Medicare HMO Institutional Payments: Improved HCFA Oversight, More Recent
Cost Data Could Reduce Overpayments (Letter Report, 09/09/98,
GAO/HEHS-98-153).

Pursuant to a congressional request, GAO reviewed the Health Care
Financing Administration's (HCFA) oversight of Medicare payments to
health maintenance organizations (HMO) for institutionalized
beneficiaries, focusing on: (1) the criteria HCFA uses to determine a
beneficiary's institutional status; (2) the methods HCFA employs to
ensure that HMOs properly classify beneficiaries as institutionalized;
and (3) whether the higher capitation rate for beneficiaries who live in
institutions is justified by higher health care costs.

GAO noted that: (1) HCFA's broad definition of institution allowed HMOs
to claim institutional status for individuals residing in facilities not
likely to house sicker-than-average seniors; (2) some of the facilities
GAO visited that HMOs had classified as institutional residences
provided no medical care but rather offered a menu of recreational
activities for seniors capable of living independently; (3) HCFA acted
on GAO's findings and those of others by narrowing the definition of
eligible institutions, effective January 1, 1998; (4) even with more
stringent criteria, however, HCFA relies on the HMOs to determine which
beneficiaries qualify for institutional status; and conducts only
limited reviews to confirm the accuracy of HMO records; (5) studies by
the Department of Health and Human Services Inspector General reviewing
the accuracy of HMO institutional status data support GAO's finding that
HCFA's reviews are not adequate to detect the extent of errors or
overpayments resulting from HMOs' misclassification of beneficiaries;
(6) the task of ensuring accurate data may be further complicated by
HCFA's policy that allows HMOs 3 years to retroactively change
institutional status data in beneficiary records; (7) the lack of a
systematic approach for identifying errors limits HCFA's efforts to
recover overpayments and ensure that appropriate payments are made to
HMOs; (8) HCFA generally waits 2 years to verify that HMOs have
corrected inaccurate recordkeeping systems, even when serious errors
have been identified; (9) HCFA continues to use 20-year-old cost data in
determining the payment rates for institutionalized enrollees and, as a
result, HCFA overcompensates HMOs for their enrolled, institutionalized
beneficiaries; (10) this overpayment problem may be corrected when HCFA
implements a revised set of risk factors in 2000; (11) however,
provisions of the Balanced Budget Act of 1997 that use 1997 rates as the
basis for 1998 and future rates effectively preclude a revision to the
institutional risk factor at this time; (12) while HCFA has revised its
definition of eligible institutions, concerns remain that HCFA's
oversight of payments for institutional status is inadequate; (13) HCFA
has no system to estimate and recover total overpayments when
institutional status errors are detected or to verify HMOs' retroactive
adjustment requests; and (14) further, HCFA does not ensure timely
review of those HMOs found to have submitted inaccurate institutional
status data, and its use of outdated cost data in determining payments
continues to overcompensate HMOs for institutionalized enrollees.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  HEHS-98-153
     TITLE:  Medicare HMO Institutional Payments: Improved HCFA 
             Oversight, More Recent Cost Data Could Reduce Overpayments
      DATE:  09/09/98
   SUBJECT:  Overpayments
             Health care programs
             Health care cost control
             Health maintenance organizations
             Extended care facilities
             Health services administration
             Long-term care
             Eligibility determinations
IDENTIFIER:  Medicare Program
             Medicare Current Beneficiary Survey
             
******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO report.  Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved.  Major          **
** divisions and subdivisions of the text, such as Chapters,    **
** Sections, and Appendixes, are identified by double and       **
** single lines.  The numbers on the right end of these lines   **
** indicate the position of each of the subsections in the      **
** document outline.  These numbers do NOT correspond with the  **
** page numbers of the printed product.                         **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************


Cover
================================================================ COVER


Report to the Permanent Subcommittee on Investigations, Committee on
Governmental Affairs, U.S.  Senate

September 1998

MEDICARE HMO INSTITUTIONAL
PAYMENTS - IMPROVED HCFA
OVERSIGHT, MORE RECENT COST DATA
COULD REDUCE OVERPAYMENTS

GAO/HEHS-98-153

Medicare HMO Institutional Payments

(101514)


Abbreviations
=============================================================== ABBREV

  BBA - Balanced Budget Act of 1997
  HCFA - Health Care Financing Administration
  HHS - Department of Health and Human Services
  HMO - health maintenance organization
  MCBS - Medicare Current Beneficiary Survey

Letter
=============================================================== LETTER


B-279730

September 9, 1998

The Honorable Susan M.  Collins
Chairman
The Honorable John Glenn
Ranking Minority Member
Permanent Subcommittee on Investigations
Committee on Governmental Affairs
United States Senate

The Congress has focused increasingly in recent years on rapidly
growing Medicare expenditures.  Between 1990 and 1997, Medicare
spending grew from $107 billion to $207 billion, or about 12 percent
per year.  Although the bulk of program outlays is for services
delivered through traditional fee-for-service Medicare, a growing
number of seniors--approximately 5 million out of 38 million Medicare
beneficiaries--receive care through health maintenance organizations
(HMO) that participate in Medicare's risk contract program.  Unlike
fee-for-service providers, which are paid on a per-claim basis, these
HMOs receive from Medicare a monthly fixed sum per enrolled
beneficiary--a capitation rate--and assume the risk of providing
beneficiary health care, regardless of the actual costs incurred. 

