Medicare HMO Enrollment: Area Differences Affected by Factors Other Than
Payment Rates (Letter Report, 05/02/97, GAO/HEHS-97-37).

Pursuant to a congressional request, GAO reviewed the factors affecting
Medicare risk health maintenance organization (HMO) enrollment, focusing
on: (1) the patterns in HMO enrollment and Medicare payment rates; (2)
selected geographical areas with higher enrollment, lower payment rates
and areas with lower enrollment, higher payment rates; and (3) how the
presence or absence of certain factors could affect enrollment.

GAO noted that: (1) Medicare payment rates to HMOs are often considered
to be the primary influence on Medicare HMO enrollment; (2) however,
GAO's analysis suggest that several other factors also play a key, and
sometimes, dominant role; (3) these factors include HMO presence, number
of Medicare beneficiaries, and employers' policies toward retiree health
benefits, and their relative importance varies across the country; (4)
moreover, in markets such as Detroit and Portland, the influence of
Medicare payment rates is not decisive; (5) enrollment in risk HMOs was
virtually nonexistent in most counties with lower Medicare payment
rates, but these lower rates were one of a constellation of factors that
make such counties unattractive business propositions for Medicare HMOs;
(6) GAO's analysis showed that these counties typically had few or no
HMOs in their health care markets; (7) lower enrollment counties were
primarily rural, only 16 percent fell within a metropolitan statistical
area (MSA), and had fewer people overall and, in particular, averaged a
small number of Medicare beneficiaries; (8) lower enrollment in risk
HMOs did not occur in every county with lower payment rates; (9) risk
HMOs enrolled large numbers of beneficiaries in 92 lower payment
counties in which factors other than payment rates were more favorable;
(10) these counties were mostly in the West, where HMOs are more
prevalent and many consumers have embraced this form of health care
delivery; (11) in contrast, higher payment rates were no guarantee that
risk HMO enrollment would also be high; (12) about one-third of the 100
counties with the highest Medicare HMO payment rates in 1995 had risk
HMO enrollments that were slight or nonexistent; (13) most of these
higher payment/lower enrollment counties were in the South, where the
presence of HMOs was limited; (14) however, several of these counties
were in three Michigan urban areas; (15) although the presence of HMOs
in the health care market was generally greater in the Michigan MSAs
than in the South, employers' provision of richer retiree health
benefits made the risk HMO option less attractive to Medicare
beneficiaries in Michigan; (16) in addition to population density and
other factors external to HMOs, HMOs' individual business strategies for
the Medicare market are likely to affect the future direction of risk
HMO enrollment; and (17) all these strategies are likely to boost risk *

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

 REPORTNUM:  HEHS-97-37
     TITLE:  Medicare HMO Enrollment: Area Differences Affected by 
             Factors Other Than Payment Rates
      DATE:  05/02/97
   SUBJECT:  Health maintenance organizations
             Health care services
             Health insurance cost control
             Insurance premiums
             Cost analysis
             Managed health care
             Health care programs
             Payments
             Employee medical benefits
             Population statistics
IDENTIFIER:  Medicare Program
             Portland (OR)
             Detroit (MI)
             Ann Arbor (MI)
             Flint (MI)
             Medigap
             Medicare Risk Contract Program
             Boston (MA)
             
******************************************************************
** 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 Honorable
John F.  Kerry, U.S.  Senate

May 1997

MEDICARE HMO ENROLLMENT - AREA
DIFFERENCES AFFECTED BY FACTORS
OTHER THAN PAYMENT RATES

GAO/HEHS-97-37

Medicare Risk HMO Enrollment

(101400)


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

  AAPCC - adjusted average per capita cost
  HCFA - Health Care Financing Administration
  HCPP - health care prepayment plan
  HMO - health maintenance organization
  MSA - metropolitan statistical area

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


B-276654

May 2, 1997

The Honorable John F.  Kerry
United States Senate

Dear Senator Kerry: 

Enrollment nationwide in the Medicare managed care program has more
than tripled in the last decade--growing from about 1 million
enrollees in 1987 to about 3.8 million in 1996--but differences in
enrollment by state and by market area are striking.  In some
metropolitan areas like Portland, Oregon, and Tucson, Arizona, the
dominant form of Medicare managed care--the health maintenance
organization (HMO)--has enrolled more than 40 percent of the Medicare
beneficiaries.  By contrast, HMO enrollment in most rural areas, and
even in some metropolitan markets, is negligible. 

Although such large disparities are sometimes attributed to areas'
Medicare payment rates for HMOs, studies have identified this as only
one of several influences at work.  In a recent report prepared at
your request, we identified a number of areas where Medicare HMO
enrollment was higher despite lower payment rates and others where
higher payments had failed to generate enrollment.\1 To explore this
issue further, you asked us to

  -- identify patterns in HMO enrollment and Medicare payment rates,
     and

  -- examine selected geographical areas--in particular, some with
     higher enrollment/lower payment rates and some with lower
     enrollment/higher payment rates--and describe how the presence
     or absence of certain factors could affect enrollment. 

To identify where beneficiaries were enrolled in HMOs and the rates
that Medicare paid to HMOs, we used data obtained from the Health
Care Financing Administration's (HCFA) Office of Managed Care.\2 For
each of 3,141 counties in 1995, the latest data available at the time
of our review, these data showed (1) the number of eligible Medicare
beneficiaries, (2) the number enrolled in managed care plans with
Medicare risk contracts,\3 and (3) the per enrollee amount that
Medicare paid plans (that is, the adjusted average per capita cost
(AAPCC) rate for the county).\4 We categorized county enrollment as
lower, intermediate, or higher, and in a similar manner categorized
counties by payment rates.  We then placed each county in one of nine
categories based on its combination of risk HMO enrollment and
payment rate.  We focused our work primarily on counties in three of
the categories:  lower enrollment/lower payment, lower
enrollment/higher payment, and higher enrollment/lower payment. 

To identify factors that are likely to affect risk HMO enrollment, we
interviewed officials at HMO plans in three markets that had
different patterns of AAPCC payments and risk HMO enrollment--Boston,
Detroit, and Portland.  To obtain more information on factors related
to enrollment, we interviewed officials at HCFA's regional offices,
six national HMO chains, four additional regional HMOs, five
management consulting firms, and five employers.  Appendix I presents
a more detailed discussion of our methodology. 


--------------------
\1 Medicare HMOs:  Rapid Enrollment Growth Concentrated in Selected
States (GAO/HEHS-96-63, Jan.  18, 1996). 

\2 HCFA, part of the Department of Health and Human Services,
administers Medicare's fee-for-service and managed care programs. 
The Office of Managed Care within HCFA provides national direction
and leadership for Medicare managed care operations. 

\3 Both HMOs and competitive medical plans enter into risk contracts
with HCFA to provide Medicare-covered services to beneficiaries who
enroll.  Although competitive medical plans are subject to regulatory
requirements similar to those for HMOs, they have greater flexibility
in setting commercial premium rates and in the services they offer to
their commercial members.  In this report, our use of the term HMO
includes competitive medical plans.  HMOs entering into risk
contracts assume all the financial risk associated with providing
Medicare-covered services to enrolled beneficiaries.  They receive a
monthly per capita premium--the adjusted average per capita cost
payment--from HCFA for each Medicare beneficiary enrolled.  These
amounts vary by county. 

\4 HCFA pays some HMOs on a cost-reimbursement basis.  This approach
is similar to reimbursement on a fee-for-service basis in that the
provider assumes no risk that fees will be insufficient to cover
costs and therefore does not have the same incentive to reduce costs. 
We did not include cost-reimbursement HMOs in our analysis. 


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

Medicare payment rates to HMOs are often considered to be the primary
influence on Medicare HMO enrollment.  However, our analysis suggests
that several other factors also play a key, and sometimes, dominant
role.  These factors include HMO presence, number of Medicare
beneficiaries, and employers' policies toward retiree health
benefits.  Their relative importance varies across the country. 
Moreover, in markets such as Detroit and Portland, the influence of
Medicare payment rates is not decisive. 

Enrollment in risk HMOs was virtually nonexistent in most counties
with lower Medicare payment rates, but these lower rates were one of
a constellation of factors that make such counties unattractive
business propositions for Medicare HMOs.  Our analysis showed that
these counties typically had few or no HMOs in their health care
markets.  Lower enrollment counties were primarily rural--only 16
percent fell within a metropolitan statistical area (MSA)--and had
fewer people overall and, in particular, averaged a small number of
Medicare beneficiaries. 

Lower enrollment in risk HMOs did not occur in every county with
lower payment rates.  Risk HMOs enrolled large numbers of
beneficiaries in 92 lower payment counties in which factors other
than payment rates were more favorable.  These counties were mostly
in the West, where HMOs are prevalent and many consumers have
embraced this form of health care delivery.  A prime example is
Portland, where about 41 percent of the Medicare beneficiaries are
enrolled in risk HMOs.  HMOs have a long history in the Portland
area, and the share of the population enrolled in them is high. 
Moreover, the risk HMO option has been available to Medicare
beneficiaries there for more than a decade.  Finally, Portland
illustrates a pattern found in some other areas:  higher enrollment
in risk HMOs has spread beyond the counties within the Portland MSA
to the counties that border Portland and three other Oregon MSAs. 

