Issues in Managed Care

Abstract

This issue of the Health Care Financing Review includes a collection of articles on managed care topics in the public programs of Medicaid and Medicare. The first article looks at developments in Medicaid in the broader context of overall changes in the managed care marketplace, while the remaining articles examine specific topics that have significant policy implications for the public sector programs.

Managed Care in Medicaid

Leading off this issue, Hurley and Draper provide an overview of the current direction of managed care in the commercial sector and the implications for Medicaid of the changes that have taken place. As the authors point out, Medicaid was an active participant in the “managed care revolution”: more than one-half of Medicaid beneficiaries are in managed care arrangements. The authors note that this level of participation partly reflects what might be termed “managed-care-friendly” features that are particular to Medicaid—such as the reality of limited provider access even in fee-for-service (FFS) Medicaid, and the ability to mandate enrollment. Medicaid and managed care have also been a felicitous pairing because of similar benefit structures—or what were once similar benefit structures—consisting of a comprehensive benefit package and limited cost sharing.

Though the authors state that there has not been turmoil in Medicaid managed care on a par with the turmoil in the commercial sector and in Medicare, there have been certain changes in Medicaid. Managed care continues to predominate as the preferred approach, but States increasingly have had to rely on Medicaid-only plans. Commercial plans have backed away from participation in Medicaid after their involvement peaked in 1998. Although there continue to be some commercial insurers with a presence in Medicaid, and there are what the authors call “investor-owned Medicaid focused plans,” market factors do not point towards a resurgence of interest in Medicaid contracting among commercial health maintenance organizations (HMOs).

As described by Hurley and Draper, the commercial managed care marketplace and Medicaid managed care appear to be heading in divergent directions—or perhaps it would be better to say that the Medicaid model has remained static while major changes have occurred in the private sector, changes which appear incompatible with the Medicaid managed care model. For example, in the commercial sector, in response to consumer preferences, health plans are offering greater choice of providers through larger panels, or through the use of point-of-service or preferred provider organization products. This has led to less utilization management, and a diminished ability to obtain favorable contracts from providers. Providers, meanwhile, were already “pushing back” in their contract negotiations and becoming reluctant to take on risk. These changes, combined with the current emphasis on health plan profitability rather than membership growth, have led to rising costs for purchasers and consumers. Enrollees are being asked to pay a greater share of costs out of pocket, with a potential movement towards a defined contribution approach of “consumer-directed products.”

The combination of factors that are changing the face of managed care in the commercial sector may produce the result that, as the authors put it, “Medicaid could actually become the last bastion of the pure HMO model.” The changes in the commercial sector do not necessarily spell trouble ahead for Medicaid, but they do signal the need for Medicaid to continue to find its own way. That is, Medicaid should continue on what appears to be its current path: continued development of improved, stable provider-sponsored plans; use of the primary care case management model as a structure for adding managed care features; and continued use of alternatives to full-risk contracting, such as partial risk arrangements and carve-outs.

In addressing the issue of whether Medicaid-only health plans can adequately serve all types of Medicaid beneficiaries, Hurley and Draper comment that “efforts to enroll beneficiaries with more extensive needs, including the severely disabled and chronically ill, may be particularly challenging for these plans because of the added resources and specialized providers needed to serve them. If they can demonstrate that they are delivering good and improving outcomes, this will offset concerns that plans for the poor are at risk of being poor quality health plans.” Having said that, how can we know whether plans are “delivering good and improving outcomes” for the most vulnerable beneficiaries? This is a question that Ireys, Thornton, and McKay address in their article—a question that to date has not been adequately answered.

Ireys et al. point out how little we know about enrollees with disabilities and chronic illnesses in Medicaid plans and whether or not their needs are being appropriately met. As the authors observe, ensuring appropriate care for individuals with disabilities or chronic conditions is “challenging under FFS or managed care arrangements.” They catalogue the “basic elements of a system of care” appropriate for a vulnerable population, explain the many ways in which individuals with disabilities and chronic care needs are different from other populations, and emphasize the importance given to a person-centered approach that allows these individuals to “shape and carry out their own care” through consumer choice and empowerment. They note that States have taken a variety of steps to attempt to improve care for this vulnerable population—for example, through standards included in managed care contracts and by working with the population and their advocates. Managed care plans have also gained greater experience in dealing with this population, and there is a greater awareness of the issue of how to identify populations with special needs and how to address their needs (e.g., through disease management programs). Another factor is the provider community and whether providers can “adopt new behaviors” addressing the complex needs of this population.