Part of the process for setting capitation rates involves risk
adjustment.  The Health Care Financing Administration (HCFA), the
agency responsible for administering Medicare, adjusts the capitation
rate for such factors as age and sex to reflect the expectation that
different classes of beneficiaries are likely to incur higher or
lower costs than the average beneficiary.  For example, HCFA has
determined that the estimated 2.6 million Medicare beneficiaries
living in nursing homes and other long-term care facilities
frequently incur greater-than-average Medicare-covered expenses. 
Consequently, the "institutional" risk adjuster generally raises
capitation payments for the approximately 50,000 Medicare HMO
enrollees residing in such facilities.  The payments are intended to
cover expected higher-than-average costs for Medicare-covered
acute-care and post-acute-care services. 

Concerned about the potential incentive for some HMOs to
inappropriately classify beneficiaries as institutionalized and the
possibility that flaws in the methodology for calculating the
institutional rate are generating excessive payments, you asked us to
examine (1) the criteria HCFA uses to determine a beneficiary's
institutional status, (2) the methods HCFA employs to ensure that
HMOs properly classify beneficiaries as institutionalized, and (3)
whether the higher capitation rate for beneficiaries who live in
institutions is justified by higher health care costs. 

To do this work, we visited selected facilities classified by HMOs as
institutions; reviewed HCFA's relevant policy and procedure manuals
at HCFA headquarters and the agency's reporting and verification
procedures at two regional offices (San Francisco and Seattle) that
cover high managed care penetration areas; and interviewed HCFA
officials, industry groups, and Medicare HMO representatives.  In
addition, we reviewed previous research on the health cost estimates
for institutional-status beneficiaries that HCFA used to set HMO
rates.  We also analyzed data from HCFA's Medicare Current
Beneficiary Survey (MCBS) to examine cost differences for
beneficiaries residing in different types of facilities.  HCFA
officials have reconciled the MCBS data with Medicare claims and
other administrative information to produce data the agency believes
to be generally accurate and complete.  Given HCFA's intensive
efforts, we did not independently verify the accuracy of the
information in the MCBS data files.  Our work was performed from
January 1997 to April 1998 in accordance with generally accepted
government auditing standards. 


   RESULTS IN BRIEF
------------------------------------------------------------ Letter :1

Until recently, HCFA's broad definition of "institution" allowed HMOs
to claim institutional status for individuals residing in facilities
not likely to house sicker-than-average seniors.  Some of the
facilities we visited that HMOs had classified as institutional
residences provided no medical care but rather offered a menu of
recreational activities for seniors capable of living independently. 
HCFA acted on our findings and those of others by narrowing the
definition of eligible institutions, effective January 1, 1998. 

Even with more stringent criteria, however, HCFA relies on the HMOs
to determine which beneficiaries qualify for institutional status. 
HCFA conducts only limited reviews, approximately every 2 years, to
confirm the accuracy of HMO records.  Studies by the Department of
Health and Human Services (HHS) Inspector General reviewing the
accuracy of HMO institutional status data support our finding that
HCFA's reviews are not adequate to detect the extent of errors or
overpayments resulting from HMOs' misclassification of beneficiaries. 
The task of ensuring accurate data may be further complicated by
HCFA's policy that allows HMOs 3 years to retroactively change
institutional status data in beneficiary records.  The lack of a
systematic approach for identifying errors limits HCFA's efforts to
recover overpayments and ensure that appropriate payments are made to
HMOs.  Moreover, HCFA generally waits 2 years to verify that HMOs
have corrected inaccurate record-keeping systems, even when serious
errors have been identified. 

Finally, HCFA continues to use 20-year-old cost data in determining
the payment rates for institutionalized enrollees.  As a result, HCFA
overcompensates HMOs for their enrolled, institutionalized
beneficiaries.  HCFA's own analysis confirms that updating the
institutional risk factor with more recent cost data, which are
available, would result in substantially lower payments for aged
institutionalized enrollees.  For example, an HMO would receive a
little more than half the current payment rate for a 74-year-old male
Medicare beneficiary classified as residing in an institution if the
institutional risk factors were updated.  This overpayment problem
may be corrected when HCFA implements a revised set of risk factors
in 2000.  However, provisions of the Balanced Budget Act of 1997
(BBA)\1 that use 1997 rates as the basis for 1998 and future rates
effectively preclude a revision to the institutional risk factor at
this time. 

While HCFA has revised its definition of eligible institutions,
concerns remain that HCFA's oversight of payments for institutional
status is inadequate.  HCFA has no system to estimate and recover
total overpayments when institutional status errors are detected or
to verify HMOs' retroactive adjustment requests.  Further, HCFA does
not ensure timely review of those HMOs found to have submitted
inaccurate institutional status data, and its use of outdated cost
data in determining payments continues to overcompensate HMOs for
institutionalized enrollees. 


--------------------
\1 Sec.  4001 of P.L.  105-33 added sec.  1853 (42 U.S.C.  1395w-23)
to the Social Security Act. 