In contrast, higher payment rates were no guarantee that risk HMO
enrollment would also be high.  About one-third of the 100 counties
with the highest Medicare HMO payment rates in 1995 had risk HMO
enrollments that were slight or nonexistent.  Most of these higher
payment/lower enrollment counties were in the South, where the
presence of HMOs was limited.  However, several of these counties
were in three Michigan urban areas:  Detroit, Ann Arbor, and Flint. 
Although the presence of HMOs in the health care market was generally
greater in the Michigan MSAs than in the South, employers' provision
of richer retiree health benefits made the risk HMO option less
attractive to Medicare beneficiaries in Michigan. 

In addition to population density and other factors external to HMOs,
HMOs' individual business strategies for the Medicare market are
likely to affect the future direction of risk HMO enrollment.  The
officials of multistate HMOs we interviewed said they are seeking to
expand their Medicare business.  Some HMOs emphasize expanding into
contiguous service areas.  Some enter into new risk contracts for
HMOs they already own, while others focus on acquiring new HMOs with
risk contracts.  All these strategies are likely to boost risk
enrollment and, sometimes, to change the market dynamics in certain
areas.  In Boston, for example, risk HMO enrollment grew by 158
percent during a 2-year period after a new player entered the market
and offered a no-cost option to Medicare beneficiaries, causing
existing HMOs to change their offerings to remain competitive. 


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

Medicare helps the elderly and certain disabled people meet the costs
of health care services.  Medicare is primarily a federally financed
and administered health insurance program, which reimburses
fee-for-service providers for each covered service rendered. 
However, Medicare also offers beneficiaries the option of enrolling
in managed care plans--mostly risk contract HMOs.  While nearly
two-thirds of beneficiaries have access to at least one Medicare HMO
that provides service in their zip code areas, only about 10 percent
of beneficiaries belong to a Medicare HMO.  Most beneficiaries are
still enrolled in the traditional fee-for-service program. 

Under Medicare's initial authority for paying HMOs that provided care
to beneficiaries, few HMOs contracted with Medicare.  As enacted in
1972, the legislation limited a participating HMO's profit potential
while losses had to be fully absorbed.  Consequently, most Medicare
HMOs chose to be paid on a cost basis.  Facing neither profit nor
loss from serving Medicare beneficiaries, cost contract HMOs lacked a
strong incentive to reduce unnecessary care and deliver care
efficiently. 

Legislation changed Medicare's payment mechanism in 1982.\5 Since
then, HMOs have been able to enter into more attractive risk
contracts with HCFA.  The HMO receives a fixed monthly capitation
payment--the AAPCC rate--for each beneficiary enrolled\6 in exchange
for providing all Medicare part A and part B services.\7 The risk to
the HMO is that for any particular beneficiary, the cost of care may
exceed the prepaid amount.\8

Each year, a risk HMO must estimate what it would charge Medicare
beneficiaries for Medicare-covered services if they were commercial
enrollees.  The estimate of the premiums it would charge to provide
such services to non-Medicare enrollees is adjusted to reflect the
differences in (1) the benefits provided to Medicare enrollees and
(2) the use of services by Medicare beneficiaries.  This estimate
(which includes the normal profit of a for-profit HMO) is used to
identify any excess profits the HMO will derive from the Medicare
business.  The HMO is permitted to retain all profits up to the level
earned on its non-Medicare business.  If the expected Medicare profit
exceeds the estimated profit on its non-Medicare business, the HMO
must either return the excess to Medicare or provide additional
supplemental benefits or reduced copayments or deductibles to
beneficiaries.  Virtually all HMOs faced with this choice have opted
to provide increased benefits, which Medicare beneficiaries can find
very attractive. 

Before an HMO can enter into a risk contract, the law requires it to
have at least 5,000 commercial enrollees.  An HMO serving primarily
rural areas must have at least 1,500 members.  In addition, the
Medicare law's "50-50 rule" states that no more than 50 percent of an
HMO's enrollment may be Medicare beneficiaries and Medicaid
recipients. 

Medicare beneficiaries enrolled in a risk HMO face a "lock-in"
requirement.  Once they enroll, they must receive virtually all their
health care services through the HMO.\9 If a beneficiary goes outside
the HMO for any health care services, neither the HMO nor Medicare is
required to pay the cost.  Exceptions are made for emergency and
similar type care, which can be obtained anywhere in the country, and
for which the HMO should pay.  A few risk HMOs now offer a "point of
service" option through which beneficiaries can receive certain
services outside the plan's network of providers but must pay more
than for "within-plan" services. 

Beneficiaries consider a number of factors when deciding whether to
enroll in a Medicare risk HMO, including (1) their familiarity with
managed care, (2) their attachment to current health care providers,
(3) whether they travel out of the area or live part of the year in
another state, and (4) likely out-of-pocket costs in a risk HMO
versus the fee-for-service program (plus the cost of a Medigap
policy\10 ).  In a risk HMO, the beneficiary may be charged a fixed
monthly premium and a copayment each time a service is used.\11
Depending on the additional benefits a Medicare HMO provides,
enrolling may be an alternative to buying Medigap insurance, often at
a lower cost to the beneficiary. 

Many employers provide health benefits to retired employees who
receive Medicare.  Some employers, faced with rising retiree health
care costs and new accounting rules,\12 believe they can reduce costs
when their Medicare-eligible retirees enroll in risk HMOs.  As a
result, some employers are providing retirees with incentives to join
risk HMOs.  They may limit their contribution to just cover the cost
of a Medicare risk HMO, thereby causing retirees who do not join an
HMO to face increased health care costs.  To remain in Medicare's
fee-for-service plan these beneficiaries may have to buy a Medigap
policy and pay the difference between the premium and the employer's
contribution. 

About 3.8 million Medicare beneficiaries were enrolled in risk HMOs
in August 1996, but enrollment was concentrated primarily in urban
areas in the West and in Florida.  Only a handful of counties in the
East, South, and Midwest had more than 5 percent of their Medicare
population enrolled in risk HMOs.  (See fig.  1.) In total, 242 of
3,141 counties had more than 5 percent enrollment in the Medicare
risk HMO program--a level that for this study we classify as "higher
enrollment." In contrast, however, 2,663 of the counties--about 85
percent--had either no beneficiaries or fewer than 1 percent enrolled
in risk HMOs.  (We classify these areas as "lower enrollment.")

   Figure 1:  Counties With Higher
   Risk HMO Enrollment, December
   1995

   (See figure in printed
   edition.)

Sources:  Medicare Market Penetration Report File, AAPCC Rate File,
and other selected files, Office of Managed Care, HCFA. 


--------------------
\5 The change was made through the Tax Equity and Fiscal
Responsibility Act of 1982 (sec.  114,
P.L.  97-248). 

\6 AAPCC rates are set at 95 percent of the average amount that HCFA
estimates it would spend reimbursing fee-for-service providers who
deliver medical care to a typical beneficiary in a county.  HCFA
adjusts payments for different categories of beneficiaries on the
basis of age, sex, Medicaid eligibility, institutional status, and
working status.  The AAPCC rates change every calendar year. 

\7 Medicare part A--Hospital Insurance--covers services provided by
hospitals, skilled nursing facilities, hospices, and home health
agencies.  Medicare part B--Supplemental Medical Insurance--covers
physician and outpatient hospital services as well as other services
and supplies received on an outpatient basis, such as laboratory
services and medical supplies. 

\8 Cost contracts are still an option.  (An HMO that is losing money
under a risk contract may switch to a cost contract.) In August 1996,
34 plans had cost contracts and had 182,494 beneficiaries enrolled. 
Also, a plan can contract with HCFA as a health care prepayment plan
(HCPP).  As an HCPP, a plan is paid on a cost basis for Medicare part
B services.  Part A services are not covered by HCPPs and remain
under the fee-for-service program.  There were 50 HCPPs with 300,108
enrollees in August 1996. 

\9 Medicare beneficiaries may choose to disenroll from a risk HMO at
any time. 

\10 Medigap insurance provides benefits that help fill the gaps in
Medicare coverage.  Federal law requires Medigap plans to offer 1 of
10 standard benefit packages--labeled plan A through plan J.  These
private, supplemental plans cover various combinations of Medicare
cost-sharing requirements.  Some plans also include services not
covered by Medicare--for example, outpatient prescription drugs. 
Plan A is the most basic while plan J is the most comprehensive. 
Insurance companies are not allowed to change the letter designations
or the combination of benefits offered.  Although the benefits are
the same for all Medigap plans, premiums may vary greatly across
companies. 

\11 This is in addition to the premium that all beneficiaries must
pay to receive physician services under Medicare part B. 

\12 Starting in 1993, Financial Accounting Standards Board rule 106
required that private-sector employers with 500 or more plan
participants treat health benefit obligations for present and future
years on an accrual instead of a pay-as-you-go basis. 