Ireys et al. find that there is very little research that has been done that can demonstrate progress made in treating this population, or which shows differences between health care systems (managed care versus FFS). The authors' enumeration of “research design challenges” has a familiar ring: the challenges include the continuously changing nature of managed care (making some research findings “irrelevant” after a system has changed), the question of what standard is used to evaluate managed care (is FFS the “gold standard” or would the standard be “an ideal managed care system”?), and the extreme diversity in this population and the diversity across State programs. They propose a coordinated national research agenda involving all the various stakeholders that will produce results that “point all stakeholder groups towards a more effective system of care” and which “maximize the link” between better outcome measures and the use of those measures by States and plans in improving performance. Given the complexities outlined in their article, improving care for this population “demands a research effort that is conceptually and operationally broad.” They suggest six areas of concentration for the research agenda and identify an “urgent need for adequate data” as an issue “transcending all of these areas.”

The authors emphasize the importance of adjusting payments to health plans to account for the “specific care needs of an enrolled population” to ensure that plans are willing to enroll particular subgroups and retain them as enrollees (as opposed to encouraging disenrollment). The article by Adams, Bronstein, and Raskin-Hood illustrates just how difficult it is to arrive at appropriate risk-adjusted payments in Medicaid.

Adams et al. examine the predictive accuracy of adjusted clinical groups (ACGs) in three States that are very different in the make-up of their Medicaid population. The ACG risk-adjustment system places individuals in 52 different categories based on health status, age, and sex. The system is “built on a hierarchy of all problems for which an individual seeks care.” A number of States use the ACG system as the basis for payment to Medicaid managed care plans. The authors use commercial and Medicaid data from three different States: two with little managed care penetration, Mississippi and Georgia, and California, which is at the other end of the penetration scale among both commercial and Medicaid populations.

The authors find essentially that ACG risk adjustment in almost all cases does offer greater predictive power than adjustments based only on age and sex for non-random groups. However, “as with most policies, the adaptation of risk adjustment must be State-specific.” This conclusion is based on the authors' having found a high degree of variation in predictive ability across the States and among different subpopulations for both age/sex and age/sex/ACG risk adjustment, resulting in potential underpayments or overpayments. The authors are particularly concerned about short-term enrollees and turnover. The expenditures for short-term enrollees (enrolled 6 months or less) are higher than for longer-term enrollees in all three States. In Mississippi, where nearly one-half of the welfare/poverty Medicaid enrollees are short-term enrollees, the per-member per-month expenditure figure that the authors compute for short-term enrollees is three times that of longer-term enrollees. When short-term enrollees are included in tests of predictive accuracy for the risk adjusters, the result is levels of error that are “markedly higher.” The authors examine different approaches to account for short-term enrollees in the risk-adjustment methodology, but none of the variations are viable for Mississippi.

The effect of short-term enrollees would again argue for an alternative, State-specific approach, such as the possibilities that the authors discuss, including the possibility of employing a concurrent/retrospective adjustment system, or the need to make additional adjustments to capitation payments, as Washington State does for new enrollees and as Maryland does for urban enrollees.

Managed Care for the Dually Eligible Beneficiary

One would think that a beneficiary who has both Medicare coverage and “full” Medicaid coverage is an ideal candidate to have all of his or her care needs coordinated in a managed care system. Such a person has a complete, comprehensive benefit package—including drug coverage and long-term care—that is essentially fully financed on a capitation basis by a combination of two payers. However, Walsh and Clark show that the coordination is rarely straightforward.

