Medicare+Choice: Plan Withdrawals Indicate Difficulty of Providing Choice
While Achieving Savings (Letter Report, 09/07/2000, GAO/HEHS-00-183).
Pursuant to a congressional request, GAO reviewed health care plans'
withdrawal from the Medicare Choice program, focusing on the: (1)
geographic distribution and the distribution among plans of enrollees
affected by the recent plan withdrawals; (2) factors associated with
plans that terminated or reduced their participation in the program; and
(3) likely role of payment rates in affecting plans' decisions.
GAO noted that: (1) of 309 plans serving Medicare beneficiaries at the
end of 1999, 99 plans terminated their contracts or reduced the number
of counties they served for the 2000 contract year, and 118 have
announced they will terminate their contracts or reduce service areas
for the 2001 contract year; (2) these withdrawals affected about 328,000
enrollees in 2000 and will affect almost 1 million enrollees in 2001;
(3) the number of enrollees affected accounts for about 5 percent of
Medicare Choice enrollees in 2000 and about 15 percent in 2001; (4) a
disproportionate number of affected enrollees live outside of major
urban areas; (5) a portion of these enrollees, approximately 79,000 in
2000 and 159,000 in 2001, will have no other Medicare managed care
option available in their area and must either switch to a non-managed
care option, if one is available in their area, or return to traditional
fee-for-service (FFS) Medicare; (6) while a new private FFS plan has
begun to offer services in many of the affected areas as an alternative
to the traditional public FFS Health Care Financing Administration
(HCFA) manages, does not offer a prescription drug benefit; (7) in
January 2000, Medicare Choice plans tended to withdraw from more
difficult to serve rural counties or large urban areas that they had
entered more recently or where they failed to attract sufficient
enrollment; (8) in 2001, the trend is essentially the same for the
service area reductions but somewhat different for the contract
terminations, which involve some older, more established plans; (9) the
pattern of Medicare Choice withdrawals shares common elements with plan
participation in the similarly choice-based health insurance program for
federal employees; (10) industry representatives contend that the
Balanced Budget Act's (BBA) payment rate changes were too severe and
that low Medicare payment rates are largely responsible for the plan
withdrawals; (11) however, since the BBA was enacted, Medicare Choice
payment rates have risen faster than per capita FFS spending; (12) in
addition, many plans have attracted beneficiaries who have
lower-than-average expected health care costs, while Medicare Choice
payments are largely based on the expected cost of beneficiaries with
average health care needs; and (13) it is unclear whether Medicare
Choice payment rate increases would affect plans' participation
decisions.
--------------------------- Indexing Terms -----------------------------
REPORTNUM: HEHS-00-183
TITLE: Medicare+Choice: Plan Withdrawals Indicate Difficulty of
Providing Choice While Achieving Savings
DATE: 09/07/2000
SUBJECT: Health care programs
Managed health care
Health insurance cost control
Health resources utilization
Medical services rates
Health insurance
IDENTIFIER: Medicare Choice Program
Federal Employees Health Benefits Program
******************************************************************
** This file contains an ASCII representation of the text of a **
** GAO Testimony. **
** **
** 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. **
** **
******************************************************************
GAO/HEHS-00-183
Appendix I: Scope, Methodology, and Data Sources
44
Appendix II: Plans Withdrawing From Medicare+Choice,
January 2000
47
Appendix III: Plans Withdrawing From Medicare+Choice,
January 2001
51
Appendix IV: Comments From the Health Care Financing
Administration
56
Appendix V: Comments From the American Association of
Health Plans
57
Appendix VI: Comments From the BlueCross BlueShield
Association
61
Appendix VII: Comments From the Health Insurance Association
of America
64
Table 1: Medicare Managed Care Enrollment and Geographic
Impact of the 2000 Withdrawals 13
Table 2: Largest Withdrawing Medicare+Choice Plans, January 2000 14
Table 3: Largest Withdrawing Medicare+Choice Plans, January 2001 15
Table 4: Key Characteristics of Plans That Terminated Their
Medicare Contracts in 2000 and 2001 21
Table 5: Average Annual Amount Spent per Enrollee on
Non-Medicare Benefits, 1999 and 2000 34
Table 6: Counties With Medicare+Choice Plans in July 1999
Affected by 2000 Withdrawals, by Year 2000 Payment Rates 36
Table 7: Counties With Medicare+Choice Plans in March 2000
Affected by 2001 Withdrawals, by Year 2001 Payment Rates 36
Table 8: Counties With Medicare Plans in July 1999 Affected by Withdrawals,
by Percentage Increase in 2000 Payment Rates 38
Figure 1: Proportion of Medicare+Choice and Affected Enrollees
in Rural, Fringe/Small Urban, and Large Urban Areas 12
Figure 2: Areas Where Sterling Option I Is Available, 2000 17
Figure 3: Plan Participation in Medicare and FEHBP, 1994-2000 19
Figure 4: Humana's Participation in Texas, 1986-2001 24
Figure 5: Plan Participation in Maryland, 1986-2001 28
Figure 6: Cumulative Increase in Average Medicare+Choice
County Rates and per Capita FFS Spending, 1997-2001 31
AAHP American Association of Health Plans
BBA Balanced Budget Act of 1997
BBRA Balanced Budget Refinement Act of 1999
BCBSA BlueCross BlueShield Association
CBO Congressional Budget Office
ESRD end-stage renal disease
FEHBP Federal Employees Health Benefits Program
FFS fee-for-service
GME graduate medical education
HCFA Health Care Financing Administration
HHS Department of Health and Human Services
HIAA Health Insurance Association of America
HMO health maintenance organization
OIG Office of Inspector General
Health, Education, and
Human Services Division
B-293906
September 7, 2000
Congressional Requesters
Through most of the 1990s, enrollment in Medicare managed care plans grew
rapidly.1 The number of beneficiaries enrolled in these plans increased from
almost 2 million in 1993 to over 6 million in 1998 (16 percent of all
Medicare beneficiaries), and the number of participating plans more than
tripled. Many beneficiaries were attracted to managed care plans because
they typically offered services not covered under Medicare's traditional
fee-for-service (FFS) program--such as routine physical examinations and
outpatient prescription drugs--and because members generally paid less out
of pocket than they would in FFS. However, many areas of the country were
not served by these plans. About 30 percent of the nation's 39 million
Medicare beneficiaries--particularly those living in rural areas--had no
alternative to the FFS program. A second problem was that Medicare was not
realizing the expected savings from managed care. A number of studies by
GAO, other government agencies, and researchers had concluded that plan
payments were not adequately adjusted to reflect the fact that plans tended
to attract beneficiaries with lower-than-average health costs.
In the Balanced Budget Act of 1997 (BBA), the Congress sought to address
some of these concerns by creating the Medicare+Choice program. To expand
Medicare beneficiaries' health plan options, the BBA included payment and
other changes to encourage the wider availability of health maintenance
organizations and permitted other types of health plans, such as preferred
provider organizations, to participate. Medicare+Choice was also expected to
improve Medicare's financial posture by better controlling spending.
Accordingly, the BBA contained provisions to temporarily slow the growth of
plan payment rates relative to FFS spending and required that future
payments better reflect the expected health care utilization of Medicare
beneficiaries enrolled in plans.
Following the implementation of Medicare+Choice, nearly 100 plans either
terminated their contracts and fully withdrew from the program or partially
withdrew by reducing the geographic areas they served for the 1999 contract
year. We previously reported that these plans tended to be recent market
entrants, had low enrollment, or faced competition from larger plans, and
that the withdrawals may have largely resulted from competitive market
forces.2 Since then, more plans have either withdrawn for the 2000 contract
year or announced that they will withdraw for the 2001 contract year.
Because of your continuing interest in the Medicare+Choice program, you
asked us to (1) determine the geographic distribution and the distribution
among plans of enrollees affected by the recent plan withdrawals, (2)
identify the factors associated with plans that terminated or reduced their
participation in the program, and (3) examine the likely role of payment
rates in affecting plans' decisions. (Requesters are listed at the end of
this letter.) To answer these questions, we analyzed enrollment and plan
participation data from the Health Care Financing Administration (HCFA) and
synthesized findings from our previous reports. Appendix I presents the
details of our methodology. Our work was done from August 1999 to September
2000 in accordance with generally accepted government auditing standards.
Of 309 plans serving Medicare beneficiaries at the end of 1999, 99 plans
terminated their contracts or reduced the number of counties they served for
the 2000 contract year, and 118 have announced they will terminate their
contracts or reduce service areas for the 2001 contract year. These
withdrawals affected about 328,000 enrollees in 2000 and will affect almost
1 million enrollees in 2001. The number of enrollees affected accounts for
about 5 percent of Medicare+Choice enrollees in 2000 and about 15 percent in
2001. A disproportionate number of affected enrollees live outside of major
urban areas. A portion of these enrollees, approximately 79,000 in 2000 and
159,000 in 2001, will have no other Medicare managed care option available
in their area and must either switch to a non-managed care option, if one is
available in their area, or return to traditional FFS Medicare. These
withdrawals can mean higher out-of-pocket costs and be disruptive for those
beneficiaries who lose access to relatively inexpensive prescription drug
benefits or must switch health care providers. While a new private FFS plan
has begun to offer services in many of the affected areas as an alternative
to the traditional public FFS HCFA manages, it does not offer a prescription
drug benefit.
Some factors are common to plan withdrawals occurring in 2000 and 2001, but
there are some differences. In January 2000, Medicare+Choice plans tended to
withdraw from more difficult to serve rural counties or large urban areas
that they had entered more recently or where they failed to attract
sufficient enrollment. In 2001, the trend is essentially the same for the
service area reductions but somewhat different for the contract
terminations, which involve some older, more established plans. The pattern
of Medicare+Choice withdrawals shares common elements with plan
participation in the similarly choice-based health insurance program for
federal employees. That program also experienced a period of rapid expansion
between 1994 and 1997 followed by the withdrawal of newer, relatively small
plans. In both 2000 and 2001, other factors, such as problems contracting
with health care providers, may also have contributed to the Medicare plan
withdrawals.
