Health Affairs

This Article Abstract Submit a response to this article Services E-mail this article to a friend Alert me to new issues of the journal

Perspectives on this article are available by clicking on the author's name below:
Joseph Antos • John Bertko • Ron Klar • Murray Ross •
Patricia Salber and Bruce Bradley

H E A L T H  A F F A I R S : M E D I C A R E  R E F O R M 
W E B  E X C L U S I V E   

28 N O V E M B E R 2001  
  

Medicare+Choice: Doubling Or Disappearing? 

A proposal to change the focus of this troubled program and use it to
reward private plans that improve quality and help manage the care of
Medicare beneficiaries with chronic diseases.

by Robert A. Berenson

ABSTRACT:

Although the changes in the program created by the Balanced Budget Act are often viewed as the reason for the current instability in the Medicare+Choice (M+C) program, in fact, health plans are having difficulties in all of their markets, not just in Medicare. It may be time to reconsider the purpose of the program and to fundamentally redesign how payments are made to managed care organizations contracting with Medicare. Two alternative approaches are suggested: treating M+C like another provider type by severing the payment linkage to spending under traditional Medicare, and overhauling the program by creating a value-based purchasing orientation rewarding plans that provide higher-quality care to beneficiaries with chronic diseases.

The balanced budget act (BBA) of 1997 was supposed to accelerate enrollment growth in Medicare's risk-based, private-sector alternatives to traditional fee-for-service (FFS) Medicare. The Congressional Budget Office (CBO) estimated that enrollment in capitated plans would be about 15 percent higher by 2005 than under the prior law that governed the Section 1876 Medicare risk program, reaching nearly 34 percent of total Medicare enrollment.¹ Analysts also thought that the BBA would produce many more private-sector insurance options and would replicate the range of options available in the commercial sector, including preferred provider organizations (PPOs), provider-sponsored organizations (PSOs), private FFS plans, and medical savings accounts (MSAs). The CBO estimated there would be 125 new PSOs by 2002.²

Now, four years later, while Tom Scully, administrator of the newly named Centers for Medicare and Medicaid Services (CMS, formerly HCFA), seeks a 30 percent Medicare+Choice (M+C) market share by 2005, others question whether the M+C program has much of a future at all.³ Since January 2000 M+C enrollment has actually been declining. The total of new types of plans is four: a PSO in New Mexico; Sterling, a private FFS plan that began enrolling last year; and two PPOs.⁴ The CMS recently announced that more than 500,000 beneficiaries will be affected by plan contract terminations or service area reductions for 2002, bringing the four-year total of beneficiaries affected by plan withdrawals to about 2.2 million.⁵

In hindsight, it seems clear that some of the BBA expectations were based more on wishful thinking than on clear-headed analysis. For example, many of the non-health maintenance organization (HMO) alternatives that do well in the private sector, such as PPOs, often do not actually manage risk but rather administer benefits for self-funded employers. Even if HMOs were not withdrawing from M+C, it is not surprising that PPOs, which use fewer cost- -containing approaches than HMOs use, would not choose to take on the serious risk of managing Medicare risk. Provider-based organizations already could manage risk as HMO subcontractors without taking on the responsibilities of actually becoming a licensed insurer and, in effect, competing with the organizations providing them business.⁶

For HMOs, competition in Medicare is very different from what they face in commercial markets. For employers concerned about escalating costs, HMOs have more cost control tools than managed indemnity and PPOs options have.⁷ In contrast, the traditional FFS Medicare program, for all of its acknowledged flaws and inefficiencies, has the market power needed to impose fairly tight reimbursement rates, when politically allowed to do so, and functions with administrative costs below 2 percent of spending.⁸

No one predicted the extraordinary success that the FFS Medicare program has had in controlling costs since passage of the BBA, thereby reducing the spending available for M+C plans, whose funding is tied, by formula, directly to traditional FFS program spending increases. At the same time, coincident with passage of the BBA, health plans have experienced major inflationary trends in health care costs in all of their product lines, increasing instability in provider networks, and a public backlash against some of their cost-containing approaches.

Although plans participating in M+C have been experiencing nearly double-digit rates of cost inflation since passage of the BBA, few have received M+C rate increases of more than 2 percent per year over that period. Although Congress has included M+C in the two "giveback" bills passed in 1999 and 2000, most of the changes were marginal and have not stabilized the M+C program. Of particular note, the health plan withdrawals and current doubts about the future of the M+C program are taking place at the same time that President George W. Bush and Republican congressional leaders are recommending competition-based Medicare reform that assumes a vibrant private health insurance sector that is actively seeking Medicare business.

This paper outlines the current payment issues and options for the M+C program. It begins with a brief history that makes the point that many of the current perceived problems with M+C are, in fact, the result of deliberate policies to solve other problems. Next it considers the most important payment issues that Congress controls, including payment adequacy, payment distribution, and risk adjustment. After identifying the immediate choices Congress has to modify payment to M+C plans, the paper concludes with a consideration of two alternative payment models for the M+C program: severing the linkage between M+C and traditional Medicare altogether, and moving to a value-based purchasing approach focused on rewarding plans for improved quality and care coordination.

