FY 2011 adjustments 
or the updates for FY 2011, please refer to the FY 2011 IRF PPS notice.
    In the FY 2012 IRF PPS final rule (76 FR 47836), we updated the IRF 
prospective payment rates, rebased and revised the RPL market basket, 
and established a new QRP for IRFs in accordance with section 
1886(j)(7) of the Act. We also consolidated, clarified, and revised 
existing policies regarding IRF hospitals and IRF units of hospitals to 
eliminate unnecessary confusion and enhance consistency. For more 
information on the policy changes implemented for FY 2012, please refer 
to the FY 2012 IRF PPS final rule.
    The FY 2013 IRF PPS notice (77 FR 44618) described the required 
adjustments to the FY 2013 prospective

[[Page 39057]]

payment rates and outlier threshold amount for IRF discharges occurring 
on or after October 1, 2012, and on or before September 30, 2013. It 
also updated the FY 2013 prospective payment rates, the CMG relative 
weights, and the average LOS values. For more information on the 
updates for FY 2013, please refer to the FY 2013 IRF PPS notice.
    In the FY 2014 IRF PPS final rule (78 FR 47860), we updated the 
prospective payment rates, the CMG relative weights, and the outlier 
threshold amount. We also updated the facility-level adjustment factors 
using an enhanced estimation methodology, revised the list of diagnosis 
codes that count toward an IRF's 60 percent rule compliance calculation 
to determine ``presumptive compliance,'' revised sections of the IRF 
patient assessment instrument (IRF-PAI), revised requirements for acute 
care hospitals that have IRF units, clarified the IRF regulation text 
regarding limitation of review, updated references to previously 
changed sections in the regulations text, and updated requirements for 
the IRF QRP. For more information on the policy changes implemented for 
FY 2014, please refer to the FY 2014 IRF PPS final rule.
    In the FY 2015 IRF PPS final rule (79 FR 45872) and the correcting 
amendments to the FY 2015 IRF PPS final rule (79 FR 59121), we updated 
the prospective payment rates, the CMG relative weights, and the 
outlier threshold amount. We also revised the list of diagnosis codes 
that count toward an IRF's 60 percent rule compliance calculation to 
determine ``presumptive compliance,'' revised sections of the IRF-PAI, 
and updated requirements for the IRF QRP. Any reference to the FY 2015 
IRF PPS final rule in this final rule also includes the provisions 
effective in the correcting amendments. For more information on the 
policy changes implemented for FY 2015, please refer to the FY 2015 IRF 
PPS final rule.
    In the FY 2016 IRF PPS final rule (80 FR 47036), we updated the 
prospective payment rates, the CMG relative weights, and the outlier 
threshold amount. We also adopted an IRF-specific market basket that 
reflects the cost structures of only IRF providers, a blended 1-year 
transition wage index based on the adoption of new OMB area 
delineations, a 3-year phase-out of the rural adjustment for certain 
IRFs due to the new OMB area delineations, and updates for the IRF QRP. 
For more information on the policy changes implemented for FY 2016, 
please refer to the FY 2016 IRF PPS final rule.
    In the FY 2017 IRF PPS final rule (81 FR 52056) and the correcting 
amendments to the FY 2017 IRF PPS final rule (81 FR 59901), we updated 
the prospective payment rates, the CMG relative weights, and the 
outlier threshold amount. We also updated requirements for the IRF QRP. 
Any reference to the FY 2017 IRF PPS final rule in this final rule also 
includes the provisions effective in the correcting amendments. For 
more information on the policy changes implemented for FY 2017, please 
refer to the FY 2017 IRF PPS final rule.
    In the FY 2018 IRF PPS final rule (82 FR 36238), we updated the 
prospective payment rates, the CMG relative weights, and the outlier 
threshold amount. We also revised the International Classification of 
Diseases, 10th Revision, Clinical Modification (ICD-10-CM) diagnosis 
codes that are used to determine presumptive compliance under the ``60 
percent rule,'' removed the 25 percent payment penalty for IRF-PAI late 
transmissions, removed the voluntary swallowing status item (Item 27) 
from the IRF-PAI, summarized comments regarding the criteria used to 
classify facilities for payment under the IRF PPS, provided for a 
subregulatory process for certain annual updates to the presumptive 
methodology diagnosis code lists, adopted the use of height/weight 
items on the IRF-PAI to determine patient body mass index (BMI) greater 
than 50 for cases of single-joint replacement under the presumptive 
methodology, and updated requirements for the IRF QRP. For more 
information on the policy changes implemented for FY 2018, please refer 
to the FY 2018 IRF PPS final rule.
    In the FY 2019 IRF PPS final rule (83 FR 38514), we updated the 
prospective payment rates, the CMG relative weights, and the outlier 
threshold amount. We also alleviated administrative burden for IRFs by 
removing the FIMTM instrument and associated Function 
Modifiers from the IRF-PAI beginning in FY 2020 and revised certain IRF 
coverage requirements to reduce the amount of required paperwork in the 
IRF setting beginning in FY 2019. Additionally, we incorporated certain 
data items located in the Quality Indicators section of the IRF-PAI 
into the IRF case-mix classification system using analysis of 2 years 
of data (FYs 2017 and 2018) beginning in FY 2020. For the IRF QRP, we 
adopted a new measure removal factor, removed two measures from the IRF 
QRP measure set, and codified a number of program requirements in our 
regulations. For more information on the policy changes implemented for 
FY 2019, please refer to the FY 2019 IRF PPS final rule.

B. Provisions of the PPACA Affecting the IRF PPS in FY 2012 and Beyond

    The PPACA included several provisions that affect the IRF PPS in 
FYs 2012 and beyond. In addition to what was previously discussed, 
section 3401(d) of the PPACA also added section 1886(j)(3)(C)(ii)(I) of 
the Act (providing for a ``productivity adjustment'' for FY 2012 and 
each subsequent fiscal year). The productivity adjustment for FY 2020 
is discussed in section VI.D. of this final rule. Section 
1886(j)(3)(C)(ii)(II) of the Act provides that the application of the 
productivity adjustment to the market basket update may result in an 
update that is less than 0.0 for a fiscal year and in payment rates for 
a fiscal year being less than such payment rates for the preceding 
fiscal year.
    Sections 3004(b) of the PPACA and section 411(b) of the Medicare 
Access and CHIP Reauthorization Act of 2015 (Pub. L. 114-10, enacted 
April 16, 2015) (MACRA) also addressed the IRF PPS. Section 3004(b) of 
PPACA reassigned the previously designated section 1886(j)(7) of the 
Act to section 1886(j)(8) of the Act and inserted a new section 
1886(j)(7) of the Act, which contains requirements for the Secretary to 
establish a QRP for IRFs. Under that program, data must be submitted in 
a form and manner and at a time specified by the Secretary. Beginning 
in FY 2014, section 1886(j)(7)(A)(i) of the Act requires the 
application of a 2 percentage point reduction to the market basket 
increase factor otherwise applicable to an IRF (after application of 
paragraphs (C)(iii) and (D) of section 1886(j)(3) of the Act) for a 
fiscal year if the IRF does not comply with the requirements of the IRF 
QRP for that fiscal year. Application of the 2 percentage point 
reduction may result in an update that is less than 0.0 for a fiscal 
year and in payment rates for a fiscal year being less than such 
payment rates for the preceding fiscal year. Reporting-based reductions 
to the market basket increase factor are not cumulative; they only 
apply for the FY involved. Section 411(b) of MACRA amended section 
1886(j)(3)(C) of the Act by adding paragraph (iii), which required us 
to apply for FY 2018, after the application of section 
1886(j)(3)(C)(ii) of the Act, an increase factor of 1.0 percent to 
update the IRF prospective payment rates.

[[Page 39058]]

C. Operational Overview of the Current IRF PPS

    As described in the FY 2002 IRF PPS final rule (66 FR 41316), upon 
the admission and discharge of a Medicare Part A Fee-for-Service (FFS) 
patient, the IRF is required to complete the appropriate sections of a 
PAI, designated as the IRF-PAI. In addition, beginning with IRF 
discharges occurring on or after October 1, 2009, the IRF is also 
required to complete the appropriate sections of the IRF-PAI upon the 
admission and discharge of each Medicare Advantage (MA) patient, as 
described in the FY 2010 IRF PPS final rule (74 FR 39762 and 74 FR 
50712). All required data must be electronically encoded into the IRF-
PAI software product. Generally, the software product includes patient 
classification programming called the Grouper software. The Grouper 
software uses specific IRF-PAI data elements to classify (or group) 
patients into distinct CMGs and account for the existence of any 
relevant comorbidities.
    The Grouper software produces a five-character CMG number. The 
first character is an alphabetic character that indicates the 
comorbidity tier. The last four characters are numeric characters that 
represent the distinct CMG number. Free downloads of the Inpatient 
Rehabilitation Validation and Entry (IRVEN) software product, including 
the Grouper software, are available on the CMS website at http://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/InpatientRehabFacPPS/Software.html.
    Once a Medicare Part A FFS patient is discharged, the IRF submits a 
Medicare claim as a Health Insurance Portability and Accountability Act 
of 1996 (Pub. L. 104-191, enacted August 21, 1996) (HIPAA) compliant 
electronic claim or, if the Administrative Simplification Compliance 
Act of 2002 (Pub. L. 107-105, enacted December 27, 2002) (ASCA) 
permits, a paper claim (a UB-04 or a CMS-1450 as appropriate) using the 
five-character CMG number and sends it to the appropriate Medicare 
Administrative Contractor (MAC). In addition, once a MA patient is 
discharged, in accordance with the Medicare Claims Processing Manual, 
chapter 3, section 20.3 (Pub. 100-04), hospitals (including IRFs) must 
submit an informational-only bill (Type of Bill (TOB) 111), which 
includes Condition Code 04 to their MAC. This will ensure that the MA 
days are included in the hospital's Supplemental Security Income (SSI) 
ratio (used in calculating the IRF LIP adjustment) for fiscal year 2007 
and beyond. Claims submitted to Medicare must comply with both ASCA and 
HIPAA.
    Section 3 of the ASCA amended section 1862(a) of the Act by adding 
paragraph (22), which requires the Medicare program, subject to section 
1862(h) of the Act, to deny payment under Part A or Part B for any 
expenses for items or services for which a claim is submitted other 
than in an electronic form specified by the Secretary. Section 1862(h) 
of the Act, in turn, provides that the Secretary shall waive such 
denial in situations in which there is no method available for the 
submission of claims in an electronic form or the entity submitting the 
claim is a small provider. In addition, the Secretary also has the 
authority to waive such denial in such unusual cases as the Secretary 
finds appropriate. For more information, see the ``Medicare Program; 
Electronic Submission of Medicare Claims'' final rule (70 FR 71008). 
Our instructions for the limited number of Medicare claims submitted on 
paper are available at http://www.cms.gov/manuals/downloads/clm104c25.pdf.
    Section 3 of the ASCA operates in the context of the administrative 
simplification provisions of HIPAA, which include, among others, the 
requirements for transaction standards and code sets codified in 45 CFR 
part 160 and part 162, subparts A and I through R (generally known as 
the Transactions Rule). The Transactions Rule requires covered 
entities, including covered health care providers, to conduct covered 
electronic transactions according to the applicable transaction 
standards. (See the CMS program claim memoranda at http://www.cms.gov/ElectronicBillingEDITrans/ and listed in the addenda to the Medicare 
Intermediary Manual, Part 3, section 3600).
    The MAC processes the claim through its software system. This 
software system includes pricing programming called the ``Pricer'' 
software. The Pricer software uses the CMG number, along with other 
specific claim data elements and provider-specific data, to adjust the 
IRF's prospective payment for interrupted stays, transfers, short 
stays, and deaths, and then applies the applicable adjustments to 
account for the IRF's wage index, percentage of low-income patients, 
rural location, and outlier payments. For discharges occurring on or 
after October 1, 2005, the IRF PPS payment also reflects the teaching 
status adjustment that became effective as of FY 2006, as discussed in 
the FY 2006 IRF PPS final rule (70 FR 47880).

D. Advancing Health Information Exchange

    The Department of Health and Human Services (HHS) has a number of 
initiatives designed to encourage and support the adoption of 
interoperable health information technology and to promote nationwide 
health information exchange to improve health care. The Office of the 
National Coordinator for Health Information Technology (ONC) and CMS 
work collaboratively to advance interoperability across settings of 
care, including post-acute care.
    To further interoperability in post-acute care, we developed a Data 
Element Library (DEL) to serve as a publicly-available centralized, 
authoritative resource for standardized data elements and their 
associated mappings to health IT standards. The DEL furthers CMS' goal 
of data standardization and interoperability. These interoperable data 
elements can reduce provider burden by allowing the use and exchange of 
healthcare data, support provider exchange of electronic health 
information for care coordination, person-centered care, and support 
real-time, data driven, clinical decision making. Standards in the Data 
Element Library (https://del.cms.gov/) can be referenced on the CMS 
website and in the ONC Interoperability Standards Advisory (ISA). The 
2019 ISA is available at https://www.healthit.gov/isa.
    The 21st Century Cures Act (Pub. L. 114-255, enacted December 13, 
2016) (Cures Act), requires HHS to take new steps to enable the 
electronic sharing of health information ensuring interoperability for 
providers and settings across the care continuum. In another important 
provision, Congress defined ``information blocking'' as practices 
likely to interfere with, prevent, or materially discourage access, 
exchange, or use of electronic health information, and established new 
authority for HHS to discourage these practices. In March 2019, ONC and 
CMS published the proposed rules, ``21st Century Cures Act: 
Interoperability, Information Blocking, and the ONC Health IT 
Certification Program,'' (84 FR 7424) and ``Interoperability and 
Patient Access'' (84 FR 7610) to promote secure and more immediate 
access to health information for patients and healthcare providers 
through the implementation of information blocking provisions of the 
Cures Act and the use of standardized application programming 
interfaces (APIs) that enable easier access to electronic health 
information. We solicited comment on the two proposed rules. We invited 
providers to

[[Page 39059]]

learn more about these important developments and how they are likely 
to affect IRFs.

II. Summary of Provisions of the Proposed Rule

    In the FY 2020 IRF PPS proposed rule, we proposed to update the IRF 
prospective payment rates for FY 2020 and to rebase and revise the IRF 
market basket to reflect a 2016 base year rather than the current 2012 
base year. We also proposed to replace the previously finalized 
unweighted motor score with a weighted motor score to assign patients 
to CMGs and remove one item from the score beginning with FY 2020 and 
to revise the CMGs and update the CMG relative weights and average LOS 
values beginning with FY 2020, based on analysis of 2 years of data 
(FYs 2017 and 2018). We also proposed to use the concurrent FY IPPS 
wage index for the IRF PPS beginning with FY 2020. We also solicited 
comments on stakeholder concerns regarding the appropriateness of the 
wage index used to adjust IRF payments. We proposed to amend the 
regulations at Sec.  412.622 to clarify that the determination as to 
whether a physician qualifies as a rehabilitation physician (that is, a 
licensed physician with specialized training and experience in 
inpatient rehabilitation) is made by the IRF.
    The proposed policy changes and updates to the IRF prospective 
payment rates for FY 2020 are as follows:
     Describe a proposed weighted motor score to replace the 
previously finalized unweighted motor score to assign a patient to a 
CMG, the removal of one item from the score, and revisions to the CMGs 
beginning on October 1, 2019, based on analysis of 2 years of data (FYs 
2017 and 2018) using the Quality Indicator items in the IRF-PAI. This 
includes proposed revisions to the CMG relative weights and average LOS 
values for FY 2020, in a budget neutral manner, as discussed in section 
III. of the FY 2020 IRF PPS proposed rule (84 FR 17244, 17249 through 
17260).
     Describe the proposed rebased and revised IRF market 
basket to reflect a 2016 base year rather than the current 2012 base 
year as discussed in section V. of the FY 2020 IRF PPS proposed rule 
(84 FR 17244, 17261 through 17273).
     Update the IRF PPS payment rates for FY 2020 by the 
proposed market basket increase factor, based upon the most current 
data available, with a proposed productivity adjustment required by 
section 1886(j)(3)(C)(ii)(I) of the Act, as described in section V. of 
the FY 2020 IRF PPS proposed rule (84 FR 17244, 17274 through 17275).
     Describe the proposed update to the IRF wage index to use 
the concurrent FY IPPS wage index and the FY 2020 proposed labor-
related share in a budget-neutral manner, as described in section V. of 
the FY 2020 IRF PPS proposed rule (84 FR 17244, 17276 through 17279).
     Describe the continued use of FY 2014 facility-level 
adjustment factors, as discussed in section IV. of the FY 2020 IRF PPS 
proposed rule (84 FR 17244, 17260 through 17261).
     Describe the calculation of the IRF standard payment 
conversion factor for FY 2020, as discussed in section V. of the FY 
2020 IRF PPS proposed rule (84 FR 17244, 17280 through 17282).
     Update the outlier threshold amount for FY 2020, as 
discussed in section VI. of the FY 2020 IRF PPS proposed rule (84 FR 
17244, 17283 through 17284).
     Update the cost-to-charge ratio (CCR) ceiling and urban/
rural average CCRs for FY 2020, as discussed in section VI. of the FY 
2020 IRF PPS proposed rule (84 FR 17244 at 17284).
     Describe the proposed amendments to the regulations at 
Sec.  412.622 to clarify that the determination as to whether a 
physician qualifies as a rehabilitation physician (that is, a licensed 
physician with specialized training and experience in inpatient 
rehabilitation) is made by the IRF, as discussed in section VII. of the 
FY 2020 IRF PPS proposed rule (84 FR 17244, 17284 through 17285).
     Updates to the requirements for the IRF QRP, as discussed 
in section VIII. of the FY 2020 IRF PPS proposed rule (84 FR 17244, 
17285 through 17330).

III. Analysis and Response to Public Comments

    We received 1,257 timely responses from the public, many of which 
contained multiple comments on the FY 2020 IRF PPS proposed rule (84 FR 
17244). The majority consisted of form letters, in which we received 
multiple copies of two types of identically-worded letters that had 
been signed and submitted by different individuals. We received 
comments from various trade associations, IRFs, individual physicians, 
therapists, clinicians, health care industry organizations, and health 
care consulting firms. The following sections, arranged by subject 
area, include a summary of the public comments that we received, and 
our responses.

IV. Refinements to the Case-Mix Classification System Beginning With FY 
2020

A. Background

    Section 1886(j)(2)(A) of the Act requires the Secretary to 
establish CMGs for payment under the IRF PPS and a method of 
classifying specific IRF patients within these groups. Under section 
1886(j)(2)(B) of the Act, the Secretary must assign each CMG an 
appropriate weighting factor that reflects the relative facility 
resources used for patients classified within the group as compared to 
patients classified within other groups. Additionally, section 
1886(j)(2)(C)(i) of the Act requires the Secretary from time to time to 
adjust the established classifications and weighting factors as 
appropriate to reflect changes in treatment patterns, technology, case-
mix, number of payment units for which payment is made under title 
XVIII of the Act, and other factors which may affect the relative use 
of resources. Such adjustments must be made in a manner so that changes 
in aggregate payments under the classification system are a result of 
real changes and are not a result of changes in coding that are 
unrelated to real changes in case mix.
    In the FY 2019 IRF PPS final rule (83 FR 38533 through 38549), we 
finalized the removal of the Functional Independence Measure 
(FIMTM) instrument and associated Function Modifiers from 
the IRF-PAI and the incorporation of an unweighted additive motor score 
derived from 19 data items located in the Quality Indicators section of 
the IRF-PAI beginning with FY 2020 (83 FR 38535 through 38536, 38549). 
As discussed in section IV.B of this final rule, based on further 
analysis to examine the potential impact of weighting the motor score, 
we proposed to replace the previously finalized unweighted motor score 
with a weighted motor score and remove one item from the score 
beginning with FY 2020.
    Additionally, as noted in the FY 2019 IRF PPS final rule (83 FR 
38534), the incorporation of the data items from the Quality Indicator 
section of the IRF-PAI into the IRF case-mix classification system 
necessitates revisions to the CMGs to ensure that IRF payments are 
calculated accurately. We finalized the use of data items from the 
Quality Indicators section of the IRF-PAI to construct the functional 
status scores used to classify IRF patients in the IRF case-mix 
classification system for purposes of establishing payment under the 
IRF PPS beginning with FY 2020, but modified our proposal based on

[[Page 39060]]

public comments to incorporate 2 years of data (FYs 2017 and 2018) into 
our analyses used to revise the CMG definitions (83 FR 38549). We 
stated that any changes to the proposed CMG definitions resulting from 
the incorporation of an additional year of data (FY 2018) into the 
analysis would be addressed in future rulemaking prior to their 
implementation beginning in FY 2020. As discussed in section III.C of 
the FY 2020 IRF PPS proposed rule (84 FR 17244, 17250 through 17260), 
we proposed to revise the CMGs based on analysis of 2 years of data 
(FYs 2017 and 2018) beginning with FY 2020. We also proposed to update 
the relative weights and average LOS values associated with the revised 
CMGs beginning with FY 2020.

B. Proposed Use of a Weighted Motor Score Beginning With FY 2020

    As noted in the FY 2019 IRF PPS final rule (83 FR 38535), the IRF 
case-mix classification system currently uses a weighted motor score 
based on FIMTM data items to assign patients to CMGs under 
the IRF PPS through FY 2019. More information on the development and 
implementation of this motor score can be found in the FY 2006 IRF PPS 
final rule (70 FR 47896 through 47900). In the FY 2019 IRF PPS final 
rule (83 FR 38535 through 38536, 38549), we finalized the incorporation 
of an unweighted additive motor score derived from 19 data items 
located in the Quality Indicators section of the IRF-PAI beginning with 
FY 2020. We did not propose a weighted motor score at the time, because 
we believed that the unweighted motor score would facilitate greater 
understanding among the provider community, as it is less complex. 
However, we also noted that we would take comments in favor of a 
weighted motor score into consideration in future analysis. In response 
to feedback we received from various stakeholders and professional 
organizations regarding the use of an unweighted motor score and 
requesting that we consider weighting the motor score, we extended our 
contract with Research Triangle Institute, International (RTI) to 
examine the potential impact of weighting the motor score. Based on 
this analysis, discussed further below, we believed that a weighted 
motor score would improve the accuracy of payments to IRFs and proposed 
to replace the previously finalized unweighted motor score with a 
weighted motor score to assign patients to CMGs beginning with FY 2020.
    The previously finalized motor score is calculated by summing the 
scores of the 19 data items, with equal weight applied to each item. 
The 19 data items are (83 FR 38535):
     GG0130A1 Eating.
     GG0130B1 Oral hygiene.
     GG0130C1 Toileting hygiene.
     GG0130E1 Shower/bathe self.
     GG0130F1 Upper-body dressing.
     GG0130G1 Lower-body dressing.
     GG0130H1 Putting on/taking off footwear.
     GG0170A1 Roll left and right.
     GG0170B1 Sit to lying.
     GG0170C1 Lying to sitting on side of bed.
     GG0170D1 Sit to stand.
     GG0170E1 Chair/bed-to-chair transfer.
     GG0170F1 Toilet transfer.
     GG0170I1 Walk 10 feet.
     GG0170J1 Walk 50 feet with two turns.
     GG0170K1 Walk 150 feet.
     GG0170M1 One step curb.
     H0350 Bladder continence.
     H0400 Bowel continence.
    In response to feedback we received from various stakeholders and 
professional organizations requesting that we consider applying weights 
to the motor score, we extended our contract with RTI to explore the 
potential of applying unique weights to each of the 19 items in the 
motor score.
    As part of their analysis, RTI examined the degree to which the 
items used to construct the motor score were related to one another and 
adjusted their weighting methodology to account for their findings. RTI 
considered a number of different weighting methodologies to develop a 
weighted index that would increase the predictive power of the IRF 
case-mix classification system while at the same time maintaining 
simplicity. RTI used regression analysis to explore the relationship of 
the motor score items to costs. This analysis was undertaken to 
determine the impact of each of the items on cost and then to weight 
each item in the index according to its relative impact on cost. Based 
on findings from this analysis, we proposed to remove the item GG0170A1 
Roll left and right from the motor score as this item was found to have 
a high degree of multicollinearity with other items in the motor score 
and would have resulted in either a negative or non-significant 
coefficient. As such, we did not believe it would be appropriate to 
include this item in the motor score calculation. Using the revised 
motor score composed of the remaining 18 items identified above, RTI 
designed a weighting methodology for the motor score that could be 
applied uniformly across all RICs. For a more detailed discussion of 
the analysis used to construct the weighted motor score, we refer 
readers to the March 2019 technical report entitled ``Analyses to 
Inform the Use of Standardized Patient Assessment Data Elements in the 
Inpatient Rehabilitation Facility Prospective Payment System'', 
available at https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/InpatientRehabFacPPS/Research.html. Findings from this analysis 
suggested that the use of a weighted motor score index slightly 
improves the ability of the IRF PPS to predict patient costs. Based on 
this analysis, we proposed to use a weighted motor score for the 
purpose of determining IRF payments.
    Table 1 shows the proposed weights for each component of the motor 
score, averaged to 1, obtained through the regression analysis.

[[Page 39061]]

[GRAPHIC] [TIFF OMITTED] TR08AU19.001

    We proposed to determine the motor score by applying each of the 
weights indicated in Table 1 to the score of each corresponding item, 
as finalized in the FY 2019 IRF PPS final rule (83 FR 38535 through 
38537), and then summing the weighted scores for each of the 18 items 
that compose the motor score.
    We received several comments on the proposal to replace the 
previously finalized unweighted motor score with a weighted motor score 
to assign patients to CMGs under the IRF PPS and our proposal to remove 
the item GG0170A1 Roll left and right from the calculation of the motor 
score beginning with FY 2020, that is, for all discharges beginning on 
or after October 1, 2019. As summarized in more detail below, with the 
exception of one comment from MedPAC, the commenters overwhelmingly 
requested that CMS delay implementation of a weighted motor score and 
use an unweighted motor score to assign patients to CMGs until we can 
more fully analyze and work with stakeholders on developing a weighted 
motor score methodology.
    In response to public comments, we carefully considered whether to 
finalize the proposed weighted motor score or go back to using an 
unweighted motor score to assign patients to CMGs. Although the 
proposed weighted motor score results in a slight improvement in the 
ability of the IRF PPS to predict patient costs and thus the accuracy 
of IRF PPS payments (less than 0.18 difference in accuracy between the 
weighted and the unweighted motor scores), we acknowledge the 
unweighted motor score is conceptually simpler and, as such, believe it 
will ease providers' transition to the use of the data items located in 
the Quality Indicators section of the IRF-PAI (also referred to as 
section GG items). Thus, we are finalizing based on public comments the 
use of an unweighted motor score to assign patients to CMGs beginning 
with FY 2020. We appreciate the commenters' suggestions on the 
weighting methodology and will take them into consideration as we 
explore possible refinements to the case-mix classification system in 
the future.
    Comment: Although several commenters noted appreciation for the 
fact that we analyzed a weighted motor score in response to their 
comments on the FY 2019 IRF PPS proposed rule (83 FR 38546), these same 
commenters expressed concerns with the actual weight values that CMS 
proposed for FY 2020, as indicated in Table 1, and stated that we 
should go back to an unweighted motor score so that we can do further 
analysis and collaborate with stakeholders to further refine the 
weighting methodology. Some commenters expressed concern that CMS might 
be proposing higher weights for the self-care items than for the 
mobility items, in contrast to the current weighted motor score, which 
weights mobility items higher than self-care items. Some commenters 
specifically requested that CMS explain why the weight for the eating 
item increased from 0.6 under the current weighting methodology to 2.7 
under the proposed methodology, and requested we explain what we 
believe this change will mean for patients with eating deficits. 
Commenters were also generally concerned by what they suggested were 
large differences in the weight value assignments between the current 
and proposed motor score.
    Response: We used simple ordinary least squares regression analysis 
of the data that IRFs submitted to us in FYs 2017 and 2018 to calculate 
the proposed weight values for the motor score, in response to 
stakeholder feedback on the FY 2019 IRF PPS proposed rule (83 FR 
38546). Commenters are correct that the proposed weights for the motor 
score items, in comparison with the current weights, shift some of the 
weight from the mobility to the self-care items. We also note that the 
proposed weights assigned to the bowel and bladder function items 
increased compared with the current weights. These changes are all 
reflective of the data the IRFs submitted to us in FYs 2017 and 2018.
    Regarding the proposed increase in the weight for the eating item, 
it is important to note key differences in the coding guidelines 
between the FIMTM eating item and the section GG eating item 
that may have contributed to the change in the relative importance of 
this item for predicting IRF costs. For item GG0130A, Eating, 
assistance with tube feedings is not considered when coding this item. 
If a patient does not eat or drink by mouth but is instead tube fed, 
item GG0130A must be coded as 88--``Not attempted due to medical 
condition or safety concerns'' or 09--``Not applicable''. Both of these 
responses would be recoded to a 01--``Dependent'' for the purposes of 
assigning the patient to a CMG. This

[[Page 39062]]

differs from the coding instructions for the FIMTM eating 
item used in the current motor score, which takes into consideration 
assistance with tube feedings in scoring the item. For example, 
according to the FIMTM instructions, a patient who could 
administer the tube feeding completely independently could receive a 
score of 7-Complete independence on the eating item.
    In regards to the suggested differences in the weight value 
assignments between the current and proposed methodologies, we note 
that in certain cases the proposed weights were divided among multiple 
items in the motor score that were found to be highly correlated to 
avoid overweighting any particular measure of function. For instance, 
the three items (GG0170I1, GG0170J1, and GG0170K1) that assess walking 
function were each assigned a proposed weight of 0.8. When summed 
together, the weight value for walking under the proposed methodology 
is 2.4, which is slightly higher than the weight value of 1.6 for the 
single walking item used in the current motor score.
    Comment: One commenter disagreed with the removal of item GG0170A1 
roll left and right from the motor score and noted it is an important 
functional task in the IRF setting. Some commenters questioned the use 
of averaging values across pairs of items that were correlated and 
inquired why the roll left and right item was removed from the motor 
score while other correlated items were not removed. Commenters also 
inquired about the use of the item ``walk 10 feet'' to derive the 
weights for the ``walk 50 feet'' and ``walk 150 feet'' items.
    Response: We appreciate the commenter's concerns regarding the 
removal of item GG0170A1 from the motor score. As described in detail 
in the technical report, ``Analyses to Inform the Use of Standardized 
Patient Assessment Data Elements in the Inpatient Rehabilitation 
Facility Prospective Payment System,'' the roll left and right item was 
found to have a high degree of multicollinearity with other 
standardized patient assessment elements and to be inversely correlated 
with costs after controlling for each of the other self-care and 
mobility items. This relationship persisted when this item was paired 
with the other correlated items. The continued inclusion of this item 
in the motor score would have resulted in either a negative or non-
significant coefficient. As such, we do not believe it is appropriate 
to include this item in the construction of the motor score. The other 
item pairs that were found to be correlated did not generate negative 
or non-significant coefficients, and were therefore maintained in the 
calculation of the motor score.
    Unlike the FIMTM instrument, the items from the quality 
indicator section of the IRF-PAI sometimes use more than one item to 
measure functional areas. As discussed in more detail in the technical 
report, we noted that a few items were found to be highly correlated. 
Because of the correlation, we proposed to use an average score for 
some items so as to avoid introducing bias or inappropriately 
overweighting any particular functional area. We note this methodology 
is consistent with the methodology used under the Patient Driven 
Payment Model (PDPM), as described in more detail in the FY 2019 SNF 
final rule (83 FR 39204) and the accompanying technical report entitled 
``Skilled Nursing Facilities Patient-Driven Payment Model Technical 
Report'' available on the CMS website at https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/SNFPPS/therapyresearch.html.
    Regarding the ``walk 10 feet'' item, that item was used to derive 
the weights for the ``walk 50 feet'' and ``walk 150 feet'' items as 
these three items were found to be highly correlated and the ``walk 150 
feet'' item had a high proportion of observations coded on admission 
with ``activity not attempted'' codes.
    Comment: Some commenters requested that CMS apply the current motor 
score weights associated with the FIMTM items to the revised 
motor score while other commenters requested that CMS postpone 
weighting the motor score until additional data can be collected and 
analyzed. While a few commenters were supportive of using a weighted 
motor score, other commenters suggested that CMS use a 1-year payment 
model or phase in the use of a weighted motor score.
    Response: We do not believe it would be appropriate to apply the 
weight values associated with the FIMTM items to the 
components of the revised motor score, as these weights would not 
accurately reflect how the various components of the revised motor 
score contribute to predicting patient costs. We used simple ordinary 
least squares regression analysis of the data that IRFs submitted to us 
in FYs 2017 and 2018 to calculate the proposed weight values for the 
revised motor score. Changes in patient demographics, treatment 
practices, technology, and other factors that may affect the relative 
use of resources in an IRF since the motor score weights were 
originally calculated have likely contributed to changes in the weight 
values applied across the self-care and mobility items. We proposed to 
apply weights to the motor score items because RTI's analysis indicated 
that a weighted motor score would improve the classification of 
patients into CMGs, which in turn would improve the accuracy of 
payments to IRFs. However, as discussed above, in response to public 
comments, we carefully considered whether to finalize the proposed 
weighted motor score or go back to using an unweighted motor score to 
assign patients to CMGs. Although the proposed weighted motor score 
results in a slight improvement in the ability of the IRF PPS to 
predict patient costs and thus the accuracy of IRF PPS payments (less 
than 0.18 difference in accuracy between the weighted and the 
unweighted motor scores), we acknowledge the unweighted motor score is 
conceptually simpler and, as such, believe it will ease providers' 
transition to the use of the data items located in the Quality 
Indicators section of the IRF-PAI (also referred to as section GG 
items). Thus, we are finalizing based on public comments the use of an 
unweighted motor score, in which each of the 18 items have a weight of 
1, to assign patients to CMGs beginning with FY 2020.
    Comment: Commenters expressed concern that the analysis performed 
by RTI did not explicitly follow the analysis conducted by RAND when 
the motor score weights were developed for FY 2006 (70 FR 47896 through 
47900) and that RTI based their analyses on 2 years of data instead of 
several years of data. Additionally, commenters requested more 
information on the other weighting methodologies that RTI considered.
    Response: We disagree with the commenters that the RAND analysis 
for FY 2006 used more years of data than RTI's analysis for the FY 2020 
proposed rule. As discussed in the FY 2006 IRF PPS final rule (70 FR 
47897), RAND performed regression analysis on less than 2 full years of 
data (calendar year (CY) 2002 and FY 2003) to derive the current motor 
score weights. In contrast, RTI used 2 full years of data (FYs 2017 and 
2018) to perform the analysis for the weighted motor score proposed in 
the FY 2020 IRF PPS proposed rule. As the FYs 2017 and 2018 data 
portrays the most recent and complete picture of patients under the IRF 
PPS, we believe it was sufficient and appropriate to utilize for the 
analysis for the proposed rule.
    While RTI utilized a different weighting methodology than was used 
by RAND in 2006, the overall model

[[Page 39063]]

prediction using the weighted motor score developed by RAND and the 
weighted motor score developed by RTI is extremely similar. The model 
using the CMGs based on the standardized patient assessment data 
elements and comorbidity tiers to predict wage-adjusted costs of care 
has an r-squared value is 0.3358, while the r-squared value is 0.3169 
for the CMGs in the current IRF PPS. This is indicative of similar 
model performance regardless of model specification. The item weights 
that the RAND work notes as ``optimally weighted'' are weights that 
were constructed separately for each RIC. These were not the weights 
that were used in the final weights developed by RAND.
    RTI also examined weighing methodologies utilizing a general linear 
model (GLM) and log transformed ordinary least squares (OLS) regression 
models, as well as the OLS model described in more detail in the 
technical report. All three models had comparable model fit and 
generated similar item weights. Based on the greater simplicity 
achieved through the use of the OLS regression model we believe using 
the OLS regression was appropriate to maintain simplicity and 
transparency in the payment system.
    Comment: Commenters disagreed with the omission of the wheelchair 
mobility items from the items used to construct the motor score.
    Response: We appreciate the commenters' concerns about wheelchair-
dependent patients. As most recently discussed in the FY 2019 IRF PPS 
final rule (83 FR 38546) in response to similar stakeholder comments, 
we explained our rationale for not including the wheelchair mobility 
items in the construction of the finalized motor score. We continue to 
believe that the higher resource needs of wheelchair dependent patients 
in IRFs will be better accounted for by not including a wheelchair item 
in the motor score at this time. Patients that are considered 
wheelchair dependent or unable to walk will be accounted for through 
the ``not attempted'' response codes captured through other items, 
especially some of the walking items, that are included in the motor 
score. In this way, we ensure that IRFs will be appropriately 
compensated for the higher costs they incur in treating wheelchair-
dependent patients. We refer readers to the FY 2019 IRF PPS final rule 
(83 FR 38546) and the technical report entitled ``Analyses to Inform 
the Use of Standardized Patient Assessment Data Elements in the 
Inpatient Rehabilitation Facility Prospective Payment System'' for more 
information on the rationale as to why this item was not included in 
the calculation of the motor score.
    Comment: Commenters expressed concern with the weighted motor score 
and questioned the reliability and validity of the weighted motor 
score. Some commenters stated that they believe the weighted and 
unweighted motor scores have shown little to no correlation with the 
weighted motor score currently in use, and therefore, questioned if the 
weighted motor score could accurately measure patient severity.
    Response: We disagree with the commenters' suggestion that 
unweighted and weighted motor scores have shown little to no 
correlation with the weighted motor score currently in use as our 
analysis shows a strong correlation between the scores. In addition, 
each of the proposed Quality Indicators data items that were included 
in the motor score were found to have statistically significant 
correlation with IRF costs. As discussed in the technical report 
``Analyses to Inform the Use of Standardized Patient Assessment Data 
Elements in the Inpatient Rehabilitation Facility Prospective Payment 
System'' the use of a weighted motor score was found to increase the 
predictive ability of the payment model.
    Comment: Commenters requested that CMS make available the data 
utilized in the analyses including patient assessment data, matching 
claims data, and additional facility and cost report data to enable 
stakeholders to replicate the analyses.
    Response: We appreciate the commenters' feedback regarding the 
types of information that would be most useful to them in replicating 
our analyses. We are unable to make patient assessment and claims data 
publicly available on the CMS website because these data contain 
personally identifiable information. However, we believe that we 
released sufficient information in the proposed rule, the accompanying 
data files, and the technical report entitled ``Analyses to Inform the 
Use of Standardized Patient Assessment Data Elements in the Inpatient 
Rehabilitation Facility Prospective Payment System,'' to enable 
stakeholders to submit meaningful comments on the underlying analyses 
and methodologies used to revise the IRF case-mix classification 
system, to pose alternative approaches, and to assess the impacts of 
the proposed revisions.
    Comment: A few commenters noted that they did not believe that CMS 
has performed the thorough data analyses and engagement with the 
provider community that are necessary prior to making significant 
changes to the existing IRF PPS. These commenters requested that we 
solicit additional feedback from the stakeholder community, including 
convening technical advisory panels (TEPs), to provide additional 
transparency into the underlying analyses and to delay implementation 
of a weighted motor score until we conduct additional engagements with 
stakeholders.
    Response: We value transparency in our processes and will continue 
to engage stakeholders in future development of payment policies. We 
appreciate the offers from stakeholders to assist in the development of 
future revisions to payment policies and we recognize the value from 
these partnerships. However, for something as analytically simple as 
running a regression analysis to determine the weights for the motor 
score items that best reflect patients' resource needs in the IRF, we 
do not believe that a TEP is necessary.
    As noted above, although the proposed weighted motor score results 
in a slight improvement in the ability of the IRF PPS to predict 
patient costs and thus the accuracy of IRF PPS payments (less than 0.18 
difference in accuracy between the weighted and the unweighted motor 
scores), we acknowledge the unweighted motor score is conceptually 
simpler and, as such, believe it will ease providers' transition to the 
use of the data items located in the Quality Indicators section of the 
IRF-PAI (also referred to as section GG items). Thus, we are finalizing 
based on public comments the use of an unweighted motor score to assign 
patients to CMGs beginning with FY 2020. We appreciate the 
stakeholders' comments on this topic and will take them into 
consideration for future analysis.
    Comment: A few commenters requested that CMS provide additional 
information regarding the provider specific impact analysis file that 
accompanied the rule, such as a data dictionary describing the data 
used to calculate the impacts.
    Response: In conjunction with the release of the FY 2020 IRF PPS 
proposed rule, we posted a provider-specific impact analysis file that 
compared estimated payments to providers for FY 2020 without the 
proposed revisions to the CMGs with estimated payments to providers for 
FY 2020 with the proposed revisions to the CMGs. We believe that this 
file gives IRFs added information to enable them to see how their 
individual payments would be affected by the proposed changes to the 
CMGs. We updated this

[[Page 39064]]

provider specific impact analysis file shortly after it was initially 
posted to include additional information regarding the underlying data 
used to calculate the provider specific impacts, and we believe that 
this additional information is responsive to commenters' requests. The 
file can be downloaded from the CMS website at https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/InpatientRehabFacPPS/IRF-Rules-and-Related-Files.html. We appreciate the commenters' suggestions 
regarding the additional types of information that would be most useful 
to them to further facilitate understanding of our analyses.
    As previously discussed, we proposed a weighted motor score as it 
was found to slightly improve the predicative ability of the case-mix 
system and thus the accuracy of IRF PPS payments. However, nearly all 
of the comments we received requested that we revert to an unweighted 
motor score for the various reasons discussed above. While we continue 
to believe that a weighted motor score is slightly more accurate, the 
difference is small, and in light of the conceptual simplicity achieved 
through the use of an unweighted motor score, which we believe will 
ease providers' transition to the use of the data items located in the 
Quality Indicators section of the IRF-PAI, we are finalizing the use of 
an unweighted motor score, in which each of the 18 items used in the 
score have an equal weight of 1, to assign patients to CMGs beginning 
with FY 2020. Additionally, we are finalizing the proposed removal of 
one item (GG0170A1 Roll left to right) from the motor score beginning 
with FY 2020. Effective for all discharges beginning on or after 
October 1, 2019, we will use an unweighted motor score as indicated in 
Table 2 to determine a beneficiary's CMG placement. 
[GRAPHIC] [TIFF OMITTED] TR08AU19.002

C. Revisions to the CMGs and Updates to the CMG Relative Weights and 
Average Length of Stay Values Beginning With FY 2020

    In the FY 2019 IRF PPS final rule (83 FR 38549), we finalized the 
use of data items from the Quality Indicators section of the IRF-PAI to 
construct the functional status scores used to classify IRF patients in 
the IRF case-mix classification system for purposes of establishing 
payment under the IRF PPS beginning with FY 2020, but modified our 
proposal based on public comments to incorporate 2 years of data (FYs 
2017 and 2018) into our analyses used to revise the CMG definitions. We 
stated that any changes to the proposed CMG definitions resulting from 
the incorporation of an additional year of data (FY 2018) into the 
analysis would be addressed in future rulemaking prior to their 
implementation beginning in FY 2020. Additionally, we stated that we 
would also update the relative weights and average LOS values 
associated with any revised CMG definitions in future rulemaking.
    As noted in the FY 2020 IRF PPS proposed rule (84 FR 17251), we 
continued our contract with RTI to support us in developing proposed 
revisions to the CMGs used under the IRF PPS based on analysis of 2 
years of data (FYs 2017 and 2018). The process RTI uses for its 
analysis, which is based on a Classification and Regression Tree (CART) 
algorithm, is described in detail in the FY 2019 IRF PPS final rule (83 
FR 38536 through 38540). RTI used this analysis to revise the CMGs 
utilizing FYs 2017 and 2018 claim and assessment data and to develop 
revised CMGs that reflect the use of the data items collected in the 
Quality Indicators section of the IRF-PAI, incorporating the proposed 
weighted motor score described in the FY 2020 IRF PPS proposed rule. 
However, as discussed in section IV.B of this final rule, we are 
finalizing based on public comments the use of an unweighted motor 
score to assign patients to a CMGs beginning in with FY 2020.
    To develop the proposed revised CMGs, RTI used CART analysis to 
divide patients into payment groups based on similarities in their 
clinical characteristics and relative costs. As part of this analysis, 
RTI imposed some typically-used constraints on the payment group 
divisions (for example, on the minimum number of cases that could be in 
the resulting payment groups and the minimum dollar payment amount 
differences between groups) to identify the optimal set of payment 
groups. For a more detailed discussion of the analysis used to revise

[[Page 39065]]

the CMGs for FY 2020, we refer readers to the March 2019 technical 
report entitled, ``Analyses to Inform the Use of Standardized Patient 
Assessment Data Elements in the Inpatient Rehabilitation Facility 
Prospective Payment System'' available at https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/InpatientRehabFacPPS/Research.html. 
Additionally, we refer readers to the FY 2020 IRF PPS proposed rule (84 
FR 17250 through 17260) for more information on the proposed revisions 
to the CMGs.
    As noted above, we are finalizing the use of an unweighted motor 
score beginning with FY 2020. As the motor score is a key input in the 
CART analysis used to revise the CMGs, the use of the unweighted motor 
score required that the CART analysis be rerun utilizing the unweighted 
motor score. RTI utilized the same methodology described in the FY 2020 
IRF PPS proposed rule (84 FR 17250 through 17260) to support us in 
developing revisions to the CMGs, incorporating the unweighted motor 
score, as described in section IV.B of this final rule. The revised 
CMGs can be found in Table 3.
    After developing the revised CMGs, RTI then calculated the relative 
weights and average LOS values for each revised CMG using the same 
methodologies that we have used to update the CMG relative weights and 
average LOS values each fiscal year since 2009 (when we implemented an 
update to this methodology). More information about the methodology 
used to update the CMG relative weights can be found in the FY 2009 IRF 
PPS final rule (73 FR 46372 through 46374). For FY 2020, we proposed to 
use the FYs 2017 and 2018 IRF claims and FY 2017 IRF cost report data 
to update the CMG relative weights and average LOS values. In 
calculating the CMG relative weights, we use a hospital-specific 
relative value method to estimate operating (routine and ancillary 
services) and capital costs of IRFs. As noted in the FY 2019 IRF PPS 
final rule (83 FR 38521), this is the same methodology that we have 
used to update the CMG relative weights and average LOS values each 
fiscal year since we implemented an update to the methodology in the FY 
2009 IRF PPS final rule (73 FR 46372 through 46374). More information 
on the methodology used to update calculate the CMG relative weights 
and average LOS values can found in the March 2019 technical report 
entitled ``Analyses to Inform the Use of Standardized Patient 
Assessment Data Elements in the Inpatient Rehabilitation Facility 
Prospective Payment System'' available at https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/InpatientRehabFacPPS/Research.html. 
Consistent with the methodology that we have used to update the IRF 
classification system in each instance in the past, we proposed to 
update the relative weights associated with the revised CMGs for FY 
2020 in a budget neutral manner by applying a budget neutrality factor 
to the standard payment amount. To calculate the appropriate budget 
neutrality factor for use in updating the FY 2020 CMG relative weights, 
we used the following steps:
    Step 1. Calculate the estimated total amount of IRF PPS payments 
for FY 2020 (with no changes to the CMG relative weights).
    Step 2. Calculate the estimated total amount of IRF PPS payments 
for FY 2020 by applying the changes to the CMGs and the associated CMG 
relative weights (as described in this final rule).
    Step 3. Divide the amount calculated in step 1 by the amount 
calculated in step 2 to determine the budget neutrality factor (1.0016) 
that would maintain the same total estimated aggregate payments in FY 
2020 with and without the changes to the CMGs and the associated CMG 
relative weights.
    Step 4. Apply the budget neutrality factor (1.0016) to the FY 2019 
IRF PPS standard payment amount after the application of the budget-
neutral wage adjustment factor.
    We note that, as we typically do, we updated our data between the 
FY 2020 IRF PPS proposed and final rules to ensure that we use the most 
recent available data in calculating IRF PPS payments. Additionally, we 
are finalizing the use of unweighted motor score beginning in with FY 
2020 which generated revisions to the CMGs and relative weights. Based 
on our analysis using this updated data and an unweighted motor score, 
we now estimate a budget neutrality factor of (1.0010) to maintain the 
same total estimated aggregate payments in FY 2020 with and without the 
changes to the CMGs and the associated CMG relative weights. For FY 
2020 we will apply the budget neutrality factor (1.0010) to the FY 2019 
IRF PPS standard payment amount after the application of the budget-
neutral wage adjustment factor.
    The relative weights and average LOS values for those revised CMGs 
(found in Table 3) were calculated using the same methodology described 
in the FY 2020 IRF PPS proposed rule, which is the same methodology 
that we have used to update the CMG relative weights and average LOS 
values each fiscal year since we implemented an update to the 
methodology in FY 2009. The revised CMGs (reflecting the unweighted 
motor score) and their respective descriptions, as well as the 
comorbidity tiers, corresponding relative weights and the average LOS 
values for each CMG and tier for FY 2020 are shown in Table 3. The 
average LOS for each CMG is used to determine when an IRF discharge 
meets the definition of a short-stay transfer, which results in a per 
diem case level adjustment. In section V.H. of this final rule, we 
discuss the proposed use of the existing methodology to calculate the 
standard payment conversion factor for FY 2020.
    We received a number of comments on the proposed revisions to the 
CMGs based on analysis of 2 years of data (FYs 2017 and 2018) and the 
proposed updates to the relative weights and average LOS values 
associated with the revised CMGs beginning with FY 2020, that is, for 
all discharges beginning on or after October 1, 2019, which are 
summarized below.
    Comment: A number of commenters were appreciative of the use of 2 
years of data to revise the CMGs; however, commenters expressed concern 
with the proposed CMG revisions and suggested that these changes could 
result in payment rate compression or a misalignment between payments 
and the costs of caring for patients. Commenters suggested payment 
compression would result in reduced payments for higher acuity patients 
and increased payments for lower acuity patients which could compromise 
access to care for patients with certain impairments. Additionally, 
some commenters questioned why there would be fewer CMGs within some 
RICs and suggested having fewer CMGs would also contribute to payment 
rate compression.
    Response: We disagree with the commenters that revisions to CMGs 
will lead to payment rate compression or could compromise access to 
care for any particular group of patients. As the revised CMGs are 
reflective of the data that IRFs submitted to us in FYs 2017 and 2018, 
we believe the revised CMGs reflect the distinct resource needs of the 
current Medicare IRF population. We believe the revised CMGs more 
accurately predict resource use in IRFs and better align payments with 
the expected costs of treating patients in the IRF setting. As such, we 
believe that the revised CMGs may in fact improve access to and quality 
of care for IRF patients by increasing the accuracy of IRF payments to 
providers.
    Regarding why some RICs would have fewer CMGs, we refer the 
commenters to the Technical Report entitled ``Analyses

[[Page 39066]]

to Inform the Use of Standardized Patient Assessment Data Elements in 
the Inpatient Rehabilitation Facility Prospective Payment System'' that 
describes in detail the analysis used to derive the CMGs and the 
criteria required to generate additional payment groups. As noted in 
the FY 2020 IRF PPS proposed rule (84 FR 17250 through 17252), RTI 
imposed some typically-used constraints in their analysis to identify 
the proposed set of payment groups. These constraints consisted of a 
minimum number of stays within a node, a 0.5 percentage point increase 
of explanatory power, and monotonicity across the CMGs within each RIC. 
We do not believe it would be appropriate to generate additional CMGs 
that did not improve the predicative ability of the model beyond what 
was produced through the CART analysis utilizing the constraints above. 
We note that while the CART analysis generated fewer CMGs within some 
RICs, it generated a greater number of CMGs within other RICs and that 
the overall number of CMGs increases through these revisions to the 
case-mix classification system. We do not believe having fewer CMGs 
within any RIC will contribute to payment rate compression as we 
believe these revisions better align payments with the expected costs 
of treating patients in IRFs.
    Additionally, we disagree with the commenters' statements that the 
CMG revisions will result in higher payments for lower acuity patients 
and reduced payments for higher acuity patients. Our analysis has found 
that higher function is associated with a slight reduction in payment 
under the revised CMGs and that lower function is associated with a 
slight increase in payments. The purpose of the proposed revisions to 
the CMGs is to align payments more appropriately with the costs of 
caring for all types of patients in IRFs. As such, we do not believe 
that the revisions will result in higher payments for lower acuity 
patients. We appreciate the commenters' concerns and will continue to 
monitor the IRF data closely to ensure that IRF payments are 
appropriately aligned with costs of care and that Medicare patients 
continue to have appropriate access to IRF services.
    Comment: Several commenters expressed concerns that the proposed 
CMG revisions could cause a significant redistribution of payments 
among IRF provides. These commenters indicated that they believe the 
section GG items make patients appear to be less severe and requested 
additional information on how patients would be redistributed among the 
revised CMGs. Additionally, commenters encouraged CMS to monitor the 
data based on these changes and to update the model if necessary in the 
future.
    Response: We agree with the commenters that the revisions to the 
CMGs may result in some redistribution of payments among providers. As 
noted in the FY 2019 IRF PPS final rule (83 FR 38547), the scales and 
coding instructions are slightly different between the item sets used 
to derive the existing CMGs and those used to derive the revised CMGs. 
As such, these differences may result in some patients grouping into 
different CMGs that more accurately account for the expected resource 
needs of the patient. While we cannot make individual Medicare 
beneficiary data publically available, we believe we released adequate 
information for stakeholders to determine how beneficiaries could be 
distributed across the revised CMGs. We appreciate the commenters' 
suggestions to conduct monitoring activities and make future updates to 
the case-mix classification system and will take this into 
consideration in the future.
    Comment: Commenters expressed concern with the use of section GG 
items to assign a patient to a CMG and suggested that these items are 
not sensitive enough and do not capture patients' true burden of care. 
Commenters also expressed concern with the reliability of the data 
collected through these items and suggested that the data is not 
accurate or valid.
    Response: As discussed in detail in the FY 2019 IRF PPS final rule 
(83 FR 38541), we believe that the data items located in the Quality 
Indicators section of the IRF-PAI are sensitive and accurately capture 
the functional and cognitive status of patients and can also be used to 
accurately assess changes in patients' functional status. As noted 
above, RTI found that the model predicting costs using the CMGs derived 
from the items located in the Quality Indicators section of the IRF-PAI 
had a slightly higher R-squared value than models using the current 
CMGs which are derived from items in the FIMTM instrument, 
indicating that the revised CMGs more accurately predict resource use 
in IRFs than the CMGs that are currently utilized. As the data 
collected in the Quality Indicators section of the IRF-PAI have been 
collected nationally for all IRFs since October 1, 2016, we believe the 
data to be accurate and valid at this time. We also believe it is the 
responsibility of the IRF to submit accurate and valid data that 
adheres to the coding guidelines detailed in the IRF-PAI training 
manual.
    Comment: Commenters expressed concern with the cognition items 
collected on the IRF-PAI and their omission from the revised CMGs. A 
few commenters noted the importance of cognitive impairment in the IRF 
setting and encouraged CMS to conduct further analysis of the 
relationship between cognitive function and resource use in the IRF 
setting and to improve the items that are used to measure cognitive 
function.
    Response: We appreciate the commenters' concerns with the cognitive 
items that are collected on the IRF-PAI. As we discussed in the FY 2019 
IRF PPS final rule (83 FR 38546), the cognitive items that we used for 
this analysis are the best ones that we have for use at the present 
time. Unfortunately, we found that including these cognitive items in 
generating the CMGs would have resulted in lower payments for patients 
with higher cognitive deficits. This result does not make sense from a 
clinical perspective, and could have the unintended consequence of 
reducing access to IRF care for more cognitively impaired 
beneficiaries. Thus, we determined that it would be better at this time 
to remove the CMG splits that were generated by the cognitive items. We 
appreciate the commenters' suggestion to incorporate improved cognition 
measures into the IRF-PAI and will take this into consideration in the 
future.
    Comment: Commenters suggested that CMS has not provided sufficient 
education, training materials, or supporting documentation regarding 
the functional items to support their use in developing a payment 
model. Some commenters suggested revisions to the existing training 
materials while other commenters requested that CMS provide additional 
training, monitor the data, and modify the case mix groupings as 
needed.
    Response: We disagree with the commenters that we have provided 
insufficient training or guidance on proper coding of this data. We 
believe we have provided adequate training opportunities for IRFs on 
coding the Quality Indicator data items, including multiple in-person 
training opportunities, webinars, on-line training and on-going help 
desk guidance. We are committed to providing information and support 
that will allow providers to accurately interpret and complete quality 
reporting items and we will continue to provide these types of 
opportunities to the IRF community. We thank the commenters for their 
suggestions to improve the training materials and we appreciate the 
commenters' suggestions to continue to monitor the data and make 
updates to

[[Page 39067]]

the case-mix classification system when necessary.
    After careful consideration of the comments received, we are 
finalizing revisions to the CMGs based on analysis of 2 years of data 
(FYs 2017 and 2018) and the incorporation of the unweighted motor score 
described in section IV.B of this final rule. The revised CMGs that 
will be effective October 1, 2019 are presented below in Table 3. We 
refer readers to Table 20 in section XIII.C of this final rule for more 
information on the distributional effects of revisions to the CMGs. For 
a provider specific impact analysis for this change, we refer readers 
to the CMS website at https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/InpatientRehabFacPPS/IRF-Rules-and-Related-Files.html. 
We are also updating the relative weights and average LOS values 
associated with the revised CMGs (reflecting an unweighted motor score) 
beginning with FY 2020.
BILLING CODE 4120-01-P

[[Page 39068]]

[GRAPHIC] [TIFF OMITTED] TR08AU19.003


[[Page 39069]]


[GRAPHIC] [TIFF OMITTED] TR08AU19.004


[[Page 39070]]


[GRAPHIC] [TIFF OMITTED] TR08AU19.005


[[Page 39071]]


[GRAPHIC] [TIFF OMITTED] TR08AU19.006

BILLING CODE 4120-01-C

V. Facility-Level Adjustment Factors

    Section 1886(j)(3)(A)(v) of the Act confers broad authority upon 
the Secretary to adjust the per unit payment rate by such factors as 
the Secretary determines are necessary to properly reflect variations 
in necessary costs of treatment among rehabilitation facilities. Under 
this authority, we currently adjust the prospective payment amount 
associated with a CMG to account for facility-level characteristics 
such as an IRF's LIP, teaching status, and location in a rural area, if 
applicable, as described in Sec.  412.624(e).
    Based on the substantive changes to the facility-level adjustment 
factors that were adopted in the FY 2014 IRF PPS final rule (78 FR 
47860, 47868 through 47872), in the FY 2015 IRF PPS final rule (79 FR 
45872, 45882 through 45883), we froze the facility-level adjustment 
factors at the FY 2014 levels for FY 2015 and all subsequent years 
(unless and until we propose to update them again through future 
notice-and-comment rulemaking). For FY 2020, we will continue to hold 
the adjustment factors at the FY 2014 levels as we continue to monitor 
the most current IRF claims data available and continue to evaluate and 
monitor the effects of the FY 2014 changes.

VI. FY 2020 IRF PPS Payment Update

A. Background

    Section 1886(j)(3)(C) of the Act requires the Secretary to 
establish an increase factor that reflects changes over time in the 
prices of an appropriate mix of goods and services included in the 
covered IRF services. According to section 1886(j)(3)(A)(i) of the Act, 
the increase factor shall be used to update the IRF prospective payment 
rates for each FY. Section 1886(j)(3)(C)(ii)(I) of the Act requires the 
application of a productivity adjustment. Thus, in the FY 2020 IRF 
proposed rule, we proposed to update the IRF PPS payments for FY 2020 
by a market basket increase factor as required by section 1886(j)(3)(C) 
of the Act based upon the most current data available, with a 
productivity adjustment as required by section 1886(j)(3)(C)(ii)(I) of 
the Act (84 FR 17261).
    We have utilized various market baskets through the years in the 
IRF PPS. For a discussion of these market baskets, we refer readers to 
the FY 2016 IRF PPS final rule (80 FR 47046).
    Beginning with FY 2016, we finalized the use of a 2012-based IRF 
market basket, using Medicare cost report (MCR) data for both 
freestanding and hospital-based IRFs (80 FR 47049 through 47068). 
Beginning with FY 2020, we proposed to rebase and revise the IRF market 
basket to reflect a 2016 base year. In the following discussion, we 
provide an overview of the proposed market basket and describe the 
methodologies used to determine the operating and capital portions of 
the proposed 2016-based IRF market basket.

B. Overview of the 2016-Based IRF Market Basket

    The 2016-based IRF market basket is a fixed-weight, Laspeyres-type 
price index. A Laspeyres price index measures the change in price, over 
time, of the same mix of goods and services purchased in the base 
period. Any changes in the quantity or mix of goods and services (that 
is, intensity) purchased over time relative to a base period are not 
measured.
    The index itself is constructed in three steps. First, a base 
period is selected (for the proposed IRF market basket, the base period 
is 2016), total base period costs are estimated for a set of mutually 
exclusive and exhaustive cost categories, and each category is 
calculated as a proportion of total costs. These proportions are called 
cost weights. Second, each cost category is matched to an appropriate 
price or wage variable, referred to as a price proxy. In nearly every 
instance where we have selected price proxies for the various market 
baskets, these price proxies are derived from publicly available 
statistical series that are published on a consistent schedule 
(preferably at least on a quarterly basis). In cases where a publicly 
available price series is not available (for example, a price index for 
malpractice insurance), we have collected price data from other sources 
and subsequently developed our own index to capture changes in prices 
for these types of costs. Finally, the cost weight for each cost 
category is multiplied by the established price proxy. The sum of these 
products (that is, the cost weights multiplied by their price levels) 
for all cost categories yields the composite index level of the market 
basket for the given time period. Repeating this step for other periods 
produces a series of market basket levels over time. Dividing the 
composite index level of one period by the composite index level for an 
earlier period produces a rate of growth in the input price index over 
that timeframe.

[[Page 39072]]

    As previously noted, the market basket is described as a fixed-
weight index because it represents the change in price over time of a 
constant mix (quantity and intensity) of goods and services needed to 
furnish IRF services. The effects on total costs resulting from changes 
in the mix of goods and services purchased after the base period are 
not measured. For example, an IRF hiring more nurses after the base 
period to accommodate the needs of patients would increase the volume 
of goods and services purchased by the IRF, but would not be factored 
into the price change measured by a fixed-weight IRF market basket. 
Only when the index is rebased would changes in the quantity and 
intensity be captured, with those changes being reflected in the cost 
weights. Therefore, we rebase the market basket periodically so that 
the cost weights reflect recent changes in the mix of goods and 
services that IRFs purchase to furnish inpatient care between base 
periods.

C. Rebasing and Revising of the IRF PPS Market Basket

    As discussed in the FY 2016 IRF PPS final rule (80 FR 47050), the 
2012-based IRF market basket reflects the Medicare cost reports for 
both freestanding and hospital-based facilities.
    Beginning with FY 2020, we proposed to rebase and revise the 2012-
based IRF market basket to a 2016 base year reflecting both 
freestanding and hospital-based IRFs. Below we provide a detailed 
description of our methodology used to develop the proposed 2016-based 
IRF market basket. This proposed methodology is generally similar to 
the methodology used to develop the 2012-based IRF market basket with 
the exception of the proposed derivation of the Home Office Contract 
Labor cost weight using the MCR data as described in section VI.C.a.(6) 
of this final rule.
1. Development of Cost Categories and Weights for the 2016-Based IRF 
Market Basket
a. Use of Medicare Cost Report Data
    We proposed a 2016-based IRF market basket that consists of seven 
major cost categories and a residual derived from the 2016 Medicare 
cost reports (CMS Form 2552-10) for freestanding and hospital-based 
IRFs. The seven cost categories are Wages and Salaries, Employee 
Benefits, Contract Labor, Pharmaceuticals, Professional Liability 
Insurance (PLI), Home Office Contract Labor, and Capital. The residual 
category reflects all remaining costs not captured in the seven cost 
categories. The 2016 cost reports include providers whose cost 
reporting period began on or after October 1, 2015, and prior to 
September 30, 2016. We selected 2016 as the base year because we 
believe that the Medicare cost reports for this year represent the most 
recent, complete set of MCR data available for developing the IRF 
market basket at the time of the proposed rule.
    Since our goal is to establish cost weights that were reflective of 
case mix and practice patterns associated with the services IRFs 
provide to Medicare beneficiaries, as we did for the 2012-based IRF 
market basket, we proposed to limit the cost reports used to establish 
the 2016-based IRF market basket to those from facilities that had a 
Medicare average LOS that was relatively similar to their facility 
average LOS. We believe that this requirement eliminates statistical 
outliers and ensures a more accurate market basket that reflects the 
costs generally incurred during a Medicare-covered stay. The Medicare 
average LOS for freestanding IRFs is calculated from data reported on 
line 14 of Worksheet S-3, part I. The Medicare average LOS for 
hospital-based IRFs is calculated from data reported on line 17 of 
Worksheet S-3, part I. We proposed to include the cost report data from 
IRFs with a Medicare average LOS within 15 percent (that is, 15 percent 
higher or lower) of the facility average LOS to establish the sample of 
providers used to estimate the 2016-based IRF market basket cost 
weights. We proposed to apply this LOS edit to the data for IRFs to 
exclude providers that serve a population whose LOS would indicate that 
the patients served are not consistent with a LOS of a typical Medicare 
patient. We note that this is the same LOS edit that we applied to 
develop the 2012-based IRF market basket. This process resulted in the 
exclusion of about eight percent of the freestanding and hospital-based 
IRF Medicare cost reports. Of those excluded, about 18 percent were 
freestanding IRFs and 82 percent were hospital-based IRFs. This ratio 
is relatively consistent with the ratio of the universe of freestanding 
to hospital-based IRF providers.
    We then used the cost reports for IRFs that met this requirement to 
calculate the costs for the seven major cost categories (Wages and 
Salaries, Employee Benefits, Contract Labor, Professional Liability 
Insurance, Pharmaceuticals, Home Office Contract Labor, and Capital) 
for the market basket. For comparison, the 2012-based IRF market basket 
utilized the Bureau of Economic Analysis Benchmark Input-Output data 
rather than MCR data to derive the Home Office Contract Labor cost 
weight. A more detailed discussion of this methodological change is 
provided in section VI.C.1.a.(6). of this final rule.
    Similar to the 2012-based IRF market basket major cost weights, the 
proposed 2016-based IRF market basket cost weights reflect Medicare 
allowable costs (routine, ancillary and capital)--costs that are 
eligible for reimbursement through the IRF PPS.
    For freestanding IRFs, total Medicare allowable costs would be 
equal to the total costs as reported on Worksheet B, part I, column 26, 
lines 30 through 35, 50 through 76 (excluding 52 and 75), 90 through 
91, and 93. For hospital-based IRFs, total Medicare allowable costs 
would be equal to the total costs for the IRF inpatient unit after the 
allocation of overhead costs (Worksheet B, part I, column 26, line 41) 
and a proportion of total ancillary costs reported on Worksheet B, part 
I, column 26, lines 50 through 76 (excluding 52 and 75), 90 through 91, 
and 93. We proposed to calculate the portion of ancillary costs 
attributable to the hospital-based IRF for a given ancillary cost 
center by multiplying total facility ancillary costs for the specific 
cost center (as reported on Worksheet B, part I, column 26) by the 
ratio of IRF Medicare ancillary costs for the cost center (as reported 
on Worksheet D-3, column 3 for hospital-based IRFs) to total Medicare 
ancillary costs for the cost center (equal to the sum of Worksheet D-3, 
column 3 for all relevant PPS [that is, IPPS, IRF, IPF and skilled 
nursing facility (SNF)]). We proposed to use these methods to derive 
levels of total costs for IRF providers. This is the same methodology 
used for the 2012-based IRF market basket. With this work complete, we 
then set about deriving cost levels for the seven major cost categories 
and then derive a residual cost weight reflecting all other costs not 
classified.
(1) Wages and Salaries Costs
    For freestanding IRFs, we proposed to derive Wages and Salaries 
costs as the sum of routine inpatient salaries, ancillary salaries, and 
a proportion of overhead (or general service cost centers in the 
Medicare cost reports) salaries as reported on Worksheet A, column 1. 
Since overhead salary costs are attributable to the entire IRF, we only 
include the proportion attributable to the Medicare allowable cost 
centers. We proposed to estimate the proportion of overhead salaries 
that are attributed to Medicare allowable costs centers by multiplying 
the ratio of Medicare allowable area salaries (Worksheet A, column 1, 
lines 50 through 76

[[Page 39073]]

(excluding 52 and 75), 90 through 91, and 93) to total salaries 
(Worksheet A, column 1, line 200) times total overhead salaries 
(Worksheet A, column 1, lines 4 through 18). This is the same 
methodology used in the 2012-based IRF market basket.
    For hospital-based IRFs, we proposed to derive Wages and Salaries 
costs as the sum of inpatient routine salary costs (Worksheet A, column 
1, line 41) for the hospital-based IRF and the overhead salary costs 
attributable to this IRF inpatient unit; and ancillary salaries plus a 
portion of overhead salary costs attributable to the ancillary 
departments utilized by the hospital-based IRF.
    We proposed to calculate hospital-based ancillary salary costs for 
a specific cost center (Worksheet A, column 1, lines 50 through 76 
(excluding 52 and 75), 90 through 91, and 93) using salary costs from 
Worksheet A, column 1, multiplied by the ratio of IRF Medicare 
ancillary costs for the cost center (as reported on Worksheet D-3, 
column 3, for IRF subproviders) to total Medicare ancillary costs for 
the cost center (equal to the sum of Worksheet D-3, column 3, for all 
relevant PPS units [that is, IPPS, IRF, IPF and a SNF]). For example, 
if hospital-based IRF Medicare physical therapy costs represent 30 
percent of the total Medicare physical therapy costs for the entire 
facility, then 30 percent of total facility physical therapy salaries 
(as reported in Worksheet A, column 1, line 66) would be attributable 
to the hospital-based IRF. We believe it is appropriate to use only a 
portion of the ancillary costs in the market basket cost weight 
calculations since the hospital-based IRF only utilizes a portion of 
the facility's ancillary services. We believe the ratio of reported IRF 
Medicare costs to reported total Medicare costs provides a reasonable 
estimate of the ancillary services utilized, and costs incurred, by the 
hospital-based IRF.
    We proposed to calculate the portion of overhead salary costs 
attributable to hospital-based IRFs by first calculating total 
noncapital overhead costs (Worksheet B, part I, columns 4-18, line 41, 
less Worksheet B, part II, columns 4-18, line 41). We then multiply 
total noncapital overhead costs by an overhead ratio equal to the ratio 
of total facility overhead salaries (as reported on Worksheet A, column 
1, lines 4-18) to total facility noncapital overhead costs (as reported 
on Worksheet A, column 1 and 2, lines 4-18). This methodology assumes 
the proportion of total costs related to salaries for the overhead cost 
center is similar for all inpatient units (that is, acute inpatient or 
inpatient rehabilitation).
    We proposed to calculate the portion of overhead salaries 
attributable to each ancillary department by first calculating total 
noncapital overhead costs attributable to each specific ancillary 
department (Worksheet B, part I, columns 4-18 less, Worksheet B, part 
II, columns 4-18). We then identify the portion of these noncapital 
overhead costs attributable to Wages and Salaries by multiplying these 
costs by the overhead ratio defined as the ratio of total facility 
overhead salaries (as reported on Worksheet A, column 1, lines 4-18) to 
total overhead costs (as reported on Worksheet A, column 1 & 2, lines 
4-18). Finally, we identified the portion of these overhead salaries 
for each ancillary department that is attributable to the hospital-
based IRF by multiplying by the ratio of IRF Medicare ancillary costs 
for the cost center (as reported on Worksheet D-3, column 3, for 
hospital-based IRFs) to total Medicare ancillary costs for the cost 
center (equal to the sum of Worksheet D-3, column 3, for all relevant 
PPS units [that is, IPPS, IRF, IPF and SNF]). This is the same 
methodology used to derive the 2012-based IRF market basket.
(2) Employee Benefits Costs
    Effective with the implementation of CMS Form 2552-10, we began 
collecting Employee Benefits and Contract Labor data on Worksheet S-3, 
part V.
    For 2016 MCR data, the majority of providers did not report data on 
Worksheet S-3, part V; particularly, approximately 48 percent of 
freestanding IRFs and 40 percent of hospital-based IRFs reported data 
on Worksheet S-3, part V. However, we believe we have a large enough 
sample to enable us to produce a reasonable Employee Benefits cost 
weight. Again, we continue to encourage all providers to report these 
data on the Medicare cost report.
    For freestanding IRFs, we proposed Employee Benefits costs would be 
equal to the data reported on Worksheet S-3, part V, column 2, line 2. 
We note that while not required to do so, freestanding IRFs also may 
report Employee Benefits data on Worksheet S-3, part II, which is 
applicable to only IPPS providers. For those freestanding IRFs that 
report Worksheet S-3, part II, data, but not Worksheet S-3, part V, we 
proposed to use the sum of Worksheet S-3, part II, lines 17, 18, 20, 
and 22, to derive Employee Benefits costs. This proposed method allows 
us to obtain data from about 30 more freestanding IRFs than if we were 
to only use the Worksheet S-3, part V, data as was done for the 2012-
based IRF market basket.
    For hospital-based IRFs, we proposed to calculate total benefit 
costs as the sum of inpatient unit benefit costs, a portion of 
ancillary benefits, and a portion of overhead benefits attributable to 
the routine inpatient unit and a portion of overhead benefits 
attributable to the ancillary departments. We proposed inpatient unit 
benefit costs be equal to Worksheet S-3, part V, column 2, line 4. We 
proposed that the portion of overhead benefits attributable to the 
routine inpatient unit and ancillary departments be calculated by 
multiplying ancillary salaries for the hospital-based IRF and overhead 
salaries attributable to the hospital-based IRF (determined in the 
derivation of hospital-based IRF Wages and Salaries costs as described 
above) by the ratio of total facility benefits to total facility 
salaries. Total facility benefits is equal to the sum of Worksheet S-3, 
part II, column 4, lines 17-25, and total facility salaries is equal to 
Worksheet S-3, part II, column 4, line 1.
(3) Contract Labor Costs
    Contract Labor costs are primarily associated with direct patient 
care services. Contract labor costs for other services such as 
accounting, billing, and legal are calculated separately using other 
government data sources as described in section VI.C.3. of this final 
rule. To derive contract labor costs using Worksheet S-3, part V, data, 
for freestanding IRFs, we proposed Contract Labor costs be equal to 
Worksheet S-3, part V, column 1, line 2. As we noted for Employee 
Benefits, freestanding IRFs also may report Contract Labor data on 
Worksheet S-3, part II, which is applicable to only IPPS providers. For 
those freestanding IRFs that report Worksheet S-3, part II data, but 
not Worksheet S-3, part V, we proposed to use the sum of Worksheet S-3, 
part II, lines 11 and 13, to derive Contract Labor costs.
    For hospital-based IRFs, we proposed that Contract Labor costs 
would be equal to Worksheet S-3, part V, column 1, line 4. As 
previously noted, for 2016 MCR data, while there were providers that 
did report data on Worksheet S-3, part V, many providers did not 
complete this worksheet. However, we believe we have a large enough 
sample to enable us to produce a reasonable Contract Labor cost weight. 
We continue to encourage all providers to report these data on the 
Medicare cost report.
(4) Pharmaceuticals Costs
    For freestanding IRFs, we proposed to calculate pharmaceuticals 
costs using

[[Page 39074]]

non-salary costs reported on Worksheet A, column 7, less Worksheet A, 
column 1, for the pharmacy cost center (line 15) and drugs charged to 
patients cost center (line 73).
    For hospital-based IRFs, we proposed to calculate pharmaceuticals 
costs as the sum of a portion of the non-salary pharmacy costs and a 
portion of the non-salary drugs charged to patient costs reported for 
the total facility. We proposed that non-salary pharmacy costs 
attributable to the hospital-based IRF would be calculated by 
multiplying total pharmacy costs attributable to the hospital-based IRF 
(as reported on Worksheet B, part I, column 15, line 41) by the ratio 
of total non-salary pharmacy costs (Worksheet A, column 2, line 15) to 
total pharmacy costs (sum of Worksheet A, columns 1 and 2 for line 15) 
for the total facility. We proposed that non-salary drugs charged to 
patient costs attributable to the hospital-based IRF would be 
calculated by multiplying total non-salary drugs charged to patient 
costs (Worksheet B, part I, column 0, line 73 plus Worksheet B, part I, 
column 15, line 73, less Worksheet A, column 1, line 73) for the total 
facility by the ratio of Medicare drugs charged to patient ancillary 
costs for the IRF unit (as reported on Worksheet D-3 for hospital-based 
IRFs, column 3, line 73) to total Medicare drugs charged to patient 
ancillary costs for the total facility (equal to the sum of Worksheet 
D-3, column 3, line 73 for all relevant PPS [that is, IPPS, IRF, IPF 
and SNF]).
(5) Professional Liability Insurance Costs
    For freestanding IRFs, we proposed that Professional Liability 
Insurance (PLI) costs (often referred to as malpractice costs) would be 
equal to premiums, paid losses and self-insurance costs reported on 
Worksheet S-2, part I, columns 1 through 3, line 118. For hospital-
based IRFs, we proposed to assume that the PLI weight for the total 
facility is similar to the hospital-based IRF unit since the only data 
reported on this worksheet is for the entire facility, as we currently 
have no means to identify the proportion of total PLI costs that are 
only attributable to the hospital-based IRF. Therefore, hospital-based 
IRF PLI costs are equal to total facility PLI (as reported on Worksheet 
S-2, part I, columns 1 through 3, line 118) divided by total facility 
costs (as reported on Worksheet A, columns 1 and 2, line 200) times 
hospital-based IRF Medicare allowable total costs. Our assumption is 
that the same proportion of expenses are used among each unit of the 
hospital.
(6) Home Office/Related Organization Contract Labor Costs
    For the 2016-based IRF market basket, we proposed to determine the 
home office/related organization contract labor costs using MCR data. 
The 2012-based IRF market basket used the 2007 Benchmark Input-Output 
(I-O) expense data published by the Bureau of Economic Analysis (BEA) 
to derive these costs (80 FR 47057). A more detailed explanation of the 
general methodology using the BEA I-O data is provided in section 
VI.C.3. of this final rule. For freestanding and hospital-based IRFs, 
we proposed to calculate the home office contract labor cost weight 
(using data reported on Worksheet S-3, part II, column 4, lines 14, 
1401, 1402, 2550, and 2551) and total facility costs (Worksheet B, part 
I, column 26, line 202). We proposed to use total facility costs as the 
denominator for calculating the home office contract labor cost weight 
as these expenses reported on Worksheet S-3, part II reflect the entire 
hospital facility. Our assumption is that the same proportion of 
expenses are used among each unit of the hospital. For the 2012-based 
IRF market basket, we calculated the home office cost weight using 
expense data for North American Industry Classification System (NAICS) 
code 55, Management of Companies and Enterprises (80 FR 47067).
(7) Capital Costs
    For freestanding IRFs, we proposed that capital costs would be 
equal to Medicare allowable capital costs as reported on Worksheet B, 
part II, column 26, lines 30 through 35, 50 through 76 (excluding 52 
and 75), 90 through 91, and 93.
    For hospital-based IRFs, we proposed that capital costs would be 
equal to IRF inpatient capital costs (as reported on Worksheet B, part 
II, column 26, line 41) and a portion of IRF ancillary capital costs. 
We calculate the portion of ancillary capital costs attributable to the 
hospital-based IRF for a given cost center by multiplying total 
facility ancillary capital costs for the specific ancillary cost center 
(as reported on Worksheet B, part II, column 26) by the ratio of IRF 
Medicare ancillary costs for the cost center (as reported on Worksheet 
D-3, column 3 for hospital-based IRFs) to total Medicare ancillary 
costs for the cost center (equal to the sum of Worksheet D-3, column 3 
for all relevant PPS [that is, IPPS, IRF, IPF and SNF]). For example, 
if hospital-based IRF Medicare physical therapy costs represent 30 
percent of the total Medicare physical therapy costs for the entire 
facility, then 30 percent of total facility physical therapy capital 
costs (as reported in Worksheet B, part II, column 26, line 66) would 
be attributable to the hospital-based IRF.
b. Final Major Cost Category Computation
    After we derive costs for the major cost categories for each 
provider using the MCR data as previously described, we proposed to 
trim the data for outliers. For the Wages and Salaries, Employee 
Benefits, Contract Labor, Pharmaceuticals, Professional Liability 
Insurance, and Capital cost weights, we first divide the costs for each 
of these six categories by total Medicare allowable costs calculated 
for the provider to obtain cost weights for the universe of IRF 
providers. We then remove those providers whose derived cost weights 
fall in the top and bottom 5 percent of provider specific derived cost 
weights to ensure the exclusion of outliers. After the outliers have 
been excluded, we sum the costs for each category across all remaining 
providers. We then divide this by the sum of total Medicare allowable 
costs across all remaining providers to obtain a cost weight for the 
2016-based IRF market basket for the given category.
    The proposed trimming methodology for the Home Office Contract 
Labor cost weight is slightly different than the proposed trimming 
methodology for the other six cost categories as described above. For 
the Home Office Contract Labor cost weight, since we are using total 
facility data rather than Medicare-allowable costs associated with IRF 
services, we proposed to trim the freestanding and hospital-based IRF 
cost weights separately. For each of the providers, we first divide the 
home office contract labor costs by total facility costs to obtain a 
Home Office Contract Labor cost weight for the universe of IRF 
providers. We then proposed to trim only the top 1 percent of providers 
to exclude outliers while also allowing providers who have reported 
zero home office costs to remain in the Home Office Contract Labor cost 
weight calculations as not all providers will incur home office costs. 
After removing these outliers, we are left with a trimmed data set for 
both freestanding and hospital-based providers. We then proposed to sum 
the costs for each category (freestanding and hospital-based) across 
all remaining providers. We next divide this by the sum of total 
facility costs across all remaining providers to obtain a freestanding 
and hospital-based cost weight. Lastly, we proposed to weight these two 
cost weights together using

[[Page 39075]]

the Medicare-allowable costs to derive a Home Office Contract Labor 
cost weight for the 2016-based IRF market basket.
    Finally, we proposed to calculate the residual ``All Other'' cost 
weight that reflects all remaining costs that are not captured in the 
seven cost categories listed.
    We received a few comments on our proposed derivation of the Home 
Office Contract Labor cost weight from the Medicare cost reports, which 
are summarized below.
    Comment: Commenters expressed concern with the proposed methodology 
change to the Home Office Contract Labor cost weight. These commenters 
stated that CMS had not provided sufficient rationale for this change 
in methodology nor has CMS provided a discussion of how these data 
points were reasonably validated and tested. One commenter requested 
that CMS provide stakeholders with more information on the rationale 
and the data validation methodologies employed in the final rule.
    The commenters expressed concern with the sample of IRFs reporting 
the home office cost data and found based on their analysis that 
reporting was between 50 to 65 percent. These commenters suggested that 
this was due to these cost report line items being an optional category 
for IRFs under Medicare cost reporting requirements. One of the 
commenters further expressed concern with the methodology and approach 
that CMS applied in determining IRF unit Home Office Contract Labor 
amounts, specifically the assumption that hospital-based IRFs utilize 
the same proportion of home office expenses as the rest of the acute 
care hospital in which it is located. The commenter stated that 
typically IRF units are a very small part of the larger parent acute 
care hospital and that the larger systems do not spend the same 
proportional time and resources on these units compared to hospital 
system as a whole. They stated that this assumption likely overstates 
the Home Office Contract Labor cost weight.
    Based on these concerns, the commenters requested that CMS not 
finalize its proposed changes to the Home Office Contract Labor cost 
category and instead finalize use of the previous methodology relating 
to this category that was used for the 2012-based market basket. One 
commenter also requested that CMS revisit this potential change with 
adequate explanation and data in future rulemaking.
    Response: We appreciate the commenters' concerns on the proposed 
methodological change for the Home Office Contract Labor cost weight. 
We proposed to revise our methodology and use the 2016 IRF MCR data to 
calculate the Home Office Contract Labor costs rather than the 2012 
Benchmark I-O data because it reflected more up-to-date data and we 
believe it to be an improvement over the use of the BEA Benchmark I-O 
data that is not specific to IRFs. The MCR data allows us to calculate 
Home Office Contract Labor Costs for freestanding and IRF hospital-
based facilities.
    We disagree with the commenters' concern that the MCR data 
completion rates for the Home Office Contract Labor costs are 
inadequate to obtain a cost weight. When developing the proposed 2016-
based IRF market basket, we conducted a thorough analysis of the MCR 
data and our proposed Home Office Contract Labor cost weight 
methodology. We found that approximately 90 percent of freestanding 
IRFs reported having a home office, of which over 50 percent reported 
home office compensation data on Worksheet S-3, part II. The 
composition of the providers (by ownership-type and region) that 
reported both wage index data (including those who do not have a home 
office) and home office contract labor cost data were similarly 
representative to all freestanding IRFs. A sensitivity analysis of 
calculating a reweighted Home Office Contract Labor cost weight based 
on ownership-type and region produced a Home Office Contract Labor cost 
weight similar to the proposed 3.7 percent weight.
    For additional sensitivity testing, recognizing that some of the 
freestanding IRFs with home offices may not have completed the 
applicable fields on the MCR, we calculated a weight using only 
freestanding IRFs that reported having a home office (Worksheet S-2, 
part I, line 140). This produced a Home Office Contract Labor cost 
weight nearly identical to the freestanding IRF 2016 cost weight using 
our proposed methodology. Based on this analysis, we believe that the 
sample of providers included in the Home Office Contract Labor cost 
weight are a technically representative sample of all IRF providers.
    Regarding IRF units, we recognize the commenter's concern that they 
represent a small proportion of the total facility. We believe that the 
assumption that IRFs utilize the same proportion of home office 
expenses as the rest of the acute care hospital is reasonable. The use 
of total facility data assumes the facility Home Office Contract Labor 
cost weight is equal to the Home Office Contract Labor cost weight for 
the IRF unit. Further analysis of the MCR data shows IRF unit direct 
patient care costs (as reported on Worksheet B, part I, column 0, line 
41) account for about one percent of total facility costs (excluding 
capital, Administrative and General (A&G), and Employee Benefit 
department costs). Similarly, A&G costs (Worksheet B, part I, column 0, 
line 5), where Home Office Contract Labor costs are likely captured, 
allocated to the IRF unit account for a similar proportion of direct 
patient care costs with about one percent of total A&G costs. We also 
found the proportion of allocated A&G costs for other larger, more 
medically-complex hospital units (such as the intensive care, surgical 
care, and operating room) were consistent with direct patient care cost 
proportions and the proportions for these units were higher than the 
proportion of the A&G expenses allocated to the IRF unit. This supports 
the commenter's claim that hospitals allocate less A&G costs to less 
medically-complex services (as measured by costs). Our proposed 
calculation would adhere to this assumption as well since the facility 
level cost weight is applied to the IRF Medicare allowable total costs 
representing these relatively less medically-complex services. 
Furthermore, the Benchmark I-O methodology used in the 2012-based IRF 
market basket also assumes that the IRF relative costs are the same as 
those of the hospital total facility. We invite the commenters to 
submit additional data that would help in this area for consideration 
in future rulemaking.
    We disagree with the commenters' request to use the Benchmark I-O 
data to calculate the Home Office Contract Labor cost weight rather 
than the proposed 2016 MCR data. We believe the proposed methodology is 
a technical improvement over the prior methodology because it 
represents more recent data that is representative compositionally and 
geographically of IRFs. It is also is the same data used to determine 
the other major cost weights in the 2016-based market basket and the 
proportion of the Home Office Contract Labor cost weight that is 
allocated to the Professional Fees: Labor-related and Professional 
Fees: Nonlabor-related cost weights. We believe the assumptions made by 
using the total facility data for the hospital-based IRFs are 
reasonable and supported by the MCR data on A&G cost allocation. 
Finally, we note that the methodological change accounts for only 0.2 
percentage point of the 2.0 percentage points change in the labor-
related share.