Industry representatives contend that the BBA's payment rate changes were
too severe and that low Medicare payment rates are largely responsible for
the plan withdrawals. However, since the BBA was enacted, Medicare+Choice
payment rates have risen faster than per capita FFS spending. In addition,
many plans have attracted beneficiaries who have lower-than-average expected
health care costs, while Medicare+Choice payments are largely based on the
expected cost of beneficiaries with average health care needs. The result is
that Medicare can pay more for a beneficiary who enrolls in a plan than if
the beneficiary had remained in FFS. As we recently reported, these
additional payments amounted to $5.2 billion, or 21 percent, more in 1998
than the FFS program would have spent to provide Medicare-covered benefits
to plans' enrollees.3 Similarly, many of the withdrawing plans reported that
Medicare's payment rates substantially exceeded their own estimated costs
(including allowed profits) of providing Medicare-covered benefits. On
average, plans that terminated their contracts in 2000 or 2001 reported
spending 22 percent of their Medicare payments, or about $1,200 per
beneficiary, on benefits that are not available in the FFS program--such as
routine physical examinations or prescription drugs.
Although industry representatives have called for Medicare+Choice payment
rate increases, it is unclear whether increases would affect plans'
participation decisions. In 2000, 7 percent of the counties with a
Medicare+Choice plan in 1999 received a payment rate increase of 10 percent
or more. Nonetheless, nearly 40 percent of these counties experienced a plan
withdrawal. This suggests that the magnitude of rate increases needed to
make participating in Medicare a sufficiently attractive business option for
some plans may not be reasonable in light of countervailing pressures to
make the Medicare program financially sustainable for the long term.
As of July 2000, about 6.2 million people--or approximately 16 percent of
Medicare's 39 million beneficiaries--were enrolled in Medicare+Choice plans.
These plans receive a fixed monthly payment for each beneficiary, regardless
of what an individual enrollee's care actually costs. Higher costs reduce a
plan's profits or result in losses, while lower costs can enable it to offer
additional benefits that help it to retain existing enrollees and attract
new enrollees. Because managed care plans have a financial incentive to
provide care efficiently, policymakers have long looked to them to curb
unnecessary spending and produce savings for Medicare. Among BBA's major
reforms to contain Medicare spending was the creation of Medicare+Choice,
which was also designed to increase the plan options available to Medicare
beneficiaries.
Before the BBA, numerous studies by us, the Physician Payment Review
Commission--which has been incorporated into the Medicare Payment Advisory
Commission--HCFA, and others demonstrated that the Medicare program spent
hundreds of millions more on beneficiaries enrolled in health plans than it
would have spent if the same individuals had remained in traditional FFS
Medicare.4 This occurred because Medicare payments were based on the
estimated cost of FFS beneficiaries with average health and were not
adequately adjusted to reflect the fact that plans tended to enroll
beneficiaries in better-than-average health who had lower health care
costs--a phenomenon known as favorable selection.
Before 1998, base payment rates to plans in each county were set at 95
percent of the estimated FFS cost of the average beneficiary. The wide
variation in local FFS expenditures, caused by local differences in both the
prices of medical services and in beneficiaries' use of services, led to
corresponding variation in these base rates. This variation may have
accounted for some of the unevenness in plan availability across the
country. Other factors, such as the higher concentration of Medicare
beneficiaries, may have prompted plans to serve primarily urban areas.
Beneficiaries in most rural areas lacked access to plans.
Beginning in 1998, the BBA substantially changed the method used to set plan
payment rates. The new method involves paying the highest of three
alternative rates: a minimum amount, or "floor"; a minimum increase over the
previous year's payment rate; or a blend of historical FFS spending in a
county and national average costs adjusted for local price levels. Some of
the new payment provisions were designed to reduce excess payments, while
others were designed for different purposes--such as increasing plan
participation in geographic areas that had low payment rates.
The BBA aims to reduce the excess in Medicare's health plan payments
primarily by holding down per capita payment increases for 5 years and by
mandating a new health-based risk adjustment system. In January 2000, HCFA
implemented a method for adjusting plan payments based on beneficiary health
status, as required by the BBA. The new method, to be phased in over time,
will pay plans more for serving Medicare beneficiaries with serious health
problems and less for serving relatively healthy ones.
The BBA also contains provisions to gradually remove graduate medical
education (GME) payments from plan payments and provide for teaching
hospitals to receive these payments directly from Medicare.5 Because GME
spending is concentrated in high-payment-rate counties, its removal is
expected to slow payment rate growth more in those areas.
Another BBA objective is to reduce the geographic disparity in payment
rates. A methodological approach known as "blending" will, over time, move
all rates closer to the national average by providing for larger payment
increases in low-rate counties and smaller payment increases in high-rate
counties. In addition, the BBA established a minimum payment rate, known as
a "floor," to encourage plans to offer services in areas that historically
had low payment rates and few participating plans--primarily rural counties.
The BBA also eliminated the requirement that no more than 50 percent of a
plan's enrollment may consist of Medicare and Medicaid beneficiaries. This
means that Medicare plans can now serve areas without first building a
commercial base.
In 1999, 45 of the 346 plans that participated in 1998 terminated their
Medicare contracts and 54 others reduced the number of counties they served.
These withdrawal decisions affected about 407,000 enrollees (7 percent of
the managed care population) who had to choose a new plan (if one was
available in their county) or switch to FFS. About 61,000 of these
enrollees, or 1 percent of the total Medicare managed care population, lived
in counties in which no other plan was offered. Even if another plan was
available, the approximately 450 beneficiaries affected by the withdrawals
who had end-stage renal disease (ESRD) had to return to FFS.6 Medicare
prohibits beneficiaries with ESRD from joining a health plan, although they
may stay in one if they develop the disease while enrolled.
Plan withdrawals can be disruptive and costly for affected beneficiaries.
Although many affected beneficiaries can enroll in another plan, this option
may require them to switch health care providers and accept different
benefit coverage. Those who return to FFS may be able to retain their
providers, but typically face out-of-pocket costs that are higher than they
incurred as managed care enrollees. For example, most plan enrollees receive
some coverage for outpatient prescription drugs, a benefit not offered in
the FFS program. Although the BBA guarantees beneficiaries affected by plan
withdrawals the right to purchase certain supplemental insurance policies
(known as Medigap), none of the guaranteed policies cover prescription
drugs.
Officials from organizations representing plans reported that the BBA
changes to the payment rates and the increased administrative burden of the
new regulations were largely responsible for the plan withdrawals. According
to the officials, Medicare payment rate increases did not keep pace with
plans' costs or medical inflation.
Our analysis indicated that a combination of market factors may have
influenced plans' participation decisions. Plans more frequently withdrew
from counties they had entered more recently, where they had attracted fewer
enrollees, or where they faced larger competitors. Some plans indicated that
they withdrew from areas where they were unsuccessful in establishing
sufficient provider networks. The effect of Medicare's payment rates on
withdrawals was much less obvious. For example, about 90 percent of
high-payment-rate counties experienced a plan withdrawal compared with only
34 percent of low-payment counties. Taken as a whole, these findings
suggested that a portion of the withdrawals may have been the result of
plans that were less able to compete effectively in certain areas.
In November 1999, the Congress passed the Balanced Budget Refinement Act of
1999 (BBRA). The BBRA contains provisions designed to encourage plan
participation in Medicare+Choice. Among other changes, the BBRA provides a
new entry bonus to plans that begin serving currently unserved areas. It
also increases plans' flexibility to vary benefits within a geographic area
and reduces some administrative requirements. In addition, the act slows the
phase-in of the new risk adjustment methodology, reducing the short-term
effect of the new methodology on plan payments. The BBRA also reduces the
length of time that a plan has to wait to reenter the program after
terminating its Medicare contract. The effect of these provisions on future
plan participation is uncertain.
Smaller Urban Areas
In 2000, 41 of 309 participating plans terminated their Medicare+Choice
contracts and another 58 plans reduced the number of counties they serve.
This pattern will continue in 2001, when 65 of 261 plans currently
participating in Medicare+Choice have announced they will terminate and
another 53 plans will change their service areas. Combined, these plan
withdrawals directly affect about 1.3 million Medicare+Choice enrollees. The
2001 withdrawals affect a much larger percentage of enrollees, approximately
15 percent, compared with the 2000 withdrawals that affected about 5 percent
of all enrollees. All affected enrollees have to choose a new plan (if a
plan accepting new enrollees is available in their county) or switch to FFS.
By 2001, almost 75 percent of the counties that had a Medicare+Choice plan
in 1999 will have been affected.7 About 238,000, or approximately 19
percent, of the affected enrollees live in counties in which no other
managed care plan is being offered. Some of these beneficiaries may have the
option of enrolling in a new private FFS plan, but the remainder will have
no alternative to the traditional FFS program. The 1,940 beneficiaries in
withdrawing plans who have ESRD must return to FFS.
Plan withdrawals in both years disproportionately affect beneficiaries
living in small urban, fringe, and rural counties.8 In 2000, approximately
65 percent of the 328,000 beneficiaries affected by the withdrawals lived in
one of these types of counties even though these areas accounted for less
than 33 percent of Medicare's managed care enrollees. (See fig. 1.) In
contrast, the effects of the 2001 withdrawals will be more widespread and
more representative of the distribution of Medicare+Choice enrollees. In
both years, beneficiaries living in less densely populated areas were also
likelier to be left only with the FFS alternative compared to affected
beneficiaries in major urban areas. (See table 1.)
A small number of plans accounted for a substantial portion of the affected
enrollees in both years--the 10 largest withdrawing plans accounting for 45
percent in 2000 and 37 percent in 2001. (See tables 2 and 3.) Whereas the
largest plans that withdrew in 2000 were concentrated in small urban,
fringe, and rural counties, the largest withdrawing plans in 2001 are more
uniformly distributed among these and major urban areas. Also, the
withdrawing plans in 2001 tend to have significantly larger enrollments than
the withdrawing plans in 2000.
Note: Approximately 8,400 of the affected enrollees for 2001 live outside of
the withdrawing plans' service areas. These enrollees are excluded from this
analysis.
Source: Medicare Compare Database, 1999 and 2000; Medicare Managed Care
Market Penetration State/County/Plan Data Files, July 1999 and Mar. 2000,
www.hcfa.gov/medicare/; Bureau of Health Professions, Area Resource File,
Feb. 1999; and files of contract terminations and service area reductions
from the Center for Health Plans and Providers at HCFA.
