Findings from the Medicaid Competition Demonstrations: a guide for states.

The Medicaid Competition Demonstrations were initiated in 1983-84 in six States (California, Florida, Minnesota, Missouri, New Jersey, and New York). State experiences in implementing the demonstrations are presented in this article. Although problems of enrolling Medicaid recipients in prepaid plans or with primary care case managers under these demonstrations proved challenging to States, lessons were learned in three key areas: program design and administration, health plan and provider relations, and beneficiary acceptance. Therefore, States considering similar programs in the future could benefit from these findings.

Introduction

In 1982, the Health Care Financing Administration (HCFA) approved demonstration projects in six States to experiment with alternative methods of organizing and financing the delivery of care to Medicaid recipients. The demonstrations were developed in response to concerns that the Medicaid program was not fully meeting its goals of access to mainstream medicine, continuity of care, and cost containment. On the one hand, there was evidence of “doctor shopping” and of high self-referral rates contributing to excessive utilization, a problem that may have been exacerbated by the most common mode of cost control, i.e., fee constraints. On the other hand, patients lacked access to primary care physicians in some locations because of low Medicaid payment rates. As a result, they often received inadequate or inappropriate (e.g., emergency room) care.

The demonstrations were intended to test a number of concepts that State and Federal officials hoped would contain costs while promoting greater continuity of care and improving or maintaining access to care. The goal was to change the incentives facing both providers and consumers under Medicaid so that program goals could be met more effectively. The demonstrations incorporated a variety of innovations of the traditional Medicaid program structure:

Capitation as a mechanism of provider payment.

Case management by a primary care physician “gatekeeper.”

Limitations on provider choice as a means of promoting efficiency and, it was hoped, competition among providers for patients.

The purpose of this article is to share the lessons learned in the demonstrations with other States that are considering making similar changes in their Medicaid programs. We first provide an overview and comparison of the demonstration sites, then we offer findings from the demonstration experience in three areas: program design and administration, health plan and provider relations, and beneficiary acceptance.

Overview of demonstration sites

Highlights of the demonstrations are provided below and summarized in Table 1. References to the series of detailed case studies available for each demonstration are cited at the end of the article. Particularly for the completed demonstrations, an indicator of State satisfaction with the projects is their current status. Three of the demonstrations have been converted to ongoing State programs (Santa Barbara, California; Missouri; and New Jersey), one has been extended (Minnesota), and two have ended (Monterey, California; and New Jersey).

Table 1

Overview of demonstration programs

Demonstration site	Dates of operation	Enrollment	Organizational structure	Eligible populations	Participating providers	Provider payment
California
Monterey County	June 1983-March 1985	Mandatory	State contracts with risk-assuming county authority, which in turn contracts with primary care physicians and hospitals	AFDC, SSI, and medically needy	Primary care physicians, clinics, and hospitals	Fee for service
Santa Barbara County	September 1983-December 1986 (ongoing State program)	Mandatory	Same as Monterey	AFDC, SSI, and medically needy	Same as Monterey	Capitation
Minnesota
Hennepin and Dakota Counties	December 1985-June 1990	Mandatory	State contracts with prepaid health plans	AFDC, SSI, and medically needy	Prepaid health plans	Capitation
Itasca County	August 1985-June 1990	Mandatory	State contracts with county	AFDC, SSI, and medically needy	Primary care physicians	Capitation
Missouri
Jackson County	November 1983-December 1986 (ongoing State program)	Mandatory	State contracts with prepaid health plans and with individual physicians	AFDC	Prepaid plans and primary care physicians	Capitation for plans Fee for service for physicians
New Jersey	June 1983-June 1987 (ongoing State program)	Voluntary	State contracts with primary care physicians and clinics	Noninstitutionalized AFDC and SSI	Primary care physicians and clinics	Capitation
New York
Monroe County	June 1985-August 1987	Mandatory	State capitates county, which contracts with intermediary, which contracts with prepaid health plans	AFDC, Home Relief	Prepaid health plans	Capitation

NOTES: Of 4 proposed modules in Florida, 1 is currently operational as a demonstration. This project has enrolled frail elderly in a prepaid program sponsored by Mt. Sinai Hospital in Miami. AFDC is Aid to Families with Dependent Children. SSI is Supplemental Security Income.

SOURCE: Adapted from Robert E. Hurley (1986).

California demonstration programs

There were two programs in California: the Monterey County Health Initiative and the Santa Barbara County Health Initiative. In both, the county established a new authority that accepted capitation payments from the State and, in turn, contracted with physicians, clinics, hospitals, and other providers. Despite this organizational similarity, the programs differed in two critical areas. The first relates to the method of provider payment: Monterey paid a fee for service plus a case-management fee for primary care physicians, while Santa Barbara capitated primary care physicians for their services and paid for referral services on a fee-for-service basis. Second, physician attitudes differed: Although there was resistance to managed care in both counties, a few well-respected physicians served on an advisory board in Santa Barbara and helped smooth the way for utilization review and other cost-containment efforts. Physician opposition to the capitation of case managers, for example, was overcome in Santa Barbara but not in Monterey.

The Monterey program was terminated in March 1985 because of cost overruns and administrative difficulties; the Santa Barbara program, however, has become an ongoing State initiative. California passed State legislation and obtained a 1915(b) waiver from HCFA allowing the Santa Barbara program to continue operating.

The Monterey program ran into trouble because physicians were paid on a fee-for-service basis at rates that exceeded Medi-Cal's, without any incentives or administrative controls to alter medical practice patterns until the second year of operation. In addition, the resulting cost overruns were not recognized for several months because the management information system was inadequate; by the time the problems were identified and controls were developed, the demonstration could not be resuscitated. An important lesson is that counties accepting financial risk for the delivery of services must behave like health maintenance organizations (HMOs) with respect to controlling and tracking utilization.

Minnesota Prepaid Competition Demonstration Project

Minnesota elected to experiment with prepayment in three counties: one urban (Hennepin, which includes the city of Minneapolis), one suburban (Dakota), and one rural (Itasca). The approach in Hennepin and Dakota Counties, which are geographically contiguous and served by many of the same HMOs, differed from that in Itasca County. The demonstration in Hennepin and Dakota required Medicaid recipients, including most aged, blind, and disabled persons, to enroll in prepaid health plans. In Itasca, the county itself was treated like a health plan; it received capitation payments from the State and, in turn, paid providers. The demonstration in Minnesota (Hennepin, Dakota, and Itasca) was delayed in starting and will end in June 1991.