[[Page 39076]]

    After careful consideration of comments, we are finalizing our 
methodology for deriving the major cost weights as proposed.
    Table 4 presents the cost weights for these major cost categories 
calculated from the Medicare cost reports for the 2016-based IRF market 
basket, as well as for the 2012-based IRF market basket.
[GRAPHIC] [TIFF OMITTED] TR08AU19.007

    As we did for the 2012-based IRF market basket, we proposed to 
allocate the Contract Labor cost weight to the Wages and Salaries and 
Employee Benefits cost weights based on their relative proportions 
under the assumption that contract labor costs are comprised of both 
wages and salaries and employee benefits. The Contract Labor allocation 
proportion for Wages and Salaries is equal to the Wages and Salaries 
cost weight as a percent of the sum of the Wages and Salaries cost 
weight and the Employee Benefits cost weight. For the proposed rule, 
this rounded percentage is 81 percent; therefore, we proposed to 
allocate 81 percent of the Contract Labor cost weight to the Wages and 
Salaries cost weight and 19 percent to the Employee Benefits cost 
weight. The 2012-based IRF market basket percentage was also 81 percent 
(80 FR 47056). We did not receive any specific public comments on our 
proposed allocation of Contract Labor. Therefore, we are finalizing our 
method of allocating Contract Labor as proposed.
    Table 5 shows the Wages and Salaries and Employee Benefit cost 
weights after Contract Labor cost weight allocation for both the 2016-
based IRF market basket and 2012-based IRF market basket.
[GRAPHIC] [TIFF OMITTED] TR08AU19.008

c. Derivation of the Detailed Operating Cost Weights
    To further divide the ``All Other'' residual cost weight estimated 
from the 2016 MCR data into more detailed cost categories, we proposed 
to use the 2012 Benchmark I-O ``Use Tables/Before Redefinitions/
Purchaser Value'' for NAICS 622000, Hospitals, published by the BEA. 
This data is publicly available at http://www.bea.gov/industry/io_annual.htm. For the 2012-based IRF market basket, we used the 2007 
Benchmark I-O data, the most recent data available at the time (80 FR 
47057).
    The BEA Benchmark I-O data are scheduled for publication every 5 
years with the most recent data available for 2012. The 2007 Benchmark 
I-O data are derived from the 2012 Economic Census and are the building 
blocks for BEA's economic accounts. Thus, they represent the most 
comprehensive and complete set of data on the economic processes or 
mechanisms by which output is produced and distributed.\1\ BEA also 
produces Annual I-O estimates; however, while based on a similar 
methodology, these estimates reflect less comprehensive and less 
detailed data sources and are subject to revision when benchmark data 
becomes available. Instead of using the less detailed Annual I-O data, 
we proposed to inflate the 2012 Benchmark I-O data forward to 2016 by 
applying the annual price changes from the respective price proxies to 
the appropriate market basket cost categories that are obtained from 
the 2012 Benchmark I-O data. We repeat this practice for each year. We 
then proposed to calculate the cost shares that each cost category 
represents of the inflated 2012 data. These resulting 2016 cost shares 
are applied to the All Other residual cost weight to obtain the 
detailed cost weights for the 2016-based IRF market basket. For 
example, the cost for Food: Direct Purchases represents 5.0 percent of 
the sum of the ``All Other'' 2012 Benchmark I-O Hospital Expenditures 
inflated to 2016; therefore, the Food: Direct Purchases cost weight 
represents 5.0 percent of the 2016-based IRF market basket's ``All 
Other'' cost category (22.2 percent), yielding a ``final'' Food: Direct 
Purchases cost weight of 1.1 percent in the 2016-based IRF market 
basket (0.05 * 22.2 percent = 1.1 percent).
---------------------------------------------------------------------------

    \1\ http://www.bea.gov/papers/pdf/IOmanual_092906.pdf.
---------------------------------------------------------------------------

    Using this methodology, we proposed to derive seventeen detailed 
IRF market

[[Page 39077]]

basket cost category weights from the 2016-based IRF market basket 
residual cost weight (22.2 percent). These categories are: (1) 
Electricity; (2) Fuel, Oil, and Gasoline; (3) Food: Direct Purchases; 
(4) Food: Contract Services; (5) Chemicals; (6) Medical Instruments; 
(7) Rubber & Plastics; (8) Paper and Printing Products; (9) 
Miscellaneous Products; (10) Professional Fees: Labor-related; (11) 
Administrative and Facilities Support Services; (12) Installation, 
Maintenance, and Repair; (13) All Other Labor-related Services; (14) 
Professional Fees: Nonlabor-related; (15) Financial Services; (16) 
Telephone Services; and (17) All Other Nonlabor-related Services. We 
note that for the 2012-based IRF market basket, we had a Water and 
Sewerage cost weight. For the 2016-based IRF market basket, we proposed 
to include Water and Sewerage costs in the Electricity cost weight due 
to the small amount of costs in this category.
    For the 2012-based IRF market basket, we used the I-O data for 
NAICS 55 Management of Companies to derive the Home Office Contract 
Labor cost weight, which were classified in the Professional Fees: 
Labor-related and Professional Fees: Nonlabor-related cost weights. As 
previously discussed, we proposed to use the MCR data to derive the 
Home Office Contract Labor cost weight, which we would further classify 
into the Professional Fees: Labor-related or Professional Fees: 
Nonlabor-related categories.
    We did not receive any specific comments on the derivation of the 
detailed operating cost weights. In this final rule, we are finalizing 
our methodology for deriving the detailed operating cost weights as 
proposed.
d. Derivation of the Detailed Capital Cost Weights
    As described in section VI.C.1.a.(6) of this final rule, we 
proposed a Capital-Related cost weight of 9.0 percent as obtained from 
the 2016 Medicare cost reports for freestanding and hospital-based IRF 
providers. We proposed to then separate this total Capital-Related cost 
weight into more detailed cost categories.
    Using 2016 Medicare cost reports, we were able to group Capital-
Related costs into the following categories: Depreciation, Interest, 
Lease, and Other Capital-Related costs. For each of these categories, 
we proposed to determine separately for hospital-based IRFs and 
freestanding IRFs what proportion of total capital-related costs the 
category represents.
    For freestanding IRFs, we proposed to derive the proportions for 
Depreciation, Interest, Lease, and Other Capital-related costs using 
the data reported by the IRF on Worksheet A-7, which is similar to the 
methodology used for the 2012-based IRF market basket.
    For hospital-based IRFs, data for these four categories were not 
reported separately for the hospital-based IRF; therefore, we proposed 
to derive these proportions using data reported on Worksheet A-7 for 
the total facility. We assumed the cost shares for the overall hospital 
are representative for the hospital-based IRF unit. For example, if 
depreciation costs make up 60 percent of total capital costs for the 
entire facility, we believe it is reasonable to assume that the 
hospital-based IRF would also have a 60 percent proportion because it 
is a unit contained within the total facility. This is the same 
methodology used for the 2012-based IRF market basket (80 FR 47057).
    To combine each detailed capital cost weight for freestanding and 
hospital-based IRFs into a single capital cost weight for the 2016-
based IRF market basket, we proposed to weight together the shares for 
each of the categories (Depreciation, Interest, Lease, and Other 
Capital-related costs) based on the share of total capital costs each 
provider type represents of the total capital costs for all IRFs for 
2016. Applying this methodology results in proportions of total 
capital-related costs for Depreciation, Interest, Lease and Other 
Capital-related costs that are representative of the universe of IRF 
providers. This is the same methodology used for the 2012-based IRF 
market basket (80 FR 47057 through 47058).
    Lease costs are unique in that they are not broken out as a 
separate cost category in the 2016-based IRF market basket. Rather, we 
proposed to proportionally distribute these costs among the cost 
categories of Depreciation, Interest, and Other Capital-Related, 
reflecting the assumption that the underlying cost structure of leases 
is similar to that of capital-related costs in general. As was done 
under the 2012-based IRF market basket, we proposed to assume that 10 
percent of the lease costs as a proportion of total capital-related 
costs represents overhead and assign those costs to the Other Capital-
Related cost category accordingly. We proposed to distribute the 
remaining lease costs proportionally across the three cost categories 
(Depreciation, Interest, and Other Capital-Related) based on the 
proportion that these categories comprise of the sum of the 
Depreciation, Interest, and Other Capital-related cost categories 
(excluding lease expenses). This resulted in three primary capital-
related cost categories in the 2016-based IRF market basket: 
Depreciation, Interest, and Other Capital-Related costs. This is the 
same methodology used for the 2012-based IRF market basket (80 FR 
47058). The allocation of these lease expenses are shown in Table 6.
    Finally, we proposed to further divide the Depreciation and 
Interest cost categories. We proposed to separate Depreciation into the 
following two categories: (1) Building and Fixed Equipment; and (2) 
Movable Equipment. We proposed to separate Interest into the following 
two categories: (1) Government/Nonprofit; and (2) For-profit.
    To disaggregate the Depreciation cost weight, we need to determine 
the percent of total Depreciation costs for IRFs that are attributable 
to Building and Fixed Equipment, which we hereafter refer to as the 
``fixed percentage.'' For the 2016-based IRF market basket, we proposed 
to use slightly different methods to obtain the fixed percentages for 
hospital-based IRFs compared to freestanding IRFs.
    For freestanding IRFs, we proposed to use depreciation data from 
Worksheet A-7 of the 2016 Medicare cost reports. However, for hospital-
based IRFs, we determined that the fixed percentage for the entire 
facility may not be representative of the hospital-based IRF unit due 
to the entire facility likely employing more sophisticated movable 
assets that are not utilized by the hospital-based IRF. Therefore, for 
hospital-based IRFs, we proposed to calculate a fixed percentage using: 
(1) Building and fixture capital costs allocated to the hospital-based 
IRF unit as reported on Worksheet B, part I, line 41; and (2) building 
and fixture capital costs for the top five ancillary cost centers 
utilized by hospital-based IRFs. We proposed to weight these two fixed 
percentages (inpatient and ancillary) using the proportion that each 
capital cost type represents of total capital costs in the 2016-based 
IRF market basket. We proposed to then weight the fixed percentages for 
hospital-based and freestanding IRFs together using the proportion of 
total capital costs each provider type represents. For both 
freestanding and hospital-based IRFs, this is the same methodology used 
for the 2012-based IRF market basket (80 FR 47058).
    To disaggregate the Interest cost weight, we determined the percent 
of total interest costs for IRFs that are attributable to government 
and nonprofit facilities, which is hereafter referred to as the 
``nonprofit percentage,'' as price pressures associated with these 
types of interest

[[Page 39078]]

costs tend to differ from those for for-profit facilities. For the 
2016-based IRF market basket, we proposed to use interest costs data 
from Worksheet A-7 of the 2016 Medicare cost reports for both 
freestanding and hospital-based IRFs. We proposed to determine the 
percent of total interest costs that are attributed to government and 
nonprofit IRFs separately for hospital-based and freestanding IRFs. We 
then proposed to weight the nonprofit percentages for hospital-based 
and freestanding IRFs together using the proportion of total capital 
costs that each provider type represents.
    We did not receive any specific public comments on the derivation 
of the detailed capital cost weights. In this final rule, we are 
finalizing our methodology for deriving the detailed capital cost 
weights as proposed. Table 6 provides the detailed capital cost share 
composition estimated from the 2016 IRF Medicare cost reports. These 
detailed capital cost share composition percentages are applied to the 
total Capital-Related cost weight of 9.0 percent explained in detail in 
section VI.C.1.a.(6) of this final rule.
[GRAPHIC] [TIFF OMITTED] TR08AU19.009

e. 2016-Based IRF Market Basket Cost Categories and Weights
    Table 7 compares the cost categories and weights for the final 
2016-based IRF market basket compared to the 2012-based IRF market 
basket.

[[Page 39079]]

[GRAPHIC] [TIFF OMITTED] TR08AU19.010

2. Selection of Price Proxies
    After developing the cost weights for the 2016-based IRF market 
basket, we selected the most appropriate wage and price proxies 
currently available to represent the rate of price change for each 
expenditure category. For the majority of the cost weights, we base the 
price proxies on U.S. Bureau of Labor Statistics (BLS) data and group 
them into one of the following BLS categories:
     Employment Cost Indexes. Employment Cost Indexes (ECIs) 
measure the rate of change in employment wage rates and employer costs 
for employee benefits per hour worked. These indexes are fixed-weight 
indexes and strictly measure the change in wage rates and employee 
benefits per hour. ECIs are superior to Average Hourly Earnings (AHE) 
as price proxies for input price indexes because they are not affected 
by shifts in occupation or industry mix, and because they measure pure 
price change and are available by both occupational group and by 
industry. The industry ECIs are based on the NAICS and the occupational 
ECIs are based on the Standard Occupational Classification System 
(SOC).
     Producer Price Indexes. Producer Price Indexes (PPIs) 
measure the average change over time in the selling prices received by 
domestic producers for their output. The prices included in the PPI

[[Page 39080]]

are from the first commercial transaction for many products and some 
services (https://www.bls.gov/ppi/).
     Consumer Price Indexes. Consumer Price Indexes (CPIs) 
measure the average change over time in the prices paid by urban 
consumers for a market basket of consumer goods and services (https://www.bls.gov/cpi/). CPIs are only used when the purchases are similar to 
those of retail consumers rather than purchases at the producer level, 
or if no appropriate PPIs are available.
    We evaluate the price proxies using the criteria of reliability, 
timeliness, availability, and relevance:
     Reliability. Reliability indicates that the index is based 
on valid statistical methods and has low sampling variability. Widely 
accepted statistical methods ensure that the data were collected and 
aggregated in a way that can be replicated. Low sampling variability is 
desirable because it indicates that the sample reflects the typical 
members of the population. (Sampling variability is variation that 
occurs by chance because only a sample was surveyed rather than the 
entire population.)
     Timeliness. Timeliness implies that the proxy is published 
regularly, preferably at least once a quarter. The market baskets are 
updated quarterly, and therefore, it is important for the underlying 
price proxies to be up-to-date, reflecting the most recent data 
available. We believe that using proxies that are published regularly 
(at least quarterly, whenever possible) helps to ensure that we are 
using the most recent data available to update the market basket. We 
strive to use publications that are disseminated frequently, because we 
believe that this is an optimal way to stay abreast of the most current 
data available.
     Availability. Availability means that the proxy is 
publicly available. We prefer that our proxies are publicly available 
because this will help ensure that our market basket updates are as 
transparent to the public as possible. In addition, this enables the 
public to be able to obtain the price proxy data on a regular basis.
     Relevance. Relevance means that the proxy is applicable 
and representative of the cost category weight to which it is applied. 
The CPIs, PPIs, and ECIs that we have selected meet these criteria. 
Therefore, we believe that they continue to be the best measure of 
price changes for the cost categories to which they would be applied.
    Table 10 lists all price proxies that we proposed to use for the 
2016-based IRF market basket. Below is a detailed explanation of the 
price proxies we proposed for each cost category weight. We did not 
receive any specific comments on our proposed price proxies for the 
2016-based IRF market basket. Therefore, in this final rule, we are 
finalizing the price proxies as proposed.
a. Price Proxies for the Operating Portion of the 2016-Based IRF Market 
Basket
(1) Wages and Salaries
    We proposed to continue to use the ECI for Wages and Salaries for 
All Civilian workers in Hospitals (BLS series code CIU1026220000000I) 
to measure the wage rate growth of this cost category. This is the same 
price proxy used in the 2012-based IRF market basket (80 FR 47060).
(2) Benefits
    We proposed to continue to use the ECI for Total Benefits for All 
Civilian workers in Hospitals to measure price growth of this category. 
This ECI is calculated using the ECI for Total Compensation for All 
Civilian workers in Hospitals (BLS series code CIU1016220000000I) and 
the relative importance of wages and salaries within total 
compensation. This is the same price proxy used in the 2012-based IRF 
market basket (80 FR 47060).
(3) Electricity
    We proposed to continue to use the PPI Commodity Index for 
Commercial Electric Power (BLS series code WPU0542) to measure the 
price growth of this cost category. This is the same price proxy used 
in the 2012-based IRF market basket (80 FR 47060).
(4) Fuel, Oil, and Gasoline
    Similar to the 2012-based IRF market basket, for the 2016-based IRF 
market basket, we proposed to use a blend of the PPI for Petroleum 
Refineries and the PPI Commodity for Natural Gas. Our analysis of the 
Bureau of Economic Analysis' 2012 Benchmark Input-Output data (use 
table before redefinitions, purchaser's value for NAICS 622000 
[Hospitals]), shows that Petroleum Refineries expenses account for 
approximately 90 percent and Natural Gas expenses account for 
approximately 10 percent of Hospitals' (NAICS 622000) total Fuel, Oil, 
and Gasoline expenses. Therefore, we proposed to use a blend of 90 
percent of the PPI for Petroleum Refineries (BLS series code 
PCU324110324110) and 10 percent of the PPI Commodity Index for Natural 
Gas (BLS series code WPU0531) as the price proxy for this cost 
category. The 2012-based IRF market basket used a 70/30 blend of these 
price proxies, reflecting the 2007 I-O data (80 FR 47060). We believe 
that these two price proxies continue to be the most technically 
appropriate indices available to measure the price growth of the Fuel, 
Oil, and Gasoline cost category in the 2016-based IRF market basket.
(5) Professional Liability Insurance
    We proposed to continue to use the CMS Hospital Professional 
Liability Index to measure changes in PLI premiums. To generate this 
index, we collect commercial insurance premiums for a fixed level of 
coverage while holding non-price factors constant (such as a change in 
the level of coverage). This is the same proxy used in the 2012-based 
IRF market basket (80 FR 47060).
(6) Pharmaceuticals
    We proposed to continue to use the PPI for Pharmaceuticals for 
Human Use, Prescription (BLS series code WPUSI07003) to measure the 
price growth of this cost category. This is the same proxy used in the 
2012-based IRF market basket (80 FR 47060).
(7) Food: Direct Purchases
    We proposed to continue to use the PPI for Processed Foods and 
Feeds (BLS series code WPU02) to measure the price growth of this cost 
category. This is the same proxy used in the 2012-based IRF market 
basket (80 FR 47060).
(8) Food: Contract Purchases
    We proposed to continue to use the CPI for Food Away From Home (BLS 
series code CUUR0000SEFV) to measure the price growth of this cost 
category. This is the same proxy used in the 2012-based IRF market 
basket (80 FR 47060 through 47061).
(9) Chemicals
    Similar to the 2012-based IRF market basket, we proposed to use a 
four part blended PPI as the proxy for the chemical cost category in 
the 2016-based IRF market basket. The proposed blend is composed of the 
PPI for Industrial Gas Manufacturing, Primary Products (BLS series code 
PCU325120325120P), the PPI for Other Basic Inorganic Chemical 
Manufacturing (BLS series code PCU32518-32518-), the PPI for Other 
Basic Organic Chemical Manufacturing (BLS series code PCU32519-32519-), 
and the PPI for Other Miscellaneous Chemical Product Manufacturing (BLS 
series code PCU325998325998). We note that the four part blended PPI 
used in the 2012-based IRF market basket is composed of the PPI for 
Industrial Gas Manufacturing (BLS series code

[[Page 39081]]

PCU325120325120P), the PPI for Other Basic Inorganic Chemical 
Manufacturing (BLS series code PCU32518-32518-), the PPI for Other 
Basic Organic Chemical Manufacturing (BLS series code PCU32519-32519-), 
and the PPI for Soap and Cleaning Compound Manufacturing (BLS series 
code PCU32561-32561-). For the 2016-based IRF market basket, we 
proposed to derive the weights for the PPIs using the 2012 Benchmark I-
O data. The 2012-based IRF market basket used the 2007 Benchmark I-O 
data to derive the weights for the four PPIs (80 FR 47061).
    Table 8 shows the weights for each of the four PPIs used to create 
the proposed blended Chemical proxy for the 2016 IRF market basket 
compared to the 2012-based blended Chemical proxy.
[GRAPHIC] [TIFF OMITTED] TR08AU19.011

(10) Medical Instruments
    We proposed to continue to use a blend of two PPIs for the Medical 
Instruments cost category. The 2012 Benchmark Input-Output data shows 
an approximate 57/43 split between Surgical and Medical Instruments and 
Medical and Surgical Appliances and Supplies for this cost category. 
Therefore, we proposed a blend composed of 57 percent of the commodity-
based PPI for Surgical and Medical Instruments (BLS series code 
WPU1562) and 43 percent of the commodity-based PPI for Medical and 
Surgical Appliances and Supplies (BLS series code WPU1563). The 2012-
based IRF market basket used a 50/50 blend of these PPIs based on the 
2007 Benchmark I-O data (80 FR 47061).
(11) Rubber and Plastics
    We proposed to continue to use the PPI for Rubber and Plastic 
Products (BLS series code WPU07) to measure price growth of this cost 
category. This is the same proxy used in the 2012-based IRF market 
basket (80 FR 47061).
(12) Paper and Printing Products
    We proposed to continue to use the PPI for Converted Paper and 
Paperboard Products (BLS series code WPU0915) to measure the price 
growth of this cost category. This is the same proxy used in the 2012-
based IRF market basket (80 FR 47061).
(13) Miscellaneous Products
    We proposed to continue to use the PPI for Finished Goods Less Food 
and Energy (BLS series code WPUFD4131) to measure the price growth of 
this cost category. This is the same proxy used in the 2012-based IRF 
market basket (80 FR 47061).
(14) Professional Fees: Labor-Related
    We proposed to continue to use the ECI for Total Compensation for 
Private Industry workers in Professional and Related (BLS series code 
CIU2010000120000I) to measure the price growth of this category. This 
is the same proxy used in the 2012-based IRF market basket (80 FR 
47061).
(15) Administrative and Facilities Support Services
    We proposed to continue to use the ECI for Total Compensation for 
Private Industry workers in Office and Administrative Support (BLS 
series code CIU2010000220000I) to measure the price growth of this 
category. This is the same proxy used in the 2012-based IRF market 
basket (80 FR 47061).
(16) Installation, Maintenance, and Repair
    We proposed to continue to use the ECI for Total Compensation for 
Civilian workers in Installation, Maintenance, and Repair (BLS series 
code CIU1010000430000I) to measure the price growth of this cost 
category. This is the same proxy used in the 2012-based IRF market 
basket (80 FR 47061).
(17) All Other: Labor-Related Services
    We proposed to continue to use the ECI for Total Compensation for 
Private Industry workers in Service Occupations (BLS series code 
CIU2010000300000I) to measure the price growth of this cost category. 
This is the same proxy used in the 2012-based IRF market basket (80 FR 
47061).
(18) Professional Fees: Nonlabor-Related
    We proposed to continue to use the ECI for Total Compensation for 
Private Industry workers in Professional and Related (BLS series code 
CIU2010000120000I) to measure the price growth of this category. This 
is the same proxy used in the 2012-based IRF market basket (80 FR 
47061).
(19) Financial Services
    We proposed to continue to use the ECI for Total Compensation for 
Private Industry workers in Financial Activities (BLS series code 
CIU201520A000000I) to measure the price growth of this cost category. 
This is the same proxy used in the 2012-based IRF market basket (80 FR 
47061).
(20) Telephone Services
    We proposed to continue to use the CPI for Telephone Services (BLS 
series code CUUR0000SEED) to measure the price growth of this cost 
category. This is the same proxy used in the 2012-based IRF market 
basket (80 FR 47061).
(21) All Other: Nonlabor-Related Services
    We proposed to continue to use the CPI for All Items Less Food and 
Energy (BLS series code CUUR0000SA0L1E) to measure the price growth of 
this cost category. This is the same proxy used in the 2012-based IRF 
market basket (80 FR 47061).
b. Price Proxies for the Capital Portion of the 2016-Based IRF Market 
Basket
(1) Capital Price Proxies Prior to Vintage Weighting
    We proposed to continue to use the same price proxies for the 
capital-related cost categories in the 2016-based

[[Page 39082]]

IRF market basket as were used in the 2012-based IRF market basket (80 
FR 47062), which are provided in Table 10 and described below. 
Specifically, we proposed to proxy:
     Depreciation: Building and Fixed Equipment cost category 
by BEA's Chained Price Index for Nonresidential Construction for 
Hospitals and Special Care Facilities (BEA Table 5.4.4. Price Indexes 
for Private Fixed Investment in Structures by Type).
     Depreciation: Movable Equipment cost category by the PPI 
for Machinery and Equipment (BLS series code WPU11).
     Nonprofit Interest cost category by the average yield on 
domestic municipal bonds (Bond Buyer 20-bond index).
     For-profit Interest cost category by the average yield on 
Moody's Aaa bonds (Federal Reserve).
     Other Capital-Related cost category by the CPI-U for Rent 
of Primary Residence (BLS series code CUUS0000SEHA).
    We believe these are the most appropriate proxies for IRF capital-
related costs that meet our selection criteria of relevance, 
timeliness, availability, and reliability. We proposed to continue to 
vintage weight the capital price proxies for Depreciation and Interest 
to capture the long-term consumption of capital. This vintage weighting 
method is similar to the method used for the 2012-based IRF market 
basket (80 FR 47062) and is described below.
(2) Vintage Weights for Price Proxies
    Because capital is acquired and paid for over time, capital-related 
expenses in any given year are determined by both past and present 
purchases of physical and financial capital. The vintage-weighted 
capital-related portion of the 2016-based IRF market basket is intended 
to capture the long-term consumption of capital, using vintage weights 
for depreciation (physical capital) and interest (financial capital). 
These vintage weights reflect the proportion of capital-related 
purchases attributable to each year of the expected life of building 
and fixed equipment, movable equipment, and interest. We proposed to 
use vintage weights to compute vintage-weighted price changes 
associated with depreciation and interest expenses.
    Capital-related costs are inherently complicated and are determined 
by complex capital-related purchasing decisions, over time, based on 
such factors as interest rates and debt financing. In addition, capital 
is depreciated over time instead of being consumed in the same period 
it is purchased. By accounting for the vintage nature of capital, we 
are able to provide an accurate and stable annual measure of price 
changes. Annual non-vintage price changes for capital are unstable due 
to the volatility of interest rate changes, and therefore, do not 
reflect the actual annual price changes for IRF capital-related costs. 
The capital-related component of the 2016-based IRF market basket 
reflects the underlying stability of the capital-related acquisition 
process.
    The methodology used to calculate the vintage weights for the 2016-
based IRF market basket is the same as that used for the 2012-based IRF 
market basket (80 FR 47062 through 47063) with the only difference 
being the inclusion of more recent data. To calculate the vintage 
weights for depreciation and interest expenses, we first need a time 
series of capital-related purchases for building and fixed equipment 
and movable equipment. We found no single source that provides an 
appropriate time series of capital-related purchases by hospitals for 
all of the above components of capital purchases. The early Medicare 
cost reports did not have sufficient capital-related data to meet this 
need. Data we obtained from the American Hospital Association (AHA) do 
not include annual capital-related purchases. However, we are able to 
obtain data on total expenses back to 1963 from the AHA. Consequently, 
we proposed to use data from the AHA Panel Survey and the AHA Annual 
Survey to obtain a time series of total expenses for hospitals. We then 
proposed to use data from the AHA Panel Survey supplemented with the 
ratio of depreciation to total hospital expenses obtained from the 
Medicare cost reports to derive a trend of annual depreciation expenses 
for 1963 through 2016. We proposed to separate these depreciation 
expenses into annual amounts of building and fixed equipment 
depreciation and movable equipment depreciation as determined earlier. 
From these annual depreciation amounts, we derive annual end-of-year 
book values for building and fixed equipment and movable equipment 
using the expected life for each type of asset category. While data is 
not available that is specific to IRFs, we believe this information for 
all hospitals serves as a reasonable alternative for the pattern of 
depreciation for IRFs.
    To continue to calculate the vintage weights for depreciation and 
interest expenses, we also need to account for the expected lives for 
Building and Fixed Equipment, Movable Equipment, and Interest for the 
2016-based IRF market basket. We proposed to calculate the expected 
lives using MCR data from freestanding and hospital-based IRFs. The 
expected life of any asset can be determined by dividing the value of 
the asset (excluding fully depreciated assets) by its current year 
depreciation amount. This calculation yields the estimated expected 
life of an asset if the rates of depreciation were to continue at 
current year levels, assuming straight-line depreciation. We proposed 
to determine the expected life of building and fixed equipment 
separately for hospital-based IRFs and freestanding IRFs, and then 
weight these expected lives using the percent of total capital costs 
each provider type represents. We proposed to apply a similar method 
for movable equipment. Using these methods, we determined the average 
expected life of building and fixed equipment to be equal to 22 years, 
and the average expected life of movable equipment to be equal to 11 
years. For the expected life of interest, we believe vintage weights 
for interest should represent the average expected life of building and 
fixed equipment because, based on previous research described in the FY 
1997 IPPS final rule (61 FR 46198), the expected life of hospital debt 
instruments and the expected life of buildings and fixed equipment are 
similar. We note that for the 2012-based IRF market basket, the 
expected life of building and fixed equipment is 23 years, and the 
expected life of movable equipment is 11 years (80 FR 47062).
    Multiplying these expected lives by the annual depreciation amounts 
results in annual year-end asset costs for building and fixed equipment 
and movable equipment. We then calculate a time series, beginning in 
1964, of annual capital purchases by subtracting the previous year's 
asset costs from the current year's asset costs.
    For the building and fixed equipment and movable equipment vintage 
weights, we proposed to use the real annual capital-related purchase 
amounts for each asset type to capture the actual amount of the 
physical acquisition, net of the effect of price inflation. These real 
annual capital-related purchase amounts are produced by deflating the 
nominal annual purchase amount by the associated price proxy as 
provided earlier in this final rule. For the interest vintage weights, 
we proposed to use the total nominal annual capital-related purchase 
amounts to capture the value of the debt instrument (including, but not 
limited to, mortgages and bonds). Using these capital-related purchase 
time series specific to each asset type, we proposed to calculate the 
vintage weights for

[[Page 39083]]

building and fixed equipment, for movable equipment, and for interest.
    The vintage weights for each asset type are deemed to represent the 
average purchase pattern of the asset over its expected life (in the 
case of building and fixed equipment and interest, 22 years, and in the 
case of movable equipment, 11 years). For each asset type, we used the 
time series of annual capital-related purchase amounts available from 
2016 back to 1964. These data allow us to derive 32, 22-year periods of 
capital-related purchases for building and fixed equipment and 
interest, and 43, 11-year periods of capital-related purchases for 
movable equipment. For each 22-year period for building and fixed 
equipment and interest, or 11-year period for movable equipment, we 
calculate annual vintage weights by dividing the capital-related 
purchase amount in any given year by the total amount of purchases over 
the entire 22-year or 11-year period. This calculation is done for each 
year in the 22-year or 11-year period and for each of the periods for 
which we have data. We then calculate the average vintage weight for a 
given year of the expected life by taking the average of these vintage 
weights across the multiple periods of data.
    We did not receive any specific public comments on our proposed 
calculation of the vintage weights for the 2016-based IRF market 
basket. Therefore, in this final rule, we are finalizing the vintage 
weights as proposed. The vintage weights for the capital-related 
portion of the 2016-based IRF market basket and the 2012-based IRF 
market basket are presented in Table 9.
[GRAPHIC] [TIFF OMITTED] TR08AU19.012

    The process of creating vintage-weighted price proxies requires 
applying the vintage weights to the price proxy index where the last 
applied vintage weight in Table 8 is applied to the most recent data 
point. We have provided on the CMS website an example of how the 
vintage weighting price proxies are calculated, using example vintage 
weights and example price indices. The example can be found at http://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/MedicareProgramRatesStats/MarketBasketResearch.html in the zip 
file titled ``Weight Calculations as described in the IPPS FY 2010 
Proposed Rule.''
c. Summary of Price Proxies of the 2016-Based IRF Market Basket
    Table 10 shows both the operating and capital price proxies for the 
2016-based IRF market basket.
BILLING CODE 4120-01-P

[[Page 39084]]

[GRAPHIC] [TIFF OMITTED] TR08AU19.013


[[Page 39085]]


BILLING CODE 4120-01-C

D. FY 2020 Market Basket Update and Productivity Adjustment

1. FY 2020 Market Basket Update
    For FY 2020 (that is, beginning October 1, 2019 and ending 
September 30, 2020), we proposed to use the 2016-based IRF market 
basket increase factor described in section V.C. of the proposed rule 
to update the IRF PPS base payment rate. Consistent with historical 
practice, we proposed to estimate the market basket update for the IRF 
PPS based on IHS Global Inc.'s (IGI's) forecast using the most recent 
available data. IGI is a nationally-recognized economic and financial 
forecasting firm with which we contract to forecast the components of 
the market baskets and MFP. In the FY 2020 IRF PPS proposed rule (84 FR 
17274), we proposed a market basket increase factor of 3.0 percent for 
FY 2020, which was based on IGI's first quarter 2019 forecast with 
historical data through fourth quarter 2018.
    In the FY 2020 IRF PPS proposed rule, we also proposed that if more 
recent data were subsequently available (for example, a more recent 
estimate of the market basket and MFP adjustment), we would use such 
data to determine the FY 2020 update in the final rule. Incorporating 
more recent data, the projected 2016-based IRF market basket increase 
factor for FY 2020 is 2.9 percent, which is based on IGI's second 
quarter 2019 forecast with historical data through first quarter 2019.
    We received several comments on our proposed market basket update 
and productivity adjustment, which are summarized below.
    Comment: Commenters supported the proposal to update the market 
basket and MFP adjustment using the latest available data, and 
encouraged CMS to update these factors using the latest available data 
as part of the release of the FY 2020 IRF PPS final rule.
    Response: We appreciate the commenters' support for updating the 
market basket and MFP adjustments using the latest available data.
    Comment: A few commenters expressed concern about the lack of 
transparency of the market basket and MFP payment updates. The 
commenters stated that the IGI forecast appears to be procured 
specifically for the purpose of CMS updating the IRF market basket and 
productivity adjustment. The commenters also noted that it is 
concerning that CMS does not provide IGI's analyses or report to the 
public given the key role the market basket and productivity adjustment 
play in updating the payment system each year and that without such 
information stakeholders are unable to evaluate the accuracy of the 
update. The commenters also mentioned that the same comment was 
submitted in the FY 2019 rulemaking process but they do not believe 
that the response was adequate since the actual analysis or report used 
to create the forecasts was not provided (83 FR 38525). The commenters 
requested that CMS release an IGI report and analysis used to update 
the IRF market basket and standard payment conversion factor.
    Response: IGI regularly produces and publishes a wide variety of 
forecasted series on a monthly or quarterly basis. These forecasts are 
derived using a framework of proprietary economic models that are 
created and updated regularly by IGI. IGI provides these forecasts to a 
wide array of clients in addition to CMS. We use a contractor for the 
price forecasts so that the forecasts are independent and reflect a 
complete economic forecasting model, a capability that we do not have. 
IGI has received multiple awards for their macroeconomic forecast 
accuracy of major economic indicators. We use IGI's price forecasts in 
all of the FFS market baskets used for payment updates and has used the 
forecasts produced by this company for many years.
    We select approximately 30 individual price proxies as inputs to 
the IRF market basket calculation. The price series are discussed in 
detail as part of the rulemaking process. In order to derive a forecast 
of the IRF market basket index, we contract with IGI to procure the 
forecasts of these individual price proxies on a quarterly basis. We 
then combine these price proxies with the market basket base year cost 
weights to derive the levels of the IRF market basket. The data sources 
and methods used to derive these cost weights are discussed in detail 
as part of the rulemaking process.
    As provided in our previous response to this comment in the FY 2019 
IRF PPS final rule (83 FR 38525), the market basket update is derived 
using: (1) The market basket base year cost weights as finalized by CMS 
through rulemaking; and (2) the most up-to-date forecast of the price 
proxies used in the market basket as forecasted by IGI. Specifically, 
for each cost category in the market basket (for example, Wages and 
Salaries, Pharmaceuticals), the level of each of these price proxies 
are multiplied by the cost weight for that cost category. The sum of 
these products (that is, weights multiplied by proxied index levels) 
for all cost categories yields the composite index level in the market 
basket in a given year.
    As acknowledged by the commenters, we provided a link from the CMS 
website to the top-line market basket updates. We also indicated that 
more detailed forecasts of the IRF market basket calculations are 
readily available by request by sending an email to [email protected] 
to request this information (83 FR 38525). Using these detailed data, 
the commenter would be able to replicate the levels of the IRF market 
basket update in the history and the forecast period. We encourage 
stakeholders to utilize these data, which we believe will address the 
commenters' concerns.
    Incorporating more recent data, the projected 2016-based IRF market 
basket update for FY 2020 is 2.9 percent. After careful consideration 
of the comments, consistent with our historical practice of estimating 
market basket increases based on the best available data, we are 
finalizing a market basket increase factor of 2.9 percent for FY 2020. 
For comparison, the current 2012-based IRF market basket is also 
projected to increase by 2.9 percent in FY 2020 based on IGI's second 
quarter 2019 forecast.
    Table 11 compares the 2016-based IRF market basket and the 2012-
based IRF market basket percent changes.

[[Page 39086]]

[GRAPHIC] [TIFF OMITTED] TR08AU19.014

2. Productivity Adjustment
    According to section 1886(j)(3)(C)(i) of the Act, the Secretary 
shall establish an increase factor based on an appropriate percentage 
increase in a market basket of goods and services. As described in 
sections VI.C and VI.D.1. of this final rule, we are finalizing an 
estimate of the IRF PPS increase factor for FY 2020 based on the 2016-
based IRF market basket. Section 1886(j)(3)(C)(ii) of the Act then 
requires that, after establishing the increase factor for a FY, the 
Secretary shall reduce such increase factor for FY 2012 and each 
subsequent FY, by the productivity adjustment described in section 
1886(b)(3)(B)(xi)(II) of the Act. Section 1886(b)(3)(B)(xi)(II) of the 
Act sets forth the definition of this productivity adjustment. The 
statute defines the productivity adjustment to be equal to the 10-year 
moving average of changes in annual economy-wide private nonfarm 
business MFP (as projected by the Secretary for the 10-year period 
ending with the applicable FY, year, cost reporting period, or other 
annual period) (the ``MFP adjustment''). The BLS publishes the official 
measure of private nonfarm business MFP. Please see http://www.bls.gov/mfp for the BLS historical published MFP data.
    MFP is derived by subtracting the contribution of labor and capital 
input growth from output growth. The projections of the components of 
MFP are currently produced by IGI, a nationally recognized economic 
forecasting firm with which CMS contracts to forecast the components of 
the market basket and MFP. For more information on the productivity 
adjustment, we refer reader to the discussion in the FY 2016 IRF PPS 
final rule (80 FR 47065).
    Using IGI's first quarter 2019 forecast, the proposed MFP 
adjustment for FY 2020 (the 10-year moving average of MFP for the 
period ending FY 2020) was 0.5 percent (84 FR 17274). Thus, in 
accordance with section 1886(j)(3)(C) of the Act, we proposed to base 
the FY 2020 market basket update, which is used to determine the 
applicable percentage increase for the IRF payments, on the most recent 
estimate of the 2016-based IRF market basket. We proposed to then 
reduce this percentage increase by the current estimate of the proposed 
MFP adjustment for FY 2020 of 0.5 percentage point (the 10-year moving 
average of MFP for the period ending FY 2020 based on IGI's first 
quarter 2019 forecast). Therefore, the proposed FY 2020 IRF update was 
2.5 percent (3.0 percent market basket update, less 0.5 percentage 
point MFP adjustment). Furthermore, we proposed that if more recent 
data are subsequently available (for example, a more recent estimate of 
the market basket and MFP adjustment), we would use such data to 
determine the FY 2020 market basket update and MFP adjustment in the 
final rule.
    We received a few comments on the application of the productivity 
adjustment, which are summarized below.
    Comment: Commenters continue to be concerned about the application 
of the productivity adjustment to IRFs. One of the commenters stated 
that they understood CMS is bound by statute to reduce the market 
basket update by a productivity adjustment factor in accordance with 
the PPACA, but they believe that IRFs are unable to generate additional 
productivity gains at a pace matching the productivity of the economy 
at large on an ongoing, consistent basis. The commenter noted that the 
services provided in IRFs are labor-intensive and the services do not 
lend themselves to continuous productivity improvements. The commenter 
also noted that IRFs are bound by unchanging labor-intensive standards 
such as the 3-hour therapy rule and other regulatory requirements that 
reduce flexibility and restrict the pursuit of certain efficiencies. 
The commenter noted that continued application of a productivity 
adjustment to payments could results in decreased beneficiary access to 
IRF services. The commenter requested that CMS continue to monitor the 
impact that the multi-factor productivity adjustments have on the IRF 
sector, provide feedback to Congress as appropriate, and reduce the 
productivity adjustment. One commenter requested that, in addition to 
monitoring its effects on overall payments, CMS should evaluate whether 
IRFs are able to achieve the same level of productivity improvement as 
workers across the U.S. economy.
    Response: We acknowledge the commenters' concerns regarding 
productivity growth at the economy-wide level and its application to 
IRFs. As the commenter acknowledges, section 1886(j)(3)(C)(ii)(I) of 
the Act requires the application of a productivity adjustment to the 
IRF PPS market basket increase factor.
    We will continue to monitor the impact of the payment updates, 
including the effects of the productivity

[[Page 39087]]

adjustment, on IRF finances, as well as beneficiary access to care.
    We note that each year, MedPAC makes an annual update 
recommendation to Congress based on a variety of measures related to 
payment adequacy, including a detailed margin analysis and analysis of 
beneficiary access to care for IRF services. For FY 2020, MedPAC 
recommended that Congress reduce the IRF PPS base rate by 5 percent and 
found that beneficiary access to care was not a concern. The ``March 
2019 Report to the Congress: Medicare Payment Policy'', chapter 10 is 
publicly available at http://www.medpac.gov/-documents-/reports.
    We would be very interested in better understanding IRF-specific 
productivity; however, the data elements required to estimate IRF 
specific multi-factor productivity are not produced at the level of 
detail that would allow this analysis. We have estimated hospital-
sector multi-factor productivity and have published the findings on the 
CMS website at https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/ReportsTrustFunds/Downloads/ProductivityMemo2016.pdf.
    After careful consideration of comments, we are incorporating more 
recent data to determine the market basket update and MFP adjustment 
for FY 2020. Using IGI's second quarter 2019 forecast, the current 
estimate of the MFP adjustment for FY 2020 (the 10-year moving average 
of MFP for the period ending FY 2020) is 0.4 percent. Thus, in 
accordance with section 1886(j)(3)(C) of the Act, we are finalizing a 
FY 2020 market basket update of 2.9 percent. We then reduce this 
percentage increase by the most recent estimate of the MFP adjustment 
for FY 2020 of 0.4 percentage point (the 10-year moving average of MFP 
for the period ending FY 2020 based on IGI's second quarter 2019 
forecast). Therefore, the final FY 2020 IRF productivity-adjusted 
market basket update is equal to 2.5 percent (2.9 percent market basket 
update, less 0.4 percentage point MFP adjustment).
    For FY 2020, the Medicare Payment Advisory Commission (MedPAC) 
recommends that a decrease of 5 percent be applied to IRF PPS payment 
rates. As discussed, and in accordance with section 1886(j)(3)(C) of 
the Act, we are finalizing an update to IRF PPS payment rates for FY 
2020 by a productivity-adjusted market basket increase factor of 2.5 
percent, as section 1886(j)(3)(C) of the Act does not provide the 
Secretary with the authority to apply a different update factor to IRF 
PPS payment rates for FY 2020.
    Comment: One commenter (MedPAC) stated that they understand that 
CMS is required to implement the statutory update of market basket less 
productivity adjustment, but that their analysis of beneficiary access 
to rehabilitative services, the supply of providers, and aggregate IRF 
Medicare margins, which have been above 11 percent since 2012, 
indicates that the Congress should reduce the IRF payment rate by 5 
percent for FY 2020.
    Response: We appreciate MedPAC's interest in the IRF increase 
factor. However, we are required to update IRF PPS payments by the 
market basket reduced by the productivity adjustment, as directed by 
section 1886(j)(3)(C) of the Act.

E. Labor-Related Share for FY 2020

    Section 1886(j)(6) of the Act specifies that the Secretary is to 
adjust the proportion (as estimated by the Secretary from time to time) 
of rehabilitation facilities' costs which are attributable to wages and 
wage-related costs, of the prospective payment rates computed under 
section 1886(j)(3) of the Act for area differences in wage levels by a 
factor (established by the Secretary) reflecting the relative hospital 
wage level in the geographic area of the rehabilitation facility 
compared to the national average wage level for such facilities. The 
labor-related share is determined by identifying the national average 
proportion of total costs that are related to, influenced by, or vary 
with the local labor market. We proposed to continue to classify a cost 
category as labor-related if the costs are labor-intensive and vary 
with the local labor market. As stated in the FY 2016 IRF PPS final 
rule (80 FR 47068), the labor-related share was defined as the sum of 
the relative importance of Wages and Salaries, Employee Benefits, 
Professional Fees: Labor-related Services, Administrative and 
Facilities Support Services, Installation, Maintenance, and Repair, All 
Other: Labor-related Services, and a portion of the Capital Costs from 
the 2012-based IRF market basket.
    Based on our definition of the labor-related share and the cost 
categories in the 2016-based IRF market basket, we proposed to include 
in the labor-related share for FY 2020 the sum of the FY 2020 relative 
importance of Wages and Salaries, Employee Benefits, Professional Fees: 
Labor-related, Administrative and Facilities Support Services, 
Installation, Maintenance, and Repair, All Other: Labor-related 
Services, and a portion of the Capital-Related cost weight from the 
2016-based IRF market basket.
    Similar to the 2012-based IRF market basket (80 FR 47067), the 
2016-based IRF market basket includes two cost categories for 
nonmedical Professional Fees (including, but not limited to, expenses 
for legal, accounting, and engineering services). These are 
Professional Fees: Labor-related and Professional Fees: Nonlabor-
related. For the 2016-based IRF market basket, we proposed to estimate 
the labor-related percentage of non-medical professional fees (and 
assign these expenses to the Professional Fees: Labor-related services 
cost category) based on the same method that was used to determine the 
labor-related percentage of professional fees in the 2012-based IRF 
market basket.
    As was done in the 2012-based IRF market basket (80 FR 47067), we 
proposed to determine the proportion of legal, accounting and auditing, 
engineering, and management consulting services that meet our 
definition of labor-related services based on a survey of hospitals 
conducted by us in 2008, a discussion of which can be found in the FY 
2010 IPPS/LTCH PPS final rule (74 FR 43850 through 43856). Based on the 
weighted results of the survey, we determined that hospitals purchase, 
on average, the following portions of contracted professional services 
outside of their local labor market:
     34 percent of accounting and auditing services.
     30 percent of engineering services.
     33 percent of legal services.
     42 percent of management consulting services.
    We proposed to apply each of these percentages to the respective 
Benchmark I-O cost category underlying the professional fees cost 
category to determine the Professional Fees: Nonlabor-related costs. 
The Professional Fees: Labor-related costs were determined to be the 
difference between the total costs for each Benchmark I-O category and 
the Professional Fees: Nonlabor-related costs. This is the same 
methodology that we used to separate the 2012-based IRF market basket 
professional fees category into Professional Fees: Labor-related and 
Professional Fees: Nonlabor-related cost categories (80 FR 47067).
    In the 2016-based IRF market basket, nonmedical professional fees 
that are subject to allocation based on these survey results represent 
4.4 percent of total costs (and are limited to those fees related to 
Accounting & Auditing, Legal, Engineering, and Management Consulting 
services). Based on our survey results, we proposed to apportion 2.8 
percentage points of the

[[Page 39088]]

4.4 percentage point figure into the Professional Fees: Labor-related 
share cost category and designate the remaining 1.6 percentage point 
into the Professional Fees: Nonlabor-related cost category.
    In addition to the professional services listed, for the 2016-based 
IRF market basket, we proposed to allocate a proportion of the Home 
Office Contract Labor cost weight, calculated using the Medicare cost 
reports as stated above, into the Professional Fees: Labor-related and 
Professional Fees: Nonlabor-related cost categories. We proposed to 
classify these expenses as labor-related and nonlabor-related as many 
facilities are not located in the same geographic area as their home 
office, and therefore, do not meet our definition for the labor-related 
share that requires the services to be purchased in the local labor 
market. For the 2012-based IRF market basket, we used the BEA I-O 
expense data for NAICS 55, Management of Companies and Enterprises, to 
estimate the Home Office Contract Labor cost weight (80 FR 47067). We 
then allocated these expenses into the Professional Fess: Labor-related 
and Professional Fees: Nonlabor-related cost categories.
    Similar to the 2012-based IRF market basket, we proposed for the 
2016-based IRF market basket to use the Medicare cost reports for both 
freestanding IRF providers and hospital-based IRF providers to 
determine the home office labor-related percentages. The MCR requires a 
hospital to report information regarding their home office provider. 
For the 2016-based IRF market basket, we proposed to start with the 
sample of IRF providers that passed the top 1 percent trim used to 
derive the Home Office Contract Labor cost weight as described in 
section VI.B. of this final rule. For both freestanding and hospital-
based providers, we proposed to multiply each provider's Home Office 
Contract Labor cost weight (calculated using data from the total 
facility) by Medicare allowable total costs. This results in an amount 
of Medicare allowable home office compensation costs for each IRF. 
Using information on the Medicare cost report, we then compare the 
location of the IRF with the location of the IRF's home office. We 
proposed to classify an IRF with a home office located in their 
respective local labor market if the IRF and its home office are 
located in the same Metropolitan Statistical Area. We then calculate 
the proportion of Medicare allowable home office compensation costs 
that these IRFs represent of total Medicare allowable home office 
compensation costs. We proposed to multiply this percentage (42 
percent) by the Home Office Contract Labor cost weight (3.7 percent) to 
determine the proportion of costs that should be allocated to the 
labor-related share. Therefore, we allocated 1.6 percentage points of 
the Home Office Contract Labor cost weight (3.7 percent times 42 
percent) to the Professional Fees: Labor-related cost weight and 2.1 
percentage points of the Home Office Contract Labor cost weight to the 
Professional Fees: Nonlabor-related cost weight (3.7 percent times 58 
percent). For the 2012-based IRF market basket, we used a similar 
methodology but we relied on provider counts rather than home office/
related organization contract labor compensation costs to determine the 
labor-related percentage (80 FR 47067).
    In summary, we apportioned 2.8 percentage points of the non-medical 
professional fees and 1.6 percentage points of the home office/related 
organization contract labor cost weights into the Professional Fees: 
Labor-related cost category. This amount was added to the portion of 
professional fees that was identified to be labor-related using the I-O 
data such as contracted advertising and marketing costs (approximately 
0.6 percentage point of total costs) resulting in a Professional Fees: 
Labor-related cost weight of 5.0 percent.
    We received several comments on the proposed labor-related share, 
which are summarized below.
    Comment: A few commenters noted that the cost weight for Home 
Office Contract Labor costs is 3.7 percent of all IRFs' costs and 
influences changes in other payment areas, such as the total labor-
related share. The commenters stated that they believe the proposed 
changes to the methodology are responsible, at least in large part, to 
the notable proposed increase of approximately 2 percent of the labor-
related share. Some of the commenters also stated that the increase in 
the labor-related share will adversely impact rural IRFs and IRFs with 
a wage index below 1.0.
    Response: The labor-related share for IRFs is derived from the 
relative importance of the labor-related cost categories. The relative 
importance for FY 2020 reflects the different rates of price change for 
each of the individual cost categories between the base year and FY 
2020. For the FY 2020 final rule, as proposed, the final labor-related 
share for FY 2020 is based on a more recent forecast of the 2016-based 
IRF market basket. Using the more recent forecast, the total difference 
between the FY 2020 labor-related share using the 2016-based IRF market 
basket and 2012-based IRF market basket is 2.0 percentage points (72.7 
percent using 2016-based IRF market basket and 70.7 percent using 2012-
based IRF market basket). This difference can be separated into two 
primary components: (1) Revision to the base year cost weights (1.4 
percentage points); and (2) revision to starting point of calculation 
of relative importance (base year) from 2012 to 2016 (0.6 percentage 
point). Of the 1.4-percentage points difference in the base year cost 
weights, just 0.2 percentage point is attributable to deriving the Home 
Office Contract Labor cost weight using the MCR data rather than the I-
O data; the remainder is due to the increase in Compensation and 
Capital cost weights (calculated using the MCR data) and the 
incorporation of the 2012 Benchmark I-O data.
    The impact of using the MCR data to calculate the Home Office 
Contract Labor cost weight is minimal because it also lowers the 
residual ``All Other'' cost weight from 25.8 percent (using the I-O 
data to calculate the Home Office Contract Labor cost weight) to 22.2 
percent (using the MCR data to calculate the Home Office Contract labor 
cost weight). The lower residual ``All Other'' cost weight then leads 
to relatively lower cost weights for Administrative and Business 
Support Services, Installation, Maintenance and Repair Services, and 
All Other: Labor-related Services (which are calculated using the 
Benchmark I-O data), each of which is also reflected in the labor-
related share.
    After careful consideration of comments, in this final rule, we are 
finalizing the 2016-based IRF market basket labor-related share cost 
weights as proposed.
    As stated previously, we proposed to include in the labor-related 
share the sum of the relative importance of Wages and Salaries, 
Employee Benefits, Professional Fees: Labor-Related, Administrative and 
Facilities Support Services, Installation, Maintenance, and Repair, All 
Other: Labor-related Services, and a portion of the Capital-Related 
cost weight from the 2016-based IRF market basket. The relative 
importance reflects the different rates of price change for these cost 
categories between the base year (2016) and FY 2020. Based on IGI's 2nd 
quarter 2019 forecast for the 2016-based IRF market basket, the sum of 
the FY 2020 relative importance for Wages and Salaries, Employee 
Benefits, Professional Fees: Labor-related, Administrative and 
Facilities Support Services, Installation Maintenance & Repair 
Services, and All Other: Labor-related Services is 68.7 percent. The 
portion of Capital costs that are influenced by the local labor market 
is estimated to be 46 percent, which is the same percentage applied to

[[Page 39089]]

the 2012-based IRF market basket (80 FR 47068). Since the relative 
importance for Capital is 8.6 percent of the 2016-based IRF market 
basket in FY 2020, we took 46 percent of 8.6 percent to determine the 
labor-related share of Capital for FY 2020 of 4.0 percent. Therefore, 
we are finalizing a total labor-related share for FY 2020 of 72.7 
percent (the sum of 68.7 percent for the operating costs and 4.0 
percent for the labor-related share of Capital).
    Table 12 shows the FY 2020 labor-related share using the final 
2016-based IRF market basket relative importance and the FY 2019 labor-
related share which was based on the 2012-based IRF market basket 
relative importance.
[GRAPHIC] [TIFF OMITTED] TR08AU19.015

F. Update to the IRF Wage Index To Use Concurrent IPPS Wage Index 
Beginning With FY 2020

1. Background
    Section 1886(j)(6) of the Act requires the Secretary to adjust the 
proportion of rehabilitation facilities' costs attributable to wages 
and wage-related costs (as estimated by the Secretary from time to 
time) by a factor (established by the Secretary) reflecting the 
relative hospital wage level in the geographic area of the 
rehabilitation facility compared to the national average wage level for 
those facilities. The Secretary is required to update the IRF PPS wage 
index on the basis of information available to the Secretary on the 
wages and wage-related costs to furnish rehabilitation services. Any 
adjustment or updates made under section 1886(j)(6) of the Act for a FY 
are made in a budget-neutral manner.
2. Update to the IRF Wage Index To Use Concurrent IPPS Wage Index 
Beginning with FY 2020
    When the IRF PPS was implemented in the FY 2002 IRF PPS final rule 
(66 FR 41358), we finalized the use of the FY IPPS wage data in the 
creation of an IRF wage index. We believed that a wage index based on 
FY IPPS wage data was the best proxy and most appropriate wage index to 
use in adjusting payments to IRFs, since both IPPS hospitals and IRFs 
compete in the same labor markets. For this reason, we believed, and 
continue to believe, that the wage data of IPPS hospitals accurately 
captures the relationship of wages and wage-related costs of IRFs in an 
area as compared with the national average. Therefore, in the FY 2002 
IRF PPS final rule, we finalized use of the FY 1997 IPPS wage data to 
develop the wage index for the IRF PPS, as that was the most recent 
final data available.
    For all subsequent years in which the IRF PPS wage index has been 
updated, we have continued to use the most recent final IPPS data 
available, which has led us to use the pre-floor, pre-reclassified FY 
IPPS wage index values from the prior fiscal year.
    In the FY 2018 IRF PPS proposed rule (82 FR 20742 through 20743), 
we included a request for information (RFI) to solicit comments from 
stakeholders requesting information on CMS flexibilities and 
efficiencies. The purpose of the RFI was to receive feedback regarding 
ways in which we could reduce burden for hospitals and physicians, 
improve quality of care, decrease costs and ensure that patients 
receive the best care. We received comments from IRF industry 
associations, state and national hospital associations, industry 
groups, representing hospitals, and individual IRF providers in 
response to the solicitation. One of the responses we received to the 
RFI suggested that there is concern among IRF stakeholders about the 
different wage index data used in the different post-acute care (PAC) 
settings. For the IRF PPS, we use a 1-year lag of the pre-floor, pre-
reclassified FY IPPS wage index, meaning that for the IRF PPS for FY 
2019, we finalized use of the FY 2018 IPPS wage index (83 FR 38527). 
However, we base the wage indexes for the SNF PPS and the LTCH PPS on 
the concurrent IPPS wage index ((83 FR 39172 through 39178) and (83 FR 
41731), respectively).
    As we look towards a more unified PACpayment system, we believe 
that standardizing the wage index data across PAC settings is 
necessary. Therefore, we proposed to change the IRF wage index 
methodology to align with other PAC settings. Specifically, we proposed 
changing from our established policy of using the pre-floor, pre-
reclassified FY IPPS wage index (that is, for FY 2020 we proposed to 
use the concurrent FY 2020 pre-floor, pre-reclassified IPPS wage index 
under the IRF PPS). This proposed change would use the concurrent IPPS 
pre-floor, pre-reclassified wage index for the IRF wage index beginning 
with FY 2020 and continuing for all subsequent years. Thus, for the FY 
2020 IRF wage index, we proposed to use the FY 2020 pre-floor, pre-
reclassified IPPS wage index, which is based on data submitted for 
hospital cost reporting periods beginning in FY 2016. We proposed to 
implement these revisions in a budget neutral manner. For more 
information

[[Page 39090]]

on the distributional impacts of this proposal, we refer readers to the 
FY 2020 IRF PPS proposed rule (84 FR 17278).
    Using the current pre-floor, pre-reclassified FY IPPS wage index 
would result in the most up-to-date wage data being the basis for the 
IRF wage index. It would also result in more consistency and equity in 
the wage index methodology used by Medicare.
    We received 7 comments on this proposal to align the data 
timeframes with that of the IPPS by using the FY 2020 pre-floor, pre-
reclassified FY IPPS wage index as the basis for the FY 2020 IRF wage 
index, which are summarized below.
    Comment: All of the commenters supported CMS' proposal to use the 
FY 2020 pre-floor, pre-reclassified FY IPPS wage index for the FY 2020 
IRF wage index. Commenters agreed that the proposed change to use the 
concurrent FY IPPS wage index data would align the wage index data 
across PAC settings and move in the direction of unified PAC payment. A 
few commenters recommended that CMS adopt other wage index policies for 
IRFs that apply to or have been proposed for IPPS hospitals, such as 
geographic reclassifications, suggesting that this would increase 
consistency and alignment across settings.
    Response: We appreciate the commenter's support for the proposal. 
We agree that finalizing this proposal is necessary as we move towards 
a more unified PAC payment system. We plan to monitor the use of the 
concurrent FY IPPS wage index data before we consider any other 
potential wage index policy changes.
    After careful consideration of the comments we received, we are 
finalizing our proposal to align the data timeframes with that of the 
IPPS by using the concurrent pre-floor, pre-reclassified IPPS wage 
index for the IRF wage index beginning with FY 2020 and continuing for 
all subsequent years. Thus, we will use the FY 2020 pre-floor, pre-
reclassified IPPS wage index as the basis for the FY 2020 IRF wage 
index (that is, for all IRF discharges beginning on or after October 1, 
2019). We will implement these revisions in a budget neutral manner. We 
refer readers to Table 20 in section XIII.C of this final rule for more 
information on the distributional effects of this change.
3. Wage Adjustment for FY 2020 Using Concurrent IPPS Wage Index Labor 
Market Area Definitions and the
    Due to our proposal to use the concurrent IPPS wage index beginning 
with FY 2020, for FY 2020, we proposed using the policy and 
methodologies described in section VI. of this final rule related to 
the labor market area definitions and the wage index methodology for 
areas with wage data. Thus, we proposed using the CBSA labor market 
area definitions and the FY 2020 pre-reclassification and pre-floor 
IPPS wage index data. In accordance with section 1886(d)(3)(E) of the 
Act, the FY 2020 pre-reclassification and pre-floor IPPS wage index is 
based on data submitted for hospital cost reporting periods beginning 
on or after October 1, 2015 and before October 1, 2016 (that is, FY 
2016 cost report data).
    The labor market designations made by the OMB include some 
geographic areas where there are no hospitals and, thus, no hospital 
wage index data on which to base the calculation of the IRF PPS wage 
index. We proposed to continue to use the same methodology discussed in 
the FY 2008 IRF PPS final rule (72 FR 44299) to address those 
geographic areas where there are no hospitals and, thus, no hospital 
wage index data on which to base the calculation for the FY 2020 IRF 
PPS wage index.
    We received one comment on this proposal, which is summarized 
below.
    Comment: One commenter requested that, until a new wage index 
system is implemented, CMS should establish a smoothing variable to be 
applied to the current IRF wage index to reduce the fluctuations IRFs 
experience annually.
    Response: Under section 1886(j)(6) of the Act, we adjust IRF PPS 
rates to account for differences in area wage levels. Any perceived 
volatility in the wage index is predicated upon volatility in actual 
wages in that area and reflects real differences in area wage levels. 
As we believe that the application of a smoothing variable would make 
the wage index values less reflective of the area wage levels, we do 
not believe it would be appropriate to implement such a change to the 
IRF wage index policy.
    After careful consideration of the comments we received, we are 
finalizing our proposal to use the policy and methodologies described 
in section VI. of this final rule related to the labor market area 
definitions and the wage index methodology for areas with wage data. 
Thus, we are finalizing the use of the CBSA labor market area 
definitions and the FY 2020 pre-reclassification and pre-floor IPPS 
wage index data. We are finalizing the continued use of the same 
methodology discussed in the FY 2008 IRF PPS final rule (72 FR 44299) 
to address those geographic areas where there are no hospitals and, 
thus, no hospital wage index data on which to base the calculation for 
the FY 2020 IRF PPS wage index.
4. Core-Based Statistical Areas (CBSAs) for the FY 2020 IRF Wage Index
    The wage index used for the IRF PPS is calculated using the pre-
reclassification and pre-floor IPPS wage index data and is assigned to 
the IRF on the basis of the labor market area in which the IRF is 
geographically located. IRF labor market areas are delineated based on 
the CBSAs established by the OMB. The current CBSA delineations (which 
were implemented for the IRF PPS beginning with FY 2016) are based on 
revised OMB delineations issued on February 28, 2013, in OMB Bulletin 
No. 13-01. OMB Bulletin No. 13-01 established revised delineations for 
Metropolitan Statistical Areas, Micropolitan Statistical Areas, and 
Combined Statistical Areas in the United States and Puerto Rico based 
on the 2010 Census, and provided guidance on the use of the 
delineations of these statistical areas using standards published in 
the June 28, 2010 Federal Register (75 FR 37246 through 37252). We 
refer readers to the FY 2016 IRF PPS final rule (80 FR 47068 through 
47076) for a full discussion of our implementation of the OMB labor 
market area delineations beginning with the FY 2016 wage index.
    Generally, OMB issues major revisions to statistical areas every 10 
years, based on the results of the decennial census. However, OMB 
occasionally issues minor updates and revisions to statistical areas in 
the years between the decennial censuses. On July 15, 2015, OMB issued 
OMB Bulletin No. 15-01, which provides minor updates to and supersedes 
OMB Bulletin No. 13-01 that was issued on February 28, 2013. The 
attachment to OMB Bulletin No. 15-01 provides detailed information on 
the update to statistical areas since February 28, 2013. The updates 
provided in OMB Bulletin No. 15-01 are based on the application of the 
2010 Standards for Delineating Metropolitan and Micropolitan 
Statistical Areas to Census Bureau population estimates for July 1, 
2012 and July 1, 2013.
    In the FY 2018 IRF PPS final rule (82 FR 36250 through 36251), we 
adopted the updates set forth in OMB Bulletin No. 15-01 effective 
October 1, 2017, beginning with the FY 2018 IRF wage index. For a 
complete discussion of the adoption of the updates set forth in OMB 
Bulletin No. 15-01, we refer readers to the FY 2018 IRF PPS final rule. 
In the FY 2019 IRF PPS final rule (83 FR 38527), we continued to use 
the OMB delineations that were adopted

[[Page 39091]]

beginning with FY 2016 to calculate the area wage indexes, with updates 
set forth in OMB Bulletin No. 15-01 that we adopted beginning with the 
FY 2018 wage index.
    On August 15, 2017, OMB issued OMB Bulletin No. 17-01, which 
provided updates to and superseded OMB Bulletin No. 15-01 that was 
issued on July 15, 2015. The attachments to OMB Bulletin No. 17-01 
provide detailed information on the update to statistical areas since 
July 15, 2015, and are based on the application of the 2010 Standards 
for Delineating Metropolitan and Micropolitan Statistical Areas to 
Census Bureau population estimates for July 1, 2014 and July 1, 2015. 
In OMB Bulletin No. 17-01, OMB announced that one Micropolitan 
Statistical Area now qualifies as a Metropolitan Statistical Area. The 
new urban CBSA is as follows:
     Twin Falls, Idaho (CBSA 46300). This CBSA is comprised of 
the principal city of Twin Falls, Idaho in Jerome County, Idaho and 
Twin Falls County, Idaho.
    The OMB bulletin is available on the OMB website at https://www.whitehouse.gov/sites/whitehouse.gov/files/omb/bulletins/2017/b-17-01.pdf.
    As we indicated in the FY 2019 IRF PPS final rule (83 FR 38528), we 
believe that it is important for the IRF PPS to use the latest labor 
market area delineations available as soon as is reasonably possible to 
maintain a more accurate and up-to-date payment system that reflects 
the reality of population shifts and labor market conditions. As 
discussed in the FY 2019 IPPS and LTCH PPS final rule (83 FR 20591), 
these updated labor market area definitions were implemented under the 
IPPS beginning on October 1, 2018. Therefore, we proposed to implement 
these revisions for the IRF PPS beginning October 1, 2019, consistent 
with our historical practice of modeling IRF PPS adoption of the labor 
market area delineations after IPPS adoption of these delineations.
    We received 2 comments on this proposal, which are summarized 
below.
    Comment: Commenters expressed concern that the IRF wage index 
values published in the FY 2020 IRF PPS proposed rule were not 
consistent with the values published in the FY 2020 IPPS proposed rule 
wage index public use file. These commenters suggested that CMS examine 
these wage index values and correct them if we find that they are in 
error prior to finalizing the use of the concurrent IPPS wage index 
data for the IRF PPS.
    Response: We identified a slight error in the proposed rule wage 
index values after the FY 2020 IRF PPS proposed rule was published. A 
programming error caused the data for all providers in a single county 
to be included twice, which affected the national average hourly rate, 
and therefore, affected nearly all wage index values. We have corrected 
the programming logic so this error cannot occur again. We also 
standardized our procedures for rounding, to ensure consistency. The 
correction to the proposed rule wage index data was not completed until 
after the comment period closed on June 17, 2019. This final rule 
reflects the corrected and updated wage index data.
    We are finalizing and implementing these revisions for the IRF PPS 
beginning October 1, 2019, consistent with our historical practice of 
modeling IRF PPS adoption of the labor market area delineations after 
IPPS adoption of these delineations.
5. Wage Adjustment
    The FY 2020 wage index tables (which, as discussed in section VI.F 
above, we base on the FY 2020 pre-reclassified, pre-floor FY 2020 IPPS 
wage index) are available on the CMS website at https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/InpatientRehabFacPPS/IRF-Rules-and-Related-Files.html. Table A is for urban areas, and Table B 
is for rural areas.
    To calculate the wage-adjusted facility payment for the payment 
rates set forth in this final rule, we would multiply the unadjusted 
federal payment rate for IRFs by the FY 2020 labor-related share based 
on the 2016-based IRF market basket (72.7 percent) to determine the 
labor-related portion of the standard payment amount. A full discussion 
of the calculation of the labor-related share is located in section 
VI.E of this final rule. We would then multiply the labor-related 
portion by the applicable IRF wage index from the tables in the 
addendum to this final rule. These tables are available on the CMS 
website at https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/InpatientRehabFacPPS/IRF-Rules-and-Related-Files.html. 
Adjustments or updates to the IRF wage index made under section 
1886(j)(6) of the Act must be made in a budget-neutral manner. We 
proposed to calculate a budget-neutral wage adjustment factor as 
established in the FY 2004 IRF PPS final rule (68 FR 45689), codified 
at Sec.  412.624(e)(1), as described in the steps below. We proposed to 
use the listed steps to ensure that the FY 2020 IRF standard payment 
conversion factor reflects the updates to the IRF wage index (based on 
the FY 2020 IPPS wage index) and the labor-related share in a budget-
neutral manner:
    Step 1. Determine the total amount of the estimated FY 2019 IRF PPS 
payments, using the FY 2019 standard payment conversion factor and the 
labor-related share and the wage indexes from FY 2019 (as published in 
the FY 2019 IRF PPS final rule (83 FR 38514)).
    Step 2. Calculate the total amount of estimated IRF PPS payments 
using the FY 2020 standard payment conversion factor and the FY 2020 
labor-related share and CBSA urban and rural wage indexes.
    Step 3. Divide the amount calculated in step 1 by the amount 
calculated in step 2. The resulting quotient is the FY 2020 budget-
neutral wage adjustment factor of 1.0076.
    Step 4. Apply the FY 2020 budget-neutral wage adjustment factor 
from step 3 to the FY 2020 IRF PPS standard payment conversion factor 
after the application of the increase factor to determine the FY 2020 
standard payment conversion factor.
    We note that we have updated our data between the FY 2020 IRF PPS 
proposed and final rules to ensure that we use the most recent 
available data in calculating IRF PPS payments. This updated data 
includes a more complete set of claims for FY 2018 and updated wage 
index data. Based on our analysis using this updated data, we now 
estimate a budget-neutral wage adjustment factor of 1.0031 for FY 2020.
    We discuss the calculation of the standard payment conversion 
factor for FY 2020 in section VI.H. of this final rule.
    We invited public comments on this proposal. However, we did not 
receive any comments on the proposed methodology for calculating the 
budget-neutral wage adjustment factor.
    As we did not receive any comments on the proposed methodology for 
calculating the budget-neutral wage adjustment factor, we are 
finalizing this policy as proposed for FY 2020.

