Demonstrations of alternative delivery systems under Medicare and Medicaid.

The current Administration supports competition as one method of helping to contain escalating costs. Proponents of competition claim many advantages to its implementation, but their claims have yet to be widely tested. Over the past several years, however, the Health Care Financing Administration has supported a number of Medicare and Medicaid demonstrations to yield information on plan participation, marketing, and reimbursement under alternative delivery systems. Much of these data are applicable to the competitive plans being considered by the Administration and Congress. This paper discusses recent findings from these projects.

Introduction

During the 15 years since enactment of the Medicare and Medicaid programs, escalating costs have been paralleled by increasing regulation aimed at containing those costs. Initially, the Nixon Administration was intrigued by health maintenance organizations (HMOs) as a competitive approach to reforming the health care delivery system. Enthusiasm diminished, however, in the wake of provider opposition (Falkson, 1980). The potential for significant Federal HMO contracting and beneficiary enrollment was further limited by many of the provisions of the HMO Act and its regulations. While there has been some opposition to specific regulatory programs, few legislative or health policy analysts argued that regulation itself was the culprit. It was generally accepted that regulation was required because the health care system did not resemble the classic economic free market model (Fuchs. 1980).

Health care costs continued to grow throughout the decade. The 96th Congress responded with the introduction of five bills which provided for greater competition and which largely repudiated the regulatory policies of the past (Enthoven, 1981). Many of these proposals, to various degrees, would reduce or eliminate programs such as health planning and professional standards review organizations. Since the new Congress convened in January 1981, interest in similar bills and in proposals such as Alain Enthoven's Consumer-Choice Health Plan (1978) has grown, and the Administration has indicated its support for the competitive approach to the dilemma of health care costs. Proponents of competition foresee several advantages to widespread implementation:

Consumers, facing economic decisions in the purchase of care, would seek the least costly providers and reduce their use of discretionary services.

Providers (oversupplied already in many specialties) would face more limited demand for their services, thus forcing them to compete for consumers by lowering their prices, increasing their quality, or both.

Knowledge about what makes health care outcomes more efficient and effective would grow due to the need to justify dollars spent as well as to improve quality (Ellwood, Malcolm, and McDonald, 1981).

Multi-billion dollar, third-party payer programs like Medicare and Medicaid would reap much needed savings, since reimbursement for beneficiaries enrolled in competitive health plans would be less than the amounts that would have been paid to fee-for-service providers.

These assumptions remain largely untested except in a few areas of the country. While the Federal Employees Health Benefits Program and a few other systems are often cited as examples of successful competitive approaches for employed persons, there is minimal experience with competitive models for Medicare and Medicaid beneficiaries.

Over the past four to six years, however, the Health Care Financing Administration (HCFA) has supported Medicare and Medicaid demonstration projects and studies which have yielded information about plan participation, marketing, and reimbursement under alternative delivery systems (Trieger, Galblum, and Riley, 1981). Much of the information is applicable to implementation of many of the competitive plans under consideration by the Administration and the Congress. This paper reviews the most recent findings from these projects.

Medicare

Plan Participation and Marketing

Most competition models assume that sufficient numbers of health plans in most areas of the country would exist and choose to serve those beneficiaries who would enroll. They also assume that a sufficient number of beneficiaries could be encouraged to join qualified health plans. Research and demonstrations lend credence to these assumptions but suggest that progress may occur over years rather than months.

Plan Participation

When HCFA released a Request for Proposal (RFP) in 1978 to solicit demonstrations involving prospective risk capitation contracts for health services for Medicare and Medicaid beneficiaries, 29 organizations responded. Reviewers deemed eight proposals in the competitive range.

Seven demonstration projects were ultimately funded, and five now have operational contracts. (See the descriptive charts in Technical Note A.) The majority of operational contractors enroll Medicare and employed populations only. While the existence of these projects does not indicate a nationwide willingness of health plans to assume risk for the elderly, these HMO demonstrations have highlighted key incentives that could be emphasized in encouraging other plans to participate in competitive systems. These incentives include:

attractive reimbursement methodology, with the opportunity to retain savings

additional revenue, leading to overall plan growth

significantly simplified and decreased regulatory and reporting requirements

opportunity to capitalize on future demographic patterns which will include an increasing segment of the nation's population over age 65.