   RISK HMO ENROLLMENT LINKED TO
   PAYMENT AND OTHER FACTORS
------------------------------------------------------------ Letter :3

Our analysis of counties grouped by their AAPCC rates and risk HMO
enrollment levels suggests that while AAPCC rates play a role, other
factors also affect enrollment.  (See table 1.) In addition to HMO
presence in the health care market, such factors as population
density, the number of Medicare beneficiaries, and employers'
policies on retiree health benefits can influence risk enrollment. 
The studies we reviewed all found that several factors--usually
including AAPCC rates and HMO presence--help account for the
differences in risk HMO enrollment from county to county.\13
Moreover, a study using recent data found that the factor with the
strongest influence was HMO presence.\14

Greater HMO presence establishes more familiarity in an area for
managed care in general and for a particular plan.  HMOs can draw on
that when trying to attract Medicare beneficiaries. 



                          Table 1
          
               Number of Counties With Lower,
           Intermediate, and Higher Enrollment in
           Risk HMOs by Lower, Intermediate, and
             Higher AAPCC Rates, December 1995

                     Enrollment in Medicare risk HMOs
                ------------------------------------------
Monthly AAPCC
payment rates             Intermediate
(per person)\a   Lower\b            \c  Higher\d     Total
--------------  --------  ------------  --------  ========
Lower (under     2,257\f           125        92     2,474
 $375)\e
Intermediate         373            93       101       567
 ($375 to
 $464)\g
Higher (top           33            18        49       100
 100 payments-
 -$464 to
 $679)\h
==========================================================
Total              2,663           236       242     3,141
----------------------------------------------------------
Note:  Payment rates are rounded to the nearest dollar. 

\a The lower, intermediate, and higher AAPCC rate and enrollment
designations were chosen for analytic purposes.  The designations are
relative and do not imply that such levels are appropriate or not
appropriate. 

\b Counties had 1 percent or less of their beneficiaries enrolled or
none enrolled. 

\c Counties had more than 1 percent but no more than 5 percent of
their beneficiaries enrolled. 

\d Counties had more than 5 percent of their beneficiaries enrolled. 

\e The average AAPCC rate was $306; the median, $310. 

\f Of the 2,257 counties, 552 had no risk HMO enrollment.  According
to HCFA data, an additional 1,604 counties had fewer than 25
enrollees. 

\g The average AAPCC rate was $408; the median, $403. 

\h The average AAPCC rate was $512; the median, $496. 

Sources:  Medicare Market Penetration Report File and AAPCC Rate
File, Office of Managed Care, HCFA. 


--------------------
\13 See Physician Payment Review Commission, Annual Report to
Congress, Physician Payment Review Commission, 1995 (Washington,
D.C.:  1995), pp.  92-94; Physician Payment Review Commission, Annual
Report to Congress, Physician Payment Review Commission, 1996
(Washington, D.C.:  1996), pp.  80-81; and Carl Serrato, Randall S. 
Brown, and Jeanette Bergeron, "Why Do So Few HMOs Offer Medicare Risk
Plans in Rural Areas?" Health Care Financing Review, Vol.  17, No.  1
(Fall 1995),
pp.  85-97. 

\14 W.  Pete Welch, "Growth in HMO Share of the Medicare Market,
1989-94," Health Affairs, Fall 1996, pp.  201-214. 


   MOST COUNTIES WITH LOWER
   PAYMENT RATES ARE LESS URBAN
   AND LACK STRONG HMO PRESENCE
------------------------------------------------------------ Letter :4

Medicare risk HMO enrollment in 2,257 counties that had lower AAPCC
rates was minimal or nonexistent, but the principal barrier to risk
HMO enrollment was not the payment level.  On average, the 2,257
counties with lower AAPCC rates and lower enrollment had fewer people
per square mile, and only about 16 percent of these counties were
urban in that they were included in an MSA.  For the most part, the
2,257 counties were mainly rural or sparsely populated and
consequently, most had few or no HMOs serving non-Medicare residents. 

Generally, HMOs thrive and exert a stronger presence in counties that
are part of metropolitan areas.  In the 10 states with the highest
percentage of people enrolled in HMOs in 1994, about 92 percent of
the population lived in an MSA.  Markets need to be of sufficient
size to generate an HMO presence in general and a risk HMO program in
particular.  According to one study, the most remote counties with
the smallest populations are not likely to be included in HMO
markets.\15 Table 2 compares characteristics of the 10 states with
the lowest level of total HMO market share (that is, commercial,
Medicaid, and Medicare enrollment) and the 10 states with the highest
total HMO market share.\16 The 10 states with the lowest total HMO
market shares averaged less than two HMOs each. 



                          Table 2
          
             Selected Characteristics of the 10
          States With Lowest and Highest Total HMO
                        Market Share

                               Data for 10     Data for 10
                               states with     states with
                                    lowest         highest
Characteristic                     share\a         share\b
----------------------------  ------------  --------------
Total HMO market share\c              1.9%           32.9%
Average number of HMOs\c               1.6            16.6
Percentage of Medicare                0.1%           17.2%
 beneficiaries enrolled in
 risk HMOs\d
Percentage of counties in an          9.7%           41.4%
 MSA
Percentage of population             38.0%           92.0%
 living in an MSA
----------------------------------------------------------
\a The 10 states with the lowest total HMO market share at the end of
1994 were Alaska, West Virginia, Wyoming, Mississippi, North Dakota,
Idaho, Montana, South Dakota, Arkansas, and Iowa. 

\b The 10 states with the highest total HMO market share at the end
of 1994 were California, Oregon, Maryland, Arizona, Massachusetts,
Rhode Island, Connecticut, Minnesota, Colorado, and New York.  We did
not include the District of Columbia, which had the ninth highest
total HMO market share in 1994. 

\c As of year-end 1994. 

\d As of December 1995. 

Sources:  Medicare Market Penetration Report File, AAPCC Rate File,
and other selected files, Office of Managed Care, HCFA; Group Health
Association of America, Patterns in HMO Enrollment, June 1995; and
HCFA, 1995 Data Compendium. 

Even if the counties that had lower payment rates and lower
enrollments had an HMO present, most lacked a Medicare population of
sufficient size to attract an HMO into the risk contract program. 
Table 3 shows these counties had an average of about 5,600 Medicare
beneficiaries.  According to officials at several multistate HMO
chains we interviewed, one of the factors they consider most when
selecting areas of the country in which to pursue the Medicare
business is the size of the Medicare population.  An official at one
company estimated that no more than 20 to 30 percent of the
beneficiaries in a new market will join a risk HMO, which suggests
that only about 1,100 to 1,700 beneficiaries in each of these
counties may join.  In addition, officials at multistate HMOs said
that they need to enroll at least 10,000 beneficiaries within a few
years to spread the risk.  In another study, HMO officials said that
they do not enroll Medicare beneficiaries in rural areas because the
Medicare population is not large enough to cover the fixed costs
associated with this coverage.\17



                          Table 3
          
            Selected Characteristics of Counties
              With Lower AAPCC Rates and Lower
                  Enrollment in Risk HMOs

Characteristic
--------------------------------------------  ------------
Lower payment/lower enrollment counties              2,257
Counties in an MSA                                     16%
Rural counties                                         84%
Median population per square mile                       31
Average Medicare beneficiaries                       5,603
Counties with 10,000 or more Medicare                  13%
 beneficiaries
----------------------------------------------------------
Sources:  Medicare Market Penetration Report File, AAPCC Rate File,
and other selected files, Office of Managed Care, HCFA; U.S.  Bureau
of the Census, County and City Data Book:  1994 (Washington, D.C.: 
U.S.  Government Printing Office, 1994); and U.S.  Bureau of the
Census data. 


--------------------
\15 See Thomas C.  Ricketts, Rebecca T.  Slifkin, and Karen D. 
Johnson-Webb, "Patterns of Health Maintenance Organization Service
Areas in Rural Counties," Health Care Financing Review, Vol.  17, No. 
1 (Fall 1995), p.  110. 

\16 We used state data on total HMO enrollment because reliable data
for all counties were not available. 

\17 See Serrato, Brown, and Bergeron, p.  93. 


   CERTAIN FACTORS FOSTER RISK HMO
   ENROLLMENT IN SOME AREAS
   DESPITE LOWER PAYMENT RATES
------------------------------------------------------------ Letter :5

In a number of counties where AAPCC payments were relatively low, the
rate of Medicare risk HMO enrollment was relatively high, our
analysis showed.  Risk HMO enrollment flourished because of other
factors, including urban settings and stronger HMO presence.  Most of
these primarily urban counties were in the West, where HMOs have a
longer history than in many other parts of the country.  The counties
in and around the Portland MSA and three other MSAs in Oregon
illustrate how factors other than payment rates serve to raise risk
HMO enrollment. 

A large number of Medicare beneficiaries were enrolled in risk HMOs
that serve more urban, lower payment counties.  About 400,000
beneficiaries--nearly 18 percent of those eligible--were enrolled in
risk HMOs in 92 counties that had lower payment rates (less than
$375) and higher enrollment (greater than 5 percent).  Most of these
92 counties were urban--or closer to urban areas--rather than rural;
about 62 percent were part of the 35 MSAs shown in figure 2.  Another
33 percent bordered these or other MSAs.  Some of the 35 MSAs were
scattered throughout the country, but most were in the West. 