Missouri Managed Health Care Project

In Jackson County (Kansas City), Aid to Families with Dependent Children (AFDC) cash recipients were required to enroll in one of five prepaid health plans (including two formed by community health centers, an IPA-type (individual practice association) HMO, and two hospital-based plans) or with individual physicians who were paid a fee for service and a case-management fee. By far, the most popular plans were the two sponsored by the teaching hospitals. The program also proved a useful introduction to prepayment for two community health centers, one of which hopes to obtain prepaid contracts in the future.

Missouri also obtained the Federal waivers necessary to operate an ongoing program, and four of the five prepaid plans that participated in the demonstration have continued serving Medicaid recipients on a prepaid basis. To do so, each met Federal requirements for participation (i.e., become a federally qualified HMO with no more than 25 percent Medicare and Medicaid enrollment or qualify as a public HMO or federally funded community health center).

New Jersey Medicaid Personal Physician Plan

The New Jersey program originally was intended to be statewide, but, by the end of the demonstration period, it was implemented only in 11 of 20 counties. New Jersey's was the only demonstration project in which recipient participation was voluntary, and it had difficulty attracting enrollment. The State contracted with individual physicians, clinics, and community health centers and paid them a capitation to provide primary care services. The program was not successful in attaining its goal of attracting new physicians to the Medicaid program, and most of the demonstration enrollees were served by a small number of traditional Medicaid providers.

New Jersey's demonstration has been converted to an ongoing program, operating as Garden State Health Plan, a State-certified HMO. Federal legislation was needed to establish this health plan because the State itself acts as the HMO.

The Monroe County (New York) MediCap Program

In Monroe County, New York State capitated a county-level intermediary known as MediCap, Inc., which in turn capitated prepaid health plans. Although the State had intended MediCap, Inc., to contract with several health plans, most declined to bid, and a single HMO (Rochester Health Network, known as RHN) provided care to demonstration participants throughout most of the project. However, RHN—a network model with multiple providers, including community health centers, individual physicians, and hospitals—allowed participants considerable choice of provider.

The New York (Monroe County) demonstration ended in August 1987 (after having received an extension from HCFA to operate through April 1988) when RHN withdrew because it was unable to reach agreement with the State on an appropriate capitation rate. As a result of declining utilization in the fee-for-service system, the State wanted to pay only a small increase for the third year, but RHN said its contracting providers faced considerably higher costs. (Providers were still unhappy with the State for having reduced payment rates between the first and second years.) Some Monroe County Medicaid recipients continue to be enrolled on a voluntary basis in another local HMO.

Florida Alternative Health Plan

Florida originally planned to implement four modules, but only two of them ever became operational. Module A, which involved Medicaid enrollment in prepaid plans, failed largely because the established HMOs found the Medicaid capitation rate too low, particularly when compared with the alternatives (Medicare capitation rates, for example, were much more generous). In addition, the plans objected to what they viewed as arbitrary enrollment caps established by the State. Module B, case management for recipients who consistently overutilize or underutilize medical care, was partially implemented but quickly converted from the demonstration to an ongoing State program. After a lengthy development period, Module C, prepaid health plans for frail elderly patients, eventually became operational at one site only. Module D, medical care vouchers offered by private insurers, failed to attract insurer interest and thus could not be implemented. Because its only operational demonstration had little in common with the other States, Florida was excluded from the Research Triangle Institute (RTI) evaluation and is addressed only selectively herein.

Shared goals and characteristics

Despite their differences, the demonstrations shared some fundamental concepts: capitation, primary care case management, and limited provider and health plan choice. These concepts were employed to create competition among providers and health plans in order to promote cost-effective health care delivery.

All of the programs used capitation as a method of payment, at least at one level. In some cases, States capitated health plans either directly (Minnesota and Missouri) or through an intermediary (New York). In other cases, States treated counties like health plans and paid them a capitation (Monterey and Santa Barbara, California; and Itasca County, Minnesota). The counties at risk in turn contracted with providers, paying them on either a capitation (Santa Barbara) or a fee-for-service (Itasca and Monterey) basis. Finally, New Jersey capitated primary care physicians directly.

Most of the demonstrations employed case management; in some, it was explicitly a requirement of physician and hospital participation. Demonstrations that contracted directly with individual physicians required them to assume case management responsibilities, i.e., to provide all primary care and to authorize specialty and inpatient referrals. In Monterey and Missouri, physicians received a case-management fee ($1.50 per patient per month in Missouri) for this service. In demonstrations that contracted with health plans, the method of case management was typically left up to the individual health plan (Minnesota and New York).

Although dubbed the “competition” demonstrations, evidence of competition among providers and/or plans was mixed at best. By and large, the demonstrations did not change the fact that States are more often in a position of recruiting health plans and providers than of selecting among competitors. In Hennepin and Dakota Counties, where the health plans were more eager for Medicaid business than in any other site, the State received nine applications, seven of which were accepted. Although other sites, such as New York, had trouble recruiting adequate numbers of participating plans, there was not noticeably more competition among participating plans for enrollees in Minnesota than in New York. This primarily reflected the plans' concerns about the inability to make money by serving the Medicaid population, resulting in a reluctance to enroll large numbers of recipients.

About the only evidence of competition (and this was not extensive) occurred among traditional Medicaid providers that had formed health plans specifically to serve this population. Two plans in Minnesota and four in Missouri fell into this category, and each of those plans focused more attention on the demonstrations because they viewed Medicaid as a major line of business.

Program design and administration

Developing and implementing the Medicaid demonstrations required a substantial administrative commitment from the respective States and/or counties. Minnesota, New Jersey, and Missouri all created State-level offices to manage the demonstration and other prepayment initiatives; New Jersey relied on existing Medicaid staff. In California (Santa Barbara and Monterey Counties) and New York (Monroe County), most administrative functions were performed by county-based authorities.

When a State decides to implement prepayment and/or primary care case management, it faces the following administrative issues:

Legal requirements, such as the need to enact State legislation and/or obtain waivers from HCFA.

Program design choices relating to the populations and services the program will cover as well as the type of payment and delivery system to be established.

Problems associated with managing enrollment and disenrollment.

The management of data flow, particularly current enrollment information and utilization data (psuedoclaims), to and from participating plans and providers.

The monitoring of participation plans with regard to financial solvency and quality assurance.

Staffing needs.

Waivers and legislation

States that establish Medicaid prepayment programs modeled on the demonstrations will generally need to apply to HCFA for waivers from certain provisions of Title XIX, Grants to States for Medical Assistance Programs. In addition, some States require approval by the State legislature (e.g., California, Minnesota, Missouri, and New York).