   BACKGROUND
------------------------------------------------------------ Letter :2

Medicare provides health insurance for nearly all elderly Americans
(those aged 65 and older) and certain of the nation's disabled.  Most
Medicare beneficiaries receive services through the fee-for-service
sector.  However, as of April 1998, roughly 15 percent of Medicare's
beneficiaries--up from about 7 percent in mid-1995--were enrolled in
risk contract HMOs.\2 Of these, about 50,000 beneficiaries are
classified as institutionalized each month.\3

HCFA, an agency within HHS, administers the Medicare program and is
responsible for ensuring that Medicare HMOs comply with data
reporting, beneficiary protection, and care delivery requirements. 
HCFA seeks to ensure that HMOs meet financial solvency and enrollment
requirements, do not earn excessive profits,\4 operate internal
quality assurance systems, and establish grievance and appeals
procedures.  HCFA also implements the capitation rate formula
authorized by legislation and calculates payments for each HMO. 
Further, HCFA is responsible for monitoring HMOs to ensure that all
Medicare requirements are met, including that HMOs' reports of
beneficiaries' institutional status are accurate.  HCFA is also
responsible for ensuring that corrective actions are taken if
overpayments, underpayments, or other errors are discovered.  HCFA
established a national policy in 1994 permitting HMOs to seek
retroactive payment adjustments--for either overpayments or
underpayments--for the prior 3-year period.\5


--------------------
\2 Other Medicare managed care plans include cost contract HMOs and
health care prepayment plans, which together enroll fewer than 2
percent of the total Medicare population.  Because Medicare pays
these plans using methods other than capitation rates, they are not
the subject of this report. 

\3 A beneficiary's institutional status can change on a monthly
basis. 

\4 HMOs are permitted to retain all profits up to the level earned on
non-Medicare business. 

\5 Technically, every institutional rate adjustment is retroactive
because it applies to the month preceding the month in which the
adjustment is being reported.  (The regional staff we contacted
considered the months prior to the "current" adjustment month as a
retroactive period.)


      SIZE AND DISTRIBUTION OF
      MEDICARE HMO POPULATION WITH
      INSTITUTIONAL STATUS
---------------------------------------------------------- Letter :2.1

About 1 percent of Medicare's roughly 5 million HMO enrollees are
classified as living in institutions, as compared with about 7
percent of beneficiaries covered under Medicare fee-for-service.  In
1997, Medicare paid $197 million more to HMOs because of their
enrollees' institutional status. 

The distribution of enrollees with institutional status varies among
HMOs and among geographic regions.  In December 1997, most
institutional enrollment rates for individual HMOs ranged from 0 to
about 10 percent, although one outlier exceeded 44 percent.  (See
fig.  1.)

   Figure 1:  Percentage of
   Institutional Medicare
   Beneficiaries Enrolled in HMOs,
   December 1997

   (See figure in printed
   edition.)

Notes:  Institutional beneficiaries are those who reside in an
eligible institution for at least 30 consecutive days. 

The range of institutional enrollment excludes HMOs participating in
certain demonstrations that, for payment purposes, classify all HMO
members as institutionalized. 

Source:  GAO analysis of HCFA data. 


      THE ROLE OF THE
      INSTITUTIONAL RISK FACTOR IN
      SETTING HMO RATES
---------------------------------------------------------- Letter :2.2

Medicare's risk contract program was designed to save Medicare money
by paying HMOs 95 percent of the amount Medicare estimated it would
spend on similar beneficiaries in the fee-for-service sector.  It was
believed that HMOs would have lower costs because of their greater
emphasis on preventive health services and their incentive to
eliminate unnecessary services.  The base capitation rate in each
county--the amount an HMO receives each month for enrolling an
average-cost beneficiary--is determined by law, largely on the basis
of Medicare's per capita fee-for-service spending.\6

An HMO's monthly capitation payment is adjusted for the expected care
costs of each individual enrolled.  To make the adjustment, HCFA
assigns weights to defined risk classes of beneficiaries on the basis
of age; sex; and disability, Medicaid, institutional, and employment
status.  The weights are expressed as ratios of the national average
per capita costs for each risk class relative to the overall national
average.  For example, in 1997, compared with the national average
weight set at 1.0, the weight assigned to the risk class for men aged
85 or older with institutional status was 2.25.  Thus, HCFA's
estimate was that institutionalized men aged 85 and older would have
health care costs that were 2.25 times the costs for the average
beneficiary.  HCFA adjusts capitation rates for most
institutionalized beneficiaries upward to reflect the expected
differential.\7

The additional monthly payment amount associated with institutional
status can be substantial.  For example, in 1998 a Los Angeles HMO
receives $618 more per month for a 65-year-old man living in an
institution than for one who is not living in an institution ($1,071
instead of $453).  (See fig.  2 for a comparison of monthly HMO
payments in Los Angeles for institutional and noninstitutional
enrollees.)

   Figure 2:  Comparison of
   Monthly HMO Payments in Los
   Angeles for Aged Institutional
   and Noninstitutional Enrollees,
   for Selected Age Groups, 1998

   (See figure in printed
   edition.)

Source:  GAO analysis of HCFA's 1998 capitation rates. 

In 2000, HCFA is required by the BBA to implement a new risk
adjustment methodology that uses direct indicators of health
status--in addition to any other demographic adjusters such as age
and sex--to better reflect differences in individuals' expected
health care costs. 


--------------------
\6 Medicare determines four capitation rates for each county:  two
rates for part A services (one for the aged and one for the disabled)
and two rates for part B services (one for the aged and one for the
disabled).  Under the Medicare program, part A covers primarily
inpatient services; part B covers physician and other services. 
Capitation payment provisions of the BBA, which established minimum
payment amounts, minimum updates, and blended local and national
rates, loosened but did not eliminate the link between local
fee-for-service spending and local capitation rates. 