However, the InterStudy demonstration indicates that plan participation may prove difficult to achieve in implementing a competitive model under Medicare in the short run. The InterStudy project is the most approximate example of a Medicare competition model yet tested. Originally, six Minneapolis HMOs agreed to participate in an experiment where InterStudy would serve as the broker. The largest HMO in Minneapolis did not choose to participate from the beginning. During negotiations between InterStudy, the HMOs, and HCFA, two HMOs dropped out because they objected to establishing a fixed premium (actuarially equivalent to coinsurance and deductible), wanting instead to collect the coinsurance and deductible as beneficiaries incurred them. The remaining four plans would not have participated in the demonstration if HCFA had not permitted enrollment to occur with health screening during the period outside the 30-day open enrollment. A further question raised by the InterStudy experience is the necessity of the broker. Without the broker to prepare educational material and manage enrollment, the process could have been confusing. However, the cost of the broker is high. The InterStudy contract, excluding service costs, will cost nearly $1 million over four years.

Marketing

The four operational demonstrations have shown clearly that incentives such as increased benefits and reduced premiums are effective in attracting the elderly to enroll in alternative health plans. Over 25,000 Medicare beneficiaries have enrolled with the participating HMOs. This number represents 32 percent of all Medicare beneficiaries currently enrolled under HMO risk or cost contracts. The demonstrations were not successful in enrolling representative numbers of disabled beneficiaries. In theory, the enrollment incentives are financed under the demonstrations by the savings retained by the HMO (that is, the difference between 95 percent of the Adjusted Average Per Capita Cost (AAPCC) and the adjusted community rate). There are several issues that remain unresolved about the use and impact of such incentives.

Assuming that the AAPCC correctly estimates fee-for-service costs, some health plans may not be able to generate enough savings to finance the additional benefits. Conversely, if the AAPCC as currently calculated is systematically over-estimating fee-for-service costs, then the demonstration savings (attributed to plan efficiency) may be illusory. In this case, additional benefits which were thought to have been financed by the savings are really increased costs to Medicare. Future reductions in the AAPCC in many plan areas would be warranted. Without substantial additional benefits, enrollment response will probably be disappointing.

For example, the Fallon and Kaiser projects, which offered a generous benefit package, enrolled a total of over 10,000 beneficiaries in their first three months of open enrollment. Conversely, three of four Minneapolis HMOs participating in the InterStudy demonstration enrolled a total of 1,000 beneficiaries during their one-month open enrollment in May 1981. Relative to Kaiser and Fallon, these three HMOs offered marginal additional benefits (for example, no copayment on the 21st through 100th day of skilled nursing care or on the 61st through 90th day of hospitalization). The one HMO in the InterStudy project (Share) which does have significant Medicare enrollment was implemented in December 1980, included over 2,000 cost contract conversions, and did minimal health screening the first two months. Table 1 compares the marketing results of Kaiser, Fallon, Marshfield, and the four InterStudy HMOs during one to three months of open enrollment. In evaluating the meaning of these figures, several questions merit consideration:

Table 1

HMO Enrollment Statistics

	Kaiser	Fallon	Marshfield1	InterStudy
				Share2	MedCenter	Nicollet-Eitel	HMO-Minnesota
Eligible Population in Service Area	142,728	56,000	18,000	200,000	200,000	200,000	200,000
Premium	$0–15.813	$7.50	$25.94	$14.95	$27.75	$16.55	$22.85
First One to Three Months Open Enrollment
• Total	6,330	3,600		4,816	212	114	55
• Conversions	1,500	None		2,000	None	None	None
Enrollment as of September 1981
• Total	7,800	5,600	8,863	5,269	602	633	387
• Conversions	1,900	None	None	2,000	None	None	None
• Total Percent Penetration	5	10	49	3	<1	<1	<1

Continuous open enrollment; figure reflects total enrollment as of September 1981.

Continuous high option open enrollment during December, January, and May, with minimal health screening the first two months.