HCFA granted section 1115 research and demonstration waivers to the demonstration sites, but requests to replicate these demonstrations are likely to be denied in the future because HCFA reserves this authority to test bona fide innovations. However, States can develop similar programs through the regular program waiver mechanism. In general, States need to apply for section 1915(b) Medicaid waivers to limit recipients' freedom of choice, offer a different Medicaid program in one area of the State, change the comparability of services among recipient categories (e.g., by adding a service), and/or make payments (such as capitation or case-management fees) that exceed the fee-for-service costs for comparable recipients.

In addition, the county-at-risk model employed by Santa Barbara, Monterey, and Itasca Counties is no longer feasible under current law. Thus, in planning a new program, States should seek guidance at an early stage as to the organizational options that are acceptable and the circumstances under which waivers are likely to be granted.

Key program design considerations

States will face a number of decisions about program design early in the planning process. Among these are:

Which Medicaid populations will be included? Missouri served AFDC (cash) only; Minnesota and Santa Barbara and Monterey Counties covered the Supplemental Security Income aged, blind, and disabled. Minnesota and California included the AFDC medically needy, and New York covered county Home Relief recipients in addition to AFDC (cash) recipients. New Jersey's program was open to any noninstitutionalized Medicaid eligible.

What kind of payment and delivery system will be established? The demonstrations represent a mixture of competitive prepaid health plans (Hennepin and Dakota Counties in Minnesota, Missouri, and New York); fee-for-service partially capitated primary care case management (Missouri and New Jersey); and counties at risk (Santa Barbara and Monterey Counties in California, and Itasca County in Minnesota).

What services will be included in the demonstration? The programs included most Medicaid-covered services. However, some States excluded items such as prescription drugs or dental care and reimbursed them on a fee-for-service basis. Minnesota largely excluded the room and board cost of institutional care (skilled nursing facility and intermediate care facility for the mentally retarded) from the capitation payments to health plans. Santa Barbara excluded certain services, including obstetrics, from the capitation payments to physicians.

The question of whether to operate a voluntary or a mandatory program, a critical policy decision, is addressed later in the Enrollment section.

In many cases, the need for adaptations in programs to achieve political accommodations became evident early on. As with the development of any new program, public or private, extensive negotiation and compromise shaped the outcome. Two common sources of political friction were intergovernmental conflicts, notably between a powerful county and the State, and conflicts with the provider community. States need to strike an appropriate balance between necessary compromise and “giving away the store” to special interests, as arguably happened in Monterey County, leading to the demise of the demonstration there. In Monterey, it was the physicians' opposition to capitation that led to a change in the method of provider payment to fee for service. In other sites, compromise led to more satisfactory outcomes.

As an example of county influence, in Minnesota, one of about six States where Medicaid is jointly administered by the State and the counties, Hennepin County was reluctant to participate in the demonstration. Among other concerns, it feared an adverse financial impact on its medical center, which is a major Medicaid provider. The county thus established several conditions for its participation in the demonstration, most of which the State ultimately accepted. Two of the more visible outcomes of this negotiation were the enrollment of only 35 percent of the target Medicaid eligibles from Hennepin County in the demonstration, leaving the remainder under fee-for-service Medicaid, and the use of an outside broker (a nonprofit organization specializing in marketing public programs) to conduct consumer education and enrollment because the county, which is responsible for eligibility determination, did not want its workers to assume this responsibility. (A similar concern for the county hospital in Monterey County resulted in agreement to payment rates that were widely criticized as too generous.)

In Missouri, primary care physicians were included as case managers and paid a fee for service only after a number of them opposed the State's plan to enroll Medicaid recipients exclusively in prepaid plans. (These physicians relied on Medicaid patients as a major source of practice revenues and were for the most part not affiliated with prepaid plans.) As a result, the Missouri Managed Health Care Project allows recipients to sign up with 1 of some 50 participating primary care physicians in lieu of enrolling in a prepaid plan (approximately 15 percent of the population exercised this option). Furthermore, the powerful retail druggists' lobby was successful in excluding prescription drugs from the program.

Not surprisingly, these political conflicts and negotiations extended the planning process and delayed startup; the Minnesota program, for example, did not begin enrolling recipients until November 1985, 16 months after the demonstration was slated to begin.

Enrollment

Enrollment requires both policy decisions and ongoing administrative attention. Three key decisions are: whether enrollment in the demonstration is mandatory or voluntary, how to educate recipients about their choices, and how recipients who fail to make a choice are assigned to a plan or provider in mandatory programs. States also face ongoing administrative issues with respect to tracking enrollment and disenrollment; these are described later in the Data management section.

A basic policy decision facing States was whether to establish a mandatory or a voluntary program. All but New Jersey chose to make enrollment mandatory, principally to increase the numbers of participants. Cost savings was the prime motivation because, historically, voluntary programs have had low participation. In addition, significant enrollment is necessary for program viability (e.g., to attract providers, justify administrative costs, and permit meaningful evaluation). Some of the States already had voluntary HMO enrollment for Medicaid recipients (New York and Minnesota) in which participation was low, and they sought more dramatic results. In New York, an exception to mandatory enrollment was made for patients who had an ongoing relationship of a year's duration or longer with a provider not included in the demonstration.

In both voluntary and mandatory programs, recipients must be informed of the health plan options. In a voluntary program, the emphasis is on the merits of prepayment relative to fee for service (e.g., one-stop shopping), whereas in a mandatory program, the focus is on the differences among the health plan options. New Jersey, the only voluntary program, initially contracted with a consulting firm to market the program, then later allowed physicians serving as case managers to educate and enroll recipients in their offices. This boosted enrollment but raises questions about biased selection as well as the potential for misleading messages and lack of uniformity in the information provided to recipients.

Most of the mandatory programs handled consumer education and enrollment at the time eligibility was determined or recertified. The advantage of conducting education and enrollment in this manner is that recipients are generally receptive and motivated to make a choice. Minnesota initially contracted consumer education and enrollment to an outside broker who contacted recipients after they had established their eligibility and were using the fee-for-service system. This process resulted in a high percent of recipients failing to choose a health plan and, consequently, needing to be assigned by the broker. In Missouri and Santa Barbara, California, recipients were informed of their health plan options in a group presentation in the county social services office at the time they established eligibility, and most recipients indicated their preferences at that time. Only a small percent refused to attend the presentations and were assigned to one of the options.