Brief History Of The TEFRA Risk Program

In the Tax Equity and Fiscal Responsibility Act (TEFRA) of 1982, Congress authorized federally qualified HMOs and nonqualified "competitive medical plans" to be paid 95 percent of the adjusted average per capita cost (AAPCC) for each county–essentially what Medicare paid for the average beneficiary in each county–reflecting different rates based on age band, sex, institutional status, and Medicaid status. The 5 percent differential from FFS payments was meant to account for the presumed greater efficiency of HMOs, thus reducing government outlays.

After an unsettled beginning that saw a major reduction of contracting plans in the late 1980s, the risk program experienced substantial growth by the mid-1990s, coinciding with HMOs' success in commercial markets. From about 1.5 million at the beginning on 1993, enrollment rose steadily and peaked at about 6.35 million in 1999, representing annual rates of increase in the 30-40 percent range and approaching one million beneficiaries per year between 1993 and 1997.⁹ At peak enrollment in 1999, twelve states had M+C enrollment exceeding 20 percent of the beneficiary population; California, with 1.5 million M+C enrollees, had 40 percent.¹⁰

It is important to note that during the period 1993-1997 average growth in Medicare spending per enrollee exceeded private health insurance spending by four percentage points (7.5 percent to 3.5 percent), reflecting HMOs' relative success in holding down costs for all of their products in that period.¹¹ Yet even though their own costs were well controlled, the AAPCC payment methodology gave Medicare risk contractors payment increases corresponding to the inflationary Medicare FFS experience.

The structure of Medicare risk contracting was such that Medicare HMOs' relative efficiencies in this period were either returned to beneficiaries in the form of richer benefits or retained by the HMOs as excess profits. Indeed, because major HMOs enjoyed a disproportionate share of operating profit from their Medicare book of business, policy analysts asked whether HMOs actually shifted costs by having high Medicare reimbursements subsidize lower premiums in their commercial accounts.¹²

Despite the growing risk-contract enrollment, problems existed with the method of payment to health plans, from the perspectives of health plan administrators, policy analysts, budgeteers, and consumer advocates alike. As summarized in the March 1997 Annual Report of the Physician Payment Review Commission (PPRC), "Many observers think current policies limit the growth of Medicare managed care by paying too little in some markets and promoting it in others by paying more than necessary to compensate plans fairly. In any case, these policies hamper Medicare's ability to benefit from the efficiencies of managed care. Moreover, they do not encourage beneficiaries to make cost-conscious choices."¹³

Risk-plan payments, based on the AAPCC methodology, varied widely across the country and could not be explained by the differences of local input prices or of the underlying health status of Medicare beneficiaries residing in the various counties. The PPRC emphasized that the differences partly reflected different practice styles and patient preferences, factors that it thought should not justify such large payment differentials and that contributed to an uneven pattern of enrollment and wide variation in health plans' ability to offer additional benefits.

The health plans themselves had difficulty with unpredictably large year-to-year variations in payment increases, which existed even though the CMS used a five-year moving average of FFS county-level per capita spending. This unpredictable annual fluctuation in county payment rates was a particular problem in small counties, where, because of small numbers of FFS beneficiaries, rates could swing widely year to year. These less populous counties, predominantly rural, were precisely the ones that, for various reasons, had few risk plans.

Also, by 1997 it had become established that Medicare beneficiaries in HMOs were healthier than those in the traditional FFS program that formed the basis for HMO payments.¹⁴ Yet by taking into account only demographic information and not health status, the payment methodology resulted in "overpaying" health plans that were experiencing favorable selection and providing a disincentive for plans to invest in approaches for the efficient, high-quality management of expensive, chronic conditions.

BBA Payment Changes And Subsequent Modifications

All of these factors led Congress to greatly alter the way it paid health plans. In brief, payment rates under the BBA were fixed at the 1997 county amount trended forward based on a formula that consisted of paying the highest of a blend of the local and national rate, a national minimum payment amount (or floor), and a guaranteed increase of 2 percent from the prior year's county rate.¹⁵ The complex payment formula was designed to reduce gradually, although not eliminate, the payment-level differences based upon historical patterns of traditional FFS program spending and to reduce year-to-year payment fluctuations. In addition, the BBA also called for the CMS to begin adjusting payments for the underlying health risk status of beneficiaries.

Also, as part of the overall goal of reducing Medicare spending to meet budget targets, the BBA placed a budget-neutrality factor into the formula, reduced the capitation rate updates by a cumulative 2.8 percent less than the projected percentage increase in FFS costs for fiscal years 1998-2003, and phased out payment associated with hospital payments for graduate medical education (GME), which would now be made directly to teaching facilities and would produce a 3 percent lower payment rate when fully phased out.

Although these changes were significant, it is important to emphasize that the BBA kept the structure of plan payments basically intact. It maintained an administrative pricing system for plans instead of moving toward any form of competitive bidding. It also maintained the county as the geographic area on which payments would be based, although now based on 1997 payment levels trended forward. Finally, the BBA continued to tie plan payment updates to the spending increases in the traditional Medicare FFS program, albeit with payment updates based on national-level FFS per capita spending rather than on county-level per capita spending.