Small
Rural urban/ Major urban Nationwide
counties fringe areas
counties
1999
Medicare+Choice
enrollment 219,000 1,774,000 4,198,000 6,191,000
Percentage of total 3.5 28.7 67.8 100
2000
Enrollees affected by
withdrawals 42,000 169,000 117,000 328,000
Percentage of total 12.7 51.7 35.6 100
Affected enrollees with
no other managed care 30,000 35,000 14,000 79,000
option
Percentage of total 37.5 44.7 17.8 100
2001
Enrollees affected by
withdrawals 69,000 363,000 493,000 925,000
Percentage of total 7.5 39.2 53.3 100
Affected enrollees with
no other managed care 44,000 93,000 22,000 159,000
option
Percentage of total 27.5 58.6 13.9 100
Note: Approximately 8,400 of the affected enrollees for 2001 live outside of
the withdrawing plans' service areas. These enrollees are excluded from this
analysis.
Source: Medicare Compare Database, 1999 and 2000; Medicare Managed Care
Market Penetration State/County/Plan Data Files, July 1999 and Mar. 2000,
www.hcfa.gov/medicare/; Bureau of Health Professions, Area Resource File,
Feb. 1999; and files of contract terminations and service area reductions
from the Center for Health Plans and Providers at HCFA.
Number of Percentage of affected
Plan name enrollees enrollees in rural, fringe,
affected or small urban counties
Humana HP of Texasa 23,500 39
Ochsner Health Plana
(La.) 17,000 100
United Health Care of
Louisiana 17,000 64
Florida Health Choice,
Inc. 14,700 88
Free State Health Plana
(Md.) 14,700 100
Blue Cross/Shield of
Arizona 14,500 28
Optima Health Plan (Va.) 13,800 2
Healthsource New
Hampshire 13,400 100
Capital Area Community HP
(N.Y.) 9,700 100
Humana Health Plan, Inc.
(Nev.) 9,500 100
Affected enrollees in 10
plans 147,800 69
Affected enrollees in 89
other withdrawing plans 180,200 61
aPlan remained in Medicare but reduced the number of counties served.
Sources: Medicare Compare Database, 1999; Medicare Managed Care Market
Penetration State/County/Plan Data Files, July 1999, www.hcfa.gov/medicare/;
and files of contract terminations and service area reductions from the
Center for Health Plans and Providers at HCFA.
Percent of affected enrollees
Plan name Total number of in rural, fringe, or small
enrollees affected
urban counties
NYLCare Health Plans
(Tex.) 71,600 42
NYLCare Health Plans
(Tex.) 56,200 21
Aetna U.S. Healthcare
(N.Y.)a 34,900 22
Free State Health Plan
(Md.) 31,400 0
HMO of Northeastern PA 30,700 100
Aetna U.S. Healthcare
(Ohio) 27,800 30
Humana Medical Plan
(Fla.)a 25,600 64
Anthem Health Plans
(Conn.) 23,500 50
Humana HP of Texasa 20,500 12
Aetna U.S. Healthcare
(Pa.)a 18,100 100
Affected enrollees in
10 plans 340,300 40
Affected enrollees in
108 other
585,000 50
withdrawing plans
Note: Approximately 8,400 of the affected enrollees for 2001 live outside of
the withdrawing plans' service areas. These enrollees are excluded from this
analysis.
aPlan will remain in Medicare but reduce the number of counties served.
Sources: Medicare Compare Database, 2000; Medicare Managed Care Market
Penetration State/County/Plan Data Files, Mar. 2000, www.hcfa.gov/medicare/;
and files of contract terminations and service area reductions from the
Center for Health Plans and Providers at HCFA.
Although some plans continue to submit applications to enter the
Medicare+Choice program or expand their service areas, the volume of
applications has decreased from 30 in 1999 to 10 in 2000. HCFA has already
approved many of the applications submitted since July 1998, including one
for a private FFS plan called Sterling Option I that initially will serve
1,221 counties in 17 states.9 The new plan's service area encompasses 940
counties, including many rural counties, previously not served by a
Medicare+Choice plan. Since the initial offering, Sterling has added 8 more
states to its service area.10 (See fig. 2.) Beneficiaries who enroll in
Sterling will pay a $55 monthly premium (in addition to the Medicare part B
premium) in exchange for reduced out-of-pocket costs for many services and
extended coverage for hospitalizations, among other benefits. However,
Sterling Option I does not offer prescription drug coverage.
Source: HCFA press release, May 2000.
in 2000; This Pattern Less Evident in 2001
Plan participation in Medicare managed care increased rapidly after 1993,
peaked in 1998, and began declining in 1999. This experience is not unique
to Medicare and, in fact, closely tracks plan participation in the Federal
Employees Health Benefits Program (FEHBP), another large program offering
multiple health plan choices. The withdrawals in 2000 followed a pattern
that is similar to the pattern of withdrawals in FEHBP, as well as the
pattern we found in our prior analysis of the 1999 Medicare plan
withdrawals.11 Nearly all of the plans that terminated their Medicare
contracts for 2000 or reduced their service areas were relatively new
entrants in their respective markets, had attracted few beneficiaries, or
had only a small share of the local Medicare managed care market. The plan
withdrawals for 2001 deviate somewhat from this pattern in that some older,
more established plans are terminating. However, the service area reductions
in 2000 and 2001 are consistent with the 1999 pattern of withdrawals. In
both years, other factors--such as plans' inability to establish sufficient
provider networks--are often evident.
Withdrawals From Medicare and FEHBP
Between 1993 and 1998, the Medicare managed care program grew rapidly and
the number of plans more than tripled--from about 110 plans to 350 plans.12
Since 1998, however, 151 plans have terminated their Medicare contracts or
announced that they will, and few new plans have joined the program. Despite
the drop in plan participation, enrollment has continued to
increase--although at a slower pace--with the result that the total number
of Medicare managed care enrollees has remained approximately the same or
even increased slightly over the past 2 years. However, the substantial
decline in plan participation next year may cause total enrollment to fall.
FEHBP experienced a similar rapid rise in the number of participating plans
followed by a decline.13 (See fig. 3.) Between 1994 and 1997, the number of
plans participating in FEHBP increased from 369 to 470. Since then, the
number of FEHBP plans has declined steadily and may fall to approximately
240 next year. This roughly 50 percent decline in the number of FEHBP plans
is similar to the approximately 57 percent decline experienced in Medicare
over the same period. However, the percentage of FEHBP enrollees affected is
substantially smaller than the percentage of Medicare+Choice enrollees
affected. In 2001, for example, FEHBP plan withdrawals are expected to
affect about 1 percent of enrollees, compared to Medicare+Choice withdrawals
affecting 15 percent of enrollees.
Source: GAO analysis of HCFA and Office of Personnel Management data.
At the same time new plans were joining the Medicare program, many existing
plans expanded their geographic service areas. Some plans entered previously
unserved rural counties while others entered urban counties with one or more
existing Medicare plans. As a result, the percentage of rural beneficiaries
with access to Medicare managed care increased from about 10 percent in 1993
to over 31 percent in 1998. Because of recent plan withdrawals, however, the
percentage of beneficiaries in rural areas with access to a Medicare managed
care plan has fallen to about 21 percent in 2000.
Urban beneficiaries, nearly all of whom already had access to at least one
plan in 1993, had a wider choice of plan options. In recent years, however,
even large urban areas have seen a decline in plan participation. The
percentage of beneficiaries living in large urban areas with access to at
least one plan has declined from 99 percent in 1999 to 97 percent in 2000
and is expected to fall again in 2001.
Low Enrollments; Pattern Less Evident in 2001
The vast majority of Medicare+Choice plans that terminated their Medicare
contracts in 2000, as opposed to reducing the number of counties they
served, were recent entrants into urban areas that already had substantial
plan participation. Many terminating plans had few beneficiaries or a
relatively small share of the local Medicare managed care enrollment. These
factors are the same ones that were associated with the 1999 withdrawals. In
2000, 38 of the 41 terminating plans were either recent entrants, had
attracted fewer than 200 enrollees, or had less than a 15 percent share of
the local Medicare plan market in each of the counties they served.14 (See
table 4.) Plans that terminated their participation in FEHBP had similar
characteristics: 42 percent of the terminating plans had fewer than 300
enrollees and many of those were recent entrants.
The pattern of plan withdrawals is different in 2001 in that some older,
larger, and more established plans are also terminating their Medicare
contracts. For example, almost 43 percent of terminating plans entered the
market before 1996 and 29 percent had total plan enrollments that exceeded
10,000 enrollees.
Number of plans
Key
characteristics Participating Terminated Participating Terminating
in July 1999 in 2000 in March 2000 in 2001
Year of entry
Before 1996 160 8 (20%) 136 28 (43%)
In 1996/97 84 15 (37%) 57 24 (37%)
Since 1998 55 18 (44%) 39 13 (20%)
Enrollment
Fewer than 1,000
enrollees 27 11 (27%) 13 3 (5%)
Between 1,000 and
10,000 enrollees 136 25 (61%) 102 43 (66%)
More than 10,000
enrollees 136 5 (12%) 117 19 (29%)
Market share
Less than 15% 82 22 (54%) 52 21 (32%)
More than 15% 217 19 (46%) 180 44 (68%)
Total 299a 41 232b 65
aMedicare+Choice plans as of July 1999.
bMedicare+Choice plans as of Mar. 2000.
Sources: Medicare Compare Database, 1999 and 2000; Medicare Managed Care
Market Penetration State/County/Plan Data Files, July 1999 and Mar. 2000,
www.hcfa.gov/medicare/; and files of contract terminations and service area
reductions for 2000 and 2001 from the Center for Health Plans and Providers
at HCFA.
Relatively Low-Enrollment Counties in 2000 and 2001
Although the patterns of contract terminations in 2000 and 2001 appear to be
somewhat different, the patterns of service area reductions in the 2 years
are similar. In both years, plans that withdrew from only a portion of the
counties they served tended to pull out of counties that they had more
recently entered or where they had relatively low enrollment. In the
majority of cases--92 percent in 2000 and 79 percent in 2001--plans withdrew
from counties where they had recently entered, where they enrolled fewer
than 200 beneficiaries, or where they enrolled fewer than 15 percent of the
Medicare managed care enrollees. This pattern was more pronounced in 2000
than in 2001, but the 2001 service area reductions still follow the same
general trend. In some cases, plans consolidated into one or more core areas
where they were most strongly established.