G. Wage Index Comment Solicitation

    Historically, we have calculated the IRF wage index values using 
unadjusted wage index values from another provider setting. 
Stakeholders have frequently commented on certain aspects of the IRF 
wage index values and their impact on payments. Therefore, we solicited 
public comments in the FY 2020 IRF PPS proposed rule (84 FR 17280) on 
concerns stakeholders may have regarding the wage index used to adjust

[[Page 39092]]

IRF payments and suggestions for possible updates and improvements to 
the geographic adjustment of IRF payments.
    We appreciate the commenters' responses to this solicitation and 
will take them into consideration for possible future policy 
development.

H. Description of the IRF Standard Payment Conversion Factor and 
Payment Rates for FY 2020

    To calculate the standard payment conversion factor for FY 2020, as 
illustrated in Table 13, we begin by applying the increase factor for 
FY 2020, as adjusted in accordance with sections 1886(j)(3)(C) of the 
Act, to the standard payment conversion factor for FY 2019 ($16,021). 
Applying the 2.5 percent increase factor for FY 2020 to the standard 
payment conversion factor for FY 2019 of $16,021 yields a standard 
payment amount of $16,422. Then, we apply the budget neutrality factor 
for the FY 2020 wage index and labor-related share of 1.0031, which 
results in a standard payment amount of $16,472. We next apply the 
budget neutrality factor for the revised CMGs and CMG relative weights 
of 1.0010, which results in the standard payment conversion factor of 
$16,489 for FY 2020.
[GRAPHIC] [TIFF OMITTED] TR08AU19.016

    We received one comment on the proposed FY 2020 standard payment 
conversion factor, which is summarized below.
    Comment: One commenter stated that the proposed rate update fails 
to cover the cost of medical inflation or payment reductions due to 
sequestration. As a result, this commenter expressed concern that their 
hospitals' financial viability and their ability to care for their 
patients will be threatened.
    Response: We appreciate this commenter's concerns. However, we note 
that the IRF PPS payment rates are updated annually by an increase 
factor that reflects changes over time in the prices of an appropriate 
mix of goods and services included in the covered IRF services, as 
required by section 1886(j)(3)(C) of the Act.
    After careful consideration of the comment we received, we are 
finalizing the IRF standard payment conversion factor of $16,489 for FY 
2020.
    After the application of the CMG relative weights described in 
section IV. of this final rule to the FY 2020 standard payment 
conversion factor ($16,489), the resulting unadjusted IRF prospective 
payment rates for FY 2020 are shown in Table 14.
BILLING CODE 4120-01-P

[[Page 39093]]

[GRAPHIC] [TIFF OMITTED] TR08AU19.017


[[Page 39094]]


[GRAPHIC] [TIFF OMITTED] TR08AU19.018

BILLING CODE 4120-01-C

H. Example of the Methodology for Adjusting the Prospective Payment 
Rates

    Table 15 illustrates the methodology for adjusting the prospective 
payments (as described in section VI. of this final rule). The 
following examples are based on two hypothetical Medicare 
beneficiaries, both classified into CMG 0104 (without comorbidities). 
The unadjusted prospective payment rate for

[[Page 39095]]

CMG 0104 (without comorbidities) appears in Table 14.
    Example: One beneficiary is in Facility A, an IRF located in rural 
Spencer County, Indiana, and another beneficiary is in Facility B, an 
IRF located in urban Harrison County, Indiana. Facility A, a rural non-
teaching hospital has a Disproportionate Share Hospital (DSH) 
percentage of 5 percent (which would result in a LIP adjustment of 
1.0156), a wage index of 0.8319, and a rural adjustment of 14.9 
percent. Facility B, an urban teaching hospital, has a DSH percentage 
of 15 percent (which would result in a LIP adjustment of 1.0454 
percent), a wage index of 0.8844, and a teaching status adjustment of 
0.0784.
    To calculate each IRF's labor and non-labor portion of the 
prospective payment, we begin by taking the unadjusted prospective 
payment rate for CMG 0104 (without comorbidities) from Table 14. Then, 
we multiply the labor-related share for FY 2020 (72.7 percent) 
described in section VI.E. of this final rule by the unadjusted 
prospective payment rate. To determine the non-labor portion of the 
prospective payment rate, we subtract the labor portion of the federal 
payment from the unadjusted prospective payment.
    To compute the wage-adjusted prospective payment, we multiply the 
labor portion of the federal payment by the appropriate wage index 
located in Tables A and B. These tables are available on the CMS 
website at https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/InpatientRehabFacPPS/IRF-Rules-and-Related-Files.html.
    The resulting figure is the wage-adjusted labor amount. Next, we 
compute the wage-adjusted federal payment by adding the wage-adjusted 
labor amount to the non-labor portion of the federal payment.
    Adjusting the wage-adjusted federal payment by the facility-level 
adjustments involves several steps. First, we take the wage-adjusted 
prospective payment and multiply it by the appropriate rural and LIP 
adjustments (if applicable). Second, to determine the appropriate 
amount of additional payment for the teaching status adjustment (if 
applicable), we multiply the teaching status adjustment (0.0784, in 
this example) by the wage-adjusted and rural-adjusted amount (if 
applicable). Finally, we add the additional teaching status payments 
(if applicable) to the wage, rural, and LIP-adjusted prospective 
payment rates. Table 15 illustrates the components of the adjusted 
payment calculation.
[GRAPHIC] [TIFF OMITTED] TR08AU19.019

    Thus, the adjusted payment for Facility A would be $28,327.82, and 
the adjusted payment for Facility B would be $28,467.16.

VII. Update to Payments for High-Cost Outliers Under the IRF PPS for FY 
2020

A. Update to the Outlier Threshold Amount for FY 2020

    Section 1886(j)(4) of the Act provides the Secretary with the 
authority to make payments in addition to the basic IRF prospective 
payments for cases incurring extraordinarily high costs. A case 
qualifies for an outlier payment if the estimated cost of the case 
exceeds the adjusted outlier threshold. We calculate the adjusted 
outlier threshold by adding the IRF PPS payment for the case (that is, 
the CMG payment adjusted by all of the relevant facility-level 
adjustments) and the adjusted threshold amount (also adjusted by all of 
the relevant facility-level adjustments). Then, we calculate the 
estimated cost of a case by multiplying the IRF's overall CCR by the 
Medicare allowable covered charge. If the estimated cost of the case is 
higher than the adjusted outlier threshold, we make an outlier payment 
for the case equal to 80 percent of the difference between the 
estimated cost of the case and the outlier threshold.
    In the FY 2002 IRF PPS final rule (66 FR 41362 through 41363), we 
discussed our rationale for setting the outlier threshold amount for 
the IRF PPS so that estimated outlier payments would equal 3 percent of 
total estimated payments. For the 2002 IRF PPS final rule, we analyzed 
various outlier policies using 3, 4, and 5 percent of the total 
estimated payments, and we concluded that an outlier policy set at 3 
percent of total estimated payments would optimize the extent to which 
we could reduce the financial risk to IRFs

[[Page 39096]]

of caring for high-cost patients, while still providing for adequate 
payments for all other (non-high cost outlier) cases.
    Subsequently, we updated the IRF outlier threshold amount in the 
FYs 2006 through 2019 IRF PPS final rules and the FY 2011 and FY 2013 
notices (70 FR 47880, 71 FR 48354, 72 FR 44284, 73 FR 46370, 74 FR 
39762, 75 FR 42836, 76 FR 47836, 76 FR 59256, 77 FR 44618, 78 FR 47860, 
79 FR 45872, 80 FR 47036, 81 FR 52056, 82 FR 36238, and 83 FR 38514, 
respectively) to maintain estimated outlier payments at 3 percent of 
total estimated payments. We also stated in the FY 2009 final rule (73 
FR 46370 at 46385) that we would continue to analyze the estimated 
outlier payments for subsequent years and adjust the outlier threshold 
amount as appropriate to maintain the 3 percent target.
    To update the IRF outlier threshold amount for FY 2020, we proposed 
to use FY 2018 claims data and the same methodology that we used to set 
the initial outlier threshold amount in the FY 2002 IRF PPS final rule 
(66 FR 41316 and 41362 through 41363), which is also the same 
methodology that we used to update the outlier threshold amounts for 
FYs 2006 through 2019. The outlier threshold is calculated by 
simulating aggregate payments and using an iterative process to 
determine a threshold that results in outlier payments being equal to 3 
percent of total payments under the simulation. To determine the 
outlier threshold for FY 2020, we estimate the amount of FY 2020 IRF 
PPS aggregate and outlier payments using the most recent claims 
available (FY 2018) and the FY 2020 standard payment conversion factor, 
labor-related share, and wage indexes, incorporating any applicable 
budget-neutrality adjustment factors. The outlier threshold is adjusted 
either up or down in this simulation until the estimated outlier 
payments equal 3 percent of the estimated aggregate payments. Based on 
an analysis of the preliminary data used for the proposed rule, we 
estimated that IRF outlier payments as a percentage of total estimated 
payments would be approximately 3.2 percent in FY 2019. Therefore, we 
proposed to update the outlier threshold amount from $9,402 for FY 2019 
to $9,935 for FY 2020 to maintain estimated outlier payments at 
approximately 3 percent of total estimated aggregate IRF payments for 
FY 2020.
    We note that, as we typically do, we updated our data between the 
FY 2020 IRF PPS proposed and final rules to ensure that we use the most 
recent available data in calculating IRF PPS payments. This updated 
data includes a more complete set of claims for FY 2018. Based on our 
analysis using this updated data, we now estimate that IRF outlier 
payments as a percentage of total estimated payments are approximately 
3.0 percent in FY 2019. Although our analysis shows that we achieved 
our goal to have estimated outlier payments equal 3.0 percent of total 
estimated aggregate IRF payments for FY 2019, we still need to adjust 
the IRF outlier threshold to reflect changes in estimated costs and 
payments for IRFs in FY 2020. That is, as discussed in section VI. of 
this final rule, we are finalizing our proposal to increase IRF PPS 
payment rates by 2.5 percent, in accordance with section 1886(j)(3)(C) 
of the Act to account for changes over time in the prices of an 
appropriate mix of goods and services included in the covered IRF 
services. Similarly, we estimate costs for IRFs in FY 2020 are expected 
to increase to account for changes over time in the prices of goods and 
services included in the covered IRF services. Therefore, we will 
update the outlier threshold amount from $9,402 for FY 2019 to $9,300 
for FY 2020 to account for the increases in IRF PPS payments and 
estimated costs and to maintain estimated outlier payments at 
approximately 3 percent of total estimated aggregate IRF payments for 
FY 2020.
    We received three comments on the proposed update to the FY 2020 
outlier threshold, which are summarized below.
    Comment: Commenters suggested that historical outlier 
reconciliation dollars should be included in the calculation of the 
fixed loss threshold under the IRF PPS.
    Response: As we did not propose a change to the methodology used to 
establish an outlier threshold for IRF PPS payments, these comments are 
outside the scope of this rule. However, we will continue to monitor 
our IRF outlier policies to ensure that they continue to compensate 
IRFs appropriately for treating unusually high-cost patients and do not 
limit access to care for patients who are likely to require unusually 
high-cost care.
    Comment: A few commenters suggested that CMS consider implementing 
a cap on the amount of outlier payments an individual IRF can receive 
under the IRF PPS. One commenter was supportive of maintaining 
estimated payments for outlier payments at approximately 3 percent 
while other commenters expressed concern with maintaining the 3 percent 
target and suggested reducing the outlier pool below 3 percent.
    Response: As we did not propose to implement a cap on the amount of 
outlier payments an individual IRF can receive under the IRF PPS, these 
comments are outside the scope of this rule. However, we note that any 
future consideration given to imposing a limit on outlier payments 
would have to carefully analyze and take into consideration the effect 
on access to IRF care for certain high-cost populations.
    As most recently discussed in the FY 2019 IRF PPS final rule (83 FR 
38532), we analyzed various outlier policies using 3, 4, and 5 percent 
of the total estimated payments for the FY 2002 IRF PPS final rule, and 
we concluded that an outlier policy set at 3 percent of total estimated 
payments would optimize the extent to which we could reduce the 
financial risk to IRFs of caring for high-cost patients, while still 
providing for adequate payments for all other (non-high cost outlier) 
cases. We continue to believe that the outlier policy of 3 percent of 
total estimated aggregate payments accomplishes this objective. We 
refer readers to the FY 2002 IRF PPS final rule (66 FR 41316, 41362 
through 41363) for more information regarding the rationale for setting 
the outlier threshold amount for the IRF PPS so that estimated outlier 
payments would equal 3 percent of total estimated payments.
    Comment: One commenter requested that CMS update the outlier 
threshold amount in the final rule using the latest available data.
    Response: We agree that we should use the most recent data 
available to calculate the outlier threshold. Therefore, as previously 
stated, we updated the data used to calculate the outlier threshold 
between the FY 2020 IRF PPS proposed and final rules.
    Having carefully considered the public comments received and also 
taking into account the most recent available data, we are finalizing 
the outlier threshold amount of $9,300 to maintain estimated outlier 
payments at approximately 3 percent of total estimated aggregate IRF 
payments for FY 2020.

B. Update to the IRF Cost-to-Charge Ratio Ceiling and Urban/Rural 
Averages for FY 2020

    Cost-to-charge ratios are used to adjust charges from Medicare 
claims to costs and are computed annually from facility-specific data 
obtained from Medicare cost reports. IRF specific cost-to-charge ratios 
are used in the development of the CMG relative weights and the 
calculation of outlier

[[Page 39097]]

payments under the IRF prospective payment system. In accordance with 
the methodology stated in the FY 2004 IRF PPS final rule (68 FR 45674, 
45692 through 45694), we proposed to apply a ceiling to IRFs' CCRs. 
Using the methodology described in that final rule, we proposed to 
update the national urban and rural CCRs for IRFs, as well as the 
national CCR ceiling for FY 2020, based on analysis of the most recent 
data that is available. We apply the national urban and rural CCRs in 
the following situations:
     New IRFs that have not yet submitted their first Medicare 
cost report.
     IRFs whose overall CCR is in excess of the national CCR 
ceiling for FY 2020, as discussed below in this section.
     Other IRFs for which accurate data to calculate an overall 
CCR are not available.
    Specifically, for FY 2020, we proposed to estimate a national 
average CCR of 0.500 for rural IRFs, which we calculated by taking an 
average of the CCRs for all rural IRFs using their most recently 
submitted cost report data. Similarly, we proposed to estimate a 
national average CCR of 0.406 for urban IRFs, which we calculated by 
taking an average of the CCRs for all urban IRFs using their most 
recently submitted cost report data. We apply weights to both of these 
averages using the IRFs' estimated costs, meaning that the CCRs of IRFs 
with higher total costs factor more heavily into the averages than the 
CCRs of IRFs with lower total costs. For this final rule, we have used 
the most recent available cost report data (FY 2017). This includes all 
IRFs whose cost reporting periods begin on or after October 1, 2016, 
and before October 1, 2017. If, for any IRF, the FY 2017 cost report 
was missing or had an ``as submitted'' status, we used data from a 
previous fiscal year's (that is, FY 2004 through FY 2016) settled cost 
report for that IRF. We do not use cost report data from before FY 2004 
for any IRF because changes in IRF utilization since FY 2004 resulting 
from the 60 percent rule and IRF medical review activities suggest that 
these older data do not adequately reflect the current cost of care. 
Using updated FY 2017 cost report data for this final rule, we estimate 
a national average CCR of 0.500 for rural IRFs, and a national average 
CCR of 0.405 for urban IRFs.
    In accordance with past practice, we proposed to set the national 
CCR ceiling at 3 standard deviations above the mean CCR. Using this 
method, we proposed a national CCR ceiling of 1.31 for FY 2020. This 
means that, if an individual IRF's CCR were to exceed this ceiling of 
1.31 for FY 2020, we would replace the IRF's CCR with the appropriate 
proposed national average CCR (either rural or urban, depending on the 
geographic location of the IRF). We calculated the proposed national 
CCR ceiling by:
    Step 1. Taking the national average CCR (weighted by each IRF's 
total costs, as previously discussed) of all IRFs for which we have 
sufficient cost report data (both rural and urban IRFs combined).
    Step 2. Estimating the standard deviation of the national average 
CCR computed in step 1.
    Step 3. Multiplying the standard deviation of the national average 
CCR computed in step 2 by a factor of 3 to compute a statistically 
significant reliable ceiling.
    Step 4. Adding the result from step 3 to the national average CCR 
of all IRFs for which we have sufficient cost report data, from step 1.
    Using the updated FY 2017 cost report data for this final rule, we 
estimate a national average CCR ceiling of 1.31, using the same 
methodology.
    We did not receive comments on the proposed update to the IRF CCR 
ceiling and the urban/rural averages for FY 2020.
    As we did not receive any comments on the proposed update to the 
IRF CCR ceiling and the urban/rural averages for FY 2020, we are 
finalizing the national average urban CCR at 0.405, the national 
average rural CCR at 0.500, and the national average CCR ceiling at 
1.31 for FY 2020.

VIII. Amendments to Sec.  412.622 To Clarify the Definition of a 
Rehabilitation Physician

    Under Sec.  412.622(a)(3)(iv), a rehabilitation physician is 
defined as ``a licensed physician with specialized training and 
experience in inpatient rehabilitation.'' The term rehabilitation 
physician is used in several other places in Sec.  412.622, with 
corresponding references to Sec.  412.622(a)(3)(iv). The definition at 
Sec.  412.622(a)(3)(iv) does not specify the level or type of training 
and experience required for a licensed physician to be designated as a 
rehabilitation physician because we believe that the IRFs are in the 
best position to make this determination for purposes of Sec.  412.622.
    Therefore, we proposed to amend the definition of a rehabilitation 
physician to clarify that the determination as to whether a physician 
qualifies as a rehabilitation physician (that is, a licensed physician 
with specialized training and experience in inpatient rehabilitation) 
is made by the IRF (84 FR 17284 through 17285). For clarity, we also 
proposed to remove this definition from Sec.  412.622(a)(3)(iv) and 
move it to a new paragraph (Sec.  412.622(c)). We also proposed to make 
corresponding technical corrections elsewhere in Sec.  
412.622(a)(3)(iv), (a)(4)(i)(A), (a)(4)(iii)(A), and (a)(5)(i) to 
remove the references to Sec.  412.622(a)(3)(iv) in those paragraphs, 
so as to reflect the new location of the definition.
    We received 1,163 comments on the proposal to clarify the 
definition of a rehabilitation physician, to move the definition from 
Sec.  412.622(a)(3)(iv) to Sec.  412.622(c), and to make corresponding 
technical corrections elsewhere in Sec.  412.622 to remove references 
to the current location of the definition in Sec.  412.622(a)(3)(iv). 
The majority of these comments consisted of form letters, in which we 
received multiple copies of two types of identically-worded letters 
that had been signed and submitted by different individuals. The 
comments we received on this are summarized below.
    Comment: Many of the commenters noted appreciation and support for 
the proposal to amend the definition of a rehabilitation physician to 
clarify that the determination as to whether a physician qualifies as a 
rehabilitation physician (that is, a licensed physician with 
specialized training and experience in inpatient rehabilitation) is 
made by the IRF. One commenter stated that while board-certified 
physiatrists play a crucial caregiver and leadership role in 
rehabilitation hospitals, they are not alone in doing so. Physicians 
representing other specialties can and do also display the leadership 
and caregiving skills and experience that clearly qualify them as a 
rehabilitation physician. One commenter indicated that CMS' proposal is 
consistent with CMS' previously stated position from 2010. Some 
commenters also stated that clarifying the regulation would reduce the 
number of claims denials by promoting a shared understanding of the 
requirements between IRFs and Medicare contractors.
    Response: We appreciate the commenters' support and agree that this 
clarification in our regulations supports our longstanding position 
that the responsibility is, and always has been, on the IRF to ensure 
that the rehabilitation physician(s) who are making the admission 
decisions and treating the patients have the necessary training and 
experience.
    Comment: Many commenters stated that they do not support CMS' 
proposal and suggested that CMS not finalize the proposed amendments to 
Sec.  412.622. These commenters requested that CMS delay any changes to 
current regulations

[[Page 39098]]

until CMS and stakeholders can work together to develop a consensus 
approach for protecting the quality and integrity of IRF care. These 
commenters stated that they believe that allowing the IRF to determine 
whether an individual physician meets the regulatory standards for a 
rehabilitation physician could increase the risks that some IRFs will 
hire or contract with unqualified or underqualified physicians, reduce 
the quality of care that patients receive in IRFs, and reduce the value 
of physiatrists. These commenters also stated that reducing the value 
of physiatrists could also deter students from wanting to pursue this 
specialty in the future. Some commenters also indicated that CMS' 
proposal, if finalized, would undermine CMS' ability to engage in 
appropriate program integrity oversight by not reviewing an IRF's 
decision to hire a particular physician to fill a rehabilitation 
physician role.
    Response: While we appreciate and share the commenters' desire to 
ensure that Medicare beneficiaries in IRFs receive the highest-quality 
care from trained and qualified physicians, we do not believe that 
merely clarifying our existing policy would reduce quality of care. The 
regulation will continue to require a rehabilitation physician to be a 
licensed physician with specialized training and experience in 
inpatient rehabilitation. We are not lowering these requirements. 
However, we continue to believe that we need to clarify our existing 
policy that the IRF makes the determination as to whether a given 
physician qualifies as a rehabilitation physician in order to eliminate 
any unnecessary uncertainty on this issue. Over the past year, we have 
received questions regarding how this provision can be enforced, and we 
believe that this clarification will promote a shared understanding of 
how we intend the enforcement to occur. We expect that IRFs will 
continue to ensure that the rehabilitation physicians treating patients 
in their facilities have the necessary training and experience in 
inpatient rehabilitation. To this end, we will continue to work with 
stakeholders to refine Medicare's IRF payment policies in the future so 
that they support IRFs in providing the highest quality care to 
beneficiaries.
    After careful consideration of the comments we received, we are 
finalizing our proposal to amend the definition of a rehabilitation 
physician to clarify that the determination as to whether a physician 
qualifies as a rehabilitation physician (that is, a licensed physician 
with specialized training and experience in inpatient rehabilitation) 
is made by the IRF. However, based on the stakeholder feedback, we will 
continue to assess whether future refinements to this policy may be 
needed.
    For clarity, we are also removing this definition from Sec.  
412.622(a)(3)(iv) and moving it to a new paragraph (Sec.  412.622(c)). 
We are also making corresponding technical corrections elsewhere in 
Sec.  412.622(a)(3)(iv), (a)(4)(i)(A), (a)(4)(iii)(A), and (a)(5)(i) to 
remove the references to Sec.  412.622(a)(3)(iv) in those paragraphs, 
so as to reflect the new location of the definition.

IX. Updates to the IRF Quality Reporting Program (QRP)

A. Background

    The IRF QRP is authorized by section 1886(j)(7) of the Act, and it 
applies to freestanding IRFs, as well as inpatient rehabilitation units 
of hospitals or critical access hospitals (CAHs) paid by Medicare under 
the IRF PPS. Under the IRF QRP, the Secretary must reduce the annual 
increase factor for discharges occurring during such fiscal year by 2 
percentage points for any IRF that does not submit data in accordance 
with the requirements established by the Secretary. For more 
information on the background and statutory authority for the IRF QRP, 
we refer readers to the FY 2012 IRF PPS final rule (76 FR 47873 through 
47874), the CY 2013 Hospital Outpatient Prospective Payment System/
Ambulatory Surgical Center (OPPS/ASC) Payment Systems and Quality 
Reporting Programs final rule (77 FR 68500 through 68503), the FY 2014 
IRF PPS final rule (78 FR 47902), the FY 2015 IRF PPS final rule (79 FR 
45908), the FY 2016 IRF PPS final rule (80 FR 47080 through 47083), the 
FY 2017 IRF PPS final rule (81 FR 52080 through 52081), the FY 2018 IRF 
PPS final rule (82 FR 36269 through 36270), and the FY 2019 IRF PPS 
final rule (83 FR 38555 through 38556).
    While we did not solicit comments on previously finalized IRF QRP 
policies, we received comments, which are summarized below.
    Comment: A few commenters stated that the IRF QRP compliance 
threshold of 95 percent for assessment-based items is too high given 
the number of data elements that have been added to the IRF-PAI, and 
requested that CMS lower it to 80 percent in alignment with other 
programs.
    Response: We did not propose any changes to the compliance 
threshold, which has been codified at Sec.  412.634(f). While these 
comments were out of scope for this rule, we will take these comments 
under consideration.

B. General Considerations Used for the Selection of Measures for the 
IRF QRP

    For a detailed discussion of the considerations we use for the 
selection of IRF QRP quality, resource use, and other measures, we 
refer readers to the FY 2016 IRF PPS final rule (80 FR 47083 through 
47084).

C. Quality Measures Currently Adopted for the FY 2021 IRF QRP

    The IRF QRP currently has 15 measures for the FY 2020 program year, 
which are set out in Table 16.

[[Page 39099]]

[GRAPHIC] [TIFF OMITTED] TR08AU19.020

    While we did not solicit comments on currently adopted measures 
(with the exception of the Discharge to Community Measure discussed in 
section IX.D.3 of this rule and the policies regarding public display 
of the Drug Regimen Review Conducted With Follow-Up for Identified 
Issues--PAC IRF QRP in section IX.I of this rule), we received several 
comments.
    Comment: A few commenters had suggestions for removing measures 
they believe were ``topped out'' according to the Hospital Inpatient 
Quality Reporting (IQR) Program definition (83 FR 20408) and did not 
demonstrate variation across facilities, including Application of 
Percent of Residents Experiencing One or More Falls with Major Injury 
(Long Stay) (NQF #0674) and Application of Percent of Long-Term Care 
Hospital Patients with an Admission and Discharge Functional Assessment 
and a Care Plan That Addresses Function (NQF #2631), and Changes in 
Skin Integrity Post-Acute Care: Pressure Ulcer/Injury. One commenter 
had suggestions for improving the training manual for the Drug Regimen 
Review measure in terms of considered clinically significant medication 
issue.
    Response: We did not propose any changes to these previously 
finalized measures, nor did we propose measure removals from the IRF 
QRP. We wish to clarify that the IRF QRP has not adopted the Hospital 
Inpatient Quality Reporting (IQR) definition of ``topped out'' in the 
measure removal criteria finalized for the IRF QRP at Sec.  412.634(2). 
We also note that we do not automatically remove high performing 
measures, and wish to reiterate that such measures may be retained for 
other specified reasons. For example, a particular measure with high 
performance rates may be retained if the measure addresses a topic 
related to quality that is so significant that we do not want to risk a 
decline in quality that could result if we removed the measure, or if 
the measure addresses a topic that is statutorily required. We will 
continue to monitor and evaluate the data from all IRF QRP measures.
    With regard to the commenter's suggestions about the Drug Regimen 
Review measure, we interpret that the commenter is requesting 
additional clarification for coding. We will take these comments into 
account as we develop training materials for the IRF QRP.

D. Adoption of Two New Quality Measures and Updated Specifications for 
a Third Quality Measure Beginning With the FY 2022 IRF QRP

    In the FY 2020 IRF PPS proposed rule (84 FR 17286 through 17291), 
we proposed to adopt two process measures for the IRF QRP that would 
satisfy section 1899B(c)(1)(E)(ii) of the Act, which requires that the 
quality measures specified by the Secretary include measures with 
respect to the quality measure domain titled ``Accurately communicating 
the existence of and providing for the transfer of health information 
and care preferences of an individual to the individual, family 
caregiver of the individual, and providers of services

[[Page 39100]]

furnishing items and services to the individual when the individual 
transitions from a PAC provider to another applicable setting, 
including a different PAC provider, a hospital, a critical access 
hospital, or the home of the individual.'' Given the length of this 
domain title, hereafter, we will refer to this quality measure domain 
as ``Transfer of Health Information.''
    The two measures we proposed to adopt are: (1) Transfer of Health 
Information to the Provider--Post-Acute Care (PAC); and (2) Transfer of 
Health Information to the Patient--Post-Acute Care (PAC). Both of these 
measures support our Meaningful Measures priority of promoting 
effective communication and coordination of care, specifically the 
Meaningful Measure area of the transfer of health information and 
interoperability.
    In addition to the two measure proposals, we proposed to update the 
specifications for the Discharge to Community--Post Acute Care (PAC) 
IRF QRP measure to exclude baseline nursing facility (NF) residents 
from the measure.
    We sought public comment on each of these proposals. These comments 
are summarized after each proposal below.
1. Transfer of Health Information to the Provider--Post-Acute Care 
(PAC) Measure
    The Transfer of Health Information to the Provider--Post-Acute Care 
(PAC) Measure that we proposed to adopt beginning with the FY 2022 IRF 
QRP is a process-based measure that assesses whether or not a current 
reconciled medication list is given to the subsequent provider when a 
patient is discharged or transferred from his or her current PAC 
setting.
a. Background
    In 2013, 22.3 percent of all acute hospital discharges were 
discharged to PAC settings, including 11 percent who were discharged to 
home under the care of a home health agency, and 9 percent who were 
discharged to SNFs.\2\ The proportion of patients being discharged from 
an acute care hospital to a PAC setting was greater among beneficiaries 
enrolled in Medicare FFS. Among Medicare FFS patients discharged from 
an acute hospital, 42 percent went directly to PAC settings. Of that 42 
percent, 20 percent were discharged to a SNF, 18 percent were 
discharged to a home health agency (HHA), 3 percent were discharged to 
an IRF, and 1 percent were discharged to an LTCH.\3\ Of the Medicare 
FFS beneficiaries with an IRF stay in FYs 2016 and 2017, an estimated 
10 percent were discharged or transferred to an acute care hospital, 51 
percent discharged home with home health services, 16 percent 
discharged or transferred to a SNF, and one percent discharged or 
transferred to another PAC setting (for example, another IRF, a 
hospice, or an LTCH).\4\
---------------------------------------------------------------------------

    \2\ Tian, W. ``An all-payer view of hospital discharge to post-
acute care,'' May 2016. Available at https://www.hcup-us.ahrq.gov/reports/statbriefs/sb205-Hospital-Discharge-Postacute-Care.jsp.
    \3\ Ibid.
    \4\ RTI International analysis of Medicare claims data for index 
stays in IRF 2016/2017. (RTI program reference: MM150).
    \5\ Kwan, J.L., Lo, L., Sampson, M., & Shojania, K.G., 
``Medication reconciliation during transitions of care as a patient 
safety strategy: a systematic review,'' Annals of Internal Medicine, 
2013, Vol. 158(5), pp. 397-403.
    \6\ Boockvar, K.S., Blum, S., Kugler, A., Livote, E., 
Mergenhagen, K.A., Nebeker, J.R., & Yeh, J., ``Effect of admission 
medication reconciliation on adverse drug events from admission 
medication changes,'' Archives of Internal Medicine, 2011, Vol. 
171(9), pp. 860-861.
    \7\ Bell, C.M., Brener, S.S., Gunraj, N., Huo, C., Bierman, 
A.S., Scales, D.C., & Urbach, D.R., ``Association of ICU or hospital 
admission with unintentional discontinuation of medications for 
chronic diseases,'' JAMA, 2011, Vol. 306(8), pp. 840-847.
    \8\ Basey, A.J., Krska, J., Kennedy, T.D., & Mackridge, A.J., 
``Prescribing errors on admission to hospital and their potential 
impact: a mixed-methods study,'' BMJ Quality & Safety, 2014, Vol. 
23(1), pp. 17-25.
    \9\ Desai, R., Williams, C.E., Greene, S.B., Pierson, S., & 
Hansen, R.A., ``Medication errors during patient transitions into 
nursing homes: characteristics and association with patient harm,'' 
The American Journal of Geriatric Pharmacotherapy, 2011, Vol. 9(6), 
pp. 413-422.
    \10\ Boling, P.A., ``Care transitions and home health care,'' 
Clinical Geriatric Medicine, 2009, Vol.25(1), pp. 135-48.
    \11\ Barnsteiner, J.H., ``Medication Reconciliation: Transfer of 
medication information across settings--keeping it free from 
error,'' The American Journal of Nursing, 2005, Vol. 105(3), pp. 31-
36.
    \12\ Arbaje, A.I., Kansagara, D.L., Salanitro, A.H., Englander, 
H.L., Kripalani, S., Jencks, S.F., & Lindquist, L.A., ``Regardless 
of age: incorporating principles from geriatric medicine to improve 
care transitions for patients with complex needs,'' Journal of 
General Internal Medicine, 2014, Vol. 29(6), pp. 932-939.
    \13\ Jencks, S.F., Williams, M.V., & Coleman, E.A., 
``Rehospitalizations among patients in the Medicare fee-for-service 
program,'' New England Journal of Medicine, 2009, Vol. 360(14), pp. 
1418-1428.
    \14\ Institute of Medicine. ``Preventing medication errors: 
quality chasm series,'' Washington, DC: The National Academies Press 
2007. Available at https://www.nap.edu/read/11623/chapter/1.
---------------------------------------------------------------------------

    The transfer and/or exchange of health information from one 
provider to another can be done verbally (for example, clinician-to-
clinician communication in-person or by telephone), paper-based (for 
example, faxed or printed copies of records), and via electronic 
communication (for example, through a health information exchange 
network using an electronic health/medical record, and/or secure 
messaging). Health information, such as medication information, that is 
incomplete or missing increases the likelihood of a patient or resident 
safety risk, and is often life-threatening.5 6 7 8 9 10 Poor 
communication and coordination across health care settings contributes 
to patient complications, hospital readmissions, emergency department 
visits, and medication errors.11 12 13 14 15 16 17 18 19 20 
Communication has been cited as the third most frequent root cause in 
sentinel events, which The Joint Commission defines \21\ as a patient 
safety event that results in death, permanent harm, or severe temporary 
harm. Failed or ineffective patient handoffs are estimated to play a 
role in 20 percent of serious preventable adverse events.\22\ When care 
transitions are enhanced through care coordination activities, such as 
expedited patient information flow, these activities can reduce 
duplication of care services and costs of care, resolve conflicting 
care plans, and prevent medical errors.23 24 25 26 27
---------------------------------------------------------------------------

    \15\ Kitson, N.A., Price, M., Lau, F.Y., & Showler, G., 
``Developing a medication communication framework across continuums 
of care using the Circle of Care Modeling approach,'' BMC Health 
Services Research, 2013, Vol. 13(1), pp. 1-10.
    \16\ Mor, V., Intrator, O., Feng, Z., & Grabowski, D.C., ``The 
revolving door of rehospitalization from skilled nursing 
facilities,'' Health Affairs, 2010, Vol. 29(1), pp. 57-64.
    \17\ Institute of Medicine. ``Preventing medication errors: 
quality chasm series,'' Washington, DC: The National Academies Press 
2007. Available at https://www.nap.edu/read/11623/chapter/1.
    \18\ Kitson, N.A., Price, M., Lau, F.Y., & Showler, G., 
``Developing a medication communication framework across continuums 
of care using the Circle of Care Modeling approach,'' BMC Health 
Services Research, 2013, Vol. 13(1), pp. 1-10.
    \19\ Forster, A.J., Murff, H.J., Peterson, J.F., Gandhi, T.K., & 
Bates, D.W., ``The incidence and severity of adverse events 
affecting patients after discharge from the hospital.'' Annals of 
Internal Medicine, 2003,138(3), pp. 161-167.
    \20\ King, B.J., Gilmore-Bykovskyi, A.L., Roiland, R.A., 
Polnaszek, B.E., Bowers, B.J., & Kind, A.J. ``The consequences of 
poor communication during transitions from hospital to skilled 
nursing facility: a qualitative study,'' Journal of the American 
Geriatrics Society, 2013, Vol. 61(7), 1095-1102.
    \21\ The Joint Commission, ``Sentinel Event Policy'' available 
at https://www.jointcommission.org/sentinel_event_policy_and_procedures/.
    \22\ The Joint Commission. ``Sentinel Event Data Root Causes by 
Event Type 2004-2015.'' 2016. Available at https://www.jointcommission.org/assets/1/23/jconline_Mar_2_2016.pdf.
    \23\ Mor, V., Intrator, O., Feng, Z., & Grabowski, D.C., ``The 
revolving door of rehospitalization from skilled nursing 
facilities,'' Health Affairs, 2010, Vol. 29(1), pp. 57-64.
    \24\ Institute of Medicine, ``Preventing medication errors: 
quality chasm series,'' Washington, DC: The National Academies 
Press, 2007. Available at https://www.nap.edu/read/11623/chapter/1.
    \25\ Starmer, A.J., Sectish, T.C., Simon, D.W., Keohane, C., 
McSweeney, M.E., Chung, E.Y., Yoon, C.S., Lipsitz, S.R., Wassner, 
A.J., Harper, M.B., & Landrigan, C.P., ``Rates of medical errors and 
preventable adverse events among hospitalized children following 
implementation of a resident handoff bundle,'' JAMA, 2013, Vol. 
310(21), pp. 2262-2270.
    \26\ Pronovost, P., M.M.E. Johns, S. Palmer, R.C. Bono, D.B. 
Fridsma, A. Gettinger, J. Goldman, W. Johnson, M. Karney, C. Samitt, 
R.D. Sriram, A. Zenooz, and Y.C. Wang, Editors. Procuring 
Interoperability: Achieving High-Quality, Connected, and Person-
Centered Care. Washington, DC, 2018 National Academy of Medicine. 
Available at https://nam.edu/wp-content/uploads/2018/10/Procuring-Interoperability_web.pdf.
    \27\ Balaban RB, Weissman JS, Samuel PA, & Woolhandler, S., 
``Redefining and redesigning hospital discharge to enhance patient 
care: a randomized controlled study,'' J Gen Intern Med, 2008, Vol. 
23(8), pp. 1228-33.

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

[[Page 39101]]

    Care transitions across health care settings have been 
characterized as complex, costly, and potentially hazardous, and may 
increase the risk for multiple adverse outcomes.28 29 The 
rising incidence of preventable adverse events, complications, and 
hospital readmissions have drawn attention to the importance of the 
timely transfer of health information and care preferences at the time 
of transition. Failures of care coordination, including poor 
communication of information, were estimated to cost the U.S. health 
care system between $25 billion and $45 billion in wasteful spending in 
2011.\30\ The communication of health information and patient care 
preferences is critical to ensuring safe and effective transitions from 
one health care setting to another.31 32
---------------------------------------------------------------------------

    \28\ Arbaje, A.I., Kansagara, D.L., Salanitro, A.H., Englander, 
H.L., Kripalani, S., Jencks, S.F., & Lindquist, L.A., ``Regardless 
of age: incorporating principles from geriatric medicine to improve 
care transitions for patients with complex needs,'' Journal of 
General Internal Medicine, 2014, Vol 29(6), pp. 932-939.
    \29\ Simmons, S., Schnelle, J., Slagle, J., Sathe, N.A., 
Stevenson, D., Carlo, M., & McPheeters, M.L., ``Resident safety 
practices in nursing home settings.'' Technical Brief No. 24 
(Prepared by the Vanderbilt Evidence-based Practice Center under 
Contract No. 290-2015-00003-I.) AHRQ Publication No. 16-EHC022-EF. 
Rockville, MD: Agency for Healthcare Research and Quality. May 2016. 
Available at https://www.ncbi.nlm.nih.gov/books/NBK384624/.
    \30\ Berwick, D.M. & Hackbarth, A.D. ``Eliminating Waste in US 
Health Care,'' JAMA, 2012, Vol. 307(14), pp.1513-1516.
    \31\ McDonald, K.M., Sundaram, V., Bravata, D.M., Lewis, R., 
Lin, N., Kraft, S.A. & Owens, D.K. Care Coordination. Vol. 7 of: 
Shojania K.G., McDonald K.M., Wachter R.M., Owens D.K., editors. 
``Closing the quality gap: A critical analysis of quality 
improvement strategies.'' Technical Review 9 (Prepared by the 
Stanford University-UCSF Evidence-based Practice Center under 
contract 290-02-0017). AHRQ Publication No. 04(07)-0051-7. 
Rockville, MD: Agency for Healthcare Research and Quality. June 
2006. Available at https://www.ncbi.nlm.nih.gov/books/NBK44015/.
    \32\ Lattimer, C., ``When it comes to transitions in patient 
care, effective communication can make all the difference,'' 
Generations, 2011, Vol. 35(1), pp. 69-72.
---------------------------------------------------------------------------

    Patients in PAC settings often have complicated medication regimens 
and require efficient and effective communication and coordination of 
care between settings, including detailed transfer of medication 
information.33 34 35 Individuals in PAC settings may be 
vulnerable to adverse health outcomes due to insufficient medication 
information on the part of their health care providers, and the higher 
likelihood for multiple comorbid chronic conditions, polypharmacy, and 
complicated transitions between care settings.36 37 
Preventable adverse drug events (ADEs) may occur after hospital 
discharge in a variety of settings including PAC.\38\ A 2014 Office of 
Inspector General report found that 10 percent of Medicare patients in 
IRFs experienced adverse events, with most of those events being 
medication related. Over 45 percent of the adverse events and temporary 
harm events were clearly or likely preventable.\39\ Medication errors 
and one-fifth of ADEs occur during transitions between settings, 
including admission to or discharge from a hospital to home or a PAC 
setting, or transfer between hospitals.40 41
---------------------------------------------------------------------------

    \33\ Starmer A.J, Spector N.D., Srivastava R., West, D.C., 
Rosenbluth, G., Allen, A.D., Noble, E.L., & Landrigen, C.P., 
``Changes in medical errors after implementation of a handoff 
program,'' N Engl J Med, 2014, Vol. 37(1), pp. 1803-1812.
    \34\ Kruse, C.S. Marquez, G., Nelson, D., & Polomares, O., ``The 
use of health information exchange to augment patient handoff in 
long-term care: a systematic review,'' Applied Clinical Informatics, 
2018, Vol. 9(4), pp. 752-771.
    \35\ Brody, A.A., Gibson, B., Tresner-Kirsch, D., Kramer, H., 
Thraen, I., Coarr, M.E., & Rupper, R., ``High prevalence of 
medication discrepancies between home health referrals and Centers 
for Medicare and Medicaid Services home health certification and 
plan of care and their potential to affect safety of vulnerable 
elderly adults,'' Journal of the American Geriatrics Society, 2016, 
Vol. 64(11), pp. e166-e170.
    \36\ Chhabra, P.T., Rattinger, G.B., Dutcher, S.K., Hare, M.E., 
Parsons, K., L., & Zuckerman, I.H., ``Medication reconciliation 
during the transition to and from long-term care settings: a 
systematic review,'' Res Social Adm Pharm, 2012, Vol. 8(1), pp. 60-
75.
    \37\ Levinson, D.R., & General, I., ``Adverse events in skilled 
nursing facilities: national incidence among Medicare 
beneficiaries.'' Washington, DC: U.S. Department of Health and Human 
Services, Office of the Inspector General, February 2014. Available 
at https://oig.hhs.gov/oei/reports/oei-06-11-00370.pdf.
    \38\ Battles J., Azam I., Grady M., & Reback K., ``Advances in 
patient safety and medical liability,'' AHRQ Publication No. 17-
0017-EF. Rockville, MD: Agency for Healthcare Research and Quality, 
August 2017. Available at https://www.ahrq.gov/sites/default/files/publications/files/advances-complete_3.pdf.
    \39\ Health and Human Services Office of Inspector General. 
Adverse Events in Rehabilitation Hospitals: National Incidence Among 
Medicare Beneficiaries. (OEI-06-14-00110). 2018. Available at 
https://oig.hhs.gov/oei/reports/oei-06-14-00110.asp.
    \40\ Barnsteiner, J.H., ``Medication Reconciliation: Transfer of 
medication information across settings--keeping it free from 
error,'' The American Journal of Nursing, 2005, Vol. 105(3), pp. 31-
36.
    \41\ Gleason, K.M., Groszek, J.M., Sullivan, C., Rooney, D., 
Barnard, C., Noskin, G.A., ``Reconciliation of discrepancies in 
medication histories and admission orders of newly hospitalized 
patients,'' American Journal of Health System Pharmacy, 2004, Vol. 
61(16), pp. 1689-1694.
---------------------------------------------------------------------------

    Patients in PAC settings are often taking multiple medications. 
Consequently, PAC providers regularly are in the position of starting 
complex new medication regimens with little knowledge of the patients 
or their medication history upon admission. Furthermore, inter-facility 
communication barriers delay resolving medication discrepancies during 
transitions of care.\42\ Medication discrepancies are common \43\ and 
found to occur in 86 percent of all transitions, increasing the 
likelihood of ADEs.44 45 46 Up to 90 percent of patients 
experience at least one medication discrepancy in the transition from 
hospital to home care, and discrepancies occur within all therapeutic 
classes of medications.47 48
---------------------------------------------------------------------------

    \42\ Patterson M., Foust J.B., Bollinger, S., Coleman, C., 
Nguyen, D., ``Inter-facility communication barriers delay resolving 
medication discrepancies during transitions of care,'' Research in 
Social & Administrative Pharmacy (2018), doi: 10.1016/
j.sapharm.2018.05.124.
    \43\ Manias, E., Annaikis, N., Considine, J., Weerasuriya, R., & 
Kusljic, S. ``Patient-, medication- and environment-related factors 
affecting medication discrepancies in older patients,'' Collegian, 
2017, Vol. 24, pp. 571-577.
    \44\ Tjia, J., Bonner, A., Briesacher, B.A., McGee, S., Terrill, 
E., Miller, K., ``Medication discrepancies upon hospital to skilled 
nursing facility transitions,'' J Gen Intern Med, 2009, Vol. 24(5), 
pp. 630-635.
    \45\ Sinvani, L.D., Beizer, J., Akerman, M., Pekmezaris, R., 
Nouryan, C., Lutsky, L., Cal, C., Dlugacz, Y., Masick, K., Wolf-
Klein, G.,''Medication reconciliation in continuum of care 
transitions: a moving target,'' J Am Med Dir Assoc, 2013, Vol. 
14(9), 668-672.
    \46\ Coleman E.A., Parry C., Chalmers S., & Min, S.J., ``The 
Care Transitions Intervention: results of a randomized controlled 
trial,'' Arch Intern Med, 2006, Vol. 166, pp. 1822-28.
    \47\ Corbett C.L., Setter S.M., Neumiller J.J., & Wood, l.D., 
``Nurse identified hospital to home medication discrepancies: 
implications for improving transitional care,'' Geriatr Nurs, 2011, 
Vol. 31(3), pp. 188-96.
    \48\ Setter S.M., Corbett C.F., Neumiller J.J., Gates, B.J., 
Sclar, D.A., & Sonnett, T.E., ``Effectiveness of a pharmacist-nurse 
intervention on resolving medication discrepancies in older patients 
transitioning from hospital to home care: impact of a pharmacy/
nursing intervention,'' Am J Health Syst Pharm, 2009, Vol. 66, pp. 
2027-31.
---------------------------------------------------------------------------

    Transfer of a medication list between providers is necessary for 
medication reconciliation interventions, which have been shown to be a 
cost-effective way to avoid ADEs by reducing errors 49 50 51

[[Page 39102]]

especially when medications are reviewed by a pharmacist using 
electronic medical records.\52\
---------------------------------------------------------------------------

    \49\ Boockvar, K.S., Blum, S., Kugler, A., Livote, E., 
Mergenhagen, K.A., Nebeker, J.R., & Yeh, J., ``Effect of admission 
medication reconciliation on adverse drug events from admission 
medication changes,'' Archives of Internal Medicine, 2011, Vol. 
171(9), pp. 860-861.
    \50\ Kwan, J.L., Lo, L., Sampson, M., & Shojania, K.G., 
``Medication reconciliation during transitions of care as a patient 
safety strategy: a systematic review,'' Annals of Internal Medicine, 
2013, Vol. 158(5), pp. 397-403.
    \51\ Chhabra, P.T., Rattinger, G.B., Dutcher, S.K., Hare, M.E., 
Parsons, K., L., & Zuckerman, I.H., ``Medication reconciliation 
during the transition to and from long-term care settings: a 
systematic review,'' Res Social Adm Pharm, 2012, Vol. 8(1), pp. 60-
75.
    \52\ Agrawal A, Wu WY. ``Reducing medication errors and 
improving systems reliability using an electronic medication 
reconciliation system,'' The Joint Commission Journal on Quality and 
Patient Safety, 2009, Vol. 35(2), pp. 106-114.
---------------------------------------------------------------------------

b. Stakeholder and Technical Expert Panel (TEP) Input
    The proposed measure was developed after consideration of feedback 
we received from stakeholders and four TEPs convened by our 
contractors. Further, the proposed measure was developed after 
evaluation of data collected during two pilot tests we conducted in 
accordance with the CMS Measures Management System Blueprint.
    Our measure development contractors constituted a TEP which met on 
September 27, 2016,\53\ January 27, 2017,\54\ and August 3, 2017 \55\ 
to provide input on a prior version of this measure. Based on this 
input, we updated the measure concept in late 2017 to include the 
transfer of a specific component of health information--medication 
information. Our measure development contractors reconvened this TEP on 
April 20, 2018 for the purpose of obtaining expert input on the 
proposed measure, including the measure's reliability, components of 
face validity, and feasibility of being implemented across PAC 
settings. Overall, the TEP was supportive of the proposed measure, 
affirming that the measure provides an opportunity to improve the 
transfer of medication information. A summary of the April 20, 2018 TEP 
proceedings titled ``Transfer of Health Information TEP Meeting 4--June 
2018'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
---------------------------------------------------------------------------

    \53\ Technical Expert Panel Summary Report: Development of two 
quality measures to satisfy the Improving Medicare Post-Acute Care 
Transformation Act of 2014 (IMPACT Act) Domain of Transfer of health 
Information and Care Preferences When an Individual Transitions to 
Skilled Nursing Facilities (SNFs), Inpatient Rehabilitation 
Facilities (IRFs), Long Term Care Hospitals (LTCHs) and Home Health 
Agencies (HHAs). Available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/Downloads/Transfer-of-Health-Information-TEP_Summary_Report_Final-June-2017.pdf.
    \54\ Technical Expert Panel Summary Report: Development of two 
quality measures to satisfy the Improving Medicare Post-Acute Care 
Transformation Act of 2014 (IMPACT Act) Domain of Transfer of health 
Information and Care Preferences When an Individual Transitions to 
Skilled Nursing Facilities (SNFs), Inpatient Rehabilitation 
Facilities (IRFs), Long Term Care Hospitals (LTCHs) and Home Health 
Agencies (HHAs). Available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/Downloads/Transfer-of-Health-Information-TEP-Meetings-2-3-Summary-Report_Final_Feb2018.pdf.
    \55\ Ibid.
---------------------------------------------------------------------------

    Our measure development contractors solicited stakeholder feedback 
on the proposed measure by requesting comment on the CMS Measures 
Management System Blueprint website, and accepted comments that were 
submitted from March 19, 2018 to May 3, 2018. The comments received 
noted overall support for the measure. Several commenters suggested 
ways to improve the measure, primarily related to what types of 
information should be included at transfer. We incorporated this input 
into development of the proposed measure. The summary report for the 
March 19 to May 3, 2018 public comment period titled ``IMPACT 
Medication Profile Transferred Public Comment Summary Report'' is 
available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
c. Pilot Testing
    The proposed measure was tested between June and August 2018 in a 
pilot test that involved 24 PAC facilities/agencies, including five 
IRFs, six SNFs, six LTCHs, and seven HHAs. The 24 pilot sites submitted 
a total of 801 records. Analysis of agreement between coders within 
each participating facility (266 qualifying pairs) indicated a 93 
percent agreement for this measure. Overall, pilot testing enabled us 
to verify its reliability, components of face validity, and feasibility 
of being implemented across PAC settings. Further, more than half of 
the sites that participated in the pilot test stated during the 
debriefing interviews that the measure could distinguish facilities or 
agencies with higher quality medication information transfer from those 
with lower quality medication information transfer at discharge. The 
pilot test summary report titled ``Transfer of Health Information 2018 
Pilot Test Summary Report'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
d. Measure Applications Partnership (MAP) Review and Related Measures
    We included the proposed measure in the IRF QRP section of the 2018 
Measures Under Consideration (MUC) List. The MAP conditionally 
supported this measure pending NQF endorsement, noting that the measure 
can promote the transfer of important medication information. The MAP 
also suggested that we consider a measure that can be adapted to 
capture bi-directional information exchange, and recommended that the 
medication information transferred include important information about 
supplements and opioids. More information about the MAP's 
recommendations for this measure is available at http://www.qualityforum.org/Publications/2019/02/MAP_2019_Considerations_for_Implementing_Measures_Final_Report_-_PAC-LTC.aspx.
    As part of the measure development and selection process, we also 
identified one NQF-endorsed quality measure similar to the proposed 
measure, titled Documentation of Current Medications in the Medical 
Record (NQF #0419, CMS eCQM ID: CMS68v8). This measure was adopted as 
one of the recommended adult core clinical quality measures for 
eligible professionals for the EHR Incentive Program beginning in 2014 
and was also adopted under the Merit-based Incentive Payment System 
(MIPS) quality performance category beginning in 2017. The measure is 
calculated based on the percentage of visits for patients aged 18 years 
and older for which the eligible professional or eligible clinician 
attests to documenting a list of current medications using all 
resources immediately available on the date of the encounter.
    The proposed Transfer of Health Information to the Provider--Post-
Acute Care (PAC) measure addresses the transfer of information whereas 
the NQF-endorsed measure #0419 assesses the documentation of 
medications, but not the transfer of such information. This is 
important as the proposed measure assesses for the transfer of 
medication information for the proposed measure calculation. Further, 
the proposed measure utilizes standardized patient assessment data 
elements (SPADEs), which is a requirement for measures specified under 
the Transfer of Health Information measure domain under section 
1899B(c)(1)(E) of the Act, whereas NQF #0419 does not.

[[Page 39103]]

    After review of the NQF-endorsed measure, we determined that the 
proposed Transfer of Health Information to the Provider--Post-Acute 
Care (PAC) measure better addresses the Transfer of Health Information 
measure domain, which requires that at least some of the data used to 
calculate the measure be collected as standardized patient assessment 
data through the post-acute care assessment instruments. Section 
1886(j)(7)(D)(i) of the Act requires that any measure specified by the 
Secretary be endorsed by the entity with a contract under section 
1890(a) of the Act, which is currently the National Quality Form (NQF). 
However, when a feasible and practical measure has not been NQF 
endorsed for a specified area or medical topic determined appropriate 
by the Secretary, section 1886(j)(7)(D)(ii) of the Act allows the 
Secretary to specify a measure that is not NQF endorsed as long as due 
consideration is given to the measures that have been endorsed or 
adopted by a consensus organization identified by the Secretary. For 
the reasons discussed previously, we believe that there is currently no 
feasible NQF-endorsed measure that we could adopt under section 
1886(j)(7)(D)(ii) of the Act. However, we note that we intend to submit 
the proposed measure to the NQF for consideration of endorsement when 
feasible.
e. Quality Measure Calculation
    The proposed Transfer of Health Information to the Provider--Post-
Acute Care (PAC) quality measure is calculated as the proportion of 
patient stays with a discharge assessment indicating that a current 
reconciled medication list was provided to the subsequent provider at 
the time of discharge. The proposed measure denominator is the total 
number of IRF patient stays ending in discharge to a subsequent 
provider, which is defined as a short-term general acute-care hospital, 
intermediate care (intellectual and developmental disabilities 
providers), home under care of an organized home health service 
organization or hospice, hospice in an institutional facility, a SNF, 
an LTCH, another IRF, an IPF, or a CAH. These health care providers 
were selected for inclusion in the denominator because they are 
identified as subsequent providers on the discharge destination item 
that is currently included on the IRF PAI. The proposed measure 
numerator is the number of IRF patient stays with an IRF-PAI discharge 
assessment indicating a current reconciled medication list was provided 
to the subsequent provider at the time of discharge. For additional 
technical information about this proposed measure, we refer readers to 
the document titled, ``Final Specifications for IRF QRP Quality 
Measures and Standardized Patient Assessment Data Elements,'' available 
at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html. The data source for the proposed 
quality measure is the IRF-PAI assessment instrument for IRF patients.
    For more information about the data submission requirements we 
proposed for this measure, we refer readers to section VIII.G.3. of 
this final rule.
    Commenters submitted the following comments related to the proposed 
rule's discussion of the IRF QRP Quality Measure Proposals beginning 
with the FY 2022 IRF QRP. A discussion of these comments, along with 
our responses, appears below. We also address comments on the proposed 
Transfer of Health Information to the Patient--Post-Acute Care measure 
(discussed further in a subsequent section of this final rule) in this 
section because commenters frequently addressed both Transfer of Health 
Information measures together.
    Response: We thank the commenters for their support of the Transfer 
of Health Information measures.
    Comment: One commenter suggested that other providers, such as 
outpatient physical therapists, should be included in the definition of 
a subsequent provider for the Transfer of Health Information to the 
Provider--Post-Acute Care measure.
    Response: We appreciate the suggestion to expand the Transfer of 
Health Information to the Provider--Post-Acute Care measure outcome to 
assess the transfer of health information to other providers such as 
outpatient physical therapists. We recognize that sharing medication 
information with outpatient providers is important, and will take into 
consideration additional providers in future measure modifications. 
Through our measure development and pilot testing we learned that 
outpatient providers cannot always be readily identified by the PAC 
provider. For this process measure, which serves as a building block 
for improving the transfer of medication information, we specified 
providers who will be involved in the care of the patient and 
medication management after discharge and can be readily identified 
through the discharge location item on the IRF-PAI. The clear 
delineation of the recipient of the medication list in the measure 
specifications will improve measure reliability and validity.
    Comment: A commenter recommended that the Transfer of Health 
Information to the Provider--Post-Acute Care measure be expanded to 
include the transfer of information that would help prevent infections 
and facilitate appropriate infection prevention and control 
interventions during care transitions in addition to the medication 
information in the finalized measure.
    Response: The Transfer of Health Information to the Provider--Post-
Acute Care measure focuses on the transfer of a reconciled medication 
list. The measure was designed after input from TEPs, public comment, 
and other stakeholders that suggested the quality measures focus on the 
transfer of the most critical pieces of information to support patient 
safety and care coordination. However, we acknowledge that the transfer 
of many other forms of health information is important, and while the 
focus of this measure is on a reconciled medication list, we hope to 
expand our measures in the future.
    Comment: Several commenters raised concerns about both of the 
Transfer of Health Information measures not being endorsed by the 
National Quality Forum (NQF). A few commenters requested that we 
consider delaying rollout of these two new measures until endorsed by 
NQF. A few commenters recommended that we only adopt measures that have 
NQF approval. One commenter was opposed to the measures because they 
have not been endorsed by NQF.
    Response: While this measure is not currently NQF-endorsed, we 
recognize that the NQF endorsement process is an important part of 
measure development. As discussed in the FY 2020 IRF PPS proposed rule 
(84 FR 17286 through 17291), we believe the measures better address the 
Transfer of Health Information measure domain, which requires that at 
least some of the data used to calculate the measure be collected as 
standardized patient assessment data through the post-acute care 
assessment instruments, than any endorsed measures. While section 
1886(j)(7)(D)(i) of the Act requires that any measure specified by the 
Secretary be endorsed by the entity with a contract under section 
1890(a) of the Act, which is currently the National Quality Form (NQF), 
when a feasible and practical measure has not been NQF endorsed for a 
specified area or medical topic determined appropriate by the

[[Page 39104]]

Secretary, section 1886(j)(7)(D)(ii) of the Act allows the Secretary to 
specify a measure that is not NQF endorsed as long as due consideration 
is given to the measures that has been endorsed or adopted by a 
consensus organization identified by the Secretary. We plan to submit 
the measure for NQF endorsement consideration as soon as feasible.
    Comment: Several commenters stated that the Transfer of Health 
Information measures will add burden. Two commenters did not support 
the measures for this reason. One commenter stated that achieving high 
performance on the measures will add administrative burden. Another 
commenter stated that the measures will add burden with no added value. 
Another commenter stated that while there will be additional burden on 
IRFs to collect and report data for these new measures, the benefit to 
patients and the CMS program outweighs the additional burden on 
providers.
    Response: We agree that the benefit to patients outweighs any 
additional burden on providers. We are also very mindful of burden that 
may occur from the collection and reporting of our measures, as 
supported by the Meaningful Measures and Patients over Paperwork 
initiatives. We emphasize that both measures are comprised of one item, 
and further, the activities associated with the measure align with 
existing requirements related to transferring information at the time 
of discharge to safeguard patients. Additionally, TEP feedback and 
pilot test found that the burden of reporting will not be significant. 
We believe that these measures will likely drive improvements in the 
transfer of medication information between providers and with patients, 
families, and caregivers.
    Comment: One commenter stated that there will be no additional 
burden to IRFs, because providing medication information as part of 
discharge planning is a Condition of Participation requirement for 
Medicaid and Medicare, and the medication list can be generated from 
the electronic medical record.
    Response: We believe that the Transfer of Health Information 
measures will not substantially increase burden because we understand 
that many hospitals already generate medication lists as a best 
practice.
    Comment: We received comments related to the validity and 
reliability of both Transfer of Health Information measures. One 
commenter suggested that CMS should ensure accuracy of these measures. 
Other commenters suggested that additional testing is needed to ensure 
that these measures will be able to differentiate among IRF providers. 
Another commenter questioned if the measures would be topped out 
shortly after adoption, since medication reconciliation is already 
completed by facilities at discharge.
    Response: Elements of validity and reliability were analyzed during 
pilot testing of these measures, with good results, including inter-
rater reliability of at least 87 percent for all tested items. Pilot 
testing also indicated that there is room for improvement for IRFs and 
other settings, so we do not expect the measure to be topped out 
shortly after adoption. As we monitor the outcomes of these measures, 
we will ensure that reliability and validity of the measures meet 
acceptable standards.
    Comment: Some commenters recommended ways in which the Transfer of 
Health Information measures specifications could be updated or changed. 
A few commenters suggested that the ``not applicable'' (NA) answer 
choice available in the home health version of the measure be made 
available in all settings, including IRFs. A few commenters also 
requested clarification about why patients discharged home under the 
care of an organized home health service or hospice would be captured 
in the denominators of both Transfer of Health information measures.
    Response: We are appreciative of the measure modification 
suggestions and clarify why the response option of N/A was considered 
only for the HH version of this measure. The coding response N/A, or 
``not applicable'' is used when the HHA was not made aware of the 
transfer in a timely manner and, therefore, the HHA is not able to 
provide the medication list at the time of transfer to the subsequent 
provider. For example, a HHA may not be immediately aware when a 
patient is taken to the emergency room. For facility settings, such as 
the IRF setting, where 24-hour care is being provided, the facility 
should always be aware and actively involved in the discharge of the 
patient, and therefore, able to provide the current reconciled 
medication list at the time of discharge. Therefore, we believed the 
coding option of ``N/A'' would not be useful in the facility-based 
measure as the facility is aware and involved in the discharge. We wish 
to note that while the N/A option is considered for the HHA version of 
the measure, the measure specifications indicate that these patients 
are not removed from the denominator. In addition, discharge to home 
under the care of an organized HHA or hospice is captured in the 
denominator of both the Transfer of Health Information to Provider and 
Transfer of Health Information to Patient measures because this type of 
discharge represents two opportunities to transfer the medication list. 
These measures aim to assure that each of these transfers is taking 
place. We refer readers to the measure specifications where updates or 
changes can be found and are available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Comment: One commenter suggested that the Transfer of Health 
Information measures should include a measure of the timeliness of the 
transfer. The commenter stated that, as currently specified, the 
measures give equal credit for information that is sent immediately and 
information sent days later.
    Response: We appreciate the suggestions that CMS develop and adopt 
measures that assess for the timeliness of transfer. We agree that 
measure concepts of this type are important and would complement the 
measures that focus for whether information was transferred at the time 
the patient leaves the facility. We clarify that the measures do not 
give credit for when information was sent, whether immediately or days 
later. This is because there may be circumstances where information may 
not be sent at the immediate time of discharge. However, the measures 
do require that information be shared with the subsequent provider and/
or the patient as close to the time of discharge as this is actionable, 
allows for shared decision making, and will increase coordinated care. 
We are not establishing a new standard of transfer at discharge; we are 
simply assessing if information was sent at the time a patient leaves 
the facility. As we move through future measure development work, we 
will consider a ``timeliness'' component for these measure concepts.
    Comment: A commenter noted that although CMS provided guidelines 
regarding what should be included in the transfer of medication 
information, the data collection on this measure does not require that 
these guidelines be met. The commenter questioned if CMS intends to 
audit IRFs to ensure that the measure values are consistent with the 
information being shared.
    Response: The Transfer of Health Information measures serve as a 
check to ensure that a reconciled medication list is provided as the 
patient changes care settings. Defining the completeness of that 
medication list is left to the discretion of the providers and patient

[[Page 39105]]

who are coordinating this care. We interpret the comment about audits 
to be referring to data validation. While we do not have a data 
validation program in place at this time, we are exploring such a 
program akin to that of the hospital QRPs. For all measures and data 
collected for the IRF QRP, we monitor and evaluate our data to assess 
for coding patterns, errors, reliability, and soundness of the data. 
Through data monitoring, we are able to assess if measure outcomes are 
consistent with the information that is collected. We note that all 
data are subject to review and audit.
    Comment: A few comments included concerns that the Transfer of 
Health Information measures are not indicative of provider quality and 
questioned the ability of the measures to improve patient outcomes. Two 
commenters did not support the measures for this reason. Commenters 
noted that the measures assess whether a medication list was 
transferred and not whether that medication list was accurate and 
received by the subsequent provider.
    Response: The Transfer of Health Information to the Provider--Post-
Acute Care and Transfer of Health Information to the Patient--Post-
Acute Care measures are process measures designed to address and 
improve an important aspect of care quality. Lack of timely transfer of 
medication information at transitions has been demonstrated to lead to 
increased risk of adverse events, medication errors, and 
hospitalizations. In addition, public commenters and our TEP members 
identified many problems and gaps in the timely transfer of medication 
information at transitions. Process measures, such as these, are 
building blocks toward improved coordinated care and discharge 
planning, providing information that will improve shared decision 
making and coordination. Further, process measures hold a lot of value 
as they delineate negative and/or positive aspects of the health care 
process. These measures will capture the quality of the process of 
medication information transfer and, we believe, help to improve those 
processes. When developing future measures, we will take into 
consideration suggestions about measures that assess the accuracy of 
the medication list and whether it was received by the subsequent 
provider.
    Comment: One commenter suggested that CMS work to identify 
interoperability solutions as a means of decreasing opportunities for 
errors by providing clinicians and patients secure access to the most 
up-to-date medication-related information. The commenter also suggests 
that if CMS is required by the IMPACT Act to adopt these measures, that 
we do so as an interim step, within a defined timeframe, while 
interoperability solutions are explored and tested.
    Response: We agree with the comments on the importance of 
interoperability solutions to support health information transfer. CMS 
and ONC are focused on improving interoperability and the timely 
sharing of information between providers, patients, families and 
caregivers. We believe that PAC provider health information exchange 
supports the goals of high quality, personalized, efficient healthcare, 
care coordination, person-centered care, and supports real-time, data 
driven, clinical decision making. We are optimistic that this measure 
will encourage the electronic transfer of current and important 
medication information at transitions. These measures and related 
efforts may help accelerate interoperability solutions. The Transfer of 
Health Information measures assess the process of medication transfer, 
which can occur through both electronic and non-electronic means. We 
clarify that these measures are an interim step in improving 
coordinated care, and we also believe that other interoperable 
solutions should be explored. Finalizing these Transfer of Health 
measures will be a first step in measuring the transfer of this 
medication-related information.
    After consideration of the public comments, we are finalizing our 
proposal to adopt the Transfer of Health Information to the Provider--
Post Acute Care (PAC) measure, under section 1899B(c)(1)(E) of the Act, 
with data collection for discharges beginning October 1, 2020.
2. Transfer of Health Information to the Patient--Post-Acute Care (PAC) 
Measure
    Beginning with the FY 2022 IRF QRP, we proposed to adopt the 
Transfer of Health Information to the Patient--Post Acute Care (PAC) 
measure, a measure that satisfies the IMPACT Act domain of Transfer of 
Health Information, with data collection for discharges beginning 
October 1, 2020. This process-based measure assesses whether or not a 
current reconciled medication list was provided to the patient, family, 
or caregiver when the patient was discharged from a PAC setting to a 
private home/apartment, a board and care home, assisted living, a group 
home, transitional living or home under care of an organized home 
health service organization, or a hospice.
a. Background
    In 2013, 22.3 percent of all acute hospital discharges were 
discharged to PAC settings, including 11 percent who were discharged to 
home under the care of a home health agency.\56\ Of the Medicare FFS 
beneficiaries with an IRF stay in FYs 2016 and 2017, an estimated 51 
percent were discharged home with home health services, 21 percent were 
discharged home with self-care, and 0.5 percent were discharged with 
home hospice services.\57\
---------------------------------------------------------------------------

    \56\ Tian, W. ``An all-payer view of hospital discharge to 
postacute care,'' May 2016. Available at https://www.hcup-us.ahrq.gov/reports/statbriefs/sb205-Hospital-Discharge-Postacute-Care.jsp.
    \57\ RTI International analysis of Medicare claims data for 
index stays in IRF 2016/2017. (RTI program reference: MM150).
---------------------------------------------------------------------------

    The communication of health information, such as a reconciled 
medication list, is critical to ensuring safe and effective patient 
transitions from health care settings to home and/or other community 
settings. Incomplete or missing health information, such as medication 
information, increases the likelihood of a patient safety risk, often 
life-threatening.58 59 60 61 62 Individuals who use PAC care 
services are particularly vulnerable to adverse health outcomes due to 
their higher likelihood of having multiple comorbid chronic conditions, 
polypharmacy, and complicated transitions between care 
settings.63 64 Upon discharge to home,

[[Page 39106]]

individuals in PAC settings may be faced with numerous medication 
changes, new medication regimes, and follow-up 
details.65 66 67 The efficient and effective communication 
and coordination of medication information may be critical to prevent 
potentially deadly adverse effects. When care coordination activities 
enhance care transitions, these activities can reduce duplication of 
care services and costs of care, resolve conflicting care plans, and 
prevent medical errors.68 69
---------------------------------------------------------------------------

    \58\ Kwan, J.L., Lo, L., Sampson, M., & Shojania, K.G., 
``Medication reconciliation during transitions of care as a patient 
safety strategy: A systematic review,'' Annals of Internal Medicine, 
2013, Vol. 158(5), pp. 397-403.
    \59\ Boockvar, K.S., Blum, S., Kugler, A., Livote, E., 
Mergenhagen, K.A., Nebeker, J.R., & Yeh, J., ``Effect of admission 
medication reconciliation on adverse drug events from admission 
medication changes,'' Archives of Internal Medicine, 2011, Vol. 
171(9), pp. 860-861.
    \60\ Bell, C.M., Brener, S.S., Gunraj, N., Huo, C., Bierman, 
A.S., Scales, D.C., & Urbach, D.R., ``Association of ICU or hospital 
admission with unintentional discontinuation of medications for 
chronic diseases,'' JAMA, 2011, Vol. 306(8), pp. 840-847.
    \61\ Basey, A.J., Krska, J., Kennedy, T.D., & Mackridge, A.J., 
``Prescribing errors on admission to hospital and their potential 
impact: A mixed-methods study,'' BMJ Quality & Safety, 2014, Vol. 
23(1), pp. 17-25.
    \62\ Desai, R., Williams, C.E., Greene, S.B., Pierson, S., & 
Hansen, R.A., ``Medication errors during patient transitions into 
nursing homes: Characteristics and association with patient harm,'' 
The American Journal of Geriatric Pharmacotherapy, 2011, Vol. 9(6), 
pp. 413-422.
    \63\ Brody, A.A., Gibson, B., Tresner-Kirsch, D., Kramer, H., 
Thraen, I., Coarr, M.E., & Rupper, R. ``High prevalence of 
medication discrepancies between home health referrals and Centers 
for Medicare and Medicaid Services home health certification and 
plan of care and their potential to affect safety of vulnerable 
elderly adults,'' Journal of the American Geriatrics Society, 2016, 
Vol. 64(11), pp. e166-e170.
    \64\ Chhabra, P.T., Rattinger, G.B., Dutcher, S.K., Hare, M.E., 
Parsons, K., L., & Zuckerman, I.H., ``Medication reconciliation 
during the transition to and from long-term care settings: A 
systematic review,'' Res Social Adm Pharm, 2012, Vol. 8(1), pp. 60-
75.
    \65\ Brody, A.A., Gibson, B., Tresner-Kirsch, D., Kramer, H., 
Thraen, I., Coarr, M.E., & Rupper, R. ``High prevalence of 
medication discrepancies between home health referrals and Centers 
for Medicare and Medicaid Services home health certification and 
plan of care and their potential to affect safety of vulnerable 
elderly adults,'' Journal of the American Geriatrics Society, 2016, 
Vol. 64(11), pp. e166-e170.
    \66\ Bell, C.M., Brener, S.S., Gunraj, N., Huo, C., Bierman, 
A.S., Scales, D.C., & Urbach, D.R., ``Association of ICU or hospital 
admission with unintentional discontinuation of medications for 
chronic diseases,'' JAMA, 2011, Vol. 306(8), pp. 840-847.
    \67\ Sheehan, O.C., Kharrazi, H., Carl, K.J., Leff, B., Wolff, 
J.L., Roth, D.L., Gabbard, J., & Boyd, C.M., ``Helping older adults 
improve their medication experience (HOME) by addressing medication 
regimen complexity in home healthcare,'' Home Healthcare Now. 2018, 
Vol. 36(1) pp. 10-19.
    \68\ Mor, V., Intrator, O., Feng, Z., & Grabowski, D. C., ``The 
revolving door of rehospitalization from skilled nursing 
facilities,'' Health Affairs, 2010, Vol. 29(1), pp. 57-64.
    \69\ Starmer, A.J., Sectish, T.C., Simon, D.W., Keohane, C., 
McSweeney, M.E., Chung, E.Y., Yoon, C.S., Lipsitz, S.R., Wassner, 
A.J., Harper, M.B., & Landrigan, C.P., ``Rates of medical errors and 
preventable adverse events among hospitalized children following 
implementation of a resident handoff bundle,'' JAMA, 2013, Vol. 
310(21), pp. 2262-2270.
---------------------------------------------------------------------------

    Finally, the transfer of a patient's discharge medication 
information to the patient, family, or caregiver is common practice and 
supported by discharge planning requirements for participation in 
Medicare and Medicaid programs.\70\ \71\ Most PAC EHR systems generate 
a discharge medication list to promote patient participation in 
medication management, which has been shown to be potentially useful 
for improving patient outcomes and transitional care.\72\
---------------------------------------------------------------------------

    \70\ CMS, ``Revision to state operations manual (SOM), Hospital 
Appendix A--Interpretive Guidelines for 42 CFR 482.43, Discharge 
Planning'' May 17, 2013. Available at https://www.cms.gov/Medicare/Provider-Enrollment-and-Certification/SurveyCertificationGenInfo/Downloads/Survey-and-Cert-Letter-13-32.pdf.
    \71\ The State Operations Manual Guidance to Surveyors for Long 
Term Care Facilities (Guidance Sec.  483.21(c)(1) Rev. 11-22-17) for 
discharge planning process. Available at https://www.cms.gov/Regulations-and-Guidance/Guidance/Manuals/downloads/som107ap_pp_guidelines_ltcf.pdf.
    \72\ Toles, M., Colon-Emeric, C., Naylor, M.D., Asafu-Adjei, J., 
Hanson, L.C., ``Connect-home: Transitional care of skilled nursing 
facility patients and their caregivers,'' Am Geriatr Soc., 2017, 
Vol. 65(10), pp. 2322-2328.
---------------------------------------------------------------------------

b. Stakeholder and Technical Expert Panel (TEP) Input
    The proposed measure was developed after consideration of feedback 
we received from stakeholders and four TEPs convened by our 
contractors. Further, the proposed measure was developed after 
evaluation of data collected during two pilot tests we conducted in 
accordance with the CMS Measures Management System Blueprint.
    Our measure development contractors constituted a TEP which met on 
September 27, 2016,\73\ January 27, 2017,\74\ and August 3, 2017 \75\ 
to provide input on a prior version of this measure. Based on this 
input, we updated the measure concept in late 2017 to include the 
transfer of a specific component of health information--medication 
information. Our measure development contractors reconvened this TEP on 
April 20, 2018 to seek expert input on the measure. Overall, the TEP 
members supported the proposed measure, affirming that the measure 
provides an opportunity to improve the transfer of medication 
information. Most of the TEP members believed that the measure could 
improve the transfer of medication information to patients, families, 
and caregivers. Several TEP members emphasized the importance of 
transferring information to patients and their caregivers in a clear 
manner using plain language. A summary of the April 20, 2018 TEP 
proceedings titled ``Transfer of Health Information TEP Meeting 4--June 
2018'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
---------------------------------------------------------------------------

    \73\ Technical Expert Panel Summary Report: Development of two 
quality measures to satisfy the Improving Medicare Post-Acute Care 
Transformation Act of 2014 (IMPACT Act) Domain of Transfer of health 
Information and Care Preferences When an Individual Transitions to 
Skilled Nursing Facilities (SNFs), Inpatient Rehabilitation 
Facilities (IRFs), Long Term Care Hospitals (LTCHs) and Home Health 
Agencies (HHAs). Available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/Downloads/Transfer-of-Health-Information-TEP_Summary_Report_Final-June-2017.pdf.
    \74\ Technical Expert Panel Summary Report: Development of two 
quality measures to satisfy the Improving Medicare Post-Acute Care 
Transformation Act of 2014 (IMPACT Act) Domain of Transfer of health 
Information and Care Preferences When an Individual Transitions to 
Skilled Nursing Facilities (SNFs), Inpatient Rehabilitation 
Facilities (IRFs), Long Term Care Hospitals (LTCHs) and Home Health 
Agencies (HHAs). Available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/Downloads/Transfer-of-Health-Information-TEP-Meetings-2-3-Summary-Report_Final_Feb2018.pdf.
    \75\ Ibid.
---------------------------------------------------------------------------

    Our measure development contractors solicited stakeholder feedback 
on the proposed measure by requesting comment on the CMS Measures 
Management System Blueprint website, and accepted comments that were 
submitted from March 19, 2018 to May 3, 2018. Several commenters noted 
the importance of ensuring that the instruction provided to patients 
and caregivers is clear and understandable to promote transparent 
access to medical record information and meet the goals of the IMPACT 
Act. The summary report for the March 19 to May 3, 2018 public comment 
period titled ``IMPACT--Medication Profile Transferred Public Comment 
Summary Report'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
c. Pilot Testing
    Between June and August 2018, we held a pilot test involving 24 PAC 
facilities/agencies, including five IRFs, six SNFs, six LTCHs, and 
seven HHAs. The 24 pilot sites submitted a total of 801 assessments. 
Analysis of agreement between coders within each participating facility 
(241 qualifying pairs) indicated an 87 percent agreement for this 
measure. Overall, pilot testing enabled us to verify its reliability, 
components of face validity, and feasibility of being implemented 
across PAC settings. Further, more than half of the sites that 
participated in the pilot test stated, during debriefing interviews, 
that the measure could distinguish facilities or agencies with higher 
quality medication information transfer from those with lower quality 
medication information transfer at discharge. The pilot test summary 
report titled ``Transfer of Health Information 2018 Pilot Test Summary 
Report'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.