\7 For all aged beneficiaries and most disabled beneficiaries,
institutional status raises the capitation rate paid to HMOs.  For a
few age groups of disabled beneficiaries, institutional status can
reduce HMO payments, however.  In these cases, therefore, HMOs have
no financial incentive to report institutional status. 


   BROAD CRITERIA HAVE ALLOWED
   HMOS TO CLAIM HIGHER PAYMENT
   RATES FOR CERTAIN SENIORS
------------------------------------------------------------ Letter :3

Until recently, HCFA has defined the term "institution" to include
skilled nursing facilities, nursing homes, sanitoriums, rest homes,
convalescent homes, long-term care hospitals, domiciliary homes,
swing-bed facilities,\8

and intermediate-care facilities.\9 HCFA based the higher payment
rates for institutional status on historical evidence that
beneficiaries living in these types of facilities had greater medical
needs and higher medical costs than those who lived in the community. 
These types of facilities, however, have evolved to serve individuals
with varied health care needs.  Thus, HCFA's broad criteria permitted
an HMO to classify virtually any residential facility as an
institution for payment purposes.  Consequently, HMOs considered as
institutions many facilities that were housing seniors whose expected
health needs were at or below those of the average Medicare
beneficiary. 

HMOs have had an incentive to broadly interpret HCFA's institution
criteria.  For example, if an HMO classified a residential facility
as an institution, the HMO would receive a much higher capitation
payment--up to $766 more per month--for every enrollee living in that
facility.  In fact, the increased institutional payments to one
relatively small HMO amounted to about an additional $135,000 for 1
month.  For HMOs with larger institutionalized enrollment, the annual
additional capitation payments could be on the order of $4 million to
$9 million. 

Medicare's payment system is based on the assumption that HMO
enrollees living in institutions generate above-average health care
costs.  However, some facilities classified by HMOs as institutions
clearly did not serve seniors with serious health problems.  For
example, among the designated institutions we visited, one (called by
its manager an "independent living facility") provided private
apartments, meals in a communal setting, and field trips to tourist
and shopping sites.  About 12 percent of the residents owned and
drove their own cars.  The facility did not provide any medical care. 
Another facility we visited--a retirement center--was characterized
by its marketing brochure as "a clean, comfortable home for those who
do not need nursing care." This facility employed a full-time
activity director and housed several residents who drove their own
cars. 

Moreover, HCFA's institutional payment policy is unclear when a
single facility offers a range of assistance levels--from independent
living arrangements to skilled nursing care.  For example, one
residential community we visited consisted of two facilities:  one, a
skilled nursing care facility, provided subacute, skilled, and
custodial care; the other, an independent living facility, provided
limited assistance, such as helping individuals get to the communal
dining room.  An HMO planned to classify the entire residential
community as an institution.  However, the residential community's
manager disagreed that the independent living facility constituted an
institution.  Although the HMO ultimately chose not to classify those
beneficiaries living in the independent facility as
institutionalized, no HCFA policy would have prevented such
classification.  Facilities of this nature pose continued challenges
in appropriately determining which beneficiaries should be classified
as institutionalized for payment purposes. 

Even though some regional HCFA officials felt that some facilities
should not have been classified as institutions, these officials
believed they had little basis for challenging any classification. 
In a 1995 memorandum to HCFA headquarters, for example, the director
of a HCFA region's managed care operations noted that "the manual
provides no guidance regarding the level of care provided to
residents [needed to qualify for institutional status]" and that
"HCFA has the obligation to provide better guidance to plans
regarding the types of facilities which may be designated as
`institutions.'"

On July 24, 1997, HCFA issued a policy letter that narrowed the
definition of eligible institutions effective January 1, 1998.  The
letter cited a history of interpretation problems in using the
broader definition as well as the concerns we, the HHS Inspector
General, and the agency itself have raised about the potential for
making improper payments to HMOs.  Under the new definition, only
specified Medicare- or Medicaid-certified institutions are included,
thus limiting eligibility to institutions qualifying under the Social
Security Act, such as skilled nursing and nursing facilities;
intermediate care facilities for the mentally retarded; and
psychiatric, rehabilitation, long-term care, and swing-bed hospitals. 
Tying eligibility to certain Medicare- and Medicaid-certified
institutions effectively rules out eligibility for independent or
low-level assisted living facilities. 

In principle, this change could significantly improve HCFA's ability
to ensure that the higher capitation rate is being paid on behalf of
only those beneficiaries likely to have higher health care needs and
costs.  However, in practice, the collocation of independent living
arrangements with eligible institutions and HCFA's infrequent and
narrow review of HMO records, as discussed in the next section, may
limit the practical impact of HCFA's new policy. 


--------------------
\8 Swing-bed facilities are hospitals that are permitted to provide
skilled nursing services to Medicare beneficiaries occupying
acute-care beds. 

\9 To be eligible for the institutional rate during a particular
month, the beneficiary must have been a resident for the 30
consecutive days immediately preceding the last day of the month. 


   HCFA OVERSIGHT IS INADEQUATE TO
   PREVENT AND RECOVER
   INAPPROPRIATE HMO PAYMENTS
------------------------------------------------------------ Letter :4

The process HCFA uses to verify that HMOs appropriately claim the
institutional payment rate is inadequate.  HCFA relies on the HMOs
themselves to identify and report the names of beneficiaries for whom
the HMOs should receive the institutional rate.  Using these mostly
unaudited, HMO-reported data, HCFA adjusts the capitation payments
for the HMOs' Medicare members who live in institutions. 