Under the demonstrations, States either prohibited direct marketing to consumers or constrained it. Except for New Jersey, States did not allow plans or providers to market their services directly. (New Jersey required providers to attend a training session and abide by certain rules when educating and enrolling patients in their own offices.) In general, plan brochures and other literature were reviewed and distributed by the States or intermediaries. The restrictions on marketing to Medicaid recipients proved not to be a significant constraint, because most health plans and other providers were not interested in actively marketing to this population. Instead, the plans were content to gain experience serving small numbers of recipients. Even for traditional Medicaid providers participating in the demonstrations, such as community health centers and public hospitals, marketing was not a key issue. Some passively attracted their clientele in adequate numbers without it, and others, who might have liked greater enrollment, lacked the resources to pay for marketing. Despite this, it is likely that some health plans, at least in a modest way, would take advantage of the ability to market their services directly to patients via advertising campaigns (mailed brochures, telephone marketing, etc.) if this were permitted.

Medicaid recipients who should have made a health plan or provider choice but did not were customarily assigned to a participating plan or provider. Recipients were randomly assigned in Minnesota, Missouri, and Santa Barbara. (In Santa Barbara, patient age was considered in order to avoid, for example, assigning pediatric patients to an inappropriate case manager.) In New York, assignment took into consideration such factors as prior provider use and geographic location. (The relationship of assignment to the potential for biased selection is discussed later in this article.)

Data management

State information systems should generally be adapted to meet the demands of a prepayment program. Under the demonstrations, two critical areas of data management were enrollment and utilization. Current enrollment data are always important in a capitated system. Utilization data (e.g., records of the volume and type of services delivered) may be collected as well, but their value to the State has been questioned. (Despite good intentions, most demonstration States had other pressing priorities and did not marshall the resources necessary to analyze the data.)

Under prepayment, participating providers and plans need prompt notification of enrollments and disenrollments. They should also receive a statement itemizing the capitation payment for each enrollee, and the two sets of data should match. This information is important because providers and plans are typically at risk for the cost of health services and, therefore, need to know whether they will be paid for services they provide or authorize. (There is also the possibility of denial of service for those who are entitled if a health plan does not have a record of new enrollees.)

Several problems made it difficult for States to provide timely enrollment data to plans: lack of priority for State computer time (i.e., queueing problems), programming errors, poorly trained staff and/or inadequate management information system (MIS) staff time, lack of necessary hardware and software, and computer failures. In general, State systems for processing eligibility data were different from those used to maintain health plan enrollment, and it was difficult to merge the two. Missouri modified its MIS for the demonstration, but most States struggled along with their existing systems.

The second area of data flow relates to tracking services utilization through pseudo or dummy claims. (HMOs that record only utilization and not costs commonly refer to these as “encounter data.”) As part of the evaluation effort, HCFA required all demonstrations to compile pseudoclaims, even though those claims were no longer the basis of payment (because the capitation rates were fixed and not related to services rendered). The data were analyzed by the HCFA-contracted evaluation team headed by the Research Triangle Institute. In addition, most States intended to use these data to compare the performance of the demonstrations with fee for service. Other States had reinsurance provisions that limited the financial liability of a plan once expenses for an individual enrollee reached a preset level, and the pseudoclaims served to determine when that level had been reached. However, it is not clear that the States actually used much of the data.

If States wish to collect utilization data, they should plan carefully for uniform reporting and efficient transfer of the information to State data bases. Under the demonstration, contracting health plans and providers had some incentive to submit pseudoclaims in accordance with State wishes because of the reinsurance provisions described earlier. Nonetheless, problems abounded. States reported inconsistencies and lack of comparability both over time and across providers and plans. Plans for their part said States were unclear in their specifications. Standardized reporting is difficult because health plans and providers have different definitions of an encounter. Also, although some health plans (particularly staff and group) may have computerized encounter data, plans that capitate medical groups do not themselves record encounters; and the chances of noncomparable data increase when they are reported by individual providers.

In addition, the States did not keep up with the input and tabulation of these data, creating a formidable backlog. Some observers question whether the States will eventually allocate the necessary resources to sort and analyze these records, which were submitted at considerable time and expense by some providers and health plans. Missouri, for example, had planned to base its payment rates in subsequent years on pseudoclaims data from early years of the demonstration. This never occurred because the State decided that its methodology was adequate and because it had concerns about possible inconsistencies in the pseudoclaims data.

Quality assurance and program monitoring

States and/or counties administering the demonstrations sought to monitor quality of care and financial solvency. Most programs devoted relatively few staff to either activity, reflecting mainly the press of routine administrative responsibilities and the lack of any obvious problems or threats to the demonstrations stemming from insolvency (except in Monterey) or poor quality care. States also recognized that prepaid health plans had their own quality assurance systems, although the majority of those delivering care under the demonstration were not prepared health plans, except in Minnesota and New York.

States approached quality of care monitoring in different ways. In New York, consumer satisfaction surveys were conducted pre- and post-demonstration. MediCap, Inc., the agency administering the project, reported improvements in consumer satisfaction levels under managed care. (However, advocates claimed that enrollees were disgruntled with restricted access to providers under the system and were overall less satisfied than before.) MediCap delegated responsibility for ongoing quality assurance efforts to Rochester Health Network, the health plan serving demonstration enrollees. Minnesota hired a quality assurance coordinator in 1986 and charged her with developing methods of assessing quality beyond the existing grievance process. Minnesota also plans to evaluate questionnaires completed by recipients who switched health plans during open enrollment to assess reasons for dissatisfaction and to compare results across plans.

In Missouri, the quality assurance department was headed by a registered nurse who reviewed medical records. Protocols for selecting and reviewing records were developed, and nurses conducted onsite audits. Both Missouri and Santa Barbara conducted consumer satisfaction surveys and analyzed recipient requests to change plans or providers to determine if quality of care was at issue. The Santa Barbara program was especially aggressive about soliciting consumer feedback on the demonstration and following up any reported problems.

The States struggled with quality assurance but ultimately did little, in part because they were unsure how to proceed. Because consumers have difficulty assessing technical quality of care and because of beneficial “sentinel” effects from monitoring quality, it is useful for States to be more proactive in this area. States could, for example, target particular diagnoses for review. They could also check ambulatory records for routine immunizations, proper followup of abnormal laboratory test results, or other indicators of clinical quality.

Staffing

States should expect their staffing needs to change under a prepaid Medicaid program. In addition to understanding the traditional Medicaid program, staff need experience dealing with managed care issues, including ratesetting, marketing and consumer education, MIS, and quality assurance. Acquiring such staff requires perseverance because of civil service hiring procedures and because the talents involved are in demand by the private sector.