Service area reductions have been more concentrated in rural areas. Despite
the floor payment rates, enacted in BBA, which make payments to plans
considerably higher than FFS costs in many rural counties, the challenge of
providing managed care in rural areas may be a significant contributing
factor. The sparseness of both beneficiaries and providers may present
difficulties for plans. Without sufficient beneficiary populations, plans
say they cannot enroll enough individuals to spread risk and cover fixed
operating costs.15 In addition, plans may have difficulty obtaining
discounts and negotiating contracts with physicians and hospitals when an
area has few competing providers.
Consolidation Behavior
Humana Health Plan of Texas, the plan with the single largest number of
affected enrollees in 2000, illustrates the consolidation behavior exhibited
by a number of the plans that reduced their service areas. Humana started
serving Medicare beneficiaries in areas around Corpus Christi, Texas, in
1986 and added San Antonio in 1988. It more recently expanded its service
area by adding a total of 23 counties in 1995, 1997, and 1999. (See fig. 4.)
In 2000, the plan withdrew from 16 of the counties, both urban and rural, it
entered in 1997 and 1999, as well as a few it entered in 1995. The plan
remained in the central counties encompassing San Antonio and Houston, both
urban areas where the plan had by far its largest concentration of
enrollment, and the Corpus Christi area. Humana remained in San Antonio
despite the fact that the county's monthly payment rate for 2000 was, on
average, $26 lower than payments in the four urban counties it dropped. The
10 counties it retained in 2000 accounted for 70 percent of its Medicare
managed care enrollees in Texas.
In 2001, Humana will consolidate even further, serving only the San Antonio
and Corpus Christi areas. This time, the 2001 monthly payment rate in San
Antonio is, on average, $147 lower than the six counties the plan is
dropping. Humana recently stated that it incurred pre-tax losses exceeding
$26 million during 1999 in the counties it will leave in 2001. However, only
a fraction of these losses may be due to providing Medicare covered
benefits. The plan is currently offering, at no additional charge, an
unlimited generic prescription drug benefit and a brand name benefit up to
$1,400 per year, in addition to some coverage for physical exams and vision
services, to the beneficiaries in these Texas counties.
Sources: Medicare Compare Database, 1999; Medicare Managed Care Market
Penetration State/County/Plan Data Files, July 1999, www.hcfa.gov/medicare/;
Historical Service Area File, Feb. 1999; and files of contract terminations
and service area reductions from the Center for Health Plans and Providers
at HCFA.
Example of Plan Consolidations
The Medicare managed care experience in Maryland illustrates both the
service area reductions that occurred in 2000 and the trend toward larger,
more established plans terminating their contracts in 2001. In 2000, plans
withdrew from recently entered rural counties while continuing to serve more
heavily populated urban areas. In 2001, these plans are continuing the
exodus from Medicare by withdrawing from these urban areas and terminating
their contracts.
Between 1986 and 1993, only one plan, Freestate Health Plan--sponsored by
Blue Cross?operated in Maryland. Its service area included only 6 of
Maryland's 24 counties, all located in Maryland's major metropolitan
area--the areas surrounding Baltimore and Washington, D.C. Over time, new
plans began operation in the state, mostly in the same Baltimore-Washington
corridor. (See fig. 5.) One plan, Optimum Choice, began offering service
statewide in 1994, followed by 2 more statewide plans in 1996. Between 1997
and 1999, however, these 3 plans reduced their service areas until, by 1999,
only Freestate continued to serve Maryland's rural counties. In 2000,
Freestate reduced its service area to the Baltimore-Washington area--its
historical core service area. Rural Maryland beneficiaries, who had a
managed care option between 1994 and 1999, were left with no alternative to
traditional FFS Medicare.16
The difficulty of serving sparsely populated rural areas may have been an
important factor in the Maryland plans' withdrawal decisions for 2000.
Freestate Health Plan, for example, withdrew from rural Caroline County
where it faced no competition and enrolled nearly one in five of the
county's beneficiaries despite charging a $75 per month enrollee premium.
However, the plan's relatively large market share in the county amounted to
only 895 enrollees. In contrast, the plan's 2 percent market share in urban
Montgomery County, an area it continued to serve, resulted in more than
2,000 enrollees. In addition, Medicare payment rates were increasing faster
in the rural counties the plan left because of the floor and the blend
provisions in the BBA.
Freestate has announced it will terminate its contract in Maryland for 2001,
leaving Kaiser Health Plan as the only remaining Medicare plan serving the
state. Freestate has said that it expects to incur losses of $7.5 million by
the end of 2000.
Sources: Medicare Compare Database, 1999; Medicare Managed Care Market
Penetration State/County/Plan Data Files, July 1999, www.hcfa.gov/medicare/;
Historical Service Area File, Feb. 1999; and files of contract terminations
and service area reductions from the Center for Health Plans and Providers
at HCFA.
Contributed to Plan Withdrawals
Although recent entry, low enrollment, or low market share are
characteristics of most withdrawing plans, in some cases plan withdrawals
appear to have little to do with these factors. In one case, a merger caused
a plan to change operations to avoid anti-trust violations and subsequently
resulted in termination of selected contracts. In other cases, plans
terminated all operations--Medicare, Medicaid, and commercial--in an area.
Finally, some plans have reported that providers in some areas are becoming
increasingly resistant to contracting with them, making it more difficult
for plans to assemble viable provider networks in certain areas. The
following examples illustrate other factors that may have contributed to
plan withdrawals.
� Aetna U.S. Healthcare acquired NYLCare Health Plans in July 1998, and
later purchased Prudential Health Care in August 1999. Because Texas
officials were concerned that Aetna would have too large a share of the
state's market after it acquired Prudential, they agreed to the purchase
under the condition that Aetna sell its NYLCare business in the state.
However, under special agreement, Aetna was allowed to continue managing the
Medicare component. Aetna subsequently terminated this contract.
� Capital Area Community Health Plan of Albany, NY, was affiliated with
Kaiser Permanente, which withdrew from all of its operations in the entire
northeast region in 2000.
� Humana terminated all of its business--commercial and Medicare--in Nevada.
� United Health Care of Louisiana was one of the first national plans to buy
out local plans in Louisiana. Local providers, who preferred dealing with
the local plans, resisted contracting with United. The plan eventually
withdrew from these areas.
� Oxford Health Plans of NY had trouble assembling a viable provider network
in one of the large counties it served, so it withdrew from that county.
Industry representatives have stated that low Medicare payments, resulting
from BBA provisions designed to control program spending, are primarily to
blame for the recent plan withdrawals. The American Association of Health
Plans contends that the BBA created a "fairness gap" by decreasing payments
to health plans relative to spending on beneficiaries in the FFS program.
However, since the BBA was enacted the increase in Medicare+Choice payment
rates has exceeded the growth in per capita FFS spending. Furthermore, our
recent study of plan payments found that Medicare paid plans $5.2 billion
(or about 21 percent) more than it would have spent in 1998 if plan
enrollees had received standard Medicare covered services through the
traditional FFS program. According to reports that plans submit to HCFA,
Medicare's payments are also substantially higher than the average plan's
projected costs of providing Medicare-covered benefits. Moreover, although
industry representatives have called for higher payment rates, the extent to
which rate increases would affect plans' decisions to participate in
Medicare is unclear. In 2000 and 2001, withdrawals have not been confined to
counties where payment rate increases, or payment rates, were low.
Increase
Between 1997--the year the BBA was enacted--and 1999, Medicare+Choice
payment rates increased on average by about 4.2 percent.17 (See fig. 6.)
Furthermore, the payment rate increase was applied to 1997 rates that HCFA
now estimates were inflated by about 3 percent because of an error in the
spending forecast used to set the rates.18 In contrast, per capita FFS
spending fell 1.7 percent during the same period.
Note: Medicare+Choice payment rate increases were calculated as the change
in county payment rates for aged beneficiaries, weighted by the number of
aged beneficiaries in each county in 1997. Medicare+Choice payment rates for
2001 were announced in Mar. 2000. Per capita FFS spending for 2000 and 2001
is based on HCFA projections.
Source: 2001 Medicare+Choice County Payment Rate Calculation Worksheet,
www.hcfa.gov/medicare/.
HCFA estimates that between 1999 and 2001, per capita FFS spending will grow
faster than Medicare+Choice payment rates. If these estimates prove
accurate, the cumulative increase in Medicare+Choice payment rates between
1997 and 2000 will still exceed the growth in per capita FFS spending, but
the gap will be much narrower. By 2001, HCFA's current projections indicate
that average spending in the traditional program will have increased 11.9
percent, while plan payment rates will have increased 10.7 percent.
Traditional FFS Program
Recently, we reported that Medicare+Choice plan payments likely exceed the
amount that beneficiaries enrolled in plans would cost in the traditional
FFS program.19 In 1998, aggregate payments exceeded enrollees' estimated FFS
costs by about 21 percent--or approximately $5.2 billion. On a per enrollee
basis, Medicare paid plans about $1,000 more than the FFS program would have
spent to provide Medicare-covered benefits.
A portion of the estimated $5.2 billion in annual excess plan payments may
diminish over time. Approximately $2 billion of these excess payments
resulted from FFS spending forecast errors built into the 1997 county
payment rates due to the BBA provisions that based future county rates on
the 1997 rates and guaranteed 2 percent minimum annual rate increases. The
effect of the 1997 forecast error will largely be mitigated by the BBA
provision that slows Medicare+Choice rate increases relative to the growth
in FFS spending between 1998 and 2002.
The bulk of the excess payments we estimated for 1998 ($3.2 billion) will
persist each year until payments on behalf of individual enrollees better
match their expected health care costs. Medicare+Choice plans attracted a
disproportionate selection of healthier and less-expensive beneficiaries
relative to traditional FFS Medicare (a phenomenon known as favorable
selection), while payment rates largely continued to reflect the expected
FFS costs of beneficiaries in average health. Consequently, we estimate that
the program spent about 13.2 percent more on plan enrollees than if they had
received services through the traditional FFS program. This year, HCFA
implemented a new methodology to adjust payments for beneficiary health
status. However, our results suggest that this new methodology, which will
be phased in over several years, may ultimately remove less than half of the
excess payments caused by favorable selection.20 HCFA expects to introduce a
more refined methodology in 2004 that may better adjust payments to reflect
enrollees' expected health care costs.
