Financial Performance and Participation in Medicaid and Medi-Cal Managed Care

Abstract

This article assesses the participation and the financial performance of licensed health maintenance organizations (HMOs) in the Medicaid market. The study found that participation by Medicaid Dominant plans has more than doubled from 11 percent in 1992 to 23 percent in 1998 while Medicaid membership in Commercial Dominant plans declined from 71 percent in 1994 to 51 percent in 1998. Both participating and non-participating plans incurred operating losses in 1998. Medi-Cal participating plans had higher operating margins than Medicaid participating plans throughout the United States. Interviews with key informants express concern about competence in program management, rate adequacy, decline in Medicaid enrollment, and turbulence forces of managed care market on Medicaid programs.

Introduction

Medicaid managed care enrollment reached 17 million beneficiaries in 1999, more than 54 percent of the Medicaid eligible population (Health Care Financing Administration, 1999). Virtually every State has implemented some form of managed care in its Medicaid program. In recent years, Medicaid managed care initiatives have been buffeted by trends in the broader managed care market and a number of States have seen participation among plans decline (Felt-Lisk and Yang, 1997; McCue et al., 1999). Some States have been unable to sustain their programs as a result of plan departures. Because health plans experienced a distinct reversal of fortune across all business lines in the mid- and late 1990s, it is difficult to determine whether Medicaid participation is a cause or a consequence of declines in financial conditions of plans.

Like nearly everything else about Medicaid, the picture of experience with managed care is a mixed one across the more than 50 programs in States and territories. The life cycle of program implementation represents an additional source of variation with some States initiating programs relatively recently, while others have very mature programs. A similar life cycle might be observed among plans, with some plans still becoming familiar with the Medicaid market, while others are well experienced in this product line, and still others have tried and failed to achieve success in this business segment. A better understanding of these trends and their causes can enable States to make appropriate responses to market changes.

The popular press has taken note of these trends (Langreth, 1998; Meyer, 1997) and prior research has cast some light on patterns of participation (Felt-Lisk et al., 1999). An earlier study attempted to examine the question of whether or not Medicaid managed care is a sustainable line of business for commercial HMOs (Hurley and McCue, 1998; McCue et al., 1999). Analyzing plan and financial data during the period from 1992 to 1996, this article found that HMO participation in Medicaid grew at an annual rate of 22 percent. However, among HMOs with a high number of Medicaid enrollees, financial performance showed a downward trend.

Using a similar analytical strategy, this article updates the earlier evidence (McCue et al., 1999) of licensed HMO participation as well as the financial performance assessment of the Medicaid market. More importantly, this article differs from the previous work by synthesizing the findings from two recent research projects that were conducted in a parallel and coordinated fashion. One project focused on the Medicaid managed care experiences of eight selected States, which represent different levels of program maturity and success (Hurley et al., 2000a). The other project explored the experiences of California's three distinct Medicaid managed care models of Medi-Cal (Hurley et al., 2000b). Common features of the States in both analyses included that they have displayed a preference for full-risk models, and are experiencing general marketplace turbulence, shrinking eligible populations, and a varied mix of participating plans. The California experience is particularly instructive in that it represents the most mature commercial managed care market in the Nation, relies on three distinct strategies for promoting managed care enrollment, and the State has witnessed few plan withdrawals from Medi-Cal (Hurley et al., 2000a, b). The findings have important implications for the long-term viability of Medicaid managed care.

Medicaid Managed Care Approaches

National

The varied approaches that States have employed in Medicaid managed care have been well detailed (Hurley and Somers, 2000; Holahan et al., 1998)). At the risk of oversimplification, they fall into two broad categories: full-risk contracting with prepaid organizations and primary care case management (PCCM) models that in nearly all instances are fee-for-service (FFS) paid program. Approximately 70 percent of all beneficiaries in managed care are in full-risk arrangements where State Medicaid programs contract with multiple competing prepaid health plans that are usually HMOs, or some variant of an HMO. Federal Medicaid regulations require that at least two plans be available to support beneficiary choice if a State wishes to make prepaid health plan enrollment mandatory. The participating plans may be organizations that have commercial and other product lines, or may serve Medicaid beneficiaries exclusively.