Problems with plan withdrawals, in the form of both contract terminations and service area reductions, surfaced in 1998 and have escalated since. Marsha Gold recently summarized the extent of plan withdrawals, as well as benefit cutbacks and premium increases, documenting that beneficiaries are facing higher premiums and reduced benefits.¹⁶ Enrollment in M+C risk plans at the end of 1999 was about 6.35 million–more than 16 percent of the eligible beneficiary population. As of 1 August 2001 enrollment had declined to 5.61 million, about 14 percent of eligible beneficiaries, with further decreases virtually certain in January 2002.¹⁷

Although most of the discussion of and proposed remedies for the troubled M+C program focus on payment, the reason for plan withdrawals is more complex. Researchers at Abt Associates and the University of Minnesota found that in the recent years of plan pullouts there was no relationship between either absolute payment levels or payment level changes and the likelihood of plans' exiting M+C. At the same time, they found the expected relationship between payment levels and benefits-that is, plans in counties with higher payment rates offer more generous benefits.¹⁸

Similarly, Randall Brown and Gold found that market-based reasons in addition to payment level were influential in whether plans participate in Medicare: historical non-Medicare managed care patterns, practice patterns, beneficiaries' care expectations and characteristics, patterns of supplemental coverage, extent and form of provider organization, concurrent goals and trends in other lines of business, state regulatory context, and geographic location of the market.¹⁹ Others have criticized the complexity of government regulations, administrative requirements, and uncertainty as factors creating an environment not hospitable to M+C contracting.

Despite yearly concerns related to withdrawals and growing public awareness of decreased benefits, for the most part Congress has made only marginal changes that have not changed the basic direction of the program.²⁰ In the 1999 Balanced Budget Refinement Act (BBRA) Congress added some funds to the M+C formula, partly by increasing spending on the FFS program, thereby increasing the national update factor on which the M+C payments are based.

The Medicare, Medicaid, and SCHIP Benefits Improvement and Protection Act (BIPA) of 2000 made more substantial changes in M+C payments, but still within the BBA framework. The major change was to raise the base floor payment to $475 and to raise the floor payment for counties in urban areas with populations over 250,000 to $525. For 2002 the base floor rate will become $500.37, and the qualifying urban county floor payment will become $553.04.²¹ Contrasting these floor payments with pre-BBA rates demonstrates how dramatic these floor payment increases have been. In 1997 about half of all beneficiaries lived in counties with rates between $390 and $530.²² Payments for certain urban and rural counties have actually doubled, but plan participation in these areas remains rare.²³ Looking to 2002, as a result of the two giveback bills that returned funding to FFS providers and M+C plans and of some actuarial reestimates, earlier this year the CMS announced the increase in the national per capita M+C growth percentage for calendar year 2002 as 7.99 percent. For 2002 about 20 percent of counties (or about 660 counties) will receive the minimum percentage increase of 2 percent, and the remaining 80 percent will receive the new floor county rates. A number of counties that otherwise would have received blended payments of greater than 2 percent under the BBA formula actually will be limited to 2 percent increases because of the BIPA priority on raising floor payments.²⁴

The Future Of M+C Under Current Law

Although current payment levels may not have been the decisive factor that determined whether M+C plans exited the market in recent years, certainly expectation of future payment is a major factor that health plans consider in determining whether to continue a commitment to the M+C program. Accordingly, a number of payment-related issues must be addressed in assessing M+C's future.

Payment adequacy. The issue of whether M+C plans are paid enough in aggregate has been the most contentious issue related to the M+C program, with many arguing simply that overpaid commercial enterprises typically do not exit markets. The American Association of Health Plans (AAHP) described a "fairness gap" in payment, projecting that by 2004 some urban areas would see M+C payments below 80 percent of FFS payment levels for those areas, thereby creating a competitive disadvantage for plans trying to attract beneficiaries from the traditional FFS program.²⁵

Health plans' assertions about a fairness gap notwithstanding, the U.S. General Accounting Office (GAO) and the US Department of Health and Human Services (HHS) Office of Inspector General (OIG) independently documented that actual payment rates for M+C plans since 1997 have in fact risen faster than per capita FFS spending, a result mostly of two things: a technical mistake in the BBA that prevented the CMS from readjusting the 1997 payment base for prior incorrect estimates, and the guaranteed 2 percent minimum increase that turned out to be more than would have been made under pre-BBA methodology.²⁶ Recently, CMS actuaries estimated that average payments in 2001 for beneficiaries enrolled in the M+C program actually will be about 98 percent of spending for those in the traditional FFS program.²⁷

Taking favorable plan selection into account contributes to even higher plan payments. The GAO has estimated that, overall, plans were paid 21 percent more in 1998 than the FFS program would have spent to provide Medicare-covered benefits to plan enrollees, about 60 percent due to the absence of risk adjustment and the rest to the statutory-based overpayments described above.²⁸