[[Page 39107]]

d. Measure Applications Partnership (MAP) Review and Related Measures
    We included the proposed measure in the IRF QRP section of the 2018 
MUC list. The MAP conditionally supported this measure pending NQF 
endorsement, noting that the measure can promote the transfer of 
important medication information to the patient. The MAP recommended 
that providers transmit medication information to patients that is easy 
to understand because health literacy can impact a person's ability to 
take medication as directed. More information about the MAP's 
recommendations for this measure is available at http://www.qualityforum.org/Publications/2019/02/MAP_2019_Considerations_for_Implementing_Measures_Final_Report_-_PAC-LTC.aspx.
    Section 1886(j)(7)(D)(i) of the Act, requires that any measure 
specified by the Secretary be endorsed by the entity with a contract 
under section 1890(a) of the Act, which is currently the NQF. However, 
when a feasible and practical measure has not been NQF endorsed for a 
specified area or medical topic determined appropriate by the 
Secretary, section 1886(j)(7)(D)(ii) of the Act allows the Secretary to 
specify a measure that is not NQF endorsed as long as due consideration 
is given to the measures that have been endorsed or adopted by a 
consensus organization identified by the Secretary. Therefore, in the 
absence of any NQF-endorsed measures that address the proposed Transfer 
of Health Information to the Patient--Post-Acute Care (PAC), which 
requires that at least some of the data used to calculate the measure 
be collected as standardized patient assessment data through PAC 
assessment instruments, we believe that there is currently no feasible 
NQF-endorsed measure that we could adopt under section 
1886(j)(7)(D)(ii) of the Act. However, we note that we intend to submit 
the proposed measure to the NQF for consideration of endorsement when 
feasible.
e. Quality Measure Calculation
    The calculation of the proposed Transfer of Health Information to 
the Patient--Post-Acute Care (PAC) measure would be based on the 
proportion of patient stays with a discharge assessment indicating that 
a current reconciled medication list was provided to the patient, 
family, or caregiver at the time of discharge.
    The proposed measure denominator is the total number of IRF patient 
stays ending in discharge to a private home/apartment, a board and care 
home, assisted living, a group home, transitional living or home under 
care of an organized home health service organization, or a hospice. 
These locations were selected for inclusion in the denominator because 
they are identified as home locations on the discharge destination item 
that is currently included on the IRF-PAI. The proposed measure 
numerator is the number of IRF patient stays with an IRF-PAI discharge 
assessment indicating a current reconciled medication list was provided 
to the patient, family, or caregiver at the time of discharge. For 
technical information about this proposed measure, we refer readers to 
the document titled ``Final Specifications for IRF QRP Quality Measures 
and Standardized Patient Assessment Data Elements,'' available at 
https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html. Data for the proposed quality 
measure would be calculated using data from the IRF-PAI assessment 
instrument for IRF patients.
    For more information about the data submission requirements we 
proposed for this measure, we refer readers to section VIII.G.3. of 
this rule.
    Commenters submitted the following comments related to the proposed 
rule's discussion of the IRF QRP Quality Measure Proposals Beginning 
with the FY 2022 IRF QRP. A discussion of these comments, along with 
our responses, appears below. We received many comments that addressed 
both of the Transfer of Health Information measures. Comments that 
applied to both measures are discussed above in IX.D.1 of this rule.
    Comment: One commenter suggested that CMS use the field's 
experience with transferring information to patients and reporting on 
this measure to disseminate best practices about how to best convey the 
medication list and suggested this include formats and informational 
elements helpful to patients and families.
    Response: We have interpreted ``the field'' to mean PAC providers. 
Facilities and clinicians should use clinical judgement to guide their 
practices around transferring information to patients and how to best 
convey the medication list, including identifying the best formats and 
informational elements. This may be determined by the patient's 
individualized needs in response to their medical condition. We do not 
determine clinical best practices standards and facilities are advised 
to refer to other sources, such as professional guidelines.
    Comment: One commenter suggested that the Transfer of Health 
Information to the Patient--Post-Acute Care (PAC) Measure require 
transfer of the medication list to both the patient and family or 
caregiver.
    Response: We agree there are times when it is appropriate for the 
IRF to provide the medication list to the patient and family and this 
decision should be based on clinical judgement. However, because it is 
not always necessary or appropriate to provide the medication list to 
both the patient and family, we are not requiring this for the measure.
    After consideration of the public comments, we are finalizing our 
proposal to adopt the Transfer of Health Information to the Patient--
Post Acute Care (PAC) measure, under section 1899B(c)(1)(E) of the Act, 
with data collection for discharges beginning October 1, 2020.
3. Update to the Discharge to Community--Post Acute Care (PAC) 
Inpatient Rehabilitation Facility (IRF) Quality Reporting Program (QRP) 
Measure
    In the FY 2020 IRF PPS proposed rule (84 FR 17291), we proposed to 
update the specifications for the Discharge to Community--PAC IRF QRP 
measure to exclude baseline nursing facility (NF) residents from the 
measure. This measure reports an IRF's risk-standardized rate of 
Medicare FFS patients who are discharged to the community following an 
IRF stay, do not have an unplanned readmission to an acute care 
hospital or LTCH in the 31 days following discharge to community, and 
who remain alive during the 31 days following discharge to community. 
We adopted this measure in the FY 2017 IRF PPS final rule (81 FR 52095 
through 52103).
    In the FY 2017 IRF PPS final rule (81 FR 52099), we addressed 
public comments recommending exclusion of IRF patients who were 
baseline NF residents, as these patients lived in a NF prior to their 
IRF stay, as these patients may not be expected to return to the 
community following their IRF stay. In the FY 2018 IRF PPS final rule 
(82 FR 36285), we addressed public comments expressing support for a 
potential future modification of the measure that would exclude 
baseline NF residents; commenters stated that the exclusion would 
result in the measure more accurately portraying quality of care 
provided by IRFs, while controlling for factors outside of IRF control.

[[Page 39108]]

    We assessed the impact of excluding baseline NF residents from the 
measure using CY 2015 and CY 2016 data, and found that this exclusion 
impacted both patient- and facility-level discharge to community rates. 
We defined baseline NF residents as IRF patients who had a long-term NF 
stay in the 180 days preceding their hospitalization and IRF stay, with 
no intervening community discharge between the NF stay and qualifying 
hospitalization for measure inclusion. Baseline NF residents 
represented 0.3 percent of the measure population after all measure 
exclusions were applied. Observed patient-level discharge to community 
rates were significantly lower for baseline NF residents (20.82 
percent) compared with non-NF residents (64.52 percent). The national 
observed patient-level discharge to community rate was 64.41 percent 
when baseline NF residents were included in the measure, increasing to 
64.52 percent when they were excluded from the measure. After excluding 
baseline NF residents, 26.9 percent of IRFs had an increase in their 
risk-standardized discharge to community rate that exceeded the 
increase in the national observed patient-level discharge to community 
rate.
    Based on public comments received and our impact analysis, we 
proposed to exclude baseline NF residents from the Discharge to 
Community--PAC IRF QRP measure beginning with the FY 2020 IRF QRP, with 
baseline NF residents defined as IRF patients who had a long-term NF 
stay in the 180 days preceding their hospitalization and IRF stay, with 
no intervening community discharge between the NF stay and 
hospitalization.
    For additional technical information regarding the Discharge to 
Community--PAC IRF QRP measure, including technical information about 
the proposed exclusion, we refer readers to the document titled ``Final 
Specifications for IRF QRP Quality Measures and Standardized Patient 
Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We sought public comment on this proposal and received several 
comments. A discussion of these comments, along with our responses, 
appears below.
    Comment: Several commenters supported the proposed exclusion of 
baseline NF residents from the Discharge to Community--PAC IRF QRP 
measure. Commenters referred to their recommendation of this exclusion 
in prior years and appreciated CMS' willingness to consider and 
implement stakeholder feedback. One commenter stated they did not 
foresee any negative impacts of the exclusion. One commenter suggested 
that CMS instead consider other quality measures for NF residents, such 
as functional status measures, to determine whether residents receive 
the appropriate standard of care they need in a long-term NF stay.
    Response: We thank the commenters for their support of the proposed 
exclusion of baseline nursing facility residents from this measure and 
for recommending other measures for consideration for baseline NF 
residents.
    Comment: MedPAC did not support the proposed exclusion of baseline 
nursing facility residents from the Discharge to Community--PAC IRF QRP 
measure. They suggested that CMS instead expand their definition of 
``return to the community'' to include baseline nursing home residents 
returning to the nursing home where they live, as this represents their 
home or community. MedPAC also stated that providers should be held 
accountable for the quality of care they provide for as much of their 
Medicare patient population as feasible.
    Response: We agree that providers should be accountable for quality 
of care for as much of their Medicare population as feasible; we 
endeavor to do this as much as possible, only specifying exclusions we 
believe are necessary for measure validity. We also believe that 
monitoring quality of care and outcomes is important for all PAC 
patients, including baseline NF residents who return to a NF after 
their PAC stay. We publicly report several long-stay resident quality 
measures on Nursing Home Compare including measures of hospitalization 
and emergency department visits.
    Community is traditionally understood as representing non-
institutional settings by policy makers, providers, and other 
stakeholders. Including long-term care NF in the definition of 
community would confuse this long-standing concept of community and 
would misalign with CMS' definition of community in patient assessment 
instruments. We conceptualized this measure using the traditional 
definition of ``community'' and specified the measure as a discharge to 
community measure, rather than a discharge to baseline residence 
measure.
    Baseline NF residents represent an inherently different patient 
population with not only a significantly lower likelihood of discharge 
to community settings, but also a higher likelihood of post-discharge 
readmissions and death compared with PAC patients who did not live in a 
NF at baseline. The inherent differences in patient characteristics and 
PAC processes and goals of care for baseline NF residents and non-NF 
residents are significant enough that we do not believe risk adjustment 
using a NF flag would provide adequate control. While we acknowledge 
that a return to nursing home for baseline NF residents represents a 
return to their home, this outcome does not align with our measure 
concept. Thus, we have chosen to exclude baseline NF residents from the 
measure.
    Comment: One commenter suggested the definition of ``long-term'' NF 
stay in the proposed measure exclusion, requesting further 
clarification in the measure specifications.
    Response: We have further clarified the definition of long-term NF 
stay in the final measure specifications, Final Specifications for IRF 
QRP Quality Measures and Standardized Patient Assessment Data 
Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html. A 
long-term NF stay is identified by the presence of a non-SNF PPS MDS 
assessment in the 180 days preceding the qualifying prior acute care 
admission and index SNF stay.
    Comment: One commenter questioned whether the methodology for 
calculating confidence intervals for performance categories used in 
public display of the Discharge to Community--PAC measures has been 
updated.
    Response: On May 31, 2019, we announced an update to the 
methodology used for calculating confidence intervals for provider 
assignment to performance categories for public display of the 
Discharge to Community--PAC measures. For more information, we refer 
readers to the ``Fact Sheet for Discharge to Community Post-Acute Care 
Measures'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/LTCH-Quality-Reporting/Downloads/Fact-Sheet-for-Discharge-to-Community-Post-Acute-Care-Measures.pdf and the ``FAQ for Discharge to Community Post-Acute Care 
Measures'' available at https://www.cms.gov/Medicare/Quality-
Initiatives-Patient-Assessment-Instruments/LTCH-Quality-Reporting/
Downloads/FAQ-for-Discharge-to-

[[Page 39109]]

Community-Post-Acute-Care-Measures.pdf.
    After consideration of the public comments, we are finalizing our 
proposal to exclude baseline NF residents from the Discharge to 
Community--PAC IRF QRP measure as proposed beginning with the FY 2020 
IRF QRP.

E. IRF QRP Quality Measures, Measure Concepts, and Standardized Patient 
Assessment Data Elements Under Consideration for Future Years: Request 
for Information

    We sought input on the importance, relevance, appropriateness, and 
applicability of each of the measures, standardized patient assessment 
data elements (SPADEs), and concepts under consideration listed in the 
Table 17 for future years in the IRF QRP.
[GRAPHIC] [TIFF OMITTED] TR08AU19.021

    While we will not be responding to specific comments submitted in 
response to this Request for Information, we intend to use this input 
to inform our future measure and SPADE development efforts.
    We received several comments on this RFI, which are summarized 
below.
    Comment: Several commenters supported the inclusion of all of the 
proposed measures and SPADEs listed in Table 17. One commenter agreed 
that the SPADE categories will provide a fuller picture of the patients 
in the IRF setting and could be used for creating and risk adjusting 
quality measures.
    Many commenters supported the dementia SPADE, since dementia can 
affect a beneficiary's ability to participate in his or her care in the 
PAC setting, in addition to managing chronic conditions and medications 
after discharge. One commenter also agreed that regularly assessing 
cognitive function and mental health status presents opportunities for 
better care and quality of life.
    One commenter did not support the cognitive complexity SPADEs, 
since there is no singular assessment tool designed to assess executive 
function and memory, and it would be overly burdensome for IRFs to 
conduct testing on every patient. The commenter recommended that CMS 
work with stakeholders to prioritize which patient conditions would 
benefit from a cognitive complexity assessment and screen for those 
cases.
    Many commenters supported the caregiver status SPADE; one commenter 
stated that regular assessment of caregivers will result in better care 
for the beneficiary and quality of life for both individuals. Another 
commenter encouraged CMS to capture caregiver status, along with the 
caregiver's willingness and ability, and account for it in discharge 
disposition outcomes.
    With regard to an opioids-based quality measure, providers had some 
concerns about unintended consequences of reporting of opioid use, 
including the over- or under-prescribing of opioids or limiting 
patients access to critical treatments for pain management.
    Many commenters were supportive of SPADEs focused on bowel and 
bladder continence. One commenter noted that this is already collected 
on admission and did not support a bowel and bladder SPADE on 
discharge, citing that IRFs already communicate continence needs at 
discharge and this would be duplicative. A few commenters had concerns 
about the burden of future measures and SPADEs. One commenter 
recommended that prior to adding measures or data elements, CMS 
reassess and analyze all of the measures and data elements currently 
collected to limit administrative burden and create a meaningful set of 
measures and data elements. Another commenter supported utilization of 
data from the suggested measures and SPADEs and suggested using 
existing data sources, such a Medicare claims data. One commenter did 
not support any future SPADE concepts that were not required by the 
IMPACT Act. Another commenter suggested that CMS should explore 
beneficiary-matching methods with the Department of Veteran's Affairs 
to collect veteran status without additional IRF data collection 
burden.
    Response: We appreciate the input provided by commenters. While we 
will not be responding to specific comments submitted in response to 
this Request for Information, we intend to use this input to inform our 
future measure and SPADE development efforts.

F. Standardized Patient Assessment Data Reporting Beginning With the FY 
2022 IRF QRP

    Section 1886(j)(7)(F)(ii) of the Act requires that, for FY 2019 and 
each subsequent fiscal year, IRFs must report standardized patient 
assessment data required under section 1899B(b)(1) of the Act. Section 
1899B(a)(1)(C) of the Act requires, in part, the Secretary to modify 
the PAC assessment instruments in order for PAC providers, including 
IRFs, to submit SPADEs under the Medicare program. Section 
1899B(b)(1)(A) of the Act requires PAC providers to submit SPADEs under 
applicable reporting provisions (which, for IRFs, is the IRF QRP) with 
respect to the admission and discharge of an

[[Page 39110]]

individual (and more frequently as the Secretary deems appropriate), 
and section 1899B(b)(1)(B) of the Act defines standardized patient 
assessment data as data required for at least the quality measures 
described in section 1899B(c)(1) of the Act and that is with respect to 
the following categories: (1) Functional status, such as mobility and 
self-care at admission to a PAC provider and before discharge from a 
PAC provider; (2) cognitive function, such as ability to express ideas 
and to understand, and mental status, such as depression and dementia; 
(3) special services, treatments, and interventions, such as need for 
ventilator use, dialysis, chemotherapy, central line placement, and 
total parenteral nutrition; (4) medical conditions and comorbidities, 
such as diabetes, congestive heart failure, and pressure ulcers; (5) 
impairments, such as incontinence and an impaired ability to hear, see, 
or swallow; and (6) other categories deemed necessary and appropriate 
by the Secretary.
    In the FY 2018 IRF PPS proposed rule (82 FR 20722 through 20739), 
we proposed to adopt SPADEs that would satisfy the first five 
categories. In the FY 2018 IRF PPS final rule (82 FR 36287 through 
36289), we summarized comments that supported our adoption of SPADEs, 
including support for our broader standardization goal and support for 
the clinical usefulness of specific proposed SPADEs. However, we did 
not finalize the majority of our SPADE proposals in recognition of the 
concern raised by many commenters that we were moving too fast to adopt 
the SPADEs and modify our assessment instruments in light of all of the 
other requirements we were also adopting under the IMPACT Act at that 
time (82 FR 36292 through 36294). In addition, commenters noted that we 
should conduct further testing of the data elements we have proposed 
(82 FR 36288).
    However, we finalized the adoption of SPADEs for two of the 
categories described in section 1899B(b)(1)(B) of the Act: (1) 
Functional status: Data elements currently reported by IRFs to 
calculate the measure Application of Percent of Long-Term Care Hospital 
Patients with an Admission and Discharge Functional Assessment and a 
Care Plan That Addresses Function (NQF #2631); and (2) Medical 
conditions and comorbidities: The data elements used to calculate the 
pressure ulcer measures, Percent of Residents or Patients with Pressure 
Ulcers That Are New or Worsened (Short Stay) (NQF #0678) and the 
replacement measure, Changes in Skin Integrity Post-Acute Care: 
Pressure Ulcer/Injury. We stated that these data elements were 
important for care planning, known to be valid and reliable, and 
already being reported by IRFs for the calculation of quality measures.
    Since we issued the FY 2018 IRF PPS final rule, IRFs have had an 
opportunity to familiarize themselves with other new reporting 
requirements that we have adopted under the IMPACT Act. We have also 
conducted further testing of the SPADEs, as described more fully below, 
and believe that this testing supports the use of the SPADEs in our PAC 
assessment instruments. Therefore, we proposed to adopt many of the 
same SPADEs that we previously proposed to adopt, along with other 
SPADEs.
    We proposed that IRFs would be required to report these SPADEs 
beginning with the FY 2022 IRF QRP. If finalized as proposed, IRFs 
would be required to report these data with respect to admission and 
discharge for Medicare Part A and Medicare Advantage patients 
discharged between October 1, 2020, and December 31, 2020 for the FY 
2022 IRF QRP. Beginning with the FY 2023 IRF QRP, we proposed that IRFs 
must report data with respect to Medicare Part A and Medicare Advantage 
admissions and discharges that occur during the subsequent calendar 
year (for example, CY 2021 for the FY 2023 IRF QRP, CY 2022 for the FY 
2024 IRF QRP).
    We also proposed that IRFs that submit the Hearing, Vision, Race, 
and Ethnicity SPADEs with respect to admission will be deemed to have 
submitted those SPADEs with respect to both admission and discharge, 
because it is unlikely that the assessment of those SPADEs at admission 
will differ from the assessment of the same SPADEs at discharge.
    In selecting the proposed SPADEs below, we considered the burden of 
assessment-based data collection and aimed to minimize additional 
burden by evaluating whether any data that is currently collected 
through one or more PAC assessment instruments could be collected as 
SPADEs. In selecting the SPADEs below, we also took into consideration 
the following factors with respect to each data element:
    (1) Overall clinical relevance;
    (2) Interoperable exchange to facilitate care coordination during 
transitions in care;
    (3) Ability to capture medical complexity and risk factors that can 
inform both payment and quality; and
    (4) Scientific reliability and validity, general consensus 
agreement for its usability.
    In identifying the SPADEs proposed below, we additionally drew on 
input from several sources, including TEPs held by our data element 
contractor, public input, and the results of a recent National Beta 
Test of candidate data elements conducted by our data element 
contractor (hereafter ``National Beta Test'').
    The National Beta Test collected data from 3,121 patients and 
residents across 143 PAC facilities (26 LTCHs, 60 SNFs, 22 IRFs, and 35 
HHAs) from November 2017 to August 2018 to evaluate the feasibility, 
reliability, and validity of the candidate data elements across PAC 
settings. The 3,121 patients and residents with an admission assessment 
included 507 in LTCHs, 1,167 in SNFs, 794 in IRFs, and 653 in HHAs. The 
National Beta Test also gathered feedback on the candidate data 
elements from staff who administered the test protocol in order to 
understand usability and workflow of the candidate data elements. More 
information on the methods, analysis plan, and results for the National 
Beta Test can be found in the document titled, ``Development and 
Evaluation of Candidate Standardized Patient Assessment Data Elements: 
Findings from the National Beta Test (Volume 2),'' available in the 
document titled, ``Development and Evaluation of Candidate Standardized 
Patient Assessment Data Elements: Findings from the National Beta Test 
(Volume 2),'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Further, to inform the proposed SPADEs, we took into account 
feedback from stakeholders, as well as from technical and clinical 
experts, including feedback on whether the candidate data elements 
would support the factors described above. Where relevant, we also took 
into account the results of the Post-Acute Care Payment Reform 
Demonstration (PAC PRD) that took place from 2006 to 2012.
    Comment: Several commenters were supportive of the SPADE proposals. 
A commenter recognized that the proposed SPADEs may influence care, 
impact case mix and risk adjustment scores, and drive planning for 
future management. Other commenters supported the proposals to add the 
proposed SPADEs to the IRF-PAI, with one noting that many of the data 
elements are already collected and reported on, and the other stating 
that the items are important to describing current IRF patients and are 
applicable to determining patient acuity. Another

[[Page 39111]]

commenter stated that data standardization as accomplished by the 
SPADEs will help facilitate appropriate payment reforms and appropriate 
quality measures.
    Response: We thank the commenters for their support. We selected 
the proposed SPADEs in part because of the attributes that the 
commenters noted, such as their ability to describe IRF patients and to 
support future quality measurement.
    Comment: Some commenters stated support but noted reservations. One 
commenter described the SPADEs as an appropriate start, but noted that 
the SPADEs cannot stand alone, and must be built upon in order to be 
useful for risk adjustment and quality measurement. Similarly, another 
commenter suggested CMS continue working with clinicians and 
researchers to ensure that the SPADEs are collecting valid, reliable, 
and useful data, and to continue to refine and explore new data 
elements for standardization.
    Response: We agree with the commenter's statement that the SPADEs 
are an appropriate start for standardization, but we disagree that they 
cannot stand alone. While we intend to evaluate the SPADEs as they are 
submitted and explore additional opportunities for standardization, we 
also believe that the SPADEs as proposed represent an important core 
set of information about clinical status and patient characteristics 
and they will be useful for quality measurement. We will continue to 
explore the use of the SPADEs across our PAC setting, continuing our 
efforts to explore the feasibility, reliability, validity, and 
usability of the data elements in our measure models and QRPs. We would 
welcome continued input, recommendations, and feedback from 
stakeholders about ways to improve assessment and quality measurement 
for PAC providers, including ways that the SPADEs could be used in the 
IRF QRP. Input can be shared with CMS through our PAC Quality 
Initiatives email address [email protected].
    Comment: One commenter noted support for the goals of the IMPACT 
Act, but expressed concern about the scope and timing of proposed 
changes, including the SPADEs. The same commenter suggested that CMS 
share with the public a data use strategy and analysis plan for the 
SPADEs so that providers better understand how CMS will assess the 
potential usability of the SPADEs to support changes to payment and 
quality programs.
    Response: We thank the commenter for the support and appreciate 
their concern about the proposed changes. We intend to monitor and 
evaluate SPADEs as they are submitted, and to continue to engage 
stakeholders around ways the SPADEs could be best used in the PAC 
quality programs. We will continue to communicate and collaborate with 
stakeholders by soliciting input on use of the SPADEs in the IRF QRP 
through future rulemaking.
    Comment: One commenter was generally critical of the set of SPADEs 
proposed, stating they fail to adequately describe a patient's clinical 
situation with regard to their level of independence, including 
swallowing function, communication, and cognitive function.
    Response: The proposed SPADEs were selected based on their overall 
clinical relevance to PAC providers, including IRFs, their ability to 
facilitate care coordination during transitions, their ability to 
capture medical complexity and risk factors, and their scientific 
reliability and validity. We have strived to balance the scope and 
level of detail of the data elements against the potential burden 
placed on patients and providers. At this time, SPADEs focused on 
impairments are limited to sensory impairments (that is, hearing and 
vision) and do not include swallowing. The patient's ability to 
communicate is also not captured with a SPADE, although we note that 
the IRF-PAI includes two data elements on communication: Expression of 
Ideas and Wants, and Understanding Verbal and Non-Verbal Content. 
However, in combination with other sections of the IRF-PAI that have 
been standardized across PAC providers, we believe the proposed SPADEs 
capture key clinical information (for example, cognitive function for 
patients who are able to communicate, as collected by the BIMS) and 
form an important foundation of standardized assessment on which to 
build.
    Comment: One commenter described several concerns about the scope 
and implementation of the National Beta Test, including the 
representativeness of IRFs included in the sample, the share of total 
IRF patients included in the National Beta Test, the reported exclusion 
of patients with communication and cognitive impairments, and the 
exclusion of non-English speaking patients, and described how these 
concerns compromise their confidence in the findings of the National 
Beta Test.
    Response: In a supplementary document to the proposed rule, we 
described key findings from the National Beta Test related to the 
proposed SPADEs. We also referred readers to an initial volume of the 
National Beta Test report that details the methodology of the field 
test (``Development and Evaluation of Candidate Standardized Patient 
Assessment Data Elements: Findings from the National Beta Test (Volume 
2),'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html). Additional 
volumes of the National Beta Test report will be available in late 
2019.
    To address the commenter's specific concerns, we note that the 
National Beta Test was designed to generate valid and robust national 
SPADE performance estimates for each of the four PAC provider types, 
which required acceptable geographic diversity, sufficient sample size, 
and reasonable coverage of the range of clinical characteristics. To 
meet these requirements, the National Beta Test was carefully designed 
so that data could be collected from a wide range of environments, 
allowing for thorough evaluation of candidate SPADE performance in all 
PAC settings. The approach included a stratified random sample, to 
maximize generalizability, and subsequent analyses included extensive 
checks on the sampling design.
    The commenter further implied that the small share of overall IRF 
admissions included in the Beta test is indicative of inadequate 
representativeness. The objective of the National Beta Test was to 
evaluate the performance of candidate SPADEs for cross-setting use. It 
is true that the proportion of IRFs may not reflect actual proportion 
in the United States, but our sampling design ensured that sufficient 
spread of IRFs across randomly selected markets, and adequate numbers 
to provide ample data with which to evaluate SPADE performance in IRFs 
relative to other settings.
    The National Beta Test did not exclude non-communicative patients/
residents; rather, it had two distinct samples, one of which focused on 
patients/residents who were able to communicate, and one of which 
focused on patient/residents who were not able to communicate. The 
assessment of non-communicative patients/residents differed primarily 
in that observational assessments were substituted for some interview 
assessments. Non-English-speaking patients were excluded from the 
National Beta Test due to feasibility constraints during the field 
test. Including limited English proficiency patients/residents in the 
sample would

[[Page 39112]]

have required the Beta test facilities to engage or involve translators 
during the test assessments. We anticipated that this would have added 
undue complexity to what facilities/agencies were being requested to 
do, and would have undermined the ability of facility/agency staff to 
complete the requested number of assessments during the study period. 
Moreover, there is strong existing evidence for the feasibility of all 
clinical patient/resident interview SPADEs included in this final rule 
(BIMS [section IX.G.1 in this final rule], Pain Interference [section 
IX.G.3 in this final rule], PHQ [section IX.G.1 in this final rule]) 
when administered in other languages, either through standard PAC 
workflow, as tested and currently collected in the MDS 3.0, or through 
rigorous translation and testing, such as the PHQ. For all these 
reasons, we determined that the performance of translated versions of 
these patient/resident interview SPADEs did not need to be further 
evaluated. In addition, because their exclusion did not threaten our 
ability to achieve acceptable geographic diversity, sufficient sample 
size, and reasonable coverage of the range of PAC patient/resident 
clinical characteristics, the exclusion of limited English proficiency 
patients/residents was not considered a limitation to interpretation of 
the National Beta Test results.
    Comment: Two commenters wanted CMS to share more information from 
the National Beta Test. One of the commenters remarked on the lack of 
information about clinical characteristics that has been shared with 
stakeholders, limiting their ability to draw conclusions about the 
data, and requested that CMS release the data from the National Beta 
Test to be analyzed by third parties. The other commenter noted that 
CMS has not shared quantitative results of the National Beta Test which 
has limited the ability of stakeholders to determine if these items 
will yield useful information for quality and/or payment purposes, and 
suggested CMS release additional information, such as response 
frequencies, and analysis from the field test to provide evidence of 
the validity and utility of the SPADEs for quality and payment.
    Response: We shared both quantitative and qualitative findings from 
the National Beta Test with stakeholders at a public meeting on 
November 27, 2018. For each SPADE proposed in this rule within the 
clinical categories in the IMPACT Act, we provided information in the 
supplementary documents to the proposed rule (the document titled 
``Proposed Specifications for IRF QRP Quality Measures and Standardized 
Patient Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html) on the feasibility and reliability based on findings from 
the National Beta Test.
    We are in the process of writing the final report for the National 
Beta Test, which includes the clinical SPADEs in this rule as well as 
additional data elements. Volume 2 of that report (``Development and 
Evaluation of Candidate Standardized Patient Assessment Data Elements. 
Findings from the National Beta Test (Volume 2)'') was posted on CMS' 
website in March 2019. The other volumes will be available in late 
2019. In addition, we are committed to making data available for 
researchers and the public to analyze, and to doing so in a way that 
protects the privacy of patients and providers who participated in the 
National Beta Test. We are in the process of creating research 
identifiable files that we anticipate will be available through a data 
use agreement sometime in 2019.
    Comment: Many commenters expressed concerns with respect to the 
standardized patient assessment data proposals. Several commenters 
stated that the standardized patient assessment data reporting 
requirements will impose significant burden on providers, given the 
volume of new standardized patient assessment data elements, and 
corresponding sub-elements, that were proposed to be added to the IRF-
PAI. One commenter noted that the addition of the proposed standardized 
patient assessment data elements would require an expanded timeline to 
implement to ensure necessary operational and workflow revisions.
    Response: We acknowledge the additional burden that the SPADEs will 
impose on providers and patients. Our development and selection process 
for the SPADEs we are adopting in this final rule prioritized data 
elements that are essential to comprehensive patient care. We maintain 
that there will be significant benefit associated with each of the 
SPADEs to providers and patients, in that they are clinically useful 
(for example, for care planning), they support patient-centered care, 
and they will promote interoperability and data exchange between 
providers. During the SPADE development process, we were cognizant of 
the changes that providers will need to make to implement these 
additions to the IRF-PAI. In the last two rules (82 FR 36287 through 
36289, 83 FR 38555), we provided information about goals, scope, and 
timeline for implementing SPADEs, as well as updated IRFs about ongoing 
development and testing of data elements through other public forums. 
We believe that IRFs have had an opportunity to familiarize themselves 
with other new reporting requirements that we have adopted under the 
IMPACT Act and prepare for additional changes.
    Comment: Some commenters expressed concern that this additional 
burden was not justified because, in their view, there was limited or 
no evidence for the SPADEs to describe case mix, measure quality, or 
improve care. One of these commenters noted that CMS has provided 
evidence of validity, reliability, and feasibility through documents 
related to the National Beta Test, but stated that CMS has not provided 
any evidence that the proposed SPADEs have the ``potential for 
improving quality'' or ``utility for describing case mix.''
    Response: The clinical SPADEs proposed in this rule were the result 
of an extensive consensus vetting process in which experts and 
stakeholders were engaged through Technical Expert Panels, Special Open 
Door Forums, and posting of interim reports and other documents on the 
CMS website. Results of these activities provide evidence that experts 
and providers believe that the proposed SPADEs have the potential for 
measuring quality, for describing case mix, and improving care. We 
refer the commenter to the most recent TEP report: A summary of the 
September 17, 2018 TEP meeting titled ``SPADE Technical Expert Panel 
Summary (Third Convening)'', which is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html. In this report, we summarize the TEP's discussion of 
individual SPADEs in which they reflect on the clinical usefulness and 
importance of the SPADEs for describing patient acuity (case mix) and 
providing high-quality clinical care (improving quality). Therefore, we 
have provided evidence that the SPADEs have the potential for improving 
quality and utility for describing case mix.
    Comment: One commenter believes that the expansion of the IRF-PAI 
assessment will prove to be intrusive and prove challenging for 
patients who are elderly, frail, in pain, or have cognitive deficits, 
causing the patients

[[Page 39113]]

to lose focus, and thus, impact the accuracy of the data.
    Response: We acknowledge that several SPADEs in this rule require 
the patient to be asked questions directly. We believe that direct 
patient assessment and patient-reported outcomes on these topics have 
benefits for providers and patients. These data elements support 
patient-centered care by soliciting the patient's perspective, and 
better information on a patient's status is expected to improve the 
care the patient receives.76 77 78 The burden the patient-
interview data elements place on patients is necessary for accurate 
assessment of the patient's status. Regarding the validity and 
performance of interview-based data elements, we note that many of 
these data elements (for example, the BIMS, PHQ, and Pain Interference 
data elements) are currently used in the MDS in SNFs. Evidence from 
that setting, as well as from the National Beta Test, demonstrates 
feasibility of these data elements for even very sick patients, such as 
many patients receiving care from IRFs.
---------------------------------------------------------------------------

    \76\ Boyce MB, Browne JP, Greenhalgh J The experiences of 
professionals with using information from patient-reported outcome 
measures to improve the quality of healthcare: A systematic review 
of qualitative research BMJ Quality & Safety 2014;23:508-518.
    \77\ Chen J, Ou L, Hollis SJ. A systematic review of the impact 
of routine collection of patient reported outcome measures on 
patients, providers and health organizations in an oncologic 
setting. BMC Health Services Research 2013;13:211.
    \78\ Marshall, S., Haywood, K. and Fitzpatrick, R. (2006), 
Impact of patient[hyphen]reported outcome measures on routine 
practice: A structured review. Journal of Evaluation in Clinical 
Practice, 12: 559-568. doi:10.1111/j.1365-2753.2006.00650.x.
---------------------------------------------------------------------------

    Comment: Commenters also stated that the time burden (as in, 
``time-to-complete'') associated with the clinical SPADEs was 
underestimated, with some commenters noting that it did not account for 
clinician time to review charts and update treatment plans or that test 
conditions do not represent conditions of day-to-day operation. One 
commenter stated that the estimated time to complete reported in the 
National Beta Test was based only on the time needed to enter a value 
on a tablet and did not include the time to evaluate the patient on 
each item. Another commenter stated that because testing conditions 
focused on cognitively intact, English-speaking patients with no speech 
or language deficits, the estimates of impact to providers' time and 
resources is inadequate.
    Response: We disagree with the commenters that the National Beta 
Test time-to-complete estimates are underestimates. Contrary to what 
one commenter noted, we wish to clarify that time-to-complete estimates 
from the National Beta Test included the time spent both to collect 
data, including the review of the medical record, if needed, and to 
enter the data elements into a tablet. We note that time-to-complete 
estimates were calculated using the data from Facility/Agency Staff 
only, and not Research Nurses, who completed more training and 
conducted more assessments overall than the Facility/Agency staff. This 
decision to calculate time-to-complete estimates from Facility/Agency 
Staff only supports our claim that the time-to-complete estimates are 
accurate reflections of the time the SPADEs will require when 
implemented by PAC providers in day-to-day operations. Contrary to 
another commenter's statement, we also wish to clarify that National 
Beta Test did exclude patients/residents who were not able to 
communicate in English, but did not categorically exclude patients with 
cognitive impairment or patients with speech or language deficits. 
Therefore, we believe that our estimates of time-to-complete capture 
the general population of IRF patients, including those with 
communication impairments.
    Comment: Some commenters recommended changes to when and how SPADEs 
would be collected in order to reduce administrative burden. These 
recommendations included collecting data only at admission when answers 
are unlikely to change between admission and discharge, adopting a 
staged implementation or only a subset of the proposed data elements, 
and that CMS explore options for obtaining these data via claims or 
voluntary reporting only, particularly as many of the proposed SPADEs 
are not relevant to IRF patients.
    Response: We appreciate the commenters' recommendations. To support 
data exchange between settings, and to support quality measurement, 
section 1899B(b)(1)(A) of the Act requires that the SPADEs be collected 
with respect to both admission and discharge. In the FY 2020 IRF PPS 
proposed rule (84 FR 17292), we proposed that IRFs that submit four 
SPADEs with respect to admission will be deemed to have submitted those 
SPADEs with respect to both admission and discharge, because we stated 
that it is unlikely that the assessment of those SPADEs at admission 
would differ from the assessment of the same SPADEs at discharge. We 
note that a patient's ability to hear or ability to see are more likely 
to change between admission and discharge than, for example, a 
patient's self-report of his or her race, ethnicity, preferred 
language, or need for interpreter services. The Hearing and Vision 
SPADEs are also different from the other SPADEs (that is, Race, 
Ethnicity, Preferred Language, and Interpreter Services) because 
evaluation of sensory status is a fundamental part of the ongoing 
nursing assessment conducted for IRF patients. Therefore, clinically 
significant changes that occur in a patient's hearing or vision status 
during the IRF stay would be captured as part of the clinical record 
and communicated to the next setting of care, as well as taken into 
account during discharge planning as a part of standard best practice.
    After consideration of public comments discussed in sections IX.G.4 
and IX.G.4.b in this final rule, we will deem IRFs that submit the 
Hearing, Vision, Race, Ethnicity, Preferred Language, and Interpreter 
Services SPADEs with respect to admission to have submitted with 
respect to both admission and discharge. We will take into 
consideration the recommendation to obtain patient data from claims 
data in future work.
    Comment: A commenter recommended that CMS limit the number and type 
of data elements implemented in the coming year, continue ongoing 
dialogue with stakeholders, and develop and implement a process to 
assess the value of specific indicators for all patient types. Another 
commenter recommended that CMS conduct a thorough analysis of SPADEs 
currently collected to determine if any current data elements could be 
eliminated. One commenter believed that CMS should not finalize the 
implementation of the SPADEs until they evaluate alternative means of 
data collection (such as via billing/claims data), or measures to 
reduce burden (such as removal of duplicative data elements and 
elimination of data collection at discharge).
    Response: We note that we adopted SPADEs in the last two rule 
cycles to support the adoption of the IRF Functional Outcomes Measures 
(Application of Percent of Long-Term Care Hospital Patients with an 
Admission and Discharge Functional Assessment and a Care Plan That 
Addresses Function (80 FR 47111); Change in Self-Care for Medical 
Rehabilitation Patients (80 FR 47117); Change in Mobility Score for 
Medical Rehabilitation Patients (80 FR 47118); Discharge Self-Care 
Score for Medical Rehabilitation Patients (80 FR 47119); Discharge 
Mobility Score for Medical Rehabilitation Patients (80 FR 47120)) and 
drug regimen review (Drug Regimen Review Conducted with Follow-Up for

[[Page 39114]]

Identified Issues (81 FR 52111)). We have also communicated about the 
SPADE development work with stakeholders over the last 2 years through 
SODFs held on June 20, 2017, September 28, 2017, December 12, 2017, 
March 28, 2018, June 19, 2018, and July 25, 2018, and at a public 
meeting of stakeholders on November 27, 2018. Therefore, our 
implementation to date has been incremental while we have strived to 
keep stakeholders apprised as to the status of ongoing SPADE 
development. We have also conducted a large-scale test of feasibility 
and reliability--the National Beta Test, described in the proposed rule 
(84 FR 17293)--which, along with the consensus vetting activities 
described in the proposals for each SPADE, provide evidence of the 
value of the SPADEs for patients across PAC settings, including IRF 
patients. We will monitor and conduct analysis on the SPADEs as they 
are submitted in order to identify any problems and to identify any 
unnecessary burden or duplication.
    Comment: One commenter recommended that CMS focus on providing 
funding and administrative support to allow improvements and 
standardization to the electronic medical record to allow effective 
interoperability across all post-acute sites.
    Response: We appreciate the commenter's recommendation. At this 
time, funding for electronic medical record adoption and support is not 
currently authorized for PAC providers.
    Final decisions on the SPADEs are given below, following more 
detailed comments on each SPADE proposal.

G. Standardized Patient Assessment Data by Category

1. Cognitive Function and Mental Status Data
    A number of underlying conditions, including dementia, stroke, 
traumatic brain injury, side effects of medication, metabolic and/or 
endocrine imbalances, delirium, and depression, can affect cognitive 
function and mental status in PAC patient and resident populations.\79\ 
The assessment of cognitive function and mental status by PAC providers 
is important because of the high percentage of patients and residents 
with these conditions,\80\ and because these assessments provide 
opportunity for improving quality of care.
---------------------------------------------------------------------------

    \79\ National Institute on Aging. (2014). Assessing Cognitive 
Impairment in Older Patients. A Quick Guide for Primary Care 
Physicians. Retrieved from https://www.nia.nih.gov/alzheimers/publication/assessing-cognitive-impairment-older-patients.
    \80\ Gage B., Morley M., Smith L., et al. (2012). Post-Acute 
Care Payment Reform Demonstration (Final report, Volume 4 of 4). 
Research Triangle Park, NC: RTI International.
---------------------------------------------------------------------------

    Symptoms of dementia may improve with pharmacotherapy, occupational 
therapy, or physical activity,81 82 83 and promising 
treatments for severe traumatic brain injury are currently being 
tested.\84\ For older patients and residents diagnosed with depression, 
treatment options to reduce symptoms and improve quality of life 
include antidepressant medication and 
psychotherapy,85 86 87 88 and targeted services, such as 
therapeutic recreation, exercise, and restorative nursing, to increase 
opportunities for psychosocial interaction.\89\
---------------------------------------------------------------------------

    \81\ Casey D.A., Antimisiaris D., O'Brien J. (2010). Drugs for 
Alzheimer's Disease: Are They Effective? Pharmacology & 
Therapeutics, 35, 208-11.
    \82\ Graff M.J., Vernooij-Dassen M.J., Thijssen M., Dekker J., 
Hoefnagels W.H., Rikkert M.G.O. (2006). Community Based Occupational 
Therapy for Patients with Dementia and their Care Givers: Randomised 
Controlled Trial. BMJ, 333(7580): 1196.
    \83\ Bherer L., Erickson K.I., Liu-Ambrose T. (2013). A Review 
of the Effects of Physical Activity and Exercise on Cognitive and 
Brain Functions in Older Adults. Journal of Aging Research, 657508.
    \84\ Giacino J.T., Whyte J., Bagiella E., et al. (2012). 
Placebo-controlled trial of amantadine for severe traumatic brain 
injury. New England Journal of Medicine, 366(9), 819-826.
    \85\ Alexopoulos G.S., Katz I.R., Reynolds C.F. 3rd, Carpenter 
D., Docherty J.P., Ross R.W. (2001). Pharmacotherapy of depression 
in older patients: A summary of the expert consensus guidelines. 
Journal of Psychiatric Practice, 7(6), 361-376.
    \86\ Arean P.A., Cook B.L. (2002). Psychotherapy and combined 
psychotherapy/pharmacotherapy for late life depression. Biological 
Psychiatry, 52(3), 293-303.
    \87\ Hollon S.D., Jarrett R.B., Nierenberg A.A., Thase M.E., 
Trivedi M., Rush A.J. (2005). Psychotherapy and medication in the 
treatment of adult and geriatric depression: Which monotherapy or 
combined treatment? Journal of Clinical Psychiatry, 66(4), 455-468.
    \88\ Wagenaar D, Colenda CC, Kreft M, Sawade J, Gardiner J, 
Poverejan E. (2003). Treating depression in nursing homes: Practice 
guidelines in the real world. J Am Osteopath Assoc. 103(10), 465-
469.
    \89\ Crespy SD, Van Haitsma K, Kleban M, Hann CJ. Reducing 
Depressive Symptoms in Nursing Home Residents: Evaluation of the 
Pennsylvania Depression Collaborative Quality Improvement Program. J 
Healthc Qual. 2016. Vol. 38, No. 6, pp. e76-e88.
---------------------------------------------------------------------------

    In alignment with our Meaningful Measures Initiative, accurate 
assessment of cognitive function and mental status of patients and 
residents in PAC is expected to make care safer by reducing harm caused 
in the delivery of care; promote effective prevention and treatment of 
chronic disease; strengthen person and family engagement as partners in 
their care; and promote effective communication and coordination of 
care. For example, standardized assessment of cognitive function and 
mental status of patients and residents in PAC will support 
establishing a baseline for identifying changes in cognitive function 
and mental status (for example, delirium), anticipating the patient's 
or resident's ability to understand and participate in treatments 
during a PAC stay, ensuring patient and resident safety (for example, 
risk of falls), and identifying appropriate support needs at the time 
of discharge or transfer. Standardized patient assessment data elements 
will enable or support clinical decision-making and early clinical 
intervention; person-centered, high quality care through facilitating 
better care continuity and coordination; better data exchange and 
interoperability between settings; and longitudinal outcome analysis. 
Therefore, reliable standardized patient assessment data elements 
assessing cognitive function and mental status are needed to initiate a 
management program that can optimize a patient's or resident's 
prognosis and reduce the possibility of adverse events.
    The data elements related to cognitive function and mental status 
were first proposed as standardized patient assessment data elements in 
the FY 2018 IRF PPS proposed rule (82 FR 20723 through 20726). In 
response to our proposals, a few commenters noted that the proposed 
data elements did not capture some dimensions of cognitive function and 
mental status, such as functional cognition, communication, attention, 
concentration, and agitation. One commenter also suggested that other 
cognitive assessments should be considered for standardization. Another 
commenter stated support for the standardized assessment of cognitive 
function and mental status, because it could support appropriate use of 
skilled therapy for beneficiaries with degenerative conditions, such as 
dementia, and appropriate use of medications for behavioral and 
psychological symptoms of dementia.
    We sought comment on our proposals to collect as standardized 
patient assessment data the following data with respect to cognitive 
function and mental status.
    Commenters submitted the following comments related to the proposed 
rule's discussion of the cognitive function and mental status data 
elements.
    Comment: A few commenters were supportive of the proposal to adopt 
the BIMS, CAM, and PHQ-2 to 9 as SPADEs on the topic of cognitive 
function and mental status. One commenter agreed that standardizing 
cognitive assessments will allow providers to identify changes in 
status, support clinical decision-making, and improve care continuity 
and interventions.
    Response: We thank the commenters for their support. We selected 
the

[[Page 39115]]

Cognitive Function and Mental Status data elements for proposal as 
standardized data in part because of the attributes that the commenters 
noted.
    Comment: A few commenters noted limitations of these SPADEs to 
fully assess all areas of cognition and mental status, particularly 
mild to moderate cognitive impairment, and performance deficits that 
may be related to cognitive impairment. Some commenters suggested CMS 
continue exploring assessment tools on the topic of cognition and to 
include a more comprehensive assessment of cognitive function for use 
in PAC settings, noting that highly vulnerable patients with a mild 
cognitive impairment cannot be readily identified through the current 
SPADEs.
    Response: We have strived to balance the scope and level of detail 
of the data elements against the potential burden placed on patients 
and providers. In our past work, we evaluated the potential of several 
different cognition assessments for use as standardized data elements 
in PAC settings. We ultimately decided on the BIMS, CAM, and PHQ-2 to 9 
data elements in our proposal as a starting point. We would welcome 
continued input, recommendations, and feedback from stakeholders about 
additional data elements for standardization, which can be shared with 
CMS through our PAC Quality Initiatives email address: 
[email protected].
    Comment: A commenter stated that cognitive assessment should be 
individualized, rather than standardized, and performed as determined 
by patient needs.
    Response: We believe that the standardized assessment of cognitive 
function is essential to achieving the goals of the IMPACT Act. We also 
wish to clarify that the proposed SPADEs are not intended to replace 
comprehensive clinical evaluation and in no way preclude providers from 
conducting further patient evaluation or assessments in their settings 
as they believe are necessary and useful.
    Comment: Regarding future use of these data elements, one commenter 
recommended that CMS monitor the use of the cognition and mental status 
SPADEs as risk adjustors and make appropriate adjustments to 
methodology as needed.
    Response: We intend to monitor data submitted via the proposed 
SPADEs and will consider these uses in the future. We will also 
continue to review recommendation and feedback from stakeholders 
regarding data elements that would both satisfy the categories listed 
in the IMPACT Act and provide meaningful data.
    Final decisions on the SPADEs are given below, following more 
detailed comments on each SPADE proposal.
 Brief Interview for Mental Status (BIMS)
    In the FY 2020 IRF PPS proposed rule (84 FR 17294 through 17295), 
we proposed that the data elements that comprise the BIMS meet the 
definition of standardized patient assessment data with respect to 
cognitive function and mental status under section 1899B(b)(1)(B)(ii) 
of the Act.
    As described in the FY 2018 IRF PPS Proposed Rule (82 FR 20723 
through 20724), dementia and cognitive impairment are associated with 
long-term functional dependence and, consequently, poor quality of life 
and increased healthcare costs and mortality.\90\ This makes assessment 
of mental status and early detection of cognitive decline or impairment 
critical in the PAC setting. The intensity of routine nursing care is 
higher for patients and residents with cognitive impairment than those 
without, and dementia is a significant variable in predicting 
readmission after discharge to the community from PAC providers.\91\
---------------------------------------------------------------------------

    \90\ Ag[uuml]ero-Torres, H., Fratiglioni, L., Guo, Z., Viitanen, 
M., von Strauss, E., & Winblad, B. (1998). ``Dementia is the major 
cause of functional dependence in the elderly: 3-year follow-up data 
from a population-based study.'' Am J of Public Health 88(10): 1452-
1456.
    \91\ RTI International. Proposed Measure Specifications for 
Measures Proposed in the FY 2017 IRF QRP NPRM. Research Triangle 
Park, NC. 2016.
---------------------------------------------------------------------------

    The BIMS is a performance-based cognitive assessment screening tool 
that assesses repetition, recall with and without prompting, and 
temporal orientation. The data elements that make up the BIMS are seven 
questions on the repetition of three words, temporal orientation, and 
recall that result in a cognitive function score. The BIMS was 
developed to be a brief, objective screening tool, with a focus on 
learning and memory. As a brief screener, the BIMS was not designed to 
diagnose dementia or cognitive impairment, but rather to be a 
relatively quick and easy to score assessment that could identify 
cognitively impaired patients, as well as those who may be at risk for 
cognitive decline and require further assessment. It is currently in 
use in two of the PAC assessments: The MDS used by SNFs and the IRF-PAI 
used by IRFs. For more information on the BIMS, we refer readers to the 
document titled ``Final Specifications for IRF QRP Quality Measures and 
Standardized Patient Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    The data elements that comprise the BIMS were first proposed as 
standardized patient assessment data elements in the FY 2018 IRF PPS 
proposed rule (82 FR 20723 through 20724). In that proposed rule, we 
stated that the proposal was informed by input we received through a 
call for input published on the CMS Measures Management System 
Blueprint website. Input submitted from August 12 to September 12, 
2016, noted support for use of the BIMS, noting that it is reliable, 
feasible to use across settings, and will provide useful information 
about patients and residents. We also stated that the data collected 
through the BIMS will provide a clearer picture of patient or resident 
complexity, help with the care planning process, and be useful during 
care transitions and when coordinating across providers. A summary 
report for the August 12 to September 12, 2016 public comment period 
titled ``SPADE August 2016 Public Comment Summary Report'' is available 
at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In response to our proposal in the FY 2018 IRF PPS proposed rule, 
we received public comments in support of the use of the BIMS, 
especially in its capacity to inform care transitions, but other 
commenters were critical, noting the limitations of the BIMS to assess 
mild cognitive impairment and ``functional'' cognition, and that the 
BIMS cannot be completed by patients and residents who are unable to 
communicate. They also stated that other cognitive assessments 
available in the public domain should be considered for 
standardization. One commenter suggested that CMS require use of the 
BIMS with respect to discharge, as well as admission.
    Subsequent to receiving comments on the FY 2018 IRF PPS rule, the 
BIMS was included in the National Beta Test of candidate data elements 
conducted by our data element contractor from November 2017 to August 
2018. Results of this test found the BIMS to be feasible and reliable 
for use with PAC patients and residents. More information about the 
performance of the BIMS in the National Beta Test can be found in the 
document titled ``Final Specifications for IRF QRP Quality Measures and

[[Page 39116]]

Standardized Patient Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In addition, our data element contractor convened a TEP on 
September 17, 2018, for the purpose of soliciting input on the proposed 
standardized patient assessment data elements and the TEP supported the 
assessment of patient or resident cognitive status with respect to both 
admission and discharge. A summary of the September 17, 2018 TEP 
meeting titled ``SPADE Technical Expert Panel Summary (Third 
Convening)'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We also held Special Open Door Forums (SODFs) and small-group 
discussions with PAC providers and other stakeholders in 2018 for the 
purpose of updating the public about our ongoing SPADE development 
efforts. Finally, on November 27, 2018, our data element contractor 
hosted a public meeting of stakeholders to present the results of the 
National Beta Test and solicit additional comments. General input on 
the testing and item development process and concerns about burden were 
received from stakeholders during this meeting and via email through 
February 1, 2019. Some commenters also expressed concern that the BIMS, 
if used alone, may not be sensitive enough to capture the range of 
cognitive impairments, including mild cognitive impairment. A summary 
of the public input received from the November 27, 2018 stakeholder 
meeting titled ``Input on Standardized Patient Assessment Data Elements 
(SPADEs) Received After November 27, 2018 Stakeholder Meeting'' is 
available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We understand the concerns raised by stakeholders that BIMS, if 
used alone, may not be sensitive enough to capture the range of 
cognitive impairments, including functional cognition and MCI, but note 
that the purpose of the BIMS data elements as SPADEs is to screen for 
cognitive impairment in a broad population. We also acknowledge that 
further cognitive tests may be required based on a patient's condition 
and will take this feedback into consideration in the development of 
future standardized patient assessment data elements. However, taking 
together the importance of assessing for cognitive status, stakeholder 
input, and strong test results, we proposed that the BIMS data elements 
meet the definition of standardized patient assessment data with 
respect to cognitive function and mental status under section 
1899B(b)(1)(B)(ii) of the Act and to adopt the BIMS data elements as 
standardized patient assessment data for use in the IRF QRP.
    Commenters submitted the following comments related to the proposed 
rule's discussion of the BIMS data elements.
    Comment: One commenter supported the collection of BIMS at both 
admission and discharge and believes it will result in more complete 
data and better care.
    Response: We thank the commenter for the support of the BIMS data 
element.
    Comment: One commenter stated that the BIMS fails to detect mild 
cognitive impairment, differentiate cognitive impairment from a 
language impairment, link impairment to functional limitation, or 
identify issues with problem solving and executive function. This 
commenter recommended use of the Development of Outpatient Therapy 
Payment Alternatives (DOTPA) items for PAC, as well as a screener 
targeting functional cognition. Another commenter also recommended CMS 
identify a better cognitive assessment and not to move forward with the 
proposal.
    Response: We recognize that the BIMS assesses components of 
cognition and does not, alone, provide a comprehensive assessment of 
potential cognitive impairment. We clarify that any SPADE is intended 
as a minimum assessment and does not limit the ability of providers to 
conduct a more comprehensive assessment of cognition to identify the 
complexities or potential impacts of cognitive impairment that the 
commenter describes.
    We evaluated the suitability of the DOTPA, as well as other 
screening tools that targeted functional cognition, by engaging our 
TEP, through ``alpha'' feasibility testing, and through soliciting 
input from stakeholders. At the second meeting of TEP in March 2017, 
members questioned the use of data elements that rely on assessor 
observation and judgment, such as DOTPA CARE tool items, and favored 
other assessments of cognition that required patient interview or 
patient actions. The TEP also discussed performance-based assessment of 
functional cognition. These are assessments that require patients to 
respond by completing a simulated task, such as ordering from a menu, 
or reading medication instructions and simulating the taking of 
medications, as required by the Performance Assessment of Self-Care 
Skills (PASS) items.
    In Alpha 2 feasibility testing, which was conducted between April 
and July 2017, we included a subset of items from the DOTPA as well as 
the PASS. Findings of that test identified several limitations of the 
DOTPA items for use as SPADEs, such as relatively long to administer (5 
to 7 minutes), especially in the LTCH setting. Assessors also indicated 
that these items had low relevance for SNF and LTCH patients. In 
addition, interrater reliability was highly variable among the DOTPA 
items, both overall and across settings, with some items showing very 
low agreement (as low as 0.34) and others showing excellent agreement 
(as high as 0.81). Similarly, findings of the Alpha 2 feasibility test 
identified several limitations of the PASS for use as SPADEs. The PASS 
was relatively time-intensive to administer (also 5 to 7 minutes), many 
patients in HHAs and IRFs needed assistance completing the PASS tasks, 
and missing data were prevalent. Unlike the DOTPA items, interrater 
reliability was consistently high overall for PASS (ranging from 0.78 
to 0.92), but the high reliability was not deemed to outweigh 
fundamental feasibility concerns related to administration challenges. 
A summary report for the Alpha 2 feasibility testing titled 
``Development and Maintenance of Standardized Cross Setting Patient 
Assessment Data for Post-Acute Care: Summary Report of Findings from 
Alpha 2 Pilot Testing'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/Downloads/Alpha-2-SPADE-Pilot-Summary-Document.pdf.
    Feedback was obtained on the DOTPA and other assessments of 
functional cognition through a call for input that was open from April 
26, 2017 to June 26, 2017. While we received support for the DOTPA, 
PASS, and other assessments of functional cognition, commenters also 
raised concerns about the reliability of the DOTPA, given that it is 
based on staff evaluation, and the feasibility of the PASS, given that 
the simulated medication task requires props, such as a medication 
bottle with printed label and pill box, which may not be accessible in 
all settings. A summary report for the April 26 to June 26, 2017 public 
comment period titled

[[Page 39117]]

``Public Comment Summary Report 2'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/Downloads/Public-Comment-Summary-Report_Standardized-Patient-Assessment-Data-Element-Work_PC2_Jan-2018.pdf.
    Based on the input from our TEP, results of alpha feasibility 
testing, and input from stakeholders, we decided to propose the BIMS 
for standardization at this time due to the body of research literature 
supporting its feasibility and validity, its relative brevity, and its 
existing use in the MDS and IRF-PAI.
    Comment: A few commenters noted that BIMS is currently collected by 
IRFs and has not been demonstrated to predict costs or differentiate 
case-mix and believes that CMS has not provided any evidence that the 
BIMS is capable of being utilized for quality purposes to support the 
collection of these data elements at discharge. Another commenter 
stated that CMS has not provided quantitative evidence that the BIMS 
data elements are capable of measuring provider performance for quality 
or of differentiating case-mix for payment.
    Response: We reiterate that the purpose of standardizing data 
elements, in accordance with the IMPACT Act, is to support care 
planning, clinical decision support, inform case-mix and quality 
measurement, support care transitions, and enable interoperable data 
exchange and data sharing between PAC settings. Before being identified 
as a SPADE, the BIMS underwent an extensive consensus vetting process 
in which experts and stakeholders were engaged through TEPs, SODFs, and 
posting of interim reports and other documents on the CMS.gov website. 
A summary of the most recent TEP meeting (September 17, 2018) titled 
``SPADE Technical Expert Panel Summary (Third Convening)'' is available 
at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html. Results of these activities 
provide evidence that experts and providers believe that the BIMS data 
elements have the potential for measuring quality, describing case mix, 
and improving care.
    Comment: A commenter believes that assessing BIMS at discharge 
would not be clinically useful and would not contribute to improved 
patient care or outcomes. The commenter noted that assessing BIMS at 
discharge was not evaluated during the National Beta Test, and objected 
to the BIMS being proposed for use at discharge.
    Response: We maintain that a standardized cognitive assessment 
using the BIMS is clinically useful and has the potential to improve 
patient care and outcomes. The commenter stated that the BIMS was not 
administered at discharge in the National Beta Test. However, the BIMS 
was in fact assessed at both admission and discharge in the National 
Beta Test. Moreover, to support data exchange between settings, and to 
support quality measurement, the IMPACT Act requires that the SPADEs be 
collected with respect to both admission and discharge. After careful 
consideration of the public comments we received, we are finalizing our 
proposal to adopt the BIMS as standardized patient assessment data 
beginning with the FY 2022 IRF QRP as proposed.
 Confusion Assessment Method (CAM)
    In the FY 2020 IRF PPS proposed rule (84 FR 17295), we proposed 
that the data elements that comprise the Confusion Assessment Method 
(CAM) meet the definition of standardized patient assessment data with 
respect to cognitive function and mental status under section 
1899B(b)(1)(B)(ii) of the Act.
    As described in the FY 2018 IRF PPS proposed rule (82 FR 20724), 
the CAM was developed to identify the signs and symptoms of delirium. 
It results in a score that suggests whether a patient or resident 
should be assigned a diagnosis of delirium. Because patients and 
residents with multiple comorbidities receive services from PAC 
providers, it is important to assess delirium, which is associated with 
a high mortality rate and prolonged duration of stay in hospitalized 
older adults.\92\ Assessing these signs and symptoms of delirium is 
clinically relevant for care planning by PAC providers.
---------------------------------------------------------------------------

    \92\ Fick, D.M., Steis, M.R., Waller, J.L., & Inouye, S.K. 
(2013). ``Delirium superimposed on dementia is associated with 
prolonged length of stay and poor outcomes in hospitalized older 
adults.'' J of Hospital Med 8(9): 500-505.
---------------------------------------------------------------------------

    The CAM is a patient assessment that screens for overall cognitive 
impairment, as well as distinguishes delirium or reversible confusion 
from other types of cognitive impairment. The CAM is currently in use 
in two of the PAC assessments: A four-item version of the CAM is used 
in the MDS in SNFs; and a six-item version of the CAM is used in the 
LTCH CARE Data Set (LCDS) in LTCHs. We proposed the four-item version 
of the CAM that assesses acute change in mental status, inattention, 
disorganized thinking, and altered level of consciousness. For more 
information on the CAM, we refer readers to the document titled ``Final 
Specifications for IRF QRP Quality Measures and Standardized Patient 
Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    The data elements that comprise the CAM were first proposed as 
standardized patient assessment data elements in the FY 2018 IRF PPS 
proposed rule (82 FR 20724). In that proposed rule, we stated that the 
proposal was informed by public input we received on the CAM through a 
call for input published on the CMS Measures Management System 
Blueprint website. Input submitted from August 12 to September 12, 2016 
noted support for use of the CAM, noting that it would provide 
important information for care planning and care coordination, and 
therefore, contribute to quality improvement. We also stated that those 
commenters had noted the CAM is particularly helpful in distinguishing 
delirium and reversible confusion from other types of cognitive 
impairment. A summary report for the August 12 to September 12, 2016 
public comment period titled ``SPADE August 2016 Public Comment Summary 
Report'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In response to our proposal in the FY 2018 IRF PPS proposed rule, 
one commenter supported use of the CAM for standardized patient 
assessment data. However, some commenters expressed concerns that the 
CAM data elements assess: The presence of behavioral symptoms, but not 
the cause; the possibility of a false positive for delirium due to 
patient cognitive or communication impairments; and the lack of 
specificity of the assessment specifications. In addition, other 
commenters noted that the CAM is not necessary because: Delirium is 
easily diagnosed without a tool; the CAM and BIMS assessments are 
redundant; and some CAM response options are not meaningful.
    Subsequent to receiving comments on the FY 2018 IRF PPS rule, the 
CAM was included in the National Beta Test of candidate data elements 
conducted by our data element contractor from November 2017 to August 
2018. Results of this test found the CAM to be feasible

[[Page 39118]]

and reliable for use with PAC patients and residents. More information 
about the performance of the CAM in the National Beta Test can be found 
in the document titled ``Final Specifications for IRF QRP Quality 
Measures and Standardized Patient Assessment Data Elements,'' available 
at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In addition, our data element contractor convened a TEP on 
September 17, 2018, for the purpose of soliciting input on the proposed 
standardized patient assessment data elements. Although they did not 
specifically discuss the CAM data elements, the TEP supported the 
assessment of patient or resident cognitive status with respect to both 
admission and discharge. A summary of the September 17, 2018 TEP 
meeting titled ``SPADE Technical Expert Panel Summary (Third 
Convening)'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We also held SODFs and small-group discussions with PAC providers 
and other stakeholders in 2018 for the purpose of updating the public 
about our ongoing SPADE development efforts. Finally, on November 27, 
2018, our data element contractor hosted a public meeting of 
stakeholders to present the results of the National Beta Test and 
solicit additional comments. General input on the testing and item 
development process and concerns about burden were received from 
stakeholders during this meeting and via email through February 1, 
2019. A summary of the public input received from the November 27, 2018 
stakeholder meeting titled ``Input on Standardized Patient Assessment 
Data Elements (SPADEs) Received After November 27, 2018 Stakeholder 
Meeting'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Taking together the importance of assessing for delirium, 
stakeholder input, and strong test results, we proposed that the CAM 
data elements meet the definition of standardized patient assessment 
data with respect to cognitive function and mental status under section 
1899B(b)(1)(B)(ii) of the Act and to adopt the CAM data elements as 
standardized patient assessment data for use in the IRF QRP.
    Commenters submitted the following comments related to the proposed 
rule's discussion of the proposed CAM data elements.
    Comment: A few commenters stated that the CAM would be redundant 
with other cognitive assessments, such as BIMS. One commenter stated 
that delirium would be assessed prior to discharge from the acute care 
setting, making the assessment of delirium at admission to the IRF 
redundant. Another commenter stated that concerns about burden 
outweighed the value that the CAM might have for some populations, and 
noted that daily physician visits and daily assessments of patients by 
the interdisciplinary team were sufficient to assess cognitive needs.
    Response: The CAM specifically screens for change in mental status, 
inattention, disorganized thinking and altered level of consciousness, 
which can indicate symptoms of delirium. These symptoms are not 
assessed by other cognitive assessments in the IRF-PAI. We believe the 
assessment of delirium at admission and discharge is important to 
informing patient care. Delirium occurs in up to half of patients/
residents receiving PAC services,\93\ and signs and symptoms of 
delirium are associated with poor functional recovery,\94\ re-
hospitalization, and mortality.\95\ Because the majority of delirium 
episodes are transient,\96\ we would not expect assessment of delirium 
prior to discharge from the acute care setting to capture all cases of 
delirium in PAC, as there may be an acute change in mental status from 
the patient's baseline or fluctuations in the patient's behaviors that 
are identified after PAC admission.
---------------------------------------------------------------------------

    \93\ Dan K. Kiely et al., ``Characteristics Associated with 
Delirium Persistence Among Newly Admitted Post-Acute Facility 
Patients,'' Journals of Gerontology: Series A (Biological Sciences 
and Medical Sciences), Vol. 59, No. 4, April 2004; Edward R. 
Marcantonio et al., ``Delirium Symptoms in Post-Acute Care: 
Prevalent, Persistent, and Associated with Poor Functional 
Recovery,'' Journal of the American Geriatrics Society, Vol. 51, No. 
1, January 2003.
    \94\ Marcantonio, Edward R., Samuel E. Simon, Margaret A. 
Bergmann, Richard N. Jones, Katharine M. Murphy, and John N. Morris, 
``Delirium Symptoms in Post-Acute Care: Prevalent, Persistent, and 
Associated with Poor Functional Recovery,'' Journal of the American 
Geriatrics Society, Vol. 51, No. 1, January 2003, pp. 4-9.
    \95\ Edward R. Marcantonio et al., Outcomes of Older People 
Admitted to Postacute Facilities with Delirium,'' Journal of the 
American Geratrics Society, Vol. 53, No. 6, June 2005.
    \96\ Cole MG, Ciampi A, Belzile E, Zhong L. Persistent delirium 
in older hospital patients: A systematic review of frequency and 
prognosis. Age Ageing 2009;38:19-26.
---------------------------------------------------------------------------

    Comment: Several commenters noted doubts about the usefulness of 
the CAM. One commenter was unsure if CAM will identify differences in 
cognitive status or measure changes during the stay resulting from 
therapeutic interventions. A few commenters stated that the CAM would 
not provide information that would be useful clinically, that it was 
not specific enough or too narrowly focused, and that it should not be 
required at discharge. Another commenter suggested that CMS not include 
the CAM as SPADE because they believe delirium is clinically apparent, 
and therefore, doubt that a standardized assessment of delirium will 
contribute to improving patient care or outcomes. Another commenter 
expressed concern that the CAM data elements would not identify 
cognitive needs that would impact quality in therapeutic intervention 
across facilities.
    Response: As with any brief screening tool, we believe that the CAM 
has value as a universal assessment to identify patients in need of 
further clinical evaluation. Delirium occurs in up to 50 percent of 
patients/residents in PAC \97\ and is associated with poor 
outcomes.98 99 Hyperactive delirium--the type of delirium 
that manifests with agitation--makes up only a quarter of delirium 
cases.100 101 Delirium more commonly manifests as 
hypoactive, or ``quiet'' delirium,\102\ suggesting that brief, 
universal screening is appropriate. Moreover, because there are 
treatments for delirium that can be developed based on medication 
review, physical examination, laboratory tests, and evaluation of 
environmental factors,\103\

[[Page 39119]]

we believe that screening for delirium would support care planning and 
care transitions for these patients.
---------------------------------------------------------------------------

    \97\ Kiely DK, Jones RN, Bergmann MA, Marcantonio ER. 
Association between psychomotor activity delirium subtypes and 
mortality among newly admitted post-acute facility patients. J 
Gerontol A Biol Sci Med Sci 2007;62:174-179.
    \98\ Marcantonio, Edward R., Samuel E. Simon, Margaret A. 
Bergmann, Richard N. Jones, Katharine M. Murphy, and John N. Morris, 
``Delirium Symptoms in Post-Acute Care: Prevalent, Persistent, and 
Associated with Poor Functional Recovery,'' Journal of the American 
Geriatrics Society, Vol. 51, No. 1, January 2003, pp. 4-9.
    \99\ Edward R. Marcantonio et al., Outcomes of Older People 
Admitted to Postacute Facilities with Delirium,'' Journal of the 
American Geratrics Society, Vol. 53, No. 6, June 2005.
    \100\ Inouye SK, Westendorp RG, Saczynski JS. Delirium in 
elderly people. Lancet 2014;383:911-922.
    \101\ Marcantonio ER. In the clinic: Delirium. Ann Intern Med 
2011;154:ITC6-1-ITC6-1.
    \102\ Yang FM, Marcantonio ER, Inouye SK, et al. 
Phenomenological subtypes of delirium in older persons: Patterns, 
prevalence, and prognosis. Psychosomatics 2009;50:248-254.
    \103\ Marcantonio ER. Delirium in Hospitalized Older Adults. N 
Engl J Med. 2017 Oct 12;377(15):1456-1466.
---------------------------------------------------------------------------

    Comment: A few commenters believe the CAM would be difficult to 
administer and raised concerns about the training that staff would 
receive in order to ensure that administration is consistent and valid.
    Response: We appreciate the commenters' recommendation to provide 
clear training for administering the CAM, and will take it into 
consideration as we revise the current training for the IRF-PAI. We 
intend to reinforce assessment tips and item rationale through 
training, open door forums, and future rulemaking efforts.
    Comment: One commenter disagreed that delirium assesses a dimension 
of cognitive function.
    Response: The CAM data elements were proposed to meet the 
definition of the standardized patient assessment data with respect to 
cognitive function and mental status. Section 1899B(b)(1)(B)(ii) of the 
Act specifies that PAC providers shall be required to submit 
standardized patient assessment data for the category of cognitive 
function, such as the ability to express ideas and to understand, and 
mental status, such as depression and dementia. A recent deterioration 
in cognitive function or present and fluctuating behaviors of 
inattention, disorganized thinking, or altered level of consciousness 
may indicate delirium.\104\ Delirium can also be misdiagnosed as 
dementia.\105\
---------------------------------------------------------------------------

    \104\ Inouye SK, van Dyck CH, Alessi CA, Balkin S, Siegal AP, 
Horwitz RI. Clarifying confusion: The confusion assessment method. A 
new method for detection of delirium. Ann Intern Med. 1990 Dec 
15;113(12):941-8.
    \105\ Marcantonio ER. Delirium in Hospitalized Older Adults. N 
Engl J Med. 2017 Oct 12;377(15):1456-1466.
---------------------------------------------------------------------------

    Comment: A commenter stated that CMS has not provided quantitative 
evidence that the CAM data elements are capable of measuring provider 
performance for quality or of differentiating case-mix for payment.
    Response: The clinical SPADEs proposed in this rule, including CAM, 
were the result of an extensive consensus vetting process. Over the 
past several years, we have engaged experts and a wide range of 
stakeholders through TEPs, Special Open Door Forums, and documents made 
available on the CMS.gov website. A summary of the most recent TEP 
meeting (September 17, 2018) titled ``SPADE Technical Expert Panel 
Summary (Third Convening)'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html. Results of these activities provide evidence that experts 
and providers believe that the proposed SPADEs, including the CAM data 
elements, have the potential for measuring quality, describing case 
mix, and improving care.
    After careful consideration of the public comments we received, we 
are finalizing our proposal to adopt the CAM as standardized patient 
assessment data beginning with the FY 2022 IRF QRP as proposed.
 Patient Health Questionnaire-2 to 9 (PHQ-2 to 9)
    In the FY 2020 IRF PPS proposed rule (84 FR 17296 through 17297), 
we proposed that the Patient Health Questionnaire-2 to 9 (PHQ-2 to 9) 
data elements meet the definition of standardized patient assessment 
data with respect to cognitive function and mental status under section 
1899B(b)(1)(B)(ii) of the Act. The proposed data elements are based on 
the PHQ-2 mood interview, which focuses on only the two cardinal 
symptoms of depression, and the longer PHQ-9 mood interview, which 
assesses presence and frequency of nine signs and symptoms of 
depression. The name of the data element, the PHQ-2 to 9, refers to an 
embedded skip pattern that transitions patients with a threshold level 
of symptoms in the PHQ-2 to the longer assessment of the PHQ-9. The 
skip pattern is described further below. As described in the FY 2018 
IRF PPS proposed rule (82 FR 20725 through 20726), depression is a 
common and under-recognized mental health condition. Assessments of 
depression help PAC providers better understand the needs of their 
patients and residents by: Prompting further evaluation after 
establishing a diagnosis of depression; elucidating the patient's or 
resident's ability to participate in therapies for conditions other 
than depression during their stay; and identifying appropriate ongoing 
treatment and support needs at the time of discharge.
    The proposed PHQ-2 to 9 is based on the PHQ-9 mood interview. The 
PHQ-2 consists of questions about only the first two symptoms addressed 
in the PHQ-9: Depressed mood and anhedonia (inability to pleasure), 
which are the cardinal symptoms of depression. The PHQ-2 has performed 
well as both a screening tool for identifying depression, to assess 
depression severity, and to monitor patient mood over 
time.106 107 If a patient demonstrates signs of depressed 
mood and anhedonia under the PHQ-2, then the patient is administered 
the lengthier PHQ-9. This skip pattern (also referred to as a gateway) 
is designed to reduce the length of the interview assessment for 
patients who fail to report the cardinal symptoms of depression. The 
design of the PHQ-2 to 9 reduces the burden that would be associated 
with requiring the full PHQ-9, while ensuring that patients and 
residents with indications of depressive symptoms based on the PHQ-2 
receive the longer assessment.
---------------------------------------------------------------------------