States use different methods for selecting plans (Hurley and McCue, 1998). In a number of States such as Virginia, Maryland, New Jersey, and Wisconsin, plans are selected based on their meeting prespecified qualifications and their willingness to accept administered rates. Other States use competitive bidding processes to make contract awards based on price and other considerations, such as Washington and Michigan. Still other States use a combination of bidding and negotiating of terms including payment rates, as in the case of Arizona. States may adopt multiple approaches within their States to reflect market structure and conditions as well as State agency and local preferences.

Most States have also made special accommodations for the protection of traditional high volume Medicaid and other so-called safety-net providers (Felt-Lisk, 2000). Such accommodations may include requirements for including these providers in health plan networks, preservation of special payment arrangements for these providers, and favorable consideration of plans that include these providers in terms of contract awards or allocation of enrollees. In some States, these providers sponsor their own health plans that enroll primarily or exclusively Medicaid members (Gray and Rowe, 2000; Brown and Sparer, 2000).

California

In California, where 48 percent of the 2.5 million beneficiaries in December 1998 were in managed care, there are three main contracting models that reflect deliberate State strategies to adopt and adapt managed care (Draper and Gold, 2000) to local conditions. The models include the Two Plan Model, the County Operated Health Systems (COHS), and the Geographic Managed Care (GMC). The Two Plan Model, found in 11 counties, includes two plans in each county selected by the State Medi-Cal agency. One of the plans is a local initiative developed and sponsored by local governments in conjunction with community-based organizations and is designed explicitly to contract with and to protect traditional safety net providers. The other plan is a commercial (mainstream) plan selected through a competitive bidding process for each county. Implementation of the Two Plan Model began in 1996 and currently 1,783,628 enrollees are in the plans in the 11 counties with 1,195,855 enrollees (67 percent) in local initiatives plans and 587,773 enrollees (33 percent) in commercial plans.

COHS is the second largest Medi-Cal managed care model and included 409,325 enrollees (17 percent) in the five systems in 1998. Although these systems vary quite substantially, structurally they are quasi-government organizations that contract with the State Medi-Cal agency to become risk-assuming intermediaries and negotiate capitation rates for most of the Medi-Cal beneficiaries residing in each county. Some of the models such as the Santa Barbara Regional Health Authority operate a distinct health plan for all the county's beneficiaries, while another model, CalOptima in Orange County operates both as a health plan and a general contractor that subcontracts with HMOs and non-HMO risk-bearing provider sponsored entities. This distinctive model of risk-based contracting with local authorities is found virtually only in California. (Draper and Gold, 2000).

The GMC Model had enrollment in 1998 of 309,867 in Sacramento and San Diego Counties. In these two counties, the Medi-Cal program contracts with multiple health plans that bid to participate and agree to accept the capitation rates set by the State agency. This approach is the one that is most similar to what is found in most other States that are operating full-risk Medicaid managed care programs. Notably, the counties with the COHS and with GMC Models do not have extensive public hospital systems.

Data and Methods

The first part of the data analysis examines, on a nationwide basis over a 7-year period, the number of HMOs, their Medicaid participation status, and their financial performance. Regarding Medicaid participation, HMOs are classified by their relative involvement, which is defined as the number of Medicaid enrollees as a percentage of a plan's total membership. Based on this percentage, each participating plan was assigned to one of three categories: Commercial Dominant (less than 26 percent Medicaid), Mixed (26-75 percent), or Medicaid Dominant (more than 75 percent). A financial analysis is conducted by these categories. These findings are presented in more detail by Felt-Lisk (2000) but comparison can easily be made to this recent work.

The second part of the data analysis presents a comparison of California HMOs and national financial data by Medicaid participation status. The third part of the data analysis provides a financial assessment of the different types of California models found in Medi-Cal managed care.

The financial assessment of HMO performance is based on three financial ratios: (1) operating margin, (2) medical loss, and (3) administrative cost ratio. These ratios are defined in accordance with the Health Plan Employer Data and Information Set (HEDIS^®) 3.0 indicators of financial stability. The operating margin ratio measures how much operating income an HMO earns from its insurance revenues, the medical loss ratio measures the proportion of insurance revenues paid out in medical claims, and the administrative cost ratio measures the proportion of insurance revenues paid out in administrative expenses.

With respect to the national data, this study examines the period from 1992-1998. The HMO database of Health Care Investment Analysts (HCIA) is the source of the financial and operating data used to explore the participation of the Medicaid market nationally from 1992-1998. For the California analysis, the study only analyzes the period from 1996-1998. Some California financial data were calculated from financial reports submitted to the California Department of Corporations, California Department of Health Services, and some from HCIA data. At the time of this article, these financial and operating data were the most recent data available.