How does one reconcile the two views–on the one hand, that payments are too low, causing plans to withdraw; on the other hand, that payments are excessive in relation to per capita FFS spending? Perhaps plans need to be able to offer substantial additional benefits, at little or no extra cost, to entice beneficiaries to give up the freedom of choice and other perceived advantages of FFS, even when facing a supplemental (Medigap) plan costing $150-$300 per month. As payments have not kept up with costs, M+C plans are not able to be as generous with additional benefits and have either withdrawn, put on capacity limits restricting new enrollment, or cut back on their benefits. In short, plans are overpaid in relation to FFS but underpaid in relation to what Medicare beneficiaries apparently are seeking in the market.²⁹

In recent years health plans have been having difficulty managing costs in all of their lines of business, partly because of the spiraling costs of prescription drugs.³⁰ In addition, insurers recently entered the phase of the insurance cycle in which they are raising premiums to restore profitability after a period of holding down premiums to try to gain market share.³¹ These two factors have combined to drive up insurance premiums, from 4.8 percent in 1999 to 11.0 percent in 2001, the largest annual increase since 1992.³²

Employers have swallowed hard and are paying these nearly -double-digit increases, with employees partly protected at least from the direct premium increases. In contrast, the structure of M+C payments, relying on a fixed, formula-driven Medicare contribution that does not take plan cost inflation into account, requires beneficiaries to pick up any additional premiums that plans may charge.³³ Faced with not being able to pass on cost inflation directly to the purchaser–in this case, Medicare–plans have adopted various strategies, some exiting M+C, others staying in the program by reducing benefits and raising individual premiums.

Payment distribution. In its 2001 Annual Report, the Medicare Payment Advisory Commission (MedPAC) reconsidered its previous recommendation to support redistribution of M+C payments from high- to low-payment areas. The logic is as follows:

MedPAC believes that Medicare payment policy should be neutral as to whether beneficiaries enroll in traditional Medicare or in M+C plans…Because health care is delivered in local markets, payment neutrality needs to be pursued at the local level. Failure to make payments equal within a local market would give one sector-either M+C or traditional FFS-an advantage over the other. Distortions in local markets could thus have the effect of limiting choice for Medicare beneficiaries. If payments are higher in one sector than the other, beneficiaries will move to the higher-payment sector if higher payment is successfully translated into a higher value product. Based on this policy rationale, MedPAC recommends that Congress should make Medicare payments for beneficiaries in the two sectors of a local market substantially equal, after accounting for risk.³⁴

This recommendation assumes a passive role for health plans in their ability to form networks and manage utilization. In fact, plans have inherent advantages in managing costs in high-cost Medicare FFS areas.³⁵ It is in those areas that excess capacity, such as hospital beds and specialist physicians, should give plans leverage to negotiate lower prices with providers and that inefficient practice patterns should permit plans to reduce excess utilization.³⁶ In fact, payments to M+C plans appear to be unrelated to plans' costs of providing the entitlement benefit package, as evidenced by the fact that the free supplementary benefits offered by M+C plans vary proportionately with their payment levels.³⁷ MedPAC itself found that plans' per capita costs rise by much less than one dollar for each dollar increase in per capita payments. One of the offered explanations was that plans serving counties with relatively high per capita costs in traditional Medicare may have greater opportunity to reduce their costs below traditional Medicare costs than do plans serving counties with relatively low traditional Medicare costs.³⁸

The converse is true also; in areas with low FFS payment there is often an undersupply of physicians and a single hospital, which allows a health plan little leverage in contract negotiations. On the service side, these areas may have pent-up demand for services because of access problems under FFS for which the HMO then becomes responsible.

Although MedPAC's "level playing field" argument has a certain logic when determining a government contribution amount pegged off of local FFS spending in a premium-support approach to Medicare restructuring, there is no need to abandon the goal of reducing payment disparities across counties and regions in Medicare's current defined-benefit structure. A prudent purchaser should not need to pay plans in high-cost areas the going FFS rate to obtain broad plan participation and extra benefits.

Risk-adjusted payments. There is broad policy agreement that risk adjustment based on health status should be a primary component of the payment system for M+C plans, whether in administered or competitive pricing, as well as in any Medicare restructuring that involves competition among health plans, such as under premium support. Indeed, despite its concerns about the CMS reliance on risk adjustment requiring plans to provide extensive encounter data, Congress, in BIPA, was willing to legislate a detailed implementation schedule that would fully implement risk adjustment by 2007.

Nevertheless, health plans, providers, and the CMS all face major burdens in complying with the requirements of an encounter-based risk adjustment system, based on multiple sites of service, including physicians' offices and outpatient hospital departments. Because of health plans' concerns, CMS administrator Scully suspended collection of ambulatory encounter data for a year and recently said that implementation of risk adjustment will remain on hold until the administration adopts a less burdensome, but still accurate, risk adjustment approach.³⁹ As a result, it is doubtful that the CMS will meet the statutory requirement to have a comprehensive, encounter-based risk adjustment system in place in 2004.