Kaiser offers a choice of four benefit packages, each with a different premium.

What are the effects of health screening?

The InterStudy HMOs are the only demonstrations which were permitted to perform health screening for high option coverage. Low option coverage is available to those who fail the health screen only during one 30-day open enrollment period per year. In addition to potential for favorable risk selection (which would be an incentive for plans to contract), the existence of the health screen may discourage some beneficiaries with a known medical condition from attempting to enroll. The other demonstration projects elected to offer only a high option benefit without health screening.

What are the effects of established physician relationships?

Prior to the demonstration, Marshfield already served 40 percent of the non-Medicare population on a prepaid basis, and Marshfield providers rendered most of the fee-for-service care in the area. Since many Medicare beneficiaries could continue to see their same providers under the demonstration, Marshfield's 49 percent Medicare penetration is not so remarkable, although its premium is high. In the results of a study to determine whether any selection bias occurred in the demonstrations' first year of open enrollment, Marshfield was the only HMO found not to have experienced significant favorable selection (Eggers, 1981). Both Kaiser and Fallon enrollees' pre-enrollment costs were 21 percent lower than their comparison groups. Since the enrollees used less medical care prior to enrolling, it is possible that physician ties prior to enrollment were not as strong as among non-enrollees.

What are the effects of marketing techniques?

While none of the demonstrations used exactly the same marketing strategies, they tested common techniques, such as mailings and public meetings. Kaiser used television advertising for the first time. The relative effectiveness of the various techniques will be assessed by Jurgovan and Blair, Inc. (JBI), HCFA's contractor to evaluate the HMO demonstrations. Although it is evident that some techniques yield a better response than others, the relative success of Kaiser, Fallon, and Marshfield compared to most of the InterStudy HMOs cannot be attributed mainly to the marketing technique used. The most effective way to ensure enrollment is to offer an attractive benefit package for a low premium.

Reimbursement

One of the primary purposes of the demonstrations was to test the feasibility of developing prospective risk capitation reimbursement methodologies and to determine whether the resulting reimbursement systems were acceptable to the HMOs and HCFA. HCFA actuaries developed a procedure for the demonstrations to prospectively determine the AAPCC (Kunkel and Powell, 1981; see Technical Note B for a description of calculation of retrospective and prospective AAPCC). An acceptable reimbursement methodology would not only save money for HCFA and beneficiaries, it would also ensure adequate revenues to the HMO. The demonstrations addressed the related question of the accuracy and usefulness of the AAPCC as a benchmark for setting Medicare rates. Some preliminary answers to these questions can now be offered.

Per Capita Risk Reimbursement Methodology

For the InterStudy and Kaiser projects, HCFA based reimbursement on 95 percent of the AAPCC and applied a ratebook methodology. The ratebook yields an appropriate rate for each age-sex cell in the AAPCC (60 cells per county for the aged and disabled populations). In the other demonstrations, HCFA used 95 percent of the AAPCC as a cap (except for Marshfield, where 99 percent was used) for the reimbursement rate. Fallon and Marshfield derived one rate for all enrollees (except ESRD) based on the adjusted community rate (ACR). Health Central, which was implemented in the fall of 1981, used an actuarial approach. While all three approaches proved workable, the percent of the AAPCC has the advantage of straightforwardness. The ACR and actuarial approaches, because they require more extensive government review and plan justification, are administratively cumbersome. In addition, plans using the ACR have usually estimated their prospective costs to match the cap represented by 95 percent of the AAPCC.

Use of the AAPCC as a Benchmark

Whether the AAPCC is used to derive a percentage rate or as a cap, some argue that it is inappropriate to use fee-for-service experience to measure the efficiency of a restructured competitive system. Unfortunately, no alternatives to the AAPCC concept are readily available for implementation at present, and even the AAPCC, as currently calculated, has several limitations:

The first level at which the AAPCC can be calculated is the county level. If an HMO's service area is not representative of the county in which it is located (a problem likely to occur in counties with large cities or distinct urban and rural sectors), serious errors in the AAPCC calculation may occur. Blue Shield of Massachusetts maintained that suburban Suffolk County costs underestimated what it would have cost to serve the elderly in urban Boston. The project failed for this reason, as well as from the organizational difficulties in starting an HMO.

