Statement of Madeleine Smith, Ph.D., Specialist in Social Legislation,
Domestic Social Policy Division, Congressional Research Service, Library of Congress

Testimony Before the Subcommittee on Health
of the House Committee on Ways and Means

Hearing on the Medicare+Choice: Lessons for Reform

May 1, 2001

There are two points that I would like to emphasize about the effects of payment reform under Medicare+Choice:

First, although the number of health maintenance organizations (HMOs) in the M+C program has declined, the proportion of Medicare beneficiaries enrolled in managed care has not changed much. In 1997, 14% were enrolled; today, 15% are enrolled. This fairly constant percentage of beneficiaries enrolled in HMOs followed a period of rapid growth in enrollment that has not continued. Fewer beneficiaries have access to HMOs nationwide, but with the entry of a private fee-for-service plan into the program, access to an M+C plan for rural beneficiaries has risen.

Second, variation in payment rates has decreased. In 1997, the highest rate was 3 ½ times the lowest rate. Today, the highest rate is 1¾ times the lowest rate. However, benefits offered by M+C plans still vary widely across the country.

In the remainder of my testimony, I will review how rates were determined before the M+C program, and major reasons for reform of the payment system. Then I will turn to a brief discussion of how rates are currently calculated. Finally, I will summarize one effect of rate reform - plan withdrawals - and changes to the M+C payment rate calculations enacted since 1997.

Pre-BBA

Medicare has included a managed care alternative to traditional fee-for-service for almost 30 years, since the 1970s. Under the risk contract program created in 1982 (Section 1876 of the Social Security Act), an HMO received a single monthly capitation payment for each of its enrollees. This payment was known as the adjusted average per capita cost, or AAPCC. In return for the monthly payment, the HMO agreed to provide or arrange for the full range of Medicare services through an organized system of affiliated physicians, hospitals, and other providers.

The Health Care Financing Administration (HCFA) calculated the AAPCC for each of the over 3,000 counties in the US. A county's AAPCC was based on the costs of providing care under traditional fee-for-service (FFS) Medicare to a beneficiary in the county. Basically, HCFA determined the average per capita costs by adding together all of the Medicare FFS expenditures for beneficiaries living in the county, and dividing this by the number of FFS beneficiaries in the county. This county-level average per capita cost was adjusted for demographic differences between the county's Medicare beneficiaries and average beneficiaries nationwide. The county rate was set equal to 95% of the AAPCC to account for savings delivered by managed care organizations through coordination of care. Actual payments to HMOs for individual enrollees were adjusted for risk, using demographic characteristics of the enrollees, such as age, gender, and residence in an institution.

Each HMO was required to submit an estimate of its costs of covering Medicare services for its Medicare enrollees. This estimate is known as the adjusted community rate (ACR), and is still submitted today. If the AAPCC was greater than the ACR, the HMO was required to reduce beneficiary cost-sharing, enhance benefits, contribute the excess to a stabilization fund, or return the funds to HCFA. Many HMOs were able to provide additional benefits, such as prescription drug coverage, without charge to an enrollee because the AAPCC exceeded their ACR.

Reasons for Payment Reform

There were at least three main reasons behind reform of the AAPCC payment method under the Balanced Budget Act of 1997 (BBA, P.L. 105-33): lack of access to a Medicare HMO in many areas; wide variation in the payments and benefits offered by HMOs; and volatility of payment rates over time.

Lack of access to an alternative to FFS Medicare was the first perceived problem. The risk contract program expanded dramatically between 1993 and 1998, when the number of plans tripled from 110 to 346. In 1998, almost three-fourths of Medicare beneficiaries had access to at least one risk plan, and almost two-thirds had a choice of plans. Still, over one-quarter of Medicare beneficiaries nationwide lacked access to a risk plan, and most of these beneficiaries were in rural areas. Over 90% of Medicare beneficiaries in rural areas lacked access to a risk plan, while all beneficiaries in central urban areas had such access. Many of the counties without plans had low AAPCCs.

A second perceived problem was wide variation in payments and benefits offered by HMOs in different areas. In 1997, the highest payment rate was 3½ times the lowest rate: $767 versus $221 monthly for an aged beneficiary. An analysis of ACRs in 1995 showed that HMOs in Miami were required to offer benefits worth over $100 per month without charging enrollees anything: the payment rate was $100 per month higher than the HMO's costs of covering Medicare's benefits. In contrast, HMOs in Minneapolis were not required to offer any additional benefits: the payment rate was equal to the HMO's costs of covering Medicare's benefits. Beneficiaries in the federal Medicare program were receiving benefits that differed across localities.

A third perceived problem was volatility of the AAPCC over time, especially in rural counties. This problem occurred because of the relatively small number of Medicare beneficiaries in some counties: today, one county has 18 Medicare beneficiaries. If one beneficiary in a sparsely populated county incurred large Medicare expenditures in one year, the average per capita costs would skyrocket. If that beneficiary recovered or died, the next year the average per capita costs could plummet. Wide variation in payment rates over time was considered one obstacle to risk plan entry into some counties.