Both the HCIA and California databases have several limitations. One limitation of the HCIA data is that the study population only includes State-licensed HMOs. Thus, these data underrepresent the entire population of fully capitated entities participating in Medicaid, particularly those in States with Section 1115¹ waivers where participation does not require an HMO license. Moreover, the findings of this article that are based on national data apply specifically to licensed HMOs. Another limitation of the HCIA data is that a small number of companies, whose HMOs are licensed in several States, have filed consolidated data in their States of operation. In the case of these plans this article used additional information from the American Association of Health Plans to classify and refine plan data for plans that filed consolidated data in various States of operation.

California data include both licensed and non-licensed HMOs. HMOs in California are licensed by the Department of Corporations under the Knox-Keene² legislation passed by California in 1975. COHS are not required to be licensed HMOs.

This article also gleans information about Medicaid participation from the interviews of multiple informants (Medicaid officials, health plan representatives, HMO and hospital associations, and advocacy groups). These informants were selected from nine States—Arizona, California, Maryland, New Jersey, Ohio, Texas, Virginia, Washington, and Wisconsin—representing various ranges of Medicaid experiences, market diversity, and plan withdrawals.

Participation and Financial Performance

Our analysis is presented in three stages. First, we update the national trends in Medicaid participation versus non-participation by plan characteristics and financial performance. Second, California HMOs and national financial data are compared with Medicaid participation and non-participation. Third, we present a financial analysis by the different types of California models.

National Trend Update

The number of plans participating in the Medicaid market escalated upward from 1992 to 1996 but leveled off in 1997 and 1998 (Table 1). However, the characteristics of plans participating in Medicaid managed care changed over this time period. The number of for-profit plans participating in Medicaid grew to 135 plans in 1997 from 43 plans in 1992. By 1998, however, the number of for-profits participating in Medicaid declined to 122 plans. The number of not-for-profit plans participating in Medicaid also expanded as well from 59 plans in 1992 to 102 plans in 1997 and exhibited a minimal decline to 100 by 1998. HMO participation in Medicaid based on Medicaid membership size as a percentage of total enrollment (Table 1) indicates that plans participating in Medicaid were primarily Commercial Dominant plans with small Medicaid enrollment throughout the study period. However, after 1994 a downward trend in Medicaid participation occurred within Commercial Dominant plans, falling from 71 percent of all plans in 1994 to 51 percent of all plans in 1998. An upward trend occurred for Medicaid Dominant plans, which more than doubled from 11 percent in 1992 to 23 percent in 1998.

Table 1. Health Plan Participation Status in Medicaid Managed Care, by Plan Charactertics: National Data 1992-1998.

Health Plan	Medicaid Participation
	1992		1993		1994		1995		1996		1997		1998
Characteristic	Yes	No	Yes	No	Yes	No	Yes	No	Yes	No	Yes	No	Yes	No
Total	102	369	116	355	150	345	188	350	225	349	237	442	226	429
Percent	21.7	78.3	24.6	65.4	30.3	69.7	34.9	65.1	39.2	60.8	34.9	65.1	34.5	65.5
For-Profit
Number	43	276	52	267	73	256	106	262	131	256	135	334	122	311
Percent	42.2	74.7	44.8	75.2	48.7	74.2	56.4	74.8	58.2	73.3	57.0	76.7	55.0	77.3
Not-for-Profit
Number	59	93	64	88	77	89	82	88	94	93	102	101	100	91
Percent	57.8	25.2	55.2	24.7	51.3	25.7	43.6	25.1	41.8	26.6	43.0	23.2	45.0	22.6
Medicaid Membership Commercial Dominant
Number	69		79		106		128		147		131		115
Percent	67.6		68.1		70.7		68		65.3		55.3		50.9
Mixed
Number	22		26		26		26		38		52		58
Percent	21.6		22.4		17.3		13.8		16.9		21.9		25.7
Dominant
Number	11		11		18		34		40		54		53
Percent	10.8		9.5		12		18.1		17.8		22.8		23.5

NOTE: Commercial Dominant is less than 26 percent of enrollees as a percentage of total plan membership; Mixed is 26-75 percent of Medicaid enrollees as a percentage of total plan membership; and Medicaid Dominant is greater than 75 percent of Medicaid enrollees as a percentage of total plan membership.