HCFA does not conduct either comprehensive or spot checks at the
institutional facilities to assess the accuracy of the institutional
status data reported by HMOs.  Instead, HCFA regional staff make site
visits to each HMO about every 2 years and examine a small sample of
beneficiary records maintained by the HMO.  Results of previous HHS
Inspector General audits and our work show that the lack of effective
oversight fails to hold HMOs accountable for submitting accurate
records and thus does not ensure that HMOs receive appropriate
payments. 


      HCFA RELIES ON ERROR-PRONE
      DATA TO DETERMINE WHETHER TO
      INCREASE HMO PAYMENT
---------------------------------------------------------- Letter :4.1

HCFA bases its institutional rate adjustments solely on HMO-reported
data.  Each HMO is responsible for establishing a system to identify
and report its institutionalized beneficiaries to receive the pay
adjustment from HCFA.  The reporting process works roughly as
follows:  Each month, the HMO identifies those members who have
resided in eligible facilities for 30 consecutive days prior to the
reporting month and sends HCFA a list of beneficiaries qualifying for
institutional status.\10 Using the HMO's information, HCFA then
develops and sends to the HMO its monthly report of the HMO's
qualifying beneficiaries (the HMO is responsible for informing HCFA
of any further changes in beneficiary status).  On the basis of the
final, HMO-corrected report, HCFA adjusts--generally substantially
increasing--the HMO's capitation payment for each institutionalized
beneficiary. 

HCFA regional staff and HMO staff concur that, for a variety of
reasons, HMO data on institutionalized beneficiaries can be
inaccurate.  The financial incentive for HMOs to classify
beneficiaries as institutionalized is one possible explanation for
inaccurate data.  Other explanations include financial incentives for
physicians to misclassify beneficiaries, inaccurate data reported by
institutions, and data entry errors by HMOs. 

Some HMOs may have difficulty ensuring accurate data because their
providers financially benefit if enrollees are classified as living
in institutions.  For example, the primary care providers in the
Minnesota-based HMO cited in a 1995 HHS Inspector General's report\11
received from 85 to 90 percent of Medicare's per capita payment,
while the HMO kept the remainder.  The HMO required these providers
to notify it when their Medicare patients entered or left an
institutional setting or otherwise changed their status.  However,
the Inspector General found that the providers failed to do so and
did not correct HMO reports sent to them for reconciliation purposes. 
As a consequence, the HMO substantially overreported the number of
enrollees living in institutions. 

HMO staff also reported difficulty obtaining accurate information on
beneficiaries' current residence in particular facilities.  HMO staff
who were responsible for verifying enrollees' institutional status
said they typically contacted facilities by phone or mail monthly to
determine which enrollees resided in those facilities.  However,
facilities housing an HMO's Medicare enrollees do not necessarily
have a contractual or financial relationship with the HMO. 
Consequently, these facilities have no compelling reason to comply
with an HMO's information requests.  HMO staff reported instances of
not learning of changes in beneficiaries' institutional status, even
when the HMOs had requested verification and received regular
responses from facility personnel. 

Data entry errors are a third possible reason for inaccurate data. 
The 1995 Inspector General's report attributed some instances of
institutional status misclassification and Medicare overpayments to
the HMO's own data entry errors. 


--------------------
\10 The list of institutional beneficiaries becomes part of a larger
report sent to HCFA that includes other demographic information about
the HMO's Medicare enrollees. 

\11 See HHS, Office of Inspector General, Review of Medicare Payments
for Institutionalized Beneficiaries (A-05-94-00053) (Washington,
D.C.:  HHS, May 1995). 


      HCFA ALLOWS HMOS A 3-YEAR
      PERIOD TO ADJUST PAYMENTS
      FOR INSTITUTIONAL ENROLLEES
---------------------------------------------------------- Letter :4.2

Verifying HMOs' historical institutional status data is even more
difficult than ensuring the accuracy of current data.  The period of
time being scrutinized is longer, and HCFA's policy of allowing HMOs
3 years to correct institutional status data and adjust payments
accordingly compounds the problem.  For example, after a 1992
monitoring review of an HMO, HCFA required the HMO to correct
problems in its procedures for verifying its Medicare enrollees'
institutional status and to conduct an audit of its own institutional
status records.  As a result of the audit, the HMO reported nearly $5
million in overbillings and more than $4.5 million in underbillings
to Medicare during a 2-1/2-year period.  HCFA accepted the results of
the HMO's self-audit and the HMO's request to repay Medicare the
difference between the two amounts (approximately $500,000).  The
accuracy of the HMO's audit results was found to be questionable,
however, when, according to one HCFA official we interviewed, the HMO
later attempted to reverify its audit findings and was unable to do
so because facilities had changed the information they had originally
reported to the HMO. 

Concerned about potential errors in HMOs' historical institutional
status data, three of the four HCFA regional offices we contacted for
this study do not permit, or permit only by exception, retroactive
reimbursements.  Officials in HCFA's central office said that the
regional offices should be following the national policy, which
allows corrections in institutional status, and related
reimbursements, for up to 3 years.  They also said, however, that
they are aware that regional offices are not doing so.  These
officials said that, because of the frequent changes in HMOs'
historical institutional status data, following the national policy
would require substantial additional regional work to validate and
update the necessary corrections to HCFA's payment system. 