The AAPCC attempts to control for differences between enrolled and unenrolled beneficiaries by applying underwriting indexes which adjust for age, sex, welfare, and institutional status. The current procedure for collecting institutional data involves a survey of all institutions in the HMO's service area to determine how many beneficiaries are institutionalized. The response rate in the second year of these surveys dropped from 70 to approximately 50 percent, and the procedure is probably too time consuming for national application.

For plans in which the prepaid business will be competing with fee-for-service business in the same organization, it is difficult to set a reimbursement rate lower than the AAPCC because the AAPCC reflects the efficiency of care already being provided by the organization. Marshfield and Kitsap Physicians Service are examples of such organizations. Because Kitsap requested rates 15 percent more than its area cost and at the same time wanted to pay physicians more than reasonable charges, HCFA did not implement the project.

Accuracy of the AAPCC

The critical reimbursement issue that has surfaced, through analysis of the demonstration projects, is the accuracy of the AAPCC. Eggers' study suggests that the current AAPCC underwriting factors may be resulting in inaccurate estimates of expected reimbursement for a given group of individuals. Despite the AAPCC adjustments, Eggers showed that Fallon and Kaiser had a favorable selection as evidenced by their enrollees' use of medical care prior to enrollment. Pre-enrollment inpatient reimbursement was respectively 23 and 25 percent less than the comparison group. Outpatient and physician reimbursement was 13 and 12 percent less. On the other hand, Marshfield's 22 percent greater outpatient and physician utilization appears to have offset a 4 percent favorable selection on inpatient reimbursement. A health status adjustment to the AAPCC may have to be considered. HCFA has recently established an internal work group to conduct further research to improve the AAPCC.

HMO Revenues

Marshfield, and possibly Fallon, are exhibiting evidence of an inability to adequately control hospital utilization. In the first seven months of its demonstration, Marshfield lost approximately $1 million. Total first year losses are expected to reach $2 million. Hospitalization rates were 700 days per 1000 beneficiaries over original estimates. Marshfield estimated that for its next contract year, it would need approximately $110 from HCFA (and $30 from each enrollee). Since HCFA determined that the Marshfield AAPCC was $88, Marshfield estimated that it would require 125 percent of the AAPCC to avert continued losses. HCFA and Marshfield have agreed to a reimbursement method for the next contract year whereby Marshfield will be reimbursed at 99 percent of the AAPCC. HCFA will provide reinsurance to partially cover losses which Marshfield may experience because of hospital utilization.

Marshfield believes that the AAPCC is not accurately reflecting costs in its area. Marshfield has also hypothesized that it has experienced adverse selection due to continuous open enrollment and because the service area, like many other rural areas, is medically underserved. It argues that the removal of barriers to carry through the prepaid program has resulted in an increase in legitimately needed services. Lastly, the demonstration may be experiencing a temporary increase in utilization because people are postponing elective services until they have enrolled in the plan. This phenomenon occurred once before with Marshfield's private program.

The situation at Fallon appears to be less serious but is still of some concern. Fallon indicates that in the first six months of its second year, it lost about $400,000 due to underestimated utilization and cost of hospitalization in its ACR. Fallon's second year ACR plus additional benefits equaled 95 percent of the AAPCC ($120.19) plus the $7.50 premium. During Fallon's first year, reimbursement equaled 91.4 percent of the Area Prevailing Cost and revenues approximately equaled costs.

Another problem affecting demonstration sites' costs was excessive out-of-plan utilization. Despite extensive education efforts by the HMOs, many beneficiaries used out-of-plan providers during the early months of their enrollment. Some HMOs elected to forgive the first occurrence with a warning that subsequent out-of-plan services would have to be paid for by the beneficiary.

Experiences like these are likely to intimidate potential future contractors, particularly if reimbursement is set lower than 95 percent of the AAPCC, or if no provision is made to protect the HMO from at least some portion of institutional risk in the first year of the contract.