    \106\ Li, C., Friedman, B., Conwell, Y., & Fiscella, K. (2007). 
``Validity of the Patient Health Questionnaire 2 (PHQ-2) in 
identifying major depression in older people.'' J of the A 
Geriatrics Society, 55(4): 596-602.
    \107\ L[ouml]we, B., Kroenke, K., & Gr[auml]fe, K. (2005). 
``Detecting and monitoring depression with a two-item questionnaire 
(PHQ-2).'' J of Psychosomatic Research, 58(2): 163-171.
---------------------------------------------------------------------------

    Components of the proposed data elements are currently used in the 
OASIS for HHAs (PHQ-2) and the MDS for SNFs (PHQ-9). For more 
information on the PHQ-2 to 9, we refer readers to the document titled 
``Final Specifications for IRF QRP Quality Measures and Standardized 
Patient Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We proposed the PHQ-2 data elements as SPADEs in the FY 2018 IRF 
proposed rule (82 FR 20725 through 20726). In that proposed rule, we 
stated that the proposal was informed by input we received from the TEP 
convened by our data element contractor on April 6 and 7, 2016. The TEP 
members particularly noted that the brevity of the PHQ-2 made it 
feasible to administer with low burden for both assessors and PAC 
patients or residents. A summary of the April 6 and 7, 2016 TEP meeting 
titled ``SPADE Technical Expert Panel Summary (First Convening)'' is 
available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    The rule proposal was also informed by public input that we 
received through a call for input published on the CMS Measures 
Management System Blueprint website. Input was submitted from August 12 
to September 12, 2016 on three versions of the PHQ depression screener: 
The PHQ-2; the PHQ-9; and

[[Page 39120]]

the PHQ-2 to 9 with the skip pattern design. Many commenters were 
supportive of the standardized assessment of mood in PAC settings, 
given the role that depression plays in well-being. Several commenters 
noted support for an approach that would use PHQ-2 as a gateway to the 
longer PHQ-9 while still potentially reducing burden on most patients 
and residents, as well as test administrators, and ensuring the 
administration of the PHQ-9, which exhibits higher specificity,\108\ 
for patients and residents who showed signs and symptoms of depression 
on the PHQ-2. A summary report for the August 12 to September 12, 2016 
public comment period titled ``SPADE August 2016 Public Comment Summary 
Report'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
---------------------------------------------------------------------------

    \108\ Arroll B, Goodyear-Smith F, Crengle S, Gunn J, Kerse N, 
Fishman T, et al. Validation of PHQ-2 and PHQ-9 to screen for major 
depression in the primary care population. Annals of family 
medicine. 2010;8(4):348-53. doi: 10.1370/afm.1139 pmid:20644190; 
PubMed Central PMCID: PMC2906530.
---------------------------------------------------------------------------

    In response to our proposal to use the PHQ-2 in the FY 2018 IRF PPS 
proposed rule (82 FR 20725 through 20726), we received comments 
agreeing to the importance of a standardized assessment of depression 
in patients and residents receiving PAC services. Commenters also 
raised concerns about the ability of the PHQ-2 to correctly identify 
all patients and residents with signs and symptoms of depression. One 
commenter supported using the PHQ-2 as a gateway assessment and 
conducting a more thorough evaluation of depression symptoms with the 
PHQ-9 if the PHQ-2 is positive. Another commenter expressed concern 
that standardized assessment of signs and symptoms of depression via 
the PHQ-2 is not appropriate in the IRF setting, as patients may have 
recently experienced acute illness or injury, and routine screening may 
lead to overprescribing of antidepressant medications. Another 
commenter expressed concern about potential conflicts between the 
results of screening assessments and documented diagnoses based on the 
expertise of physicians and other clinicians. In response to these 
comments, we carried out additional testing, and we provide our 
findings below.
    Subsequent to receiving comments on the FY 2018 IRF PPS rule, the 
PHQ-2 to 9 was included in the National Beta Test of candidate data 
elements conducted by our data element contractor from November 2017 to 
August 2018. Results of this test found the PHQ-2 to 9 to be feasible 
and reliable for use with PAC patients and residents. More information 
about the performance of the PHQ-2 to 9 in the National Beta Test can 
be found in the document titled ``Final Specifications for IRF QRP 
Quality Measures and Standardized Patient Assessment Data Elements,'' 
available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In addition, our data element contractor convened a TEP on 
September 17, 2018, for the purpose of soliciting input on the PHQ-2 to 
9. The TEP was supportive of the PHQ-2 to 9 data element set as a 
screener for signs and symptoms of depression. The TEP's discussion 
noted that symptoms evaluated by the full PHQ-9 (for example, 
concentration, sleep, appetite) had relevance to care planning and the 
overall well-being of the patient or resident, but that the gateway 
approach of the PHQ-2 to 9 would be appropriate as a depression 
screening assessment, as it depends on the well-validated PHQ-2 and 
focuses on the cardinal symptoms of depression. A summary of the 
September 17, 2018 TEP meeting titled ``SPADE Technical Expert Panel 
Summary (Third Convening)'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We also held SODFs and small-group discussions with PAC providers 
and other stakeholders in 2018 for the purpose of updating the public 
about our on-going SPADE development efforts. Finally, on November 27, 
2018, our data element contractor hosted a public meeting of 
stakeholders to present the results of the National Beta Test and 
solicit additional comments. General input on the testing and item 
development process and concerns about burden were received from 
stakeholders during this meeting and via email through February 1, 
2019. A summary of the public input received from the November 27, 2018 
stakeholder meeting titled ``Input on Standardized Patient Assessment 
Data Elements (SPADEs) Received After November 27, 2018 Stakeholder 
Meeting'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Taking together the importance of assessing for depression, 
stakeholder input, and test results, we proposed that the PHQ-2 to 9 
data elements meet the definition of standardized patient assessment 
data with respect to cognitive function and mental status under section 
1899B(b)(1)(B)(ii) of the Act and to adopt the PHQ-2 to 9 data elements 
as standardized patient assessment data for use in the IRF QRP.
    Commenters submitted the following comments related to the proposed 
rule's discussion of the PHQ-2 to 9 data elements.
    Comment: Some commenters supported the inclusion of the PHQ-2 to 9. 
One of these commenters was particularly supportive of the use of the 
2-item gateway in the PHQ-2 to 9 approach to improve efficiency.
    Response: We thank the commenters for their support of the PHQ-2 to 
9, including the gateway approach as a way to decrease burden for 
providers and patients.
    Comment: One commenter was unsure if PHQ-2 to 9 will identify 
differences in cognitive status or measure changes during the stay 
resulting from therapeutic interventions. Another commenter expressed 
concern that the PHQ-2 to 9 data elements would not identify cognitive 
needs that would impact quality in therapeutic intervention across 
facilities.
    Response: As with any brief screening tool, we believe that the 
PHQ-2 to 9 has value as a universal assessment to identify patients in 
need of further clinical evaluation. We believe that applying a brief, 
standardized assessment of depression across PAC settings, including 
IRFs, will improve detection based on the PHQ-2 to 9 interview. A 
universal depression screening is expected to improve patient outcomes 
by increasing the likelihood that depression will be identified and 
treated in IRF patients. The proposal of the PHQ-2 to 9 was the result 
of an extensive consensus vetting process in which experts and 
stakeholders were engaged through TEPs, SODFs, and posting of interim 
reports and other documents on CMS.gov. These experts and stakeholders 
were supportive of the clinical usefulness of the PHQ-2 to 9 
assessment. A summary of the most recent TEP meeting (September 17, 
2018) titled ``SPADE Technical Expert Panel Summary (Third Convening)'' 
is available at https://www.cms.gov/Medicare/Quality-Initiatives-
Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/
IMPACT-Act-of-

[[Page 39121]]

2014/IMPACT-Act-Downloads-and-Videos.html.
    Comment: A few commenters raised concerns about administration of 
the PHQ-2 to 9 to IRF patients. One commenter noted that patients in 
acute rehabilitation may have limited attention and working memory that 
affects their ability to complete the PHQ-2 to 9. Another commenter 
noted doubts that PHQ-9 is a good tool for IRFs because of the 
likelihood of false positives, given patients who are adjusting to 
recent injuries, surgeries, conditions, and various disabilities. 
Rather, the commenter believes that assessment by rehabilitation 
psychologists, who have specialty training in working with 
rehabilitation populations, would provide a comprehensive evaluation 
and informed treatment plan. Another commenter expressed concerns about 
the use of the PHQ in short-stay IRF patients, suggesting that being 
assessed for depression, especially if assessed multiple times, will 
affect the patient's perception of how they should be experiencing 
their situation.
    Response: We recognize the challenges faced by patients receiving 
care from IRF providers. We believe that the PHQ-2 to 9 is the most 
accurate and appropriate depression screening for the PAC population, 
including patients in IRFs, and that assessing for depression is 
necessary for high-quality clinical care. As stated in our proposal 
above, the PHQ-2 has performed well as a screening tool for identifying 
depression, to assess depression severity, and to monitor patient mood 
over time.\109\ \110\ Additionally, the PHQ-2 and PHQ-9 instruments 
have been validated in primary care populations against a gold standard 
diagnostic interview.\111\ We believe this prior validation research 
generalizes to the IRF population. We also note that, regardless of the 
LOS of patients, the timeframe over which they may have been 
experiencing signs and symptoms of depression, and the types of 
circumstances that have led to their IRF stay, it is the responsibility 
of the IRF to deliver high quality care for all the symptoms or 
conditions a patient may have. The expectation that the episode of care 
will be short does not exempt an IRF from screening and treating 
patients for the full range of physical and mental health problems. 
Similarly, if a patient self-reports a significant number of depressive 
symptoms, we do not believe that they should be considered to be a 
``false positive'' because of, for example, a recent trauma or acute 
care stay. As a screening tool, the PHQ-2 to 9 is intended to capture 
likely depression to have those patients referred for further 
evaluation, which will ascertain if their condition is consistent with 
the full diagnostic criteria for a major depressive disorder. Moreover, 
standardized screening for the signs and symptoms of depression with 
the PHQ-2 to 9 does not preclude or provide a substitute for assessment 
by rehabilitation psychologist or other clinicians, as deemed 
appropriate by a patient's care team.
---------------------------------------------------------------------------

    \109\ Li, C., Friedman, B., Conwell, Y., & Fiscella, K. (2007). 
``Validity of the Patient Health Questionnaire 2 (PHQ-2) in 
identifying major depression in older people.'' J of the A 
Geriatrics Society, 55(4): 596-602.
    \110\ L[ouml]we, B., Kroenke, K., & Gr[auml]fe, K. (2005). 
``Detecting and monitoring depression with a two-item questionnaire 
(PHQ-2).'' J of Psychosomatic Research, 58(2): 163-171.
    \111\ Arroll B, Goodyear-Smith F, Crengle S, Gunn J, Kerse N, 
Fishman T, et al. Validation of PHQ-2 and PHQ-9 to screen for major 
depression in the primary care population. Annals of family 
medicine. 2010;8(4):348-353.
---------------------------------------------------------------------------

    Comment: Several commenters cited concerns related to the findings 
from the National Beta Test related to the PHQ-2 to 9, namely, that 
testing found it to be burdensome for staff and patients and the 
wording difficult to understand.
    Response: We acknowledge that some assessors in the National Beta 
Test noted concerns regarding the burden of the PHQ-2 to 9 for staff 
and patients and that the wording of some items was challenging for 
patients to understand. In the National Beta Test, the PHQ-2 to 9 was 
one of a collection of mood assessments, meaning that assessors and 
patients completed additional questions about depressed mood and well-
being immediately before and after the PHQ-2 to 9. We believe that the 
perception of burden of the PHQ-2 to 9 was in part due to the larger 
mood assessment section included in the National Beta Test. Despite the 
burden and administration challenges noted by National Beta Test 
assessors, assessors generally appreciated the clinical utility and 
relevance of the PHQ-2 to 9 and noted the importance of standardizing 
the assessment of depressive symptoms.
    Comment: Additional concerns about administration focused on the 
patient interview format of the PHQ-2 to 9. Some commenters raised 
concerns about administering the PHQ-2 to 9 to patients with severe 
cognitive deficits, prior mental health issues, or non-communicative 
conditions. One commenter suggested that CMS develop exemptions from 
repeated screenings for short stay patients, and for patients whose 
medical or cognitive status make it inappropriate to administer the 
PHQ-2 to 9. Another commenter suggested that the PHQ-2 to 9 have an 
option to be self-administered by the patient via a patient-friendly 
paper and pencil layout, which would reduce time burden placed on 
assessors.
    Response: We appreciate commenters' concerns that administering the 
PHQ-2 to 9 to patients whose medical or cognitive status make it 
inappropriate to administer. The guidance for completing the data 
elements will include instructions that if the patient is rarely or 
never understood verbally, in writing, or using another method, the 
PHQ-2 to 9 interview will not be completed and the assessor code the 
responses to the first two items (Little interest or pleasure in doing 
things; Feeling down, depressed, or hopeless) as 9 (no response). We 
will take the suggestion to explore the possibility for patient self-
administration of the PHQ-2 to 9 into consideration in future SPADE 
development work.
    Comment: One commenter noted confusion about how depression relates 
to cognitive function.
    Response: Section 1899(b)(1)(B)(ii) of the Act specifies the 
category of ``cognitive function, such as ability to express ideas and 
to understand, and mental status, such as depression and dementia.'' We 
proposed the PHQ-2 to 9 data elements to meet the definition of the 
standardized patient assessment data with respect to cognitive function 
and mental status, particularly the ``mental status'' topic within that 
category.
    After careful consideration of the public comments we received, we 
are finalizing our proposal to adopt the PHQ-2 to 9 data elements as 
standardized patient assessment data beginning with the FY 2022 IRF QRP 
as proposed.
2. Special Services, Treatments, and Interventions Data
    Special services, treatments, and interventions performed in PAC 
can have a major effect on an individual's health status, self-image, 
and quality of life. The assessment of these special services, 
treatments, and interventions in PAC is important to ensure the 
continuing appropriateness of care for the patients and residents 
receiving them, and to support care transitions from one PAC provider 
to another, an acute care hospital, or discharge. In alignment with our 
Meaningful Measures Initiative, accurate assessment of special 
services, treatments, and interventions of patients and residents 
served by PAC providers is expected to make care safer by reducing harm 
caused in the delivery of care; promote effective prevention and 
treatment of chronic disease; strengthen person and

[[Page 39122]]

family engagement as partners in their care; and promote effective 
communication and coordination of care.
    For example, standardized assessment of special services, 
treatments, and interventions used in PAC can promote patient and 
resident safety through appropriate care planning (for example, 
mitigating risks such as infection or pulmonary embolism associated 
with central intravenous access), and identifying life-sustaining 
treatments that must be continued, such as mechanical ventilation, 
dialysis, suctioning, and chemotherapy, at the time of discharge or 
transfer. Standardized assessment of these data elements will enable or 
support: Clinical decision-making and early clinical intervention; 
person-centered, high quality care through, for example, facilitating 
better care continuity and coordination; better data exchange and 
interoperability between settings; and longitudinal outcome analysis. 
Therefore, reliable data elements assessing special services, 
treatments, and interventions are needed to initiate a management 
program that can optimize a patient's or resident's prognosis and 
reduce the possibility of adverse events.
    A TEP convened by our data element contractor provided input on the 
proposed data elements for special services, treatments, and 
interventions. In a meeting held on January 5 and 6, 2017, this TEP 
found that these data elements are appropriate for standardization 
because they would provide useful clinical information to inform care 
planning and care coordination. The TEP affirmed that assessment of 
these services and interventions is standard clinical practice, and 
that the collection of these data by means of a list and checkbox 
format would conform with common workflow for PAC providers. A summary 
of the January 5 and 6, 2017 TEP meeting titled ``SPADE Technical 
Expert Panel Summary (Second Convening)'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Comments on the category of special services, treatments, and 
interventions were also submitted by stakeholders during the FY 2018 
IRF PPS proposed rule (82 FR 20726 through 20736) public comment 
period. One commenter supported adding the SPADEs for special services, 
treatments, and interventions. Others stated labor costs and staff 
burden would increase for data collection. The Medicare Payment 
Advisory Commission (MedPAC) suggested that a few other high-cost 
services, such as cardiac monitoring and specialty bed/surfaces, may 
warrant consideration for inclusion in future collection efforts. One 
commenter believes that the low frequency of the special services, 
treatments, and interventions in the IRF setting makes them not worth 
assessing for patients given the cost of data collection and reporting. 
A few commenters noted that many of these data elements should be 
obtainable from administrative data (that is, coding and Medicare 
claims), and therefore, assessing them through patient record review 
would be duplicated effort.
    Information on data element performance in the National Beta Test, 
which collected data between November 2017 and August 2018, is reported 
within each data element proposal below. Clinical staff who 
participated in the National Beta Test supported these data elements 
because of their importance in conveying patient or resident 
significant health care needs, complexity, and progress. However, 
clinical staff also noted that, despite the simple ``check box'' format 
of these data element, they sometimes needed to consult multiple 
information sources to determine a patient's or resident's treatments.
    We sought comment on our proposals to collect as standardized 
patient assessment data the following data with respect to special 
services, treatments, and interventions.
    Commenters submitted the following comments related to the proposed 
rule's discussion of special services, treatments, and interventions 
data elements.
    Comment: One commenter was supportive of collecting these data 
elements, noting that collection will help to better inform CMS and IRF 
providers on the severity and needs of patients in this setting.
    Response: We thank the commenter for the support of these items. We 
selected the Special Services, Treatments, and Interventions data 
elements for proposal as standardized data in part because of the 
attributes noted.
    Comment: Some commenters were concerned about the reliability of 
the Special Services, Treatments, and Interventions data elements, 
noting that the results of the National Beta Test indicated that these 
data elements had a low interrater reliability kappa statistic relative 
to other data elements in the test.
    Response: In the category of Special Services, Treatments, and 
Interventions, for SPADEs where kappas could be calculated, 1 data 
element and 2 sub-elements demonstrated overall reliabilities in the 
moderate range (0.41-0.60) and only 1 sub-element demonstrated an 
overall reliability in the slight/poor range (0.00-0.20). These overall 
reliabilities were as follows: 0.60 for the Therapeutic Diet data 
element; 0.55 for the ``Continuous'' sub-element of Oxygen Therapy; 
0.46 for the ``Other'' sub-element of IV Medications; and 0.13 for the 
``Anticoagulant'' sub-element of IV Medications. However, the overall 
reliabilities for all other data elements and sub-elements where kappas 
could be calculated were substantial/good or excellent/almost perfect. 
When looking at percent agreement--an alternative measure of interrater 
agreement--values of overall percent agreement for all Special 
Services, Treatments, and Interventions SPADEs and sub-elements ranged 
from 80 to 100 percent.
    Comment: Commenters also noted concern around the burden of 
completing these data elements, in particular because of their low 
frequency of occurrence in IRF settings. To reduce burden around 
collection of this information, commenters recommended that CMS explore 
obtaining this data via claims. Additionally, one commenter added that 
if these data elements are finalized, they should be collected at 
discharge only, to reduce administrative burden.
    Response: We appreciate the commenters' concern for burden on 
clinical staff due to completing assessments with respect to both 
admission and discharge. We believe that assessment of various special 
services, treatments, and interventions received by patients in the IRF 
setting will provide important information for care planning and 
resource use in IRFs. The assessments of the special services, 
treatments, and interventions with multiple responses are formatted as 
a ``check all that apply'' format. Therefore, when treatments do not 
apply--as the commenters note, this is the case for many IRF patients--
the assessor need only check one row for ``None of the Above.'' We will 
take under consideration the commenters' recommendation to explore the 
feasibility of collecting information on special services, treatments, 
and interventions through claims-based data. Regarding the 
recommendation to collect these SPADEs at discharge only, we state that 
it is clinically appropriate and important to the ultimate usefulness 
of these SPADEs that they are collected with respect to both admission 
and

[[Page 39123]]

discharge. For example, for patients coming from acute care or from the 
community, the admission assessment establishes a baseline for the IRF 
stay. For all patients, the admission assessment ensures that each 
patient is systematically assessed for a broad range of health and 
well-being issues, which we expect to inform care planning.
    Comment: One commenter expressed concern that the Special Services, 
Treatments, and Interventions data elements assess the presence or 
absence of something rather than the clinical rationale or patient 
outcomes. This commenter stressed the importance of bringing this 
assessment to ``the next level'' in order to determine impact of these 
treatments on patients' outcomes.
    Response: We agree with commenter's concern that recording the 
presence or absence of certain treatments is only a first step in 
characterizing the complexity that is often the cause of a patient's 
receipt of special services, treatments, and interventions. We clarify 
that all the SPADEs we proposed were intended as a minimum assessment 
and do not limit the ability of providers to conduct a more 
comprehensive evaluation of a patient's situation to identify the 
potential impacts on outcomes that the commenter describes.
    Comment: One commenter noted that the item numbering in the Special 
Services, Treatments, and Interventions data elements is extremely 
confusing and needs to be reworked.
    Response: Several patient assessment tools have traditionally 
combined letters and numbers, along with labels, to distinguish between 
data elements. The proposed data elements in the Special Services, 
Treatments, and Interventions section follow the conventions 
established by CMS. However, we will take this feedback into 
consideration in our evaluation and refinement of patient assessment 
instruments.
    Final decisions on the SPADEs are given below, following more 
detailed comments on each SPADE proposal.
 Cancer Treatment: Chemotherapy (IV, Oral, Other)
    In the FY 2020 IRF PPS proposed rule (84 FR 17297 through 17299), 
we proposed that the Chemotherapy (IV, Oral, Other) data element meets 
the definition of standardized patient assessment data with respect to 
special services, treatments, and interventions under section 
1899B(b)(1)(B)(iii) of the Act.
    As described in the FY 2018 IRF PPS proposed rule (82 FR 20726 
through 20727), chemotherapy is a type of cancer treatment that uses 
drugs to destroy cancer cells. It is sometimes used when a patient has 
a malignancy (cancer), which is a serious, often life-threatening or 
life-limiting condition. Both intravenous (IV) and oral chemotherapy 
have serious side effects, including nausea/vomiting, extreme fatigue, 
risk of infection due to a suppressed immune system, anemia, and an 
increased risk of bleeding due to low platelet counts. Oral 
chemotherapy can be as potent as chemotherapy given by IV and can be 
significantly more convenient and less resource-intensive to 
administer. Because of the toxicity of these agents, special care must 
be exercised in handling and transporting chemotherapy drugs. IV 
chemotherapy is administered either peripherally, or more commonly, 
given via an indwelling central line, which raises the risk of 
bloodstream infections. Given the significant burden of malignancy, the 
resource intensity of administering chemotherapy, and the side effects 
and potential complications of these highly-toxic medications, 
assessing the receipt of chemotherapy is important in the PAC setting 
for care planning and determining resource use. The need for 
chemotherapy predicts resource intensity, both because of the 
complexity of administering these potent, toxic drug combinations under 
specific protocols, and because of what the need for chemotherapy 
signals about the patient's underlying medical condition. Furthermore, 
the resource intensity of IV chemotherapy is higher than for oral 
chemotherapy, as the protocols for administration and the care of the 
central line (if present) for IV chemotherapy require significant 
resources.
    The Chemotherapy (IV, Oral, Other) data element consists of a 
principal data element (Chemotherapy) and three response option sub-
elements: IV chemotherapy, which is generally resource-intensive; Oral 
chemotherapy, which is less invasive and generally requires less 
intensive administration protocols; and a third category, Other, 
provided to enable the capture of other less common chemotherapeutic 
approaches. This third category is potentially associated with higher 
risks and is more resource intensive due to delivery by other routes 
(for example, intraventricular or intrathecal). If the assessor 
indicates that the patient is receiving chemotherapy on the principal 
Chemotherapy data element, the assessor would then indicate by which 
route or routes (for example, IV, Oral, Other) the chemotherapy is 
administered.
    A single Chemotherapy data element that does not include the 
proposed three sub-elements is currently in use in the MDS in SNFs. For 
more information on the Chemotherapy (IV, Oral, Other) data element, we 
refer readers to the document titled ``Final Specifications for IRF QRP 
Quality Measures and Standardized Patient Assessment Data Elements,'' 
available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    The Chemotherapy data element was first proposed as a standardized 
patient assessment data element in the FY 2018 IRF PPS proposed rule 
(82 FR 20726 through 20727). In that proposed rule, we stated that the 
proposal was informed by input we received through a call for input 
published on the CMS Measures Management System Blueprint website. 
Input submitted from August 12 to September 12, 2016 noted support for 
the IV Chemotherapy data element and suggested it be included as 
standardized patient assessment data. We also stated that those 
commenters had noted that assessing the use of chemotherapy services is 
relevant to share across the care continuum to facilitate care 
coordination and care transitions and noted the validity of the data 
element. Commenters also noted the importance of capturing all types of 
chemotherapy, regardless of route, and stated that collecting data only 
on patients and residents who received chemotherapy by IV would limit 
the usefulness of this standardized data element. A summary report for 
the August 12 to September 12, 2016 public comment period titled 
``SPADE August 2016 Public Comment Summary Report'' is available at 
https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In response to our proposal in the FY 2018 IRF PPS proposed rule, 
we received public comments in support of the special services, 
treatments, and interventions data elements in general; no additional 
comments were received that were specific to the Chemotherapy data 
element.
    Subsequent to receiving comments on the FY 2018 IRF PPS rule, the 
Chemotherapy data element was included in the National Beta Test of 
candidate data elements conducted by our data element contractor from 
November 2017 to August 2018. Results

[[Page 39124]]

of this test found the Chemotherapy data element to be feasible and 
reliable for use with PAC patients and residents. More information 
about the performance of the Chemotherapy data element in the National 
Beta Test can be found in the document titled ``Final Specifications 
for IRF QRP Quality Measures and Standardized Patient Assessment Data 
Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In addition, our data element contractor convened a TEP on 
September 17, 2018, for the purpose of soliciting input on the proposed 
standardized patient assessment data elements. Although the TEP members 
did not specifically discuss the Chemotherapy data element, the TEP 
members supported the assessment of the special services, treatments, 
and interventions included in the National Beta Test with respect to 
both admission and discharge. A summary of the September 17, 2018 TEP 
meeting titled ``SPADE Technical Expert Panel Summary (Third 
Convening)'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We also held SODFs and small-group discussions with PAC providers 
and other stakeholders in 2018 for the purpose of updating the public 
about our ongoing SPADE development efforts. Finally, on November 27, 
2018, our data element contractor hosted a public meeting of 
stakeholders to present the results of the National Beta Test and 
solicit additional comments. General input on the testing and item 
development process and concerns about burden were received from 
stakeholders during this meeting and via email through February 1, 
2019. A summary of the public input received from the November 27, 2018 
stakeholder meeting titled ``Input on Standardized Patient Assessment 
Data Elements (SPADEs) Received After November 27, 2018 Stakeholder 
Meeting'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Taking together the importance of assessing for chemotherapy, 
stakeholder input, and strong test results, we proposed that the 
Chemotherapy (IV, Oral, Other) data element with a principal data 
element and three sub-elements meet the definition of standardized 
patient assessment data with respect to special services, treatments, 
and interventions under section 1899B(b)(1)(B)(iii) of the Act and to 
adopt the Chemotherapy (IV, Oral, Other) data element as standardized 
patient assessment data for use in the IRF QRP.
    A commenter submitted the following comment related to the proposed 
rule's discussion of the Chemotherapy data element.
    Comment: One commenter agreed that it is important to know if a 
patient is receiving chemotherapy for cancer and the method of 
administration, but also expressed concern about the lack of an 
association with a patient outcome. This commenter noted that 
implications of chemotherapy for patients needing speech-language 
pathology services include chemotherapy-related cognitive impairment, 
dysphagia, and speech- and voice-related deficits.
    Response: We appreciate the commenter's concern. We agree with the 
commenter that chemotherapy can create related treatment needs for 
patients, such as the examples noted by the commenter. However, we 
believe that it is not feasible for SPADEs to capture all of a 
patient's needs related to any given treatment, and we maintain that 
the Special Services, Treatments, and Interventions SPADEs provide a 
common foundation of clinical assessment, which can be built on by the 
individual provider or a patient's care team.
    After careful consideration of the public comments we received, we 
are finalizing our proposal to adopt the Chemotherapy (IV, Oral, Other) 
data element as standardized patient assessment data beginning with the 
FY 2022 IRF QRP as proposed.
 Cancer Treatment: Radiation
    In the FY 2020 IRF PPS proposed rule (84 FR 17299), we proposed 
that the Radiation data element meets the definition of standardized 
patient assessment data with respect to special services, treatments, 
and interventions under section 1899B(b)(1)(B)(iii) of the Act.
    As described in the FY 2018 IRF PPS proposed rule (82 FR 20727 
through 20728), radiation is a type of cancer treatment that uses high-
energy radioactivity to stop cancer by damaging cancer cell DNA, but it 
can also damage normal cells. Radiation is an important therapy for 
particular types of cancer, and the resource utilization is high, with 
frequent radiation sessions required, often daily for a period of 
several weeks. Assessing whether a patient or resident is receiving 
radiation therapy is important to determine resource utilization 
because PAC patients and residents will need to be transported to and 
from radiation treatments, and monitored and treated for side effects 
after receiving this intervention. Therefore, assessing the receipt of 
radiation therapy, which would compete with other care processes given 
the time burden, would be important for care planning and care 
coordination by PAC providers.
    The proposed data element consists of the single Radiation data 
element. The Radiation data element is currently in use in the MDS in 
SNFs. For more information on the Radiation data element, we refer 
readers to the document titled ``Final Specifications for IRF QRP 
Quality Measures and Standardized Patient Assessment Data Elements,'' 
available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    The Radiation data element was first proposed as a standardized 
patient assessment data element in the FY 2018 IRF PPS proposed rule 
(82 FR 20727 through 20728). In that proposed rule, we stated that the 
proposal was informed by input we received through a call for input 
published on the CMS Measures Management System Blueprint website. 
Input submitted from August 12 to September 12, 2016 noted support for 
the Radiation data element, noting its importance and clinical 
usefulness for patients and residents in PAC settings, due to the side 
effects and consequences of radiation treatment on patients and 
residents that need to be considered in care planning and care 
transitions, the feasibility of the item, and the potential for it to 
improve quality. A summary report for the August 12 to September 12, 
2016 public comment period titled ``SPADE August 2016 Public Comment 
Summary Report'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In response to our proposal in the FY 2018 IRF PPS proposed rule, 
we received public comments in support of the special services, 
treatments, and interventions data elements in general; no additional 
comments were received

[[Page 39125]]

that were specific to the Radiation data element.
    Subsequent to receiving comments on the FY 2018 IRF PPS rule, the 
Radiation data element was included in the National Beta Test of 
candidate data elements conducted by our data element contractor from 
November 2017 to August 2018. Results of this test found the Radiation 
data element to be feasible and reliable for use with PAC patients and 
residents. More information about the performance of the Radiation data 
element in the National Beta Test can be found in the document titled 
``Final Specifications for IRF QRP Quality Measures and Standardized 
Patient Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In addition, our data element contractor convened a TEP on 
September 17, 2018, for the purpose of soliciting input on the proposed 
standardized patient assessment data elements. Although the TEP members 
did not specifically discuss the Radiation data element, the TEP 
members supported the assessment of the special services, treatments, 
and interventions included in the National Beta Test with respect to 
both admission and discharge. A summary of the September 17, 2018 TEP 
meeting titled ``SPADE Technical Expert Panel Summary (Third 
Convening)'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We also held SODFs and small-group discussions with PAC providers 
and other stakeholders in 2018 for the purpose of updating the public 
about our ongoing SPADE development efforts. Finally, on November 27, 
2018, our data element contractor hosted a public meeting of 
stakeholders to present results of the National Beta Test and solicit 
additional comments. General input on the testing and item development 
process and concerns about burden were received from stakeholders 
during this meeting and via email through February 1, 2019. A summary 
of the public input received from the November 27, 2018 stakeholder 
meeting titled ``Input on Standardized Patient Assessment Data Elements 
(SPADEs) Received After November 27, 2018 Stakeholder Meeting'' is 
available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Taking together the importance of assessing for radiation, 
stakeholder input, and strong test results, we proposed that the 
Radiation data element meets the definition of standardized patient 
assessment data with respect to special services, treatments, and 
interventions under section 1899B(b)(1)(B)(iii) of the Act and to adopt 
the Radiation data element as standardized patient assessment data for 
use in the IRF QRP.
    A commenter submitted the following comment related to the proposed 
rule's discussion of the Radiation data element.
    Comment: One commenter expressed concern that the Radiation data 
element assesses whether a patient is receiving radiation for cancer 
treatment, but does not identify the rationale for and outcomes 
associated with radiation. The commenter noted that implications of 
radiation for patients needing speech-language pathology services 
include reduced head and neck range of motion due to radiation or 
severe fibrosis, scar bands, and reconstructive surgery complications 
and that these can impact both communication and swallowing abilities.
    Response: We appreciate the commenter's concern. We agree with the 
commenter that radiation can create related treatment needs for 
patients, such as the examples noted by the commenter. However, we 
believe that it is not feasible for SPADEs to capture all of a 
patient's needs related to any given treatment, and we maintain that 
the Special Services, Treatments, and Interventions SPADEs provide a 
common foundation of clinical assessment, which can be built on by the 
individual provider or a patient's care team.
    After careful consideration of the public comments we received, we 
are finalizing our proposal to adopt the Radiation data element as 
standardized patient assessment data beginning with the FY 2022 IRF QRP 
as proposed.
 Respiratory Treatment: Oxygen Therapy (Intermittent, 
Continuous, High-concentration Oxygen Delivery System)
    In the FY 2020 IRF PPS proposed rule (84 FR 17299 through 17300), 
we proposed that the Oxygen Therapy (Intermittent, Continuous, High-
concentration Oxygen Delivery System) data element meets the definition 
of standardized patient assessment data with respect to special 
services, treatments, and interventions under section 
1899B(b)(1)(B)(iii) of the Act.
    As described in the FY 2018 IRF PPS proposed rule (82 FR 20728), we 
proposed a similar data element related to oxygen therapy. Oxygen 
therapy provides a patient or resident with extra oxygen when medical 
conditions such as chronic obstructive pulmonary disease, pneumonia, or 
severe asthma prevent the patient or resident from getting enough 
oxygen from breathing. Oxygen administration is a resource-intensive 
intervention, as it requires specialized equipment such as a source of 
oxygen, delivery systems (for example, oxygen concentrator, liquid 
oxygen containers, and high-pressure systems), the patient interface 
(for example, nasal cannula or mask), and other accessories (for 
example, regulators, filters, tubing). The data element proposed here 
captures patient or resident use of three types of oxygen therapy 
(intermittent, continuous, and high-concentration oxygen delivery 
system), which reflects the intensity of care needed, including the 
level of monitoring and bedside care required. Assessing the receipt of 
this service is important for care planning and resource use for PAC 
providers.
    The proposed data element, Oxygen Therapy, consists of the 
principal Oxygen Therapy data element and three response option sub-
elements: Continuous (whether the oxygen was delivered continuously, 
typically defined as > =14 hours per day); Intermittent; or High-
concentration Oxygen Delivery System. Based on public comments and 
input from expert advisors about the importance and clinical usefulness 
of documenting the extent of oxygen use, we added a third sub-element, 
high-concentration oxygen delivery system, to the sub-elements, which 
previously included only intermittent and continuous. If the assessor 
indicates that the patient is receiving oxygen therapy on the principal 
oxygen therapy data element, the assessor then would indicate the type 
of oxygen the patient receives (for example, Intermittent, Continuous, 
High-concentration oxygen delivery system).
    These three proposed sub-elements were developed based on similar 
data elements that assess oxygen therapy, currently in use in the MDS 
in SNFs (``Oxygen Therapy''), previously used in the OASIS (``Oxygen 
(intermittent or continuous)''), and a data element tested in the PAC 
PRD that focused on intensive oxygen therapy (``High O2 Concentration 
Delivery System with FiO2 > 40 percent''). For more information on the 
proposed Oxygen

[[Page 39126]]

Therapy (Continuous, Intermittent, High-concentration oxygen delivery 
system) data element, we refer readers to the document titled ``Final 
Specifications for IRF QRP Quality Measures and Standardized Patient 
Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    The Oxygen Therapy (Intermittent, Continuous) data element was 
first proposed as standardized patient assessment data in the FY 2018 
IRF PPS proposed rule (82 FR 20728). In that proposed rule, we stated 
that the proposal was informed by input we received on the single data 
element, Oxygen (inclusive of intermittent and continuous oxygen use), 
through a call for input published on the CMS Measures Management 
System Blueprint website. Input submitted from August 12 to September 
12, 2016, noted the importance of the Oxygen data element, noting 
feasibility of this item in PAC, and the relevance of it to 
facilitating care coordination and supporting care transitions, but 
suggesting that the extent of oxygen use be documented. A summary 
report for the August 12 to September 12, 2016 public comment period 
titled ``SPADE August 2016 Public Comment Summary Report'' is available 
at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In response to our proposal in the FY 2018 IRF PPS proposed rule, 
we received public comments in support of the special services, 
treatments, and interventions data elements in general; no additional 
comments were received that were specific to the Oxygen Therapy 
(Intermittent, Continuous) data element.
    Subsequent to receiving comments on the FY 2018 IRF PPS rule, the 
Oxygen Therapy data element was included in the National Beta Test of 
candidate data elements conducted by our data element contractor from 
November 2017 to August 2018. Results of this test found the Oxygen 
Therapy data element to be feasible and reliable for use with PAC 
patients and residents. More information about the performance of the 
Oxygen Therapy data element in the National Beta Test can be found in 
the document titled ``Final Specifications for IRF QRP Quality Measures 
and Standardized Patient Assessment Data Elements,'' available at 
https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In addition, our data element contractor convened a TEP on 
September 17, 2018, for the purpose of soliciting input on the proposed 
standardized patient assessment data elements. Although the TEP did not 
specifically discuss the Oxygen Therapy data element, the TEP supported 
the assessment of the special services, treatments, and interventions 
included in the National Beta Test with respect to both admission and 
discharge. A summary of the September 17, 2018 TEP meeting titled 
``SPADE Technical Expert Panel Summary (Third Convening)'' is available 
at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We also held SODFs and small-group discussions with PAC providers 
and other stakeholders in 2018 for the purpose of updating the public 
about our ongoing SPADE development efforts. Finally, on November 27, 
2018, our data element contractor hosted a public meeting of 
stakeholders to present the results of the National Beta Test and 
solicit additional comments. General input on the testing and item 
development process and concerns about burden were received from 
stakeholders during this meeting and via email through February 1, 
2019. A summary of the public input received from the November 27, 2018 
stakeholder meeting titled ``Input on Standardized Patient Assessment 
Data Elements (SPADEs) Received After November 27, 2018 Stakeholder 
Meeting'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Taking together the importance of assessing oxygen therapy, 
stakeholder input, and strong test results, we proposed that the Oxygen 
Therapy (Intermittent, Continuous, High-concentration Oxygen Delivery 
System) data element with a principal data element and three sub-
elements meets the definition of standardized patient assessment data 
with respect to special services, treatments, and interventions under 
section 1899B(b)(1)(B)(iii) of the Act and to adopt the Oxygen Therapy 
(Intermittent, Continuous, High-concentration Oxygen Delivery System) 
data element as standardized patient assessment data for use in the IRF 
QRP.
    We invited public comment on this proposal. While we received 
support from some commenters on the Special Services, Treatments, and 
Interventions section (IX.G.2 in this final rule) and its proposals as 
a whole (section IX.F in this final rule), we did not receive any 
specific comments on the Oxygen Therapy (Intermittent, Continuous, 
High-concentration Oxygen Delivery System) data element in particular.
    After careful consideration of the public comments we received on 
the category of Special Services, Treatments, and Interventions, we are 
finalizing our proposal to adopt the Oxygen Therapy (Intermittent, 
Continuous, High-Concentration Oxygen Delivery System) data element as 
standardized patient assessment data beginning with the FY 2022 IRF QRP 
as proposed.
 Respiratory Treatment: Suctioning (Scheduled, as Needed)
    In the FY 2020 IRF PPS proposed rule (84 FR 17300 through 17302), 
we proposed that the Suctioning (Scheduled, As needed) data element 
meets the definition of standardized patient assessment data with 
respect to special services, treatments, and interventions under 
section 1899B(b)(1)(B)(iii) of the Act.
    As described in the FY 2018 IRF PPS proposed rule (82 FR 20728 
through 20729), suctioning is a process used to clear secretions from 
the airway when a person cannot clear those secretions on his or her 
own. It is done by aspirating secretions through a catheter connected 
to a suction source. Types of suctioning include oropharyngeal and 
nasopharyngeal suctioning, nasotracheal suctioning, and suctioning 
through an artificial airway such as a tracheostomy tube. Oropharyngeal 
and nasopharyngeal suctioning are a key part of many patients' or 
residents' care plans, both to prevent the accumulation of secretions 
than can lead to aspiration pneumonias (a common condition in patients 
and residents with inadequate gag reflexes), and to relieve 
obstructions from mucus plugging during an acute or chronic respiratory 
infection, which often lead to desaturations and increased respiratory 
effort. Suctioning can be done on a scheduled basis if the patient is 
judged to clinically benefit from regular interventions, or can be done 
as needed when secretions become so prominent that gurgling or choking 
is noted, or a sudden desaturation occurs from a mucus plug. As 
suctioning is generally performed by a care provider

[[Page 39127]]

rather than independently, this intervention can be quite resource 
intensive if it occurs every hour, for example, rather than once a 
shift. It also signifies an underlying medical condition that prevents 
the patient from clearing his/her secretions effectively (such as after 
a stroke, or during an acute respiratory infection). Generally, 
suctioning is necessary to ensure that the airway is clear of 
secretions which can inhibit successful oxygenation of the individual. 
The intent of suctioning is to maintain a patent airway, the loss of 
which can lead to death or complications associated with hypoxia.
    The Suctioning (Scheduled, As needed) data element consists of a 
principal data element, and two sub-elements: Scheduled and As needed. 
These sub-elements capture two types of suctioning. Scheduled indicates 
suctioning based on a specific frequency, such as every hour. As needed 
means suctioning only when indicated. If the assessor indicates that 
the patient is receiving suctioning on the principal Suctioning data 
element, the assessor would then indicate the frequency (for example, 
Scheduled, As needed). The proposed data element is based on an item 
currently in use in the MDS in SNFs which does not include our proposed 
two sub-elements, as well as data elements tested in the PAC PRD that 
focused on the frequency of suctioning required for patients and 
residents with tracheostomies (``Trach Tube with Suctioning: Specify 
most intensive frequency of suctioning during stay [Every __hours]''). 
For more information on the Suctioning data element, we refer readers 
to the document titled ``Final Specifications for IRF QRP Quality 
Measures and Standardized Patient Assessment Data Elements,'' available 
at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    The Suctioning data element was first proposed as standardized 
patient assessment data elements in the FY 2018 IRF PPS proposed rule 
(82 FR 20728 through 20729). In that proposed rule, we stated that the 
proposal was informed by input we received through a call for input 
published on the CMS Measures Management System Blueprint website. 
Input submitted from August 12 to September 12, 2016 noted support for 
the Suctioning data element. The input noted the feasibility of this 
item in PAC, and the relevance of this data element to facilitating 
care coordination and supporting care transitions.
    We also stated that those commenters had suggested that we examine 
the frequency of suctioning to better understand the use of staff time, 
the impact on a patient or resident's capacity to speak and swallow, 
and intensity of care required. Based on these comments, we decided to 
add two sub-elements (Scheduled and As needed) to the suctioning 
element. The proposed Suctioning data element includes both the 
principal Suctioning data element that is included on the MDS in SNFs 
and two sub-elements, Scheduled and As needed. A summary report for the 
August 12 to September 12, 2016 public comment period titled ``SPADE 
August 2016 Public Comment Summary Report'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In response to our proposal in the FY 2018 IRF PPS proposed rule, 
we received public comments in support of the special services, 
treatments, and interventions data elements in general; no additional 
comments were received that were specific to the Suctioning data 
element. Subsequent to receiving comments on the FY 2018 IRF PPS rule, 
the Suctioning data element was included in the National Beta Test of 
candidate data elements conducted by our data element contractor from 
November 2017 to August 2018. Results of this test found the Suctioning 
data element to be feasible and reliable for use with PAC patients and 
residents. More information about the performance of the Suctioning 
data element in the National Beta Test can be found in the document 
titled ``Final Specifications for IRF QRP Quality Measures and 
Standardized Patient Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In addition, our data element contractor convened a TEP on 
September 17, 2018, for the purpose of soliciting input on the proposed 
standardized patient assessment data elements. Although the TEP did not 
specifically discuss the Suctioning data element, the TEP supported the 
assessment of the special services, treatments, and interventions 
included in the National Beta Test with respect to both admission and 
discharge. A summary of the September 17, 2018 TEP meeting titled 
``SPADE Technical Expert Panel Summary (Third Convening)'' is available 
at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We also held SODFs and small-group discussions with PAC providers 
and other stakeholders in 2018 for the purpose of updating the public 
about our ongoing SPADE development efforts. Finally, on November 27, 
2018, our data element contractor hosted a public meeting of 
stakeholders to present the results of the National Beta Test and 
solicited additional comments. General input on the testing and item 
development process and concerns about burden were received from 
stakeholders during this meeting and via email through February 1, 
2019. A summary of the public input received from the November 27, 2018 
stakeholder meeting titled ``Input on Standardized Patient Assessment 
Data Elements (SPADEs) Received After November 27, 2018 Stakeholder 
Meeting'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Taking together the importance of assessing for suctioning, 
stakeholder input, and strong test results, we proposed that the 
Suctioning (Scheduled, As needed) data element with a principal data 
element and two sub-elements meets the definition of standardized 
patient assessment data with respect to special services, treatments, 
and interventions under section 1899B(b)(1)(B)(iii) of the Act and to 
adopt the Suctioning (Scheduled, As needed) data element as 
standardized patient assessment data for use in the IRF QRP.
    A commenter submitted the following comment related to the proposed 
rule's discussion of the Suctioning data element.
    Comment: One commenter requested that this data element also assess 
the frequency of suctioning, as it can impact resource utilization and 
potential medication changes in the plan of care.
    Response: We appreciate the commenter's feedback that the response 
options for this data element may not fully capture impacts to resource 
utilization and care plans. The Suctioning data element does include 
sub-elements to identify if suctioning is performed on a ``Scheduled'' 
or ``As Needed'' basis, but it does not directly

[[Page 39128]]

assess the frequency of suctioning by, for example, asking an assessor 
to specify how often suctioning is scheduled. As finalized, this data 
element differentiates between patients who only occasionally need 
suctioning, and patients for whom assessment of suctioning needs is a 
frequent and routine part of the care (that is, where suctioning is 
performed on a schedule according to physician instructions). In our 
work to identify standardized data elements, we have strived to balance 
the scope and level of detail of the data elements against the 
potential burden placed on patients and providers. However, we clarify 
that any SPADE is intended as a minimum assessment and does not limit 
the ability of providers to conduct a more comprehensive evaluation of 
a patient's situation to identify the potential impacts on outcomes 
that the commenter describes.
    After careful consideration of the public comments we received, we 
are finalizing our proposal to adopt the Suctioning (Scheduled, As 
needed) data element as standardized patient assessment data beginning 
with the FY 2022 IRF QRP as proposed.
 Respiratory Treatment: Tracheostomy Care
    In the FY 2020 IRF PPS proposed rule (84 FR 17302), we proposed 
that the Tracheostomy Care data element meets the definition of 
standardized patient assessment data with respect to special services, 
treatments, and interventions under section 1899B(b)(1)(B)(iii) of the 
Act.
    As described in the FY 2018 IRF PPS proposed rule (82 FR 20729 
through 20730), a tracheostomy provides an air passage to help a 
patient or resident breathe when the usual route for breathing is 
obstructed or impaired. Generally, in all of these cases, suctioning is 
necessary to ensure that the tracheostomy is clear of secretions, which 
can inhibit successful oxygenation of the individual. Often, 
individuals with tracheostomies are also receiving supplemental 
oxygenation. The presence of a tracheostomy, albeit permanent or 
temporary, warrants careful monitoring and immediate intervention if 
the tracheostomy becomes occluded or if the device used becomes 
dislodged. While in rare cases the presence of a tracheostomy is not 
associated with increased care demands (and in some of those instances, 
the care of the ostomy is performed by the patient) in general the 
presence of such as device is associated with increased patient risk, 
and clinical care services will necessarily include close monitoring to 
ensure that no life-threatening events occur as a result of the 
tracheostomy. In addition, tracheostomy care, which primarily consists 
of cleansing, dressing changes, and replacement of the tracheostomy 
cannula (tube), is a critical part of the care plan. Regular cleansing 
is important to prevent infection, such as pneumonia, and to prevent 
any occlusions with which there are risks for inadequate oxygenation.
    The proposed data element consists of the single Tracheostomy Care 
data element. The proposed data element is currently in use in the MDS 
in SNFs (``Tracheostomy care''). For more information on the 
Tracheostomy Care data element, we refer readers to the document titled 
``Final Specifications for IRF QRP Quality Measures and Standardized 
Patient Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    The Tracheostomy Care data element was first proposed as a 
standardized patient assessment data element in the FY 2018 IRF PPS 
proposed rule (82 FR 20729 through 20730). In that proposed rule, we 
stated that the proposal was informed by input we received on the 
Tracheostomy Care data element through a call for input published on 
the CMS Measures Management System Blueprint website. Input submitted 
from August 12 to September 12, 2016 noted support for this data 
element, noting the feasibility of this item in PAC, and the relevance 
of this data element to facilitating care coordination and supporting 
care transitions. A summary report for the August 12 to September 12, 
2016 public comment period titled ``SPADE August 2016 Public Comment 
Summary Report'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In response to our proposal in the FY 2018 IRF PPS proposed rule, 
we received public comments in support of the special services, 
treatments, and interventions data elements in general; no additional 
comments were received that were specific to the Tracheostomy Care data 
element.
    Subsequent to receiving comments on the FY 2018 IRF PPS rule, the 
Tracheostomy Care data element was included in the National Beta Test 
of candidate data elements conducted by our data element contractor 
from November 2017 to August 2018. Results of this test found the 
Tracheostomy Care data element to be feasible and reliable for use with 
PAC patients and residents. More information about the performance of 
the Tracheostomy Care data element in the National Beta Test can be 
found in the document titled ``Final Specifications for IRF QRP Quality 
Measures and Standardized Patient Assessment Data Elements,'' at 
https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In addition, our data element contractor convened a TEP on 
September 17, 2018, for the purpose of soliciting input on the proposed 
standardized patient assessment data elements. Although the TEP did not 
specifically discuss the Tracheostomy Care data element, the TEP 
supported the assessment of the special services, treatments, and 
interventions included in the National Beta Test with respect to both 
admission and discharge. A summary of the September 17, 2018 TEP 
meeting titled ``SPADE Technical Expert Panel Summary (Third 
Convening)'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We also held SODFs and small-group discussions with PAC providers 
and other stakeholders in 2018 for the purpose of updating the public 
about our ongoing SPADE development efforts. Finally, on November 27, 
2018, our data element contractor hosted a public meeting of 
stakeholders to present the results of the National Beta Test and 
solicit additional comments. General input on the testing and item 
development process and concerns about burden were received from 
stakeholders during this meeting and via email through February 1, 
2019. A summary of the public input received from the November 27, 2018 
stakeholder meeting titled ``Input on Standardized Patient Assessment 
Data Elements (SPADEs) Received After November 27, 2018 Stakeholder 
Meeting'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Taking together the importance of assessing for tracheostomy care, 
stakeholder input, and strong test results, we proposed that the

[[Page 39129]]

Tracheostomy Care data element meets the definition of standardized 
patient assessment data with respect to special services, treatments, 
and interventions under section 1899B(b)(1)(B)(iii) of the Act and to 
adopt the Tracheostomy Care data element as standardized patient 
assessment data for use in the IRF QRP.
    We invited public comment on this proposal. While we received 
support from some commenters on Special Services, Treatments, and 
Interventions as a whole (section IX.G.2 in this final rule), we did 
not receive any specific comments on Tracheostomy Care data element.
    After careful consideration of the public comments we received on 
the category of Special Services, Treatments, and Interventions, we are 
finalizing our proposal to adopt the Tracheostomy Care data element as 
standardized patient assessment data beginning with the FY 2022 IRF QRP 
as proposed.
 Respiratory Treatment: Non-Invasive Mechanical Ventilator 
(BiPAP, CPAP)
    In the FY 2020 IRF PPS proposed rule (84 FR 17303), we proposed 
that the Non-invasive Mechanical Ventilator (Bilevel Positive Airway 
Pressure [BiPAP], Continuous Positive Airway Pressure [CPAP]) data 
element meets the definition of standardized patient assessment data 
with respect to special services, treatments, and interventions under 
section 1899B(b)(1)(B)(iii) of the Act.
    As described in the FY 2018 IRF PPS proposed rule (82 FR 20730), 
BiPAP and CPAP are respiratory support devices that prevent the airways 
from closing by delivering slightly pressurized air via electronic 
cycling throughout the breathing cycle (BiPAP) or through a mask 
continuously (CPAP). Assessment of non-invasive mechanical ventilation 
is important in care planning, as both CPAP and BiPAP are resource-
intensive (although less so than invasive mechanical ventilation) and 
signify underlying medical conditions about the patient or resident who 
requires the use of this intervention. Particularly when used in 
settings of acute illness or progressive respiratory decline, 
additional staff (for example, respiratory therapists) are required to 
monitor and adjust the CPAP and BiPAP settings and the patient or 
resident may require more nursing resources.
    The proposed data element, Non-invasive Mechanical Ventilator 
(BiPAP, CPAP), consists of the principal Non-invasive Mechanical 
Ventilator data element and two response option sub-elements: BiPAP and 
CPAP. If the assessor indicates that the patient is receiving non-
invasive mechanical ventilation on the principal Non-invasive 
Mechanical Ventilator data element, the assessor would then indicate 
which type (for example, BiPAP, CPAP). Data elements that assess non-
invasive mechanical ventilation are currently included on LCDS for the 
LTCH setting (``Non-invasive Ventilator (BiPAP, CPAP)''), and the MDS 
for the SNF setting (``Non-invasive Mechanical Ventilator (BiPAP/
CPAP)''). For more information on the Non-invasive Mechanical 
Ventilator (BiPAP, CPAP) data element, we refer readers to the document 
titled ``Final Specifications for IRF QRP Quality Measures and 
Standardized Patient Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    The Non-invasive Mechanical Ventilator data element was first 
proposed as standardized patient assessment data elements in the FY 
2018 IRF PPS proposed rule (82 FR 20730). In that proposed rule, we 
stated that the proposal was informed by input we received through a 
call for input published on the CMS Measures Management System 
Blueprint website. Input submitted from August 12 to September 12, 2016 
on a single data element, BiPAP/CPAP, that captures equivalent clinical 
information but uses a different label than the data element currently 
used in the MDS in SNFs and LCDS, noted support for this data element, 
noting the feasibility of these items in PAC, and the relevance of this 
data element for facilitating care coordination and supporting care 
transitions. In addition, we also stated that some commenters supported 
separating out BiPAP and CPAP as distinct sub-elements, as they are 
therapies used for different types of patients and residents. A summary 
report for the August 12 to September 12, 2016 public comment period 
titled ``SPADE August 2016 Public Comment Summary Report'' is available 
at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In response to our proposal in the FY 2018 IRF PPS proposed rule, 
we received public comments in support of the special services, 
treatments, and interventions data elements in general. One commenter 
noted appreciation of the revisions to the Non-invasive Mechanical 
Ventilator data element in response to comments submitted during a 
public input period held from August 12 to September 12, 2016.
    Subsequent to receiving comments on the FY 2018 IRF PPS rule, the 
Non-invasive Mechanical Ventilator data element was included in the 
National Beta Test of candidate data elements conducted by our data 
element contractor from November 2017 to August 2018. Results of this 
test found the Non-invasive Mechanical Ventilator data element to be 
feasible and reliable for use with PAC patients and residents. More 
information about the performance of the Non-invasive Mechanical 
Ventilator data element in the National Beta Test can be found in the 
document titled ``Final Specifications for IRF QRP Quality Measures and 
Standardized Patient Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In addition, our data element contractor convened a TEP on 
September 17, 2018, for the purpose of soliciting input on the proposed 
standardized patient assessment data elements. Although the TEP did not 
specifically discuss the Non-invasive Mechanical Ventilator data 
element, the TEP supported the assessment of the special services, 
treatments, and interventions included in the National Beta Test with 
respect to both admission and discharge. A summary of the September 17, 
2018 TEP meeting titled ``SPADE Technical Expert Panel Summary (Third 
Convening)'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We also held SODFs and small-group discussions with PAC providers 
and other stakeholders in 2018 for the purpose of updating the public 
about our ongoing SPADE development efforts. Finally, on November 27, 
2018, our data element contractor hosted a public meeting of 
stakeholders to present the results of the National Beta Test and 
solicit additional comments. General input on the testing and item 
development process and concerns about burden were received from 
stakeholders during this meeting and via email through February 1, 
2019. A summary of the public input received from the November 27, 2018 
stakeholder meeting titled ``Input on Standardized

[[Page 39130]]

Patient Assessment Data Elements (SPADEs) Received After November 27, 
2018 Stakeholder Meeting'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Taking together the importance of assessing for non-invasive 
mechanical ventilation, stakeholder input, and strong test results, we 
proposed that the Non-invasive Mechanical Ventilator (BiPAP, CPAP) data 
element with a principal data element and two sub-elements meets the 
definition of standardized patient assessment data with respect to 
special services, treatments, and interventions under section 
1899B(b)(1)(B)(iii) of the Act and to adopt the Non-invasive Mechanical 
Ventilator (BiPAP, CPAP) data element as standardized patient 
assessment data for use in the IRF QRP.
    We invited public comment on this proposal. While we received 
support from some commenters on Special Services, Treatments, and 
Interventions as a whole (section IX.G.2 in this final rule), we did 
not receive any specific comments on the Non-invasive Mechanical 
Ventilator (BiPAP, CPAP) data element.
    After careful consideration of the public comments we received on 
the category of Special Services, Treatments, and Interventions, we are 
finalizing our proposal to adopt the Non-invasive Mechanical Ventilator 
(BiPAP, CPAP) data element as standardized patient assessment data 
beginning with the FY 2022 IRF QRP as proposed.
 Respiratory Treatment: Invasive Mechanical Ventilator
    In the FY 2020 IRF PPS proposed rule (84 FR 17304), we proposed 
that the Invasive Mechanical Ventilator data element meets the 
definition of standardized patient assessment data with respect to 
special services, treatments, and interventions under section 
1899B(b)(1)(B)(iii) of the Act.
    As described in the FY 2018 IRF PPS proposed rule (82 FR 20730 
through 20731), invasive mechanical ventilation includes ventilators 
and respirators that ventilate the patient through a tube that extends 
via the oral airway into the pulmonary region or through a surgical 
opening directly into the trachea. Thus, assessment of invasive 
mechanical ventilation is important in care planning and risk 
mitigation. Ventilation in this manner is a resource-intensive therapy 
associated with life-threatening conditions without which the patient 
or resident would not survive. However, ventilator use has inherent 
risks requiring close monitoring. Failure to adequately care for the 
patient or resident who is ventilator dependent can lead to iatrogenic 
events such as death, pneumonia, and sepsis. Mechanical ventilation 
further signifies the complexity of the patient's underlying medical or 
surgical condition. Of note, invasive mechanical ventilation is 
associated with high daily and aggregate costs.\112\
---------------------------------------------------------------------------

    \112\ Wunsch, H., Linde-Zwirble, W.T., Angus, D.C., Hartman, 
M.E., Milbrandt, E.B., & Kahn, J.M. (2010). ``The epidemiology of 
mechanical ventilation use in the United States.'' Critical Care Med 
38(10): 1947-1953.
---------------------------------------------------------------------------

    The proposed data element, Invasive Mechanical Ventilator, consists 
of a single data element. Data elements that capture invasive 
mechanical ventilation are currently in use in the MDS in SNFs and LCDS 
in LTCHs. For more information on the Invasive Mechanical Ventilator 
data element, we refer readers to the document titled ``Final 
Specifications for IRF QRP Quality Measures and Standardized Patient 
Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    The Invasive Mechanical Ventilator data element was first proposed 
as a standardized patient assessment data element in the FY 2018 IRF 
PPS proposed rule (82 FR 20730 through 20731). In that proposed rule, 
we stated that the proposal was informed by input we received on data 
elements that assess invasive ventilator use and weaning status that 
were tested in the PAC PRD (``Ventilator--Weaning'' and ``Ventilator--
Non-Weaning'') through a call for input published on the CMS Measures 
Management System Blueprint website. Input submitted from August 12 to 
September 12, 2016, noted support for this data element, highlighting 
the importance of this information in supporting care coordination and 
care transitions. We also stated that some commenters had expressed 
concern about the appropriateness for standardization given: The 
prevalence of ventilator weaning across PAC providers; the timing of 
administration; how weaning is defined; and how weaning status in 
particular relates to quality of care. These public comments guided our 
decision to propose a single data element focused on current use of 
invasive mechanical ventilation only, which does not attempt to capture 
weaning status. A summary report for the August 12 to September 12, 
2016 public comment period titled ``SPADE August 2016 Public Comment 
Summary Report'' we received is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In response to our proposal in the FY 2018 IRF PPS proposed rule, 
we received public comments in support of the special services, 
treatments, and interventions data elements in general. Two commenters 
noted their appreciation of the revisions to the Invasive Mechanical 
Ventilator data element in response to comments submitted during a 
public input period held from August 12 to September 12, 2016.
    Subsequent to receiving comments on the FY 2018 IRF PPS rule, the 
Invasive Mechanical Ventilator data element was included in the 
National Beta Test of candidate data elements conducted by our data 
element contractor from November 2017 to August 2018. Results of this 
test found the Invasive Mechanical Ventilator data element to be 
feasible and reliable for use with PAC patients and residents. More 
information about the performance of the Invasive Mechanical Ventilator 
data element in the National Beta Test can be found in the document 
titled ``Final Specifications for IRF QRP Quality Measures and 
Standardized Patient Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In addition, our data element contractor convened a TEP on 
September 17, 2018, for the purpose of soliciting input on the proposed 
standardized patient assessment data element. Although the TEP did not 
specifically discuss the Invasive Mechanical Ventilator data element, 
the TEP supported the assessment of the special services, treatments, 
and interventions included in the National Beta Test with respect to 
both admission and discharge. A summary of the September 17, 2018 TEP 
meeting titled ``SPADE Technical Expert Panel Summary (Third 
Convening)'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.

[[Page 39131]]

    We also held SODFs and small-group discussions with PAC providers 
and other stakeholders in 2018 for the purpose of updating the public 
about our ongoing SPADE development efforts. Finally, on November 27, 
2018, our data element contractor hosted a public meeting of 
stakeholders to present results of the National Beta Test and solicit 
additional comments. General input on the testing and item development 
process and concerns about burden were received from stakeholders 
during this meeting and via email through February 1, 2019. A summary 
of the public input received from the November 27, 2018 stakeholder 
meeting titled ``Input on Standardized Patient Assessment Data Elements 
(SPADEs) Received After November 27, 2018 Stakeholder Meeting'' is 
available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Taking together the importance of assessing for invasive mechanical 
ventilation, stakeholder input, and strong test results, we proposed 
that the Invasive Mechanical Ventilator data element that assesses the 
use of an invasive mechanical ventilator meets the definition of 
standardized patient assessment data with respect to special services, 
treatments, and interventions under section 1899B(b)(1)(B)(iii) of the 
Act and to adopt the Invasive Mechanical Ventilator data element as 
standardized patient assessment data for use in the IRF QRP.
    A commenter submitted the following comment related to the proposed 
rule's discussion of the Invasive Mechanical Ventilator data element.
    Comment: One commenter noted disappointment over seeing that the 
SPADE for invasive mechanical ventilator only assesses whether or not a 
patient is on a mechanical ventilator. The commenter suggested CMS 
consider collecting data to track functional outcomes related to 
progress towards independence in communication and swallowing.
    Response: We have attempted to balance the scope and level of 
detail of the data elements against the potential burden placed on 
patients and providers. We believe that assessing the use of an 
invasive mechanical ventilator will be a useful point of information to 
inform care planning and further assessment, such as related to 
functional outcomes, as the commenter suggests.
    After careful consideration of the public comments we received, we 
are finalizing our proposal to adopt the Invasive Mechanical Ventilator 
data element as standardized patient assessment data beginning with the 
FY 2022 IRF QRP as proposed.
 Intravenous (IV) Medications (Antibiotics, Anticoagulants, 
Vasoactive Medications, Other)
    In the FY 2020 IRF PPS proposed rule (84 FR 17305 through 17306), 
we proposed that the IV Medications (Antibiotics, Anticoagulants, 
Vasoactive Medications, Other) data element meets the definition of 
standardized patient assessment data with respect to special services, 
treatments, and interventions under section 1899B(b)(1)(B)(iii) of the 
Act.
    As described in the FY 2018 IRF PPS proposed rule (82 FR 20731 
through 20732), when we proposed a similar data element related to IV 
medications, IV medications are solutions of a specific medication (for 
example, antibiotics, anticoagulants) administered directly into the 
venous circulation via a syringe or intravenous catheter. IV 
medications are administered via intravenous push, single, 
intermittent, or continuous infusion through a catheter placed into the 
vein. Further, IV medications are more resource intensive to administer 
than oral medications, and signify a higher patient complexity (and 
often higher severity of illness).
    The clinical indications for each of the sub-elements of the IV 
Medications data element (Antibiotics, Anticoagulants, Vasoactive 
Medications, and Other) are very different. IV antibiotics are used for 
severe infections when the bioavailability of the oral form of the 
medication would be inadequate to kill the pathogen or an oral form of 
the medication does not exist. IV anticoagulants refer to anti-clotting 
medications (that is, ``blood thinners''). IV anticoagulants are 
commonly used for hospitalized patients who have deep venous 
thrombosis, pulmonary embolism, or myocardial infarction, as well as 
those undergoing interventional cardiac procedures. Vasoactive 
medications refer to the IV administration of vasoactive drugs, 
including vasopressors, vasodilators, and continuous medication for 
pulmonary edema, which increase or decrease blood pressure or heart 
rate. The indications, risks, and benefits of each of these classes of 
IV medications are distinct, making it important to assess each 
separately in PAC. Knowing whether or not patients and residents are 
receiving IV medication and the type of medication provided by each PAC 
provider will improve quality of care.
    The IV Medications (Antibiotics, Anticoagulants, Vasoactive 
Medications, and Other) data element we proposed consists of a 
principal data element (IV Medications) and four response option sub-
elements: Antibiotics, Anticoagulants, Vasoactive Medications, and 
Other. The Vasoactive Medications sub-element was not proposed in the 
FY 2018 IRF PPS proposed rule (82 FR 20731 through 20732). We added the 
Vasoactive Medications sub-element to our proposal in order to 
harmonize the proposed IV Mediciations element with the data currently 
collected in the LCDS.
    If the assessor indicates that the patient is receiving IV 
medications on the principal IV Medications data element, the assessor 
would then indicate which types of medications (for example, 
Antibiotics, Anticoagulants, Vasoactive Medications, Other). An IV 
Medications data element is currently in use on the MDS in SNFs and 
there is a related data element in OASIS that collects information on 
Intravenous and Infusion Therapies. For more information on the IV 
Medications (Antibiotics, Anticoagulants, Vasoactive Medications, 
Other) data element, we refer readers to the document titled ``Final 
Specifications for IRF QRP Quality Measures and Standardized Patient 
Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    An IV Medications data element was first proposed as standardized 
patient assessment data element in the FY 2018 IRF PPS proposed rule 
(82 FR 20731 through 20732). In that proposed rule, we stated that the 
proposal was informed by input we received on Vasoactive Medications 
through a call for input published on the CMS Measures Management 
System Blueprint website. Input submitted from August 12 to September 
12, 2016 supported this data element with one noting the importance of 
this data element in supporting care transitions. We also stated that 
those commenters had criticized the need for collecting specifically 
Vasoactive Medications, giving feedback that the data element was too 
narrowly focused. In addition, public comment received indicated that 
the clinical significance of vasoactive medications administration 
alone was not high enough in PAC to merit mandated assessment, noting 
that related and more useful information could be captured in an item 
that assessed all IV medication use. A

[[Page 39132]]

summary report for the August 12 to September 12, 2016 public comment 
period titled ``SPADE August 2016 Public Comment Summary Report'' is 
available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In response to our proposal in the FY 2018 IRF PPS proposed rule, 
we received public comments in support of the special services, 
treatments, and interventions data elements in general; no additional 
comments were received that were specific to the IV Medications data 
element.
    Subsequent to receiving comments on the FY 2018 IRF PPS rule, the 
IV Medications data element was included in the National Beta Test of 
candidate data elements conducted by our data element contractor from 
November 2017 to August 2018. Results of this test found the IV 
Medications data element to be feasible and reliable for use with PAC 
patients and residents. More information about the performance of the 
IV Medications data element in the National Beta Test can be found in 
the document titled ``Final Specifications for IRF QRP Quality Measures 
and Standardized Patient Assessment Data Elements,'' available at 
https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In addition, our data element contractor convened a TEP on 
September 17, 2018, for the purpose of soliciting input on the proposed 
standardized patient assessment data elements. Although the TEP did not 
specifically discuss the IV Medications data element, the TEP supported 
the assessment of the special services, treatments, and interventions 
included in the National Beta Test with respect to both admission and 
discharge. A summary of the September 17, 2018 TEP meeting titled 
``SPADE Technical Expert Panel Summary (Third Convening)'' is available 
at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We also held SODFs and small-group discussions with PAC providers 
and other stakeholders in 2018 for the purpose of updating the public 
about our ongoing SPADE development efforts. Finally, on November 27, 
2018, our data element contractor hosted a public meeting of 
stakeholders to present the results of the National Beta Test and 
solicit additional comments. General input on the testing and item 
development process and concerns about burden were received from 
stakeholders during this meeting and via email through February 1, 
2019. A summary of the public input received from the November 27, 2018 
stakeholder meeting titled ``Input on Standardized Patient Assessment 
Data Elements (SPADEs) Received After November 27, 2018 Stakeholder 
Meeting'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Taking together the importance of assessing for IV medications, 
stakeholder input, and strong test results, we proposed that the IV 
Medications (Antibiotics, Anticoagulants, Vasoactive Medications, 
Other) data element with a principal data element and four sub-elements 
meets the definition of standardized patient assessment data with 
respect to special services, treatments, and interventions under 
section 1899B(b)(1)(B)(iii) of the Act and to adopt the IV Medications 
(Antibiotics, Anticoagulants, Vasoactive Medications, Other) data 
element as standardized patient assessment data for use in the IRF QRP.
    Commenters submitted the following comments related to the proposed 
rule's discussion of the IV Medications data elements.
    Comment: One commenter noted that the IV Medications data elements 
seem redundant of the proposed High-Risk Drug Classes: Use and 
Indication data elements.
    Response: We wish to clarify that the IV Medications data element 
collects information on medications received by IV only, with sub-
elements specific to antibiotics, anticoagulants, and vasoactive 
medications only. In contrast, the High Risk Drug Classes: Use and 
Indication data element collects information on medications received by 
any route, only for six specific drug classes, and collects information 
on the presence of an indication. We believe the overlap between these 
SPADEs is minimal, as it would only occur when a medication in a high-
risk drug class is delivered by IV. Additionally, in this case, the 
High-Risk Drug Classes: Use and Indication data element would assess 
the presence of an indication in the patient's medical record, which 
the IV Medications data element does not do.
    Comment: Commenters were concerned about the performance of the IV 
Medications data element in the National Beta Test, noting that its 
reliability was only fair to good and poor for the anticoagulation sub-
element.
    Response: The kappa for the overarching IV Medications data element 
was 0.70 across settings, which falls in the range of ``substantial/
good'' agreement. The IV Medications sub-element that had a ``slight/
poor'' reliability (in the range of 0.00-0.20) was the IV 
Anticoagulants sub-element (kappa = 0.13). The Other IV Medications 
sub-element had ``moderate'' reliability (kappa = 0.46). Consultation 
with assessors suggested that the low kappa for the IV Anticoagulants 
sub-element was likely due to inconsistent interpretation of the coding 
instructions. Having identified the likely source of the relatively 
lower interrater reliability, we are confident that with proper 
training of IRFs on how to report the data elements, the reliability of 
these sub-elements will be improved.
    After careful consideration of the public comments we received, we 
are finalizing our proposal to adopt the IV Medications (Antibiotics, 
Anticoagulants, Vasoactive Medications, Other) data element as 
standardized patient assessment data beginning with the FY 2022 IRF QRP 
as proposed.
 Transfusions
    In the FY 2020 IRF PPS proposed rule (84 FR 17306), we proposed 
that the Transfusions data element meets the definition of standardized 
patient assessment data with respect to special services, treatments, 
and interventions under section 1899B(b)(1)(B)(iii) of the Act.
    As described in the FY 2018 IRF PPS proposed rule (82 FR 20732), 
transfusion refers to introducing blood or blood products into the 
circulatory system of a person. Blood transfusions are based on 
specific protocols, with multiple safety checks and monitoring required 
during and after the infusion in case of adverse events. Coordination 
with the provider's blood bank is necessary, as well as documentation 
by clinical staff to ensure compliance with regulatory requirements. In 
addition, the need for transfusions signifies underlying patient 
complexity that is likely to require care coordination and patient 
monitoring, and impacts planning for transitions of care, as 
transfusions are not performed by all PAC providers.
    The proposed data element consists of the single Transfusions data 
element. A

[[Page 39133]]

data element on transfusion is currently in use in the MDS in SNFs 
(``Transfusions'') and a data element tested in the PAC PRD (``Blood 
Transfusions'') was found feasible for use in each of the four PAC 
settings. For more information on the Transfusions data element, we 
refer readers to the document titled ``Final Specifications for IRF QRP 
Quality Measures and Standardized Patient Assessment Data Elements,'' 
available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    The Transfusions data element was first proposed as a standardized 
patient assessment data element in the FY 2018 IRF PPS proposed rule 
(82 FR 20732). In response to our proposal in the FY 2018 IRF PPS 
proposed rule, we received public comments in support of the special 
services, treatments, and interventions data elements in general; no 
additional comments were received that were specific to the 
Transfusions data element.
    Subsequent to receiving comments on the FY 2018 IRF PPS rule, the 
Transfusions data element was included in the National Beta Test of 
candidate data elements conducted by our data element contractor from 
November 2017 to August 2018. Results of this test found the 
Transfusions data element to be feasible and reliable for use with PAC 
patients and residents. More information about the performance of the 
Transfusions data element in the National Beta Test can be found in the 
document titled ``Final Specifications for IRF QRP Quality Measures and 
Standardized Patient Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In addition, our data element contractor convened a TEP on 
September 17, 2018, for the purpose of soliciting input on the proposed 
standardized patient assessment data elements. Although the TEP did not 
specifically discuss the Transfusions data element, the TEP supported 
the assessment of the special services, treatments, and interventions 
included in the National Beta Test with respect to both admission and 
discharge. A summary of the September 17, 2018 TEP meeting titled 
``SPADE Technical Expert Panel Summary (Third Convening)'' is available 
at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We also held SODFs and small-group discussions with PAC providers 
and other stakeholders in 2018 for the purpose of updating the public 
about our ongoing SPADE development efforts. Finally, on November 27, 
2018, our data element contractor hosted a public meeting of 
stakeholders to present the results of the National Beta Test and 
solicit additional comments. General input on the testing and item 
development process and concerns about burden were received from 
stakeholders during this meeting and via email through February 1, 
2019. A summary of the public input received from the November 27, 2018 
stakeholder meeting titled ``Input on Standardized Patient Assessment 
Data Elements (SPADEs) Received After November 27, 2018 Stakeholder 
Meeting'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Taking together the importance of assessing for transfusions, 
stakeholder input, and strong test results, we proposed that the 
Transfusions data element meets the definition of standardized patient 
assessment data with respect to special services, treatments, and 
interventions under section 1899B(b)(1)(B)(iii) of the Act and to adopt 
the Transfusions data element as standardized patient assessment data 
for use in the IRF QRP.
    Commenters submitted the following comments related to the proposed 
rule's discussion of the Transfusions data element.
    Comment: One commenter applauded CMS for including the Transfusions 
data element, noting that it will provide information on care planning, 
clinical decision making, patient safety, care transitions, and 
resource use in IRFs and will contribute to higher quality and 
coordinated care for patients who rely on these life-saving treatments.
    Response: We thank the commenter for their support. We selected the 
Transfusions data element for proposal as standardized data in part 
because of the attributes that the commenter noted.
    Comment: One commenter was concerned that IRFs will not have the 
resources needed to provide patients with access to blood transfusions 
and requested that CMS consider whether payments to IRFs are adequate 
to cover the cost of this resource intensive, specialized service.
    Response: We wish to clarify that this item is finalized only to 
collect information on the complexity of the patient and resources the 
patient requires. At this time, this item will not be used for any 
payment purposes, and thus we are not able to comment on cost of this 
service. We wish to clarify that this SPADE is not intended to measure 
the ability of an IRF to provide in-house transfusions, only to capture 
the services a given patient may be receiving. Further, for patients 
who require services related to blood transfusions, information 
collected by this data element is a part of common clinical workflow, 
and thus, we believe that burden on resource intensity would not be 
affected by the standardization of this data element.
    After careful consideration of the public comments we received, we 
are finalizing our proposal to adopt the Transfusions data element as 
standardized patient assessment data beginning with the FY 2022 IRF QRP 
as proposed.
 Dialysis (Hemodialysis, Peritoneal Dialysis)
    In the FY 2020 IRF PPS proposed rule (84 FR 17306 through 17307), 
we proposed that the Dialysis (Hemodialysis, Peritoneal dialysis) data 
element meets the definition of standardized patient assessment data 
with respect to special services, treatments, and interventions under 
section 1899B(b)(1)(B)(iii) of the Act.
    As described in the FY 2018 IRF PPS proposed rule (82 FR 20732 
through 20733), dialysis is a treatment primarily used to provide 
replacement for lost kidney function. Both forms of dialysis 
(hemodialysis and peritoneal dialysis) are resource intensive, not only 
during the actual dialysis process but before, during, and following. 
Patients and residents who need and undergo dialysis procedures are at 
high risk for physiologic and hemodynamic instability from fluid shifts 
and electrolyte disturbances, as well as infections that can lead to 
sepsis. Further, patients or residents receiving hemodialysis are often 
transported to a different facility, or at a minimum, to a different 
location in the same facility for treatment. Close monitoring for fluid 
shifts, blood pressure abnormalities, and other adverse effects is 
required prior to, during, and following each dialysis session. Nursing 
staff typically perform peritoneal dialysis at the bedside, and as with 
hemodialysis, close monitoring is required.