Another risk adjustment issue of particular interest to plans complaining of inadequate payment is whether risk-adjusted payments were designed in the BBA to potentially reduce payment to plans. Although the CBO initially did not score savings from the BBA's introduction of risk-adjusted payments to M+C plans, it believed that "adding a health-status adjuster could further reduce capitation rates relative to per capita fee-for-service costs."⁴⁰ In 1999 the CBO contradicted its earlier view, indicating that it had previously assumed that risk adjustment would be implemented on a redistributive but spending-neutral basis. As rationale for this position, it argued that the BBA-imposed reduction in M+C payment rates relative to rates in the traditional program was about the amount of the expected overpayment attributed to risk adjustment.⁴¹

In relation to implementing risk adjustment, it might be appropriate at this time to reconsider the 5 percent reduction off of applicable FFS rates that determine actual M+C payments. The 5 percent now provides the government with some cushion for the absence of health status-based risk-adjusted payments. Consistent with a view that M+C plans should be competing with the traditional FFS program on a level playing field, the 5 percent discount could be phased out, in conjunction with phasing in the comprehensive, health status-based risk adjustment system.

Available Policy Choices

In all likelihood, the expected congressional fight on what to do about M+C will take place on familiar terrain. One side will argue that the government already is overpaying plans in aggregate and plans in high-payment areas in particular. Additional benefits, such as prescription drugs, should be generally available in the program and not a function of where a beneficiary happens to live. To the extent that the BBA payment formula needs modification at all, it would be to modify the budget-neutrality provision to move more money to blend counties, away from high-payment counties.

The other side will argue that providing beneficiaries with a choice of private insurance plans is a fundamental goal of the M+C program and that maintaining an infrastructure of competing private plans is necessary as a platform for premium-support restructuring. Paying more in the short term is necessary to create the structure for price competition discipline that will reduce costs in the longer term. This side will suggest a series of BBA "fixes," including payment equality with local FFS rates, removal or phasing out of the TEFRA 5 percent discount, recalculation upward of the national rate in the blend formula to include only counties where there are M+C enrollees, and whatever else can be modified to increase spending flow to current and potential M+C plans.

With the two sides holding fundamentally different viewpoints on the performance of FFS Medicare, the desirability of premium-support restructuring, and whether prescription drugs should become an entitled benefit, the fight over M+C payment is an important one, but the likely legislative compromise is likely to satisfy neither side. Accordingly, I suggest two other M+C payment models that would shift the debate away from the BBA payment formulation and offer alternatives to the current legislative stalemate.

Severing The Payment Linkage To FFS

The new MedPAC recommendation calling for payment equivalency at the local level has one practical virtue: Payment equivalency is a straightforward formulation and more easily determined-it is simply data driven-than one that attempts to disaggregate the factors over which health plans presumably have little direct control (enrollee health status and input prices) and those they are supposed to strongly influence (utilization patterns). Indeed, the difficulty over getting the administrative pricing "right" for any particular area provides one of the main rationales for letting plans determine their own prices based on their knowledge of their own costs of care, by way of competitive bidding.

But perhaps the formal linkage between spending for M+C plans and traditional Medicare should be removed rather than reinforced. In the early to mid-1990s, when the FFS program was experiencing nearly double-digit spending increases, contracting risk HMOs were paid far more than their own medical cost trends. Conversely, after the BBA the FFS program experienced a dramatic drop in the rate of spending, giving health plans too small a government contribution, at least in relation to their own cost increases.

Furthermore, there is evidence that FFS Medicare's and M+C plans' ability to manage costs do not always move in tandem because of cost shifting. Although economists maintain that cost shifting theoretically should not occur, the experience in recent years suggests that the phenomenon is real, at least for hospitals.⁴² In pre-BBA years, when hospital profit margins were high, hospitals could enter into relatively unfavorable contracts with managed care plans, with Medicare providing a financial cushion. For a brief period in the aftermath of the BBA, the combined impact of private and public payers' reimbursement reductions threatened hospital margins.⁴³ But now, hospitals are renegotiating for higher payment rates from HMOs and even dropping disadvantageous contracts.⁴⁴

It seems odd that M+C plan payment increases are not based on their own performance but rather are tied directly to the annual success or failure of an "any-willing-provider" government purchaser–FFS Medicare–that must meet social and political objectives and does not have the freedom to compete as a private insurer would. The FFS program is sometimes permitted to use its market clout but is not permitted to use effective managed care tools to control costs. Also, as noted earlier, health insurers, including HMOs, are subject to the insurance cycle. In contrast, FFS Medicare responds to the legislative cycle of alternating payment reductions and givebacks. Finally, payment increases to M+C plans that offer prescription drugs are artificially low because FFS program cost experience does not include inflationary drug costs, which M+C plans offer as supplemental benefits.

Until there is more comprehensive restructuring of the program, M+C might be viewed as another provider type, with payment updates based on its own characteristics and performance, much as Congress treats hospitals, physicians, or home health agencies. Of course, this approach would fly in the face of concern that payment in Medicare already is too "siloed," with payment rules and levels based too much on historical costs experienced by each unique provider type rather than on what is needed to provide services efficiently. Nevertheless, considering each provider type on its own terms is the basis of most Medicare payments today. The way M+C plans are paid stands out as the anomaly.