      HCFA'S OVERSIGHT FAILS TO
      ENSURE ACCURATE PAYMENTS
---------------------------------------------------------- Letter :4.3

Our review and the Inspector General audits underscore the need for
HCFA to improve its oversight of the HMO data used in determining
Medicare payments to HMOs.  HMOs' records are normally checked by
HCFA only during routine monitoring visits, which occur about every 2
years.  During a monitoring visit, HCFA staff focus primarily on
whether the HMO has a data verification system in place.  That is,
they review the HMO's policies and procedures for both updating the
HCFA report on institutionalized beneficiaries and contacting
facilities to verify residence and length-of-stay information.  HCFA
regional staff also contact a few facilities to confirm the residence
and length of stay of some beneficiaries.  Specifically, HCFA
protocol requires regional staff to verify the status of 30 enrollees
living in at least three different institutions and to contact three
of the institutions. 

HCFA's verification practices may be too superficial to determine
whether HMOs accurately report beneficiaries' institutional status. 
For example, after HCFA reviewed one Minnesota-based HMO's
institutional reporting procedures and records and found no problems,
an Inspector General audit of the same HMO revealed significant
errors.  The Inspector General examined the records of 100 enrollees
randomly selected from the 1,941 Medicare beneficiaries the HMO
listed as living in an institution during April 1994.\12 By checking
the HMO's records against those of the institutions, the Inspector
General determined that 15 of the 100 beneficiaries did not reside in
the listed institution.  The Inspector General also checked
historical records and found that some of the 15 misclassified
beneficiaries had never lived in an institution while enrolled in the
HMO.  In some cases, the HMO had misclassified the beneficiaries and
collected the institutional payment rate for over 5 years.  Total
overpayments for the 15 misclassified beneficiaries amounted to
$93,252. 

In 1993, the Inspector General cited two Massachusetts-based HMOs for
receiving enhanced payments on the basis of HMO data that
inaccurately classified beneficiaries as institutionalized.\13
Moreover, the HMOs' internal reporting systems did not accurately
reflect the discharge dates of some institutionalized beneficiaries. 
The Inspector General identified overpayments of about $215,000 for
the two HMOs over roughly a 2-year period ending June 30, 1993.\14

When HCFA staff identify faulty HMO data on institutional
beneficiaries, the agency frequently does little to determine the
full extent of the errors or the total overpayments generated by the
faulty data.  HCFA generally requires only that the HMO develop a
corrective action plan describing how the HMO intends to generate
better data.  In some cases, HCFA also requires HMOs to self-audit
their prior institutional reporting. 

After HCFA identifies HMO data errors, mandates corrective actions,
and approves a corrective action plan, it often waits 2 years or more
before verifying HMO compliance.  Once HCFA has approved an action
plan to correct an identified problem, the agency typically does not
check to determine whether the HMO has implemented the plan until
HCFA staff conduct the next routine monitoring review.  Sometimes
this monitoring review is delayed beyond the routine 2-year schedule,
even when a serious reporting problem was found to have existed
earlier.  Such was the case for the Minnesota-based HMO cited in a
previous example.  HCFA did not review the HMO's institutional
records until the fall of 1997, over 2 years after the Inspector
General reported the HMO's inaccurate record-keeping and resulting
overpayments, even though the HMO continued to maintain a rate of
institutionalized enrollment that was five times the national
average. 

The Inspector General is completing a study designed to determine the
extent of institutional status misreporting and to project total
national overpayments.  The Inspector General is reviewing the
institutional records at eight HMOs to determine whether
beneficiaries resided in the facilities listed in the HMOs' records
for the dates indicated.  If data errors are found, the Inspector
General intends to project and recoup overpayments from these
specific HMOs and also use the projections to estimate national
overpayments.  Preliminary results indicate data problems at five of
the eight HMOs.  Because the Inspector General's study does not
attempt to determine whether the listed facilities fit HCFA's
criteria for an eligible institution, the study's overpayment
estimate may understate the full extent of the problem. 


--------------------
\12 See HHS, Review of Medicare Payments for Institutionalized
Beneficiaries, May 1995. 

\13 See HHS, Office of the Inspector General, Premium Payments for
Medicare Beneficiaries Enrolled in Risk-Based Health Maintenance
Organizations, (A-01-93-00500) (Washington, D.C.:  Dec.  1993). 

\14 The Inspector General also identified approximately $115,000 in
premium payments made on behalf of Medicare beneficiaries who were
deceased.  In some cases, the beneficiaries had been deceased for
over 2 years. 


      HCFA LACKS A SYSTEMATIC
      APPROACH TO RECOVERING
      OVERPAYMENTS
---------------------------------------------------------- Letter :4.4

HCFA's procedures do not ensure that Medicare overpayments are
recovered when HMO data reporting errors are found.  In such cases,
HCFA requires HMOs to improve data reporting in the future, but often
the agency makes no attempt to estimate and recover overpayments
resulting from the faulty data.  HCFA sometimes, but not always,
requires HMOs to perform self-audits and bases payment adjustments on
the results.  However, beyond the limited number of beneficiary
records reviewed during routine monitoring visits, HCFA does not
attempt to verify HMO data or the results of HMOs' self-audits. 