Benefits
Medicare+Choice payment rates not only surpass what the FFS program would
spend to provide Medicare-covered benefits to plans' enrollees, but data
submitted by plans show that rates also generally exceed plans' estimated
costs to provide those same benefits. As part of the annual contracting
process, each Medicare+Choice plan is required to project its per enrollee
cost of providing Medicare-covered benefits.21 If estimated Medicare
payments exceed a plan's projected costs, the plan must use the difference
to provide additional benefits during the contract year or contribute to an
escrow account and use the funds to provide benefits in future years. To
fulfill Medicare's requirement, plans choose to provide additional
benefits--such as routine vision care, dental care, and coverage for
outpatient prescription drugs--that are not covered in the traditional FFS
program.
In their 1999 contract submissions, the average plan--including plans that
withdrew in 2000--projected that its costs would be substantially less than
its Medicare payment. On average, plans estimated that they could provide
Medicare-covered services for about 89 percent of Medicare's payment.22
Plans indicated that they would provide additional benefits to make up the
difference. Most plans' benefit packages exceeded the minimum requirements.
Consequently, the average plan in 1999 estimated it would spend about $1,300
per enrollee, an amount equal to about 22.5 percent of its Medicare payment,
on benefits that are not covered in the FFS program. Among plans that
terminated their contracts or reduced their service areas in 2000, the
average annual amount spent on additional benefits was slightly lower--about
$1,100, or 21.6 percent of Medicare's payment. (See table 5.)
Plans' contract submissions for 2000 exhibited a similar pattern of
additional benefits.23 Plans that will terminate their contracts in 2001
projected that they would spend an average of about $1,200 per enrollee, or
22 percent of their Medicare payment, on additional benefits in 2000. Plans
that will reduce their service areas projected they would spend slightly
less, about $1,000 or 18 percent of their Medicare payment, on additional
benefits. In contrast, spending on additional benefits was estimated at
nearly $1,500 per enrollee, or about 25 percent of 2000 Medicare payments,
for plans that will remain in the program in 2001.
Amount spent per enrollee
1999 2000
Plans remaining in Medicare+Choice for 2001 1,308 1,259
Plans terminating their contracts for 2001 1,164 1,202
Plans reducing their service areas for 2001 1,080 1,010
Note: The number of enrollees in each plan was used to compute the weighted
average.
Source: 1999 and 2000 Adjusted Community Rate Proposals submitted by
Medicare+Choice plans.
Interpret
The effect of payment rates on Medicare+Choice plan participation is
ambiguous. While changes in payment rates are an important influence on
plans' participation decisions, we found that plan withdrawals were not
limited to counties with low payment rates.
On the one hand, plan withdrawals appear to be more extensive in the 2 years
with lower payment rate increases. In both 1999 and 2001, county rates
increased by an average of 2 percent. In 1999, plan withdrawals affected 42
percent of counties that previously had a managed care plan, and in 2001
plan withdrawals will affect 58 percent of such counties. In contrast, a
smaller proportion of counties--approximately 37 percent--were affected in
2000 when rates increased by about 4 percent. The extensiveness of plan
withdrawals may also be related to the gap between average county rate
increases and the change in expected per capita FFS spending. For example,
the projected increase in per capita FFS spending is much higher for 2001
than it was for 1999 and withdrawals will be more extensive.24 Therefore,
withdrawals may moderate after 2002 when payment rate increases will mirror
expected increases in per capita FFS spending except for adjustments to
correct prior spending forecast errors.
The relationship across counties between plan participation and payment
rates, and rate increases, is not clear. Both high-payment rate and
low-payment rate counties are affected by the 2000 and 2001 plan
withdrawals, although the relationship between payment rates and withdrawals
is somewhat different in the two years. In 2000, approximately 39 percent of
the non-floor counties that had at least one plan in 1999--those with
payment rates set above the minimum payment of $402--were affected by a plan
withdrawal. A slightly higher proportion of counties in the middle payment
categories were affected compared to the proportion of affected counties in
the highest rate category and the rate category just above the floor. (See
table 6.) In 2001, about 80 to 90 percent of counties in the higher payment
ranges, but less than two-thirds of the counties in the lower payment ranges
will be affected. (See table 7.) The 2001 withdrawal pattern is similar to
the one that occurred in 1999 in that a disproportionate number of high
payment rate counties were affected by withdrawals. In both 2000 and 2001,
floor counties that previously had a Medicare+Choice plan will be
proportionately less affected by the withdrawals compared to counties that
receive payment rates above the floor. However, the difference between floor
and nonfloor counties is less pronounced in the 2001 withdrawals.
Year 2000 payment ratesa
$402
Total
(floor $402−$505 $505−$608 $608−$711$711−$814
rate)
Counties
with plan(s)
affected by 8 187 109 22 3 329
2000
withdrawals
Counties
with plan(s)64 522 243 48 10 887
in July 1999
Percentage
of all
counties
with plan(s)13% 36% 45% 46% 30% 37%
in July 1999
affected by
withdrawals
Notes: Only Medicare+Choice plans are used for this analysis. Payment rates
are rounded to the nearest dollar. The actual categories are $401.61,
$401.62-$504.78, $504.79-$607.96, $607.97-$711.14, and $711.15-$814.32.
a Payment rate categories were determined by dividing the range into 4 equal
intervals (with a payment range of approximately $103). The floor rate is
the fifth category.
Sources: Medicare Compare Database, 1999; Medicare Managed Care Market
Penetration State/County/Plan Data Files, July 1999, www.hcfa.gov/medicare/;
and files of contract terminations and service area reductions from the
Center for Health Plans and Providers at HCFA.
Year 2001 payment ratesa
$415 Total
(floor $415−$519 $519−$623 $623−$727$727−$831
rate)
Counties
with plan(s)
affected by 39 242 139 36 9 465
2001
withdrawals
Counties
with plan(s)
in March 90 444 217 45 10 806
2000
Percentage
of all
counties
with plan(s)
in March 43% 55% 64% 80% 90% 58%
2000
affected by
withdrawals
Notes: Only Medicare+Choice plans are used for this analysis. Payment rates
are rounded to the nearest dollar. The actual categories are $415.01,
$415.02-$518.91, $518.92-$622.81, $622.82-$726.71, and $726.72-$830.61.
aPayment rate categories were determined by creating four equal rate
categories (with a payment range of approximately $104). The floor rate is
the fifth category.
Sources: Medicare Compare Database, 2000; Medicare Managed Care Market
Penetration State/County/Plan Data Files, Mar. 2000, www.hcfa.gov/medicare/;
and files of contract terminations and service area reductions from the
Center for Health Plans and Providers at HCFA.
The relationship between payment rate increases and plan participation in a
particular county is unclear. In 2000 and 2001 floor counties may have been
less affected by the withdrawals because the BBA substantially increased
payment rates in those counties, and those rates remain considerably above
the average cost of Medicare benefits in the traditional FFS program.
Between 1997 and 2001, payment rates in floor counties increased by 27
percent. In contrast, payment rate increases have been more modest in
nonfloor counties, around 11 percent.25 However, the pattern of plan
withdrawals in 2000 suggests that even relatively large payment rate
increases may not be enough to keep some plans in certain counties. While
county payment rates increased by an average of 4 percent in 2000, the BBA's
rate "blending" provision increased rates by 10 percent or more in certain
counties.26 Nonetheless, 40 percent of these counties with large increases
were affected by plan withdrawals in 2000--about the same as the percentage
of affected counties among those that received the lowest (2 percent) rate
increase.27 (See table 8.) Some areas may have too few beneficiaries or
providers to support multiple plans, or even a single plan. Moreover, plans
that fail to attract a sufficient number of enrollees will not realize their
revenue goals even if payments are adequate on a per capita basis.
Percentage change in payment
rates, 1999-2000
Total
2% 2-5%a 5-10%a More than
10%
Number of counties affected by
2000 withdrawals 72 96 137 24 329
Number of counties with plan(s)
in 1999 179 213 435 60 887
Percentage of counties with
plan(s) in 1999 affected by the 40% 45% 31% 40% 37%
withdrawals
aThe range for each of these categories begins slightly above the lower
number and includes the higher number.
Sources: Medicare Risk Monthly Payment Rates, 2000, www.hcfa.gov/; Medicare
Compare Database, 1999, www.hcfa.gov/; Medicare Managed Care Market
Penetration State/County/Plan Data Files, July 1999, www.hcfa.gov/medicare/;
and files of contract nonrenewals and service area reductions from the
Center for Health Plans and Providers at HCFA.
Medicare+Choice is at a crossroads. Because of contract terminations and
service area reductions, by January 2001 more than 1.6 million beneficiaries
will have had to switch to a different plan or the traditional
fee-for-service program since 1999. Industry representatives contend that
payment rate increases are necessary to keep the program viable.
However, the Medicare+Choice program has already been expensive for
taxpayers. As our work on payment rates shows, the vast majority of plans
have gotten paid more for their Medicare enrollees than the government would
have paid had these enrollees remained in the traditional fee-for-service
program. Raising payment rates to a level sufficient to retain the plans
leaving Medicare would mean increasing the excess that currently exists in
payments for plan enrollees relative to their expected fee-for-service
costs. In areas of the country where there are few beneficiaries and
providers are in short supply, no reasonable payment rate increase is likely
to entice plans to participate in Medicare. Thus, a trade-off exists between
the significant additional costs that would be needed to keep more plans in
the program and the benefits of providing more beneficiaries with options
for accessing Medicare covered services. Such a trade-off raises questions
about the equity of providing a greater array of benefits to a fraction of
the Medicare beneficiary population. In our view, efforts to protect the
viability of Medicare+Choice plans come at the expense of ensuring
Medicare's financial sustainability over the long term.