SOURCE: Health Care Investment Analysts: Data from health maintenance organization database, Baltimore, Maryland. Computations prepared by McCue et al.

The financial performance by size of Medicaid enrollment as a percentage of total enrollment (Table 2) shows higher operating margins in 1992 and 1993 for Medicaid Dominant plans. After 1994, operating margins for Medicaid Dominant plans declined considerably, from 0.01 in 1994 to -0.07 in 1997.

Table 2. Health Plan Financial Performance Measures, by Percentage of Membership in Medicaid: National Data 1992-1998.

Medicaid Membership Plan	1992	1993	1994	1995	1996	1997	1998
	Medical Loss Ratios
None	0.87	0.85	0.85	0.88	0.90	0.92	0.93
Commercial Dominant	0.90	0.88	0.86	0.88	0.91	0.93	0.93
Mixed	0.84	0.84	0.84	0.87	0.89	0.90	0.92
Medicaid Dominant	0.85	0.82	0.83	0.82	0.83	0.88	0.89
	Administrative Cost Ratios
None	0.13	0.13	0.15	0.15	0.16	0.17	0.16
Commercial Dominant	0.12	0.12	0.14	0.14	0.13	0.14	0.14
Mixed	0.14	0.16	0.13	0.16	0.18	0.16	0.16
Medicaid Dominant	0.18	0.19	0.17	0.19	0.22	0.19	0.18
	Operating Margins
None	0.01	0.02	0.01	-0.02	-0.08	-0.09	-0.09
Commercial Dominant	0.01	0.02	0.02	0.00	-0.02	-0.03	-0.04
Mixed	0.02	0.02	0.02	-0.02	-0.03	-0.05	-0.08
Medicaid Dominant	0.04	0.03	0.01	-0.04	-0.08	-0.07	-0.03

NOTE: Commercial Dominant is less than 26 percent of enrollees as a percentage of total plan membership; Mixed is 26-75 percent of Medicaid enrollees as a percentage of total plan membership; and Medicaid Dominant is greater than 75 percent of Medicaid enrollees as a percentage of total plan membership.

SOURCE: Health Care Investment Analysts: Data from health maintenance organization database, Baltimore, Maryland. Computations prepared by McCue et al.

Higher administrative costs were incurred each year by Medicaid Dominant plans. In 1996, their administrative cost ratio was 22 percent and declined to 18 percent in 1998. Conversely, Commercial Dominant plans had lower administrative costs each year ranging from 12 percent in 1992 to 14 percent in 1998.

Financial Performance: California Versus U.S.

Operating margins were consistently better for plans participating in Medicaid in California than plans participating in Medicaid in the U.S. between 1996 and 1998.³ Medi-Cal participating plans had break-even margins of 0.001 and 0.004, respectively for 1997 and 1998 compared with operating losses for Medicaid participating U.S. plans. In California, Medi-Cal participating plans had higher operating margin ratios for all 3 years than non-participating plans and the U.S. California non-participating plans incurred operating losses for all 3 years; these losses were lower than those for non-participating U.S. plans.

For 1997 and 1998, administrative cost ratios were consistently lower for plans participating in Medicaid in California than participating U.S plans. For participating Medi-Cal plans, the administrative cost ratio decreased to 0.112 in 1998 from 0.155 in 1996. For non-participating Medi-Cal plans, the administrative cost ratio was also lower than the U.S. non-participating plans. Medical loss ratios increased for Medicaid participating plans, Medi-Cal plans, and U.S. plans; however, U.S. plans experienced a greater increase than Medi-Cal plans. For Medi-Cal participating plans medical loss ratio increased from 0.836 in 1996 to 0.868 in 1998 while U.S. participating plans, medical loss ratio increased from 0.894 in 1996 to 0.921 in 1998. For non-participating Medi-Cal plans, the medical loss ratio incurred a greater increase from 0.842 in 1996 to 0.889 in 1998 while non-participating U.S. plans increased from 0.904 in 1996 to 0.928 in 1998. In all instances, the operating margins for participating plans in California were superior to those found among participating plans in the rest of the Nation as a whole. In addition, the operating losses for non-participating plans in California were lower than non-participating plans in the rest of the Nation as a whole.