Other problems were more technical. The AAPCC was calculated based on average FFS Medicare costs. The costs of care provided to Medicare beneficiaries by Veterans Affairs (VA) or Department of Defense (DOD)facilities were excluded from the calculation. This could depress a county's AAPCC. AAPCCs also included payments for disproportionate-share hospitals (DSH) and graduate medical education (GME) even though some questioned whether HMOs were passing these funds through to hospitals.

Payments under M+C

In order to address some of these problems, BBA 97 included a new payment rate formula. The M+C rate in a county was set at the highest of 3 amounts:

• a floor, or minimum amount, set at $367 in 1998;
• a blend, or average, of local and national rates;
• a minimum update representing a 2% increase over the prior year's rate.

The blend calculation used the 1997 AAPCCs as the base local rate. National rates were an average of local rates, adjusted to reflect differences in input prices in each county. A portion of GME payments was excluded from the local rates used to compute the blend, beginning with 20% in 1998 and rising to 100% by 2002. The blend was phased-in. In 1998, 90% was based on local rates and 10% on the national rate; in 2003 and thereafter, 50% will be based on local rates and 50% on the national rate.

The formula included a floor and minimum update to alter the immediate effects of blending local and national rates. The floor increased rates in low payment counties more quickly than would occur through blending of local and national rates. The minimum update was included to cushion the effects of blending on high payment counties. At the time of enactment, analysis projected that over 80% of counties would be receiving blend payment rates by 2003. Among remaining counties, 16% would receive floor rates and 2% would receive minimum updates.

Payment rates were affected by other provisions in BBA 97, including statutory reductions in the national per capita growth percentage used to compute the local rate and the floor, and the budget neutrality provision which requires that aggregate M+C payments equal total payments that would have been made without changes to the formula. Both of these components were meant to guarantee budgetary savings. The M+C payment formula removed funding of GME from the calculation, but left DSH payments in the formula. No adjustments were made to account for care received through VA or DOD facilities. Finally, HCFA was required to implement a new risk adjustment system, based on the health status of beneficiaries, beginning in 2000.

Plan Withdrawals and Legislative Responses

The M+C program has now experienced three waves of plan withdrawals and service area reductions, effective at the onset of the M+C program in 1999, and annually since then. Interspersed between announced withdrawals have come two legislative responses, the Balanced Budget Refinement Act of 1999 (BBRA, P.L 106-113) and the Medicare, Medicaid, and SCHIP Benefits Improvement and Protection Act of 2000 (BIPA, P.L. 106-554).

Not all HMOs that had operated under the predecessor program chose to convert to the M+C program in 1999. According to HCFA, the 66 organizations that withdrew or reduced service areas affected slightly more than 400,000 beneficiaries in risk plans in 1998, about 6% of all risk enrollees. Slightly more than 50,000, less than 1% of risk plan enrollees, did not have access to another managed care plan and were forced to return to traditional FFS Medicare.

Plans announced further withdrawals and service area reductions in 1999 and 2000. Of the approximately 300 plans serving Medicare beneficiaries at the end of 1999, 99 plans withdrew or reduced service areas for the 2000 contract year, and 118 withdrew or reduced service areas for the 2001 contract year (GAO, 2000). These changes affected about 5% of M+C enrollees in 2000, and about 15% in 2001. About one-fourth of affected beneficiaries in 2000, and 15% in 2001, had no other managed care option available.

Why did plans withdraw completely or reduce service areas? Industry representatives believe that inadequate M+C payment rates are a principal cause of plan withdrawals. HCFA contends that withdrawals reflect strategic business decisions by M+C organizations that transcend payment rate issues. Studies of withdrawals by CRS, GAO and others have found that in 2000 M+C plans tended to withdraw from rural counties, where they may have had difficulty maintaining provider networks, and large urban areas, which they had recently entered or where they lacked sufficient enrollment. Similar results were found for 2001, with the added withdrawal of some plans with more extensive program participation. GAO notes that the pattern of M+C withdrawals resembles the experience of the Federal Employees Health Benefits program (FEHBP), with rapid expansion of plan participation between 1994 and 1997, followed by withdrawals of more recent entrants with few enrollees. A recent report from InterStudy indicates similar events in the general HMO market. In 1999, 83 HMOs (12%) ceased operations, many through merger, but 29 HMOs failed. The industry experienced its first annual decline in enrollment in nearly 30 years. Rural areas accounted for the greatest loss in enrollment, and 91% of HMO enrollees now live in urban areas. The boom cycle experienced by HMOs in the mid-1990s came to a close.

Congress acted to increase M+C payment rates. The BBRA in 1999 made a few modest changes to raise future plan payments by decreasing the scheduled reduction in the national per capita M+C growth percentage, and by reducing assessments for beneficiary education. It established bonus payments for plans that enter areas where no other plan is in operation, to encourage participation in rural areas, and it slowed down the Secretary's scheduled phase-in of risk adjustment. The BIPA in 2000 made more substantial changes to increase payments. For 2001, the floor rate was raised to $475 per month in lower populated areas, and $525 in areas with population of more than 250,000. The minimum increase in rates was raised from 2% to 3% for 2001. BIPA also extended the current risk adjustment method until 2003 (when a new risk adjustment method will be phased-in), and expanded the new entry bonus payments to encourage participation. Many other provisions with less general impact on payment rates were included. One notable BIPA provision allows M+C plans to offer reductions in the Medicare Part B premium as an additional benefit to enrollees, beginning in 2003.