M+C plans would like to have their inflationary costs passed through in increased Medicare payments. However, current FFS payment models no longer compensate providers for utilization increases, based on the policy view that providers should be financially responsible for managing what they do have control over: volume and intensity of services. Although the payment update approaches for the various providers vary somewhat, they all typically pass through increases in input prices, often referred to as a "market basket" of input prices specific to the relevant health sector, and sometimes provide an inflation factor pegged off of a neutral, nonmedical measure, such as the Consumer Price Index (CPI) or gross domestic product (GDP).⁴⁵

It is important to note that the current administered price formula for M+C plans does pass through the increase in input prices that hospitals and physicians face-for example, wages for hospital personnel-at least in the calculation of the national rate used for blending. However, M+C plans are affected only indirectly by changes in input prices faced by providers. Instead, the plans face input price changes that result from contracting with providers, such as the price of a hospital day.⁴⁶ As implied in the previous discussion on cost shifting, the input prices faced by providers and those faced by plans may vary greatly.

Congress could periodically adjust the update factor as it does with other providers to balance competing goals of cost containment and access; the factor should be calibrated based on comparison to the traditional Medicare program cost trends. Although this process would continue some degree of payment level uncertainty, at least the payment updates would take into account factors relevant to the business of managed care.

Paying M+C plans like other providers would still require a decision on whether base payments should follow county-level payments for the traditional program. There is now at least some evidence about how much plan costs vary from plan payments and how underlying health status risk varies by county. Using these factors, one could attempt to achieve a political compromise, again phased in over time, on the level of base payments. An alternative approach would be to provide an annual percentage increase pegged off of the average national per capita cost. In this way, plans in high-payment areas would get the same dollar increase, but a lower percentage increase, as plans in low-payment areas.

Recognizing And Rewarding Quality Of Care

Although proponents and detractors of managed care cite different studies to buttress their arguments about the quality of care provided by HMOs, what is striking is that in a broad comparison of managed care versus FFS, the statistically significant differences reported are fairly small and far smaller than the interplan variations found comparing Health Plan Employer Data and Information Set (HEDIS) scores or the state-level variations in the FFS program that the CMS found in looking at twenty-four process-of-care measures of quality.⁴⁷ Recently, in an update of previous reviews, Robert Miller and Harold Luft again concluded that quality-of-care findings for HMO plans were roughly comparable to those for non-HMO plans for the years 1997-2000. Importantly, they found no difference to the overall pattern of findings when they divided observations into the three categories of cancer, heart, and other.⁴⁸

Thus, for the most part, quality in M+C plans mirrors quality in FFS Medicare: Neither is very good.⁴⁹ Although the BBA set out a series of quality-related requirements for M+C coordinated care plans, plans are complaining about the burden of carrying out straightforward quality improvement projects.⁵⁰ The CMS has used its authority to implement risk adjustment to make a tentative foray into this area, by increasing risk-adjusted payments for health plans that meet national performance measures for managing patients with congestive heart failure under the hospital-only encounter data system that is supposed to end in 2004.

A few large employers and purchasing coalitions have adopted value purchasing approaches to health plan contracting that explicitly recognize and reward those plans that perform measurably better than their competitors on quality of care. Notably, General Motors goes even further by varying the level of salaried employee contribution to premiums based on how plans perform on quality and on how they manage costs.⁵¹

In administering the M+C program, the CMS lacks virtually all of the tools of value purchasing. For example, FFS Medicare cannot engage in selective contracting to channel patients to higher-quality providers, cannot contract with disease management companies and care coordinators to complement what individual physicians are able to provide beneficiaries, and cannot bundle payments among various providers to foster integrated care.

Unfortunately, Medicare risk plans have not tried to distinguish themselves for quality. Without risk adjustment of payment based on health status, any plan that tried to do so would undoubtedly have suffered financially. Further, in the face of the managed care backlash, many plans are giving up even basic managed care approaches and are beginning to resemble indemnity insurance companies of the past.⁵²

It is difficult to find a compelling policy rationale for overpaying health plans only to do what the traditional Medicare program does reasonably well: administer benefits and pay claims efficiently. Rather, in the areas of quality improvement and chronic care coordination, the M+C program offers opportunities that the traditional Medicare program lacks. So far, the M+C debate has been mostly about whether M+C plans should get more government money, not about whether Medicare and beneficiaries are getting enough value from M+C plans for their current level of spending. It might make sense, however, to invest in what these plans potentially are better able to do than FFS Medicare is.

But such an investment would require a radically altered payment and regulatory regime for M+C plans. To become a value purchaser of M+C services, first and foremost, the CMS would expedite implementation of risk-adjusted payment to M+C plans. In addition, it would have new authority and discretion to pick winners and losers, such as paying M+C plans differentially based on quality measures, actively encouraging and even providing incentives to beneficiaries to select better-performing plans, or negotiating specific performance contracts with plans that are best able to manage chronic, high-cost illnesses and providing meaningful incentives for beneficiaries to use these plans. Needless to say, the legal, administrative, and most of all political barriers to this kind of restructuring are formidable. Health plans would have to see the value purchasing approach as in their own best interest, and Congress would have to grant the agency an unprecedented amount of autonomy to use administrative discretion in M+C contract terms and conditions.