The use of outside consultants to supplement State staff created problems in some cases. Such consultants can provide valuable services, as in Minnesota and Missouri, where they were used during the development phase for assistance in both planning and ratesetting. It is important that State staff have the technical ability to work in a collegial fashion with consultants and supervise their efforts.

The experiences in New Jersey and California illustrate the problems of using outside consultants. Santa Barbara and Monterey initially hired the same consulting firm to develop and run a management information system. Santa Barbara eventually assumed the responsibility in-house after encountering problems. In Monterey, the contractor promised more than it could deliver. Although a system for paying claims was up and running within about 3 months, it took a year before timely utilization reports were available to providers. The cost overruns that accumulated during this time were largely responsible for the demise of the Monterey demonstration. New Jersey relied heavily on a consulting firm to conduct many operational aspects of the program and lacked the technical staff to supervise the firm. In both New Jersey and Santa Barbara, contracting helped alleviate problems in the short run but detracted from the building of necessary internal capacity.

Health plan and provider issues

Most of the demonstrations succeeded in soliciting the participation of health plans and providers. The major exception was Florida, where a combination of factors (mainly low payment rates but also the exclusion of health plans from the planning process and restrictions on market share) discouraged plans from bidding. However, neither New Jersey nor New York obtained as many participating plans (New York) or providers (New Jersey) as desired. It is noteworthy that, in general, most of the providers and plans had participated in Medicaid prior to the demonstrations. Features of Medicaid prepayment that particularly relate to providers and plans include:

The planning process and provider recruitment.

Ratesetting and the availability of reinsurance.

The potential for biased selection.

Service delivery issues.

Planning process and provider recruitment

States that involved health plans and providers (as well as consumer advocates and other interested parties) in the process of planning for the demonstration took longer to become operational but may have benefited in the long run. Minnesota is a case in point; after a planning process in which the State consulted extensively with interested parties, the program was implemented with widespread community acceptance and understanding of its intent. (The downside to a thorough planning process is delay, and some observers felt that “tinkering” in Minnesota and New York unduly delayed startup.) In contrast, Monterey became operational quickly before necessary systems (e.g., for monitoring utilization) were in place; in order to defuse the concerns of physicians and recruit case managers, the program acquiesced to high payment rates. In retrospect, more careful planning and a strategy for recruiting providers that did not place the demonstration in financial jeopardy might have paid off.

States also benefit from the goodwill and understanding on the part of health plans and providers that can result from giving them an opportunity to be consulted. (Problems and delays affecting health plans and providers are inevitable during the startup phase, and the existence of good working relations can reduce frictions.) Finally, by consulting with the health plans or providers, the State gains a better understanding of their objectives and concerns. The lack of such understanding was one of the problems that doomed the Florida effort almost from the outset.

Health plans and providers participated in the demonstrations for many reasons, but chief among them were the goals of expanding or maintaining market share and gaining experience with Medicaid prepayment. However, other variables also came into play. Some health plans participated simply because their physicians wanted them to (e.g., Prevention Plus in Missouri), often reflecting a concern that they might otherwise lose patients. Sometimes the plans' payment levels to physicians were higher than those in the Medicaid fee-for-service program. In other cases, the State's willingness to guarantee eligibility for 6 months for demonstration enrollees convinced health plans that it was worth the administrative burden. Individual physicians were sometimes attracted to prepaid programs instead of fee for service by the significant reduction in paperwork, although in actuality the paperwork was not always lessened (physicians in Santa Barbara, for example, complained about the paperwork associated with treatment and referral authorizations).

Ratesetting

States must balance the need for payment rates generous enough to attract plans and providers against the goal of containing costs. Similarly, a balance must be struck between requiring contractors (e.g., health plans) to bear some risk for the cost of health care for recipients and asking them, particularly those with little prepayment experience, to assume more risk than they can afford. Reinsurance, often provided by the State, reduces the degree of exposure.

The major factor determining providers' willingness to serve Medicaid patients on a prepaid basis is the perceived adequacy of the payment rate. In the demonstrations, the starting point for the development of rates was the fee-for-service experience. Following the Medicare program's precedent, the States' objective was to pay providers, plans, or counties at risk a percent of estimated fee-for-service costs (typically 90 or 95 percent) because this was believed to ensure savings. Importantly, by establishing a capitation rate in this manner, States set the stage for holding down future costs to a percentage increase. In some cases, the intent was to use fee-for-service experience only at the start of the program and constrain future year increases.

States based rates on the fee-for-service experience by rating category for the Medicaid population in some prior year or years. The number of categories (rate cells) used varied from 2 in Missouri (AFDC adult and AFDC child) to more than 70 in Minnesota (reflecting age, sex, eligibility category, residence in an institution, etc.). These data were then actuarially adjusted for any program differences between fee-for-service Medicaid and the demonstration (e.g., changes in covered services, eligibility, and geographic area). Where applicable, the rates were reduced to reflect State-provided reinsurance protections. Finally, the rates were trended forward to the appropriate year.

Predictably, rates were a common source of friction between health plans or other providers and States. Complaints included:

Inadequate documentation of ratesetting methodology.

Trend factors viewed as unfairly low.

Too many or too few rate cells and lack of homogeneity within rate cells.

Inappropriate geographic base (e.g., use of statewide averages for a locality or, conversely, reliance or too small a sample of local recipients instead of statewide data).

Disagreement over rates that are derived from the fee-for-service experience is likely to increase over time as the base year recedes, practice patterns change, adjustments are made to reflect benefit changes, and so on. Demonstration health plans and providers were dismayed when, in some cases, the second and third year rates declined as a result of reduced expenditures in the fee-for-service program (New York and Missouri). Despite a clear rationale for the reductions, providers and/or plans saw their own costs rising and viewed this phenomenon as evidence of the inadequacy of the ratesetting process.

The New York demonstration ended in the wake of a dispute over rates. After the first year, the State reduced capitation rates by 11 percent; after the second year, it proposed a modest increase (about 4 percent), but providers affiliated with the major participating health plan felt they were losing money and were unwilling to continue without a substantial payment increase. A separate problem arose in Minnesota, where the capitation levels rose by the same fixed percent each year as did fees in the regular program. The plans felt that they were disadvantaged because their total revenues were capped, whereas fee-for-service provider revenues could (and did) rise as a result of increases in volume and intensity of services.

Despite complaints from a number of plans and providers who reported losing money, many were satisfied with the rates or, at least, lacked grounds for challenging them. Furthermore, many recognized that the States were developing capitation rates for the first time and were trying to be fair.