Medicaid

Organizing Delivery Systems

Currently, most Medicaid beneficiaries can receive their care under fee-for-service from any provider willing to serve them. To the extent that the beneficiary is unfamiliar with proper patterns of utilization, and within State limits and prior authorization requirements, the care provided to individuals can be quite fragmented and expensive. With only 55 current HMO Medicaid contracts in 17 States, the influence of HMOs is minimal. Almost all competitive models propose to alter this situation by increasing the role of organized delivery systems in providing care to Medicaid beneficiaries. Frequently, these proposals use capitation to stimulate cost-effective management; however, they all involve the HMO principle of assuming responsibility for care for an identified group of individuals.

State and Plan Participation

While alternative delivery systems appear feasible, problems with plan participation, in particular, may slow their establishment and inhibit viability. The demonstrations indicate that willingness of States, provider groups, and plans to participate in alternative organized delivery systems is largely subject to the following factors:

Plans will more willingly become involved if they believe the potential for economic loss is remote.

Plans may not participate if they are required to make major administrative modifications.

States may not be willing or able to make changes in their administrative systems which are small in scope but critical to the successful management of organized systems.

Plans that enroll large numbers of Medicaid beneficiaries may have difficulty in implementing a program for private subscribers in the same organization.

The requirement that enrollment of Medicare or Medicaid beneficiaries cannot exceed 50 percent of total enrollment in prepaid plans may be a significant barrier to providing care in urban areas. Raising the limit to 75 percent, as authorized by the Omnibus Budget Reconciliation Act of 1981, will not significantly reduce the problem. However, eliminating the requirement will require States to more actively monitor quality of care provided in these settings.

The experiences of two projects, Massachusetts Case Management and Multnomah County Project Health, illustrate these conclusions.

The Case Management project was developed on the assumption that a variety of provider groups could be encouraged to assume responsibility for providing care to an enrolled population, both on a fee-for-service and a risk basis. The State of Massachusetts issued a Request for Proposals (RFP) to solicit up to six participants. Four contracts were awarded, none of which involved potential financial loss for the sites. The State was unable to interest any risk-sharing sites until late 1980, when two such contracts were signed. (However, the risk corridors for these two sites are narrow and the enrolled population set at a maximuim of 500 families each.) By the end of the first year at the group practice risk-sharing site, a high disenrollment rate had raised State concerns about contract renewal. This site dropped out of the project in October 1981, mainly because of disagreement over the second year's incentive rate.

Once the four non-risk, fee-for-service sites became operational, various barriers to the success of the program became apparent. Sites were slow to document all referrals for off-site care. The State subsequently installed computer edits to reject off-site claims, but without a concerted site effort to properly document referrals made by their own providers, there was no way to know whether the off-site service was a legitimate referral or out-of-plan. This problem has been most difficult to resolve at sites not oriented toward primary and preventive care. Since the Reconciliation Act authorized the Secretary to approve waivers requested by States to implement case management programs, many States may elect to follow the Massachusetts models.

Project Health provides another example of the issues faced in implementing a multi-provider based health plan system. Project Health was based on county contracts with prepaid health plans from which non-institutionalized medically needy recipients had to choose to receive services. (During the first two years, recipients could also choose a fee-for-service delivery system.) Participating prepaid health plans informed Project Health of their charges for providing a specified benefit package. The project established a beneficiary premium which was dependent upon, but not equal to, the difference between the chosen plan's premiums and those of the lowest cost plan. For the highest income level eligible to participate, the monthly premium for individuals ranged from $0 to $22 in 1979. The Blue Cross/Blue Shield plan, which had the highest costs and premiums, dropped out of the system after three years, claiming that it experienced unfavorable selection. This plan was one of the most frequently chosen by enrollees. Independent evaluators of Project Health observed that a greater proportion of high service users joined the open panel plans; however, this problem was not discovered in the closed panel plans. (For a complete overview of the operation and results of Project Health, see Jurgovan and Blair, 1981.)