Financial Performance of Medi-Cal Plans

For participants in the Two-Plan Model and the COHS, operating margins increased from 1996 to 1998, while participants in the GMC Model experienced a decline in the operating margin ratio (Table 3). The operating margin ratio for participants in the Two-Plan Model experienced a significant turnaround from -0.066 in 1996 to 0.059 in 1998, while the operating margin ratio for the COHS grew slightly from -0.001 in 1996 to 0.001 in 1998. Higher profits for the Two-Plan Model participants may stem from lower administrative costs and also from growth and maturity as these plans have had less experience in most cases than the plans in the other two models. During the 3-year period, the administrative cost ratio decreased for Two-Plan participants from 0.291 in 1996 to 0.101 in 1998. Higher initial administrative cost ratios among the Two-Plan Model contractors may be due to high startup costs for new local initiatives. The COHS administrative cost ratios were lower than the other models and remained around 0.07 over the 3-year period. The COHS was expected to have lower administrative costs since they incur no marketing costs.

Table 3. Health Plan Financial Performance, by California Models and Data: 1996-1998.

Model	1996	1997	1998
	Operating Margin
Two-Plan	-0.066	0.034	0.059
GMC	0.100	-0.001	-0.008
COHS	-0.001	0.014	0.001
	Administrative Cost Ratio
Two-Plan	0.291	0.123	0.101
GMC	0.159	0.122	0.119
COHS	0.070	0.075	0.074
	Medical Loss Ratio
Two-Plan	0.776	0.843	0.849
GMC	0.829	0.868	0.88
COHS	0.902	0.900	0.908

NOTES: California's Two-Plan Model includes two plans in each county selected by the State Medical Agency—a local initiative and a commercial plan. GMC is Geographical Managed Care. COHS is County Operated Health Systems.

SOURCES: Health Care Investment Analysts: Data from health maintenance organization database; California Department of Corporations, and California Department of Health Services, Sacramento, California. Computations prepared by McCue et al.

Each of the three model types incurred higher medical loss ratios over time while the Two-Plan Model had the lowest medical loss ratio in 1998. For participants in the Two-Plan Model, the medical loss ratio climbed from 0.776 in 1996 to 0.849 in 1998. Two-Plan contractors, especially local initiatives may have been more cautious in allocating premium revenues to medical expenses in the first 2 years of operation. Conversely, the medical loss ratio for COHS participants did not change over the 3-year period and remained around 0.90.

Key Informant Interviews

Similar interview protocols were developed and used in both the national and California studies with stakeholders from Medicaid officials, health plan representatives, HMO and hospital associations, and advocacy groups. The components of the interview protocol included questions related to model design, program management, market environment, plan characteristics, contracting, and rate issues. More than 60 telephone and in-person interviews were conducted across the two studies during 1999 in July and September.

Common Themes

There are many similarities across State experiences, including those in California. The general managed care market place is seen as highly turbulent and necessitating both vigilance and adaptability at the State level. Even States that have not seen substantial numbers of withdrawals, such as Arizona and Wisconsin, expressed some anxiety regarding the potential for forces beyond what is happening in Medicaid to disrupt and destabilize their programs. There was also widespread concern about declining numbers of Medicaid beneficiaries (which were still dropping sharply at the time of the interviews). This trend made the Medicaid market less appealing for plans that rely heavily or exclusively on Medicaid membership—such as the COHS and local initiative plans in California and resulted in hardships for these programs.

Most plans and other informants expressed concern about program management competence, the capacity and reliability of Medicaid agencies, and the uncertainty that this represents in securing long-term business relationships with external entities. Several States have experienced loss of key personnel, often to the managed care industry. Observers expressed concern about whether States without adequate leadership and expertise can adjust and adapt to market place changes or whether States may simply give up on prepaid managed care arrangements if turbulence continues. Because these interviews were occurring at the same time withdrawals in the Medicare market were growing, some plans noted these developments raise broader concerns about public sector contracting. Not surprisingly, there was a good deal of concern expressed about rates and contracts by all States. Debate over rate adequacy is a major issue in a number of States and this is unlikely to change.

A related concern among plans is that the administrative costs of Medicaid are not adequately subsidized and States have seemed intent on adding to these costs with new requirements. In other cases, plans are often forced to achieve standards set by States that go well beyond those that State FFS programs were achieving. For some plans, this may ultimately mean that the level of effort to comply with Medicaid requirements will no longer be sustainable. These plans are likely to be those with predominantly commercial membership that see Medicaid requirements as poorly conformed to private sector demands and thus representing additive costs, as implied in the national data trends analysis. Medicaid-specialized plans that are heavily dependent on this single line of business can and must build their systems and infrastructures around these program requirements. This is consistent with the findings on plan profitability as shown in Table 2.