Reinsurance

Reinsurance (or stop-loss) protection limits the risk that plans face either on a per enrollee basis or in the aggregate. This protection was a major concern of smaller and less experienced health plans, and the State's willingness to offer such coverage was in some cases a determining factor in their participation. In Missouri, for example, community health centers viewed this protection as vital to ensure their participation as prepaid health plans. Missouri originally considered a simple catastrophic reinsurance arrangement whereby the State agreed to pay annual per recipient costs above $20,000 of Medicaid-allowable expenses. However, the hospitals and community health centers that formed plans to participate in the demonstration had no experience accepting financial risk and wanted more protection. The State then developed special risk pools to accommodate their concerns, specifically, a pool to fund (on a fee-for-service basis) high-risk deliveries and neonatal intensive care for patients hospitalized more than 9 days, a pool for adverse selection (eventually abandoned because a method for measuring adverse selection was lacking), and another to fund any additional births after the plan had absorbed the costs for a certain number of deliveries. The plans agreed to have a small amount deducted from their monthly capitation in return.

In other cases, for example New York, large plans accustomed to bearing risk preferred to handle reinsurance themselves, ostensibly at lower cost. Although States sometimes permitted individual providers and plans to decide whether to purchase reinsurance, more often than not a minimum level of coverage was required, and its cost was automatically deducted from the payment rate. Santa Barbara experienced more catastrophic claims than expected, and the State responded by increasing the per enrollee reinsurance limit from $15,000 to $25,000, leaving the county authority with less protection.

Minnesota had both individual and aggregate reinsurance provisions. For individual patients with over $15,000 (AFDC) or $30,000 (aged, blind, disabled) in cumulative annual hospital expenses based on Medicaid allowable charges, the State paid 80 percent of the cost. The aggregate reinsurance provision applied only during the first 2 years of the demonstration. For the AFDC population (capitated at 90 percent of fee for service), the State agreed to pay 50 percent of first-year losses based on actual costs between 90 and 110 percent of fee for service. Above 110 percent, the plans were fully at risk. In the second year, the State shared one-half of the loss between 90 and 100 percent (rather than between 90 and 110 percent). Similar provisions applied for the aged, blind, and disabled population. Capitation rates were reduced to account for the per enrollee reinsurance protection but not for the aggregate risk-sharing provisions.

Potential for biased selection

Biased selection refers to the systematic enrollment into a prepaid system of individuals who are healthier (favorable selection) or sicker (adverse selection) than average. The result is underpayment or overpayment by a rate structure that assumes average risk. The consequences of biased selection differ, depending on whether enrollment is voluntary or mandatory. In voluntary programs, the State may be concerned that the plans or providers at risk obtain a disproportionately healthy enrollment, leaving the sicker and higher utilizing enrollees in the fee-for-service system. For example, in New Jersey, a voluntary demonstration, it is possible for healthier Medicaid recipients to join while sicker recipients stay in the fee-for-service system (or vice versa). In mandatory programs, the concern is for biased selection across providers or plans.

Providers and plans in some demonstrations claim to be victims of adverse selection. Such assertions have not been fully documented, although in analyses conducted in Missouri as part of the RTI evaluation no evidence was found of adverse selection at that site. Adverse selection could be more likely in situations where the participating providers are hospital-based and/or specialize in the treatment of such chronic conditions as arthritis and diabetes because these providers may attract chronically ill patients. In addition, health plans with particularly broad provider panels (such as many Blue Cross and Blue Shield plans) could be adversely selected against because patients with multiple physician relationships are believed to be more likely to choose the plan that allows them to maintain these relationships.

The consumer education process, if not unbiased, could also steer recipients to certain plans, a concern that some plans expressed.

Finally, the assignment of recipients who do not elect a provider offers another opportunity for biased enrollment. The assignment process can be random, as in Minnesota and Missouri, or subjectively determined, as in New York. If social service workers are responsible for assigning recipients to health plans, they may make biased assignments, relying on their own impressions of the participating providers.

Service delivery

It was hoped that the demonstrations would contribute to the state of the art in delivering services to the populations covered under Medicaid. This expectation was, for the most part, not realized, perhaps because participating plans and providers did not perceive it worth their while to invest in developing methods for improving care to these populations for a 3-year demonstration at the Medicaid payment rates. In addition, both States and plans or providers were hesitant about enrolling the aged and disabled populations, who generally use services much more intensively than AFDC recipients.

Few of the plans adapted their delivery systems to meet the special needs of, for example, patients with chronic conditions. Established HMOs generally have limited experience with patients who suffer from mental and physical handicaps, mental retardation, and chronic mental conditions. These patients may need more continuous and intensive medical supervision than prepaid plans typically provide. As discussed more fully in the next section, the demonstrations appeared not to result in innovations in this area.

By the same token, some public teaching hospitals that traditionally served the Medicaid population were also slow to adapt. In these hospitals, the teaching curriculum typically takes precedence; and, thus, prepaid enrollees are likely to use physician specialty care and ancillary services at a higher rate than enrollees in private plans, making cost control difficult. Further, the hospitals have often been slow to introduce effective utilization review and patient-management procedures (e.g., prior authorization). As a result, it is business as usual for many plans operated by teaching hospitals.

State Medicaid programs have developed systems and policies over the years that are not necessarily compatible with prepaid care. Examples of problems were:

Court-ordered treatment, often for mental health services or substance abuse cases, which Medicaid traditionally reimburses. Prepaid plans do not customarily pay for any nonemergency care that they have not specifically authorized, but the State will not pay for services included in the capitation. Thus, the party responsible for the cost of court-ordered care may be in dispute.

Overlapping responsibilities for the case management of special populations, such as the mentally retarded. For example, in Minnesota, county workers traditionally performed this service and, in some cases, were unwilling to cooperate with the health plans.

Managing nonmedical benefits under a medical case management model, such as the problem in New York of day treatment for children with emotional, physical, or developmental handicaps. Such treatment is covered by Medicaid in New York, and its cost was part of the capitation rate. However, the treatment is expensive (estimated at $9,000 per year), and the referral decision is, by law, made by the day treatment program with input from other parties, sometimes including the child's health plan. This creates a dilemma: Although the primary care physician is not equipped for some decisions, removing the decision from the plan is of questionable fairness if the plan is at risk.