Of the 6 million dually eligible beneficiaries (a count that includes beneficiaries eligible only for premium or cost-sharing assistance), a small number are enrolled in health plans set up to coordinate Medicare and Medicaid coverage (e.g., the Program for All-Inclusive Care for the Elderly). About 5 percent of dually eligible beneficiaries are enrolled in “traditional” Medicare+Choice (M+C) plans. States (sometimes on a county-by-county basis) establish the rules as to whether dually eligible beneficiaries may enroll in a Medicaid plan, an M+C plan, or both. Enrollment rules run the full gamut of possibilities: depending on the State, beneficiaries can/cannot be in two plans; beneficiaries can only be in two plans if they are related/unrelated plans; etc.

There are often valid reasons that explain why the coordination process is not entirely straightforward. For example, Medicaid is the payer of last resort, meaning that Medicare coverage is primary for dually eligible beneficiaries. To the extent that extra benefits are paid for by Medicare through a capitation to an M+C organization (e.g., some level of outpatient prescription drug coverage may be part of a basic M+C plan), a State should have no financial obligation for what otherwise would have been a Medicaid benefit. As the authors point out, coordination of benefits in such circumstances can be quite complicated, if, for example, the M+C drug benefit is limited in terms of dollar coverage, has a restricted formulary, or involves copayments that Medicaid beneficiaries should not have to pay.

Walsh and Clark point out that there are significant information gaps for beneficiaries as well as for health plans (both in terms of absence of data and lags in obtaining information), and that certain rules act as impediments to good information dissemination. Similarly, effective case management is hampered by not knowing the status of enrollees and by the difficulties of coordinating benefits across programs. The authors recommend that there be better and more timely information dissemination to beneficiaries as well as to health plans (e.g., on-line access to eligibility information). Beneficiaries should understand their financial liability and should receive information specifically prepared for dually eligible beneficiaries. They also suggest that CMS and the States develop innovative approaches to address problems, following the lead of California in its decision to pay M+C premiums for Medicaid enrollees after M+C organizations in much of the State stopped offering zero premium plans. Finally, Walsh and Clark advise against the practice of having beneficiaries enrolled in a Medicaid managed care plan combined with FFS Medicare, and advise against the practice of having beneficiaries enrolled in two unrelated health plans.

Managed Care in Medicare

From a certain point of view, the managed care revolution never gained a firm grip on Medicare. While in the commercial sector, among employer-sponsored individuals with health care coverage, a mere 5 percent are in non-managed care arrangements (Kaiser Family Foundation/Health Research and Educational Trust, 2002), Medicare continues to be primarily a FFS program, with only about 15 percent of beneficiaries enrolled in managed care. This is not to say that the recent history of Medicare managed care has not had its share of dramatic changes in a relatively short period of time. Enrollment in Medicare health plans burgeoned in the late 1990s, followed not much later by a steep decline in the number of participating plans and recent declines in the total number of enrollees.

The decline in plan participation in M+C has been attributed to the payment rate changes of the Balanced Budget Act (BBA) of 1997, which provided for increases in payments to historically lower-paid counties while providing only minimum payment increases in areas with higher payment rates (which generally were the counties with the highest Medicare HMO enrollment). The decline in the generosity of M+C benefit packages since 1999, and the reduced availability of M+C plans that do not charge enrollees a premium, have also been attributed to the payment changes. However, factors other than payment have also played a role in the ups and downs of M+C since the BBA, according to the U.S. General Accounting Office (1999, 2000).

Congress has modified the BBA M+C payment rules in subsequent legislation enacted since the passage of the act. The Benefits Improvement and Protection Act of 2000 (BIPA) provided for an increase in the BBA floor payment, an add-on to the minimum payment increase, and the introduction of a new floor payment in counties within metropolitan statistical areas with populations exceeding 250,000 residents. The last provision in particular resulted in substantial payment increases in many counties with M+C plans. BIPA was not passed until December 2000, with new payments effective March 2001. Because of the availability of additional funds that could be used to change premiums and benefits (or could be passed on to providers, or retained by the Federal Government for the health plans to use at a later date), plans were required to resubmit benefit and premium proposals to replace those already approved by CMS.