In commenting on our report, HCFA stated that our findings confirmed its own
analysis of Medicare+Choice plan withdrawals. HCFA noted that the pattern of
withdrawals, analyzed at the corporation level instead of at the individual
plan level, reinforces our finding that factors besides payment rates likely
influenced plans' participation decisions. For example, HCFA said that in
2001, 54 percent of Aetna's Medicare+Choice enrollees and 69 percent of
Cigna's enrollees will be affected by plan withdrawals, but less than 2
percent of Pacificare's enrollees and only 0.1 percent of Kaiser's enrollees
will be affected. The agency contends that these differences provide
evidence that the withdrawals reflect corporations' strategic business
decisions that go beyond Medicare payment adequacy. HCFA also said that it
believes the Administration's proposal to provide a prescription drug
benefit to all enrollees would both reduce inequities in benefit
availability and increase payments to Medicare+Choice plans that cover
prescription drugs. (HCFA's comments appear in app. IV.)
We also provided representatives of the American Association of Health Plans
(AAHP), the BlueCross BlueShield Association (BCBSA), and the Health
Insurance Association of America (HIAA) an opportunity to comment on the
report. All three groups disagreed with our conclusions and stated that our
report did not touch on important issues relevant to plan withdrawals. They
also said that withdrawals can be costly for beneficiaries because
Medicare+Choice plans typically provide preventive care services and other
benefits that are not covered in the traditional FFS program. (AAHP's,
BCBSA's, and HIAA's comments appear in apps. V, VI, and VII.)
AAHP, BCBSA, and HIAA believe that inadequate Medicare+Choice payment rates
are a principal cause of plan withdrawals. BCBSA stated that many plans
could not afford to continue providing sufficient additional benefits
(beyond those covered in FFS) to attract beneficiaries. All three industry
groups stated that it is inappropriate to compare Medicare+Choice payment
rate increases with changes in per capita FFS spending (as we did in fig. 6)
because plans' costs have been growing faster than per capita FFS spending.
HIAA said that FFS spending slowed only as a result of BBA's unprecedented
reductions in Medicare reimbursements and that the Congress began correcting
these reductions with the enactment of BBRA in 1999. BCBSA commented that
the comparison is unfair because the traditional program can control costs
in ways that are unavailable to plans. In our report, we acknowledge that
plans typically provide benefits that are not available in the FFS program.
However, we found that Medicare+Choice payments substantially exceeded
plans' projected costs (including normal profits) of providing
Medicare-covered benefits and that plans contracted with Medicare to use the
difference to provide benefits that are not available in the FFS program.
Furthermore, the contention that plans' costs have grown more rapidly than
per capita FFS spending, or that plans have a limited ability to control
their own cost increases, does not alter our finding that Medicare+Choice
payments exceed the estimated amount that the traditional program would
spend on the individuals enrolled in plans.
AAHP and HIAA stated that our methodology for estimating the FFS costs of
plan enrollees, based on enrollees' prior use of services in the FFS
program, underestimates the health care costs of plan enrollees and
therefore overestimates excess payments to plans. In developing our
methodology, however, we employed assumptions that would tend to
underestimate excess payments.28 Therefore, we believe our findings likely
represent a lower bound on the estimated excess payments plans receive and
the potential savings from improved risk adjustment. HIAA stated that
services are overutilized in the FFS program and that by using FFS spending
as a comparison we overestimated the degree of favorable selection and the
extent of excess payments to plans. In our analysis, we did not attempt to
quantify an appropriate level of care. If services are overutilized in the
FFS program, a comparison of plan payments with a more efficient delivery
system might indicate less favorable selection, but it would not alter our
finding that current Medicare+Choice payment rates--largely based on FFS
spending patterns--exceed the estimated cost of providing Medicare-covered
benefits in the FFS program.
AAHP, BCBSA, and HIAA said that we did not address the issue of regulatory
burden in our report. They believe that recent regulations have increased
plans' administrative costs and discouraged plan participation. Because many
of the recent regulations resulted from provisions in BBA designed to
increase plan accountability, facilitate informed choice and plan
comparisons, protect beneficiary rights, or foster quality improvement
efforts, a comprehensive analysis of this issue would require an assessment
of the regulations' benefits as well as their costs. Such an analysis was
beyond the scope of our report.
Finally, AAHP and BCBSA stressed that plans typically provide benefits not
covered in the traditional FFS program and that plan withdrawals are not
only disruptive for beneficiaries but can also result in beneficiaries
having to pay more in out-of-pocket costs. Although we agree, and did
discuss this issue in the report, it was not the focus of our study.29
We are sending copies of this report to the Honorable Nancy-Ann Min DeParle,
Administrator of the Health Care Financing Administration, and other
interested parties who request them.
If you or your staffs have any questions about this report, please call me
on (202) 512-7114 or Laura A. Dummit, Associate Director, on (202) 512-7119.
Other major contributors included George Duncan, Beverly Ross, and Susanne
Seagrave under the direction of James C. Cosgrove.
William J. Scanlon
Director, Health Financing and
Public Health Issues
List of Requesters
The Honorable Charles E. Grassley
Chairman
The Honorable John B. Breaux
Ranking Minority Member
Special Committee on Aging
United States Senate
The Honorable William V. Roth, Jr.
Chairman
The Honorable Daniel Patrick Moynihan
Ranking Minority Member
Committee on Finance
United States Senate
The Honorable John D. Dingell
Ranking Minority Member
Committee on Commerce
House of Representatives
The Honorable William M. Thomas
Chairman
The Honorable Pete Stark
Ranking Minority Member
Subcommittee on Health
Committee on Ways and Means
House of Representatives
Scope, Methodology, and Data Sources
We reviewed pertinent laws, regulations, HCFA policies, and research by
others to obtain information on the Medicare+Choice program, including
revisions to the payment methodologies. To obtain different perspectives on
why plans withdrew or reduced their service areas, we interviewed officials
at HCFA's regional offices and representatives from the American Association
of Health Plans and Blue Cross/Blue Shield of Maryland, one of the plans
that withdrew. To do our analysis, we obtained data files from HCFA, which
the agency uses to compute Medicare+Choice plan payments and which are
widely used by researchers.
To identify counties with a plan in 1999, we used HCFA's 1999 Medicare
Compare Database combined with HCFA's July 1999 Medicare Managed Care Market
Penetration for All Medicare Plan Contractors?Quarterly State/County/Plan
Data Files. We excluded cost, demonstration, and health care prepayment
plans from our analysis and used only those plans identified as
Medicare+Choice. We concluded that a plan was offered in a particular county
only if both databases agreed. The count of enrollees by plan by county in a
plan's service area as of July 1999 was obtained from the State/County/Plan
Penetration Files, except in four cases where plans reduced their service
areas and withdrew from only part of a county. In these cases, we obtained
the actual number of enrollees affected from HCFA's Center for Health Plans
and Providers.
Similarly, we identified counties with a plan in 2000 using HCFA's 2000
Medicare Compare Database combined with HCFA's March 2000 Medicare Managed
Care Market Penetration for All Medicare Plan Contractors?Quarterly
State/County/Plan Data Files. Again, we excluded cost, demonstration, and
health care prepayment plans from our analysis and used only those plans
identified as Medicare+Choice. We concluded that a plan was offered in a
particular county only if both databases agreed. HCFA's Center for Health
Plans and Providers gave us a list of contract consolidations that occurred
in 2000, and we adjusted our information accordingly. The count of enrollees
by plan by county in a plan's service area as of March 2000 was obtained
from the State/County/Plan Penetration Files.
To analyze the changes in plan participation in the Medicare+Choice program
in 2000 and 2001, we used HCFA data on Medicare+Choice plan contracts. In
July 1999, HCFA provided us with a list of plans that had announced they
were withdrawing from the program or reducing their service areas as of
January 1, 2000, and the counties and number of enrollees affected. In July
2000, HCFA provided us with the same information for plans that had
announced changes for 2001.
We excluded Guam, Puerto Rico, and the Virgin Islands from all county-level
analyses. In some of the analyses, the same counties are defined as separate
entities if plans can contract with them separately. For example, Los
Angeles County, California, is divided into Los Angeles-1 and Los Angeles-2;
they are counted separately because plans may contract with them separately.
The independent cities of Virginia are also counted as separate counties
because their payment rates differ from those of their counties, and plans
contract to serve these areas as if they were independent counties.
We classified counties as urban, rural, or small urban/fringe using the
rural/urban continuum codes in the February 1999 Area Resource File, which
we obtained from the Bureau of Health Professions, Health Resources and
Services Administration of the Department of Health and Human Services. We
defined urban counties as the central counties of metropolitan areas of 1
million population or more and rural counties as all nonmetropolitan
counties. Finally, included in the small urban/fringe counties are counties
in metropolitan areas of less than 1 million population and fringe counties
of metropolitan areas of 1 million population or more. The February 1999
Area Resource File combines the Virginia independent cities into their
original counties and does not report separate rural/urban continuum codes
for these. We kept these cities separate in keeping with the HCFA data, and
we assigned these cities the same rural/urban continuum codes as their
original counties.
To analyze geographic differences in beneficiaries' access to a plan from
1993 to 1998, we used the December 1993-1998 State/County/Plan Penetration
Files and deleted all plan/county combinations where a plan enrolled fewer
than 10 enrollees. Because we were not able to obtain actual contract
information on plan service areas before 1997, this provided an
approximation of plans' service areas. We then used the same urban, rural,
and small urban/fringe county designations as before from the February 1999
Area Resource File to determine the percentage of beneficiaries with access
to a Medicare+Choice plan in these different areas.
We obtained county-level payment rate information for 1997 through 2001 for
Medicare risk plans and Medicare+Choice plans, including payment reductions
resulting from the removal of graduate medical education (GME) spending,
from HCFA's Web site. In addition, we used a February 1999 file from HCFA's
Office of Information Systems containing historical county-level information
on the year that plans first entered individual counties.