A random sample of records of beneficiaries listed as living in
institutions can be useful in projecting and recovering total
Medicare overpayments.  For example, in the case of the
Minnesota-based HMO discussed earlier, the Inspector General found
that the status of 15 out of 100 randomly selected beneficiaries
classified by the HMO as living in institutions had been misreported. 
The overpayments associated with the 15 beneficiaries amounted to
$93,252.  On the basis of the random sample, the Inspector General
projected that the HMO had inappropriately received at least
$861,000, and perhaps as much as $2.8 million, from January 1989
through September 1994 for all enrollees misclassified as living in
institutions.  The Inspector General's findings enabled HCFA to
recoup about $861,000 from the HMO.\15


--------------------
\15 The Inspector General estimated that the HMO had been overpaid by
$1,810,021 during the 5-year period.  However, every statistical
projection has an associated "confidence interval." In this case, the
Inspector General determined that the overpayments ranged somewhere
between $861,615 and $2,758,428.  HCFA adopted a conservative
approach and required the HMO to repay the lower amount. 


   HCFA DATA SHOW MEDICARE
   OVERPAYING FOR
   INSTITUTIONALIZED BENEFICIARIES
------------------------------------------------------------ Letter :5

HCFA's most recent data show that the current institutional risk
adjuster substantially overcompensates HMOs for the institutionalized
beneficiaries they serve.  As a result, in July 1997, HCFA proposed
new weights for the institutional risk adjuster to more accurately
reflect the health care costs of institutional beneficiaries. 
However, in September of 1997, HCFA halted implementation of the new
weights, announcing that provisions of the recently passed BBA
precluded the agency from modifying any of the risk factors' weights
at that time.  Nonetheless, HCFA's new criteria for eligible
institutions--which exclude facilities housing beneficiaries with
relatively low expected health care costs--should help reduce
overpayments to HMOs serving institutional beneficiaries. 


      NEW DATA WOULD SUBSTANTIALLY
      LOWER CAPITATION PAYMENTS
      FOR INSTITUTIONALIZED
      BENEFICIARIES
---------------------------------------------------------- Letter :5.1

In the course of our review, HCFA developed new cost estimates for
institutionalized beneficiaries that were based on the 1993 MCBS
data.  The expected health care costs for institutionalized
beneficiaries, based on the 1993 MCBS, were much lower than those
estimated from the 1974-76 survey data, which are currently used to
set the risk factor for institutional beneficiaries.  Using the new
cost data, HCFA calculated lower adjustments to the capitation
payments for aged institutionalized beneficiaries.\16 For example, a
Medicare HMO that enrolls a 74-year-old male beneficiary living in an
institution in Los Angeles receives a monthly payment of about $1,307
in 1998.  If HCFA had implemented its revised rates, the HMO would be
receiving about $761 per month\17 --an amount that more accurately
reflects the expected costs associated with institutionalized
beneficiaries.  Table 1 shows that the Medicare part A component of
the monthly capitation payments would have fallen by as much as 24
percent for beneficiaries aged 85 and older and by as much as 62
percent for beneficiaries aged 65 to 84.  The decrease in the part B
component would have been somewhat less. 



                                Table 1
                
                   Comparison of Current Weights and
                   Weights HCFA Calculated Using More
                        Recent Cost Data for the
                         Institutionalized Aged

                                                             Percentag
                                                              e change
                                                                  from
                                                              existing
                                             Existi  Propos         to
                                                 ng      ed   proposed
                                             weight  weight    payment
Age                                               s       s       rate
-------------------------------------------  ------  ------  ---------
Medicare part A--male
----------------------------------------------------------------------
65 -69                                         1.75    0.80        -54
70 -74                                         2.25    1.05        -53
75 -79                                         2.25    1.40        -38
80 -84                                         2.25    1.75        -22
85 +                                           2.25    2.15         -4

Medicare part A--female
----------------------------------------------------------------------
65 -69                                         1.45    0.55        -62
70 -74                                         1.80    0.75        -58
75 -79                                         2.10    1.05        -50
80 -84                                         2.10    1.30        -38
85 +                                           2.10    1.60        -24

Medicare part B--male
----------------------------------------------------------------------
65 -69                                         1.60    1.10        -31
70 -74                                         1.80    1.40        -22
75 -79                                         1.95    1.65        -15
80 -84                                         1.95    1.80         -8
85 +                                           1.95    1.85         -5

Medicare part B--female
----------------------------------------------------------------------
65 -69                                         1.50    1.05        -30
70 -74                                         1.65    1.30        -21
75 -79                                         1.65    1.45        -12
80 -84                                         1.65    1.60         -3
85 +                                           1.65    1.65          0
----------------------------------------------------------------------
Source:  HCFA. 

Although HCFA announced plans in July 1997 to recalculate the weights
of the current demographic risk factors, including the institutional
risk adjuster, it halted this effort after the enactment of the BBA
in August 1997.  HCFA reverted to the old factors for the 1998 rate
calculations because the BBA specified a new methodology for setting
the basic capitation rate in each county that explicitly used the
established 1997 county rates as a base.  HCFA officials stated that
the new weights could only have been applied to capitation payment
calculations if the weights had also been used in the calculation of
the county rates. 


--------------------
\16 Because risk factors represent relative weights, a change in any
one risk factor must be offset by changes in others.  In this
instance, the dramatic drop in the institutional risk factor for the
small number of institutional beneficiaries would have been
accompanied by a very small increase in the risk factors for the more
numerous noninstitutional beneficiaries.  While overall Medicare
payments, theoretically, might not have been affected, payments to
individual HMOs could have changed substantially. 

\17 This estimate is based on the 1998 county rate for Los Angeles,
which would be somewhat different had HCFA adopted new weights for
the risk adjustment factors. 