Pizer and Frakt take advantage of the unique situation presented by the BIPA payment increases to evaluate the effect of payment and other factors—in particular, competition among plans—on plan behavior with respect to setting premiums and designing benefit packages. They point out that the BIPA situation permits them to overcome the problem of health plan costs' being unknown or unobservable for researchers looking at pricing behavior. They reason that there should not have been any significant change in costs in the short period of time between the BIPA payment increases and the ratesetting decisions health plans made prior to BIPA, which were to be effective on January 1, 2001. That is, cost would not have been a major factor in any differences in premium and benefits proposed for January 1, 2001, as compared with those proposed a short time later, to be effective March 1, 2001.

The authors found that higher payment rates are in fact associated with lower premiums and higher levels of extra benefits. Other factors associated with lower premiums include the level of intensity in the use of hospital care in an area, and even a variable representing differences between January and March, controlled for changes in payment rates—which the authors interpret as an indicator that “plans may have judged the post-BIPA climate to be more promising, leading to renewed efforts to attract enrollees” through reduced premiums and higher benefits. The finding that the authors emphasize, however, is that for premiums “the intensity of competition appears to have had stronger effects than changes in payment rates.”

As the authors comment, their findings “have a series of critical policy implications.” That is, their findings indicate that “it is possible for the Medicare Program to increase benefits (in M+C) without increasing spending or shifting additional costs to beneficiaries.” This would be done by promoting competition, which the authors note as being possible through the competitive pricing approaches of various Medicare reform proposals or the yet-to-be-implemented demonstration of competitive pricing in Medicare. For now, however, Pizer and Frakt comment that the more likely scenario that may induce greater competition is the new statutory provision (enacted in BIPA, coincidentally, but effective for 2003) that allows M+C plans to provide rebates on the Part B premium of their enrollees (up to 100 percent of the premium).

Returning to the issue of plan withdrawals, one of the greatest concerns has been the effect of M+C withdrawals on beneficiaries. Between 1998 and 2002, about 2.3 million beneficiaries have been affected by M+C plan termination. In some cases, beneficiaries have been affected more than once across the 5-year period, as multiple health plans have left particular areas. Rural areas have been disproportionately affected by withdrawals. Considering that total M+C enrollment has never reached the 7 million mark, actions that affect over 2 million beneficiaries are a matter of great concern.

In the case of a plan withdrawal, M+C enrollees are able return to FFS Medicare (with certain rights to Medigap coverage), and some affected enrollees may be able to join a remaining M+C plan in their area. The article by Booske, Lynch, and Riley documents the “short-term anxiety, higher costs, and/or disruptions in health care,” when M+C withdrawals occur, and the particularly disruptive (and, in some cases, adverse) effect on vulnerable populations. Their findings are based on a survey of nearly 3,400 beneficiaries between March and June 2001 (i.e., among those affected by non-renewal activity at the end of 2000, when 327,000 beneficiaries were affected by a non-renewal). As a result of plan withdrawals, a majority of beneficiaries reported being concerned about getting care and were concerned about their ability to pay for health care. Vulnerable subgroups were disproportionately represented among those with such concerns. One in five beneficiaries reported losing prescription drug coverage, and 15 percent of beneficiaries reported not obtaining medicines that had been prescribed. One in five beneficiaries reported that they delayed seeking needed care because they were concerned about the cost.

The authors identify information dissemination and comprehension as especially troublesome issues for affected enrollees. Even though 96 percent of enrollees (a rate much higher than an earlier study) recall receiving a letter from their health plan regarding an impending withdrawal, 40 percent of beneficiaries stated that they had insufficient information about their coverage options. Vulnerable groups (such as the very aged and the disabled), and black people, were less likely to report that they had received adequate information. As an indication of the confusion over what would happen with their health care, more than one-quarter of beneficiaries “thought they would end up with no health insurance.”

The high level of confusion appears to have had consequences that may have resulted in beneficiaries not making optimum choices. The authors pointed out that for many beneficiaries enrollment in an available M+C plan continues to be an attractive option when compared with FFS Medicare. However, they found that even though 80 percent of affected beneficiaries lived in an area with a surviving M+C plan, only 52 percent of surveyed enrollees report that they enrolled in another HMO (which, the authors comment, provides an overestimate of Medicare HMO enrollment when compared with CMS administrative records). Vulnerable subgroups were less likely to be aware of supplemental health insurance options that might be available. Individuals who reported having enough information about their status were more likely to obtain supplemental coverage.