Emergency room use and the difficulty faced by providers and plans attempting to change patients' reliance on the emergency room for nonurgent care. In New York, for example, primary care providers complained that hospital emergency room staff were reluctant to adapt their procedures in order to identify demonstration enrollees and contact case managers for authorization, even in nonurgent situations. In light of New York law requiring hospitals to ensure that any person who comes to the emergency department is seen by a physician, the health plans felt they could not deny payment altogether even if they denied authorization for treatment. Instead, they agreed to pay hospitals a triage fee ($20-$40) for any demonstration enrollee who was seen in the emergency room in this situation.

Particularly difficult for States and health plans and providers has been the question of whether and how to serve the disabled and chronically ill under the demonstration. Although States would like to include these populations in the demonstrations, they tend to have established provider relationships, require specialized services, and otherwise pose challenges to case management. Also, health plans and providers, even if they have the necessary expertise, may hesitate to accept risk for such potentially high-cost populations.

Only California, Minnesota, and New York intended to serve Medicaid disabled eligibles under the demonstration; all three encountered difficulties. New York never succeeded in developing a protocol for enrolling the disabled that all parties (the State, MediCap, the health plans and providers, and advocacy groups) would endorse and, as a result, they were never enrolled.

California capitated the Santa Barbara Health Authority for the disabled, as for other eligibility categories. However, on a case-by-case basis, the Health Authority designated as “special class” those individuals who were deemed difficult to case manage; instead, it reimbursed primary care case managers on a fee-for-service basis (rather than by capitation) for their care. Special class recipients included patients with acquired immunodeficiency syndrome (AIDS), the long-term institutionalized, spend-down cases, renal dialysis patients, and others for whom primary care case managers were unwilling to bear risk.

In Minnesota, the State required participating health plans to serve either the aged or the disabled and blind in addition to AFDC recipients. Four of the seven plans elected to serve the disabled, about 60 percent of whom joined Blue Cross and Blue Shield (BC/BS). BC/BS had the largest network of participating physicians and other providers, which allowed beneficiaries to maintain many of their existing provider relationships. However, after 2 years, BC/BS withdrew from the demonstration because of financial losses. Because this affected a large fraction of the disabled, the State decided to return the disabled to fee-for-service Medicaid rather than ask patients to elect a new health plan for the remaining year of the demonstration.

Beneficiary and advocate issues

Medicaid beneficiaries and their advocates expressed a number of concerns about the demonstration projects relating to access and quality of health care. In this section, program design and service delivery issues are addressed, particularly for vulnerable populations, and the results of information on consumer responses collected by the States to date are relayed.

Program design

Features of the demonstrations to which some health and welfare advocates objected included mandatory participation (a feature of all sites except New Jersey), random assignment to health plans or providers for recipients who do not make an election (Minnesota, Missouri, Santa Barbara, and Monterey), and, in mandatory programs, included restrictions on switching providers or case managers.

The opposition to mandatory participation stems from the belief that Medicaid recipients should have freedom of choice in selecting providers and from the view that prepaid health plans or case managers are not appropriate for at least some segments of the Medicaid population (particularly patients with complex medical and/or social problems). The degree of opposition to mandatory enrollment in the demonstrations varied. In Minnesota, where more than one-half the population of the Minneapolis-St. Paul metropolitan area belongs to prepaid plans, there was little questioning of the mandatory nature of the program. In New York, by contrast, where HMOs had little market share in 1982 (although Monroe County was ahead of the rest of the State), there was considerable debate by the State legislature. In Missouri, where there was little familiarity with prepaid health care in 1982, mandatory enrollment was accepted after a minimum of debate.

A second area of concern is random assignment to health plans for recipients who fail to choose a plan in manadatory enrollment programs. In Minnesota, the proportion of demonstration enrollees randomly assigned to health plans average more than 30 percent. From a recipient standpoint, random assignment is problematic because of its implications for access to care, both geographically and in terms of access to the most appropriate providers for individual patient needs. However, in the view of one Medicaid administrator, there is a limit on how much protection should be afforded a recipient who fails to make a choice.

Random assignment to health plans was adopted in Minnesota and Missouri as the means of enrolling recipients who failed to choose a health plan. In Minnesota, recipients had the right to change plans within 60 days of being assigned as well as annually thereafter during open enrollment. Despite such safeguards, some advocates (and providers) have suggested that, instead, recipients should be assigned to health plans or case managers based on the providers they have used in the past, as ascertained from Medicaid claims data. Others would rely on geographic proximity to the recipient's residence. Finally, medical history and patient age have been proposed as a basis for assignment by matching beneficiary needs with provider expertise. In New York, where assignment is not random, a combination of geographic proximity and prior use was employed by the administrative agent, MediCap, Inc. In Minnesota, it was believed that methods of nonrandom assignment applied systematically would be unduly burdensome for State staff and risked generating biased selection among health plans. Missouri considered assignment based on prior use, but in an analysis of patient records, it was found that many recipients had previously used multiple primary care providers, and it was often unclear which providers should be selected for the case management function.

Finally, the fact that recipients in mandatory programs were “locked in” to a particular health plan or case manager was problematic in the eyes of some participants and observers. They argued that many Medicaid recipients, especially those who were assigned, might not seek care during the 30- or 60-day period after enrollment, when they could still exercise their option to switch. Thus, by the time they realized they wanted to switch, they would be locked in and unable to change plans or providers for some fixed period of time (between 6 months and a year). On the other hand, in Santa Barbara, where the restrictions on changing case managers were minimal, case management was impeded by frequent switching among providers. As noted earlier, “doctor-shopping” was one of the practices the demonstrations were designed to discourage.

Special concerns about the demonstration projects were raised with respect to vulnerable populations such as the aged, the physically disabled, and, especially, the chronically mentally ill and the mentally retarded. Because most of these individuals are aged or disabled rather than AFDC recipients, these issues arose less frequently in AFDC-only programs. In some States, political alliances developed between recipient advocacy groups (e.g., for the mentally ill) and the provider associations most threatened by the new initiative (e.g., private psychologists).

Concerns raised in Minnesota, where the program included residents of nursing homes and intermediate care facilities for the mentally retarded as well as the ambulatory aged and disabled living in the community, included the following:

Health plans participating in the demonstration have had little experience with the mentally retarded. Providers who lack the expertise and/or willingness to deal with the mentally retarded may not take the extra time needed to explain issues to patients, leading to confusion, fear, and noncompliance.

Ready access to care for the chronically mentally ill (e.g., those with schizophrenia, manic depression, and personality disorders) is important because of the need for ongoing medication. Patients are often not motivated to stay on medication, and, if access to the physician or the drugstore is inconvenient, the risk of noncompliance is greater.