   Figure 2:  35 MSAs With One or
   More Counties With Lower AAPCC
   Rates and Higher Enrollment in
   Risk HMOs, December 1995

   (See figure in printed
   edition.)

\a One county in the Boston metropolitan area--Bristol County--had a
lower AAPCC payment and higher enrollment.  The county is on the
Rhode Island border near the Providence MSA.  Generally, the counties
in the Boston MSA had relatively high AAPCC rates.  We discuss the
Boston MSA later in this report. 

\b One county in the Washington MSA--Loudoun County, Virginia--had a
lower AAPCC payment and higher enrollment.  The county is on the West
Virginia border and is primarily rural with the exception of the
Leesburg and Sterling Park areas, which are within commuting distance
of Washington, D.C. 

Sources:  Medicare Market Penetration Report File, AAPCC Rate File,
and other selected files, Office of Managed Care, HCFA. 

In many of the 35 MSAs, HMO presence was high, which is favorable to
enrollment in risk HMOs.  Table 4 illustrates the total HMO
enrollment for 10 of the 35 MSAs.  Every county included in 8 of
these 10 MSAs had lower AAPCC payments and higher enrollment in risk
HMOs--the two exceptions were the Santa Fe MSA (where one of two
counties did not have lower AAPCC payments and higher enrollment) and
the Denver MSA (where three of five counties did not have lower AAPCC
payments and higher enrollment).  While most of the 10 MSAs had a
relatively high percentage of their total populations enrolled in
HMOs, enrollment was particularly high in four Oregon MSAs.  In the
Oregon MSAs, total HMO enrollment ranged from about 42 percent in
Portland-Vancouver to about 56 percent in Medford-Ashland. 



                          Table 4
          
               Risk HMO Enrollment, Total HMO
            Enrollment, and Number of HMOs in 10
                         MSAs, 1995

                    Percentage
                   of eligible
                      Medicare    Percentage
                  beneficiarie     estimated
                    s enrolled     total HMO     Number of
                       in risk    enrollment  HMOs serving
MSA/state               HMOs\a      in MSA\b         MSA\c
----------------  ------------  ------------  ------------
Portland-                 41.3          41.8            10
 Vancouver, OR-
 WA
Albuquerque, NM           33.6          33.2             5
Denver, CO                30.3          27.4             9
Salem, OR                 19.5          41.9             6
Boulder-                  18.6          49.2             4
 Longmont, CO
Pueblo, CO                17.3          26.0             4
Santa Fe, NM              16.6           8.2             4
Medford-                  12.8          55.9             5
 Ashland, OR
Colorado                  11.7          19.1             4
 Springs, CO
Eugene-                   10.8          52.0             5
 Springfield, OR
----------------------------------------------------------
\a As of December. 

\b As of January, includes estimated commercial, Medicare, and
Medicaid enrollment. 

\c The same HMO can serve more than one MSA.  Data were as of
January. 

Sources:  Medicare Market Penetration Report File, AAPCC Rate File,
and other selected files, Office of Managed Care, HCFA; and
InterStudy, Competitive Edge, 5.2 ed.  (Minneapolis, Minn.:  1995). 


      PORTLAND AREA'S EXPERIENCE
      ILLUSTRATES ROLE OF GREATER
      HMO PRESENCE
---------------------------------------------------------- Letter :5.1

We took a closer look at Portland, an area where higher enrollment in
risk HMOs coexisted with an equally strong total HMO presence and
lower AAPCC payment rates.  Portland's risk HMO enrollment rate of
about 41.3 percent was among the highest in the country, even though
its payment rates were relatively low.\18

As table 5 shows, Portland and the three other Oregon MSAs have had
an active risk program.  Four of the counties in the Portland market
rank among the top seven counties nationwide in terms of the
percentage of their Medicare beneficiaries enrolled in risk HMOs. 



                          Table 5
          
             Risk HMO Enrollment for 13 Oregon
                  Counties, December 1995

                                             Percentage of
                                                  eligible
                                                  Medicare
                                             beneficiaries
                                               enrolled in
                              County             risk HMOs
                              ------------  --------------
MSA
----------------------------------------------------------
Portland-Vancouver\a          Columbia                45.2
                              Washington              44.8
                              Multnomah               44.7
                              Clackamas               44.1
                              Yamhill\                21.8
Salem                         Marion                  20.4
                              Polk                    15.0
Medford-Ashland               Jackson                 12.8
Eugene-Springfield            Lane                    10.8

Non-MSA
----------------------------------------------------------
Borders Salem and Eugene-     Benton                  35.8
 Springfield
Borders Portland-Vancouver    Clatsop                 16.2
Borders Salem and Eugene-     Linn                    15.2
 Springfield
Borders Eugene-Springfield    Douglas                  6.0
 and Medford-Ashland
----------------------------------------------------------
\a Clark County, Washington, is also a part of the Portland-Vancouver
MSA.  It had risk enrollment of 30.2 percent. 

Sources:  Medicare Market Penetration Report File, AAPCC Rate File,
and other selected files, Office of Managed Care, HCFA; and U.S. 
Bureau of the Census, County and City Data Book:  1994 (Washington,
D.C.:  U.S.  Government Printing Office, 1994). 

Portland and Oregon's three other MSAs are clustered along the
Interstate 5 corridor in the western part of the state where risk HMO
enrollment has concentrated.  Figure 3 shows that the nine counties
in the four Oregon MSAs--Portland-Vancouver, Salem,
Eugene-Springfield, and Medford-Ashland--all had higher enrollment in
risk HMOs.  Where higher enrollment has occurred outside the MSAs, in
every case it has been in a county bordering one of Oregon's MSAs.  A
substantial share of the populations in the bordering counties has
access to HMO providers in the neighboring MSAs. 

   Figure 3:  13 Oregon Counties
   With Lower AAPCC Rates and
   Higher Enrollments in Risk
   HMOs, December 1995

   (See figure in printed
   edition.)

Sources:  Medicare Market Penetration Report File, AAPCC Rate File,
and other selected files, Office of Managed Care, HCFA; and U.S. 
Bureau of the Census, County and City Data Book:  1994 (Washington,
D.C.:  U.S.  Government Printing Office, 1994). 

About 75 percent of Oregon's Medicare beneficiaries live in the 13
counties.  HMOs have thus far expressed no interest in expanding into
the eastern parts of the state because the numbers of beneficiaries
needed to induce an HMO to enter into a risk contract with HCFA
appear to be absent. 

In the Portland area, enrollment of Medicare beneficiaries in risk
HMOs seems to have been facilitated by consumer acceptance of HMOs. 
The Portland community's familiarity with HMOs--Kaiser introduced the
concept in Portland in the 1940s--is believed to have increased the
willingness of beneficiaries to participate in the risk program. 
Officials of two Oregon risk HMOs with whom we spoke both cited
beneficiaries' familiarity with the HMO concept as the primary reason
for the high enrollment rate.  Consistent with consumers' apparently
favorable attitude toward HMOs, beneficiaries were willing to enroll
in risk HMOs even though most charge a premium.\19

Because Portland's HMOs were well-established for many years, they
could participate in the risk program when it began.  This early
participation has increased beneficiary acceptance of HMOs and
allowed more time for enrollment to develop and grow, according to
HMO officials we spoke with.  Kaiser participated in a Medicare risk
demonstration project in 1980 and has been enrolling beneficiaries
ever since.  Even before 1980, Kaiser enrolled beneficiaries through
a cost contract with HCFA.  Two other HMOs in the Portland area began
to participate in the risk program in the mid-1980s. 

Employers have played a role in fostering Portland's risk enrollment. 
About half of Kaiser's risk HMO enrollees come from employer groups. 
As employees retire and age into Medicare, they are able to remain
covered by Kaiser, the plan through which they may have been
receiving their health care coverage for many years.  In contrast,
PacifiCare, which has about as many risk enrollees as Kaiser,
receives only about 5 percent of its enrollment from employer groups. 
Its enrollment grew more than Kaiser's between 1993 and 1995,
however, primarily because of its marketing to the beneficiaries not
enrolled through employers. 


--------------------
\18 Like Portland, and the surrounding MSAs in Oregon, several MSAs
in other western states had relatively low AAPCC rates and higher
enrollment in risk HMOs.  Appendix II provides additional information
on these and similar areas. 

\19 HMOs that project Medicare profit rates near or below those for
commercial enrollees are unlikely to offer the richer level of
benefits that HMOs offer where the AAPCC rate and profit levels are
considerably higher. 


   LACK OF HMO PRESENCE CAN IMPEDE
   RISK HMO ENROLLMENT WHERE
   PAYMENTS ARE HIGHER
------------------------------------------------------------ Letter :6

Higher payment rates were no guarantee that risk HMO enrollment would
also be higher.  Our analysis showed that a third of the counties
ranked among the highest AAPCC payment areas in the country had no,
or virtually no, Medicare beneficiaries enrolled in risk HMOs.  About
82 percent of these counties were in the South, where HMOs generally
have not achieved the high levels of enrollment attained by those in
the West.  The lack of a strong HMO presence contributed to the lower
enrollment in risk HMOs that occurred in these counties. 