There is a common view that the BBA's creation of the Medicare+Choice program is largely responsible for the current instability in the program. Indeed, Gold gave the M+C program a grade of "'D' if not an 'F.'"⁵³ However, finding fault with the BBA deflects attention from the stark reality that health plans no longer are managing costs successfully. Plans are experiencing difficulties that apply to all of their products in all of their markets. Only the manifestations are different. Given this experience, it seems time to reconsider the purpose of Medicare contracting with private plans. Instead of viewing plans as a primary vehicle for reducing program costs or providing additional benefits, it might be more appropriate, instead, to reward private plans that improve quality and help manage the care for Medicare beneficiaries with chronic diseases. Changing the focus of the program will require severing the current payment linkage between M+C plans and the traditional Medicare program and designing a new regulatory and payment regime for contracting private plans.

A first draft of this paper was supported by the Health Insurance Reform Project (HIRP) at the George Washington University, a program supported by the Robert Wood Johnson Foundation. The author thanks participants in a meeting sponsored by HIRP for their helpful comments, in particular, Bruce Bradley and Bryan Dowd. In addition, the author thanks Nancy-Ann Min DeParle, Parashar Patel, David Helms, Michael O'Grady, and an anonymous reviewer for their review and comments, and Anna Weinstein for support in preparing the paper.