Plans Withdrawing From Medicare+Choice, January 2000
(Continued From Previous Page)
Plan name Total number of enrollees
affected
Arkansas
Healthsource, Arkansas, Inc. 2,100
United Healthcare of Arkansas, Inc.a 200
Arizona
Blue Cross/Shield of Arizona 14,500
United Healthcare of Arizonaa 5,200
Premier Healthcare, Inc.a 4,500
Humana Health Plan, Inc.a 3,900
Health Plan of Nevada, Inc.a 3,200
California
Blue Cross of Californiaa 3,600
Cigna Healthcare of Californiaa 3,400
Pacificare of California, Inc.a 3,100
United Healthcare of California 1,600
National Med, Inc.a 1,400
Aetna U.S. Healthcare of Californiaa 900
Health Neta 200
Health Neta 0
Colorado
Cigna Healthcare of Colorado, Inc. 6,000
Qual-Med, Inc., Denver 5,000
HMO Colorado, Inc.a 2,700
Qual-Med, Inc., Pueblo 500
Qual-Med, Inc., Colorado Springs 200
Connecticut
Kaiser Foundation Health Plan of Connecticut 3,400
Physicians Health Service of Connecticut,
Inc.a 3,100
Connecticare, Inc.a 2,100
Florida
Florida Health Choice, Inc. 14,700
Community Health Care Systems, Inc. 3,200
Humana Medical Plan, Inc.a 3,000
HIP Health Plan of Florida, Inc. 2,600
Cigna Healthcare of Florida, Inc.a 2,100
Av-Med Health Plan, Inc.a 1,900
Av-Med Health Plan, Inc.a 1,600
Health Options, Inc.a 20
Georgia
United Healthcare of Georgia, Inc.a 600
Kaiser Foundation Health Plan of Georgia,
Inc.a 600
Iowa
Exclusive Healthcare, Inc. 1,500
Illinois
Humana Health Plan, Inc.a 1,300
Accord Health Plan 700
Kansas
Humana Kansas City, Inc.a 700
United Healthcare of the Midwest, Inc. 500
Cigna Healthcare of Kansas/Missouri 300
Louisiana
Ochsner Health Plana 17,000
United Health Care of Louisiana 17,000
Massachusetts
Harvard Pilgrim Health Carea 3,400
United Health Plans of New Englanda 2,100
Maryland
Free State Health Plana 14,700
United Healthcare of the Mid-Atlantica 600
Minnesota
Medicaa 2,800
Missouri
Cigna Healthcare of Kansas/Missouri 1,200
United Healthcare of the Midwest, Inc. 800
Nebraska
Exclusive Healthcare, Inc. 5,500
New Hampshire
Healthsource New Hampshire, Inc. 13,400
New Jersey
United Healthcare of New Jersey, Inc. 2,300
Physicians Health Services of New Jersey,
Inc. 2,200
Cigna Healthcare of New Jersey, Inc.a 1,400
Cigna Healthcare of New Jersey, Inc. 40
New Mexico
Presbyterian Health Plana 20
Nevada
Humana Health Plan, Inc. 9,500
New York
Capital Area Community Health Plan 9,700
Oxford Health Plans (New York) Inc.a 9,100
Capital Area Community Health Plan 7,800
Physicians Health Service of New York, Inc.a 4,200
United Healthcare of New York, Inc.a 3,400
Capital Area Community Health Plan 2,500
Vytra Healthcare 900
Cigna Healthcare of New Yorka 800
Hum/Healthcare Systems, Inc.a 300
Ohio
Pacificare of Ohio, Inc.a 4,200
Cigna Healthcare of Ohio, Inc. 3,700
Summacare, Inc.a 1,700
Community Health Plan of Ohio a 1,400
Qualchoice Health Plan a 1,300
Aetna U.S. Healthcare, Inc. a 600
Health Alliance Plan of Michigan a 0
Family Health Plan, Inc. a 0
Oklahoma
Bluelincs HMO, Inc. a 1,200
Oregon
Pacificare of Oregon II 1,400
Pacificare of Oregon, Inc. a 1,000
Providence Health Plan a 700
Pennsylvania
Healthamerica of Central Pennsylvaniaa 800
Cigna Healthcare of Pennsylvania, Inc. 0
Rhode Island
United Health Plans of New England, Inc. a 2,100
South Carolina
Companion Healthcare Corporation 1,100
Tennessee
United Healthcare of Tennessee, Inc. 500
United Healthcare of Tennessee, Inc. 100
Tennessee Health Care Network, Inc. a 40
United Healthcare of Tennessee, Inc. 20
Texas
Humana Health Plan of Texas a 23,500
Healthcare Partners Plans, Inc. 4,400
Cigna Healthcare of Texas, Inc. 2,700
HMO Blue, Northeast Texas 1,300
Humana Health Plan of Texas a 300
United Healthcare of Texas, Inc. 80
Virginia
Optima Health Plan 13,800
Healthkeepers, Inc. 2,900
Washington
Pacificare of Washington, Inc. a 6,000
Group Health Cooperative of Puget Sound 3,300
Healthplus a 1,400
Group Health Cooperative of Puget Sound a 900
Options Health Care, Inc. a 90
First Choice Health Plan a 10
Wisconsin
Humana Wisconsin Health Organization
Insurance Corporation 4,200
Primecare Health Plan, Inc.a 2,600
West Virginia
Carelink 0
aPlan remained in Medicare but reduced the number of counties served.
Sources: Medicare Compare Database, 1999; Medicare Managed Care Market
Penetration State/County/Plan Data Files, July 1999, www.hcfa.gov/medicare/;
Bureau of Health Professions, Area Resource File, Feb. 1999; and files of
contract nonrenewals and service area reductions from the Center for Health
Plans and Providers at HCFA.
Plans Withdrawing From Medicare+Choice, January 2001
(Continued From Previous Page)
Plan name Total number of enrollees
affected
Alabama
Health Partners of Alabama, Inc.a 2,500
Arkansas
HMO Partnersa 300
Arizona
United Healthcare of Arizona 12,300
Intergroup Prepaid Health Service of
Arizonaa 7,500
Cigna Healthcare of Arizona, Inc.a 4,300
Pacificare of Arizona, Inc.a 200
California
Cigna Healthcare of California 16,400
Aetna U.S. Healthcare of California 14,500
Blue Cross of Californiaa 5,500
National Med, Inc.a 5,200
California Physicians Servicesa 4,300
Health Neta 3,100
Maxicare, A California Corporationa 3,000
Blue Cross of California 400
Colorado
Pacificare of Colorado, Inc.a 4,000
Mutual of Omaha of Colorado, Inc. 500
Connecticut
Anthem Health Plans, Inc. (CT) 23,500
Aetna U.S. Healthcare, Inc. 14,500
Cigna Healthcare of Connecticut, Inc. 10,500
Physicians Health Service of Connecticuta 2,700
District of Columbia
Free State Health Plan 700
Cigna Healthcare Mid-Atlantic, Inc. 200
Delaware
Cigna Healthcare of Delaware, Inc. 3,600
Florida
Humana Medical Plan, Inc.a 25,600
Prudential Health Care Plan, Inc. 10,200
Cigna Healthcare of Florida, Inc. 8,600
Prudential Health Care Plan, Inc. 8,300
Cigna Healthcare of Florida, Inc. 7,300
Prudential Health Care Plan, Inc. 7,200
United Healthcare of Florida, Inc.a 7,000
Aetna U.S. Healthcare, Inc. 6,900
Prudential Health Care Plan, Inc. 5,200
Av-Med Health Plan, Inc.a 900
Physicians Healthcare Plans, Inc.a 300
Preferred Medical Plan, Inc. 200
Georgia
Cigna Healthcare of Georgia, Inc. 10,000
Aetna U.S. Healthcare of Georgia, Inc. 6,700
United Healthcare of Georgia, Inc. 2,800
HMO Georgia, Inc.a 300
Illinois
Humana Health Plan, Inc.a 7,200
Aetna Health Plan of Illinois, Inc. 6,400
United Healthcare of the Midwest, Inc.a 4,200
Mercy Health Plans of Missouria 400
Indiana
MaxiCare Indiana, Inc. 6,100
Anthem Insurance Companies, Inc. 2,500
Aetna Health Plan of Illinois, Inc. 500
Kentucky
Aetna U.S. Healthcare, Inc. 3,800
Pacificare of Ohio, Inc. 2,700
Humana Health Plan of Ohio, Inc. 2,700
Louisiana
HMO Louisiana, Inc. 9,400
Ochsner Health Plana 5,900
Aetna U.S. Healthcare, Inc. 5,200
Gulf South Health Plans, Inc.a 4,600
Maxicare Louisiana, Inc.a 0
Massachusetts
Harvard Pilgrim Health Carea 10,800
United Health Plans of New England, Inc.a 10,100
Fallon Community Health Plan, Inc.a 900
Maryland
Free State Health Plan 30,700
United Healthcare of the Mid-Atlantic 15,100
Cigna Healthcare Mid-Atlantic, Inc. 7,300
Maine
NYLCare Health Plans of Maine, Inc. 1,600
Michigan
Care Choices HMOa 100
Minnesota
Medica Health Plans 14,300
Missouri
HMO Missouri, Inc. 4,800
Group Health Plan, Inc.a 3,500
Humana Kansas City, Inc.a 1,800
North Carolina
Wellpath Select, Inc. 3,800
Qualchoice of North Carolina, Inc.a 80
New Hampshire
Harvard Pilgrim Health Care of New
Hampshirea 500
New Jersey
Oxford Health Plans (New Jersey), Inc.a 7,000
Cigna Healthcare of New Jersey, Inc. 4,400
Amerihealth HMO, Inc.a 800
Qualmed Plans for Health, Inc.a 200
New Mexico
Lovelace Health Plan, Inc.a 5,800
St. Joseph Healthcare PSO, Inc. 5,600
Presbyterian Health Plana 2,400
Qualmed, New Mexico 2,100
New York
Aetna U.S. Healthcare, Inc.a 34,900
HIP of Greater New Yorka 11,300
Cigna Healthcare of New York 7,200
MDNY Healthcare, Inc. 6,500
Health Services Medical Corps Central New
York 2,300
Physicians Health Service of New York,
Inc.a 2,100
Ohio
Aetna U.S. Healthcare, Inc. 27,800
Prudential Health Care Plan of Northern
Ohio 14,400
Humana Health Plan of Ohio, Inc. 8,500
Aetna U.S. Healthcare, Inc. 5,600
Pacificare of Ohio, Inc. 2,900
Paramount Care, Inc.a 2,900
Community Health of Ohioa 2,400
Kaiser Foundation HP of Ohioa 700
Summacare, Inc.a 600
Oklahoma
Bluelincs HMO, Inc. 3,600
Community Care HMO, Inc.a 3,600
Oregon
Providence Health Plana 5,800
Pennsylvania
HMO of Northeastern Pennsylvania, Inc. 30,700
Aetna U.S. Healthcarea 18,100
Penn State Geisinger Health Plana 16,600
Keystone Health Plan Central, Inc.a 6,100
Healthcentral, Inc. 5,900
Healthguard of Lancaster, Inc. 5,200
Healthamerica of Central Pennsylvaniaa 3,400
Qualmed Plans for Health, Western
Pennsylvania 3,000
Qualmed Plans for Health, Inc.a 500
Rhode Island
United Health Plans of New England, Inc.a 1,700
Tennessee
Health 1*2*3 13,700
Tennessee Health Care Network, Inc. 3,300
Healthsource Tennessee, Inc. 2,900
Texas
NYLCare Health Plans, Inc. 71,600
NYLCare Health Plans, Inc. 56,200
Humana HP of Texasa 20,500
Prudential Health Care Plan, Inc. 6,500
Seton Health Plan 6,100
Presbyterian Health Plan, Inc. 5,400
Methodistcare, Inc. 5,100
Pacificare of Texas, Inc.a 5,000
Cigna Healthcare of Texas, Inc. 2,500
Texas Health Choice, L.C.a 1,500
Prudential Health Care Plan, Inc., San
Antonio 333
Virginia
Cigna Healthcare of Virginia, Inc. 13,700
Cigna Healthcare Mid-Atlantic, Inc. 1,000
Washington
Aetna U.S. Healthcare, Inc. 10,600
Regencecare 7,900
Healthplus 7,100
First Choice Health Plan 3,600
Providence Health Plana 1,600
Pacificare of Washington, Inc.a 1,400
Wisconsin
Network Health Plan of Wisconsin, Inc. 1,400
Note: Approximately 8,400 of the affected enrollees for 2001 live outside of
the withdrawing plans' service areas. These enrollees are excluded from this
analysis.