Booske et al. conclude with a discussion of how to improve beneficiary understanding of the consequences of a plan withdrawal. They point out that the issue may not be that beneficiaries lacked information, but rather that they had too much information or were confused by the information. Particularly for vulnerable subgroups, they suggest that information about the effect of an M+C withdrawal “should be tailored to meet the specific needs and interests of those subgroups.”

How to convey information to Medicare beneficiaries is an issue that Harris-Kojetin, Jaël, Smith, Kosiak, and Brown deal with in their article on the value of reporting voluntary M+C disenrollment rates and the reasons for voluntary disenrollment. Such reporting was required by the BBA, and CMS began reporting disenrollment rates in 2000, followed by the reporting of reasons for disenrollment in 2002. The authors describe the results of focus groups and one-on-one interviews testing informational material over a 2-year period. The intent was to produce material that would “enhance user understanding, usability, trust, and interest in using the materials in plan choice.”

They found that beneficiaries had difficulties in a number of areas. The results, as they state, are “somewhat sobering,” but they help CMS determine the direction of education efforts to encourage beneficiaries to use data and reports on health plan quality in making choices. Among other findings, they report that the very terms they initially used in the material were problematic unless they were explained or illustrated. For example, the term “disenrollment” and even the term “average” were not immediately understood by all beneficiaries.

One point to make is that beneficiaries were concerned about the source of information on disenrollment rates and reasons. Beneficiaries trusted the information more when they were told that the Federal Government, rather than health plans, provided the information. They conclude by suggesting that, because beneficiaries turn to this kind of information when it is needed (e.g., only when they are about to enroll in an HMO), the information should be targeted to particular beneficiaries or should be provided in particular settings rather than having “broad-based dissemination strategies currently mandated by Congress.”

On the question of rates of voluntary disenrollment, Cox, Lanyi, and Strabic examine whether the design of M+C benefit packages influences enrollee ratings of plans and influences voluntary disenrollment decisions. Specifically, do the types of extra benefits offered by M+C plans have such an influence, and which benefits have the greatest effect on ratings and voluntary disenrollment rates? Cox et al. note that past studies indicate a possible connection between benefits offered and satisfaction levels (based on the drug coverage offered) and between benefits—particularly premium levels—and rates of disenrollment. Establishing a direct effect attributable to the benefit package is not entirely straightforward. The authors point out that plan ratings and disenrollment rates vary by subpopulations, and they therefore examine the data at the overall level and at the subpopulation level. Access to particular benefits also varies by subgroup, partly because some subgroups are concentrated in particular market areas.

With regard to benefits and their influence on plan ratings, the authors found “few statistically significant relationships.” Vision services were associated with higher overall plan ratings, while drug coverage or hearing services were associated with lower ratings; and several specific elements of plan ratings (as opposed to overall ratings) were affected by cost-sharing levels. With regard to voluntary disenrollment rates, the authors found that market competition had the greatest impact. Organizations that are the sole health plan in their area have lower disenrollment rates compared with organizations operating in multi-plan areas. The effects of extra benefits were “mixed,” with dental benefits and health education significantly associated with lower disenrollment rates. Vision and drug benefits were associated with higher rates of disenrollment. The authors explain this finding as possibly related to the nature of the drug and vision benefits—benefits which can have limits applied, which, when exhausted, induce “benefit shopping” (enrollees leaving one health plan to join another after exhaustion of a drug coverage limit, for example).

One problem that Cox et al. faced was that they were unable to determine which options beneficiaries chose when they were offered multiple benefit packages by their M+C organization. They assumed that enrollees would be “economically rational purchasers and select a package with the greatest ‘value’.” The survey that Hileman, Moroz, Wrightson, and Kim report on in this issue fills in the information gap by providing enrollment counts for individuals who can choose among various options offered by their M+C plans.