NOTES

1. S. Christensen, "Medicare+Choice Provisions in the Balanced Budget Act of 1997," Health Affairs (July/Aug 1998): 224-231.
2. Federal Register 63, no. 123 (26 June 1998).
3. R. Pear, "Medicare Shift toward H.M.O.'s Is Planned," New York Times, 5 June 2001.
4. Saint Joseph's Health System, the New Mexico PSO, had actually planned to terminate its 2001 contract until given a second chance with passage of BIPA. The CMS summary data only include PPOs that are state-licensed as PPOs. There are two M+C plans that currently offer PPO products. Gary Bailey, director, Health Plan Benefits Group, Center for Beneficiary Choices, Centers for Medicare and Medicaid Services, personal communication, 27 September 2001.
5. M. Gold, "Medicare+Choice: An Interim Report Card," Health Affairs (July/Aug 2001): 120-138; and S. Okie, "More HMOs Quit Medicare Plan," Washington Post, 22 September 2001.
6. M. Gold, "Medicare+Choice."
7. Even with this advantage, HMOs are losing market share to PPOs. See J. Gabel et al., "Job-Based Health Insurance in 2001: Inflation Hits Double Digits, Managed Care Retreats," Health Affairs (Sep/Oct 2001): 180-186.
8. CMS, Medicare: A Profile (Washington: US Department of Health and Human Services, July 2000).
9. CMS Medicare Managed Care Contract (MMCC) Plans Monthly Summary Report, <www.hcfa.gov/stats/mmcc.htm> (13 September 2001).
10. Ibid.
11. CMS, Medicare: A Profile.
12. R. Feldman, D. Wholey, and J.B. Christianson, "Do Medicare HMOs Cost Shift?" Inquiry (Fall 1998): 315-331, did not find evidence of cost shifting from commercial premiums onto Medicare.
13. Physician Payment Review Commission, Annual Report to Congress, 1997 (Washington: PPRC, 1997).
14. R.S. Brown et al., The Medicare Risk Program for HMOs-Final Summary Report on Findings from the Evaluation, Final Report, Contract no. 500-88-0006, prepared for the CMS (Princeton, N.J.: Mathematica Policy Research, 1993); PPRC, Annual Report to Congress, 1996 (Washington: PPRC, 1996); and G. Riley et al., "Health Status of Medicare Enrollees in HMOs and Fee-for-Service in 1994," Health Care Financing Review 17, no. 4 (1996): 65-75.
15. To guarantee the minimal 2 percent update, budget-neutrality provisions were placed on the blended rate process. Because of budget-neutrality, so-called blended counties were funded in only one year, 2000, since passage of the BBA.
16. Gold, "Medicare+Choice."
17. CMS, MMCC Plans Monthly Summary Report. Of total M+C enrollees, 18,000 were in Sterling, the private FFS plan.
18. R. Coulam et al., "Studies to Support the Medicare Competitive Pricing Demonstration" (Draft report, Contract no. HCFA-500-92-0014, prepared for the CMS by Abt Associates and the University of Minnesota, 30 April 2001).
19. RS Brown and M.R. Gold, "What Drives Medicare Managed Care Growth?" Health Affairs (Nov/Dec 1999): 140-149.
20. A potentially important change representing a new approach beyond the BBA structure was the provision permitting plans to offer beneficiaries rebates off of their Part B premiums, effective in 2003. The rebate provision was consistent with approaches of both the Clinton administration's Medicare reform proposal and the Breaux-Frist legislation and represents a move toward introducing price competition within the M+C program.
21. CMS rate announcement, March 2001, <www.HCFA.gov/stats/hmorates/ aapcc.htm>(13 September 2001).
22. PPRC, Annual Report to Congress, 1997.
23. Gold, "Medicare+Choice."
24. CMS rate announcement.
25. K. Ignagni, "Putting Principles First: A Better Way to Carry Out a Demonstration," Health Affairs (Sep/Oct 2000): 44-48.
26. Office of Inspector General, Department of Health and Human Services, Adequacy of Medicare's Managed Care Payments after the Balanced Budget Act, Pub. no. A-14-00-00212 (Washington: DHHS, September 2000).
27. Medicare Payment Advisory Commission, "Reconciling Medicare+Choice Payments and Fee-for-Service Spending," Report to the Congress: Medicare Payment Policy (Washington: MedPAC, March 2001), chap. 7.
28. GAO, Medicare+Choice: Payments Exceed Cost of Benefits in Fee-for-Service, Adding Billions to Spending, Pub. no. GAO/HEHS-00-161 (Washington: GAO, August 2000).
29. Ibid.
30. C. Hogan, P.B. Ginsburg, and J.R. Gabel, "Tracking Health Care Costs: Inflation Returns," Health Affairs (Nov/Dec 2000): 217-223.
31. Gabel et al., "Job-Based Health Insurance in 2001."
32. Ibid.
33. For retirees in M+C, employers may be picking up some or all of the premium increases.
34. MedPAC, Report to the Congress: Medicare Payment Policy.
35. N.M. DeParle and R.A. Berenson, "The Need for Demonstrations to Test New Ideas," Health Affairs (Sep/Oct 2000): 57-59.
36. J.E. Wennberg, The Dartmouth Atlas of Health Care in the United States (Chicago: American Hospital Publishing, 1998).
37. B. Dowd, R. Coulam, and R. Feldman, "A Tale of Four Cities: Medicare Reform and Competitive Pricing," Health Affairs (Sep/Oct 2000): 9-29.
38. MedPAC, Report to the Congress: Improving Risk Adjustment in Medicare (Washington: MedPAC, November 2000).
39. Medicine and Health, 17 September 2001.
40. S. Christensen, "Medicare+Choice Provisions in the Balanced Budget Act of 1997," CBO Staff Memorandum (Washington: Congressional Budget Office, 12 November 1997); and S. Christensen, "Medicare+Choice Provisions in the Balanced Budget Act of 1997," Health Affairs (July/Aug 1998): 224-231.
41. CBO, "Medicare Projections and the President's Medicare Proposals," in An Analysis of the President's Budgetary Proposals for Fiscal Year 2000 (Washington: CBO, April 1999), chap. 3. At the time of the BBA, the projected savings from health status-based risk adjustment were not known with precision. It turned out that the initial risk adjuster based on encounters from inpatient hospital stays only was projected to save about 6 percent, but projected savings from comprehensive risk adjustment will likely be much more than the 6 percent equivalence that CBO pointed to a year and a half after the BBA was passed.
42. Feldman et al., "Do Medicare HMOs Cost Shift?"
43. MedPAC, Report to Congress: Selected Medicare Issues (Washington: MedPAC, June 2000).
44. M. Freudenheim, "Medical Costs Surge as Hospitals Force Insurers to Raise Payments," New York Times, 25 May 2001; and R. Cunningham, "Hospital Finance: Signs of 'Pushback' amid Resurgent Cost Pressures," Health Affairs (Mar/Apr 2001): 233-240.
45. MedPAC, Report to Congress: Medicare Payment Policy (Washington: MedPAC, March 1999).
46. Currently, there is no verified database of prices for hospital stays or physician fees for which health plans contract. Surveys can form the basis for making such determinations. But the technical difficulties in separating out input prices from utilization factors in managed care contracting with providers are likely to be formidable.
47. CMS, "Medicare Health Plan Compare," 2000,<www.medicare.gov/mph Compare/home.asp> (13 September 2001); and S.F. Jencks et al., "Quality of Medical Care Delivered to Medicare Beneficiaries: A Profile at State and National Levels," Journal of the American Medical Association 284, no. 13 (2000): 1670-1676.
48. R.H. Miller and H.S. Luft, "HMO Plan Performance Update: An Analysis of Recently Published Literature (1997-2000)" (Report prepared for the Council on the Economic Impact of Health System Change, Eighth Princeton Conference: The Future of Managed Care, Princeton, New Jersey, May 2001).
49. M.A. Schuster, E.A. McGlynn, and R.H. Brook, "How Good Is the Quality of Health Care in the United States?" Milbank Quarterly 76, no. 4 (1998): 517-563.
50. In the BBRA Congress reduced quality-related requirements applicable to PPOs.
51. P. Salber, "The General Motors Approach to Value Purchasing" (Presentation at National Health Care Purchasing Institute Advanced Purchasing Workshop: Advancing Health Plan Accountability and Quality: An Introduction to the NBCH V8 Common RFI Tool, Washington, D.C., 21-22 June 2001).
52. W.A. Zelman and R.A. Berenson, The Managed Care Blues and How to Cure Them (Washington: Georgetown University Press, 1998.)
53. Gold, "Medicare+Choice."

Bob Berenson is senior adviser at the Academy for Health Services Research and Health Policy and adjunct professor at the University of North Carolina School of Public Health and the Duke University Fuqua School of Business.

2001 Project HOPE–The People-to-People Health Foundation, Inc.