      VARIATION IN EXPECTED COSTS
      FOR INSTITUTIONAL
      BENEFICIARIES PROVIDES
      OPPORTUNITIES FOR
      OVERPAYMENT
---------------------------------------------------------- Letter :5.2

Although HCFA uses the institutional risk adjuster to take into
account the expected higher costs of health care for
institutionalized beneficiaries, research on the risk adjuster
indicates that institutional residence is actually only weakly
related to a beneficiary's expected health care costs.  In a 1977
report, HCFA staff suggested that the average medical expenditures
for institutionalized beneficiaries could vary widely by type of
facility "to the extent that legal requirements and administrative
policies of institutions differentiate among the characteristics of
their residence." Our own analysis of the 1992-94 MCBS, the most
recent data available at the time of our analysis, found substantial
differences in Medicare costs among beneficiaries living in
institutions.  The average annual Medicare cost for beneficiaries in
nursing homes--at about $8,000, for example--was more than $3,700
higher than the average annual cost for beneficiaries in assisted
living facilities. 

HMOs could benefit financially if they were able to draw their
institutional populations disproportionately from those types of
institutions whose average beneficiary costs were lower than those of
other institutions.  HCFA's new definition of eligible institutions
includes certified nursing facilities but generally excludes assisted
living facilities.  This narrower definition could potentially
improve the accuracy of HMO payments for the beneficiaries they serve
by limiting the potential variation in average expected health care
costs among different types of institutions. 


   CONCLUSIONS
------------------------------------------------------------ Letter :6

Recent data clearly show that HMOs can be overcompensated for the
institutional beneficiaries they enroll.  Although provisions of the
BBA prevent HCFA from eliminating these excess payments at this time,
HCFA will have an opportunity to fully address this problem when it
develops a new set of risk adjusters, mandated by the BBA, to be
implemented in 2000. 

By tightening the definition of what constitutes an institution, HCFA
has taken a step toward improving the accuracy of HMO payments.  For
example, HMOs should no longer receive enhanced capitation payments
for serving beneficiaries in independent living facilities. 
Nonetheless, given HCFA's HMO monitoring practices, it is doubtful
that the agency can quickly or effectively determine the extent to
which HMOs are complying with the new definition. 

Moreover, the Medicare program remains open to potential abuse by
HMOs because HCFA performs only infrequent and limited checks of
HMO-reported data.  HCFA's use of unaudited HMO data to determine
payments to HMOs engenders little confidence in the accuracy of the
data and resulting payments.  HCFA also lacks a systematic approach
for identifying and recovering total overpayments once HMO reporting
errors are discovered.  Instead, HCFA typically requires HMOs only to
develop corrective plans to gather and report more accurate data in
the future.  Even when serious HMO reporting errors--resulting in
substantial overpayments--have been discovered, HCFA may wait 2 years
or more before checking to see if the HMO has implemented a revised
data gathering and reporting system. 


   RECOMMENDATIONS TO THE
   ADMINISTRATOR OF THE HEALTH
   CARE FINANCING ADMINISTRATION
------------------------------------------------------------ Letter :7

To better protect the integrity of Medicare capitation payments, we
recommend that the HCFA Administrator take the following actions: 

  -- Establish a system to estimate and recover total overpayments
     when institutional status data errors are detected. 

  -- Allow HMOs to revise records and claim retroactive payment
     adjustments for beneficiaries with institutional status only
     when HMO records have been verified by an independent third
     party. 

  -- Conduct timely follow-up reviews of those HMOs found to have
     submitted inaccurate institutional status data. 

  -- Use more recent cost data to calculate the institutional risk
     adjuster in the event HCFA continues to include institutional
     status as a part of its new risk adjustment methodology. 


   AGENCY COMMENTS
------------------------------------------------------------ Letter :8

HCFA agreed with our recommendations to improve the integrity of
capitation payments for institutionalized beneficiaries.  HCFA noted
several initiatives it is considering to improve oversight and
rate-setting methods.  We believe that these initiatives are a step
in the right direction but that HCFA must remain committed to
implementing the new methodologies.  The full text of HCFA's comments
appears in appendix I. 


---------------------------------------------------------- Letter :8.1

As arranged with your office, unless you publicly announce the
contents of this report earlier, we plan no further distribution
until 30 days after its issue date.  At that time, we will send
copies to the Secretary of Health and Human Services; the Director,
Office of Management and Budget; the Administrator of the Health Care
Financing Administration; and other interested parties.  We will also
make copies available to others upon request. 

This work was done under the direction of James Cosgrove, Assistant
Director.  If you or your staff have any questions about this report,
please contact Mr.  Cosgrove at (202) 512-7029 or me at (202)
512-7114.  Other GAO contacts and staff acknowledgments are listed in
appendix II. 

William J.  Scanlon
Director, Health Financing
 and Systems Issues




(See figure in printed edition.)Appendix I
COMMENTS FROM THE HEALTH CARE
FINANCING ADMINISTRATION
============================================================== Letter 



(See figure in printed edition.)



(See figure in printed edition.)


GAO CONTACTS AND STAFF
ACKNOWLEDGMENTS
========================================================== Appendix II

GAO CONTACTS

James C.  Cosgrove, Assistant Director, (202) 512-7029
Marie Cushing James, Co-Project Manager, (202) 512-3597
Bonnie Hall, Co-Project Manager

STAFF ACKNOWLEDGMENTS

The following team members also made important contributions to this
report:  Hannah F.  Fein, Senior Evaluator; Robert DeRoy, Assistant
Director; and George Bogart, Senior Attorney. 


*** End of document. ***