Project Health has shown that medically needy beneficiaries are willing and able to share in the cost of prepayment plan premiums and take this obligation seriously. The project collected 80 percent of enrollee debts without any special effort. This willingness to share costs was particularly evident when it was necessary for a beneficiary to maintain an existing provider-patient relationship. That is, beneficiaries were willing to pay higher premiums to join the plan where they could continue to be treated by their family doctors. A survey of Project Health enrollees revealed that the primary reason for choice among participating prepaid plans was to continue existing or prior relationships with providers. The survey also indicated that recipients would have considered price a major choice factor if Project Health had not subsidized most of the differential costs between plans. (Jurgovan and Blair, Inc., 1981).

Project Health did not achieve its initial goal of covering all categories of individuals, including private health insurance subscribers and Medicare beneficiaries. The county was never able to convince the State to include the categorically needy in Project Health. This was partly due to the State's reluctance to turn over responsibility to the county. Coordination of payers in organized delivery systems has also proven difficult in other projects with this goal. The Massachusetts Blue Shield HMO demonstration never progressed much beyond problems with Medicare, but started out planning to include all payers. Health Central had planned to include both Medicare and Medicaid enrollees, but was unable to reach agreement with the State on a Medicaid rate.

One of the primary criticisms leveled against the California Prepaid Health Plan (PHP) program of the early and mid-1970s was that California's rate-setting efforts failed to conform to accepted standards of actuarial analysis. Indeed, a key problem faced by any State when it decides to contract with prepaid plans is how to set per, capita rates for the various aid categories. Building upon earlier reports prepared by the Martin Segal Company, State staff subsequently developed an actuarial rate-setting model based on demographic, utilization, and cost factors. This work was done under a grant called the Prepaid Health Research, Evaluation, and Demonstration (PHRED) project, awarded by the Department of Health and Human Services (then the Department of Health, Education, and Welfare) to assist California in developing model systems for States to use in managing PHP programs. In addition to testing the rate-setting model in the State rate-setting process, PHRED prepared a manual to assist Medicaid agencies with the policy and technical issues of actuarial rate-setting for prepaid Medicaid contracts (Leighton, 1978).

California's rate-setting problems and the education of State staff in actuarial techniques illustrate a significant barrier that inhibits some States in contracting with HMOs. The problem should not be underestimated in designing competitive schemes involving Medicaid eligibles. For example, after the Governor of Massachusetts announced his proposal in April 1981 to restructure the Medicaid program under fixed budget contracts with intermediaries and provider organizations, the State discovered a variety of problems with the incentive reimbursement procedures for its Case Management program. Such technical difficulties, along with legislative and public opposition to sweeping reform, led the State in September 1981 to announce that the fixed budget proposal was being replaced with the more modest goal of expanded “managed care” systems. (See statement of Governor Edward King announcing preliminary Federal approval of the fixed budget proposal for the Medicaid program; see also Knox, Richard A., “Massachusetts Medicaid Plan Sent to Washington,” Boston Globe, April 7, 1981, and Knox, “Medicaid Plan Scaled Down,” Boston Globe, September 23, 1981.)

Marketing

HMOs often cite marketing and enrollment problems as major obstacles to enrolling Medicaid beneficiaries. Many States and HMOs further indicate that standard marketing approaches are ineffective with the Medicaid population. Reasons for this include a lack of direct access for the HMO to eligible populations due to confidentiality restrictions in Title XIX regulations, a high turnover rate in enrollment, a lack of incentive for beneficiaries to enroll, and a lack of information for beneficiaries about the HMO option.

The PHRED project, which is completing the membership studies component of its grant, has shed some light on the comparative costs and benefits of alternative marketing strategies. The demonstration was based on the assumption that the joint eligibility process for welfare and Medicaid affords the best possibility for replacing door-to-door solicitation as a cost-effective method of enrolling the number of recipients necessary for viable contracts. PHRED tested marketing methods for various plans in seven welfare offices in California. The welfare office methods included 1) a printed brochure with no personal explanation, 2) a film, 3) a personal presentation by a county eligibility worker, 4) a similar presentation by a specially trained member of the PHRED staff, and 5) a personal presentation by an HMO sales representative. In addition, literature was mailed to all eligibles in the geographic areas served by the demonstration sites' welfare offices. The plans' current marketing approaches (door-to-door, member referrals, conversions) served as the control.