M+C has generally been viewed as a type of individual insurance market in which beneficiaries select a plan of their choice, choosing among competing M+C plans, and deciding between M+C enrollment and FFS Medicare with or without supplemental (Medigap) coverage. Of course, different Medicare subpopulations have to consider various factors that affect their choices or factors that constrain them from having choices that are available to other beneficiaries. The dually eligible population, as previously discussed, is one such group. Another group is those beneficiaries who have supplemental coverage available to them through a retiree health benefits plan sponsored by an employer or union.

Some employers and unions use M+C as the vehicle for retiree health care coverage. Until now, the number of M+C enrollees who have retiree coverage through M+C plans was not known, though it was clear that some health plans had high numbers of retirees among their M+C enrollees. The article by Hileman et al. presents the results of a survey that Actuarial Research Corporation (ARC) completed under a contract with CMS. ARC obtained information for 2001 and 2002 from M+C contractors operating in 2002. Plans that left the program at the end of 2001 were not surveyed. The survey had a very high response rate: 99 percent of M+C enrollment is represented in the findings.

ARC found that, across all M+C plans as of 2002, not quite one in five enrollees (18 percent) are group-sponsored retirees. However, there is significant variation across health plans and across geographic areas. Among the 74 percent of contractors with some group enrollment, the plan-level proportion ranged from less than 1 percent to more than 90 percent. Some small plans had very large group enrollment numbers, and the largest plans also had a relatively high proportion of group enrollment.

In what is not a surprise, the ARC survey found that group enrollees have more generous benefit packages than individual Medicare beneficiaries. The authors also examined benefit generosity for individual beneficiaries in relation to M+C payment rates and imputed FFS expenditure rates (county-level FFS rates computed from historical FFS rate data published by CMS, trended forward to 2001 and 2002). They found a relationship that seems counterintuitive, with areas that have M+C payment rates higher than “actual” FFS costs providing less generous benefits, and areas where M+C payments are lower than FFS costs providing more generous benefits. The authors explain this as a reflection of greater competition among those areas that have received only minimum updates in their payments since the 1997 BBA, while areas receiving higher payment increases under the BBA and post-BBA provisions are less likely to have competition among M+C plans because managed care has been “less successful” in gaining a foothold in such areas.

A bright note for the future of M+C is that Hileman et al. perceive that a “more competitive M+C environment could be developing,” with enrollment growth occurring in the smaller plans.

Conclusion

There are two themes that recur in the articles collected in this issue of the Review. One theme is competition among health plans, which several authors discuss directly. The other unifying theme is that almost all the articles, either directly or indirectly, have something to say about the individuals enrolled in managed care plans—their needs with respect to the type of care they receive, their information needs, their behavior as purchasers of health care, and how they have been affected by changes of the recent past.

To put what the articles say about competition into the frame of reference of enrollees: competition benefits enrollees. Medicare beneficiaries are more likely to have better choices (lower premiums, better benefits) offered to them by M+C plans when there is competition, and they can exercise choice when there is competition. And the greater the level of competition, the greater the likelihood that beneficiaries will exercise choice (as illustrated by the greater likelihood of disenrollment, as reported by Cox et al., and the finding by Booske et al. that enrollees affected by non-renewals are more likely to join another M+C plan if there are multiple plans in an area with high penetration). However, in order to exercise optimum choice and to make competition work effectively, beneficiaries need to have information that they find usable, as made clear by both Booske et al. and Harris-Kojetin et al.

Health plans can of course “manage” their competition. As both Pizer and Frakt, and Hileman et al., comment, it appears that there are signs that M+C plans are in search of new enrollment and are therefore offering more “competitive” benefit packages. Health plans can also design their benefit packages, and their contracting and marketing strategies, with a view towards selection, which does not serve the interests of beneficiaries or purchasers. As Adams et al. point out, appropriate payment rates are intended both to ensure that health plans will participate as contractors as well as to ensure that favorable selection is averted and unfavorable selection is adequately compensated. If a goal of Medicare and Medicaid is to have managed care serve the needs of the vulnerable populations of these public programs, payment policy has to be consistent with that goal. Enrollment of these populations also needs to be facilitated through other means in addition to payment policy—by recognizing and understanding their special needs (Ireys et al.) and by streamlining administrative mechanisms (Walsh and Clark).