aPlan remained in Medicare but reduced the number of counties served.
Sources: Medicare Compare Database, 2000; Medicare Managed Care Market
Penetration State/County/Plan Data Files, Mar. 2000, www.hcfa.gov/medicare/;
and files of contract terminations and service area reductions from the
Center for Health Plans and Providers at HCFA.
Comments From the Health Care Financing Administration
Comments From the American Association of Health Plans
Comments From the BlueCross BlueShield Association
Comments From the Health Insurance Association of America
(101850)
Table 1: Medicare Managed Care Enrollment and Geographic
Impact of the 2000 Withdrawals 13
Table 2: Largest Withdrawing Medicare+Choice Plans, January 2000 14
Table 3: Largest Withdrawing Medicare+Choice Plans, January 2001 15
Table 4: Key Characteristics of Plans That Terminated Their
Medicare Contracts in 2000 and 2001 21
Table 5: Average Annual Amount Spent per Enrollee on
Non-Medicare Benefits, 1999 and 2000 34
Table 6: Counties With Medicare+Choice Plans in July 1999
Affected by 2000 Withdrawals, by Year 2000 Payment Rates 36
Table 7: Counties With Medicare+Choice Plans in March 2000
Affected by 2001 Withdrawals, by Year 2001 Payment Rates 36
Table 8: Counties With Medicare Plans in July 1999 Affected by Withdrawals,
by Percentage Increase in 2000 Payment Rates 38
Figure 1: Proportion of Medicare+Choice and Affected Enrollees
in Rural, Fringe/Small Urban, and Large Urban Areas 12
Figure 2: Areas Where Sterling Option I Is Available, 2000 17
Figure 3: Plan Participation in Medicare and FEHBP, 1994-2000 19
Figure 4: Humana's Participation in Texas, 1986-2001 24
Figure 5: Plan Participation in Maryland, 1986-2001 28
Figure 6: Cumulative Increase in Average Medicare+Choice
County Rates and per Capita FFS Spending, 1997-2001 31
1. Prior to 1998, Medicare managed care plans that were paid a fixed amount
per enrollee were known as risk plans. Other types of Medicare plans
included cost contract and health care prepayment plans, which differed
substantially from risk plans both in how they operated and how Medicare
paid them. Nonrisk plans are being phased out. Risk plans, along with new
types of plans authorized by the Balanced Budget Act of 1997, are now known
as Medicare+Choice plans. When we refer to a "plan" in this report, we mean
either a risk plan or a Medicare+Choice plan.
2. Medicare Managed Care Plans: Many Factors Contribute to Recent
Withdrawals; Plan Interest Continues (GAO/HEHS-99-91, Apr. 27, 1999).
3. Medicare+Choice: Payments Exceed Cost of Benefits in Fee-for-Service,
Adding Billions to Spending (GAO/HEHS-00-161, Aug. 23, 2000).
4. Our studies include Medicare Plans: HCFA Can Promptly Eliminate Hundreds
of Millions in Excess Payments (GAO/HEHS-97-16, Apr. 25, 1997). Other
studies include G. Riley, C. Tudor, Y. Chiang, and M. Ingber, "Health Status
of Medicare Enrollees in HMOs and Fee-for-Service in 1994," Health Care
Financing Review, Vol. 17, No. 4 (Summer 1996), pp. 65-76 and Physician
Payment Review Commission, "Risk Selection and Risk Adjustment in Medicare,"
Annual Report to Congress, ch. 15 (Washington, D.C.: Physician Payment
Review Commission, 1996).
5. Medicare FFS payments include payments to teaching hospitals for
Medicare's share of the costs of providing GME, such as resident and faculty
salaries and overhead costs related to teaching activities.
6. ESRD is the stage of kidney impairment that is considered irreversible
and requires either regular dialysis treatments or a kidney transplant to
maintain life. It is an uncommon but expensive disease.
7. Of the approximately 3,100 counties in the United States, about 28
percent had a plan in 1999. In total, 644 counties are affected by the 2000
and 2001 withdrawals.
8. Small urban counties are defined as those counties located in urban areas
with populations under 1 million. Fringe counties are nonurban counties
adjacent to metropolitan areas.
9. A private FFS plan is an insurance plan under contract to HCFA that pays
providers for each covered service they deliver and allows enrollees to
obtain health care services from any provider willing to accept the plan's
payments (which are based on the Medicare FFS payment schedule). Medicare
pays the plan a fixed amount per enrollee. The plan may collect a monthly
premium from enrollees and require cost-sharing.
10. As of Sept. 1, 2000, fewer than half of the enrollees affected by recent
plan withdrawals will have this option available in their areas.
11. Medicare Managed Care Plans: Many Factors Contribute to Recent
Withdrawals; Plan Interest Continues (GAO/HEHS-99-91, Apr. 27, 1999).
12. Plan participation in Medicare has grown substantially, but not
consistently, since the risk contract program began in 1985. There was a
rapid increase in the number of participating plans between 1985 and 1987,
from about 100 to 165. By 1991, the number of plans had fallen to 93.
13. Federal Employees Health Program: Reasons Health Maintenance
Organizations Withdrew in 1999 and 2000 (GAO/GGD-00-100, May 2, 2000).
14. We define a plan's market share as its percentage of total
Medicare+Choice enrollment in an area.
15. Because plans are paid a fixed amount per enrollee, they rely on a large
enrollment population to spread the costs of the relatively few very costly
enrollees.
16. Plan mergers and contract terminations have reduced to four the number
of plans serving Maryland in 2000.
17. Calculated as the change in county payment rates for aged beneficiaries,
weighted by the number of aged beneficiaries in each county in 1997.
18. See GAO/HEHS-00-161 for a discussion of the forecast error and its
consequences.
19. Medicare+Choice: Payments Exceed Cost of Fee-for-Service Benefits,
Adding Billions to Spending (Aug. 23, 2000, GAO/HEHS-00-161).
20. Based on the first year of plan-submitted data, HCFA estimates that the
new methodology would reduce average payments by 5.9 percent if fully
implemented. In 2000, only 10 percent of plan payments were adjusted using
the new method. Consequently, plan payments were reduced by less than 1
percent. In 2002, the portion of plan payments adjusted by the new
methodology will rise to 20 percent.
21. These projections are included in each plan's adjusted community rate
proposal. Costs are calculated on the basis of how much a plan would charge
a commercial customer to provide the same benefit package if its members had
the same expected use of services as Medicare beneficiaries, and therefore
these "costs" include expected profits from commercial customers.
22. According to the Department of Health and Human Services (HHS) Office of
Inspector General (OIG), the difference may be even greater. In several plan
audits, the OIG concluded that the plans had overstated their Medicare
costs. HHS OIG, Administrative Costs Submitted by Risk-Based Health
Maintenance Organizations on the Adjusted Community Rate Proposals Are
Highly Inflated, A-14-97-00202 (July 1998).
23. Plans' adjusted community rate proposals for 2001 have not yet been
approved. Although an analysis of these proposals would provide updated
information on plans' projected costs, this information will be available
only for those plans that will participate in Medicare+Choice next year.
24. When the 1999 county rates were established in 1998 (and plans made
their 1999 participation decisions) per capita FFS spending was expected to
increase.
25. Floor counties are defined as those counties assigned the minimum rate
of $415 in 2001. The average increase is computed for those counties with
Medicare plans in 1997, weighted by total plan enrollment in 1997.
County-level data on FFS spending since 1997 are not available.
26. In 2001, rates in about 31 percent of counties will equal the floor
payment rate of $415. The remaining county rates will equal the minimum 2
percent increase over the 2000 rates. The blended rate will not be paid in
any county. The BBA included a budget neutrality requirement that specifies
that total payments based on the minimum increase, the floor rate, and the
blended rates must equal the aggregate payments that would have been made if
payments were based on area specific rates only. Only blended rates may be
adjusted to fulfill the budget neutrality requirement. In 2001, funding the
minimum increase and floor amounts will push aggregate spending above the
budget neutrality amount. Therefore, the blended rates cannot be funded.
27. Among counties that had a Medicare+Choice plan in 1999.
28. Our methodology for estimating these costs is described in
Medicare+Choice: Payments Exceed Cost of FFS Benefits, Adding Billions to
Spending (GAO/HEHS-00-161, Aug. 23, 2000).
29. A recent report by the Department of Health and Human Services Office of
Inspector General, HMO Withdrawals: Impact on Medicare Beneficiaries,
OEI-04-00-00390 (Aug. 2000), addresses this issue directly and contains the
results of surveys of beneficiaries affected by the 2000 withdrawals.
*** End of document. ***