Many health plans treat principally employed persons and their families and have little experience with the physically disabled. Providers unaccustomed to treating the handicapped may lack the attitudinal awareness and sensitivity needed to work with this group.

In Santa Barbara, many concerns have been raised about the physically disabled. Advocates believe health care services for the disabled are compromised under the program in three main respects:

Access to care, including the inability to self-refer for specialty services and the problem of limited numbers of primary care physicians accepting new Medi-Cal patients. The latter problem forces disabled clients into the county clinics, which may be difficult to reach, are too large and complex, and fail to provide continuity of care. These factors are believed to cause delays in recipient care-seeking.

Appropriateness of case management by primary care physicians, because persons with severe or multiple disabilities often require routine treatment by specialists.

Authorization of durable medical equipment, which advocates say requires so much time that patients have had to do without necessary equipment while authorization requests are being processed. In addition, there have been disagreements about the interpretation of medical necessity criteria with respect to equipment for the disabled (e.g., whether an electric wheelchair is necessary to maintain independent function).

Consumer satisfaction and grievances

Overall, consumer reactions to the demonstrations appeared to range from neutral to positive, although only limited data are available and the definition and reporting of grievances varied widely. In Minnesota, fewer than 20 grievances were filed in 1986, 7 of which went to a formal hearing. In Santa Barbara, by contrast, grievances numbered in the hundreds. This reflects differences in the grievance process (including how aggressively grievances were sought) more than real differences in patient satisfaction; Santa Barbara devoted more effort to assessing and reporting grievances than did any other site.

Most of the demonstrations actively solicited beneficiary feedback. Missouri surveyed participants annually, and New York conducted pre- and post-enrollment surveys. Minnesota developed a common form used by the State, counties, and all the health plans to take telephone complaints; the results will be tabulated and published. Some of the results of the States' own efforts are not yet available. The Missouri survey results, which have been tabulated, suggested that the majority of patients were satisfied with their care. In 1986, 30 percent of respondents said they were more satisfied with care under the demonstration than previously, 56 percent reported no change, and 12 percent were less satisfied. The New York MediCap data also indicate that satisfaction with the Medicaid program in Monroe County was at least as high after prepayment was initiated as before.

Santa Barbara has conduced the most extensive assessment of grievances. In 1985, the most recent year for which detailed data are available, 546 grievances were filed, a 13-percent increase over 1984. The average monthly enrolled population was 20,400 in 1985, implying a grievance rate of about 3 percent. (In 1986, however, the number of grievances dropped to 295 when it was made easier for patients to change case managers.) A breakdown of 1985 grievances by subject is provided in Table 2.

Table 2

Percent distribution of grievances, by type of grievance: Santa Barbara Health Authority, 1985

Type of grievance	Percent of total grievances filed
Total	100
Dissatisfaction with assignment to a case manager (after time to reselect elapsed)	22
Lack of reliable transportation or handicapped accessibility	15
Acceptability of care (dissatisfaction with other than medical aspects of care)	25
Dissatisfaction with medical care	23
Denied requests for treatment authorizations or payment for Medi-Cal covered services	4
Other	11

SOURCE: New Directions for Policy: Personal communication. Washington, D.C. 1988.

Conclusions

The fact that most of the States initiating demonstrations under the HCFA solicitation have sought to convert them to permanent programs testifies to their success in the view of State officials. Furthermore, even some of the problems and failures encountered by certain demonstrations represented learning experiences that had identifiable impact on State policies. At the same time, some of the limitations of the approaches attempted must be recognized.

Although the demonstrations were labeled “procompetitive,” the extent of competition that resulted was limited. First, several of the programs entailed placing counties (e.g., Itasca) or county authorities (e.g., Santa Barbara and Monterey) at risk rather than having plans compete against one another. Under this arrangement, competition could still occur among individual providers, such as primary care physicians, but the evidence that this happened is lacking. This is not to represent a judgment on the merits of the approach; rather, we would simply observe that any resulting cost-containment effects were the result of mechanisms such as physician risk-sharing arrangements or administrative controls that have little to do with competition per se.

Perhaps most important is that, even where multiple prepaid health plans within a given community in theory competed for enrollees, the extent of competition was limited. Minnesota, New York, and Missouri adopted such programs with the intent of replicating some of the competitive dynamics that surround private enrollees. HMOs and other prepaid health plans compete heavily for private enrollees on the basis of price and scope of services (as well as on factors such as quality and beneficiary convenience, e.g., provider location and choice, waiting times for appointments, and attractiveness of the office). However, most private enrollees face premium differentials when they select from among various health plans, whereas Medicaid recipients do not. Furthermore, because Medicaid benefits are comprehensive in many States, plans have little opportunity to improve benefits. Competition can occur, but primarily on the basis of service, such as convenient access to providers and short waiting times. Nonetheless, if plans are reluctant to participate in the program in the first place, even this form of competition may not be apparent.

In addition, the opportunities for cost savings may be limited. HMOs, case management, and other approaches to managed care achieve their savings primarily by constraining utilization, whether by instituting strong utilization review programs or placing financial incentives on providers. However, plans contracting with Medicaid may be at a disadvantage in two respects. First, the rates they pay providers are commonly above those paid by Medicaid fee-for-service programs. Indeed, one of the reasons that providers in some instances encourage the plans with which they are affiliated to participate is to benefit from payment levels that exceed those of the regular Medicaid program. Second, managed care, regardless of who undertakes it, entails some administrative costs that the regular Medicaid program does not incur.

States to date have been most comfortable entering into prepaid arrangements for some population groups and for certain services. The States were most comfortable covering AFDC cash recipients, although there were some notable exceptions. New York covered county Home Relief recipients in addition to AFDC. Minnesota and California served the medically needy and the aged and disabled in addition to AFDC cash beneficiaries. However, both States limited the risk borne by health plans and providers. In Minnesota, the plans were primarily at risk for physician and ancillary services for institutionalized patients, with most routine room and board costs excluded from the capitation and paid directly by the State. In Monterey and Santa Barbara, providers were reimbursed for the care of so-called “special class” recipients, many of whom were disabled, on a fee-for-service basis. Recipients were designated special class on a case-by-case basis; among those included were the long-term institutionalized, AIDS patients, spend-down cases, and renal dialysis patients.

Finally, startup times should also not be underestimated, for both political and technical reasons. Initial enrollment in New York and Minnesota occurred a full 2 years after the start of the demonstrations, although several of the other demonstrations became operational in less than a year. This raises the question of whether a 3-year demonstration is too short in light of both the startup and the winddown times.