Effects of Payment Reform

After a little over 2 years, have problems identified with the AAPCC been fixed? Lack of access was seen as a consequence of low payment rates. The BBA raised the floor to $367 per month, and the BIPA raised it again to $475/$525. Has access increased? In 1997, there were over 300 risk HMOs, and in 1998, there were 346. Today there are 179 M+C plans. The number of plans has dropped to about half.

Although the number of plans has decreased significantly, the proportion of beneficiaries enrolled has not changed much. In 1997, about 5.2 million Medicare beneficiaries (14%) were enrolled in risk plans. This increased to 6.2 million beneficiaries, or almost 17%, by 1998. In April 2001, there were 5.7 million Medicare+Choice enrollees, representing about 15% of the Medicare population.

Thirty-three percent (33%) of Medicare beneficiaries lacked access to a risk plan in 1997, including 91% of beneficiaries in rural areas. By 2001, 37% overall lacked access to an HMO, including about 85% of beneficiaries in rural areas. With the entry of a private FFS plan, offered by Sterling Life Insurance Company, into the M+C program, access has increased. Sterling now offers coverage in over half of the states and counties in the country, where 38% of all beneficiaries reside, including 57% of beneficiaries living outside metropolitan areas. Sterling provides access to 18% of beneficiaries who would not otherwise have an M+C option.

Many counties not served by a managed care plan are served by Sterling, as shown in the map.⁽¹⁾ Sterling tends to operate in lower payment counties which previously did not have an M+C option. Sterling may serve these areas for several possible reasons. First, M+C payment rates may be higher than the average cost of fee-for-service care, especially in floor counties. Second, Sterling receives a 5% bonus for the first year, and 3% bonus for the second year, for serving counties that previously were not served by an M+C plan. Third, private fee-for-service does not require a network of providers, which is difficult to assemble and maintain in rural areas. Sterling pays providers the same rate as traditional Medicare, and does not permit providers to bill beneficiaries additional amounts (i.e., balance bill). Sterling provides very few additional benefits beyond the Medicare defined benefit package, but may reduce cost sharing for some beneficiaries, depending on the exact services used. Currently, Sterling charges enrollees a $65 monthly premium; enrollees must continue to pay the Medicare Part B premium of $50 per month.

Another goal of payment reform was to decrease the variation in payment rates and benefits. This has occurred. In 1997, the highest payment rate was 3½ times the lowest rate. Today, the highest rate is 1¾ times the lowest rate ($834 versus $475), and the spread is even lower across metropolitan areas (about 1.6 times, $834 versus $525). This narrowing of differences in payment rates has been achieved by raising the minimum payment, or floor, while restraining growth in the highest paid counties to a 2% (3% in 2001 only) increase per year. (Managed care plans have argued that their costs have risen much more than 2% annually. HCFA projects an increase of 15.4% in nationwide per capita Medicare costs from 1997 to 2001. Plans receiving minimum updates over this period saw rates increase by 9.3%.) Additionally, as the payment gap has narrowed, benefits under M+C generally have declined. In 1999, 61% of beneficiaries had access to plans that charged no additional premium, and 54% had access to a plan that charged no additional premium while including drug coverage. By 2001, only 37% of beneficiaries had access to a $0 premium plan; only 26% had access to a $0 premium plan with drug coverage.

Recall the difference in benefits available in Miami and Minneapolis in 1997. Differences persist today. Several plans in Miami charge enrollees no additional premium and include full coverage of prescription drugs, both generic and brand name, for drugs on the plan's formulary. Contrast this to Minneapolis, where there are four M+C plans, three HMOs and Sterling FFS. Only one HMO offers any prescription drug coverage. For $81 per month, Minneapolis enrollees are covered for $100 in total drug expenditures every 3 months, for a total of $400 of coverage per year. In both cities, the HMO plans reduce cost-sharing and provide additional benefits, such as physical exams, eye care, and dental care.

Finally, payment reform was intended to reduce volatility in payments over time. Certainly payments have not decreased, as they did prior to M+C, but very large increases have occurred in some areas as a result of increases in the payment floor. Some counties saw rates rise over 100% between 1997 and 2001. The most recent rise in floors produced an increase of 14% in rates in non-metropolitan areas (from $415 in 2000 to $475 in 2001) and 26% in metropolitan areas (from $415 to $525). Moreover, some plans are receiving an additional 5% bonus increase in rates because they entered previously unserved areas.

This concludes my testimony. I thank the Committee for this opportunity to discuss M+C payment rates and will be happy to answer your questions.

1. The map illustrates service areas of coordinated care plans and Sterling Life Insurance's private fee-for-service plan. Coordinated care plans include health maintenance organizations, preferred provider organizations, and provider sponsored organizations. Some coordinated care plans were not accepting new enrollees in February 2001. These plans are not included in the analysis summarized by the map.