[[Page 39134]]

    The proposed data element, Dialysis (Hemodialysis, Peritoneal 
dialysis) consists of the principal Dialysis data element and two 
response option sub-elements: Hemodialysis and Peritoneal dialysis. If 
the assessor indicates that the patient is receiving dialysis on the 
principal Dialysis data element, the assessor would then indicate which 
type (Hemodialysis or Peritoneal dialysis). The principal Dialysis data 
element is currently included on the MDS in SNFs and the LCDS for LTCHs 
and assesses the overall use of dialysis.
    As the result public feedback described below, in the proposed 
rule, we proposed a data element that includes the principal Dialysis 
data element and two sub-elements (Hemodialysis and Peritoneal 
dialysis). For more information on the Dialysis data element, we refer 
readers to the document titled ``Final Specifications for IRF QRP 
Quality Measures and Standardized Patient Assessment Data Elements,'' 
available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    The Dialysis data element was first proposed as standardized 
patient assessment data in the FY 2018 IRF PPS proposed rule (82 FR 
20732 through 20733). In that proposed rule, we stated that the 
proposal was informed by input we received on a singular Hemodialysis 
data element through a call for input published on the CMS Measures 
Management System Blueprint website. Input submitted from August 12 to 
September 12, 2016 supported the assessment of hemodialysis and 
recommended that the data element be expanded to include peritoneal 
dialysis. We also stated that those commenters had supported the 
singular Hemodialysis data element, noting the relevance of this 
information for sharing across the care continuum to facilitate care 
coordination and care transitions, the potential for this data element 
to be used to improve quality, and the feasibility for use in PAC. In 
addition, we received comments that the item would be useful in 
improving patient and resident transitions of care. We also noted that 
several commenters had stated that peritoneal dialysis should be 
included in a standardized data element on dialysis and recommended 
collecting information on peritoneal dialysis in addition to 
hemodialysis. The rationale for including peritoneal dialysis from 
commenters included the fact that patients and residents receiving 
peritoneal dialysis will have different needs at post-acute discharge 
compared to those receiving hemodialysis or not having any dialysis. 
Based on these comments, the Hemodialysis data element was expanded to 
include a principal Dialysis data element and two sub-elements, 
Hemodialysis and Peritoneal dialysis. We proposed the version of the 
Dialysis element that includes two types of dialysis. A summary report 
for the August 12 to September 12, 2016 public comment period titled 
``SPADE August 2016 Public Comment Summary Report'' is available at 
https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In response to our proposal in the FY 2018 IRF PPS proposed rule, 
we received comments in support of the special services, treatments, 
and interventions data elements in general. One commenter noted that 
they appreciated the revisions to the Dialysis data element in response 
to comments submitted during a public input period held from August 12 
to September 12, 2016.
    Subsequent to receiving comments on the FY 2018 IRF PPS rule, the 
Dialysis data element was included in the National Beta Test of 
candidate data elements conducted by our data element contractor from 
November 2017 to August 2018. Results of this test found the Dialysis 
data element to be feasible and reliable for use with PAC patients and 
residents. More information about the performance of the Dialysis data 
element in the National Beta Test can be found in the document titled 
``Final Specifications for IRF QRP Quality Measures and Standardized 
Patient Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In addition, our data element contractor convened a TEP on 
September 17, 2018, for the purpose of soliciting input on the proposed 
standardized patient assessment data elements. Although they did not 
specifically discuss the Dialysis data element, the TEP supported the 
assessment of the special services, treatments, and interventions 
included in the National Beta Test with respect to both admission and 
discharge. A summary of the September 17, 2018 TEP meeting titled 
``SPADE Technical Expert Panel Summary (Third Convening)'' is available 
at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We also held SODFs and small-group discussions with PAC providers 
and other stakeholders in 2018 for the purpose of updating the public 
about our ongoing SPADE development efforts. Finally, on November 27, 
2018, our data element contractor hosted a public meeting of 
stakeholders to present the results of the National Beta Test and 
solicit additional comments. General input on the testing and item 
development process and concerns about burden were received from 
stakeholders during this meeting and via email through February 1, 
2019. A summary of the public input received from the November 27, 2018 
stakeholder meeting titled ``Input on Standardized Patient Assessment 
Data Elements (SPADEs) Received After November 27, 2018 Stakeholder 
Meeting'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Taking together the importance of assessing for dialysis, 
stakeholder input, and strong test results, we proposed that the 
Dialysis (Hemodialysis, Peritoneal dialysis) data element with a 
principal data element and two sub-elements meets the definition of 
standardized patient assessment data with respect to special services, 
treatments, and interventions under section 1899B(b)(1)(B)(iii) of the 
Act and to adopt the Dialysis (Hemodialysis, Peritoneal dialysis) data 
element as standardized patient assessment data for use in the IRF QRP.
    We invited public comment on this proposal. While we received 
support from some commenters on this Special Services, Treatments, and 
Interventions as a whole (section IX.G.2 in this final rule), we did 
not receive any specific comments on the Dialysis (Hemodialysis, 
Peritoneal dialysis) data element.
    After careful consideration of the public comments we received on 
the category of Special Services, Treatments, and Interventions, we are 
finalizing our proposal to adopt the Dialysis (Hemodialysis, Peritoneal 
dialysis) data element as standardized patient assessment data 
beginning with the FY 2022 IRF QRP as proposed.

[[Page 39135]]

 Intravenous (IV) Access (Peripheral IV, Midline, Central Line)
    In the FY 2020 IRF PPS proposed rule (84 FR 17307 through 17308), 
we proposed that the IV Access (Peripheral IV, Midline, Central line) 
data element meets the definition of standardized patient assessment 
data with respect to special services, treatments, and interventions 
under section 1899B(b)(1)(B)(iii) of the Act.
    As described in the FY 2018 IRF PPS proposed rule (82 FR 20733 
through 20734), patients or residents with central lines, including 
those peripherally inserted or who have subcutaneous central line 
``port'' access, always require vigilant nursing care to keep patency 
of the lines and ensure that such invasive lines remain free from any 
potentially life-threatening events such as infection, air embolism, or 
bleeding from an open lumen. Clinically complex patients and residents 
are likely to be receiving medications or nutrition intravenously. The 
sub-elements included in the IV Access data elements distinguish 
between peripheral access and different types of central access. The 
rationale for distinguishing between a peripheral IV and central IV 
access is that central lines confer higher risks associated with life-
threatening events such as pulmonary embolism, infection, and bleeding.
    The proposed data element, IV Access (Peripheral IV, Midline, 
Central line), consists of the principal IV Access data element and 
three response option sub-elements: Peripheral IV, Midline, and Central 
line. The proposed IV Access data element is not currently included on 
any of the PAC assessment instruments. For more information on the IV 
Access data element, we refer readers to the document titled ``Final 
Specifications for IRF QRP Quality Measures and Standardized Patient 
Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    The IV Access data element was first proposed as standardized 
patient assessment data elements in the FY 2018 IRF PPS proposed rule 
(82 FR 20733 through 20734). In that proposed rule, we stated that the 
proposal was informed by input we received on one of the PAC PRD data 
elements, Central Line Management, through a call for input published 
on the CMS Measures Management System Blueprint website. A central line 
is a type of IV access. Input submitted from August 12 to September 12, 
2016 supported the assessment of central line management and 
recommended that the data element be broadened to also include other 
types of IV access. Several commenters noted feasibility and importance 
for facilitating care coordination and care transitions. However, a few 
commenters recommended that the definition of this data element be 
broadened to include peripherally inserted central catheters (``PICC 
lines'') and midline IVs. Based on public comment feedback and in 
consultation with expert input, described below, we created an 
overarching IV Access data element with sub-elements for other types of 
IV access in addition to central lines (that is, peripheral IV and 
midline). This expanded version of IV Access is the data element being 
proposed. A summary report for the August 12 to September 12, 2016 
public comment period titled ``SPADE August 2016 Public Comment Summary 
Report'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In response to our proposal in the FY 2018 IRF PPS proposed rule, 
we received public comments in support of the special services, 
treatments, and interventions data elements in general. One commenter 
noted appreciation of the revisions to the IV Access data element in 
response to comments submitted during a public input period held from 
August 12 to September 12, 2016.
    Subsequent to receiving comments on the FY 2018 IRF PPS rule, the 
IV Access data element was included in the National Beta Test of 
candidate data elements conducted by our data element contractor from 
November 2017 to August 2018. Results of this test found the IV Access 
data element to be feasible and reliable for use with PAC patients and 
residents. More information about the performance of the IV Access data 
element in the National Beta Test can be found in the document titled 
``Final Specifications for IRF QRP Quality Measures and Standardized 
Patient Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In addition, our data element contractor convened a TEP on 
September 17, 2018, for the purpose of soliciting input on the proposed 
standardized patient assessment data elements. Although the TEP did not 
specifically discuss the IV Access data element, the TEP supported the 
assessment of the special services, treatments, and interventions 
included in the National Beta Test with respect to both admission and 
discharge. A summary of the September 17, 2018 TEP meeting titled 
``SPADE Technical Expert Panel Summary (Third Convening)'' is available 
at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We also held SODFs and small-group discussions with PAC providers 
and other stakeholders in 2018 for the purpose of updating the public 
about our ongoing SPADE development efforts. Finally, on November 27, 
2018, our data element contractor hosted a public meeting of 
stakeholders to present results of the National Beta Test and solicit 
additional comments. General input on the testing and item development 
process and concerns about burden were received from stakeholders 
during this meeting and via email through February 1, 2019. A summary 
of the public input received from the November 27, 2018 stakeholder 
meeting titled ``Input on Standardized Patient Assessment Data Elements 
(SPADEs) Received After November 27, 2018 Stakeholder Meeting'' is 
available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Taking together the importance of assessing for IV access, 
stakeholder input, and strong test results, we proposed that the IV 
access (Peripheral IV, Midline, Central line) data element with a 
principal data element and three sub-elements meets the definition of 
standardized patient assessment data with respect to special services, 
treatments, and interventions under section 1899B(b)(1)(B)(iii) of the 
Act and to adopt the IV Access (Peripheral IV, Midline, Central line) 
data element as standardized patient assessment data for use in the IRF 
QRP.
    We invited public comment on this proposal. While we received 
support from some commenters on this Special Services, Treatments, and 
Interventions as a whole (section IX.G.2 in this final rule), we did 
not receive any specific comments on the IV Access (Peripheral IV, 
Midline, Central line) data element.
    After careful consideration of the public comments we received on 
the

[[Page 39136]]

category of Special Services, Treatments, and Interventions, we are 
finalizing our proposal to adopt the IV Access (Peripheral IV, Midline, 
Central line) data element as standardized patient assessment data 
beginning with the FY 2022 IRF QRP as proposed.
 Nutritional Approach: Parenteral/IV Feeding
    In the FY 2020 IRF PPS proposed rule (84 FR 17308 through 17309), 
we proposed that the Parenteral/IV Feeding data element meets the 
definition of standardized patient assessment data with respect to 
special services, treatments, and interventions under section 
1899B(b)(1)(B)(iii) of the Act.
    As described in the FY 2018 IRF PPS proposed rule (82 FR 20734), 
parenteral nutrition/IV feeding refers to a patient or resident being 
fed intravenously using an infusion pump, bypassing the usual process 
of eating and digestion. The need for IV/parenteral feeding indicates a 
clinical complexity that prevents the patient or resident from meeting 
his or her nutritional needs enterally, and is more resource intensive 
than other forms of nutrition, as it often requires monitoring of blood 
chemistries and the maintenance of a central line. Therefore, assessing 
a patient's or resident's need for parenteral feeding is important for 
care planning and resource use. In addition to the risks associated 
with central and peripheral intravenous access, total parenteral 
nutrition is associated with significant risks, such as air embolism 
and sepsis.
    The proposed data element consists of the single Parenteral/IV 
Feeding data element. The proposed Parenteral/IV Feeding data element 
is currently in use in the MDS in SNFs, and equivalent or related data 
elements are in use in the LCDS, IRF-PAI, and OASIS. We proposed to 
rename the existing Tube/Parenteral feeding item in the IRF-PAI to be 
the Parenteral/IV Feeding data element. For more information on the 
Parenteral/IV Feeding data element, we refer readers to the document 
titled ``Final Specifications for IRF QRP Quality Measures and 
Standardized Patient Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    The Parenteral/IV Feeding data element was first proposed as a 
standardized patient assessment data element in the FY 2018 IRF PPS 
proposed rule (82 FR 20734). In that proposed rule, we stated that the 
proposal was informed by input we received on Total Parenteral 
Nutrition (an item with nearly the same meaning as the proposed data 
element, but with the label used in the PAC PRD), through a call for 
input published on the CMS Measures Management System Blueprint 
website. Input submitted from August 12 to September 12, 2016 supported 
this data element, noting its relevance to facilitating care 
coordination and supporting care transitions. After the public comment 
period, the Total Parenteral Nutrition data element was renamed 
Parenteral/IV Feeding, to be consistent with how this data element is 
referred to in the MDS in SNFs. A summary report for the August 12 to 
September 12, 2016 public comment period titled ``SPADE August 2016 
Public Comment Summary Report'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In response to our proposal in the FY 2018 IRF PPS proposed rule, 
we received comments in support of the special services, treatments, 
and interventions data elements in general; no additional comments were 
received that were specific to the Parenteral/IV Feeding data element.
    Subsequent to receiving comments on the FY 2018 IRF PPS rule, the 
Parenteral/IV Feeding data element was included in the National Beta 
Test of candidate data elements conducted by our data element 
contractor from November 2017 to August 2018. Results of this test 
found the Parenteral/IV Feeding data element to be feasible and 
reliable for use with PAC patients and residents. More information 
about the performance of the Parenteral/IV Feeding data element in the 
National Beta Test can be found in the document titled ``Final 
Specifications for IRF QRP Quality Measures and Standardized Patient 
Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In addition, our data element contractor convened a TEP on 
September 17, 2018, for the purpose of soliciting input on the proposed 
standardized patient assessment data elements. Although the TEP did not 
specifically discuss the Parenteral/IV Feeding data element, the TEP 
supported the assessment of the special services, treatments, and 
interventions included in the National Beta Test with respect to both 
admission and discharge. A summary of the September 17, 2018 TEP 
meeting titled ``SPADE Technical Expert Panel Summary (Third 
Convening)'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We also held SODFs and small-group discussions with PAC providers 
and other stakeholders in 2018 for the purpose of updating the public 
about our ongoing SPADE development efforts. Finally, on November 27, 
2018, our data element contractor hosted a public meeting of 
stakeholders to present the results of the National Beta Test and 
solicit additional comments. General input on the testing and item 
development process and concerns about burden were received from 
stakeholders during this meeting and via email through February 1, 
2019. A summary of the public input received from the November 27, 2018 
stakeholder meeting titled ``Input on Standardized Patient Assessment 
Data Elements (SPADEs) Received After November 27, 2018 Stakeholder 
Meeting'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Taking together the importance of assessing for parenteral/IV 
feeding, stakeholder input, and strong test results, we proposed that 
the Parenteral/IV Feeding data element meets the definition of 
standardized patient assessment data with respect to special services, 
treatments, and interventions under section 1899B(b)(1)(B)(iii) of the 
Act and to adopt the Parenteral/IV Feeding data element as standardized 
patient assessment data for use in the IRF QRP.
    A commenter submitted the following comment related to the proposed 
rule's discussion of the Parenteral/IV Feeding data element.
    Comment: One commenter was supportive of collecting this data 
element, but noted that it should not be a substitute for capturing 
information related to swallowing which reflects additional patient 
complexity and resource use.
    Response: We thank the commenter for their support and appreciate 
the concerns raised. We agree that the Parenteral/IV Feeding SPADE 
should not be used as a substitute for an assessment of a patient's 
swallowing

[[Page 39137]]

function. The proposed SPADEs are not intended to replace comprehensive 
clinical evaluation and in no way preclude providers from conducting 
further patient evaluation or assessments in their settings as they 
believe are necessary and useful. We agree that information related to 
swallowing can capture patient complexity. However, we also note that 
Parenteral/IV Feeding data element captures a different construct than 
an evaluation of swallowing. That is, the Parenteral/IV Feeding data 
element captures a patient's need to receive calories and nutrients 
intravenously, while an assessment of swallowing would capture a 
patient's functional ability to safely consume food/liquids orally for 
digestion in their gastrointestinal tract.
    After careful consideration of the public comments we received on 
the category of Special Services, Treatments, and Interventions, we are 
finalizing our proposal to adopt the Parenteral/IV Feeding data element 
as standardized patient assessment data beginning with the FY 2022 IRF 
QRP as proposed.
 Nutritional Approach: Feeding Tube
    In the FY 2020 IRF PPS proposed rule (84 FR 17309 through 17310), 
we proposed that the Feeding Tube data element meets the definition of 
standardized patient assessment data with respect to special services, 
treatments, and interventions under section 1899B(b)(1)(B)(iii) of the 
Act.
    As described in the FY 2018 IRF PPS proposed rule (82 FR 20734 
through 20735), the majority of patients admitted to acute care 
hospitals experience deterioration of their nutritional status during 
their hospital stay, making assessment of nutritional status and method 
of feeding if unable to eat orally very important in PAC. A feeding 
tube can be inserted through the nose or the skin on the abdomen to 
deliver liquid nutrition into the stomach or small intestine. Feeding 
tubes are resource intensive, and therefore, are important to assess 
for care planning and resource use. Patients with severe malnutrition 
are at higher risk for a variety of complications.\113\ In PAC 
settings, there are a variety of reasons that patients and residents 
may not be able to eat orally (including clinical or cognitive status).
---------------------------------------------------------------------------

    \113\ Dempsey, D.T., Mullen, J.L., & Buzby, G.P. (1988). ``The 
link between nutritional status and clinical outcome: can 
nutritional intervention modify it?'' Am J of Clinical Nutrition, 
47(2): 352-356.
---------------------------------------------------------------------------

    The proposed data element consists of the single Feeding Tube data 
element. The Feeding Tube data element is currently included in the MDS 
for SNFs, and in the OASIS for HHAs, where it is labeled Enteral 
Nutrition. A related data element, collected in the IRF-PAI for IRFs 
(Tube/Parenteral Feeding), assesses use of both feeding tubes and 
parenteral nutrition. We proposed to rename the existing Tube/
Parenteral feeding item in the IRF-PAI to the Feeding Tube data 
element. For more information on the Feeding Tube data element, we 
refer readers to the document titled ``Final Specifications for IRF QRP 
Quality Measures and Standardized Patient Assessment Data Elements,'' 
available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    The Feeding Tube data element was first proposed as a standardized 
patient assessment data element in the FY 2018 IRF PPS proposed rule 
(82 FR 20734 through 20735). In that proposed rule, we stated that the 
proposal was informed by input we received on an Enteral Nutrition data 
element (the Enteral Nutrition data item is the same as the data 
element we proposed, but is used in the OASIS under a different name) 
through a call for input published on the CMS Measures Management 
System Blueprint website. Input submitted from August 12 to September 
12, 2016 supported the data element, noting the importance of assessing 
enteral nutrition status for facilitating care coordination and care 
transitions. After the public comment period, the Enteral Nutrition 
data element used in public comment was renamed Feeding Tube, 
indicating the presence of an assistive device. A summary report for 
the August 12 to September 12, 2016 public comment period titled 
``SPADE August 2016 Public Comment Summary Report'' is available at 
https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In response to our proposal in the FY 2018 IRF PPS proposed rule, 
we received public comments in support of the special services, 
treatments, and interventions data elements in general. In addition, a 
commenter recommended that the term ``enteral feeding'' be used instead 
of ``feeding tube''.
    Subsequent to receiving comments on the FY 2018 IRF PPS rule, the 
Feeding Tube data element was included in the National Beta Test of 
candidate data elements conducted by our data element contractor from 
November 2017 to August 2018. Results of this test found the Feeding 
Tube data element to be feasible and reliable for use with PAC patients 
and residents. More information about the performance of the Feeding 
Tube data element in the National Beta Test can be found in the 
document titled ``Final Specifications for IRF QRP Quality Measures and 
Standardized Patient Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In addition, our data element contractor convened a TEP on 
September 17, 2018, for the purpose of soliciting input on the proposed 
standardized patient assessment data elements. Although the TEP did not 
specifically discuss the Feeding Tube data element, the TEP supported 
the assessment of the special services, treatments, and interventions 
included in the National Beta Test with respect to both admission and 
discharge. A summary of the September 17, 2018 TEP meeting titled 
``SPADE Technical Expert Panel Summary (Third Convening)'' is available 
at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We also held SODFs and small-group discussions with PAC providers 
and other stakeholders in 2018 for the purpose of updating the public 
about our ongoing SPADE development efforts. Finally, on November 27, 
2018, our data element contractor hosted a public meeting of 
stakeholders to present the results of the National Beta Test and 
solicit additional comments. General input on the testing and item 
development process and concerns about burden were received from 
stakeholders during this meeting and via email through February 1, 
2019. A summary of the public input received from the November 27, 2018 
stakeholder meeting titled ``Input on Standardized Patient Assessment 
Data Elements (SPADEs) Received After November 27, 2018 Stakeholder 
Meeting'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Taking together the importance of assessing for feeding tubes, 
stakeholder input, and strong test results, we

[[Page 39138]]

proposed that the Feeding Tube data element meets the definition of 
standardized patient assessment data with respect to special services, 
treatments, and interventions under section 1899B(b)(1)(B)(iii) of the 
Act and to adopt the Feeding Tube data element as standardized patient 
assessment data for use in the IRF QRP.
    Commenters submitted the following comments related to the proposed 
rule's discussion of the Feeding Tube data element.
    Comment: One commenter noted that in addition to identifying if the 
patient is on a feeding tube or not, it would be important to assess 
the patient's progression towards oral feeding within this data 
element, as this impacts the tube feeding regimen.
    Response: We agree that progression to oral feeding is important 
for care planning and transfer. At this time, we are finalizing a 
singular Feeding Tube SPADE, which assesses the nutritional approach 
only and does not capture the patient's prognosis with regard to oral 
feeding. We wish to clarify that the proposed SPADEs are not intended 
to replace comprehensive clinical evaluation and in no way preclude 
providers from conducting further patient evaluation or assessments in 
their settings as they believe are necessary and useful. We will take 
this recommendation into consideration in future work on standardized 
data elements.
    Comment: One commenter noted that this data element should 
designate between percutaneous endoscopic gastrostomy (PEG) tube and 
nasogastric (NG) tube because the different routes of access have 
different levels of resource requirements.
    Response: We appreciate the commenter's suggestion, but we have 
decided to maintain the singular Feeding Tube SPADE. We agree that 
different routes of access may have different levels of resource 
requirements. However, we do not believe collecting this level of 
information about nutritional therapies via a SPADE would be 
significantly more clinically useful or supportive of care transitions 
than the singular Feeding Tube SPADE. However, we will take this 
suggestion into consideration in future refinement of the clinical 
SPADEs.
    After careful consideration of the public comments we received on 
the category of Special Services, Treatments, and Interventions, we are 
finalizing our proposal to adopt the Feeding Tube data element as 
standardized patient assessment data beginning with the FY 2022 IRF QRP 
as proposed.
 Nutritional Approach: Mechanically Altered Diet
    In the FY 2020 IRF PPS proposed rule (84 FR 17310 through 17311), 
we proposed that the Mechanically Altered Diet data element meets the 
definition of standardized patient assessment data with respect to 
special services, treatments, and interventions under section 
1899B(b)(1)(B)(iii) of the Act.
    As described in the FY 2018 IRF PPS proposed rule (82 FR 20735 
through 20736), the Mechanically Altered Diet data element refers to 
food that has been altered to make it easier for the patient or 
resident to chew and swallow, and this type of diet is used for 
patients and residents who have difficulty performing these functions. 
Patients with severe malnutrition are at higher risk for a variety of 
complications.\114\
---------------------------------------------------------------------------

    \114\ Dempsey, D.T., Mullen, J.L., & Buzby, G.P. (1988). ``The 
link between nutritional status and clinical outcome: can 
nutritional intervention modify it?'' Am J of Clinical Nutrition, 
47(2): 352-356.
---------------------------------------------------------------------------

    In PAC settings, there are a variety of reasons that patients and 
residents may have impairments related to oral feedings, including 
clinical or cognitive status. The provision of a mechanically altered 
diet may be resource intensive, and can signal difficulties associated 
with swallowing/eating safety, including dysphagia. In other cases, it 
signifies the type of altered food source, such as ground or puree that 
will enable the safe and thorough ingestion of nutritional substances 
and ensure safe and adequate delivery of nourishment to the patient. 
Often, patients and residents on mechanically altered diets also 
require additional nursing support, such as individual feeding or 
direct observation, to ensure the safe consumption of the food product. 
Therefore, assessing whether a patient or resident requires a 
mechanically altered diet is important for care planning and resource 
identification.
    The proposed data element consists of the single Mechanically 
Altered Diet data element. The proposed data element is currently 
included on the MDS for SNFs. A related data element (``Modified food 
consistency/supervision'') is currently included on the IRF-PAI for 
IRFs. Another related data element is included in the OASIS for HHAs 
that collects information about independent eating that requires ``a 
liquid, pureed or ground meat diet.'' We proposed to replace the 
existing Modified food consistency/supervision data element in the IRF-
PAI to the Mechanically Altered Diet data element. For more information 
on the Mechanically Altered Diet data element, we refer readers to the 
document titled ``Final Specifications for IRF QRP Quality Measures and 
Standardized Patient Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    The Mechanically Altered Diet data element was first proposed as a 
standardized patient assessment data element in the FY 2018 IRF PPS 
proposed rule (82 FR 20735 through 20736). In response to our proposal 
in the FY 2018 IRF PPS proposed rule, we received public comments in 
support of the special services, treatments, and interventions data 
elements in general; no additional comments were received that were 
specific to the Mechanically Altered Diet data element.
    Subsequent to receiving comments on the FY 2018 IRF PPS rule, the 
Mechanically Altered Diet data element was included in the National 
Beta Test of candidate data elements conducted by our data element 
contractor from November 2017 to August 2018. Results of this test 
found the Mechanically Altered Diet data element to be feasible and 
reliable for use with PAC patients and residents. More information 
about the performance of the Mechanically Altered Diet data element in 
the National Beta Test can be found in the document titled ``Final 
Specifications for IRF QRP Quality Measures and Standardized Patient 
Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In addition, our data element contractor convened a TEP on 
September 17, 2018, for the purpose of soliciting input on the proposed 
standardized patient assessment data elements. Although the TEP did not 
specifically discuss the Mechanically Altered Diet data element, the 
TEP supported the assessment of the special services, treatments, and 
interventions included in the National Beta Test with respect to both 
admission and discharge. A summary of the September 17, 2018 TEP 
meeting titled ``SPADE Technical Expert Panel Summary (Third 
Convening)'' is available at https://www.cms.gov/Medicare/Quality-
Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-
Initiatives/IMPACT-Act-of-2014/

[[Page 39139]]

IMPACT-Act-Downloads-and-Videos.html.
    We also held SODFs and small-group discussions with PAC providers 
and other stakeholders in 2018 for the purpose of updating the public 
about our ongoing SPADE development efforts. Finally, on November 27, 
2018, our data element contractor hosted a public meeting of 
stakeholders to present the results of the National Beta Test and 
solicit additional comments. General input on the testing and item 
development process and concerns about burden were received from 
stakeholders during this meeting and via email through February 1, 
2019. A summary of the public input received from the November 27, 2018 
stakeholder meeting titled ``Input on Standardized Patient Assessment 
Data Elements (SPADEs) Received After November 27, 2018 Stakeholder 
Meeting'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Taking together the importance of assessing for mechanically 
altered diet, stakeholder input, and strong test results, we proposed 
that the Mechanically Altered Diet data element meets the definition of 
standardized patient assessment data with respect to special services, 
treatments, and interventions under section 1899B(b)(1)(B)(iii) of the 
Act and to adopt the Mechanically Altered Diet data element as 
standardized patient assessment data for use in the IRF QRP.
    Commenters submitted the following comments related to the proposed 
rule's discussion of the Mechanically Altered Diet data element.
    Comment: Commenters were concerned about the performance of this 
data element in the National Beta Test, noting that its reliability was 
only moderate in IRF settings.
    Response: We provided supplementary information with the proposed 
rule on the reliability of the SPADEs, described by the kappa statistic 
and by the ``percent agreement'' between assessor, another measure of 
reliability that is in some cases more accurate than the kappa 
statistic, depending on the underlying distribution. (The document 
titled ``Proposed Specifications for IRF QRP Quality Measures and 
Standardized Patient Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html). In this document, we stated that 
the interrater reliability for Mechanically Altered Diet data element, 
as measured by kappa, was ``substantial/good'' across the four PAC 
provider types (LTCH, SNF, HHA, and IRF) in which it was tested (kappa 
= 0.65) and ``moderate'' in the IRF setting (kappa = 0.53). However, 
percent agreement for the data element was 93 percent across all PAC 
settings in the National Beta Test (that is, HHA, IRF, LTCH, and SNF) 
and 89 percent in the IRF setting. That is, when assessing if patients 
required a mechanically altered diet, the facility staff and the 
external research nurse agreed 89 percent of the time for IRF patients.
    Comment: One commenter was concerned that the Mechanically Altered 
Diet data element does not capture clinical complexity and does not 
provide any insight into resource allocation because it only measures 
whether the patient needs a mechanically altered diet and not, for 
example, the extent of help a patient needs in consuming his or her 
meal.
    Response: We believe that assessing patients' needs for 
mechanically altered diets captures one piece of information about 
resource intensity. That is, patients with this special nutritional 
requirement may require additional nutritional planning services, 
special meals, and staff to ensure that meals are prepared and served 
in the way the patient needs. Additional factors that would affect 
resource allocation, such as those noted by the commenter, are not 
captured by this data element. We have attempted to balance the scope 
and level of detail of the data elements against the potential burden 
placed on providers who must complete the assessment. We will take this 
suggestion into consideration in future refinement of the clinical 
SPADEs.
    After careful consideration of the public comments we received, we 
are finalizing our proposal to adopt the Mechanically Altered Diet data 
element as standardized patient assessment data beginning with the FY 
2022 IRF QRP as proposed.
 Nutritional Approach: Therapeutic Diet
    In the FY 2020 IRF PPS proposed rule (84 FR 17311 through 17312), 
we proposed that the Therapeutic Diet data element meets the definition 
of standardized patient assessment data with respect to special 
services, treatments, and interventions under section 
1899B(b)(1)(B)(iii) of the Act.
    As described in the FY 2018 IRF PPS proposed rule (82 FR 20736), a 
therapeutic diet refers to meals planned to increase, decrease, or 
eliminate specific foods or nutrients in a patient's or resident's 
diet, such as a low-salt diet, for the purpose of treating a medical 
condition. The use of therapeutic diets among patients and residents in 
PAC provides insight on the clinical complexity of these patients and 
residents and their multiple comorbidities. Therapeutic diets are less 
resource intensive from the bedside nursing perspective, but do signify 
one or more underlying clinical conditions that preclude the patient 
from eating a regular diet. The communication among PAC providers about 
whether a patient is receiving a particular therapeutic diet is 
critical to ensure safe transitions of care.
    The proposed data element consists of the single Therapeutic Diet 
data element. This data element is currently in use in the MDS in SNFs. 
For more information on the Therapeutic Diet data element, we refer 
readers to the document titled ``Final Specifications for IRF QRP 
Quality Measures and Standardized Patient Assessment Data Elements,'' 
available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    The Therapeutic Diet data element was first proposed as a 
standardized patient assessment data element in the FY 2018 IRF PPS 
proposed rule (82 FR 20736). In response to our proposal in the FY 2018 
IRF PPS proposed rule, we received public comments in support of the 
special services, treatments, and interventions data elements in 
general. One commenter recommended that the definition of Therapeutic 
Diet be aligned with the Academy of Nutrition and Dietetics' definition 
and that ``medically altered diet'' be added to the list of nutritional 
approaches.
    Subsequent to receiving comments on the FY 2018 IRF PPS rule, the 
Therapeutic Diet data element was included in the National Beta Test of 
candidate data elements conducted by our data element contractor from 
November 2017 to August 2018. Results of this test found the 
Therapeutic Diet data element to be feasible and reliable for use with 
PAC patients and residents. More information about the performance of 
the Therapeutic Diet data element in the National Beta Test can be 
found in the document titled ``Final Specifications for IRF QRP Quality 
Measures and Standardized Patient Assessment Data Elements,'' available 
at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-

[[Page 39140]]

Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-
of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In addition, our data element contractor convened a TEP on 
September 17, 2018, for the purpose of soliciting input on the proposed 
standardized patient assessment data elements. Although the TEP did not 
specifically discuss the Therapeutic Diet data element, the TEP 
supported the assessment of the special services, treatments, and 
interventions included in the National Beta Test with respect to both 
admission and discharge. A summary of the September 17, 2018 TEP 
meeting titled ``SPADE Technical Expert Panel Summary (Third 
Convening)'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We also held SODFs and small-group discussions with PAC providers 
and other stakeholders in 2018 for the purpose of updating the public 
about our ongoing SPADE development efforts. Finally, on November 27, 
2018, our data element contractor hosted a public meeting of 
stakeholders to present the results of the National Beta Test and 
solicit additional comments. General input on the testing and item 
development process and concerns about burden were received from 
stakeholders during this meeting and via email through February 1, 
2019. A summary of the public input received from the November 27, 2018 
stakeholder meeting titled ``Input on Standardized Patient Assessment 
Data Elements (SPADEs) Received After November 27, 2018 Stakeholder 
Meeting'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Taking together the importance of assessing for therapeutic diet, 
stakeholder input, and strong test results, we proposed that the 
Therapeutic Diet data element meets the definition of standardized 
patient assessment data with respect to special services, treatments, 
and interventions under section 1899B(b)(1)(B)(iii) of the Act and to 
adopt the Therapeutic Diet data element as standardized patient 
assessment data for use in the IRF QRP.
    We invited public comment on this proposal. While we received 
support from some commenters on Special Services, Treatments, and 
Interventions as a whole (section IX.G.2 in this final rule), we did 
not receive any specific comments on the Therapeutic Diet data element.
    After careful consideration of the public comments we received on 
the category of Special Services, Treatments, and Interventions, we are 
finalizing our proposal to adopt the Therapeutic Diet data element as 
standardized patient assessment data beginning with the FY 2022 IRF QRP 
as proposed.
 High-Risk Drug Classes: Use and Indication
    In the FY 2020 IRF PPS proposed rule (84 FR 17312 through 17314), 
we proposed that the High-Risk Drug Classes: Use and Indication data 
element meets the definition of standardized patient assessment data 
with respect to special services, treatments, and interventions under 
section 1899B(b)(1)(B)(iii) of the Act.
    Most patients and residents receiving PAC services depend on short- 
and long-term medications to manage their medical conditions. However, 
as a treatment, medications are not without risk; medications are, in 
fact, a leading cause of adverse events. A study by the U.S. Department 
of Health and Human Services found that 31 percent of adverse events 
that occurred in 2008 among hospitalized Medicare beneficiaries were 
related to medication.\115\ Moreover, changes in a patient's condition, 
medications, and transitions between care settings put patients at risk 
of medication errors and adverse drug events (ADEs). ADEs may be caused 
by medication errors such as drug omissions, errors in dosage, and 
errors in dosing frequency.\116\
---------------------------------------------------------------------------

    \115\ U.S. Department of Health and Human Services. Office of 
Inspector General. Daniel R. Levinson. Adverse Events in Hospitals: 
National Incidence Among Medicare Beneficiaries. OEI-06-09-00090. 
November 2010.
    \116\ Boockvar KS, Liu S, Goldstein N, Nebeker J, Siu A, Fried 
T. Prescribing discrepancies likely to cause adverse drug events 
after patient transfer. Qual Saf Health Care. 2009;18(1):32-6.
---------------------------------------------------------------------------

    ADEs are known to occur across different types of healthcare 
settings. For example, the incidence of ADEs in the outpatient setting 
has been estimated at 1.15 ADEs per 100 person-months,\117\ while the 
rate of ADEs in the long-term care setting is approximately 9.80 ADEs 
per 100 resident-months.\118\ In the hospital setting, the incidence 
has been estimated at 15 ADEs per 100 admissions.\119\ In addition, 
approximately half of all hospital-related medication errors and 20 
percent of ADEs occur during transitions within, admission to, transfer 
to, or discharge from a hospital.\120\ \121\ \122\ ADEs are more common 
among older adults, who make up most patients receiving PAC services. 
The rate of emergency department visits for ADEs is three times higher 
among adults 65 years of age and older compared to that among those 
younger than age 65.\123\
---------------------------------------------------------------------------

    \117\ Gandhi TK, Seger AC, Overhage JM, et al. Outpatient 
adverse drug events identified by screening electronic health 
records. J Patient Saf 2010;6:91-6.doi:10.1097/PTS.0b013e3181dcae06.
    \118\ Gurwitz JH, Field TS, Judge J, Rochon P, Harrold LR, 
Cadoret C, et al. The incidence of adverse drug events in two large 
academic long-term care facilities. Am J Med. 2005; 118(3):2518. Epub 2005/03/05. https://doi.org/10.1016/j.amjmed.2004.09.018 PMID: 15745723.
    \119\ Hug BL, Witkowski DJ, Sox CM, Keohane CA, Seger DL, Yoon 
C, Matheny ME, Bates DW. Occurrence of adverse, often preventable, 
events in community hospitals involving nephrotoxic drugs or those 
excreted by the kidney. Kidney Int. 2009; 76:1192-1198. [PubMed: 
19759525].
    \120\ Barnsteiner JH. Medication reconciliation: transfer of 
medication information across settings-keeping it free from error. J 
Infus Nurs. 2005;28(2 Suppl):31-36.
    \121\ Rozich J, Roger, R. Medication safety: one organization's 
approach to the challenge. Journal of Clinical Outcomes Management. 
2001(8):27-34.
    \122\ Gleason KM, Groszek JM, Sullivan C, Rooney D, Barnard C, 
Noskin GA. Reconciliation of discrepancies in medication histories 
and admission orders of newly hospitalized patients. Am J Health 
Syst Pharm. 2004;61(16):1689-1695.
    \123\ Shehab N, Lovegrove MC, Geller AI, Rose KO, Weidle NJ, 
Budnitz DS. US emergency department visits for outpatient adverse 
drug events, 2013-2014. JAMA. doi: 10.1001/jama.2016.16201.
---------------------------------------------------------------------------

    Understanding the types of medication a patient is taking, and the 
reason for its use, are key facets of a patient's treatment with 
respect to medication. Some classes of drugs are associated with more 
risk than others.\124\ We proposed one High-Risk Drug Class data 
element with six sub-elements. The response options that correspond to 
the six medication classes are: Anticoagulants, antiplatelets, 
hypoglycemics (including insulin), opioids, antipsychotics, and 
antibiotics. These drug classes are high-risk due to the adverse 
effects that may result from use. In particular, bleeding risk is 
associated with anticoagulants and antiplatelets; \125\ \126\ fluid 
retention, heart failure, and lactic acidosis are associated with 
hypoglycemics; \127\

[[Page 39141]]

misuse is associated with opioids; \128\ fractures and strokes are 
associated with antipsychotics; \129\ \130\ and various adverse events, 
such as central nervous systems effects and gastrointestinal 
intolerance, are associated with antimicrobials,\131\ the larger 
category of medications that include antibiotics. Moreover, some 
medications in five of the six drug classes included in this data 
element are included in the 2019 Updated Beers Criteria[supreg] list as 
potentially inappropriate medications for use in older adults.\132\ 
Finally, although a complete medication list should record several 
important attributes of each medication (for example, dosage, route, 
stop date), recording an indication for the drug is of crucial 
importance.\133\
---------------------------------------------------------------------------

    \124\ Ibid.
    \125\ Shoeb M, Fang MC. Assessing bleeding risk in patients 
taking anticoagulants. J Thromb Thrombolysis. 2013;35(3):312-319. 
doi: 10.1007/s11239-013-0899-7.
    \126\ Melkonian M, Jarzebowski W, Pautas E. Bleeding risk of 
antiplatelet drugs compared with oral anticoagulants in older 
patients with atrial fibrillation: a systematic review and 
meta[hyphen]analysis. J Thromb Haemost. 2017;15:1500-1510. DOI: 
10.1111/jth.13697.
    \127\ Hamnvik OP, McMahon GT. Balancing Risk and Benefit with 
Oral Hypoglycemic Drugs. The Mount Sinai journal of medicine, New 
York. 2009; 76:234-243.
    \128\ Naples JG, Gellad WF, Hanlon JT. The Role of Opioid 
Analgesics in Geriatric Pain Management. Clin Geriatr Med. 
2016;32(4):725-735.
    \129\ Rigler SK, Shireman TI, Cook-Wiens GJ, Ellerbeck EF, 
Whittle JC, Mehr DR, Mahnken JD. Fracture risk in nursing home 
residents initiating antipsychotic medications. J Am Geriatr Soc. 
2013; 61(5):715-722. [PubMed: 23590366].
    \130\ Wang S, Linkletter C, Dore D et al. Age, antipsychotics, 
and the risk of ischemic stroke in the Veterans Health 
Administration. Stroke 2012;43:28-31. doi:10.1161/
STROKEAHA.111.617191.
    \131\ Faulkner CM, Cox HL, Williamson JC. Unique aspects of 
antimicrobial use in older adults. Clin Infect Dis. 2005;40(7):997-
1004.
    \132\ American Geriatrics Society 2019 Beers Criteria Update 
Expert Panel. American Geriatrics Society 2019 Updated Beers 
Criteria for Potentially Inappropriate Medication Use in Older 
Adults. J Am Geriatr Soc 2019; 00:1-21.
    \133\ Li Y, Salmasian H, Harpaz R, Chase H, Friedman C. 
Determining the reasons for medication prescriptions in the EHR 
using knowledge and natural language processing. AMIA Annu Symp 
Proc. 2011; 2011:768-76.
---------------------------------------------------------------------------

    The High-Risk Drug Classes: Use and Indication data element 
requires an assessor to record whether or not a patient is taking any 
medications within the six drug classes. The six response options for 
this data element are high-risk drug classes with particular relevance 
to PAC patients and residents, as identified by our data element 
contractor. The six data element response options are Anticoagulants, 
Antiplatelets, Hypoglycemics, Opioids, Antipsychotics, and Antibiotics. 
For each drug class, the assessor is required to indicate if the 
patient is taking any medications within the class, and, for drug 
classes in which medications were being taken, whether indications for 
all drugs in the class are noted in the medical record. For example, 
for the response option Anticoagulants, if the assessor indicates that 
the patient has received anticoagulant medication, the assessor would 
then indicate if an indication is recorded in the medication record for 
the anticoagulant(s).
    The High-Risk Drug Classes: Use and Indication data element that is 
being proposed as a SPADE was developed as part of a larger set of data 
elements to assess medication reconciliation, the process of obtaining 
a patient's multiple medication lists and reconciling any 
discrepancies. For more information on the High-Risk Drug Classes: Use 
and Indication data element, we refer readers to the document titled 
``Final Specifications for IRF QRP Quality Measures and Standardized 
Patient Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We sought public input on the relevance of conducting assessments 
on medication reconciliation and specifically on the proposed High-Risk 
Drug Classes: Use and Indication data element. Our data element 
contractor presented data elements related to medication reconciliation 
to the TEP convened on April 6 and 7, 2016. The TEP supported a focus 
on high-risk drugs, because of higher potential for harm to patients 
and residents, and were in favor of a data element to capture whether 
or not indications for medications were recorded in the medical record. 
A summary of the April 6 and 7, 2016 TEP meeting titled ``SPADE 
Technical Expert Panel Summary (First Convening)'' is available at 
https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html. Medication reconciliation data 
elements were also discussed at a second TEP meeting on January 5 and 
6, 2017, convened by our data element contractor. At this meeting, the 
TEP agreed about the importance of evaluating the medication 
reconciliation process, but disagreed about how this could be 
accomplished through standardized assessment. The TEP also disagreed 
about the usability and appropriateness of using the Beers Criteria to 
identify high-risk medications.\134\ A summary of the January 5 and 6, 
2017 TEP meeting titled ``SPADE Technical Expert Panel Summary (Second 
Convening)'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
---------------------------------------------------------------------------

    \134\ American Geriatrics Society 2015 Beers Criteria Update 
Expert Panel. American Geriatrics Society. Updated Beers Criteria 
for Potentially Inappropriate Medication Use in Older Adults. J Am 
Geriatr Soc 2015; 63:2227-2246.
---------------------------------------------------------------------------

    We also solicited public input on data elements related to 
medication reconciliation during a public input period from April 26 to 
June 26, 2017. Several commenters noted support for the medication 
reconciliation data elements that were put on display, noting the 
importance of medication reconciliation in preventing medication errors 
and stated that the items seemed feasible and clinically useful. A few 
commenters were critical of the choice of 10 drug classes posted during 
that comment period, stating that ADEs are not limited to high-risk 
drugs, and raised issues related to training assessors to correctly 
complete a valid assessment of medication reconciliation. A summary 
report for the April 26 to June 26, 2017 public comment period titled 
``SPADE May-June 2017 Public Comment Summary Report'' is available at 
https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    The High-Risk Drug Classes: Use and Indication data element was 
included in the National Beta Test of candidate data elements conducted 
by our data element contractor from November 2017 to August 2018. 
Results of this test found the High-Risk Drug Classes: Use and 
Indication data element to be feasible and reliable for use with PAC 
patients and residents. More information about the performance of the 
High-Risk Drug Classes: Use and Indication data element in the National 
Beta Test can be found in the document titled ``Final Specifications 
for IRF QRP Quality Measures and Standardized Patient Assessment Data 
Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In addition, our data element contractor convened a TEP on 
September 17, 2018, for the purpose of soliciting input on the proposed 
standardized patient assessment data elements. The TEP acknowledged the 
challenges of assessing medication safety, but were supportive of some 
of the data elements focused on medication reconciliation that were 
tested in the National Beta Test. The TEP was especially supportive of 
the focus on the six high-risk drug classes and using these classes to 
assess

[[Page 39142]]

whether the indication for a drug is recorded. A summary of the 
September 17, 2018 TEP meeting titled ``SPADE Technical Expert Panel 
Summary (Third Convening)'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We also held SODFs and small-group discussions with PAC providers 
and other stakeholders in 2018 for the purpose of updating the public 
about our ongoing SPADE development efforts. These activities provided 
updates on the field-testing work and solicited feedback on data 
elements considered for standardization, including the High-Risk Drug 
Classes: Use and Indication data element. One stakeholder group was 
critical of the six drug classes included as response options in the 
High-Risk Drug Classes: Use and Indication data element, noting that 
potentially risky medications (for example, muscle relaxants) are not 
included in this list; that there may be important differences between 
drugs within classes (for example, more recent versus older style 
antidepressants); and that drug allergy information is not captured. 
Finally, on November 27, 2018, our data element contractor hosted a 
public meeting of stakeholders to present the results of the National 
Beta Test and solicit additional comments. General input on the testing 
and item development process and concerns about burden were received 
from stakeholders during this meeting and via email through February 1, 
2019. Additionally, one commenter questioned whether the time to 
complete the High-Risk Drug Classes: Use and Indication data element 
would differ across settings. A summary of the public input received 
from the November 27, 2018 stakeholder meeting titled ``Input on 
Standardized Patient Assessment Data Elements (SPADEs) Received After 
November 27, 2018 Stakeholder Meeting'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Taking together the importance of assessing high-risk drugs and for 
whether or not indications are noted for high-risk drugs, stakeholder 
input, and strong test results, we proposed that the High-Risk Drug 
Classes: Use and Indication data element meets the definition of 
standardized patient assessment data with respect to special services, 
treatments, and interventions under section 1899B(b)(1)(B)(iii) of the 
Act and to adopt the High-Risk Drug Classes: Use and Indication data 
element as standardized patient assessment data for use in the IRF QRP.
    Commenters submitted the following comments related to the proposed 
rule's discussion of the High-Risk Drug Classes: Use and Indication 
data element.
    Comment: Some commenters noted that the proposed High-Risk Drug 
Classes: Use and Indication data elements are redundant of the existing 
standards in the Hospital Conditions of Participation (CoPs) and that 
requiring the collection of these data elements would be duplicative, 
unnecessary, and at odds with the Meaningful Measures framework.
    Response: We disagree that assessing the extent to which 
medications from certain drug classes are being taken and the extent to 
which indications are recorded for medications in these classes is 
redundant with the existing CoPs. The CoPs provide guidance on clinical 
practice, while the proposed SPADEs attempt to collect information 
about individual patients in order to understand clinical acuity and to 
populate a core set of information that can be exchanged with the 
patient across care transitions.
    Comment: Commenters noted that because adverse drug events (ADEs) 
are not limited to high-risk drugs, this data element has limited 
utility.
    Response: We acknowledge that not all ADEs are associated with 
``high-risk'' drugs, and we also note that medications in the named 
drug classes are mostly used in a safe manner. Prescribed high-risk 
medications are defined as a ``proximate factor'' to preventable ADEs 
by the Joint Commission.\135\ However, the Joint Commission's 
conceptual model of preventable ADEs also includes provider, patient, 
health care system, organization, and technical factors, all of which 
present many opportunities for disrupting preventable ADEs. We have 
decided to focus on a selection of drug classes that are commonly used 
by older adults and are related to ADEs which are clinically 
significant, preventable, and measurable. Anticoagulants, antibiotics, 
and diabetic agents have been implicated in an estimated 46.9 percent 
(95 percent CI, 44.2 percent-49.7 percent) of emergency department 
visits for adverse drug events.\136\ Among older adults (aged >=65 
years), three drug classes (anticoagulants, diabetic agents, and opioid 
analgesics) have been implicated in an estimated 59.9 percent (95 
percent CI, 56.8 percent-62.9 percent) of ED visits for adverse drug 
events.\137\ Further, antipsychotic medications have been identified as 
a drug class for which there is a need for increased outreach and 
educational efforts to reduce use among older adults.
---------------------------------------------------------------------------

    \135\ Chang A, Schyve PM, Croteau RJ, O'Leary DS, Loeb JM. The 
JCAHO patient safety event taxonomy: A standardized terminology and 
classification schema for near misses and adverse events. Int J Qual 
Health Care. 2005;17(2):95-105.
    \136\ Shehab N, Lovegrove MC, Geller AI, Rose KO, Weidle NJ, 
Budnitz DS. US emergency department visits for outpatient adverse 
drug events, 2013-2014. JAMA 2016;316(2):2115-2125.
    \137\ Shehab N, Lovegrove MC, Geller AI, Rose KO, Weidle NJ, 
Budnitz DS. US emergency department visits for outpatient adverse 
drug events, 2013-2014. JAMA 2016;316(2):2115-2125.
---------------------------------------------------------------------------

    Comment: One commenter was concerned with the addition of the High-
Risk Drug Classes: Use and Indication data elements, noting that 
providers should be granted clinical judgment to effectively treat 
patients without CMS monitoring of medications used for treatment.
    Response: The proposed SPADEs attempt to collect information about 
individual patients to understand clinical acuity and to populate a 
core set of information that can be exchanged with the patient across 
care transitions. The intent of these data elements is not to monitor 
prescribing practices, but rather to assess the extent to which 
indications are noted for medications in certain drug classes.
    Comment: A few commenters noted that the High-Risk Drug Class: Use 
and Indication data elements seemed redundant with other SPADEs (that 
is, IV Medications) and measures (that is, Provision of Current 
Reconciled Medication List to Subsequent Provider at Discharge), or 
duplicative of existing standards in the Hospital CoPs related to 
procurement, preparation, and administration of drugs, which creates 
unnecessary burden.
    Response: The High-Risk Drugs: Use and Indications data element 
captures unique information compared to the other SPADEs and measures 
to which the commenters referred. With regard to the reference to the 
measure Provision of Current Reconciled Medication List to Subsequent 
Provider at Discharge, we wish to clarify that the High-Risk Drug 
Classes: Use and Indication data elements capture medications taken by 
any route and focuses on a select set of drug classes, not the act of 
communicating a complete medication list. To the extent that the 
activities captured by the High-Risk Drugs: Use and Indications data 
element are already being performed by providers as part of

[[Page 39143]]

the Hospital CoPs, we believe that reporting of this data elements 
should be easily integrated into existing workflow.
    Comment: One commenter noted that medication indications are 
typically documented in narrative notes by the medical staff and would 
therefore be difficult to collect.
    Response: We maintain that collecting information on the presence 
of indications in the medical record is clinically important 
information that can inform care planning and support care transitions. 
It is the responsibility of IRF providers to record patient data in a 
way that is useful and appropriate to meet clinical and administrative 
needs. It is possible that the adoption of this SPADE and related 
reporting requirement will promote a more efficient method for 
documenting the clinical indication for each medication.
    After careful consideration of the public comments we received, we 
are finalizing our proposal to adopt the High-Risk Drug Classes: Use 
and Indication data element as standardized patient assessment data 
beginning with the FY 2022 IRF QRP as proposed.
3. Medical Condition and Comorbidity Data
    Assessing medical conditions and comorbidities is critically 
important for care planning and safety for patients and residents 
receiving PAC services, and the standardized assessment of selected 
medical conditions and comorbidities across PAC providers is important 
for managing care transitions and understanding medical complexity.
    In this section we discuss our proposals for data elements related 
to the medical condition of pain as standardized patient assessment 
data. Appropriate pain management begins with a standardized 
assessment, and thereafter establishing and implementing an overall 
plan of care that is person-centered, multi-modal, and includes the 
treatment team and the patient. Assessing and documenting the effect of 
pain on sleep, participation in therapy, and other activities may 
provide information on undiagnosed conditions and comorbidities and the 
level of care required, and do so more objectively than subjective 
numerical scores. With that, we assess that taken separately and 
together, these proposed data elements are essential for care planning, 
consistency across transitions of care, and identifying medical 
complexities including undiagnosed conditions. We also conclude that it 
is the standard of care to always consider the risks and benefits 
associated with a personalized care plan, including the risks of any 
pharmacological therapy, especially opioids.\138\ We also conclude that 
in addition to assessing and appropriately treating pain through the 
optimum mix of pharmacologic, non-pharmacologic, and alternative 
therapies, while being cognizant of current prescribing guidelines, 
clinicians in partnership with patients are best able to mitigate 
factors that contribute to the current opioid crisis.\139\ \140\ \141\
---------------------------------------------------------------------------

    \138\ Department of Health and Human Services: Pain Management 
Best Practices Inter-Agency Task Force. Draft Report on Pain 
Management Best Practices: Updates, Gaps, Inconsistencies, and 
Recommendations. Accessed April 1, 2019. https://www.hhs.gov/sites/default/files/final-pmtf-draft-report-on-pain-management%20-best-practices-2018-12-12-html-ready-clean.pdf.
    \139\ Department of Health and Human Services: Pain Management 
Best Practices Inter-Agency Task Force. Draft Report on Pain 
Management Best Practices: Updates, Gaps, Inconsistencies, and 
Recommendations. Accessed April 1, 2019. https://www.hhs.gov/sites/default/files/final-pmtf-draft-report-on-pain-management%20-best-practices-2018-12-12-html-ready-clean.pdf.
    \140\ Fishman SM, Carr DB, Hogans B, et al. Scope and Nature of 
Pain- and Analgesia-Related Content of the United States Medical 
Licensing Examination (USMLE). Pain Med Malden Mass. 2018;19(3):449-
459. doi:10.1093/pm/pnx336.
    \141\ Fishman SM, Young HM, Lucas Arwood E, et al. Core 
competencies for pain management: results of an interprofessional 
consensus summit. Pain Med Malden Mass. 2013;14(7):971-981. 
doi:10.1111/pme.12107.
---------------------------------------------------------------------------

    In alignment with our Meaningful Measures Initiative, accurate 
assessment of medical conditions and comorbidities of patients and 
residents in PAC is expected to make care safer by reducing harm caused 
in the delivery of care; promote effective prevention and treatment of 
chronic disease; strengthen person and family engagement as partners in 
their care; and promote effective communication and coordination of 
care. The SPADEs will enable or support: Clinical decision-making and 
early clinical intervention; person-centered, high quality care 
through: facilitating better care continuity and coordination; better 
data exchange and interoperability between settings; and longitudinal 
outcome analysis. Therefore, reliable data elements assessing medical 
conditions and comorbidities are needed to initiate a management 
program that can optimize a patient's or resident's prognosis and 
reduce the possibility of adverse events.
    We sought comment that applies specifically to the standardized 
patient assessment data for the category of medical conditions and co-
morbidities. We did not receive any comments on the category of medical 
conditions and co-morbidities.
    Final decisions on the SPADEs are given below, following more 
detailed comments on each SPADE proposal.
 Pain Interference (Pain Effect on Sleep, Pain Interference 
With Therapy Activities, and Pain Interference With Day-to-Day 
Activities)
    In acknowledgement of the opioid crisis, we specifically sought 
comment on whether or not we should add these pain items in light of 
those concerns. Commenters were asked to address to what extent the 
collection of the SPADEs described below through patient queries might 
encourage providers to prescribe opioids.
    In the FY 2020 IRF PPS proposed rule (84 FR 17314 through 17316), 
we proposed that a set of three data elements on the topic of Pain 
Interference (Pain Effect on Sleep, Pain Interference with Therapy 
Activities, and Pain Interference with Day-to-Day Activities) meet the 
definition of standardized patient assessment data with respect to 
medical condition and comorbidity data under section 1899B(b)(1)(B)(iv) 
of the Act.
    The practice of pain management began to undergo significant 
changes in the 1990s because the inadequate, non-standardized, non-
evidence-based assessment and treatment of pain became a public health 
issue.\142\ In pain management, a critical part of providing 
comprehensive care is performance of a thorough initial evaluation, 
including assessment of both the medical and any biopsychosocial 
factors causing or contributing to the pain, with a treatment plan to 
address the causes of pain and to manage pain that persists over 
time.\143\ Quality pain management, based on current guidelines and 
evidence-based practices, can minimize unnecessary opioid prescribing 
both by offering alternatives or supplemental treatment to opioids and 
by clearly stating when they may be appropriate, and how to utilize 
risk-benefit analysis for opioid and non-opioid treatment 
modalities.\144\
---------------------------------------------------------------------------

    \142\ Institute of Medicine. Relieving Pain in America: A 
Blueprint for Transforming Prevention, Care, Education, and 
Research. Washington (DC): National Academies Press (US); 2011. 
http://www.ncbi.nlm.nih.gov/books/NBK91497/.
    \143\ Department of Health and Human Services: Pain Management 
Best Practices Inter-Agency Task Force. Draft Report on Pain 
Management Best Practices: Updates, Gaps, Inconsistencies, and 
Recommendations. Accessed April 1, 2019. https://www.hhs.gov/sites/default/files/final-pmtf-draft-report-on-pain-management%20-best-practices-2018-12-12-html-ready-clean.pdf.
    \144\ National Academies. Pain Management and the Opioid 
Epidemic: Balancing Societal and Individual Benefits and Risks of 
Prescription Opioid Use. Washington DC: National Academies of 
Sciences, Engineering, and Medicine.; 2017.

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

[[Page 39144]]

    Pain is not a surprising symptom in PAC patients and residents, 
where healing, recovery, and rehabilitation often require regaining 
mobility and other functions after an acute event. Standardized 
assessment of pain that interferes with function is an important first 
step towards appropriate pain management in PAC settings. The National 
Pain Strategy called for refined assessment items on the topic of pain, 
and describes the need for these improved measures to be implemented in 
PAC assessments.\145\ Further, the focus on pain interference, as 
opposed to pain intensity or pain frequency, was supported by the TEP 
convened by our data element contractor as an appropriate and 
actionable metric for assessing pain. A summary of the September 17, 
2018 TEP meeting titled ``SPADE Technical Expert Panel Summary (Third 
Convening)'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
---------------------------------------------------------------------------

    \145\ National Pain Strategy: A Comprehensive Population-Health 
Level Strategy for Pain. https://iprcc.nih.gov/sites/default/files/HHSNational_Pain_Strategy_508C.pdf.
---------------------------------------------------------------------------

    We appreciate the important concerns related to the misuse and 
overuse of opioids in the treatment of pain and to that end we note 
that in the proposed rule we have also proposed a SPADE that assess for 
the use of, as well as importantly the indication for the use of, high-
risk drugs, including opioids. Further, in the FY 2017 IRF PPS final 
rule (81 FR 52111) we adopted the Drug Regimen Review Conducted With 
Follow-Up for Identified Issues--Post Acute Care (PAC) IRF QRP measure 
which assesses whether PAC providers were responsive to potential or 
actual clinically significant medication issue(s), which includes 
issues associated with use and misuse of opioids for pain management, 
when such issues were identified.
    We also note that the proposed SPADE related to pain assessment are 
not associated with any particular approach to management. Since the 
use of opioids is associated with serious complications, particularly 
in the elderly,\146\ \147\ \148\ an array of successful non-
pharmacologic and non-opioid approaches to pain management may be 
considered. PAC providers have historically used a range of pain 
management strategies, including non-steroidal anti-inflammatory drugs, 
ice, transcutaneous electrical nerve stimulation (TENS) therapy, 
supportive devices, acupuncture, and the like. In addition, non-
pharmacological interventions for pain management include, but are not 
limited to, biofeedback, application of heat/cold, massage, physical 
therapy, stretching and strengthening exercises, chiropractic, 
electrical stimulation, radiotherapy, and ultrasound.\149\ \150\ \151\
---------------------------------------------------------------------------

    \146\ Chau, D. L., Walker, V., Pai, L., & Cho, L. M. (2008). 
Opiates and elderly: use and side effects. Clinical interventions in 
aging, 3(2), 273-8.
    \147\ Fine, P. G. (2009). Chronic Pain Management in Older 
Adults: Special Considerations. Journal of Pain and Symptom 
Management, 38(2): S4-S14.
    \148\ Solomon, D. H., Rassen, J. A., Glynn, R. J., Garneau, K., 
Levin, R., Lee, J., & Schneeweiss, S. (2010). Archives Internal 
Medicine, 170(22):1979-1986.
    \149\ Byrd L. Managing chronic pain in older adults: a long-term 
care perspective. Annals of Long-Term Care: Clinical Care and Aging. 
2013;21(12):34-40.
    \150\ Kligler, B., Bair, M.J., Banerjea, R. et al. (2018). 
Clinical Policy Recommendations from the VHA State-of-the-Art 
Conference on Non-Pharmacological Approaches to Chronic 
Musculoskeletal Pain. Journal of General Internal Medicine, 33(Suppl 
1): 16. https://doi.org/10.1007/s11606-018-4323-z.
    \151\ Chou, R., Deyo, R., Friedly, J., et al. (2017). 
Nonpharmacologic Therapies for Low Back Pain: A Systematic Review 
for an American College of Physicians Clinical Practice Guideline. 
Annals of Internal Medicine, 166(7):493-505.
---------------------------------------------------------------------------

    We believe that standardized assessment of pain interference will 
support PAC clinicians in applying best-practices in pain management 
for chronic and acute pain, consistent with current clinical 
guidelines. For example, the standardized assessment of both opioids 
and pain interference would support providers in successfully tapering 
the dosage regimens in patients/residents who arrive in the PAC setting 
with long-term opioid use off of opioids onto non-pharmacologic 
treatments and non-opioid medications, as recommended by the Society 
for Post-Acute and Long-Term Care Medicine,\152\ and consistent with 
HHS's 5-Point Strategy To Combat the Opioid Crisis \153\ which includes 
``Better Pain Management.''
---------------------------------------------------------------------------

    \152\ Society for Post-Acute and Long-Term Care Medicine (AMDA). 
(2018). Opioids in Nursing Homes: Position Statement. https://paltc.org/opioids%20in%20nursing%20homes.
    \153\ https://www.hhs.gov/opioids/about-the-epidemic/hhs-response/index.html.
---------------------------------------------------------------------------

    The Pain Interference data elements consist of three data elements: 
Pain Effect on Sleep, Pain Interference with Therapy Activities, and 
Pain Interference with Day-to-Day Activities. Pain Effect on Sleep 
assesses the frequency with which pain affects a resident's sleep. Pain 
Interference with Therapy Activities assesses the frequency with which 
pain interferes with a resident's ability to participate in therapies. 
The Pain Interference with Day-to-Day Activities assesses the extent to 
which pain interferes with a resident's ability to participate in day-
to-day activities excluding therapy.
    A similar data element on the effect of pain on activities is 
currently included in the OASIS. A similar data element on the effect 
on sleep is currently included in the MDS instrument. For more 
information on the Pain Interference data elements, we refer readers to 
the document titled ``Final Specifications for IRF QRP Quality Measures 
and Standardized Patient Assessment Data Elements,'' available at 
https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We sought public input on the relevance of conducting assessments 
on pain and specifically on the larger set of Pain Interview data 
elements included in the National Beta Test. The proposed data elements 
were supported by comments from the TEP meeting held by our data 
element contractor on April 7 to 8, 2016. The TEP affirmed the 
feasibility and clinical utility of pain as a concept in a standardized 
assessment. The TEP agreed that data elements on pain interference with 
ability to participate in therapies versus other activities should be 
addressed. Further, during a more recent convening of the same TEP on 
September 17, 2018, the TEP supported the interview-based pain data 
elements included in the National Beta Test. The TEP members were 
particularly supportive of the items that focused on how pain 
interferes with activities (that is, Pain Interference data elements), 
because understanding the extent to which pain interferes with function 
would enable clinicians to determine the need for appropriate pain 
treatment. A summary of the September 17, 2018 TEP meeting titled 
``SPADE Technical Expert Panel Summary (Third Convening)'' is available 
at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We held a public input period in 2016 to solicit feedback on the 
standardization of pain and several other items that were under 
development in prior efforts. From the prior public comment period, we 
included several pain data elements (Pain Effect on Sleep; Pain 
Interference--Therapy Activities; Pain Interference--Other Activities) 
in a second call for public input, open from

[[Page 39145]]

April 26 to June 26, 2017. The items we sought comment on were modified 
from all stakeholder and test efforts. Commenters provided general 
comments about pain assessment in general in addition to feedback on 
the specific pain items. A few commenters shared their support for 
assessing pain, the potential for pain assessment to improve the 
quality of care, and for the validity and reliability of the data 
elements. Commenters affirmed that the item of pain and the effect on 
sleep would be suitable for PAC settings. Commenters' main concerns 
included redundancy with existing data elements, feasibility and 
utility for cross-setting use, and the applicability of interview-based 
items to patients and residents with cognitive or communication 
impairments, and deficits. A summary report for the April 26 to June 
26, 2017 public comment period titled ``SPADE May-June 2017 Public 
Comment Summary Report'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    The Pain Interference data elements were included in the National 
Beta Test of candidate data elements conducted by our data element 
contractor from November 2017 to August 2018. Results of this test 
found the Pain Interference data elements to be feasible and reliable 
for use with PAC patients and residents. More information about the 
performance of the Pain Interference data elements in the National Beta 
Test can be found in the document titled ``Final Specifications for IRF 
QRP Quality Measures and Standardized Patient Assessment Data 
Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In addition, our data element contractor convened a TEP on 
September 17, 2018 for the purpose of soliciting input on the 
standardized patient assessment data elements. The TEP supported the 
interview-based pain data elements included in the National Beta Test. 
The TEP members were particularly supportive of the items that focused 
on how pain interferes with activities (that is, Pain Interference data 
elements), because understanding the extent to which pain interferes 
with function would enable clinicians to determine the need for pain 
treatment. A summary of the September 17, 2018 TEP meeting titled 
``SPADE Technical Expert Panel Summary (Third Convening)'' is available 
at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We also held Special Open Door Forums and small-group discussions 
with PAC providers and other stakeholders in 2018 for the purpose of 
updating the public about our on-going SPADE development efforts. 
Finally, on November 27, 2018, our data element contractor hosted a 
public meeting of stakeholders to present the results of the National 
Beta Test and solicit additional comments. General input on the testing 
and item development process and concerns about burden were received 
from stakeholders during this meeting and via email through February 1, 
2019. Additionally, one commenter noted strong support for the Pain 
data elements and was encouraged by the fact that this portion of the 
assessment goes beyond merely measuring the presence of pain. A summary 
of the public input received from the November 27, 2018 stakeholder 
meeting titled ``Input on Standardized Patient Assessment Data Elements 
(SPADEs) Received After November 27, 2018 Stakeholder Meeting'' is 
available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Taking together the importance of assessing for the effect of pain 
on function, stakeholder input, and strong test results, we proposed 
that the three Pain Interference data elements (Pain Effect on Sleep, 
Pain Interference with Therapy Activities, and Pain Interference with 
Day-to-Day Activities) meet the definition of standardized patient 
assessment data with respect to medical conditions and comorbidities 
under section 1899B(b)(1)(B)(iv) of the Act and to adopt the Pain 
Interference data elements (Pain Effect on Sleep; Pain Interference 
with Therapy Activities; and Pain Interference with Day-to-Day 
Activities) as standardized patient assessment data for use in the IRF 
QRP.
    Commenters submitted the following comments related to our proposal 
to adopt the Pain Interference (Pain Effect on Sleep, Pain Interference 
with Therapy Activities, and Pain Interference with Day-to-Day 
Activities) data elements.
    Comment: A few commenters noted support for the Pain Interference 
data element, noting that the data element will provide a useful and 
more accurate assessment of a patient's ability to function, and that 
understanding the impact of pain on therapy and other activities, 
including sleep, can improve the quality of care, which in turn will 
support providers in their ability to provide effective pain management 
services.
    Response: We thank the commenters for their support of the Pain 
Interference data element.
    Comment: A commenter noted that the proposed Pain Interference 
SPADEs document pain frequency, but stated that it is important to 
identify both pain frequency and pain intensity.
    Response: We wish to clarify, the Pain Interference interview data 
elements question the patient on the frequency with which pain 
interferes with sleep, therapy, or non-therapy activities. These data 
elements therefore combine the concepts of frequency and intensity, 
with the measure of intensity being interference with the named 
activities. Self-reported measures of pain intensity are often 
criticized for being infeasible to standardize. In these data elements, 
we use interference with activities as an alternative to inquiring 
about intensity.
    Comment: A commenter expressed concerns about the suitability of 
the Pain Interference data elements for use in patients with cognitive 
and communication deficits and recommended CMS consider the use of non-
verbal means to allow patients to respond to SPADEs related to pain.
    Response: We appreciate the commenter's concern surrounding pain 
assessment with patients with cognitive and communication deficits. The 
Pain Interference interview SPADEs require that a patient be able to 
communicate, whether verbally, in writing, or using another method; 
assessors may use non-verbal means to administer the questions (for 
example, providing the questions and response in writing for a patient 
with severe hearing impairment). Patients who are unable to communicate 
by any means would not be required to complete the Pain Interference 
interview SPADEs. However, evidence suggests that pain presence can be 
reliably assessed in non-communicative patients through structural 
observational protocols. To that end, we tested observational pain 
presence elements in the National Beta Test, but have chosen not to 
propose those data elements as SPADEs at this time. We will take the 
commenter's concern into consideration as the SPADEs are monitored and 
refined in the future.