Few Medicare beneficiaries were enrolled in risk HMOs in one-third of
the counties that were among the highest payment areas.  Our analysis
showed that 1 percent or less of the eligible beneficiaries were
enrolled in risk HMOs in 33 counties that numbered among the top 100
AAPCC payment areas in 1995.  (See table 1 for our payment/enrollment
analysis.) Twenty-seven of these counties were in the South; of
these, 12 were located in 8 MSAs--Atlanta,
Biloxi-Gulfport-Pascagoula, Birmingham, Chattanooga, Lubbock,
Nashville, Savannah, and Stubenville-Weirton--and 9 bordered these or
other MSAs. 

HMO presence, a factor facilitating risk HMO enrollment, was
relatively low in most of the eight southern MSAs.  Even larger
southern metropolitan areas like Atlanta and Birmingham, with about
15 percent total HMO enrollment, had lower Medicare risk HMO
enrollment in most of their areas.  Atlanta, where 3 of the 20
counties in the MSA had higher payment rates, still had risk HMO
enrollment below 1 percent.  Only one of the four counties in the
Birmingham MSA had a higher payment rate, but risk HMO enrollment in
this county and two others was minimal.\20


--------------------
\20 Jefferson County, the primary county in the Birmingham MSA, had
an intermediate payment rate (between $375 and $464) and risk HMO
enrollment of 7.3 percent.  The remaining three counties in the
Birmingham MSA had risk enrollment of less than 1 percent. 


   EMPLOYER POLICIES CAN HINDER
   RISK HMO ENROLLMENT WHEN OTHER
   FACTORS ARE FAVORABLE
------------------------------------------------------------ Letter :7

In southeastern Michigan, higher payment rates in the Detroit, Flint,
and Ann Arbor MSAs have not been enough to stimulate risk HMO
enrollment.  However, this modest enrollment performance cannot be
ascribed to weak HMO presence.  Compared with HMOs in some southern
MSAs, HMOs in Michigan have generated stronger total enrollment--as
high as about 26 percent in Flint and Ann Arbor.  While the
combination of attractive payment rates, a large potential market,
and an active HMO industry are all present, the Medicare risk program
has been slow to take root.  In these Michigan MSAs, the retiree
benefits provided by the automobile industry reduces the
attractiveness of risk HMOs and contributes to lower enrollment by
Medicare beneficiaries. 

Six of the 33 counties that had lower enrollment in risk HMOs,
despite being among the 100 highest payment areas, were in Michigan. 
All six of these counties were in three MSAs--Ann Arbor, Detroit, and
Flint.  In addition to being higher payment, urban areas, these three
MSAs had a fairly strong HMO presence.  Both Ann Arbor and Flint had
total HMO enrollment of about 26 percent; in Detroit, enrollment
approached 21 percent.  While low compared with some areas in the
West, this level of total HMO enrollment was considerably higher than
in most of the eight southern MSAs that also had higher payment rates
and lower enrollment in risk HMOs.  Despite the presence of these
factors, only about 0.5 percent of the eligible Medicare
beneficiaries in the three Michigan MSAs were enrolled in risk HMOs. 

A key factor in the Michigan MSAs, according to HMO officials, is the
benefit package available to retired autoworkers and retirees from
other firms that patterned their benefit packages on the auto
industry's.  These benefit packages provide employer-sponsored
comprehensive coverage of health benefits with little or no
out-of-pocket payments for the retiree.  As officials at the HMOs we
interviewed in Detroit confirmed, beneficiaries receiving these
benefits have little incentive to switch to risk HMOs. 

Excluding beneficiaries covered by rich benefit packages, the
Medicare population in the Michigan MSAs is still relatively large
and hence attractive to risk HMOs.  Two plans whose officials we
spoke with had left the risk program in the late 1980s because they
were losing money but now have risk contract applications pending
with HCFA and hope to begin enrolling beneficiaries in 1997.  The
HMOs plan to target Medicare beneficiaries not covered by rich
retiree health benefits.  One HMO suggested that by the end of 1998
it expects about 30,000 Medicare beneficiaries to be enrolled. 
However, if the HMO could contract with one of the "big three"
automakers, it expects that number to double. 


   HMO BUSINESS STRATEGY MAY
   AFFECT RISK ENROLLMENT
------------------------------------------------------------ Letter :8

Several multistate HMOs have been pursuing the Medicare risk market. 
As the market for employer-sponsored health coverage has become more
saturated, HMOs have realized that the Medicare market is large and
potentially lucrative, that the demand for a Medicare risk product is
increasing, and that competitors are ready to move into the market. 
In addition, HMOs have been encouraged by increases in the AAPCC
payment rates in some areas.  Finally, publicly traded HMO companies,
which are especially concerned with short-term profits, are seeking
new ways in which to expand and grow. 

Companies' business strategies for expanding their involvement with
the risk HMO program include enlarging existing risk HMO contracts
through service area expansions, applying for new risk contracts, and
acquiring other HMOs.  Officials of one large multistate HMO
articulated a strategy of expanding from existing service areas into
contiguous areas and contiguous states.  While acquisitions were for
the most part made to obtain a share of the commercial managed care
market, with the decision to start a risk program coming later, some
acquisitions were made specifically with the Medicare risk market in
mind.  For example, PacifiCare Health Systems announced in August
1996 that it would buy FHP International in part to ensure its
dominance in the Medicare risk market. 

According to the officials we interviewed at multistate HMOs, their
risk HMO expansion efforts targeted markets with certain
characteristics.  Markets were more attractive if they had a
concentrated number of beneficiaries and limited competition for
them.  To spread the risk, most said that an HMO must have the
opportunity to enroll at least 10,000 beneficiaries in a few years. 
Also, officials at the HMOs said that the payment rate in a market
must be high enough for a risk plan to be financially viable. 
Furthermore, markets must have a concentration of non-Medicare
beneficiaries to be attractive because HMOs must have at least 5,000
commercial enrollees to apply for a risk contract.  And the more
people who are enrolled in the commercial sector of a local market,
the easier it is to enroll beneficiaries because of the familiarity
with managed care. 

Officials at multistate HMOs had different views of the role that
payment rates play in their decisions to move into or out of a
market.  Officials at one multistate company said that while they
consider the AAPCC payment rate when making a decision, a low payment
would not automatically disqualify a market.  For example, they said
they were considering expanding into a low-payment market because an
employer group had specifically expressed interest in the risk
program in that area.  Another company indicated that the payment
rate needs to be high enough to adequately pay health care providers. 
Officials at a third multistate company believed that there were no
parts of the country where the payment rate would be too low for them
to enter if the large numbers of Medicare beneficiaries needed for
the program to be successful were present.  Officials at two
multistate HMOs held divergent views on the AAPCC rate reduction that
would induce their companies to leave the risk HMO program.  While
officials at one company, citing "slim margins," thought a moderate
rate reduction would induce them to switch to a Medicare cost
contract, officials at another company stated that they would
consider leaving the risk contract program only if the rate reduction
was "drastic."

Expansion by dominant Medicare HMO companies can help fuel the growth
of risk enrollment in locations where it has been slow.  In Boston,
for example, HMO and HCFA officials credit the entrance of a
PacifiCare risk product with sparking the growth of Boston's risk
market.  As the PacifiCare product began obtaining market share,
local HMOs realized they needed to target the Medicare market more
aggressively if they wanted to stay competitive.  Their pursuit of
the Boston market resulted in greatly increased risk enrollment. 


      BOSTON'S RISK PROGRAM
      STIMULATED WHEN
      BENEFICIARIES COULD ENROLL
      AT NO COST
---------------------------------------------------------- Letter :8.1

The Boston risk market is an attractive market because of the large
base of Medicare beneficiaries, some of the highest AAPCC rates in
the country, and a strong HMO presence.  However, the number of risk
HMOs pursuing the Boston market peaked during 1987 through 1988, then
declined until 1993, when only three HMOs were left serving the
market.  According to HCFA officials, a number of HMOs dropped out of
the risk program because they did not understand how to manage the
senior population and did not control enrollee costs. 

The market started to turn around in Boston in 1994.  According to
HCFA and HMO officials, HMOs were returning to the risk market or
entering for the first time because the industry was learning how to
manage the Medicare business.  However, it took a new entrant to the
risk market, challenging the market share of the established HMOs,
for risk enrollment in Boston to take off. 

Tufts Associated reentered Boston's risk market in 1994,\21 with a
franchise of SecureHorizons, the Medicare HMO product of
California-based PacifiCare.  To attract beneficiaries, Tufts began
offering a zero-premium risk product--Medicare beneficiaries could
enroll without having to pay any additional premiums to the HMO.  The
introduction of a zero-premium product transformed the Boston market,
according to HCFA and HMO officials.  Tufts' action created
competition among HMOs for Medicare beneficiaries and spurred
stronger beneficiary demand for risk products.  As a result,
enrollment in the MSA grew 158 percent between 1993 and 1995.  Prior
to the introduction of the zero-premium product, HMO risk premiums
had ranged from $90 to $120 per month, and according to HCFA
officials, were hardly distinguishable from premiums for Medigap
insurance. 