Some Key Differences

The interviews conducted in California displayed a relatively strong sense of program stability, at least in terms of the Medi-Cal market and plan exiting has been very limited. Both the commercial and Medicaid-only plans have been relatively financially successful in the Medi-Cal market. This is all the more surprising in light of the fact that California Medicaid capitation rates are among the lowest in the Nation (Holahan, Rangarajan, and Schirmer, 1998). This may reflect higher levels of experience and acceptance of managed care in California, especially among providers who seem to be more passive price takers than found in non-west coast markets. It is also possible that the substantially higher disproportionate share hospital payments made to California hospitals makes them more willing to accept lower payments from health plans (Holahan, Rangarajan, and Schirmer, 1998).

In addition, it appears that plans in California are more likely to engage in downstreaming of financial risk with provider organizations in both the commercial and Medi-Cal markets (Kaiser Family Foundation, 2000). Such arrangements may aid plans by fixing their medical risk and shifting various administrative functions to their provider organizations. This would be consistent with the lower medical loss ratios and administrative costs found in the California plans in general and in the Medicaid/Medi-Cal plans in particular. Another consideration is that the local, community-based character of models such as the COHS and local initiative plans may improve the ability of plans to develop plan-provider relationships that make negotiations less adversarial. Conversely, reports of financial instability among medical groups in California raise questions about whether some of these relationships are going to be sustainable (Kaiser Family Foundation, 2000).

It may also be that through its multiple models, California has been able to successfully “manage its market” in such a way to ensure greater stability. Competition among plans is eliminated or greatly muted in the COHS and Two-Plan Model counties and consequently fewer dollars are spent in administrative activities like marketing. These models are also intended to preserve traditional Medicaid providers and to demonstrate sensitivity to community differences and desires, and they appear to be achieving this goal. Even the GMC Model counties have community-based advisory groups. As such, all three of these models may contribute to increased community involvement in local health affairs. They are seen as bolstering community infrastructure for local problem solving including support for the health care for uninsured persons as well as Medicaid beneficiaries.

Discussion and Implications

The updated findings on national trends confirm that withdrawal from participation continues among plans (primarily for-profit plans) and that an increasing number of remaining plans are serving predominantly Medicaid beneficiaries. However, these patterns vary across States.

Evolving Marketplace

As Felt-Lisk et al. (1999) have noted, the plans that are withdrawing appear to be ones that had limited Medicaid memberships and thus, the impact of their withdrawal is less significant than their actual numbers might suggest. Our data and interviews suggest that reasons for withdrawal are related to both a deterioration in the overall profitability of plans during the period and an inability of plans with small Medicaid enrollments to make this product profitable. The net effect of these withdrawals may be a positive one on the remaining plans whose memberships grow as a result of this attrition.

It is interesting to note that most of the individuals interviewed expect the number of plans to continue to decline and subsequently level off as States reach a kind of equilibrium point, especially if States make concerted efforts to maintain their HMO program. Some suggest that the extent of attrition reflects weak initial selection processes, or an inability or unwillingness to exclude entrants even when it was clear not all would succeed or stay. Others suggest that States that have less instability—California's Two-Plan and COHS Models would be a good example of this—are ones that designed their programs with a goal of limiting contractors and ensuring that participating plans had sufficient volume to gain expertise and economies of scale.

The shakeout among participants in some States suggests that Medicaid will almost certainly become more reliant on predominantly Medicaid plans (Felt-Lisk et al., 1999). Informants suggested that these plans may be centered around or perhaps sponsored by safety net providers or they could be from among an apparently growing number of multi-State Medicaid-only plans. A third type might be community based entities like the COHS and local initiatives, which are a confederation of public and quasi-public organizations—though there are substantial regulatory obstacles to developing the COHS in other States.

Other observers interviewed in this article highlight the fact that some States have been able to engage and sustain the participation of large local or regional health plans (including Blue Cross in some States). Blue Cross in California is a good illustration of a major commercial plan that has made substantial commitments to Medicaid participation. In principle, these plans have large networks and substantial management infrastructure already in place that can more easily support a Medicaid product. However, some commercial plans appear to have developed distinct and largely separate Medicaid product lines and networks, so the extent to which these plans offer a mainstream product might be challenged.