[[Page 39146]]

    Comment: A commenter expressed concerns about how CMS might use 
these data elements, noting particular concern that collection of these 
data elements may inappropriately translate into an assessment of 
quality, and that data collection on this topic could create incentives 
that directly or indirectly interfere with treatment decisions.
    Response: We appreciate the commenter's concern related to wanting 
to understand how we will use the SPADEs in the future. We intend to 
continue to communicate and collaborate with stakeholders about how the 
SPADEs will be used in the IRF QRP, as those plans are developed, by 
soliciting input during the development process and establishing use of 
the SPADEs in payment and quality programs through future rulemaking.
    After careful consideration of the public comments we received, we 
are finalizing our proposal to adopt the Pain Interference (Pain Effect 
on Sleep, Pain Interference with Therapy Activities, and Pain 
Interference with Day-to-Day Activities) data elements as standardized 
patient assessment data beginning with the FY 2022 IRF QRP as proposed.
4. Impairment Data
    Hearing and vision impairments are conditions that, if unaddressed, 
affect activities of daily living, communication, physical functioning, 
rehabilitation outcomes, and overall quality of life. Sensory 
limitations can lead to confusion in new settings, increase isolation, 
contribute to mood disorders, and impede accurate assessment of other 
medical conditions. Failure to appropriately assess, accommodate, and 
treat these conditions increases the likelihood that patients and 
residents will require more intensive and prolonged treatment. Onset of 
these conditions can be gradual, so individualized assessment with 
accurate screening tools and follow-up evaluations are essential to 
determining which patients and residents need hearing- or vision-
specific medical attention or assistive devices and accommodations, 
including auxiliary aids and/or services, and to ensure that person-
directed care plans are developed to accommodate a patient's or 
resident's needs. Accurate diagnosis and management of hearing or 
vision impairment would likely improve rehabilitation outcomes and care 
transitions, including transition from institutional-based care to the 
community. Accurate assessment of hearing and vision impairment would 
be expected to lead to appropriate treatment, accommodations, including 
the provision of auxiliary aids and services during the stay, and 
ensure that patients and residents continue to have their vision and 
hearing needs met when they leave the facility.
    In alignment with our Meaningful Measures Initiative, we expect 
accurate and individualized assessment, treatment, and accommodation of 
hearing and vision impairments of patients and residents in PAC to make 
care safer by reducing harm caused in the delivery of care; promote 
effective prevention and treatment of chronic disease; strengthen 
person and family engagement as partners in their care; and promote 
effective communication and coordination of care. For example, 
standardized assessment of hearing and vision impairments used in PAC 
will support ensuring patient safety (for example, risk of falls), 
identifying accommodations needed during the stay, and appropriate 
support needs at the time of discharge or transfer. Standardized 
assessment of these data elements will: Enable or support clinical 
decision-making and early clinical intervention; person-centered, high 
quality care (for example, facilitating better care continuity and 
coordination); better data exchange and interoperability between 
settings; and longitudinal outcome analysis. Therefore, reliable data 
elements assessing hearing and vision impairments are needed to 
initiate a management program that can optimize a patient's or 
resident's prognosis and reduce the possibility of adverse events.
    Comments on the category of impairments were also submitted by 
stakeholders during the FY 2018 IRF PPS proposed rule (82 FR 20737 
through 20739) public comment period. A commenter stated hearing and 
vision assessments should be administered at the beginning of the 
assessment process to provide evidence about any sensory deficits that 
may affect the patient's ability to participate in the assessment and 
to allow the assessor to offer an assistive device.
    We sought comment on our proposals to collect as standardized 
patient assessment data the following data with respect to impairments. 
We did not receive any comments on the category of impairments.
    Final decisions on the SPADEs are given below, following more 
detailed comments on each SPADE proposal.
 Hearing
    In the FY 2020 IRF PPS proposed rule (84 FR 17317 through 17318), 
we proposed that the Hearing data element meets the definition of 
standardized patient assessment data with respect to impairments under 
section 1899B(b)(1)(B)(v) of the Act.
    As described in the FY 2018 IRF PPS proposed rule (82 FR 20737 
through 20738), accurate assessment of hearing impairment is important 
in the PAC setting for care planning and resource use. Hearing 
impairment has been associated with lower quality of life, including 
poorer physical, mental, social functioning, and emotional 
health.154 155 Treatment and accommodation of hearing 
impairment led to improved health outcomes including, but not limited 
to, quality of life.\156\ For example, hearing loss in elderly 
individuals has been associated with depression and cognitive 
impairment,157 158 159 higher rates of incident cognitive 
impairment and cognitive decline,\160\ and less time in occupational 
therapy.\161\ Accurate assessment of hearing impairment is important in 
the PAC setting for care planning and defining resource use.
---------------------------------------------------------------------------

    \154\ Dalton DS, Cruickshanks KJ, Klein BE, Klein R, Wiley TL, 
Nondahl DM. The impact of hearing loss on quality of life in older 
adults. Gerontologist. 2003;43(5):661-668.
    \155\ Hawkins K, Bottone FG, Jr., Ozminkowski RJ, et al. The 
prevalence of hearing impairment and its burden on the quality of 
life among adults with Medicare Supplement Insurance. Qual Life Res. 
2012;21(7):1135-1147.
    \156\ Horn KL, McMahon NB, McMahon DC, Lewis JS, Barker M, 
Gherini S. Functional use of the Nucleus 22-channel cochlear implant 
in the elderly. The Laryngoscope. 1991;101(3):284-288.
    \157\ Sprinzl GM, Riechelmann H. Current trends in treating 
hearing loss in elderly people: a review of the technology and 
treatment options--a mini-review. Gerontology. 2010;56(3):351-358.
    \158\ Lin FR, Thorpe R, Gordon-Salant S, Ferrucci L. Hearing 
Loss Prevalence and Risk Factors Among Older Adults in the United 
States. The Journals of Gerontology Series A: Biological Sciences 
and Medical Sciences. 2011;66A(5):582-590.
    \159\ Hawkins K, Bottone FG, Jr., Ozminkowski RJ, et al. The 
prevalence of hearing impairment and its burden on the quality of 
life among adults with Medicare Supplement Insurance. Qual Life Res. 
2012;21(7):1135-1147.
    \160\ Lin FR, Metter EJ, O'Brien RJ, Resnick SM, Zonderman AB, 
Ferrucci L. Hearing Loss and Incident Dementia. Arch Neurol. 
2011;68(2):214-220.
    \161\ Cimarolli VR, Jung S. Intensity of Occupational Therapy 
Utilization in Nursing Home Residents: The Role of Sensory 
Impairments. J Am Med Dir Assoc. 2016;17(10):939-942.
---------------------------------------------------------------------------

    The proposed data element consists of the single Hearing data 
element. This data consists of one question that assesses level of 
hearing impairment. This data element is currently in use in the MDS in 
SNFs. For more information on the Hearing data element, we refer 
readers to the document titled ``Final Specifications for IRF QRP 
Quality Measures and Standardized Patient Assessment Data Elements,'' 
available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-
Assessment-

[[Page 39147]]

Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/
IMPACT-Act-Downloads-and-Videos.html.
    The Hearing data element was first proposed as a standardized 
patient assessment data element in the FY 2018 IRF PPS proposed rule 
(82 FR 20737 through 20738). In that proposed rule, we stated that the 
proposal was informed by input we received on the PAC PRD form of the 
data element (``Ability to Hear'') through a call for input published 
on the CMS Measures Management System Blueprint website. Input 
submitted from August 12 to September 12, 2016 recommended that 
hearing, vision, and communication assessments be administered at the 
beginning of patient assessment process. A summary report for the 
August 12 to September 12, 2016 public comment period titled ``SPADE 
August 2016 Public Comment Summary Report'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In response to our proposal in the FY 2018 IRF PPS proposed rule, 
we received public comments in support of adopting the Hearing data 
element for standardized cross-setting use, noting that it would help 
address the needs of patient and residents with disabilities and that 
failing to identify impairments during the initial assessment can 
result in inaccurate diagnoses of impaired language or cognition and 
can invalidate other information obtained from patient assessment.
    Subsequent to receiving comments on the FY 2018 IRF PPS rule, the 
Hearing data element was included in the National Beta Test of 
candidate data elements conducted by our data element contractor from 
November 2017 to August 2018. Results of this test found the Hearing 
data element to be feasible and reliable for use with PAC patients and 
residents. More information about the performance of the Hearing data 
element in the National Beta Test can be found in the document titled 
``Final Specifications for IRF QRP Quality Measures and Standardized 
Patient Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In addition, our data element contractor convened a TEP on January 
5 and 6, 2017 for the purpose of soliciting input on all the SPADEs, 
including the Hearing data element. The TEP affirmed the importance of 
standardized assessment of hearing impairment in PAC patients and 
residents. A summary of the January 5 and 6, 2017 TEP meeting titled 
``SPADE Technical Expert Panel Summary (Second Convening)'' is 
available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We also held Special Open Door Forums and small-group discussions 
with PAC providers and other stakeholders in 2018 for the purpose of 
updating the public about our ongoing SPADE development efforts. 
Finally, on November 27, 2018, our data element contractor hosted a 
public meeting of stakeholders to present the results of the National 
Beta Test and solicit additional comments. General input on the testing 
and item development process and concerns about burden were received 
from stakeholders during this meeting and via email through February 1, 
2019. Additionally, a commenter noted support for the Hearing data 
element and suggested administration at the beginning of the patient 
assessment to maximize utility. A summary of the public input received 
from the November 27, 2018 stakeholder meeting titled ``Input on 
Standardized Patient Assessment Data Elements (SPADEs) Received After 
November 27, 2018 Stakeholder Meeting'' is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Due to the relatively stable nature of hearing impairment, it is 
unlikely that a patient's score on this assessment would change between 
the start and end of the IRF stay. Therefore, we proposed that IRFs 
that submit the Hearing data element with respect to admission will be 
deemed to have submitted with respect to both admission and discharge.
    Taking together the importance of assessing for hearing, 
stakeholder input, and strong test results, we proposed that the 
Hearing data element meets the definition of standardized patient 
assessment data with respect to impairments under section 
1899B(b)(1)(B)(v) of the Act and to adopt the Hearing data element as 
standardized patient assessment data for use in the IRF QRP.
    Commenters submitted the following comments related to our proposal 
for the Hearing data element.
    Comment: A few commenters supported the collection of information 
on hearing impairment. One of these commenters also suggested that CMS 
consider how hearing impairment impacts a patient's ability to respond 
to the assessment tool in general.
    Response: We thank the commenters for their support of the Hearing 
data element. We intend to reinforce assessment tips and item rationale 
through training, open door forums, and future rulemaking efforts.
    In the existing guidance manual for the IRF-PAI, we offer tips for 
administration that direct assessors to take appropriate steps to 
accommodate sensory and communication impairments when conducting the 
assessment.
    Comment: Some commenters expressed concern that severely impaired 
hearing occurs infrequently in IRF patients, thereby limiting the 
utility of the data collected.
    Response: The Hearing SPADE consists of one data element completed 
by the assessor based primarily on interacting with the patient and 
reviewing the medical record. Given the low burden of reporting the 
Hearing data element, and despite severe hearing impairment occurring 
in a small proportion of IRF patients, we believe it is important to 
systematically assess for hearing impairment in order to improve 
clinical care and care transitions.
    Comment: One commenter recommended adding ``unable to assess'' as a 
response option, which the commenter believes would be the appropriate 
choice if the patient is comatose or is unable to effectively answer 
questions related to an assessment of their hearing.
    Response: We appreciate the commenter's recommendation. The 
assessment of hearing is completed based on observing the patient 
during assessment, patient interactions with others, reviewing medical 
record documentation, and consulting with patient's family and other 
staff, in addition to interviewing the patient, so it can be completed 
when the patient is unable to effectively answer questions related to 
an assessment of their hearing.
    After careful consideration of the public comments we received, we 
are finalizing our proposal to adopt the Hearing data element as 
standardized patient assessment data beginning with the FY 2022 IRF QRP 
as proposed.
 Vision
    In the FY 2020 IRF PPS proposed rule (84 FR 17318 through 17319), 
we proposed that the Vision data element

[[Page 39148]]

meets the definition of standardized patient assessment data with 
respect to impairments under section 1899B(b)(1)(B)(v) of the Act.
    As described in the FY 2018 IRF PPS proposed rule (82 FR 20738 
through 20739), evaluation of an individual's ability to see is 
important for assessing for risks such as falls and provides 
opportunities for improvement through treatment and the provision of 
accommodations, including auxiliary aids and services, which can 
safeguard patients and residents and improve their overall quality of 
life. Further, vision impairment is often a treatable risk factor 
associated with adverse events and poor quality of life. For example, 
individuals with visual impairment are more likely to experience falls 
and hip fracture, have less mobility, and report depressive 
symptoms.162 163 164 165 166 167 168 Individualized initial 
screening can lead to life-improving interventions such as 
accommodations, including the provision of auxiliary aids and services, 
during the stay and/or treatments that can improve vision and prevent 
or slow further vision loss.
---------------------------------------------------------------------------

    \162\ Colon-Emeric CS, Biggs DP, Schenck AP, Lyles KW. Risk 
factors for hip fracture in skilled nursing facilities: Who should 
be evaluated? Osteoporos Int. 2003;14(6):484-489.
    \163\ Freeman EE, Munoz B, Rubin G, West SK. Visual field loss 
increases the risk of falls in older adults: The Salisbury eye 
evaluation. Invest Ophthalmol Vis Sci. 2007;48(10):4445-4450.
    \164\ Keepnews D, Capitman JA, Rosati RJ. Measuring patient-
level clinical outcomes of home health care. J Nurs Scholarsh. 
2004;36(1):79-85.
    \165\ Nguyen HT, Black SA, Ray LA, Espino DV, Markides KS. 
Predictors of decline in MMSE scores among older Mexican Americans. 
J Gerontol A Biol Sci Med Sci. 2002;57(3):M181-185.
    \166\ Prager AJ, Liebmann JM, Cioffi GA, Blumberg DM. Self-
reported Function, Health Resource Use, and Total Health Care Costs 
Among Medicare Beneficiaries With Glaucoma. JAMA ophthalmology. 
2016;134(4):357-365.
    \167\ Rovner BW, Ganguli M. Depression and disability associated 
with impaired vision: The MoVies Project. J Am Geriatr Soc. 
1998;46(5):617-619.
    \168\ Tinetti ME, Ginter SF. The nursing home life-space 
diameter. A measure of extent and frequency of mobility among 
nursing home residents. J Am Geriatr Soc. 1990;38(12):1311-1315.
---------------------------------------------------------------------------

    In addition, vision impairment is often a treatable risk factor 
associated with adverse events which can be prevented and accommodated 
during the stay. Accurate assessment of vision impairment is important 
in the IRF setting for care planning and defining resource use.
    The proposed data element consists of the single Vision data 
element (Ability To See in Adequate Light) that consists of one 
question with five response categories. The Vision data element that we 
proposed for standardization was tested as part of the development of 
the MDS and is currently in use in that assessment in SNFs. Similar 
data elements, but with different wording and fewer response option 
categories, are in use in the OASIS. For more information on the Vision 
data element, we refer readers to the document titled ``Final 
Specifications for IRF QRP Quality Measures and Standardized Patient 
Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    The Vision data element was first proposed as a standardized 
patient assessment data element in the FY 2018 IRF PPS proposed rule 
(82 FR 20738 through 20739).
    In that proposed rule, we stated that the proposal was informed by 
input we received on the Ability to See in Adequate Light data element 
(version tested in the PAC PRD with three response categories) through 
a call for input published on the CMS Measures Management System 
Blueprint website. Although the data element in public comment differed 
from the proposed data element, input submitted from August 12 to 
September 12, 2016 supported assessing vision in PAC settings and the 
useful information a vision data element would provide.
    We also stated that commenters had noted that the Ability to See 
item would provide important information that would facilitate care 
coordination and care planning, and consequently improve the quality of 
care. Other commenters suggested it would be helpful as an indicator of 
resource use and noted that the item would provide useful information 
about the abilities of patients and residents to care for themselves. 
Additional commenters noted that the item could feasibly be implemented 
across PAC providers and that its kappa scores from the PAC PRD support 
its validity. Some commenters noted a preference for MDS version of the 
Vision data element in SNFs over the form put forward in public 
comment, citing the widespread use of this data element. A summary 
report for the August 12 to September 12, 2016 public comment period 
titled ``SPADE August 2016 Public Comment Summary Report'' is available 
at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In response to our proposal in the FY 2018 IRF PPS proposed rule, 
we received a comment supporting having a standardized patient 
assessment data element for vision across PAC settings, but it stated 
the proposed data element captures only basic information for risk 
adjustment, and more detailed information would need to be collected to 
use it as an outcome measure.
    Subsequent to receiving comments on the FY 2018 IRF PPS rule, the 
Vision data element was included in the National Beta Test of candidate 
data elements conducted by our data element contractor from November 
2017 to August 2018. Results of this test found the Vision data element 
to be feasible and reliable for use with PAC patients and residents. 
More information about the performance of the Vision data element in 
the National Beta Test can be found in the document titled ``Final 
Specifications for IRF QRP Quality Measures and Standardized Patient 
Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In addition, our data element contractor convened a TEP on January 
5 and 6, 2017 for the purpose of soliciting input on all the SPADEs 
including the Vision data element. The TEP affirmed the importance of 
standardized assessment of vision impairment in PAC patients and 
residents. A summary of the January 5 and 6, 2017 TEP meeting titled 
``SPADE Technical Expert Panel Summary (Second Convening)'' is 
available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We also held SODFs and small-group discussions with PAC providers 
and other stakeholders in 2018 for the purpose of updating the public 
about our ongoing SPADE development efforts. Finally, on November 27, 
2018, our data element contractor hosted a public meeting of 
stakeholders to present the results of the National Beta Test and 
solicit additional comments. General input on the testing and item 
development process and concerns about burden were received from 
stakeholders during this meeting and via email through February 1, 
2019. Additionally, a commenter noted support for the Vision data 
element and suggested administration at the beginning of the patient 
assessment to maximize utility. A summary of the public input received 
from the November 27, 2018 stakeholder meeting

[[Page 39149]]

titled ``Input on Standardized Patient Assessment Data Elements 
(SPADEs) Received After November 27, 2018 Stakeholder Meeting'' is 
available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    Due to the relatively stable nature of vision impairment, it is 
unlikely that a patient's score on this assessment would change between 
the start and end of the IRF stay. Therefore, we proposed that IRFs 
that submit the Vision data element with respect to admission will be 
deemed to have submitted with respect to both admissions and discharge.
    Taking together the importance of assessing for vision, stakeholder 
input, and strong test results, we proposed that the Vision data 
element meets the definition of standardized patient assessment data 
with respect to impairments under section 1899B(b)(1)(B)(v) of the Act 
and to adopt the Vision data element as standardized patient assessment 
data for use in the IRF QRP.
    Commenters submitted the following comments related to the proposed 
rule's discussion of the Vision data element.
    Comment: A few commenters supported the collection of information 
on vision impairment. One of the commenters noted that the collection 
of information on vision impairment would support the identification 
and appropriate treatment of vision problems, which they stated were 
prevalent and undertreated.
    Response: We thank the commenters for their support.
    Comment: One commenter recommended that a doctor of optometry 
should play a lead role in conducting vision assessments, and that 
vision assessments done by other clinicians should also obtain the 
patient's own assessment of his or her vision, such as used by the 
Centers for Disease Control and Prevention (CDC) Behavioral Risk 
Factors Surveillance System survey, which questions patients ``Do you 
have serious difficulty seeing, even when wearing glasses?'' This 
commenter expressed concerns about the proposed SPADE being subjective 
and risks of mis-categorizing patients.
    Response: We appreciate the commenter's recommendation about how to 
assess for vision impairment. We do not require that a certain type of 
clinician complete assessments; the SPADEs have been developed so that 
any clinician who is trained in the administration of the assessment 
will be able to administer it correctly. The proposed item relies on 
the assessor's evaluation of the patient's vision, which has the 
advantage of reducing burden placed on the patient. We will take the 
recommendation to use patient-reported vision impairment assessment 
into consideration in the development of future assessments.
    Comment: Some commenters expressed concern that severely impaired 
vision occurs infrequently in IRF patients, thereby limiting the 
utility of the data collected.
    Response: The Vision SPADE consists of one data element completed 
by the assessor based primarily on interacting with the patient and 
reviewing the medical record. Given the low burden of the Vision data 
element, and despite severe vision impairment occurring in a small 
proportion of IRF patients, we believe it is important to 
systematically assess for vision impairment in order to improve 
clinical care and care transitions.
    Comment: A commenter recommended that CMS require a vision 
assessment at discharge, noting that vision impairment could be related 
to challenges in medication management and compliance with written 
follow-up instructions for care.
    Response: We appreciate the commenter's feedback. We agree that 
adequate vision--or the accommodations and assistive technology needed 
to compensate for vision impairment--is important to patient safety in 
the community, in part for the reasons the commenter mentions. In the 
FY 2020 IRF PPS proposed rule (84 FR 17292), we proposed that IRFs that 
submitted the Vision SPADE with respect to admission will be deemed to 
have submitted with respect to both admission and discharge; we stated 
that it is unlikely that the assessment of this SPADEs at admission 
would differ from the assessment at discharge. Vision assessment, 
collected via the Vision SPADE with respect to admission, will provide 
information that will support the patient's care while in the IRF. Out 
of consideration for the burden of data collection, and with an 
understanding that significant clinical changes to a patient's vision 
will be documented in the medical record as part of routine clinical 
practice, we are finalizing our proposal that IRFs that submit the 
Vision SPADE with respect to admission will be deemed to have submitted 
with respect to both admission and discharge. We note that during the 
discharge planning process, it is incumbent on IRF providers to make 
reasonable assurances that the patient's needs will be met in the next 
care setting, including in the home.
    Comment: One commenter recommended adding ``unable to assess'' as a 
response option, which the commenter believes would be the appropriate 
choice if the patient is comatose or is unable to effectively answer 
questions related to an assessment of their vision.
    Response: We appreciate the commenter's recommendation. However, 
the assessment of vision is completed based on consulting with 
patient's family and other staff, observing the patient including 
requesting the patient to read text or examine pictures or numbers in 
addition to interviewing the patient about their vision abilities. 
These other sources/methods can be used to complete the assessment of 
vision when the patient is unable to effectively answer questions 
related to an assessment of their vision.
    After careful consideration of the public comments we received, we 
are finalizing our proposal to adopt the Vision data element as 
standardized patient assessment data beginning with the FY 2022 IRF QRP 
as proposed.
4. New Category: Social Determinants of Health
a. Social Determinants of Health Data Collection To Inform Measures and 
Other Purposes
    Section 2(d)(2)(A) of the IMPACT Act requires CMS to assess 
appropriate adjustments to quality measures, resource measures and 
other measures, and to assess and implement appropriate adjustments to 
payment under Medicare, based on those measures, after taking into 
account studies conducted by ASPE on social risk factors (described 
below) and other information, and based on an individual's health 
status and other factors. Paragraph (C) of section 2(d)(2) of the 
IMPACT Act further requires the Secretary to carry out periodic 
analyses, at least every 3 years, based on the factors referred to 
paragraph (A) so as to monitor changes in possible relationships. 
Paragraph (B) of section 2(d)(2) of the IMPACT Act requires CMS to 
collect or otherwise obtain access to data necessary to carry out the 
requirement of the paragraph (both assessing adjustments described 
above in such paragraph (A) and for periodic analyses in such paragraph 
(C)). Accordingly we proposed to use our authority under paragraph (B) 
of section 2(d)(2) of the IMPACT Act to establish a new data source for 
information to

[[Page 39150]]

meet the requirements of paragraphs (A) and (C) of section 2(d)(2) of 
the IMPACT Act. In this rule, we proposed to collect and access data 
about social determinants of health (SDOH) in order to perform CMS' 
responsibilities under paragraphs (A) and (C) of section 2(d)(2) of the 
IMPACT Act, as explained in more detail below. Social determinants of 
health, also known as social risk factors, or health-related social 
needs, are the socioeconomic, cultural and environmental circumstances 
in which individuals live that impact their health. We proposed to 
collect information on seven proposed SDOH SPADE data elements relating 
to race, ethnicity, preferred language, interpreter services, health 
literacy, transportation, and social isolation; a detailed discussion 
of each of the proposed SDOH data elements is found in section 
VII.G.5.b. of this rule.
    We also proposed to use the assessment instrument for the IRF QRP, 
the IRF-PAI, described as a PAC assessment instrument under section 
1899B(a)(2)(B) of the Act, to collect these data via an existing data 
collection mechanism. We believe this approach will provide CMS with 
access to data with respect to the requirements of section 2(d)(2) of 
the IMPACT Act, while minimizing the reporting burden on PAC health 
care providers by relying on a data reporting mechanism already used 
and an existing system to which PAC health care providers are already 
accustomed.
    The IMPACT Act includes several requirements applicable to the 
Secretary, in addition to those imposing new data reporting obligations 
on certain PAC providers as discussed in IX.G.4.b. of this final rule. 
Paragraphs (A) and (B) of sections 2(d)(1) of the IMPACT Act require 
the Secretary, acting through the Office of the Assistant Secretary for 
Planning and Evaluation (ASPE), to conduct two studies that examine the 
effect of risk factors, including individuals' socioeconomic status, on 
quality, resource use and other measures under the Medicare program. 
The first ASPE study was completed in December 2016 and is discussed 
below, and the second study is to be completed in the fall of 2019. We 
recognize that ASPE, in its studies, is considering a broader range of 
social risk factors than the SDOH data elements in this proposal, and 
address both PAC and non-PAC settings. We acknowledge that other data 
elements may be useful to understand, and that some of those elements 
may be of particular interest in non-PAC settings. For example, for 
beneficiaries receiving care in the community, as opposed to an in-
patient facility, housing stability and food insecurity may be more 
relevant. We will continue to take into account the findings from both 
of ASPE's reports in future policy making.
    One of the ASPE's first actions under the IMPACT Act was to 
commission the National Academies of Sciences, Engineering, and 
Medicine (NASEM) to define and conceptualize socioeconomic status for 
the purposes of ASPE's two studies under section 2(d)(1) of the IMPACT 
Act. The NASEM convened a panel of experts in the field and conducted 
an extensive literature review. Based on the information collected, the 
2016 NASEM panel report titled, ``Accounting for Social Risk Factors in 
Medicare Payment: Identifying Social Risk Factors'', concluded that the 
best way to assess how social processes and social relationships 
influence key health-related outcomes in Medicare beneficiaries is 
through a framework of social risk factors instead of socioeconomic 
status. Social risk factors discussed in the NASEM report include 
socioeconomic position, race, ethnicity, gender, social context, and 
community context. These factors are discussed at length in chapter 2 
of the NASEM report, titled ``Social Risk Factors.'' \169\ Consequently 
NASEM framed the results of its report in terms of ``social risk 
factors'' rather than ``socioeconomic status'' or ``sociodemographic 
status.'' The full text of the ``Social Risk Factors'' NASEM report is 
available for reading on the website at https://www.nap.edu/read/21858/chapter/1.
---------------------------------------------------------------------------

    \169\ National Academies of Sciences, Engineering, and Medicine. 
2016. Accounting for social risk factors in Medicare payment: 
Identifying social risk factors. Chapter 2. Washington, DC: The 
National Academies Press.
---------------------------------------------------------------------------

    Each of the data elements we proposed to collect and access under 
our authority under section 2(d)(2)(B) of the IMPACT Act is identified 
in the 2016 NASEM report as a social risk factor that has been shown to 
impact care use, cost and outcomes for Medicare beneficiaries. CMS uses 
the term social determinants of health (SDOH) to denote social risk 
factors, which is consistent with the objectives of Healthy People 
2020.\170\
---------------------------------------------------------------------------

    \170\ Social Determinants of Health. Healthy People 2020. 
https://www.healthypeople.gov/2020/topics-objectives/topic/social-determinants-of-health. (February 2019).
---------------------------------------------------------------------------

    ASPE issued its first Report to Congress, titled ``Social Risk 
Factors and Performance Under Medicare's Value-Based Purchasing 
Programs,'' under section 2(d)(1)(A) of the IMPACT Act on December 21, 
2016.\171\ Using NASEM's social risk factors framework, ASPE focused on 
the following social risk factors, in addition to disability: (1) Dual 
enrollment in Medicare and Medicaid as a marker for low income; (2) 
residence in a low-income area; (3) Black race; (4) Hispanic ethnicity; 
and (5) residence in a rural area. ASPE acknowledged that the social 
risk factors examined in its report were limited due to data 
availability. The report also noted that the data necessary to 
meaningfully attempt to reduce disparities and identify and reward 
improved outcomes for beneficiaries with social risk factors have not 
been collected consistently on a national level in PAC settings. Where 
these data have been collected, the collection frequently involves 
lengthy questionnaires. More information on the Report to Congress on 
Social Risk Factors and Performance under Medicare's Value-Based 
Purchasing Programs, including the full report, is available on the 
website at https://aspe.hhs.gov/social-risk-factors-and-medicares-value-based-purchasing-programs-reports.
---------------------------------------------------------------------------

    \171\ U.S. Department of Health and Human Services, Office of 
the Assistant Secretary for Planning and Evaluation. 2016. Report to 
Congress: Social Risk Factors and Performance Under Medicare's 
Value-Based Payment Programs. Washington, DC.
---------------------------------------------------------------------------

    Section 2(d)(2) of the IMPACT Act relates to CMS activities and 
imposes several responsibilities on the Secretary relating to quality, 
resource use, and other measures under Medicare. As mentioned 
previously, under paragraph (A) of section 2(d)(2) of the IMPACT Act, 
the Secretary is required, on an ongoing basis, taking into account the 
ASPE studies and other information, and based on an individual's health 
status and other factors, to assess appropriate adjustments to quality, 
resource use, and other measures, and to assess and implement 
appropriate adjustments to Medicare payments based on those measures. 
Section 2(d)(2)(A)(i) of the IMPACT Act applies to measures adopted 
under sections (c) and (d) of section 1899B of the Act and to other 
measures under Medicare. However, CMS' ability to perform these 
analyses, and assess and make appropriate adjustments is hindered by 
limits of existing data collections on SDOH data elements for Medicare 
beneficiaries. In its first study in 2016, in discussing the second 
study, ASPE noted that information relating to many of the specific 
factors listed in the IMPACT Act, such as health literacy, limited 
English proficiency, and Medicare beneficiary activation, are not 
available in Medicare data.

[[Page 39151]]

    Paragraph 2(d)(2)(A) of the IMPACT Act specifically requires the 
Secretary to take the studies and considerations from ASPE's reports to 
Congress, as well as other information as appropriate, into account in 
assessing and implementing adjustments to measures and related payments 
based on measures in Medicare. The results of the ASPE's first study 
demonstrated that Medicare beneficiaries with social risk factors 
tended to have worse outcomes on many quality measures, and providers 
who treated a disproportionate share of beneficiaries with social risk 
factors tended to have worse performance on quality measures. As a 
result of these findings, ASPE suggested a three-pronged strategy to 
guide the development of value-based payment programs under which all 
Medicare beneficiaries receive the highest quality healthcare services 
possible. The three components of this strategy are to: (1) Measure and 
report quality of care for beneficiaries with social risk factors; (2) 
set high, fair quality standards for care provided to all 
beneficiaries; and (3) reward and support better outcomes for 
beneficiaries with social risk factors. In discussing how measuring and 
reporting quality for beneficiaries with social risk factors can be 
applied to Medicare quality payment programs, the report offered nine 
considerations across the three-pronged strategy, including enhancing 
data collection and developing statistical techniques to allow 
measurement and reporting of performance for beneficiaries with social 
risk factors on key quality and resource use measures.
    Congress, in section 2(d)(2)(B) of the IMPACT Act, required the 
Secretary to collect or otherwise obtain access to the data necessary 
to carry out the provisions of paragraph (2) of section 2(d) of the 
IMPACT Act through both new and existing data sources. Taking into 
consideration NASEM's conceptual framework for social risk factors 
discussed above, ASPE's study, and considerations under section 
2(d)(1)(A) of the IMPACT Act, as well as the current data constraints 
of ASPE's first study and its suggested considerations, we proposed to 
collect and access data about SDOH under section 2(d)(2) of the IMPACT 
Act. Our collection and use of the SDOH data described in section 
IX.G.4.b. of this final rule, under section 2(d)(2) of the IMPACT Act 
would be independent of our proposal below (in section IX.G.4.b. of 
this final rule) and our authority to require submission of that data 
for use as SPADE under section 1899B(a)(1)(B) of the Act.
    Accessing standardized data relating to the SDOH data elements on a 
national level is necessary to permit CMS to conduct periodic analyses, 
to assess appropriate adjustments to quality measures, resource use 
measures, and other measures, and to assess and implement appropriate 
adjustments to Medicare payments based on those measures. We agree with 
ASPE's observations, in the value-based purchasing context, that the 
ability to measure and track quality, outcomes, and costs for 
beneficiaries with social risk factors over time is critical as 
policymakers and providers seek to reduce disparities and improve care 
for these groups. Collecting the data as proposed will provide the 
basis for our periodic analyses of the relationship between an 
individual's health status and other factors and quality, resource use, 
and other measures, as required by section 2(d)(2) of the IMPACT Act, 
and to assess appropriate adjustments. These data will also permit us 
to develop the statistical tools necessary to maximize the value of 
Medicare data, reduce costs and improve the quality of care for all 
beneficiaries. Collecting and accessing SDOH data in this way also 
supports the three-part strategy put forth in the first ASPE report, 
specifically ASPE's consideration to enhance data collection and 
develop statistical techniques to allow measurement and reporting of 
performance for beneficiaries with social risk factors on key quality 
and resource use measures.
    For the reasons discussed above, we proposed under section 2(d)(2) 
of the IMPACT Act, to collect the data on the following SDOH: (1) Race, 
as described in section VII.G.4.b.(1) of this rule; (2) Ethnicity, as 
described in section VII.G.4.b.(1) of this rule; (3) Preferred 
Language, as described in section VII.G.4.b.(2) of this rule; (4) 
Interpreter Services, as described in section VII.G.4.b.(2) of this 
rule; (5) Health Literacy, as described in section VII.G.4.b.(3) of 
this rule; (6) Transportation, as described in section VII.G.4.b.(4) of 
this rule; and (7) Social Isolation, as described in section 
VII.G.4.b.(5) of this rule. These data elements are discussed in more 
detail below in section VII.G.4.b of this rule. A detailed discussion 
of the comments we received, along with our responses is included in 
each section.
b. Standardized Patient Assessment Data
    Section 1899B(b)(1)(B)(vi) of the Act authorizes the Secretary to 
collect SPADEs with respect to other categories deemed necessary and 
appropriate. Below we proposed to create a Social Determinants of 
Health SPADE category under section 1899B(b)(1)(B)(vi) of the Act. In 
addition to collecting SDOH data for the purposes outlined above under 
section 2(d)(2)(B), we also proposed to collect as SPADE these same 
data elements (race, ethnicity, preferred language, interpreter 
services, health literacy, transportation, and social isolation) under 
section 1899B(b)(1)(B)(vi) of the Act. We believe that this proposed 
new category of Social Determinants of Health will inform provider 
understanding of individual patient risk factors and treatment 
preferences, facilitate coordinated care and care planning, and improve 
patient outcomes. We proposed to deem this category necessary and 
appropriate, for the purposes of SPADE, because using common standards 
and definitions for PAC data elements is important in ensuring 
interoperable exchange of longitudinal information between PAC 
providers and other providers to facilitate coordinated care, 
continuity in care planning, and the discharge planning process from 
PAC settings.
    All of the Social Determinants of Health data elements we proposed 
under section 1899B(b)(1)(B)(vi) of the Act have the capacity to take 
into account treatment preferences and care goals of patients, and to 
inform our understanding of patient complexity and risk factors that 
may affect care outcomes. While acknowledging the existence and 
importance of additional social determinants of health, we proposed to 
assess some of the factors relevant for patients receiving PAC that PAC 
settings are in a position to impact through the provision of services 
and supports, such as connecting patients with identified needs with 
transportation programs, certified interpreters, or social support 
programs.
    We proposed to adopt the following seven data elements as SPADE 
under the proposed Social Determinants of Health category: Race, 
ethnicity, preferred language, interpreter services, health literacy, 
transportation, and social isolation. To select these data elements, we 
reviewed the research literature, a number of validated assessment 
tools and frameworks for addressing SDOH currently in use (for example, 
Health Leads,\172\ NASEM, Protocol for Responding to and Assessing 
Patients' Assets, Risks, and Experiences (PRAPARE), and ICD-10), and we 
engaged in discussions with stakeholders. We also prioritized balancing 
the reporting burden for PAC providers with our policy objective to 
collect SPADEs that will inform care

[[Page 39152]]

planning and coordination and quality improvement across care settings. 
Furthermore, incorporating SDOH data elements into care planning has 
the potential to reduce readmissions and help beneficiaries achieve and 
maintain their health goals.
---------------------------------------------------------------------------

    \172\ Health Leads. Available at https://healthleadsusa.org/.
---------------------------------------------------------------------------

    We also considered feedback received during a listening session 
that we held on December 13, 2018. The purpose of the listening session 
was to solicit feedback from health systems, research organizations, 
advocacy organizations and state agencies and other members of the 
public on collecting patient-level data on SDOH across care settings, 
including consideration of race, ethnicity, spoken language, health 
literacy, social isolation, transportation, sex, gender identity, and 
sexual orientation. We also gave participants an option to submit 
written comments. A full summary of the listening session, titled 
``Listening Session on Social Determinants of Health Data Elements: 
Summary of Findings,'' includes a list of participating stakeholders 
and their affiliations, and is available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    We solicited comment on these proposals.
    Commenters submitted the following comments related to the proposed 
rule's discussion of SDOH SPADEs. A discussion of these comments, along 
with our responses, appears below.
    Comment: One commenter supported the incorporation of SDOH in the 
IRF QRP, in the interest of promoting access and assuring high-quality 
care for all beneficiaries. The commenter also encouraged CMS to be 
mindful of meaningful data collection and the potential impact for data 
overload. Since SDOH have impacts far beyond the post[hyphen]acute care 
setting, the commenter cautioned data collection that cannot be readily 
gathered, shared, or replicated beyond the PAC setting.
    The commenter also encouraged CMS to consider leveraging data 
points collected during primary care visits by using social risk factor 
data captured during those encounters. They pointed out that the 
ability to have a hospital's or physician's EHR also collect, capture, 
and exchange segments of this information is powerful. The commenter 
recommended that CMS take a holistic view of SDOH across the care 
continuum so that all care settings may gather, collect or leverage 
this data efficiently and in way that maximizes its impact.
    Response: We agree that collecting SDOH data elements can be useful 
in identifying and addressing health disparities. We also agree that 
CMS should be mindful that data elements selected are useful. The 
proposed SDOH SPADEs are aligned with SDOH identified in the 2016 NASEM 
report, which was commissioned by ASPE. Regarding the commenter's 
suggestion that CMS consider how it can align existing and future SDOH 
data collection to minimize burden on providers, we agree that it is 
important to minimize duplication of effort and will take this under 
advisement for future policy development.
    Comment: One commenter recommended that CMS consider admission 
assessment for certain SPADEs as also fulfilling the discharge 
assessment requirement. The commenter supported the inclusion of the 
SDOH SPADEs and recommended that CMS require these items be assessed at 
some point during the patient's stay instead of during the admission 
assessment time window. The commenter recommended that any SDOH SPADES 
finalized should be assessed at any point during the stay.
    Response: We disagree with the commenters regarding SDOH SPADES 
should be assessed at any point during the stay. Each of the SDOH SPADE 
data elements will assist with care planning when the patient is 
admitted. It is important for providers to identify a patient's needs 
in order to better inform the patient's care decisions made during and 
after the stay, including a patient's unique risk factors and treatment 
preferences.
    Comment: Commenters were generally in favor of the concept of 
collecting SDOH data elements and provided that, if implemented 
appropriately, the data could be useful in identifying and addressing 
health care disparities, as well as refining the risk adjustment of 
outcome measures. However, some of the commenters suggested CMS not to 
finalize the proposed policy until CMS can address important issues 
around the potential future uses of these elements and the requirements 
around data collection for certain elements. The commenters provided 
that CMS did not state explicitly in the rule whether it anticipates 
the SDOH SPADEs will be used in adjusting measures and believe that the 
IMPACT Act's requirements make it likely the SPADEs will be considered 
for use in future adjustments. The commenters recommended CMS to be 
circumspect and transparent in its approaches to incorporating the data 
elements proposed in payment and quality adjustments, such as by 
collecting stakeholder feedback before implementing any adjustments.
    Response: We appreciate the commenters for recognizing that 
collecting SDOH data elements can be useful in identifying and address 
health disparities. We intend to use this data to assess the impact 
that the social determinants of health have on health outcomes. We will 
continue to work with stakeholders to promote transparency and support 
providers who serve vulnerable populations, promote high quality care, 
and refine and further implement SDOH SPADE. We appreciate the comment 
on collecting stakeholder feedback before implementing any adjustments 
to measures based on the SDOH SPADE. Collection of this data will help 
us in identifying potential disparities, conducting analyses, and 
assessing whether any adjustments are needed. Any future policy 
development based on this data would be done transparently, and involve 
solicitation of stakeholder feedback through the notice and comment 
rulemaking process as appropriate.
    Comment: Several commenters recommended that CMS include disability 
status as a SDOH that contributes to overall patient access to care, 
health status, outcomes, and many other determinants of health since it 
is already included in some Medicare risk adjustment. The commenters 
stated that ASPE's report to Congress entitled ``Social Risk Factors 
and Performance Under Medicare's Value-Based Purchasing Programs'' 
reported that disability is an independent predictor of poor mental and 
physical health outcomes and that individuals with disabilities may 
receive lower-quality preventive care.
    Response: We appreciate the comments and suggestions provided by 
the commenters. We agree that it is important to understand and meet 
the needs of patients with disabilities. While disability is not being 
currently assessed through the SPADE, it is comprehensively assessed as 
part of existing protocols around care plans and health goals. However, 
as we continue to evaluate SDOH SPADEs, we will keep commenters' 
feedback in mind and may consider these suggestions in future 
rulemaking.
    Comment: One commenter supported CMS's proposal to collect SDOH 
data within SPADEs but was concerned that all of these new elements may 
be burdensome. The commenter recommended that CMS require data 
collection on race, ethnicity, preferred

[[Page 39153]]

language, and interpreter services, and make data collection on health 
literacy, transportation, and social isolation voluntary for now and 
have the requirement phased into future rulemaking. The commenter noted 
that this would give IRFs an opportunity to adjust to the new data 
collection methods, while signaling their importance as entities that 
are currently collecting information on SDOH are experiencing various 
workflow, privacy, and other challenges. The commenter recommended that 
CMS consider including the collection of housing status in the future 
as individuals with unmet housing needs, such as homelessness or 
substandard housing, have higher health care costs and can be at risk 
for readmissions.
    Response: We thank the commenter for their comment. As discussed 
above, section 2(d)(2)(B) of the IMPACT Act requires the Secretary to 
collect or otherwise obtain access to the data necessary to carry out 
the provisions of paragraph (2) of section 2(d) of the IMPACT Act 
through both new and existing data sources. Accessing standardized data 
relating to the SDOH data elements on a national level is necessary to 
permit CMS to conduct periodic analyses, to assess appropriate 
adjustments to quality measures, resource use measures, and other 
measures, and to assess and implement appropriate adjustments to 
Medicare payments based on those measures. Collecting the data as 
proposed will provide the basis for our periodic analyses of the 
relationship between an individual's health status and other factors 
and quality, resource use, and other measures, as required by section 
2(d)(2) of the IMPACT Act, and to assess appropriate adjustments. 
Regarding the suggestion that CMS consider a housing status SPADE data 
element in future rulemaking efforts, we appreciate this feedback and 
will consider this suggestion in future rulemaking efforts on SPADE 
SDOH data elements.
(1) Race and Ethnicity
    The persistence of racial and ethnic disparities in health and 
health care is widely documented, including in PAC 
settings.173 174 175 176 177 Despite the trend toward 
overall improvements in quality of care and health outcomes, the Agency 
for Healthcare Research and Quality, in its National Healthcare Quality 
and Disparities Reports, consistently indicates that racial and ethnic 
disparities persist, even after controlling for factors such as income, 
geography, and insurance.\178\ For example, racial and ethnic 
minorities tend to have higher rates of infant mortality, diabetes and 
other chronic conditions, and visits to the emergency department, and 
lower rates of having a usual source of care and receiving 
immunizations such as the flu vaccine.\179\ Studies have also shown 
that African Americans are significantly more likely than white 
Americans to die prematurely from heart disease and stroke.\180\ 
However, our ability to identify and address racial and ethnic health 
disparities has historically been constrained by data limitations, 
particularly for smaller populations groups such as Asians, American 
Indians and Alaska Natives, and Native Hawaiians and other Pacific 
Islanders.\181\
---------------------------------------------------------------------------

    \173\ 2017 National Healthcare Quality and Disparities Report. 
Rockville, MD: Agency for Healthcare Research and Quality; September 
2018. AHRQ Pub. No. 18-0033-EF.
    \174\ Fiscella, K. and Sanders, M.R. Racial and Ethnic 
Disparities in the Quality of Health Care. (2016). Annual Review of 
Public Health. 37:375-394.
    \175\ 2018 National Impact Assessment of the Centers for 
Medicare & Medicaid Services (CMS) Quality Measures Reports. 
Baltimore, MD: U.S. Department of Health and Human Services, Centers 
for Medicare and Medicaid Services; February 28, 2018.
    \176\ Smedley, B.D., Stith, A.Y., & Nelson, A.R. (2003). Unequal 
treatment: Confronting racial and ethnic disparities in health care. 
Washington, DC, National Academy Press.
    \177\ Chase, J., Huang, L. and Russell, D. (2017). Racial/ethnic 
disparities in disability outcomes among post-acute home care 
patients. J of Aging and Health. 30(9):1406-1426.
    \178\ National Healthcare Quality and Disparities Reports. 
(December 2018). Agency for Healthcare Research and Quality, 
Rockville, MD. http://www.ahrq.gov/research/findings/nhqrdr/index.html.
    \179\ National Center for Health Statistics. Health, United 
States, 2017: With special feature on mortality. Hyattsville, 
Maryland. 2018.
    \180\ HHS. Heart disease and African Americans. 2016b. (October 
24, 2016). http://minorityhealth.hhs.gov/omh/browse.aspx?lvl=4&lvlid=19.
    \181\ National Academies of Sciences, Engineering, and Medicine; 
Health and Medicine Division; Board on Population Health and Public 
Health Practice; Committee on Community-Based Solutions to Promote 
Health Equity in the United States; Baciu A, Negussie Y, Geller A, 
et al., editors. Communities in Action: Pathways to Health Equity. 
Washington (DC): National Academies Press (US); 2017 Jan 11. 2, The 
State of Health Disparities in the United States. Available at 
https://www.ncbi.nlm.nih.gov/books/NBK425844/.
---------------------------------------------------------------------------

    The ability to improve understanding of and address racial and 
ethnic disparities in PAC outcomes requires the availability of better 
data. There is currently a Race and Ethnicity data element, collected 
in the MDS, LCDS, IRF-PAI, and OASIS, that consists of a single 
question, which aligns with the 1997 Office of Management and Budget 
(OMB) minimum data standards for federal data collection efforts.\182\ 
The 1997 OMB Standard lists five minimum categories of race: (1) 
American Indian or Alaska Native; (2) Asian; (3) Black or African 
American; (4) Native Hawaiian or Other Pacific Islander; (5) and White. 
The 1997 OMB Standard also lists two minimum categories of ethnicity: 
(1) Hispanic or Latino; and (2) Not Hispanic or Latino. The 2011 HHS 
Data Standards requires a two-question format when self-identification 
is used to collect data on race and ethnicity. Large federal surveys 
such as the National Health Interview Survey, Behavioral Risk Factor 
Surveillance System, and the National Survey on Drug Use and Health, 
have implemented the 2011 HHS race and ethnicity data standards. CMS 
has similarly updated the Medicare Current Beneficiary Survey, Medicare 
Health Outcomes Survey, and the Health Insurance Marketplace 
Application for Health Coverage with the 2011 HHS data standards. More 
information about the HHS Race and Ethnicity Data Standards are 
available on the website at https://minorityhealth.hhs.gov/omh/browse.aspx?lvl=3&lvlid=54.
---------------------------------------------------------------------------

    \182\ ``Revisions to the Standards for the Classification of 
Federal Data on Race and Ethnicity (Notice of Decision)''. Federal 
Register 62:210 (October 30, 1997) pp. 58782-58790. Available at 
https://www.govinfo.gov/content/pkg/FR-1997-10-30/pdf/97-28653.pdf.
---------------------------------------------------------------------------

    We proposed to revise the current Race and Ethnicity data element 
for purposes of this proposal to conform to the 2011 HHS Data Standards 
for person-level data collection, while also meeting the 1997 OMB 
minimum data standards for race and ethnicity. Rather than one data 
element that assesses both race and ethnicity, we proposed two separate 
data elements: One for Race and one for Ethnicity, that would conform 
with the 2011 HHS Data Standards and the 1997 OMB Standard. In 
accordance with the 2011 HHS Data Standards a two-question format would 
be used for the proposed race and ethnicity data elements.
    The proposed Race data element asks, ``What is your race? We 
proposed to include fourteen response options under the race data 
element: (1) White; (2) Black or African American; (3) American Indian 
or Alaska Native; (4) Asian Indian; (5) Chinese; (6) Filipino; (7) 
Japanese; (8) Korean; (9) Vietnamese; (10) Other Asian; (11) Native 
Hawaiian; (12) Guamanian or Chamorro; (13) Samoan; and (14) Other 
Pacific Islander.
    The proposed Ethnicity data element asks, ``Are you Hispanic, 
Latino/a, or Spanish origin?'' We proposed to include five response 
options under the ethnicity data element: (1) Not of Hispanic, Latino/
a, or Spanish origin; (2) Mexican, Mexican American,

[[Page 39154]]

Chicano/a; (3) Puerto Rican; (4) Cuban; and (5) Another Hispanic, 
Latino, or Spanish Origin. We are including the addition of ``of'' to 
the Ethnicity data element to read, ``Are you of Hispanic, Latino/a, or 
Spanish origin?''
    We believe that the two proposed data elements for race and 
ethnicity conform to the 2011 HHS Data Standards for person-level data 
collection, while also meeting the 1997 OMB minimum data standards for 
race and ethnicity, because under those standards, more detailed 
information on population groups can be collected if those additional 
categories can be aggregated into the OMB minimum standard set of 
categories.
    In addition, we received stakeholder feedback during the December 
13, 2018 SDOH listening session on the importance of improving response 
options for race and ethnicity as a component of health care 
assessments and for monitoring disparities. Some stakeholders 
emphasized the importance of allowing for self-identification of race 
and ethnicity for more categories than are included in the 2011 HHS 
Standard to better reflect state and local diversity, while 
acknowledging the burden of coding an open-ended health care assessment 
question across different settings.
    We believe that the proposed modified race and ethnicity data 
elements more accurately reflect the diversity of the U.S. population 
than the current race/ethnicity data element included in MDS, LCDS, 
IRF-PAI, and OASIS.183 184 185 186 We believe, and research 
consistently shows, that improving how race and ethnicity data are 
collected is an important first step in improving quality of care and 
health outcomes. Addressing disparities in access to care, quality of 
care, and health outcomes for Medicare beneficiaries begins with 
identifying and analyzing how SDOH, such as race and ethnicity, align 
with disparities in these areas.\187\ Standardizing self-reported data 
collection for race and ethnicity allows for the equal comparison of 
data across multiple healthcare entities.\188\ By collecting and 
analyzing these data, CMS and other healthcare entities will be able to 
identify challenges and monitor progress. The growing diversity of the 
U.S. population and knowledge of racial and ethnic disparities within 
and across population groups supports the collection of more granular 
data beyond the 1997 OMB minimum standard for reporting categories. The 
2011 HHS race and ethnicity data standard includes additional detail 
that may be used by PAC providers to target quality improvement efforts 
for racial and ethnic groups experiencing disparate outcomes. For more 
information on the Race and Ethnicity data elements, we refer readers 
to the document titled ``Final Specifications for IRF QRP Measures and 
Standardized Patient Assessment Data Elements,'' available at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
---------------------------------------------------------------------------

    \183\ Penman-Aguilar, A., Talih, M., Huang, D., Moonesinghe, R., 
Bouye, K., Beckles, G. (2016). Measurement of Health Disparities, 
Health Inequities, and Social Determinants of Health to Support the 
Advancement of Health Equity. J Public Health Manag Pract. 22 Suppl 
1: S33-42.
    \184\ Ramos, R., Davis, J.L., Ross, T., Grant, C.G., Green, B.L. 
(2012). Measuring health disparities and health inequities: Do you 
have REGAL data? Qual Manag Health Care. 21(3):176-87.
    \185\ IOM (Institute of Medicine). 2009. Race, Ethnicity, and 
Language Data: Standardization for Health Care Quality Improvement. 
Washington, DC: The National Academies Press.
    \186\ ``Revision of Standards for Maintaining, Collecting, and 
Presenting Federal Data on Race and Ethnicity: Proposals From 
Federal Interagency Working Group (Notice and Request for 
Comments).'' Federal Register 82: 39 (March 1, 2017) p. 12242.
    \187\ National Academies of Sciences, Engineering, and Medicine; 
Health and Medicine Division; Board on Population Health and Public 
Health Practice; Committee on Community-Based Solutions to Promote 
Health Equity in the United States; Baciu A, Negussie Y, Geller A, 
et al., editors. Communities in Action: Pathways to Health Equity. 
Washington (DC): National Academies Press (US); 2017 Jan 11. 2, The 
State of Health Disparities in the United States. Available at 
https://www.ncbi.nlm.nih.gov/books/NBK425844/.
    \188\ IOM (Institute of Medicine). 2009. Race, Ethnicity, and 
Language Data: Standardization for Health Care Quality Improvement. 
Washington, DC: The National Academies Press.
---------------------------------------------------------------------------

    In an effort to standardize the submission of race and ethnicity 
data among IRFs, HHAs, SNFs and LTCHs, for the purposes outlined in 
section 1899B(a)(1)(B) of the Act, while minimizing the reporting 
burden, we proposed to adopt the Race and Ethnicity data elements 
described above as SPADEs with respect to the proposed Social 
Determinants of Health category.
    Specifically, we proposed to replace the current Race/Ethnicity 
data element with the proposed Race and Ethnicity data elements on the 
IRF-PAI. We also proposed that IRFs that submit the Race and Ethnicity 
data elements with respect to admission will be considered to have 
submitted with respect to discharge as well, because it is unlikely 
that the results of these assessment findings will change between the 
start and end of the IRF stay, making the information submitted with 
respect to a patient's admission the same with respect to a patient's 
discharge.
    We solicited comment on these proposals.
    Commenters submitted the following comments related to the proposed 
rule's discussion of the Race and Ethnicity SPADEs. A discussion of 
these comments, along with our responses, appears below.
    Comment: Some commenters noted that the response options for race 
do not align with those used in other government data, such as the U.S. 
Census or the Office of Management and Budget (OMB). The commenters 
also stated these responses are not consistent with the recommendations 
made in the 2009 Institute of Medicine report. The commenters pointed 
out that IOM report recommended using broader OMB race categories and 
granular ethnicities chosen from a national standard set that can be 
``rolled up'' into the broader categories. The commenters stated that 
it is unclear how CMS chose the 14 response options under the race data 
element and the five options under the ethnicity element and worried 
that these response options would add to the confusion that already may 
exist for patients about what terms like ``race'' and ``ethnicity'' 
mean for the purposes of health care data collection. The commenters 
also noted that CMS should confer directly with experts on the issue to 
ensure patient assessments are collecting the right data in the right 
way before these SDOH SPADEs are finalized.
    Response: The proposed Race and Ethnicity categories align with and 
are rolled up into the 1997 OMB minimum data standards and conforming 
with the 2011 HHS Data Standards as described in the implementation 
guidance titled ``U.S. Department of Health and Human Services 
Implementation Guidance on Data Collection Standards for Race, 
Ethnicity, Sex, Primary Language, and Disability Status'' at https://aspe.hhs.gov/basic-report/hhs-implementation-guidance-data-collection-standards-race-ethnicity-sex-primary-language-and-disability-status. As 
stated in the proposed rule, the 14 race categories and the 5 ethnicity 
categories conform with the 2011 HHS Data Standards for person-level 
data collection, which were developed in fulfillment of section 4302 of 
the Affordable Care Act that required the Secretary of HHS to establish 
data collection standards for race, ethnicity, sex, primary language, 
and disability status. Through the HHS Data Council, which is the 
principal, senior internal Departmental forum and advisory body to the 
Secretary on health and human

[[Page 39155]]

services data policy and coordinates HHS data collection and analysis 
activities, the Section 4302 Standards Workgroup was formed. The 
Workgroup included representatives from HHS, the OMB, and the Census 
Bureau. The Workgroup examined current federal data collection 
standards, adequacy of prior testing, and quality of the data produced 
in prior surveys; consulted with statistical agencies and programs; 
reviewed OMB data collection standards and the Institute of Medicine 
(IOM) Report Race, Ethnicity, and Language Data Collection: 
Standardization for Health Care Quality Improvement; sought input from 
national experts; and built on its members' experience with collecting 
and analyzing demographic data. As a result of this Workgroup, a set of 
data collection standards were developed, and then published for public 
comment. This set of data collection standards is referred to as the 
2011 HHS Data Standards.\189\ As described in the implementation 
guidance provided above, the categories of race and ethnicity under the 
2011 HHS Data Standards allow for more detailed information to be 
collected and the additional categories under the 2011 HHS Data 
Standards can be aggregated into the OMB minimum standards set of 
categories.
---------------------------------------------------------------------------

    \189\ HHS Data Standards. Available at https://aspe.hhs.gov/basic-report/hhs-implementation-guidance-data-collection-standards-race-ethnicity-sex-primary-language-and-disability-status.
---------------------------------------------------------------------------

    As noted in the FY 2020 IRF PPS proposed rule (84 FR 17321 through 
17323), we conferred with experts by conducting a listening session 
regarding the proposed SDOH data elements regarding the importance of 
improving response options for race and ethnicity as a component of 
health care assessments and for monitoring disparities. Some 
stakeholders emphasized the importance of allowing for self-
identification of race and ethnicity for more categories than are 
included in the 2011 HHS Data Standards to better reflect state and 
local diversity.
    Comment: A commenter recommended that CMS consider the implications 
of having PAC providers collect Race and Ethnicity codes that vary from 
the Race and Ethnicity codes collected by other healthcare providers, 
specifically acute-care hospitals. The commenter noted that unless all 
care providers are expected to utilize the uniform 2011 HHS Data 
Standards, the consistency and accuracy of race and ethnicity data 
across settings will likely be unreliable and problematic. Another 
commenter provided that the proposed list of response options for Race 
may not include all races that should be reflected, for example, Native 
African and Middle Eastern. In addition, the item should include 
``check all that apply'' to ensure accurate and complete data 
collection. The commenter encouraged CMS to refine the list of response 
options for Race and provide a rationale for the final list of response 
options.
    Response: We thank the commenter and agree that it is important to 
collect race and ethnicity data in a consistent way. The race and 
ethnicity categories that were proposed align with the 2011 HHS Data 
Standards and are rolled up into the 1997 OMB minimum data standards, 
which can be found at https://aspe.hhs.gov/basic-report/hhs-implementation-guidance-data-collection-standards-race-ethnicity-sex-primary-language-and-disability-status. For example, the 1997 OMB 
minimum data standard for Hispanic is the roll up category for the 
following response options on the 2011 HHS Data Standards: Mexican, 
Mexican American, Chicano/a; Puerto Rican; Cuban; another Hispanic, 
Latino, or Spanish origin. However, we will take the comment under 
advisement for future consideration. We also note that the option for 
``check all that apply'' is available for providers to choose from the 
list of response options.
    Comment: A commenter supported the opportunities to better account 
for SDOH in the diagnosis and treatment of patients but is concerned by 
the specificity of several of the seven proposed element for data 
collection for example, collection of race by Japanese, Chinese, 
Korean, etc. The commenter's concern is with the added burden in 
collecting the level of specificity outlined, and the commenter 
requested that CMS provide more detailed guidance in the final rule 
regarding how this information should be collected and shared in 
compliance with HIPAA. Further, the commenter asked that the agency 
outlines its expectations for how this newly collected information will 
be used by Medicare for payment and public reporting.
    Response: For the Race and Ethnicity SPADE, this data should be 
completed based on the response of the patient. It is important to ask 
the patient to select the category or categories that most closely 
correspond to their race and ethnicity. Respondents should be offered 
the option of selecting one or more race and ethnicity categories. 
Observer identification or medical record documentation may not be 
used.
    The SDOH data elements that will be collected will assist with care 
coordination and with evaluating the impact of disparities. With 
respect to how the data will be used for payment and public reporting, 
any potential future use of the data for these purposes would be done 
through future rulemaking.
    SDOH data elements should be treated the same as other data 
collected on the assessment tool. As to any specific HIPAA questions, 
we appreciate the commenter's commitment to compliance with the HIPAA 
requirements, but note that the Office for Civil Rights (OCR) is tasked 
with implementing and enforcing HIPAA, not CMS. Commenters should 
consult appropriate counsel in instances in which they are unsure of 
their HIPAA status, or the permissibility of a disclosure under the 
HIPAA Privacy Rule. In doing so, commenters may wish to consult 45 CFR 
164.103 (definition of ``required by law'') and Sec.  164.512(a) 
(allowing ``required by law'' disclosures).
(2) Preferred Language and Interpreter Services
    More than 64 million Americans speak a language other than English 
at home, and nearly 40 million of those individuals have limited 
English proficiency (LEP).\190\ Individuals with LEP have been shown to 
receive worse care and have poorer health outcomes, including higher 
readmission rates.191 192 193 Communication with individuals 
with LEP is an important component of high quality health care, which 
starts by understanding the population in need of language services. 
Unaddressed language barriers between a patient and provider care team 
negatively affects the ability to identify and address individual 
medical and non-medical care needs, to convey and understand clinical 
information, as well as discharge and follow up instructions, all of 
which are necessary for providing high quality care. Understanding the 
communication assistance needs of patients with LEP, including 
individuals who are Deaf or hard of

[[Page 39156]]

hearing, is critical for ensuring good outcomes.
---------------------------------------------------------------------------

    \190\ U.S. Census Bureau, 2013-2017 American Community Survey 5-
Year Estimates.
    \191\ Karliner LS, Kim SE, Meltzer DO, Auerbach AD. Influence of 
language barriers on outcomes of hospital care for general medicine 
inpatients. J Hosp Med. 2010 May-Jun;5(5):276-82. doi: 10.1002/
jhm.658.
    \192\ Kim EJ, Kim T, Paasche-Orlow MK, et al. Disparities in 
Hypertension Associated with Limited English Proficiency. J Gen 
Intern Med. 2017 Jun;32(6):632-639. doi: 10.1007/s11606-017-3999-9.
    \193\ National Academies of Sciences, Engineering, and Medicine. 
2016. Accounting for social risk factors in Medicare payment: 
Identifying social risk factors. Washington, DC: The National 
Academies Press.
---------------------------------------------------------------------------

    Presently, the preferred language of patients and residents and 
need for interpreter services are assessed in two PAC assessment tools. 
The LCDS and the MDS use the same two data elements to assess preferred 
language and whether a patient or resident needs or wants an 
interpreter to communicate with health care staff. The MDS initially 
implemented preferred language and interpreter services data elements 
to assess the needs of SNF residents and patients and inform care 
planning. For alignment purposes, the LCDS later adopted the same data 
elements for LTCHs. The 2009 NASEM (formerly Institute of Medicine) 
report on standardizing data for health care quality improvement 
emphasizes that language and communication needs should be assessed as 
a standard part of health care delivery and quality improvement 
strategies.\194\
---------------------------------------------------------------------------

    \194\ IOM (Institute of Medicine). 2009. Race, Ethnicity, and 
Language Data: Standardization for Health Care Quality Improvement. 
Washington, DC: The National Academies Press.
---------------------------------------------------------------------------

    In developing our proposal for a standardized language data element 
across PAC settings, we considered the current preferred language and 
interpreter services data elements that are in LCDS and MDS. We also 
considered the 2011 HHS Primary Language Data Standard and peer-
reviewed research. The current preferred language data element in LCDS 
and MDS asks, ``What is your preferred language?'' Because the 
preferred language data element is open-ended, the patient or resident 
is able to identify their preferred language, including American Sign 
Language (ASL). Finally, we considered the recommendations from the 
2009 NASEM (formerly Institute of Medicine) report, ``Race, Ethnicity, 
and Language Data: Standardization for Health Care Quality 
Improvement.'' In it, the committee recommended that organizations 
evaluating a patient's language and communication needs for health care 
purposes, should collect data on the preferred spoken language and on 
an individual's assessment of his/her level of English proficiency.
    A second language data element in LCDS and MDS asks, ``Do you want 
or need an interpreter to communicate with a doctor or health care 
staff?'' and includes yes or no response options. In contrast, the 2011 
HHS Primary Language Data Standard recommends either a single question 
to assess how well someone speaks English or, if more granular 
information is needed, a two-part question to assess whether a language 
other than English is spoken at home and if so, identify that language. 
However, neither option allows for a direct assessment of a patient's 
or resident's preferred spoken or written language nor whether they 
want or need interpreter services for communication with a doctor or 
care team, both of which are an important part of assessing patient/
resident needs and the care planning process. More information about 
the HHS Data Standard for Primary Language is available on the website 
at https://minorityhealth.hhs.gov/omh/browse.aspx?lvl=3&lvlid=54.
    Research consistently recommends collecting information about an 
individual's preferred spoken language and evaluating those responses 
for purposes of determining language access needs in health care.\195\ 
However, using ``preferred spoken language'' as the metric does not 
adequately account for people whose preferred language is ASL, which 
would necessitate adopting an additional data element to identify 
visual language. The need to improve the assessment of language 
preferences and communication needs across PAC settings should be 
balanced with the burden associated with data collection on the 
provider and patient. Therefore we proposed to retain the Preferred 
Language and Interpreter Services data elements currently in use on the 
MDS and LCDS on the IRF-PAI.
---------------------------------------------------------------------------

    \195\ Guerino, P. and James, C. Race, Ethnicity, and Language 
Preference in the Health Insurance Marketplaces 2017 Open Enrollment 
Period. Centers for Medicare & Medicaid Services, Office of Minority 
Health. Data Highlight: Volume 7--April 2017. Available at https://www.cms.gov/About-CMS/Agency-Information/OMH/Downloads/Data-Highlight-Race-Ethnicity-and-Language-Preference-Marketplace.pdf.
---------------------------------------------------------------------------

    In addition, we received feedback during the December 13, 2018 
listening session on the importance of evaluating and acting on 
language preferences early to facilitate communication and allowing for 
patient self-identification of preferred language. Although the 
discussion about language was focused on preferred spoken language, 
there was general consensus among participants that stated language 
preferences may or may not accurately indicate the need for interpreter 
services, which supports collecting and evaluating data to determine 
language preference, as well as the need for interpreter services. An 
alternate suggestion was made to inquire about preferred language 
specifically for discussing health or health care needs. While this 
suggestion does allow for ASL as a response option, we do not have data 
indicating how useful this question might be for assessing the desired 
information and thus we are not including this question in our 
proposal.
    Improving how preferred language and need for interpreter services 
data are collected is an important component of improving quality by 
helping PAC providers and other providers understand patient needs and 
develop plans to address them. For more information on the Preferred 
Language and Interpreter Services data elements, we refer readers to 
the document titled ``Final Specifications for IRF QRP Measures and 
Standardized Patient Assessment Data Elements,'' available on the 
website at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In an effort to standardize the submission of language data among 
IRFs, HHAs, SNFs and LTCHs, for the purposes outlined in section 
1899B(a)(1)(B) of the Act, while minimizing the reporting burden, we 
proposed to adopt the Preferred Language and Interpreter Services data 
elements currently used on the MDS and LCDS, and described above, as 
SPADEs with respect to the Social Determinants of Health category. We 
proposed to add the current Preferred Language and Interpreter Services 
data elements from the MDS and LCDS to the IRF-PAI.
    We solicited comment on these proposals.
    Commenters submitted the following comments related to the proposed 
rule's discussion of Preferred Language and Interpreter Services 
SPADEs. A discussion of these comments, along with our responses, 
appears below.
    Comment: Some commenters noted that, if finalized, IRFs should only 
need to submit data on the race and ethnicity SPADEs with respect to 
admission and would not need to collect and report again at discharge, 
as it is unlikely that patient status for these elements will change. 
The commenters believe that a patient's preferred language and need for 
an interpreter also are unlikely to change between admission and 
discharge; thus, the commenter urged CMS to require collection of these 
SDOH SPADEs with respect to admission only.
    Response: We thank the commenters for the comment. With regard to 
the submission of the Preferred Language SPADE and the Interpreter 
Services SPADE, we agree with the commenters that it is unlikely that 
the assessment of Preferred Language and Interpreter

[[Page 39157]]

Services at admission would differ from assessment at discharge. As 
discussed in previous response for Vision and Hearing, we believe that 
the submission of preferred language and the need for an interpreter is 
similar to the submission of Race, Ethnicity, Hearing, and Vision 
SPADES.
    We account for this change to the Collection of Information 
requirements for the IRF QRP in XIV.C of this final rule. Based on the 
comments received, and for the reasons discussed, we are finalizing 
that the Preferred Language and Interpreter Services SPADEs be 
collected as proposed with the modification that we will deem IRFs that 
submit these two SPADEs with respect to admission to have submitted 
with respect to both admission and discharge.
(3) Health Literacy
    The Department of Health and Human Services defines health literacy 
as ``the degree to which individuals have the capacity to obtain, 
process, and understand basic health information and services needed to 
make appropriate health decisions.'' \196\ Similar to language 
barriers, low health literacy can interfere with communication between 
the provider and patient and the ability for patients or their 
caregivers to understand and follow treatment plans, including 
medication management. Poor health literacy is linked to lower levels 
of knowledge about health, worse health outcomes, and the receipt of 
fewer preventive services, but higher medical costs and rates of 
emergency department use.\197\
---------------------------------------------------------------------------

    \196\ U.S. Department of Health and Human Services, Office of 
Disease Prevention and Health Promotion. National action plan to 
improve health literacy. Washington (DC): Author; 2010.
    \197\ National Academies of Sciences, Engineering, and Medicine. 
2016. Accounting for social risk factors in Medicare payment: 
Identifying social risk factors. Washington, DC: The National 
Academies Press.
---------------------------------------------------------------------------

    Health literacy is prioritized by Healthy People 2020 as an 
SDOH.\198\ Healthy People 2020 is a long-term, evidence-based effort 
led by the Department of Health and Human Services that aims to 
identify nationwide health improvement priorities and improve the 
health of all Americans. Although not designated as a social risk 
factor in NASEM's 2016 report on accounting for social risk factors in 
Medicare payment, the NASEM noted that health literacy is impacted by 
other social risk factors and can affect access to care, as well as 
quality of care and health outcomes.\199\ Assessing for health literacy 
across PAC settings would facilitate better care coordination and 
discharge planning. A significant challenge in assessing the health 
literacy of individuals is avoiding excessive burden on patients and 
health care providers. The majority of existing, validated health 
literacy assessment tools use multiple screening items, generally with 
no fewer than four, which would make them burdensome if adopted in MDS, 
LCDS, IRF-PAI, and OASIS. The Single Item Literacy Screener (SILS) 
question questions, ``How often do you need to have someone help you 
when you read instructions, pamphlets, or other written material from 
your doctor or pharmacy?'' Possible response options are: (1) Never; 
(2) Rarely; (3) Sometimes; (4) Often; and (5) Always. The SILS 
question, which assesses reading ability, (a primary component of 
health literacy), tested reasonably well against the 36 item Short Test 
of Functional Health Literacy in Adults (S-TOFHLA), a thoroughly vetted 
and widely adopted health literacy test, in assessing the likelihood of 
low health literacy in an adult sample from primary care practices 
participating in the Vermont Diabetes Information 
System.200 201 The S-TOFHLA is a more complex assessment 
instrument developed using actual hospital related materials such as 
prescription bottle labels and appointment slips, and often considered 
the instrument of choice for a detailed evaluation of health 
literacy.\202\ Furthermore, the S-TOFHLA instrument is proprietary and 
subject to purchase for individual entities or users.\203\ Given that 
SILS is publicly available, shorter and easier to administer than the 
full health literacy screen, and research found that a positive result 
on the SILS demonstrates an increased likelihood that an individual has 
low health literacy, we proposed to use the single-item reading 
question for health literacy in the standardized data collection across 
PAC settings. We believe that use of this data element will provide 
sufficient information about the health literacy of IRF patients to 
facilitate appropriate care planning, care coordination, and 
interoperable data exchange across PAC settings.
---------------------------------------------------------------------------

    \198\ Social Determinants of Health. Healthy People 2020. 
https://www.healthypeople.gov/2020/topics-objectives/topic/social-determinants-of-health. (February 2019).
    \199\ U.S. Department of Health & Human Services, Office of the 
Assistant Secretary for Planning and Evaluation. Report to Congress: 
Social Risk Factors and Performance Under Medicare's Value-Based 
Purchasing Programs. Available at https://aspe.hhs.gov/pdf-report/report-congress-social-risk-factors-and-performance-under-medicares-value-based-purchasing-programs. Washington, DC: 2016.
    \200\ Morris, N.S., MacLean, C.D., Chew, L.D., & Littenberg, B. 
(2006). The Single Item Literacy Screener: evaluation of a brief 
instrument to identify limited reading ability. BMC family practice, 
7, 21. doi:10.1186/1471-2296-7-21.
    \201\ Brice, J.H., Foster, M.B., Principe, S., Moss, C., Shofer, 
F.S., Falk, R.J., Ferris, M.E., DeWalt, D.A. (2013). Single-item or 
two-item literacy screener to predict the S-TOFHLA among adult 
hemodialysis patients. Patient Educ Couns. 94(1):71-5.
    \202\ University of Miami, School of Nursing & Health Studies, 
Center of Excellence for Health Disparities Research. Test of 
Functional Health Literacy in Adults (TOFHLA). (March 2019). 
Available at https://elcentro.sonhs.miami.edu/research/measures-library/tofhla/index.html.
    \203\ Nurss, J.R., Parker, R.M., Williams, M.V., &Baker, D.W. 
David W. (2001). TOFHLA. Peppercorn Books & Press. Available at 
http://www.peppercornbooks.com/catalog/information.php?info_id=5.
---------------------------------------------------------------------------

    In addition, we received feedback during the December 13, 2018 SDOH 
listening session on the importance of recognizing health literacy as 
more than understanding written materials and filling out forms, as it 
is also important to evaluate whether patients understand their 
conditions. However, the NASEM recently recommended that health care 
providers implement health literacy universal precautions instead of 
taking steps to ensure care is provided at an appropriate literacy 
level based on individualized assessment of health literacy.\204\ Given 
the dearth of Medicare data on health literacy and gaps in addressing 
health literacy in practice, we recommend the addition of a health 
literacy data element.
---------------------------------------------------------------------------

    \204\ Hudson, S., Rikard, R.V., Staiculescu, I. & Edison, K. 
(2017). Improving health and the bottom line: The case for health 
literacy. In Building the case for health literacy: Proceedings of a 
workshop. Washington, DC: The National Academies Press.
---------------------------------------------------------------------------

    The proposed Health Literacy data element is consistent with 
considerations raised by NASEM and other stakeholders and research on 
health literacy, which demonstrates an impact on health care use, cost, 
and outcomes.\205\ For more information on the proposed Health Literacy 
data element, we refer readers to the document titled ``Final 
Specifications for IRF QRP Measures and Standardized Patient Assessment 
Data Elements,'' available on the website at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
---------------------------------------------------------------------------

    \205\ National Academies of Sciences, Engineering, and Medicine. 
2016. Accounting for Social Risk Factors in Medicare Payment: 
Identifying Social Risk Factors. Washington, DC: The National 
Academies Press.
---------------------------------------------------------------------------

    In an effort to standardize the submission of health literacy data 
among IRFs, HHAs, SNFs and LTCHs, for the purposes outlined in section

[[Page 39158]]

1899B(a)(1)(B) of the Act, while minimizing the reporting burden, we 
proposed to adopt SILS question described above for the Health Literacy 
data element as SPADE under the Social Determinants of Health Category. 
We proposed to add the Health Literacy data element to the IRF-PAI.
    We solicited comment on this proposals. A discussion of these 
comments, along with our responses, appears below.
    Comment: Some commenters noted that, if finalized, IRFs should only 
need to submit data on the race and ethnicity SPADEs with respect to 
admission and would not need to collect and report again at discharge, 
as it is unlikely that patient status for these elements will change. 
The commenters believe that a patient's health literacy is unlikely to 
change between admission and discharge; thus, the commenter urged CMS 
to require collection of all SDOH SPADEs with respect to admission 
only.
    Response: We disagree with the commenters that it is unlikely 
patient status for health literacy will change from admission to 
discharge. Unlike the Vision, Hearing, Race, Ethnicity, Preferred 
Language, and Interpreter Services SPADEs, we believe that the response 
to this data element may change from admission to discharge for some 
patients. Health literacy can impact a patient's ability to manage 
their conditions, and it something that should be taken into account 
when developing care plans. The collection of the Health Literacy SPADE 
at discharge is to support patients, whose circumstances may have 
changed over the duration of their admission, in having the appropriate 
supports post-discharge. Therefore, the health literacy data element 
should be collected at both admission and discharge given the impact 
this could have on health outcomes and care planning.
    Comment: One commenter stated that the health literacy question 
could be improved to capture whether the patient can read, understand, 
and implement/respond to the information. In addition, the commenter 
stated that the question does not take into account whether a patient's 
need for help is due to limited vision, which is different from the 
purpose of the separate Vision Impairment data element. Another 
possible question the commenter suggested was ``How often do you have 
difficulty?'' The commenter suggested that a single construct may not 
be sufficient for this area, depending on the aspect of health literacy 
that CMS intends to identify.
    Response: We thank the commenters for the comment on the health 
literacy data element. We agree that knowing whether a patient has a 
reading or comprehension challenge, or limited vision would be helpful. 
However, we specifically proposed data elements that have been tested. 
We were also mindful to try and limit the potential burden of asking 
additional questions related to health literacy. The SILS Health 
Literacy data element that we proposed performed well when tested, and 
it minimizes concerns related to burden by requiring one instead of 
multiple questions on health literacy.206 207 If commenters 
have examples of SDOH questions that have been cognitively tested, we 
would welcome that feedback as we seek to refine SDOH SPADE data 
elements in future rulemaking.
---------------------------------------------------------------------------

    \206\ Morris, N.S., MacLean, C.D., Chew, L.D., & Littenberg, B. 
(2006). The Single Item Literacy Screener: Evaluation of a brief 
instrument to identify limited reading ability. BMC family practice, 
7, 21. doi:10.1186/1471-2296-7-21.
    \207\ Brice, J.H., Foster, M.B., Principe, S., Moss, C., Shofer, 
F.S., Falk, R.J., Ferris, M.E., DeWalt, D.A. (2013). Single-item or 
two-item literacy screener to predict the S-TOFHLA among adult 
hemodialysis patients. Patient Educ Couns. 94(1):71-5.
---------------------------------------------------------------------------

(4) Transportation
    Transportation barriers commonly affect access to necessary health 
care, causing missed appointments, delayed care, and unfilled 
prescriptions, all of which can have a negative impact on health 
outcomes.\208\ Access to transportation for ongoing health care and 
medication access needs, particularly for those with chronic diseases, 
is essential to successful chronic disease management. Adopting a data 
element to collect and analyze information regarding transportation 
needs across PAC settings would facilitate the connection to programs 
that can address identified needs. We therefore proposed to adopt as 
SPADE a single transportation data element that is from the Protocol 
for Responding to and Assessing Patients' Assets, Risks, and 
Experiences (PRAPARE) assessment tool and currently part of the 
Accountable Health Communities (AHC) Screening Tool.
---------------------------------------------------------------------------

    \208\ Syed, S.T., Gerber, B.S., and Sharp, L.K. (2013). 
Traveling Towards Disease: Transportation Barriers to Health Care 
Access. J Community Health. 38(5): 976-993.
---------------------------------------------------------------------------

    The proposed Transportation data element from the PRAPARE tool 
questions, ``Has lack of transportation kept you from medical 
appointments, meetings, work, or from getting things needed for daily 
living?'' The three response options are: (1) Yes, it has kept me from 
medical appointments or from getting my medications; (2) Yes, it has 
kept me from non-medical meetings, appointments, work, or from getting 
things that I need; and (3) No. The patient would be given the option 
to select all responses that apply. We proposed to use the 
transportation data element from the PRAPARE Tool, with permission from 
National Association of Community Health Centers (NACHC), after 
considering research on the importance of addressing transportation 
needs as a critical SDOH.\209\
---------------------------------------------------------------------------

    \209\ Health Research & Educational Trust. (2017, November). 
Social determinants of health series: Transportation and the role of 
hospitals. Chicago, IL. Available at www.aha.org/transportation.www.aha.org/transportation.
---------------------------------------------------------------------------

    The proposed data element is responsive to research on the 
importance of addressing transportation needs as a critical SDOH and 
would adopt the Transportation item from the PRAPARE tool.\210\ This 
data element comes from the national PRAPARE social determinants of 
health assessment protocol, developed and owned by NACHC, in 
partnership with the Association of Asian Pacific Community Health 
Organization, the Oregon Primary Care Association, and the Institute 
for Alternative Futures. Similarly the Transportation data element used 
in the AHC Screening Tool was adapted from the PRAPARE tool. The AHC 
screening tool was implemented by the Center for Medicare and Medicaid 
Innovation's AHC Model and developed by a panel of interdisciplinary 
experts that looked at evidence-based ways to measure SDOH, including 
transportation. While the transportation access data element in the AHC 
screening tool serves the same purposes as our proposed SPADE 
collection about transportation barriers, the AHC tool has binary yes 
or no response options that do not differentiate between challenges for 
medical versus non-medical appointments and activities. We believe that 
this is an important nuance for informing PAC discharge planning to a 
community setting, as transportation needs for non-medical activities 
may differ than for medical activities and should be taken into 
account.\211\ We believe that use of this data element will provide 
sufficient information about transportation barriers to medical and 
non-medical care for IRF patients to facilitate appropriate discharge 
planning and care coordination across PAC settings. As such, we 
proposed to adopt the Transportation data element from PRAPARE. More 
information about

[[Page 39159]]

development of the PRAPARE tool is available on the website at https://protect2.fireeye.com/url?k=7cb6eb44-20e2f238-7cb6da7b-0cc47adc5fa2-1751cb986c8c2f8c&u=http://www.nachc.org/prapare.
---------------------------------------------------------------------------

    \210\ Health Research & Educational Trust. (2017, November). 
Social determinants of health series: Transportation and the role of 
hospitals. Chicago, IL. Available at www.aha.org/transportation.
    \211\ Northwestern University. (2017). PROMIS Item Bank v. 1.0--
Emotional Distress--Anger--Short Form 1.
---------------------------------------------------------------------------