--------------------
\21 Tufts Associated HMO began operations in the Boston area in 1981
and started a risk program in 1985.  However, it dropped out of the
risk program in 1988.  According to Tufts officials, the HMO was
unsuccessful in the risk program because it did not develop a
specifically tailored program to manage seniors as a distinctly
different population from the non-elderly and failed to manage them
effectively.  As a result, Tufts had lost money.  Like other HMOs,
Tufts began reconsidering the risk market in 1992 as it became
apparent that some HMOs were able to manage their risk programs
profitably.  Tufts negotiated a franchise of PacifiCare's
SecureHorizons product as a way of gaining the expertise necessary to
run a successful Medicare risk program. 


   CONCLUSIONS
------------------------------------------------------------ Letter :9

Despite the considerable momentum of risk HMO enrollment growth, its
uneven pattern across the country focuses attention on understanding
why such disparities occur.  Although the linkage of Medicare payment
rates to risk HMO enrollment may be important in some areas, dramatic
differences in enrollment are often associated with other factors. 
The presence of HMOs, the density of population, and the number of
Medicare beneficiaries, especially those familiar with managed health
care, all facilitate growth in enrollment--and their absence hinders
it.  In addition, the health care benefits provided by employers in a
market area can affect beneficiaries' willingness to enroll in risk
HMOs.  The rapid growth in risk HMO enrollment during the past
several years, which has occurred without any major federal policy
changes, is likely to continue as employers encourage retirees to
join HMOs and as HMOs pursue varied strategies for expanding their
Medicare business. 


---------------------------------------------------------- Letter :9.1

We provided a draft of this report to officials in HCFA's Office of
Managed Care.  These officials agreed with the information presented. 
We are sending copies of this report to the Secretary of Health and
Human Services and other interested parties, and we will make copies
available to others on request.  If you or your staff have any
questions, please call me at (202) 512-7114 or Michael F.  Gutowski,
Assistant Director, at (202) 512-7128.  Other major contributors to
this report include Howard Cott, Aleta Hancock, Joseph Petko, Wayne
Turowski, and Joan Vogel. 

Sincerely yours,

Jonathan Ratner
Associate Director, Health Financing and
 Systems Issues


METHODOLOGY
=========================================================== Appendix I

Health maintenance organizations (HMO) enter into risk contracts with
the Health Care Financing Administration (HCFA) to provide
Medicare-covered services to beneficiaries who enroll.\22 Risk plans
assume all the financial risk associated with providing
Medicare-covered services to enrolled beneficiaries in return for a
monthly per capita premium--the adjusted average per capita cost
(AAPCC) payment--from HCFA for each Medicare beneficiary enrolled. 
The amount of these AAPCC-based payments varies by county.  HCFA pays
some HMOs on a cost-reimbursement basis, with these HMOs assuming no
risk that fees will be insufficient to cover costs; our work focused
only on managed care plans that had entered into risk contracts with
HCFA. 

We first obtained data from the Office of Managed Care at HCFA
headquarters.  For each of 3,141 counties in the nation, these data
showed the (1) number of eligible Medicare beneficiaries, (2) number
of beneficiaries enrolled in managed care plans with a risk contract
with HCFA, and (3) AAPCC rate paid by HCFA for each Medicare
beneficiary enrolled in a risk plan.  These data were as of December
for 1993, 1994, and 1995, the latest data available at the time of
our review.  Data did not include Guam, Puerto Rico, or the Virgin
Islands. 

To identify counties with similar risk enrollment and AAPCC rates, we
determined the percentage of Medicare beneficiaries enrolled in
Medicare risk plans in December 1995 and the AAPCC rates paid by HCFA
in 1995 for each of the 3,141 counties.  Using these data, we placed
each county into one of nine categories based on whether it had a
higher, intermediate, or lower AAPCC rate and higher, intermediate,
or lower Medicare risk enrollment. 

We placed counties with the top 100 AAPCC rates in 1995 in the higher
category.  These rates ranged from $463.89 to $678.90.  There were
2,474 counties with AAPCC rates under $375.00 that we considered as
having lower payments (their rates ranged from $177.32 to $374.86\23
).  We defined the remaining 567 counties as having intermediate
AAPCC rates.  Again using December 1995 data, we placed the 2,663
counties with no Medicare risk enrollment or enrollment of 1 percent
or less in the lower enrollment category.  Of these 2,663 counties,
618 had no Medicare risk enrollment at all, and 2,045 had enrollments
of 1 percent or less.  For the purposes of our study, we placed the
242 counties that had enrollments of more than 5 percent in the
higher enrollment category.  Although 5 percent would not be
considered high for private sector managed care enrollment, it is
high for the Medicare risk program--only about 8 percent of the
counties had Medicare risk enrollments of more than 5 percent.  We
placed the remaining 236 counties in the intermediate category
(enrollments greater than 1 percent and less than or equal to 5
percent).  For the distribution of the 3,141 counties in the nine
categories, see table 1. 

We focused our study on the following three categories:  counties
that had (1) lower AAPCC rates and lower risk enrollment (2,257
counties), (2) lower AAPCC rates and higher risk enrollment (92
counties), and (3) higher AAPCC rates and lower risk enrollment (33
counties).  Using HCFA data, we converted the county data into
metropolitan statistical area (MSA) data for the three study
categories. 

In addition to analyzing data for the three study categories, we
selected three markets in which to perform more detailed work:  (1)
Portland, Oregon--a market in the lower AAPCC rate and higher
enrollment category; (2) Detroit--a market in the higher AAPCC rate
and lower enrollment category; and (3) Boston--a market in which risk
enrollment grew considerably in a relatively short time.  In these
markets, we interviewed officials at the applicable HCFA regional
offices and at selected HMOs that had entered into or were planning
to enter into Medicare risk contracts. 

Finally, we interviewed officials at all 10 HCFA regional offices, 6
national HMO chains, and 12 regional HMOs to obtain more information
and get opinions on (1) the Medicare risk program in general, (2)
enrollment trends in particular, (3) reasons managed care plans and
beneficiaries participate in Medicare risk programs, (4) factors that
made the risk market attractive, and (5) factors affecting risk
program enrollment.  All of the 18 HMOs had risk contracts or were
planning to enter into contracts with HCFA.  Also, we interviewed
representatives from five management consulting firms involved in
bringing employers and Medicare risk plans together to cover
Medicare-eligible retirees and five employers regarding their efforts
to enroll a portion of their retirees in Medicare risk plans.  Table
I.1 shows the HMOs, employers, and management consultants we
interviewed. 



                         Table I.1
          
            National HMO Chains, Regional HMOs,
           Employers, and Management Consultants
                         Contacted

Group contacted                   Location
--------------------------------  ------------------------
National HMO chains
----------------------------------------------------------
FHP, Inc.                         Fountain Valley, CA

Health Systems International      Woodland Hills, CA

Kaiser Foundation Health Plan,    Oakland, CA
Inc.

PacifiCare Health Systems         Cypress, CA

Prudential Health Care Senior     Roseland, NJ
Care

United HealthCare Corporation     Minnetonka, MN


Regional HMOs
----------------------------------------------------------
Blue Care Network of Southeast    Southfield, MI
Michigan

FHP of Utah                       Salt Lake City, UT

Harvard Pilgrim Health Care       Dedham, MA

Health Alliance Plan              Detroit, MI

Intermountain Health Care of      Salt Lake City, UT
Utah

Kaiser Foundation Health Plan,    Oakland, CA
Northern California Region

Kaiser Foundation Health Plan of  Portland, OR
the Northwest

Mercy Health Plan                 Farmington Hills, MI

PacifiCare of Oregon              Lake Oswego, OR

Tufts/SecureHorizons              Waltham, MA

United HealthCare of Utah         Salt Lake City, UT

U.S. Healthcare, New England      Burlington, MA
Region


Employers
----------------------------------------------------------
Chevron Corporation               San Francisco, CA

General Motors Corporation        Detroit, MI

McKesson Corporation              San Francisco, CA

Sears, Roebuck and Company        Hoffman Estates, IL

University of California          Oakland, CA


Management consultants
----------------------------------------------------------
Foster Higgins and Company, Inc.  Los Angeles, CA

KPMG Peat Marwick                 San Francisco, CA

Price Waterhouse                  San Francisco, CA

Towers Perrin                     San Francisco, CA; New
                                  York City, NY

William Mercer, Inc.              Richmond, VA
----------------------------------------------------------
We performed our work between February 1996 and February 1997 in
accordance with generally accepted government auditing standards. 


--------------------
\22 Managed care plans that enter into risk contracts include both
HMOs and competitive medical plans.  Although competitive medical
plans are subject to regulatory requirements similar to those for
HMOs, they have greater flexibility in setting commercial premium
rates and the services offered to their commercial members.  In this
report, our use of the term HMO includes competitive medical plans. 

\23 HCFA data showed that 13 areas, most of which were in Alaska, had
AAPCC rates of zero. 