The future of Medicaid managed care will hinge largely on the issue of rate adequacy. Here, the picture is somewhat more promising and reveals the willingness of some States to respond constructively to market challenges. Nationally, the managed care industry is experiencing an economic turnaround, brought on in part by sizable increase in commercial product premiums (Felt-Lisk, 1999). Additionally, States have demonstrated growing attentiveness to the concerns of plans and display increasing sophistication in methods of ratesetting (Freudenheim, 2000). Not only are they addressing whether the levels of rates are appropriate but also whether the contractual demands and performance standards are commensurate with what plans are being paid. Several of the plans noted that States are showing more awareness of the administrative cost impact that their impositions may be having on plans and hope that this awareness is translated into more reasonable and flexible requirements. However, both States and plans commented on the additional burdens that the pending regulations under the Balanced Budget Act of 1997 will impose.

Tracking and Interpreting Change

The volatility of the managed care market and the high degree of dependence that Medicaid agencies now have on prepaid health plans underscore the importance of tracking trends in plan participation closely. There remains a need for a national database that includes all health plans, both licensed and non-licensed, participating in Medicaid that is more complete than databases created only from National Association of Insurance Commissioners filings. The database should have information specifically operating performance on the Medicaid product line. Such detail has now been added to the National Association of Insurance Commissioners filings for licensed plans and is expected to become more reliable over time. Data for all Medicaid-serving plans may be available through the Medicaid agency or health plan regulating body in many States, but it is not aggregated nationally. Such data will be unaudited in most instances but will provide industrywide benchmarks and support trend analyses.

More research is needed to understand the costs of the administrative components of Medicaid managed care products. The study found high administrative costs among Medicaid Dominant plans. This outcome may be attributed to turnover among Medicaid enrollees and the imposition of extensive contract and reporting requirements. A reversal of this trend occurred in 1998 when Medicaid Dominant plans incurred lower operating losses than Commercial Dominant plans. Declining administrative costs may have contributed to these lower losses; however, the administrative cost ratio of 18 percent for Medicaid Dominant plans still exceeds Commercial Dominant and mixed plans. These Medicaid Dominant plans are smaller in terms of overall median enrollment of 45,500 compared with median enrollment of 142,600 for Commercial Dominant plans that have a small Medicaid enrollment as a percentage of total enrollment. Smaller total enrollments hinder the ability of these Medicaid Dominant plans to achieve economies of scale in administrative costs.

There remains substantial variation that is difficult to explain at this time—with rather striking differences across plans and States. This may be due to size, plan maturity, program demands, or other factors. Indepth cost accounting studies would be useful to ascertain if the Medicaid product is more costly to administer than commercial products, and to identify the relative costs of various administrative components. This type of research might also assist in determining if the downstreaming of financial risk and delegation of functions to provider organizations are contributing to the markedly low administrative costs found in California.

Additional work in developing more refined studies on rates and rate adequacy is necessary. The study by Holahan, Rangarajan, and Schirmer (1998) established a previously unavailable benchmark with which to begin to make cross-State capitation rate comparisons. The study had a number of acknowledged limitations that will be improved in subsequent iterations. The relationship of rate variation to program participation and plan financial performance has not been studied indepth because of the limited availability of data.

Finally, both of the studies described here focus on participation and financial performance indicators, rather than how well plans are meeting access, use, and quality goals. These questions are beyond the scope of this analysis, but they are of critical importance in assessing the value of the overall Medicaid managed care strategy as well as appraising individual plan performance. This remains problematic on two levels. National summary data are not yet available in a standardized format to support comprehensive cross-State comparisons. Likewise, many States continue to struggle with basic data collection and reporting from plans and remain far from obtaining meaningful performance data to evaluate their managed care programs. On the other hand, some of the high performing States interviewed in this study, have developed mature, steady-State programs which have turned their attention to concerted quality improvement initiatives.

Conclusion

As Medicaid agencies have become more dependent on plans that voluntarily are participating in the Medicaid managed care market, the need to track change in participation among plans grows in importance. This participation appears to be affected by both general financial performance as well as specific Medicaid managed care results. This study updates prior research and casts new light on significant trends underway in Medicaid managed care. It also illustrates that States must have responded to market place changes in terms of both program design and operational differences. Such adaptability will continue to be essential to maintain viable models of managed care.