    In addition, we received stakeholder feedback during the December 
13, 2018 SDOH listening session on the impact of transportation 
barriers on unmet care needs. While recognizing that there is no 
consensus in the field about whether providers should have 
responsibility for resolving patient transportation needs, discussion 
focused on the importance of assessing transportation barriers to 
facilitate connections with available community resources.
    Adding a Transportation data element to the collection of SPADE 
would be an important step to identifying and addressing SDOH that 
impact health outcomes and patient experience for Medicare 
beneficiaries. For more information on the Transportation data element, 
we refer readers to the document titled ``Final Specifications for IRF 
QRP Measures and Standardized Patient Assessment Data Elements,'' 
available on the website at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In an effort to standardize the submission of transportation data 
among IRFs, HHAs, SNFs and LTCHs, for the purposes outlined in section 
1899B(a)(1)(B) of the Act, while minimizing the reporting burden, we 
proposed to adopt the Transportation data element described above as 
SPADE with respect to the proposed Social Determinants of Health 
category. If finalized as proposed, we would add the Transportation 
data element to the IRF-PAI.
    We solicited comment on these proposals. A discussion of these 
comments, along with our responses, appears below.
    Comment: One commenter supported the collection of data to capture 
the reason(s) transportation affects a patient's access to health care. 
The commenter appreciated the inclusion of these items on the IRF-PAI 
and encouraged exploration of quality measures in this area as 
transportation is an extremely important instrumental activity of daily 
living to effectively transition to the community.
    Response: We thank the commenter and we will consider this feedback 
as we continue to improve and refine the SPADEs.
    Comment: Some commenters noted that, if finalized, IRFs should only 
need to submit data on the race and ethnicity SPADEs with respect to 
admission and would not need to collect and report again at discharge, 
as it is unlikely that patient status for these elements will change. 
The commenters believe that a patient's access to transportation is 
unlikely to change between admission and discharge; thus, the commenter 
suggested CMS to require collection of all SDOH SPADEs with respect to 
admission only.
    Response: We disagree with the commenters that stated that access 
to transportation will always be the same from admission to discharge. 
Unlike the Vision, Hearing, Race, Ethnicity, Preferred Language, and 
Interpreter Services SPADEs, we believe that the response to this data 
element is likely to change from admission to discharge for some 
patients. For example, a patient could lose a family member or 
caregiver between admission and discharge, which could impact his or 
her access to transportation and impact how the patient responds to the 
access to transportation SPADE data element. Therefore, we believe that 
the response to this SDOH data element is likely to change from 
admission to discharge for some patients and we proposed to collect 
this SPADE data element with respect to both admission and discharge.
    As outlined in the FY 2020 IRF PPS proposed rule, multiple studies 
have demonstrated that access to transportation has an impact on the 
health of patients (84 FR 17325). Therefore, it is important for 
providers to be able to identify a patient's needs when the patient is 
admitted and when the patient is discharged in order to better inform 
the patient's care decisions made during and after the stay, including 
understanding the patient's unique risk factors and treatment 
preferences. Because of this, we are requiring that the Access to 
Transportation data element be assessed with respect to both admission 
and discharge.
(5) Social Isolation
    Distinct from loneliness, social isolation refers to an actual or 
perceived lack of contact with other people, such as living alone or 
residing in a remote area.212 213 Social isolation tends to 
increase with age, is a risk factor for physical and mental illness, 
and a predictor of mortality.214 215 216 PAC providers are 
well-suited to design and implement programs to increase social 
engagement of patients, while also taking into account individual needs 
and preferences. Adopting a data element to collect and analyze 
information about social isolation in IRFs and across PAC settings 
would facilitate the identification of patients who are socially 
isolated and who may benefit from engagement efforts.
---------------------------------------------------------------------------

    \212\ Tomaka, J., Thompson, S., and Palacios, R. (2006). The 
Relation of Social Isolation, Loneliness, and Social Support to 
Disease Outcomes Among the Elderly. J of Aging and Health. 18(3): 
359-384.
    \213\ Social Connectedness and Engagement Technology for Long-
Term and Post-Acute Care: A Primer and Provider Selection Guide. 
(2019). Leading Age. Available at https://www.leadingage.org/white-papers/social-connectedness-and-engagement-technology-long-term-and-post-acute-care-primer-and#1.1.
    \214\ Landeiro, F., Barrows, P., Nuttall Musson, E., Gray, A.M., 
and Leal, J. (2017). Reducing Social Loneliness in Older People: A 
Systematic Review Protocol. BMJ Open. 7(5): e013778.
    \215\ Ong, A.D., Uchino, B.N., and Wethington, E. (2016). 
Loneliness and Health in Older Adults: A Mini-Review and Synthesis. 
Gerontology. 62:443-449.
    \216\ Leigh-Hunt, N., Bagguley, D., Bash, K., Turner, V., 
Turnbull, S., Valtorta, N., and Caan, W. (2017). An overview of 
systematic reviews on the public health consequences of social 
isolation and loneliness. Public Health. 152:157-171.
---------------------------------------------------------------------------

    We proposed to adopt as SPADE a single social isolation data 
element that is currently part of the AHC Screening Tool. The AHC item 
was selected from the Patient-Reported Outcomes Measurement Information 
System (PROMIS[supreg]) Item Bank on Emotional Distress and questions, 
``How often do you feel lonely or isolated from those around you?'' The 
five response options are: (1) Never; (2) Rarely; (3) Sometimes; (4) 
Often; and (5) Always.\217\ The AHC Screening Tool was developed by a 
panel of interdisciplinary experts that looked at evidence-based ways 
to measure SDOH, including social isolation. More information about the 
AHC Screening Tool is available on the website at https://innovation.cms.gov/Files/worksheets/ahcm-screeningtool.pdf.
---------------------------------------------------------------------------

    \217\ Northwestern University. (2017). PROMIS Item Bank v. 1.0--
Emotional Distress--Anger--Short Form 1.
---------------------------------------------------------------------------

    In addition, we received stakeholder feedback during the December 
13, 2018 SDOH listening session on the value of receiving information 
on social isolation for purposes of care planning. Some stakeholders 
also recommended assessing social isolation as an SDOH as opposed to 
social support.
    The proposed Social Isolation data element is consistent with NASEM 
considerations about social isolation as a function of social 
relationships that impacts health outcomes and increases mortality 
risk, as well as the current work of a NASEM committee examining how 
social isolation and loneliness

[[Page 39160]]

impact health outcomes in adults 50 years and older. We believe that 
adding a Social Isolation data element would be an important component 
of better understanding patient complexity and the care goals of 
patients, thereby facilitating care coordination and continuity in care 
planning across PAC settings. For more information on the Social 
Isolation data element, we refer readers to the document titled ``Final 
Specifications for IRF QRP Measures and Standardized Patient Assessment 
Data Elements,'' available on the website at https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/IMPACT-Act-of-2014/IMPACT-Act-Downloads-and-Videos.html.
    In an effort to standardize the submission of social isolation data 
among IRFs, HHAs, SNFs and LTCHs, for the purposes outlined in section 
1899B(a)(1)(B) of the Act, while minimizing the reporting burden, we 
proposed to adopt the Social Isolation data element described above as 
SPADE with respect to the proposed Social Determinants of Health 
category. We proposed to add the Social Isolation data element to the 
IRF-PAI.
    We sought public comment on this proposal. A discussion of these 
comments, along with our responses, appears below.
    Comment: Commenters agreed with CMS that SDOH data could provide 
Medicare with valuable information about the role that non-clinical 
factors play in PAC patient outcomes and that the addition of the SDOH 
SPADEs will facilitate communication between PAC settings and other 
health care providers. A commenter noted that common standards and 
definitions are important for interoperability and communication across 
providers and encouraged CMS to ensure that the SDOH elements collected 
in IRF settings are aligned with future proposed SDOH data collection 
requirements in other settings. One commenter stated that there is 
increasing attention on the critical role that social factors play in 
individual and population health and that addressing health-related 
social needs through enhanced clinical-community linkages can improve 
health outcomes and reduce costs. Another commenter was also pleased 
that CMS is looking at SDOH and believes it is a positive step toward 
identifying disparities in health care.
    Response: We thank the commenters for the comments.
    Comment: Some commenters noted that, if finalized, IRFs should only 
need to submit data on the race and ethnicity SPADEs with respect to 
admission and would not need to collect and report again at discharge, 
as it is unlikely that patient status for these elements will change. 
The commenters believe that a patient's response to social isolation is 
unlikely to change between admission and discharge; thus, the commenter 
suggested CMS to require collection of all SDOH SPADEs with respect to 
admission only.
    Response: We disagree with the commenters that stated that the 
response to the Social Isolation data element will be the same from 
admission to discharge. Unlike the Vision, Hearing, Race, Ethnicity, 
Preferred Language, and Interpreter Services SPADEs, we believe that 
the response to this data element is likely to change from admission to 
discharge for some patients. For example, a patient could lose a family 
member or caregiver between admission and discharge, which could impact 
their response to the Social Isolation data element. Therefore, we 
proposed to collect this SPADE data element with respect to both 
admission and discharge. As outlined in the FY 2020 IRF PPS proposed 
rule, multiple studies have demonstrated that social isolation has an 
impact on the health of patients (84 FR 17325 through 17326). 
Therefore, it is important for providers to be able to identify a 
patient's needs when the patient is admitted and when the patient is 
discharged in order to better inform the patient's care decisions made 
during and after the stay, including understanding the patient's unique 
risk factors and treatment preferences. Because of this, we are 
requiring that the Social Isolation data element be assessed at both 
admission and discharge.
    Comment: One commenter stated that the proposed question on social 
isolation may have a very different answer based on the time horizon 
considered by the beneficiary as beneficiaries who are newly admitted 
to an IRF may have experienced differing levels of social isolation 
over the preceding week due to interactions with health care providers, 
emergency providers, and friends or family visiting due to 
hospitalization. The commenter believes this question could be improved 
by adding a timeframe to the question. For example, ``How often have 
you felt lonely or isolated from those around you in the past 6 
months?''
    Response: We thank the commenter for this comment. The Social 
Isolation data element assesses whether a patient has experienced 
social isolation in the past 6 months to a year. The social isolation 
question proposed is currently part of the Accountable Health 
Communities (AHC) Screening Tool. The AHC item was selected from the 
Patient-Reported Outcomes Measurement Information System 
(PROMIS[supreg]) Item Bank on Emotional Distress.
    Comment: A commenter suggested that collecting SDOH SPADEs that 
have no clinical value, such as transportation and social isolation 
during an assigned period of either admission or discharge, is a 
significant concern. The commenter stated that at admission, the focus 
should be on assessing the patient's medical needs and plan of care, 
and at discharge, the focus shifts to patient's transition plan and 
caregiver education. As there are already multiple required assessments 
on the IRF-PAI, the SDOH SPADEs would add burden and recommended that 
any SDOH SPADEs finalized should be assessed at any point during the 
stay.
    Response: We disagree with the commenters that the Social Isolation 
and Transportation data elements have no value. As proposed in the 
transportation and social isolation section, multiple studies have 
demonstrated that access to transportation and social isolation have an 
impact on the health of patients.218 219 For example, access 
to transportation is important to medication access. Similarly, social 
isolation is a predictor of mortality. Therefore, it is important for 
providers to identify a patient's needs both at admission and discharge 
in order to better inform the patient's care decisions made during and 
after the stay, including a patient's unique risk factors and treatment 
preferences. To minimize burden, we proposed to collect this data 
element with respect to admission and discharge, rather than more 
frequently.
---------------------------------------------------------------------------

    \218\ Syed, S.T., Gerber, B.S., and Sharp, L.K. (2013). 
Traveling Towards Disease: Transportation Barriers to Health Care 
Access. J Community Health. 38(5): 976-993.
    \219\ Leigh-Hunt, N., Bagguley, D., Bash, K., Turner, V., 
Turnbull, S., Valtorta, N., and Caan, W. (2017). An overview of 
systematic reviews on the public health consequences of social 
isolation and loneliness. Public Health. 152:157-171.
---------------------------------------------------------------------------

    After consideration of the public comments, we are finalizing our 
proposals to collect SDOH data for the purposes of section 2(d)(2)(B) 
of the IMPACT Act and section 1899B(b)(1)(B)(vi) of the Act as follows. 
With regard to Race, Ethnicity, Health Literacy, Transportation, and 
Social Isolation, we are finalizing our proposals as proposed. In 
response to stakeholder comments, we are revising our proposed policies 
and finalizing

[[Page 39161]]

that IRFs that submit the Preferred Language and Interpreter Services 
SPADEs with respect to admission will be deemed to have submitted with 
respect to both admission and discharge.

H. Form, Manner, and Timing of Data Submission Under the IRF QRP

1. Background
    We refer readers to Sec.  412.634(b) for information regarding the 
current policies for reporting IRF QRP data.
2. Update to the CMS System for Reporting Quality Measures and 
Standardized Patient Assessment Data and Associated Procedural 
Proposals
    IRFs are currently required to submit IRF-PAI data to CMS using the 
Quality Improvement and Evaluation System (QIES) Assessment and 
Submission Processing (ASAP) system. We will be migrating to a new 
internet Quality Improvement and Evaluation System (iQIES) that will 
enable real-time upgrades, and we proposed to designate that system as 
the data submission system for the IRF QRP beginning October 1, 2019. 
We proposed to revise Sec.  412.634(a)(1) by replacing ``Certification 
and Survey Provider Enhanced Reports (CASPER)'' with ``CMS designated 
data submission''. We proposed to revise Sec.  412.634(d)(1) by 
replacing the reference to ``Quality Improvement and Evaluation System 
Assessment Submission and Processing (QIES ASAP) system'' with ``CMS 
designated data submission system''. We proposed to revise Sec.  
412.634(d)(5) by replacing reference to the ``QIES ASAP'' with ``CMS 
designated data submission''. We proposed to revise Sec.  412.634(f)(1) 
by replacing ``QIES'' with ``CMS designated data submission system''. 
In addition, we proposed to notify the public of any future changes to 
the CMS designated system using subregulatory mechanisms, such as 
website postings, listserv messaging, and webinars.
    We invited public comment on our proposals.
    Comment: One commenter supported this proposal and recommended that 
CMS begin educating and preparing IRFs for the transition as soon as 
possible.
    Response: We thank the commenter for their support and appreciate 
the importance of educating for this transition. Information regarding 
the transition to iQIES and instructions for onboarding has been 
provided to IRFs and will be ongoing. Training resources are currently 
available on You-Tube at https://go.cms.gov/iQIES_Training and 
additional help content for users is available within iQIES. Ongoing 
technical support via email is also available at [email protected].
    After consideration of the public comments, we are finalizing our 
proposal to revise Sec.  412.634(a)(1), Sec.  412.634(d)(1), Sec.  
412.634(d)(5), and Sec.  412.634(f)(1) as proposed. We are also 
finalizing our proposal to notify the public of any future changes to 
the CMS designated system using subregulatory mechanisms, such as 
website postings, listserv messaging, and webinars.
3. Schedule for Reporting the Transfer of Health Information Quality 
Measures Beginning With the FY 2022 IRF QRP
    As discussed in section VIII.D. of this final rule, we proposed to 
adopt the Transfer of Health Information to the Provider--Post-Acute 
Care (PAC) and Transfer of Health Information to the Patient--Post-
Acute Care (PAC) quality measures beginning with the FY 2022 IRF QRP. 
We also proposed that IRFs would report the data on those measures 
using the IRF-PAI. IRFs would be required to collect data on both 
measures for Medicare Part A and Medicare Advantage patients beginning 
with patients discharged on or after October 1, 2020. We refer readers 
to the FY 2018 IRF PPS final rule (82 FR 36291 through 36292) for the 
data collection and submission timeframes that we finalized for the IRF 
QRP.
    We sought public comment on this proposal and did not receive any 
comments.
    We are finalizing our proposal that IRFs report the data on 
Transfer of Health Information to the Provider--Post-Acute Care (PAC) 
and Transfer of Health Information to the Patient--Post-Acute Care 
(PAC) quality measures using the IRF-PAI as proposed. IRFs will be 
required to collect data on both measures for Medicare Part A and 
Medicare Advantage patients beginning with patients discharged on or 
after October 1, 2020.
4. Schedule for Reporting Standardized Patient Assessment Data Elements 
Beginning With the FY 2022 IRF QRP
    As discussed in section IV.F. of the proposed rule, we proposed to 
adopt SPADEs beginning with the FY 2022 IRF QRP. We proposed that IRFs 
would report the data using the IRF-PAI. Similar to the proposed 
schedule for reporting the Transfer of Health Information to the 
Provider--Post-Acute Care (PAC) and Transfer of Health Information to 
the Patient--Post-Acute Care (PAC) quality measures, IRFs would be 
required to collect the SPADEs for all Medicare Part A and Medicare 
Advantage patients discharged on or after October 1, 2020, at both 
admission and discharge. IRFs that submit data with respect to 
admission for the Hearing, Vision, Race, and Ethnicity SPADEs would be 
considered to have submitted data with respect to discharges. We refer 
readers to the FY 2018 IRF PPS final rule (82 FR 36291 through 36292) 
for the data collection and submission timeframes that we finalized for 
the IRF QRP.
    We sought public comment on this proposal and did not receive any 
comments.
    We are finalizing our proposal that IRFs must submit the SPADEs for 
all Medicare Part A and Medicare Advantage patients discharged on or 
after October 1, 2020, with respect to both admission and discharge, 
using the IRF-PAI. IRFs that submit data with respect to admission for 
the Hearing, Vision, Preferred Language, Interpreter Services, Race, 
and Ethnicity SPADEs will be considered to have submitted data with 
respect to discharges.
5. Data Reporting on Patients for the IRF Quality Reporting Program 
Beginning With the FY 2022 IRF QRP
    We received public input suggesting that the quality measures used 
in the IRF QRP should be calculated using data collected from all IRF 
patients, regardless of the patients' payer. This input was provided to 
us via comments requested about quality measure development on the CMS 
Measures Management System Blueprint website,\220\ as well as through 
comments we received from stakeholders via our IRF QRP mailbox, and 
feedback received from the NQF-convened MAP as part of their 
recommendations on Coordination Strategy for Post-Acute Care and Long-
Term Care Performance Measurement.\221\ Further, in the FY 2018 IRF PPS 
proposed rule (82 FR 20740), we sought input on expanding the reporting 
of quality measures to include all patients, regardless of payer, so as 
to ensure that the IRF QRP makes publicly available information 
regarding the quality of the services furnished to the IRF population 
as a whole, rather

[[Page 39162]]

than just those patients who have Medicare.
---------------------------------------------------------------------------

    \220\ Public Comment Summary Report Posting for Transfer of 
Health Information and Care Preferences. https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Post-Acute-Care-Quality-Initiatives/Downloads/Development-of-Cross-Setting-Transfer-of-Health-Information-Quality-Meas.pdf.
    \221\ MAP Coordination Strategy for Post-Acute Care and Long-
Term Care Performance Measurement. Feb 2012. http://www.qualityforum.org/Publications/2012/02/MAP_Coordination_Strategy_for_Post-Acute_Care_and_Long-Term_Care_Performance_Measurement.aspx.
---------------------------------------------------------------------------

    In response to that request for public input, several commenters, 
including MedPAC, submitted comments stating that they would be 
supportive of an effort to collect data specified under the IRF QRP 
from all IRF patients regardless of their payer. Many commenters noted 
that this would not be overly burdensome, as most of their 
organizations' members currently complete the IRF-PAI on all patients, 
regardless of their payer. A few commenters had concerns, including 
recommending that CMS continue to align the patient assessment 
instruments across PAC settings and whether the use of the data would 
outweigh any additional reporting burden. For a more detailed 
discussion, we refer readers to the FY 2018 IRF final rule (82 FR 
36292). We have taken these concerns under consideration in proposing 
this policy.
    Further, given that we do not have access to other payer claims, we 
believe that the most accurate representation of the quality provided 
in IRFs would be best conveyed using data collected via the IRF-PAI on 
all IRF patients, regardless of payer, for the purposes of the IRF QRP. 
Medicare is the primary payer for approximately 60 percent of IRF 
patients.\222\
---------------------------------------------------------------------------

    \222\ National Academies of Sciences, Engineering, and Medicine. 
2016. Accounting for social risk factors in Medicare payment: 
Identifying social risk factors. Washington, DC: The National 
Acadiemies Press.
---------------------------------------------------------------------------

    We also believe that data reporting on standardized patient 
assessment data elements using IRF-PAI should include all IRF patients 
for the same reasons for collecting data on all residents for the IRF 
QRP's quality measures: To promote higher quality and more efficient 
health care for Medicare beneficiaries and all patients receiving IRF 
services, for example through the exchange of information and 
longitudinal analysis of the data. With that, we believe that 
collecting quality measure and standardized patient assessment data via 
the IRF-PAI on all IRF patients ensures that quality care is provided 
for Medicare beneficiaries, and patients receiving IRF services as a 
whole. While we appreciate that collecting quality data on all patients 
regardless of payer may create additional burden, we also note that the 
effort to separate out Medicare beneficiaries from other patients is 
also burdensome.
    Collecting data on all IRF patients will provide us with the most 
robust, accurate reflection of the quality of care delivered to 
Medicare beneficiaries as compared with non-Medicare patients and 
residents, and we intend to display the calculation of this data on IRF 
Compare in the future. Accordingly, we proposed that IRFs collect data 
on all IRF patients to ensure that all patients, regardless of their 
payer, are receiving the same care and that provider metrics measure 
performance across the spectrum of patients.
    Therefore, to meet the quality reporting requirements for IRFs for 
the FY 2022 payment determination and each subsequent year, we proposed 
to expand the reporting of IRF-PAI data used for the IRF QRP to include 
data on all patients, regardless of their payer, beginning with 
patients discharged on or after October 1, 2020 for the FY 2022 IRF QRP 
and the IRF-PAI V4.0, effective October 1, 2020.
    We sought public comment on this proposal and received several 
comments, which are discussed below.
    Comment: Many commenters, including MedPAC, supported the proposal 
to expand the reporting of quality measures to all patients regardless 
of payer, agreeing that quality care should be a goal for all patients. 
Several commenters agreed that most providers already complete an IRF-
PAI for all patients. MedPAC also cautioned that any future Medicare 
payment adjustments related to performance should be based only on 
outcomes for Medicare beneficiaries. One commenter stated that this 
approach is consistent with other quality programs and offers consumers 
a fuller picture of quality of care. One commenter recommended 
including quality data about all payers on IRF Compare, and another 
commenter supported the proposal but suggested CMS to allow adequate 
time to review and validate data before it is made public and allow 
data on IRF Compare to be analyzed by payer.
    Response: We thank commenters for their support and appreciate 
suggestions for implementing this policy.
    Comment: A few commenters requested additional details about how 
this proposal would be implemented. One commenter suggested that CMS 
verify comprehensive data submission on all patients to avoid ``cherry-
picking'' patients. A few commenters recommended that CMS delay this 
proposal and study how this additional data affects quality measure 
performance.
    Response: We appreciate the commenters' request for more details 
regarding the implementation of this proposal, how data submission will 
be verified to avoid cherry-picking, and how this data will affect 
quality measure performance. We acknowledge the commenters' concerns 
about the proposal's implementation timeline and the request to delay 
the proposal; however instead of delaying, we plan to use the comments 
received during this rulemaking cycle to bring a new all-payer policy 
proposal in the future. Therefore, after consideration of the public 
comments we received on these issues, we have decided that at this 
time, we will not finalize this proposal. We agree that it would be 
useful to assess further how to best implement the collection of data 
for all payers for the IRF QRP.
    Comment: Many commenters had concerns about the burden of 
collecting quality data on all patients regardless of payer, citing 
that it contradicted the Patients over Paperwork initiative. One 
commenter suggested that CMS make this requirement voluntary and to 
conduct an analysis on the administrative burden on IRFs. Another 
commenter suggested that the Collection of Information section should 
contain an estimate of burden required for this reporting.
    Response: We do not believe that that the intent of this policy 
contradicts the Patients over Paperwork initiative, which aims to 
simplify the documentation required for our programs. However, the all 
payer proposal would have imposed a new reporting burden on IRFs. We 
are sensitive to the issue of burden associated with data collection 
and acknowledge the commenters' concerns about the additional burden 
required to collect quality data on all patients. Although we believe 
that the reporting of all-payer data under the IRF QRP would add value 
to the program and provide a more accurate representation of the 
quality provided by IRFs, we believe we need to better quantify the new 
reporting burden on IRFs from this proposal for stakeholders to submit 
comments. Therefore, after consideration of the public comments, we 
received on these issues, we have decided that at this time, we will 
not finalize this proposal. We agree that this burden should be 
accounted for and we will estimate this burden in future rulemaking.
    Comment: One commenter questioned whether IRFs support this 
proposal. Another commenter was concerned that this proposal would add 
complexity to CMS' administration of the IRF QRP compliance 
determination process. One commenter was concerned that quality data 
would be skewed because younger, non-Medicare patients have more room 
for improvement compared to older patients.
    Response: We do not believe this will add complexity to the IRF QRP

[[Page 39163]]

compliance determination process, since adding more patients will not 
change the overall process that we follow with regard to determining 
compliance. With regard to IRF support for this proposal, we sought 
input on this topic in the FY 2018 IRF PPS proposed rule (82 FR 20740) 
and we received several supportive comments. With regard to the 
commenter's concerns that quality data would be skewed because younger 
non-Medicare patients have more room for improvement, we note that risk 
adjustment is currently used for many quality measures, including 
measures that focus on improvement, such as the functional outcome 
measures. We take patient characteristics, such as age, into 
consideration when developing measures, and these are included as risk 
adjustors for the functional outcome measures.
    Comment: Several commenters did not support the proposal, citing 
concerns about patient privacy. Some commenters suggested that 
collecting quality data from non-Medicare beneficiaries would be a 
violation of the Health Insurance Portability and Accountability Act of 
1996 (HIPAA) since it is not required for reimbursement purposes. 
Another commenter was concerned that CMS' collection of, and possible 
disclosing of, sensitive health information from non-Medicare patients 
without consent may violate the Privacy Act of 1974, the E-Government 
Act of 2002, and other state level privacy acts. The commenter suggests 
amending Sec.  412.608(a) to require the clinician at the IRF to 
provide the Privacy Act Statement and other information to non-Medicare 
patients.
    Other commenters questioned how CMS would keep this non-Medicare 
data secure and were concerned that CMS could work with other payers to 
de-identify this data. A few commenters recommended informing non-
Medicare beneficiaries of this reporting and to use only de-identified 
data. A few commenters requested more details from CMS about the scope 
of data collection, including non-quality information on the IRF-PAI.
    Response: We appreciate the commenters' concerns but disagree that 
this proposal is a violation of HIPAA, Privacy Act of 1974, and e-
Government Act of 2002. IRF-PAI data is collected under an existing 
system of records notice (66 FR 56682). Any disclosure of the data will 
be made in accordance with the Privacy Act and those routine uses 
outlined in the SORN. Medicare patients are currently given a Privacy 
Act Statement and would be given to every patient under the IRF QRP. 
Section 208 of the e-Government Act of 2002 requires federal agencies 
to perform Privacy Impact Assessments when acquiring or developing new 
information technology or making substantial changes to existing 
information technology that involves the collection maintenance, or 
dissemination of information in identifiable form. Because we are not 
acquiring or developing new information technology, or making 
substantial changes to existing information technology under this 
proposal, we disagree that this policy violates the e-Government Act.
    With regard to questions about how CMS would keep data non-Medicare 
data secure, we safeguard the IRF-PAI data in a secure data system. The 
system limits data access to authorized users and monitors such users 
to ensure against unauthorized data access or disclosures. This system 
conforms to all applicable federal laws and regulations as well as 
federal government, Department of Health & Human Services (HHS), and 
CMS policies and standards as they relate to information security and 
data privacy. The applicable laws and regulations include, but are not 
limited to: The Privacy Act of 1974; the Federal Information Security 
Management Act of 2002; the Computer Fraud and Abuse Act of 1986; the 
Health Insurance Portability and Accountability Act of 1996; the E-
Government Act of 2002; the Clinger-Cohen Act of 1996; the Medicare 
Modernization Act of 2003; and the corresponding implementing 
regulations. With regard to the scope of data collection, IRFs would be 
required to submit quality measure and standardized patient assessment 
data elements required by the IRF QRP. After consideration of the 
public comments we received on these issues, we have decided that at 
this time, we will not finalize this proposal. We appreciate concerns 
raised by providers and will take them into consideration for future 
rulemaking.
    Comment: One commenter questioned whether CMS has the statutory 
authority to require IRFs to submit IRF-PAI data for the IRF QRP for 
all patients, regardless of payer, citing that it is inconsistent with 
section 1886(j)(2)(D) of the Act because data from non-Medicare IRF 
patients are not ``necessary'' for administering the IRF PPS. The 
commenter further noted that Sec.  412.604(c) currently requires IRFs 
to complete an IRF-PAI for all Medicare Part A and Part C patients that 
an IRF admits or discharges and does not address reporting for non-
Medicare patients.
    Response: We believe that we generally have authority to collect 
all payer data for the IRF QRP under section 1886(j)(7) of the Act. We 
also note that with respect to the data submitted in accordance with 
section 1886(j)(7)(F) of the Act, the statute expressly requires that 
data on quality measures specified under section 1899B(c)(1) of the Act 
be submitted using the IRF PAI, to the extent possible, and that SPADE 
required under section 1899B(b)(1) of the Act be submitted using the 
IRF PAI. No all payer data collected for the IRF QRP would be used for 
purposes of administering the IRF PPS.
    We appreciate the support offered by some commenters for our 
proposal to collect data on all IRF patients regardless of payer so as 
to ensure that the IRF QRP makes publicly available information 
regarding the quality of the services furnished to Medicare 
beneficiaries, as well as to the IRF population as a whole. However, we 
also acknowledge the concerns raised by some commenters with respect to 
the administrative challenges of implementing all payer data 
collection, the need to account for the burden related to this policy, 
as well as the need for us to provide further detail and training to 
IRFs. We continue to believe that the collection of quality data to 
include all patients would help to ensure that Medicare patients 
receive the same quality of care as other patients who are treated by 
IRFs.
    Therefore, after careful consideration of the public comments we 
received, we will not finalize the proposal to expand the reporting of 
IRF quality data to include all patients, regardless of payer, at this 
time. We plan to use the comments we received on this proposal to help 
inform a future all payer proposal.

I. Policies Regarding Public Display of Measure Data for the IRF QRP

    Section 1886(j)(7)(E) of the Act requires the Secretary to 
establish procedures for making the IRF QRP data available to the 
public after ensuring that IRFs have the opportunity to review their 
data prior to public display. Measure data are currently displayed on 
the Inpatient Rehabilitation Facility Compare website, an interactive 
web tool that assists individuals by providing information on IRF 
quality of care. For more information on IRF Compare, we refer readers 
to the website at https://www.medicare.gov/inpatientrehabilitationfacilitycompare/. For a more detailed discussion 
about our

[[Page 39164]]

policies regarding public display of IRF QRP measure data and 
procedures for the opportunity to review and correct data and 
information, we refer readers to the FY 2017 IRF PPS final rule (81 FR 
52125 through 52131).
    In the proposed rule, we proposed to begin publicly displaying data 
for the Drug Regimen Review Conducted With Follow-Up for Identified 
Issues--PAC IRF QRP measure beginning CY 2020 or as soon as technically 
feasible. We finalized the Drug Regimen Review Conducted With Follow-Up 
for Identified Issues--PAC IRF QRP measure in the FY 2017 IRF PPS final 
rule (81 FR 52111 through 52116).
    Data collection for this assessment-based measure began with 
patients discharged on or after October 1, 2018. We proposed to display 
data based on four rolling quarters, initially using discharges from 
January 1, 2019 through December 31, 2019 (Quarter 1 2019 through 
Quarter 4 2019). To ensure the statistical reliability of the data, we 
proposed that we would not publicly report an IRF's performance on the 
measure if the IRF had fewer than 20 eligible cases in any four 
consecutive rolling quarters. IRFs that have fewer than 20 eligible 
cases would be distinguished with a footnote that states, ``The number 
of cases/patient stays is too small to publicly report.''
    We sought public comment on these proposals and received several, 
which are summarized below.
    Comment: Several commenters supported the proposal to begin 
publicly displaying data for the Drug Regimen Review Conducted With 
Follow-Up for Identified Issues--PAC IRF QRP measure in CY 2020 or as 
soon as technically feasible, including the exception for IRFs with 
fewer than 20 eligible cases. One commenter clarified that its support 
is contingent on the measure not utilizing performance categories.
    Response: We appreciate the commenter's support.
    After consideration of the public comments, we are finalizing our 
proposal to begin publicly displaying data for the Drug Regimen Review 
Conducted With Follow-Up for Identified Issues--PAC IRF QRP measure 
beginning CY 2020 or as soon as technically feasible.

J. Removal of the List of Compliant IRFs

    In the FY 2016 IRF PPS final rule (80 FR 47125 through 47127), we 
finalized that we would publish a list of IRFs that successfully met 
the reporting requirements for the applicable payment determination on 
the IRF QRP website and update the list on an annual basis. We have 
received feedback from stakeholders that this list offers minimal 
benefit. Although the posting of successful providers was the final 
step in the applicable payment determination process, it does not 
provide new information or clarification to the providers regarding 
their annual payment update status. Therefore, we proposed that we will 
no longer publish a list of compliant IRFs on the IRF QRP website, 
effective beginning with the FY 2020 payment determination.
    We sought public comment on this proposal and received several 
comments.
    Comment: One commenter supported this proposal, but suggested that 
CMS make this information available to stakeholders upon request in the 
interest of transparency.
    Response: We thank commenters for their support. At this time, we 
do not plan to make the list of compliant IRFs available upon request, 
in alignment with other QRPs that do not provide this list. We believe 
stakeholders can find sufficient quality information about IRFs on the 
IRF compare website.
    Comment: Several commenters did not support the proposal removal of 
the list of compliant IRFs. One commenter agreed that the list was not 
relevant to IRF providers in reviewing their own compliance status, but 
stated that it could be of interest to patients and other IRFs. Other 
commenters recommended posting the list because it is helpful for large 
health systems to quickly determine which hospitals are compliant. One 
commenter further suggested that the list continue to be posted in a 
standardized manner across the various QRPs to improve transparency.
    Response: We acknowledge commenters' concerns about removing the 
requirement to post the list of compliant IRFs. Patients and consumers 
can still find information about IRF quality on the IRF Compare 
website. We do not believe that removing this list will have a negative 
impact for IRFs, since the list does not give any new information to 
IRF providers or health providers about their own compliance status. We 
also note that other QRPs do not require posting of a list of compliant 
facilities.
    After consideration of the comments, we are finalizing our proposal 
and will no longer publish a list of compliant IRFs on the IRF QRP 
website, beginning with the FY 2020 payment determination.

K. Method for Applying the Reduction to the FY 2020 IRF Increase Factor 
for IRFs That Fail To Meet the Quality Reporting Requirements

    As previously noted, section 1886(j)(7)(A)(i) of the Act requires 
the application of a 2-percentage point reduction of the applicable 
market basket increase factor for payments for discharges occurring 
during such fiscal year for IRFs that fail to comply with the quality 
data submission requirements.
    We proposed to apply a 2-percentage point reduction to the 
applicable FY 2020 proposed market basket increase factor in 
calculating an adjusted FY 2020 proposed standard payment conversion 
factor to apply to payments for only those IRFs that failed to comply 
with the data submission requirements. As previously noted, application 
of the 2-percentage point reduction may result in an update that is 
less than 0.0 for a fiscal year and in payment rates for a fiscal year 
being less than such payment rates for the preceding fiscal year. Also, 
reporting-based reductions to the market basket increase factor will 
not be cumulative; they will only apply for the FY involved.
    We invited public comment on the proposed method for applying the 
reduction to the FY 2020 IRF increase factor for IRFs that fail to meet 
the quality reporting requirements, which are summarized below.
    Comment: Some commenters suggested that CMS provide flexibility in 
its application of the IRF QRP payment penalty for IRFs who make a 
good-faith effort to comply and submit quality reporting data.
    Response: We interpret the commenter's suggestion that we take into 
consideration case by case exceptions and apply leniency for providers 
have attempted but failed to submit their quality reporting data for 
the IRF QRP. We are unable to provide flexibility with respect to the 2 
percent payment penalty; as noted previously, section 1886(j)(7) of the 
Act requires the Secretary to reduce the annual increase factor for 
IRFs that fail to comply with the quality data submission requirements. 
While we did not seek comment on flexibilities on which the penalty is 
applied, we note that we have provided flexibility where the failure of 
the IRF to comply with the requirements of the IRF QRP stemmed from 
circumstances beyond its control. For example, we have finalized 
policies that grant exceptions or extensions for IRFs if we determine 
that a systemic problem with one of our data collection systems 
affected the ability of IRFs to submit data (79 FR 45920). We have also

[[Page 39165]]

adopted policies (78 FR 47920) that allow us to grant exemptions or 
extensions to an IRF if it has experienced an extraordinary 
circumstance beyond its control. In addition, we set the reporting 
compliance threshold at 95 percent rather than at 100 percent to data 
to for account for the rare instances when assessment data collection 
and submission maybe impossible, such as when patients have been 
discharged emergently, or against medical advice.
    Table 18 shows the calculation of the adjusted FY 2020 standard 
payment conversion factor that will be used to compute IRF PPS payment 
rates for any IRF that failed to meet the quality reporting 
requirements for the applicable reporting period.
[GRAPHIC] [TIFF OMITTED] TR08AU19.022

    After consideration of the comments, we are finalizing our proposal 
to apply a 2-percentage point reduction to the applicable FY 2020 
proposed market basket increase factor in calculating an adjusted FY 
2020 proposed standard payment conversion factor to apply to payments 
for only those IRFs that failed to comply with the data submission 
requirements.

X. Miscellaneous Comments

    We received several comments that were outside the scope of the FY 
2020 IRF PPS proposed rule. Specifically, we received comments 
regarding the processes for updating the IRF facility-level adjustment 
factors and the transparency of these updates, the application of a 
cost-of-living adjustment for IRFs located in Alaska and Hawaii, the 
need for CMS education and instruction on the appropriate IGC/ICD 
coding on the IRF-PAI, re-evaluating and phasing out the 60 percent 
rule as criteria for IRF admission, and federal funding for universal 
health care. We thank commenters for bringing these issues to our 
attention, and we will take these comments into consideration for 
potential policy refinements.

XI. Provisions of the Final Regulations

    In this final rule, we are adopting the provisions set forth in the 
FY 2020 IRF PPS proposed rule (84 FR 17244).
    Specifically:
     We will adopt an unweighted motor score to assign patients 
to CMGs, the removal of one item from the score, and revisions to the 
CMGs beginning on October 1, 2019, based on analysis of 2 years of data 
(FYs 2017 and 2018) using the Quality Indicator items in the IRF-PAI. 
This includes revisions to the CMG relative weights and average LOS 
values for FY 2020, in a budget neutral manner, as discussed in section 
IV. of this final rule.
     We will rebase and revise the IRF market basket to reflect 
a 2016 base year rather than the current 2012 base year as discussed in 
section VI. of this FY 2020 IRF PPS final rule.
     We will update the IRF PPS payment rates for FY 2020 by 
the market basket increase factor, based upon the most current data 
available, with a productivity adjustment required by section 
1886(j)(3)(C)(ii)(I) of the Act, as described in section VI. of this 
final rule.
     We will update to the IRF wage index to use the concurrent 
FY IPPS wage index and the FY 2020 labor-related share in a budget-
neutral manner, as described in section VI. of this final rule.
     The facility-level adjustments will remain frozen at the 
FY 2014 levels for FY 2015 and all subsequent years, as discussed in 
section V. of this final rule.
     We will calculate the final IRF standard payment 
conversion factor for FY 2020, as discussed in section VI. of this 
final rule.
     We will update the outlier threshold amount for FY 2020, 
as discussed in section VII. of this final rule.
     We will update the CCR ceiling and urban/rural average 
CCRs for FY 2020, as discussed in section VII. of this final rule.
     We will amend the regulations at Sec.  412.622 to clarify 
that the determination as to whether a physician qualifies as a 
rehabilitation physician (that is, a licensed physician with 
specialized training and experience in inpatient rehabilitation) is 
made by the IRF, as discussed in section VIII. of this final rule.
     We will adopt updates requirements to the IRF QRP, as 
discussed in section IX. of this final rule.

XII. Collection of Information Requirements

A. Statutory Requirement for Solicitation of Comments

    Under the Paperwork Reduction Act of 1995 (PRA), we are required to 
provide 60-day notice in the Federal Register and solicit public 
comment before a collection of information requirement is submitted to 
the OMB for review and approval. To fairly evaluate whether an 
information collection should be approved by OMB, section 3506(c)(2)(A) 
of the PRA requires that we solicit comment on the following issues:
     The need for the information collection and its usefulness 
in carrying out the proper functions of our agency.
     The accuracy of our estimate of the information collection 
burden.
     The quality, utility, and clarity of the information to be 
collected.
     Recommendations to minimize the information collection 
burden on the affected public, including automated collection 
techniques.
    This final rule makes reference to associated information 
collections that are not discussed in the regulation text contained in 
this document.

B. Collection of Information Requirements for Updates Related to the 
IRF QRP

    An IRF that does not meet the requirements of the IRF QRP for a 
fiscal year will receive a 2 percentage point reduction to its 
otherwise applicable

[[Page 39166]]

annual increase factor for that fiscal year. Information is not 
currently available to determine the precise number of IRFs that will 
receive less than the full annual increase factor for FY 2020 due to 
non-compliance with the requirements of the IRF QRP.
    We believe that the burden associated with the IRF QRP is the time 
and effort associated with complying with the requirements of the IRF 
QRP. As of July 15, 2019, there are approximately 1,122 IRFs reporting 
quality data to CMS. For the purposes of calculating the costs 
associated with the collection of information requirements, we obtained 
mean hourly wages for these staff from the U.S. Bureau of Labor 
Statistics' May 2018 National Occupational Employment and Wage 
Estimates (http://www.bls.gov/oes/current/oes_nat.htm). To account for 
overhead and fringe benefits, we have doubled the hourly wage. These 
amounts are detailed in Table 19.
[GRAPHIC] [TIFF OMITTED] TR08AU19.023

    As discussed in section VIII.D. of this final rule, we are adopting 
two new measures, (1) Transfer of Health Information to the Provider--
Post-Acute Care (PAC); and (2) Transfer of Health Information to the 
Patient--Post-Acute Care (PAC), beginning with the FY 2022 IRF QRP. As 
a result, the estimated burden and cost for IRFs for complying with 
requirements of the FY 2022 IRF QRP will increase. Specifically, we 
believe that there will be a 1.2 minute addition in clinical staff time 
to report data per patient stay. We estimate 411,622 discharges from 
1,122 IRFs annually. This equates to an increase of 8,232 hours in 
burden for all IRFs (0.02 hours per assessment x 411,622 discharges). 
Given 0.7 minutes of RN time at $70.72 per hour and 0.5 minutes of LVN 
time at $43.96 per hour, we estimate that the total cost will be 
increased by $437 per IRF annually, or $490,314 for all IRFs annually. 
This increase in burden will be accounted for in the information 
collection under OMB control number (0938-0842), which expires December 
31, 2021.
    In addition, we are finalizing our proposal to add the standardized 
patient assessment data elements described in section VIII.F of this 
final rule beginning with the FY 2022 IRF QRP. As a result, the 
estimated burden and cost for IRFs for complying with requirements of 
the FY 2022 IRF QRP will be increased. Specifically, we believe that 
there will be an addition of 7.8 minutes on admission, and 10.95 
minutes on discharge, for a total of 18.8 minutes of additional 
clinical staff time to report data per patient stay. Note that this is 
a decrease from the proposed 11.1 minutes at discharge because of the 
changes in section XIII.G.4.2 of this final rule. We estimate 411,622 
discharges from 1,122 IRFs annually. This equates to an increase of 
122,995 hours in burden for all IRFs (0.3 hours per assessment x 
409,982 discharges). Given 11.3 minutes of RN time at $70.72 per hour 
and 7.5 minutes of LVN time at $43.96 per hour, we estimate that the 
total cost will be increased by $6,902 per IRF annually, or $7,744,044 
for all IRFs. This increase in burden will be accounted for in the 
information collection under OMB control number (0938-0842), which 
expires December 31, 2021.
    In summary, the newly adopted IRF QRP quality measures and 
standardized patient assessment data elements will result in a burden 
addition of $7,339 per IRF annually, and $8,234,450 for all IRFs 
annually.

XIII. Regulatory Impact Analysis

A. Statement of Need

    This final rule updates the IRF prospective payment rates for FY 
2020 as required under section 1886(j)(3)(C) of the Act. It responds to 
section 1886(j)(5) of the Act, which requires the Secretary to publish 
in the Federal Register on or before the August 1 that precedes the 
start of each fiscal year, the classification and weighting factors for 
the IRF PPS's CMGs, and a description of the methodology and data used 
in computing the prospective payment rates for that fiscal year.
    This final rule also implements sections 1886(j)(3)(C) of the Act. 
Section 1886(j)(3)(C)(ii)(I) of the Act requires the Secretary to apply 
a MFP adjustment to the market basket increase factor. The productivity 
adjustment applies to FYs from 2012 forward.
    Furthermore, this final rule also adopts policy changes under the 
statutory discretion afforded to the Secretary under section 1886(j)(7) 
of the Act. Specifically, we are rebasing and revising the IRF market 
basket to reflect a 2016 base year rather than the current 2012 base 
year, revising the CMGs, making a technical correction to the 
regulatory language to indicate that the determination of whether a 
treating physician has specialized training and experience in inpatient 
rehabilitation is made by the IRF and updating regulatory language 
related to IRF QRP data collection.

B. Overall Impact

    We have examined the impacts of this rule as required by Executive 
Order 12866 on Regulatory Planning and Review (September 30, 1993), 
Executive Order 13563 on Improving Regulation and Regulatory Review 
(January 18, 2011), the Regulatory Flexibility Act (RFA) (September 19, 
1980, Pub. L. 96-354), section 1102(b) of the Act, section 202 of the 
Unfunded Mandates Reform Act of 1995 (March 22, 1995; Pub. L. 104-4), 
Executive Order 13132 on Federalism (August 4, 1999), the Congressional 
Review Act (5 U.S.C. 804(2) and Executive Order 13771 on Reducing 
Regulation and Controlling Regulatory Costs (January 30, 2017).
    Executive Orders 12866 and 13563 direct agencies to assess all 
costs and benefits of available regulatory alternatives and, if 
regulation is necessary, to select regulatory approaches that maximize 
net benefits (including potential economic, environmental, public 
health and safety effects, distributive impacts, and equity). Section 
3(f) of Executive Order 12866 defines a ``significant regulatory 
action'' as an action that is likely to result in a rule: (1) Having an 
annual effect on the economy of $100 million or more in any 1 year, or 
adversely and materially affecting a sector of the economy, 
productivity, competition, jobs, the environment, public health or 
safety, or state, local or tribal governments or communities (also

[[Page 39167]]

referred to as ``economically significant''); (2) creating a serious 
inconsistency or otherwise interfering with an action taken or planned 
by another agency; (3) materially altering the budgetary impacts of 
entitlement grants, user fees, or loan programs or the rights and 
obligations of recipients thereof; or (4) raising novel legal or policy 
issues arising out of legal mandates, the President's priorities, or 
the principles set forth in the Executive Order.
    A regulatory impact analysis (RIA) must be prepared for major rules 
with economically significant effects ($100 million or more in any 1 
year). We estimate the total impact of the policy updates described in 
this final rule by comparing the estimated payments in FY 2020 with 
those in FY 2019. This analysis results in an estimated $210 million 
increase for FY 2020 IRF PPS payments. Additionally we estimate that 
costs associated with the proposals to update the reporting 
requirements under the IRF QRP result in an estimated $8.2 million 
addition in costs in FY 2020 for IRFs. We estimate that this rulemaking 
is ``economically significant'' as measured by the $100 million 
threshold, and hence also a major rule under the Congressional Review 
Act. Also, the rule has been reviewed by OMB. Accordingly, we have 
prepared a Regulatory Impact Analysis that, to the best of our ability, 
presents the costs and benefits of the rulemaking.

C. Anticipated Effects

1. Effects on IRFs
    The RFA requires agencies to analyze options for regulatory relief 
of small entities, if a rule has a significant impact on a substantial 
number of small entities. For purposes of the RFA, small entities 
include small businesses, nonprofit organizations, and small 
governmental jurisdictions. Most IRFs and most other providers and 
suppliers are small entities, either by having revenues of $7.5 million 
to $38.5 million or less in any 1 year depending on industry 
classification, or by being nonprofit organizations that are not 
dominant in their markets. (For details, see the Small Business 
Administration's final rule that set forth size standards for health 
care industries, at 65 FR 69432 at http://www.sba.gov/sites/default/files/files/Size_Standards_Table.pdf, effective March 26, 2012 and 
updated on February 26, 2016.) Because we lack data on individual 
hospital receipts, we cannot determine the number of small proprietary 
IRFs or the proportion of IRFs' revenue that is derived from Medicare 
payments. Therefore, we assume that all IRFs (an approximate total of 
1,120 IRFs, of which approximately 55 percent are nonprofit facilities) 
are considered small entities and that Medicare payment constitutes the 
majority of their revenues. The HHS generally uses a revenue impact of 
3 to 5 percent as a significance threshold under the RFA. As shown in 
Table 20, we estimate that the net revenue impact of this final rule on 
all IRFs is to increase estimated payments by approximately 2.5 
percent. The rates and policies set forth in this final rule will not 
have a significant impact (not greater than 3 percent) on a substantial 
number of small entities. Medicare Administrative Contractors are not 
considered to be small entities. Individuals and states are not 
included in the definition of a small entity.
    In addition, section 1102(b) of the Act requires us to prepare a 
regulatory impact analysis if a rule may have a significant impact on 
the operations of a substantial number of small rural hospitals. This 
analysis must conform to the provisions of section 604 of the RFA. For 
purposes of section 1102(b) of the Act, we define a small rural 
hospital as a hospital that is located outside of a Metropolitan 
Statistical Area and has fewer than 100 beds. As discussed in detail 
below in this section, the rates and policies set forth in this final 
rule will not have a significant impact (not greater than 3 percent) on 
a substantial number of rural hospitals based on the data of the 136 
rural units and 11 rural hospitals in our database of 1,122 IRFs for 
which data were available.
    Section 202 of the Unfunded Mandates Reform Act of 1995 (Pub. L. 
104-04, enacted March 22, 1995) (UMRA) also requires that agencies 
assess anticipated costs and benefits before issuing any rule whose 
mandates require spending in any 1 year of $100 million in 1995 
dollars, updated annually for inflation. In 2019, that threshold is 
approximately $154 million. This final rule does not mandate any 
requirements for State, local, or tribal governments, or for the 
private sector.
    Executive Order 13132 establishes certain requirements that an 
agency must meet when it issues a proposed rule (and subsequent final 
rule) that imposes substantial direct requirement costs on state and 
local governments, preempts state law, or otherwise has federalism 
implications. As stated, this final rule will not have a substantial 
effect on state and local governments, preempt state law, or otherwise 
have a federalism implication.
    Executive Order 13771, titled Reducing Regulation and Controlling 
Regulatory Costs, was issued on January 30, 2017 and requires that the 
costs associated with significant new regulations ``shall, to the 
extent permitted by law, be offset by the elimination of existing costs 
associated with at least two prior regulations.'' This final rule is 
considered an E.O. 13771 regulatory action. We estimate that this rule 
would generate $6.18 million in annualized cost, discounted at 7 
percent relative to year 2016, over a perpetual time horizon. Details 
on the estimated costs of this rule can be found in the preceding 
analyses.
2. Detailed Economic Analysis
    This final rule updates to the IRF PPS rates contained in the FY 
2019 IRF PPS final rule (83 FR 38514). Specifically, this final rule 
updates the CMG relative weights and average LOS values, the wage 
index, and the outlier threshold for high-cost cases. This final rule 
applies a MFP adjustment to the FY 2020 IRF market basket increase 
factor in accordance with section 1886(j)(3)(C)(ii)(I) of the Act. 
Further, this final rule rebases and revises the IRF market basket to 
reflect a 2016 base year rather than the current 2012 base year, 
revises the CMGs based on FYs 2017 and 2018 data and amends the 
regulatory language to clarify that the determination of whether a 
treating physician has specialized training and experience in inpatient 
rehabilitation is made by the IRF.
    We estimate that the impact of the changes and updates described in 
this final rule will be a net estimated increase of $210 million in 
payments to IRF providers. This estimate does not include the 
implementation of the required 2 percentage point reduction of the 
market basket increase factor for any IRF that fails to meet the IRF 
quality reporting requirements (as discussed in section IX.K. of this 
final rule). The impact analysis in Table 20 of this final rule 
represents the projected effects of the updates to IRF PPS payments for 
FY 2020 compared with the estimated IRF PPS payments in FY 2019. We 
determine the effects by estimating payments while holding all other 
payment variables constant. We use the best data available, but we do 
not attempt to predict behavioral responses to these changes, and we do 
not make adjustments for future changes in such variables as number of 
discharges or case-mix.
    We note that certain events may combine to limit the scope or 
accuracy of our impact analysis, because such an analysis is future-
oriented and, thus,

[[Page 39168]]

susceptible to forecasting errors because of other changes in the 
forecasted impact time period. Some examples could be legislative 
changes made by the Congress to the Medicare program that would impact 
program funding, or changes specifically related to IRFs. Although some 
of these changes may not necessarily be specific to the IRF PPS, the 
nature of the Medicare program is such that the changes may interact, 
and the complexity of the interaction of these changes could make it 
difficult to predict accurately the full scope of the impact upon IRFs.
    In updating the rates for FY 2020, we are adopting standard annual 
revisions described in this final rule (for example, the update to the 
wage and market basket indexes used to adjust the federal rates). We 
are also implementing a productivity adjustment to the FY 2020 IRF 
market basket increase factor in accordance with section 
1886(j)(3)(C)(ii)(I) of the Act. We estimate the total increase in 
payments to IRFs in FY 2020, relative to FY 2019, will be approximately 
$210 million.
    This estimate is derived from the application of the FY 2020 IRF 
market basket increase factor, as reduced by a productivity adjustment 
in accordance with section 1886(j)(3)(C)(ii)(I) of the Act, which 
yields an estimated increase in aggregate payments to IRFs of $210 
million. Outlier payments are estimated to remain at 3 percent in FY 
2020. Therefore, we estimate that these updates will result in a net 
increase in estimated payments of $210 million from FY 2019 to FY 2020.
    The effects of the updates that impact IRF PPS payment rates are 
shown in Table 20. The following updates that affect the IRF PPS 
payment rates are discussed separately below:
     The effects of the update to the outlier threshold amount, 
from approximately 3.0 percent to 3.0 percent of total estimated 
payments for FY 2020, consistent with section 1886(j)(4) of the Act.
     The effects of the annual market basket update (using the 
IRF market basket) to IRF PPS payment rates, as required by sections 
1886(j)(3)(A)(i) and (j)(3)(C) of the Act, including a productivity 
adjustment in accordance with section 1886(j)(3)(C)(i)(I) of the Act.
     The effects of applying the budget-neutral labor-related 
share and wage index adjustment, as required under section 1886(j)(6) 
of the Act.
     The effects of the budget-neutral changes to the CMGs, 
relative weights and average LOS values, under the authority of section 
1886(j)(2)(C)(i) of the Act.
     The total change in estimated payments based on the FY 
2020 payment changes relative to the estimated FY 2019 payments.
3. Description of Table 20
    Table 20 shows the overall impact on the 1,122 IRFs included in the 
analysis.
    The next 12 rows of Table 20 contain IRFs categorized according to 
their geographic location, designation as either a freestanding 
hospital or a unit of a hospital, and by type of ownership; all urban, 
which is further divided into urban units of a hospital, urban 
freestanding hospitals, and by type of ownership; and all rural, which 
is further divided into rural units of a hospital, rural freestanding 
hospitals, and by type of ownership. There are 975 IRFs located in 
urban areas included in our analysis. Among these, there are 697 IRF 
units of hospitals located in urban areas and 278 freestanding IRF 
hospitals located in urban areas. There are 147 IRFs located in rural 
areas included in our analysis. Among these, there are 136 IRF units of 
hospitals located in rural areas and 11 freestanding IRF hospitals 
located in rural areas. There are 393 for-profit IRFs. Among these, 
there are 357 IRFs in urban areas and 36 IRFs in rural areas. There are 
616 non-profit IRFs. Among these, there are 526 urban IRFs and 90 rural 
IRFs. There are 113 government-owned IRFs. Among these, there are 92 
urban IRFs and 21 rural IRFs.
    The remaining four parts of Table 20 show IRFs grouped by their 
geographic location within a region, by teaching status, and by DSH PP. 
First, IRFs located in urban areas are categorized for their location 
within a particular one of the nine Census geographic regions. Second, 
IRFs located in rural areas are categorized for their location within a 
particular one of the nine Census geographic regions. In some cases, 
especially for rural IRFs located in the New England, Mountain, and 
Pacific regions, the number of IRFs represented is small. IRFs are then 
grouped by teaching status, including non-teaching IRFs, IRFs with an 
intern and resident to average daily census (ADC) ratio less than 10 
percent, IRFs with an intern and resident to ADC ratio greater than or 
equal to 10 percent and less than or equal to 19 percent, and IRFs with 
an intern and resident to ADC ratio greater than 19 percent. Finally, 
IRFs are grouped by DSH PP, including IRFs with zero DSH PP, IRFs with 
a DSH PP less than 5 percent, IRFs with a DSH PP between 5 and less 
than 10 percent, IRFs with a DSH PP between 10 and 20 percent, and IRFs 
with a DSH PP greater than 20 percent.
    The estimated impacts of each policy described in this rule to the 
facility categories listed are shown in the columns of Table 20. The 
description of each column is as follows:
     Column (1) shows the facility classification categories.
     Column (2) shows the number of IRFs in each category in 
our FY 2020 analysis file.
     Column (3) shows the number of cases in each category in 
our FY 2020 analysis file.
     Column (4) shows the estimated effect of the adjustment to 
the outlier threshold amount.
     Column (5) shows the estimated effect of the update to the 
IRF labor-related share and wage index, in a budget-neutral manner.
     Column (6) shows the estimated effect of the update to the 
CMGs, relative weights, and average LOS values, in a budget-neutral 
manner.
     Column (7) compares our estimates of the payments per 
discharge, incorporating all of the policies reflected in this final 
rule for FY 2020 to our estimates of payments per discharge in FY 2019.
    The average estimated increase for all IRFs is approximately 2.5 
percent. This estimated net increase includes the effects of the IRF 
market basket increase factor for FY 2020 of 2.9 percent, reduced by a 
productivity adjustment of 0.4 percentage point in accordance with 
section 1886(j)(3)(C)(ii)(I) of the Act. There is no change in 
estimated IRF outlier payments from the update to the outlier threshold 
amount. Since we are making the updates to the IRF wage index and the 
CMG relative weights in a budget-neutral manner, they will not be 
expected to affect total estimated IRF payments in the aggregate. 
However, as described in more detail in each section, they will be 
expected to affect the estimated distribution of payments among 
providers.
BILLING CODE 4120-01-P

[[Page 39169]]

[GRAPHIC] [TIFF OMITTED] TR08AU19.024


[[Page 39170]]


BILLING CODE 4120-01-C
4. Impact of the Update to the Outlier Threshold Amount
    The estimated effects of the update to the outlier threshold 
adjustment are presented in column 4 of Table 20. In the FY 2019 IRF 
PPS final rule (83 FR 38531 through 38532), we used FY 2017 IRF claims 
data (the best, most complete data available at that time) to set the 
outlier threshold amount for FY 2019 so that estimated outlier payments 
would equal 3 percent of total estimated payments for FY 2019.
    For the FY 2020 IRF PPS proposed rule (84 FR 17244), we used 
preliminary FY 2018 IRF claims data, and, based on that preliminary 
analysis, we estimated that IRF outlier payments as a percentage of 
total estimated IRF payments would be 3.2 percent in FY 2019. As we 
typically do between the proposed and final rules each year, we updated 
our FY 2018 IRF claims data to ensure that we are using the most recent 
available data in setting IRF payments. Therefore, based on updated 
analysis of the most recent IRF claims data for this final rule, we now 
estimate that IRF outlier payments as a percentage of total IRF 
payments as 3.0 in FY 2019. Thus, we are adjusting the outlier 
threshold amount in this final rule to maintain total estimated outlier 
payments equal to 3 percent of total estimated payments in FY 2020.
    The impact of this outlier adjustment update (as shown in column 4 
of Table 20) is to maintain estimated overall payments to IRFs at 3 
percent.
5. Impact of the CBSA Wage Index and Labor-Related Share
    In column 5 of Table 20, we present the effects of the budget-
neutral update of the wage index and labor-related share. The changes 
to the wage index and the labor-related share are discussed together 
because the wage index is applied to the labor-related share portion of 
payments, so the changes in the two have a combined effect on payments 
to providers. As discussed in section VI.E. of this final rule, we are 
updating the labor-related share from 70.5 percent in FY 2019 to 72.7 
percent in FY 2020.
6. Impact of the Update to the CMG Relative Weights and Average LOS 
Values
    In column 6 of Table 20, we present the effects of the budget-
neutral update of the CMGs, relative weights and average LOS values. In 
the aggregate, we do not estimate that these updates will affect 
overall estimated payments of IRFs. However, we do expect these updates 
to have small distributional effects.
7. Effects of the Requirements for the IRF QRP for FY 2020
    In accordance with section 1886(j)(7)(A) of the Act, the Secretary 
must reduce by 2 percentage points the market basket increase factor 
otherwise applicable to an IRF for a fiscal year if the IRF does not 
comply with the requirements of the IRF QRP for that fiscal year. In 
section VIII.J of this final rule, we discuss the method for applying 
the 2 percentage point reduction to IRFs that fail to meet the IRF QRP 
requirements.
    As discussed in section VIII.D. of this final rule, we are 
finalizing our proposal to add two measures to the IRF QRP: (1) 
Transfer of Health Information to the Provider--Post-Acute Care (PAC); 
and (2) Transfer of Health Information to the Patient--Post-Acute Care 
(PAC), beginning with the FY 2022 IRF QRP. We are also finalizing our 
proposal to add standardized patient assessment data elements, as 
discussed in section IV.G of this final rule. We describe the estimated 
burden and cost reductions for both of these measures in section VIII.C 
of this final rule. In summary, the changes to the IRF QRP will result 
in a burden addition of $7,339 per IRF annually, and $8,234,450 for all 
IRFs annually.
    We intend to continue to closely monitor the effects of the IRF QRP 
on IRFs and to help perpetuate successful reporting outcomes through 
ongoing stakeholder education, national trainings, IRF announcements, 
website postings, CMS Open Door Forums, and general and technical help 
desks.
8. Effects of the Amending Sec.  412.622(a)(3)(iv) To Clarify the 
Definition of a Rehabilitation Physician
    As discussed in section VIII. of this final rule, we are amending 
Sec.  412.622(a)(3)(iv) to clarify that the determination as to whether 
a physician qualifies as a rehabilitation physician (that is, a 
licensed physician with specialized training and experience in 
inpatient rehabilitation) is made by the IRF. We do not expect this to 
have any effect on the quality of care that beneficiaries receive in 
IRFs because we continue to require that the rehabilitation physicians 
caring for patients in IRFs be licensed physicians with specialized 
training and experience in inpatient rehabilitation. We expect IRFs to 
continue ensuring that the rehabilitation physicians meet these 
requirements. Although we do not currently collect data from IRFs on 
the physicians specialties that are providing care to patients in IRFs, 
we do not expect this to change as a result of the amendments we are 
making to Sec.  412.622(a)(3)(iv). However, we will continue to monitor 
the quality of care beneficiaries receive in IRFs, and will initiate 
appropriate actions through future rulemaking if we observe any 
declines in quality of care in IRFs.
    As this is merely clarifying our existing policy regarding the 
definition of a rehabilitation physician in Sec.  412.622(a)(3)(iv), we 
do not expect this to result in any financial impacts for the Medicare 
contractors, IRFs, other providers, or for the Medicare program. 
However, we expect that this clarification may ease some administrative 
burden for IRFs and for Medicare contractors by making it easier for 
IRF providers to document their decisions regarding the licensed 
physicians in their facilities that meet the regulatory definition of a 
rehabilitation physician and for the Medicare contractors to continue 
to accept the IRFs' decisions in this regard. We are unable at this 
time to quantify how much administrative burden may have existed 
because of the previous ambiguity surrounding the definition of a 
rehabilitation physician, but we are hopeful that this clarification 
will alleviate any administrative burden that might have existed 
before.
    We expect this clarification to enhance Medicare's program 
integrity efforts in this area by eliminating uncertainty surrounding 
the definition of a rehabilitation physician.

D. Alternatives Considered

    The following is a discussion of the alternatives considered for 
the IRF PPS updates contained in this final rule.
    Section 1886(j)(3)(C) of the Act requires the Secretary to update 
the IRF PPS payment rates by an increase factor that reflects changes 
over time in the prices of an appropriate mix of goods and services 
included in the covered IRF services.
    We are adopting a market basket increase factor for FY 2020 that is 
based on a rebased and revised market basket reflecting a 2016 base 
year. We considered the alternative of continuing to use the IRF market 
basket without rebasing to determine the market basket increase factor 
for FY 2020. However, we typically rebase and revise the market baskets 
for the various PPS every 4 to 5 years so that the cost weights and 
price proxies reflect more recent data. Therefore, we believe it is 
more technically appropriate to use a 2016-based IRF market basket 
since it allows for the FY 2020 market basket increase

[[Page 39171]]

factor to reflect a more up-to-date cost structure experienced by IRFs.
    As noted previously in this final rule, section 
1886(j)(3)(C)(ii)(I) of the Act requires the Secretary to apply a 
productivity adjustment to the market basket increase factor for FY 
2020. Thus, in accordance with section 1886(j)(3)(C) of the Act, we are 
updating the IRF prospective payments in this final rule by 2.5 percent 
(which equals the 2.9 percent estimated IRF market basket increase 
factor for FY 2020 reduced by a 0.4 percentage point productivity 
adjustment as determined under section 1886(b)(3)(B)(xi)(II) of the Act 
(as required by section 1886(j)(3)(C)(ii)(I) of the Act)).
    As we finalized in the FY 2019 IRF PPS final rule (83 FR 38514) use 
of the Quality Indicators items in determining payment and the 
associated CMG and CMG relative weight revisions using 2 years of data 
(FYs 2017 and 2018) beginning with FY 2020, we did not consider any 
alternative to proposing these changes.
    However, we did consider whether or not to apply a weighting 
methodology to the IRF motor score that was finalized in the FY 2019 
IRF PPS final rule (83 FR 38514) to assign patients to CMGs beginning 
in FY 2020. As described in the FY 2020 IRF PPS proposed rule (84 FR 
17244, 17249 through 17260), we explored the use of a weighted motor 
score, as requested by stakeholders. Our analysis showed that weighting 
the motor score would improve the accuracy of payments under the IRF 
PPS. The improved accuracy combined with the requests from stakeholders 
to explore a weighted methodology led us to propose to use a weighted 
motor score to assign patients to CMGs beginning on October 1, 2019. 
However, in light of the many concerned stakeholder comments on the FY 
2020 IRF PPS proposed rule that requested that we go back to an 
unweighted motor score methodology until we can more fully analyze a 
weighted motor score, the fact that the improvement in accuracy using 
the weighted motor score is small, and the greater simplicity achieved 
through the use of an unweighted motor score, we are finalizing an 
unweighted motor score, in which each of the 18 items have a weight of 
1, beginning October 1, 2019. We will continue to analyze weighted 
motor score approaches and will consider possible revisions to the 
motor score for future rulemaking.
    We considered not removing the item GG0170A1 Roll left and right 
from the composition of the motor score. However, this item was found 
to be very collinear with other items in the motor score and did not 
behave as expected in the models. Therefore, we believe it is 
appropriate to remove this item from the construction of the motor 
score.
    We considered updating facility-level adjustment factors for FY 
2020. However, as discussed in more detail in the FY 2015 final rule 
(79 FR 45872), we believe that freezing the facility-level adjustments 
at FY 2014 levels for FY 2015 and all subsequent years (unless and 
until the data indicate that they need to be further updated) will 
allow us an opportunity to monitor the effects of the substantial 
changes to the adjustment factors for FY 2014, and will allow IRFs time 
to adjust to the previous changes.
    We considered not updating the IRF wage index to use the concurrent 
fiscal year's IPPS wage index and instead continuing to use a 1-year 
lag of the IPPS wage index. However, we believe that updating the IRF 
wage index based on the concurrent fiscal year's IPPS wage index will 
better align the data across acute and PAC settings in support of our 
efforts to move toward more unified Medicare payments across PAC 
settings.
    We considered maintaining the existing outlier threshold amount for 
FY 2020. However, the outlier threshold must be adjusted to reflect 
changes in estimated costs and payments for IRFs in FY 2020. 
Consequently, we are adjusting the outlier threshold amount in this 
final rule to maintain total outlier payments equal to 3 percent of 
aggregate estimated payments in FY 2020.
    We considered not amending Sec.  412.622(a)(3)(iv) to clarify that 
the determination as to whether a physician qualifies as a 
rehabilitation physician (that is, a licensed physician with 
specialized training and experience in inpatient rehabilitation) is 
made by the IRF. Instead, we considered addressing this issue through 
subregulatory means, such as issuing guidance to the Medicare 
contractors. However, we believe that it is important to clarify this 
definition in regulation to ensure that IRF providers and Medicare 
contractors have a shared understanding of these regulatory 
requirements and to make certain that there is no room for further 
ambiguity on this point.
    In addition, we considered addressing this issue by amending Sec.  
412.622(a)(3)(iv) to add further specificity to the definition of a 
rehabilitation physician. However, we did not take this approach 
because we continue to believe that the IRFs are in the best position 
to make the determination as to which licensed physicians meet the 
requirements for purposes of Sec.  412.622, and we did not want to 
inadvertently affect access to IRF care for beneficiaries. However, we 
will continue to monitor this policy and engage with stakeholders to 
determine if further specificity of these requirements may be warranted 
in the future.

E. Regulatory Review Costs

    If regulations impose administrative costs on private entities, 
such as the time needed to read and interpret this final rule, we 
should estimate the cost associated with regulatory review. Due to the 
uncertainty involved with accurately quantifying the number of entities 
that will review the rule, we assume that the total number of unique 
commenters on the FY 2020 IRF PPS proposed rule will be the number of 
reviewers of this final rule. We acknowledge that this assumption may 
understate or overstate the costs of reviewing this final rule. It is 
possible that not all commenters reviewed the FY 2020 IRF PPS proposed 
rule in detail, and it is also possible that some reviewers chose not 
to comment on the proposed rule. For these reasons we thought that the 
number of past commenters would be a fair estimate of the number of 
reviewers of this final rule.
    We also recognize that different types of entities are in many 
cases affected by mutually exclusive sections of this final rule, and 
therefore, for the purposes of our estimate we assume that each 
reviewer reads approximately 50 percent of the rule. We sought comments 
on this assumption.
    Using the wage information from the BLS for medical and health 
service managers (Code 11-9111), we estimate that the cost of reviewing 
this rule is $107.38 per hour, including overhead and fringe benefits 
(https://www.bls.gov/oes/current/oes_nat.htm). Assuming an average 
reading speed, we estimate that it would take approximately 2 hours for 
the staff to review half of this final rule. For each IRF that reviews 
the rule, the estimated cost is $218.72 (2 hours x $109.36). Therefore, 
we estimate that the total cost of reviewing this regulation is 
$274,931.04 ($218.72 x 1,257 reviewers).
    We received one comment on the proposed methodology for estimating 
the total cost of reviewing this regulation which is summarized below.
    Comment: One commenter suggested that CMS should take into 
consideration the number of times the proposed rule has been downloaded 
in estimating the cost of reviewing this regulation.
    Response: The regulatory review cost is an estimate that makes 
several assumptions such as average reading speed and number of the 
people who

[[Page 39172]]

read the document, etc. For more than 2 years, we have used the number 
of comments received as a proxy for the number of staff members who 
review the document. This assumption is well accepted by the general 
public. The number of comments received is a more reasonable proxy than 
the number of downloads since those who provide comments must actually 
read the rule, as those that download the rule may not read the rule.

F. Accounting Statement and Table

    As required by OMB Circular A-4 (available at http://www.whitehouse.gov/sites/default/files/omb/assets/omb/circulars/a004/a-4.pdf), in Table 21, we have prepared an accounting statement showing 
the classification of the expenditures associated with the provisions 
of this final rule. Table 21 provides our best estimate of the increase 
in Medicare payments under the IRF PPS as a result of the updates 
presented in this final rule based on the data for 1,122 IRFs in our 
database. In addition, Table 21 presents the costs associated with the 
new IRF QRP requirements for FY 2020.
[GRAPHIC] [TIFF OMITTED] TR08AU19.025

G. Conclusion

    Overall, the estimated payments per discharge for IRFs in FY 2020 
are projected to increase by 2.5 percent, compared with the estimated 
payments in FY 2019, as reflected in column 7 of Table 20.
    IRF payments per discharge are estimated to increase by 2.4 percent 
in urban areas and 4.4 percent in rural areas, compared with estimated 
FY 2019 payments. Payments per discharge to rehabilitation units are 
estimated to increase 5.0 percent in urban areas and 5.7 percent in 
rural areas. Payments per discharge to freestanding rehabilitation 
hospitals are estimated to increase 0.2 percent in urban areas and 
decrease 2.1 percent in rural areas.
    Overall, IRFs are estimated to experience a net increase in 
payments as a result of the policies in this final rule. The largest 
payment increase is estimated to be a 6.8 percent increase for rural 
government IRFs and rural IRFs located in the West South Central 
region. The analysis above, together with the remainder of this 
preamble, provides a Regulatory Impact Analysis.
    In accordance with the provisions of Executive Order 12866, this 
regulation was reviewed by the Office of Management and Budget.

List of Subjects in 42 CFR Part 412

    Administrative practice and procedure, Health facilities, Medicare, 
Puerto Rico, Reporting and recordkeeping requirements.

    For the reasons set forth in the preamble, the Centers for Medicare 
& Medicaid Services amends 42 CFR chapter IV as set forth below:

PART 412--PROSPECTIVE PAYMENT SYSTEMS FOR INPATIENT HOSPITAL 
SERVICES

0
1. The authority citation for part 412 is revised to read as follows:

    Authority:  42 U.S.C. 1302 and 1395hh.

0
2. Section 412.622 is amended by revising paragraphs (a)(3)(iv), 
(a)(4)(i)(A), (a)(4)(iii)(A), and (a)(5)(i) and adding paragraph (c) to 
read as follows:


Sec.  412.622   Basis of payment.

    (a) * * *
    (3) * * *
    (iv) Requires physician supervision by a rehabilitation physician. 
The requirement for medical supervision means that the rehabilitation 
physician must conduct face-to-face visits with the patient at least 3 
days per week throughout the patient's stay in the IRF to assess the 
patient both medically and functionally, as well as to modify the 
course of treatment as needed to maximize the patient's capacity to 
benefit from the rehabilitation process. The post-admission physician 
evaluation described in paragraph (a)(4)(ii) of this section may count 
as one of the face-to-face visits.
    (4) * * *
    (i) * * *
    (A) It is conducted by a licensed or certified clinician(s) 
designated by a rehabilitation physician within the 48 hours 
immediately preceding the IRF admission. A preadmission screening that 
includes all of the required elements, but that is conducted more than 
48 hours immediately preceding the IRF admission, will be accepted as 
long as an update is conducted in person or by telephone to update the 
patient's medical and functional status within the 48 hours immediately 
preceding the IRF admission and is documented in the patient's medical 
record.
* * * * *
    (iii) * * *
    (A) It is developed by a rehabilitation physician with input from 
the interdisciplinary team within 4 days of the patient's admission to 
the IRF.
* * * * *
    (5) * * *
    (i) The team meetings are led by a rehabilitation physician and 
further consist of a registered nurse with specialized training or 
experience in rehabilitation; a social worker or case manager (or 
both); and a licensed or certified therapist from each therapy 
discipline involved in treating the patient. All team members must have 
current knowledge of the patient's medical and functional status. The 
rehabilitation physician may lead the interdisciplinary team meeting 
remotely via a mode of communication such as video or telephone 
conferencing.
* * * * *
    (c) Definitions. As used in this section--
    Rehabilitation physician means a licensed physician who is 
determined by the IRF to have specialized training and experience in 
inpatient rehabilitation.

0
3. Section 412.634 is amended by revising paragraphs (a)(1), (d)(1) and 
(5), and (f)(1) to read as follows:

[[Page 39173]]

Sec.  412.634   Requirements under the Inpatient Rehabilitation 
Facility (IRF) Quality Reporting Program (QRP).

    (a) * * *
    (1) For the FY 2018 payment determination and subsequent years, an 
IRF must begin reporting data under the IRF QRP requirements no later 
than the first day of the calendar quarter subsequent to 30 days after 
the date on its CMS Certification Number (CCN) notification letter, 
which designates the IRF as operating in the CMS designated data 
submission system.
* * * * *
    (d) * * *
    (1) IRFs that do not meet the requirement in paragraph (b) of this 
section for a program year will receive a written notification of non-
compliance through at least one of the following methods: The CMS 
designated data submission system, the United States Postal Service, or 
via an email from the Medicare Administrative Contractor (MAC).
* * * * *
    (5) CMS will notify IRFs, in writing, of its final decision 
regarding any reconsideration request through at least one of the 
following methods: CMS designated data submission system, the United 
States Postal Service, or via an email from the Medicare Administrative 
Contractor (MAC).
* * * * *
    (f) * * *
    (1) IRFs must meet or exceed two separate data completeness 
thresholds: One threshold set at 95 percent for completion of required 
quality measures data and standardized patient assessment data 
collected using the IRF-PAI submitted through the CMS designated data 
submission system; and a second threshold set at 100 percent for 
measures data collected and submitted using the CDC NHSN.
* * * * *

    Dated: July 23, 2019.
Seema Verma,
Administrator, Centers for Medicare & Medicaid Services.
    Dated: July 25, 2019.
Alex M. Azar II,
Secretary, Department of Health and Human Services.
[FR Doc. 2019-16603 Filed 7-31-19; 4:15 pm]
 BILLING CODE 4120-01-P