RISK HMO ENROLLMENT SUBSTANTIAL IN
SEVERAL WESTERN MARKETS WITH
STRONGER HMO PRESENCE
========================================================== Appendix II

Counties in Oregon were not the only counties where risk HMOs had
enrolled substantial numbers of Medicare beneficiaries despite having
lower payment rates.  Other counties in or bordering several western
MSAs also followed this pattern.  As in Oregon, these more urban
counties were located in or bordered MSAs that usually had a strong
total HMO presence--an important factor that can affect risk
enrollment.  The Albuquerque and Santa Fe MSAs in New Mexico and the
Denver, Boulder-Longmont, Colorado Springs, and Pueblo MSAs in
Colorado most closely followed the pattern exhibited in Oregon. 
Clearly, factors other than payment rates affected risk HMO
enrollment in these MSAs. 

Risk HMO enrollment patterns for lower payment/higher enrollment
counties in Washington and Arizona were not nearly as clear as for
those in Oregon.  In Washington, where HMOs generally had a strong
presence, counties exhibited the lower payment/higher risk enrollment
mix, but that enrollment was not as clearly concentrated.  Higher
risk enrollment extended beyond the counties located in MSAs and
bordering MSAs with higher enrollment in risk HMOs to counties not
directly adjacent to these MSAs.  Counties in two Arizona
MSAs--Tucson and Phoenix-Mesa--had higher risk HMO enrollment and, in
Tucson in particular, had a strong HMO presence.  Their payment rates
were in the intermediate category.  Counties bordering the MSAs,
however, had the higher enrollment/lower payment pattern.  The higher
enrollment/lower payment pattern also appeared in counties located in
MSAs in seven other states in different parts of the country. 

NEW MEXICO AND COLORADO:  HIGHER
RISK HMO ENROLLMENT CLUSTERED IN
AND AROUND MSAS

Counties in both New Mexico and Colorado exhibited the pattern
described previously for Oregon.  HMO presence was generally strong
and higher levels of risk HMO enrollment were coupled with lower
payment rates in several of the more urban counties in MSAs with a
few adjoining counties also having higher enrollments in risk HMOs. 
Elsewhere in the two states, enrollment rates were considerably
lower. 

In New Mexico, the counties with higher enrollments in Medicare risk
HMOs were in and around the Albuquerque and Santa Fe MSAs and had
lower AAPCC rates.  (See fig.  II.1.) About 48 percent of New
Mexico's Medicare beneficiaries lived in the six counties in and
around the two MSAs, but about 98 percent of the risk HMO enrollees
in the state lived there.  Risk HMO enrollment was particularly high
in the Albuquerque MSA--about 34 percent of the Medicare
beneficiaries in the MSA's three counties were enrolled. 

   Figure II.1:  Six New Mexico
   Counties With Higher Enrollment
   in Risk HMOs and Lower AAPCC
   Rates, December 1995

   (See figure in printed
   edition.)

Sources:  Medicare Market Penetration Report File, AAPCC Rate File,
and other selected files, Office of Managed Care, HCFA; and U.S. 
Bureau of the Census, County and City Data Book:  1994 (Washington,
D.C.:  U.S.  Government Printing Office, 1994). 

In Colorado, about two-thirds of the Medicare beneficiaries and about
99 percent of the Medicare risk HMO enrollees lived in and around
four MSAs--Boulder-Longmont, Colorado Springs, Denver, and Pueblo. 
(See fig.  II.2.) Risk enrollment was highest in the Denver
MSA--about 30 percent of the Medicare beneficiaries were enrolled in
December 1995.  Table II.1 shows risk HMO enrollment figures for
counties in and around selected New Mexico and Colorado MSAs. 

   Figure II.2:  13 Colorado
   Counties With Higher Enrollment
   in Risk HMOs and Lower and
   Intermediate AAPCC Rates,
   December 1995

   (See figure in printed
   edition.)

Sources:  Medicare Market Penetration Report File, AAPCC Rate File,
and other selected files, Office of Managed Care, HCFA; and U.S. 
Bureau of the Census, County and City Data Book:  1994 (Washington,
D.C.:  U.S.  Government Printing Office, 1994). 



                         Table II.1
          
           Risk HMO Enrollment for Six New Mexico
           and Eight Colorado Counties, December
                            1995

                                                Percentage
                                               of eligible
                                                  Medicare
                                              beneficiarie
                                                s enrolled
                                County        in risk HMOs
                                ------------  ------------
New Mexico
----------------------------------------------------------
Albuquerque                     Sandoval              37.6
                                Valencia              35.8
                                Bernalillo            32.8
Santa Fe                        Santa Fe\a            18.4
Borders Albuquerque and Santa   Torrance              17.6
 Fe
Borders Albuquerque and Santa   Rio Arriba             9.4
 Fe

Colorado
----------------------------------------------------------
Denver                          Jefferson             31.2
                                Douglas\b             22.8
Boulder-Longmont                Boulder               18.6
Pueblo                          Pueblo                17.3
Colorado Springs                El Paso               11.7
Borders Denver and Colorado     Elbert                10.0
 Springs
Borders Denver                  Park                   9.7
Borders Denver and Colorado     Teller                 6.7
 Springs
----------------------------------------------------------
\a The other county in the Santa Fe MSA--Los Alamos--had 4.2 percent
risk HMO enrollment. 

\b The three other counties in the Denver MSA had intermediate AAPCC
rates and higher enrollments in risk HMOs:  Adams County, 34.4
percent enrollment; Arapahoe County, 29.9 percent; and Denver County,
28.3 percent.  Two counties bordering the Denver MSA also had
intermediate AAPCC rates and higher enrollments in risk HMOs:  Gilpin
County, 15 percent enrollment, and Clear Creek County, 11 percent. 

Sources:  Medicare Market Penetration Report File, AAPCC Rate File,
and other selected files, Office of Managed Care, HCFA; and U.S. 
Bureau of the Census, County and City Data Book:  1994 (Washington,
D.C.:  U.S.  Government Printing Office, 1994). 

Of the six New Mexico and Colorado MSAs shown in table II.2, the two
with the highest enrollment in risk HMOs--Albuquerque and
Denver--also had a strong HMO presence. 



                         Table II.2
          
               Risk HMO Enrollment, Total HMO
           Enrollment, and Number of HMOs in New
               Mexico and Colorado MSAs, 1995

                      Percentage
                     of eligible
                        Medicare    Percentage
                    beneficiarie     estimated   Number of
                      s enrolled     total HMO        HMOs
                         in risk    enrollment     serving
MSA/state                 HMOs\a      in MSA\b       MSA\c
------------------  ------------  ------------  ----------
Albuquerque, NM             33.6          33.2           5
Denver, CO                  30.3          27.4           9
Boulder-Longmont,           18.6          49.2           4
 CO
Pueblo, CO                  17.3          26.0           4
Santa Fe, NM                16.6           8.2           4
Colorado Springs,           11.7          19.1           4
 CO
----------------------------------------------------------
\a Data as of December. 

\b Includes estimated commercial, Medicare, and Medicaid enrollment;
data as of January. 

\c The same HMO can serve more than one MSA; data as of January. 

Sources:  Medicare Market Penetration Report File and other selected
files, Office of Managed Care, HCFA; and InterStudy, Competitive
Edge, 5.2 ed.  (Minneapolis, Minn.:  1995). 

WASHINGTON:  RISK HMO ENROLLMENT
EXTENDS BEYOND MSAS

As in Oregon, New Mexico, and Colorado, higher enrollment in risk
HMOs in Washington was primarily concentrated in counties with lower
AAPCC rates that were also in and around
MSAs--Seattle-Bellevue-Everett,\24 Tacoma, Olympia, and Bremerton in
the western part of the state and Spokane in the eastern part.\25 But
as figure II.3 shows, in Washington higher enrollment in risk HMOs
also exists in counties that are neither in nor adjacent to MSAs with
higher enrollment. 

   Figure II.3:  19 Washington
   Counties With Higher Enrollment
   in Risk HMOs and Lower and
   Intermediate AAPCC Rates,
   December 1995

   (See figure in printed
   edition.)

Sources:  Medicare Market Penetration Report File, AAPCC Rate File,
and other selected files, Office of Managed Care, HCFA; and U.S. 
Bureau of the Census, County and City Data Book:  1994 (Washington,
D.C.:  U.S.  Government Printing Office, 1994). 

Table II.3 shows that HMO presence was relatively strong in several
of the six Washington MSAs.  Overall, total HMO enrollment ranged
from about 50 percent in Olympia to about 11 percent in Tacoma. 



                         Table II.3
          
               Risk HMO Enrollment, Total HMO
           Enrollment, and Number of HMOs in Six
                   Washington MSAs, 1995

                      Percentage
                     of eligible
                        Medicare    Percentage
                    beneficiarie     estimated   Number of
                      s enrolled     total HMO        HMOs
                         in risk    enrollment     serving
MSA/state                 HMOs\a      in MSA\b       MSA\c
------------------  ------------  ------------  ----------
Portland-                   41.3          41.8          10
 Vancouver, OR-
 WA\d
Seattle-Bellevue-           24.7          17.3           8
 Everett, WA
Olympia, WA                 20.3          50.3           6
Tacoma, WA                  12.4          11.1           4
Bremerton, WA                8.8          18.5           3
Spokane, WA                  6.8          36.0           7
----------------------------------------------------------
\a Data as of December. 

\b Includes estimated commercial, Medicare, and Medicaid enrollment;
data as of January. 

\c The same HMO can serve more than one MSA; data as of January. 

\d Only one Washington county--Clark--is in this MSA.  The remaining
counties in this MSA are in Oregon. 

Sources:  InterStudy, Competitive Edge, 5.2 ed.  (Minneapolis, Minn.: 
1995); and Medicare Market Penetration Report File and other selected
files, Office of Managed Care, HCFA. 

ARIZONA:  RISK HMO ENROLLMENT
CONCENTRATED IN MSAS WITH HIGHER
PAYMENT RATES

In Arizona, higher enrollment in risk HMOs was primarily concentrated
in two MSAs--Tucson and Phoenix-Mesa.  The counties in the MSAs all
had higher AAPCC rates than those in such MSAs as Albuquerque and
Portland.  The payment rates in Tucson and Phoenix-Mesa fell in the
intermediate category.  But for the counties outside Arizona's MSAs
where enrollment was higher for risk HMOs, the payment rates were
usually lower.  Figure II.4 shows the Arizona counties with higher
enrollments in risk HMOs. 

   Figure II.4:  Nine Arizona
   Counties With Higher Enrollment
   in Risk HMOs and Lower and
   Intermediate AAPCC Rates,
   December 1995

   (See figure in printed
   edition.)

Sources:  Medicare Market Penetration Report File, AAPCC Rate File,
and other selected files, Office of Managed Care, HCFA; and U.S. 
Bureau of the Census, County and City Data Book:  1994 (Washington,
D.C.:  U.S.  Government Printing Office, 1994). 

Tucson, which had an especially high enrollment for risk HMOs--nearly
42 percent--also had a strong HMO presence as table II.4 shows. 



                         Table II.4
          
             Risk HMO Enrollment and Total HMO
           Enrollment for Nine Arizona Counties,
                            1995

                                  Percentage
                                 of eligible
                                    Medicare    Percentage
                                beneficiarie     estimated
                                  s enrolled     total HMO
                                     in risk    enrollment
                    County            HMOs\a      in MSA\b
                    ----------  ------------  ------------
Tucson              Pima\c              41.7          42.0
Phoenix-Mesa        Maricopa\c          34.7          25.3
                    Pinal\c             33.7
Borders Tucson and  Graham              28.5            \d
 Phoenix-Mesa
Borders Tucson      Santa Cruz          26.0            \d
Borders neither     Greenlee            24.4            \d
 Tucson nor
 Phoenix-Mesa
Borders Tucson      Cochise             21.8            \d
Borders Phoenix-    Gila\c              21.9            \d
 Mesa
Las Vegas           Mohave\c            10.0          20.5
----------------------------------------------------------
\a Data as of December. 

\b Includes estimated commercial, Medicare, and Medicaid enrollment;
data as of January. 

\c This county had an intermediate AAPCC rate--between $375 and $464. 

\d Not applicable. 

Sources:  Medicare Market Penetration Report File, AAPCC Rate File,
and other selected files, Office of Managed Care, HCFA; U.S.  Bureau
of the Census, County and City Data Book:  1994 (Washington, D.C.: 
U.S.  Government Printing Office, 1994); and InterStudy, Competitive
Edge, 5.2 ed.  (Minneapolis, Minn.:  1995). 

HIGHER ENROLLMENT IN RISK HMOS IN
OTHER MSAS WITH LOWER PAYMENT
RATES

Counties in MSAs in California, Florida, Hawaii, Minnesota, Oklahoma,
Pennsylvania, and Texas had lower AAPCC rates but higher percentages
of beneficiaries enrolled in risk HMOs.  Minneapolis-St.  Paul had a
large number of Medicare beneficiaries enrolled in risk HMOs.  Risk
enrollment in several southern California and Florida MSAs was also
higher despite lower AAPCC payment rates.  Even parts of several MSAs
in Pennsylvania had higher enrollments in risk HMOs despite lower
payment rates. 

Table II.5 shows the risk HMO enrollment rates for the counties in
the Minneapolis-St.  Paul MSA.  Risk enrollment was higher--about 19
percent--compared with many areas of the country but not nearly as
high as in several western MSAs even though total HMO enrollment in
the Minneapolis-St.  Paul MSA was close to 40 percent. 



                         Table II.5
          
          Risk HMO Enrollment for All Counties in
           the Minneapolis-St. Paul MSA, December
                            1995

                                             Percentage of
                                         eligible Medicare
                                             beneficiaries
                                          enrolled in risk
County/state                                          HMOs
--------------------------------------  ------------------
Anoka, MN                                             26.1
Ramsey, MN\a                                          22.8
Hennepin, MN                                          20.0
Washington, MN                                        19.2
Dakota, MN                                            17.0
Scott, MN                                              9.3
Carver, MN                                             8.4
Sherburne, MN                                          6.9
Chisago, MN                                            6.4
Isanti, MN                                           3.6\b
Wright, MN                                           3.2\b
St. Croix, WI                                        1.1\b
Pierce, WI                                           0.6\c
----------------------------------------------------------
\a This county had an intermediate AAPCC rate--between $375 and $464. 

\b Intermediate risk enrollment. 

\c Lower risk enrollment. 

Sources:  Medicare Market Penetration Report File, AAPCC Rate File,
and other selected files, Office of Managed Care, HCFA. 

Risk enrollment patterns were less clear in the remaining six states. 
Table II.6 compares the total HMO enrollment and risk HMO enrollment
where at least one county in an MSA had more than 5 percent of its
Medicare beneficiaries enrolled in a risk HMO and where the AAPCC
rate was in the lower payment category.  These MSAs had varying
degrees of risk enrollment ranging from being higher in Florida to
lower in Pennsylvania. 



                         Table II.6
          
           17 MSAs With One or More Counties That
            Received Lower or Intermediate AAPCC
               Rates and Had Higher Risk HMO
                      Enrollment, 1995

                    Percentage
                   of eligible
                      Medicare    Percentage
                  beneficiarie     estimated
                    s enrolled     total HMO     Number of
                       in risk    enrollment  HMOs serving
MSA/state               HMOs\a      in MSA\b         MSA\c
----------------  ------------  ------------  ------------
Daytona Beach,            31.8          14.7             3
 FL
Santa Barbara-            31.2          23.1             6
 Santa Maria-
 Lompoc, CA
San Luis Obispo-          28.6           2.9             2
 Atascadero-
 Paso Robles, CA
San Antonio, TX         27.1\d          12.5             7
Fort Worth-             14.7\e          14.1             8
 Arlington, TX
Gainesville, FL           11.9          13.9             1
Tulsa, OK               11.2\f          13.4             5
Williamsport, PA          11.2           7.9             2
Fresno, CA                 9.9          22.0             8
State College,             9.4          11.3             3
 PA
Honolulu, HI               9.1          23.3             7
Ocala, FL                  9.0            \g            \g
Yuba City, CA            7.5\h           5.5             3
Scranton-Wilkes          5.9\i          10.3             2
 Barre-
 Hazleton, PA
Oklahoma City,             5.1          11.4             5
 OK
Austin-San               4.8\j          26.3             4
 Marcos, TX
Harrisburg-              3.1\k          14.1             2
 Lebanon-
 Carlisle, PA
----------------------------------------------------------
\a Data as of December. 

\b Includes estimated commercial, Medicare, and Medicaid enrollment;
data as of January. 

\c The same HMO can serve more than one MSA; data as of January. 

\d Includes Bexar County, which has higher risk HMO enrollment and an
intermediate AAPCC rate. 

\e Includes Tarrant County, which has higher risk HMO enrollment and
an intermediate AAPCC rate. 

\f Includes Tulsa and Wagoner counties, which have higher risk HMO
enrollment and intermediate AAPCC rates. 

\g Data not available. 

\h Includes Yuba County, which had an intermediate AAPCC rate. 

\i Includes Lackawanna, Luzerne, and Wyoming counties, which all have
higher risk HMO enrollment and intermediate AAPCC rates. 

\j Includes Travis County, which had a risk HMO enrollment rate of
6.7 percent. 

\k Lebanon County had a risk HMO enrollment rate of 5.2 percent.  The
three remaining counties in the MSA had the following risk enrollment
and AAPCC rate combinations:  Cumberland County, intermediate
enrollment/lower AAPCC; Dauphin County, intermediate enrollment and
AAPCC; and Perry County, lower enrollment/intermediate AAPCC. 

Sources:  Medicare Market Penetration Report File, AAPCC Rate File,
and other selected files, Office of Managed Care, HCFA; U.S.  Bureau
of the Census, County and City Data Book:  1994 (Washington, D.C.: 
U.S.  Government Printing Office, 1994); and InterStudy, Competitive
Edge, 5.2 ed.  (Minneapolis, Minn.:  1995). 


--------------------
\24 One county, King, had higher enrollment in risk HMOs and
intermediate AAPCC rates.  King is part of the
Seattle-Bellevue-Everett MSA. 

\25 Clark County, Washington, is part of the Portland-Vancouver MSA. 


*** End of document. ***