PHRED's preliminary evaluation of the 5,913 choices made in welfare offices under the demonstration yielded the following initial findings (Owen, 1980):

Marketing in the welfare offices can yield a sufficient enrollment of Medicaid AFDC beneficiaries to permit elimination of door-to-door solicitation without adversely affecting the viability of the health plan contracts.

All methods tested produce enrollments, but some methods are more uniformly productive than others. Results of some seem to reflect local conditions rather than performance of the method itself.

Costs for marketing in the welfare office are lower than the plans' current methods. Welfare office enrollments cost about $9 per person enrolled, whereas door-to-door marketing costs between $45 and $50 per person enrolled.

Medicaid recipients in the welfare offices are able to make informed choices despite the pressure for funds, food, or other immediate needs which bring them to apply for welfare. Of 5,913 presentations, only 41 recipients did not choose.

The Case Management project was not specifically designed to permit comparison of the effectiveness of alternative marketing techniques. All families who enrolled shared in the anticipated savings by receiving a $7 check every month, which, along with guaranteed access to a provider, was expected to attract enrollments. With these selling points, the project used several different methods to inform AFDC families about the program. Methods were tailored to particular sites, depending largely on the sites' willingness to expand enrollment and permit non-site users to enroll. At some sites, concern about developing a welfare image prevented the State from marketing the program more actively. As with PHRED, a successful strategy at one site was not necessarily as productive at another. Initially, Case Management marketing efforts focused on mailings and personal presentations by State employees in the site waiting rooms. Later one site used some public service radio spots and marketing in the welfare office. Direct marketing in the waiting room was more effective than mailings. In preliminary survey results, recipients did not report the $7 check to be of central concern; however, the State and sites always received phone calls when enrollees' checks were issued late.

The effectiveness of other marketing approaches, such as six months' guaranteed eligibility, remain to be evaluated. Preliminary analysis based on limited California data indicates that States would still save money if they offered guaranteed eligibility to their HMO enrollees (Celum, Newacheck, and Showstack, 1981).

For Project Health, Multnomah County assumed a broker approach to manage the program. The broker role involved contract negotiation, enrollment counseling, and screening out high risk cases. The final evaluation of the project indicates that the enrollment counseling was the most successful aspect of the project, with benefits including 1) informed but mandatory consumer choice, 2) cost-effectiveness, and 3) limited possibility for intentional favorable selection by health plans. Survey respondents who were served by Project Health reported more satisfaction than categorically needy persons (Jurgovan and Blair, 1981).

Quality Assurance

Prior to the mid-1970s, States had a great deal of flexibility in signing and monitoring HMO contracts (U.S. Senate, 1978). To ensure that the management deficiencies would not recur, the Federal government awarded the PHRED grant and Congress passed the 1976 amendments to the Social Security Act. These amendments required all Medicaid contractors to meet the requirements of Title XIII for qualified HMOs, with few exceptions. Furthermore, these requirements prohibited a qualified HMO from having more than 50 percent of enrollees whose premiums are paid for by Medicare or Medicaid. The Omnibus Budget Reconciliation Act of 1981 has moved away from these restrictions. States will again be permitted to sign prepaid risk contracts with any organization that can prove its solvency and can make the same services accessible to enrollees as are available to non-enrollees who receive their care in the same area. In addition, the Medicare/Medicaid enrollment limit has been raised to 75 percent. With this trend toward decreased Federal monitoring, and increased pressure on States to save taxpayers' money, it will be important to watch for evidence of serious under-provision of services to enrollees.

The PHRED project developed and tested two approaches that States could adopt to monitor the quality of care provided by HMOs. One system relies completely on computer modules which collect and edit encounter data, screen the data against criteria for quality of care and generate potential exceptions to the criteria (California Department of Health Services, 1981). The second approach relies on the acquisition of selected information from medical records (Marks ). The system reports “passes” and “failures” to the criteria as programmed in a mini-computer which nurse abstractors can use onsite. This system was developed and tested by Kaiser Research Center.

Both of these approaches hold promise in monitoring quality in future alternative delivery systems, either by the providers, the State, or Federal regulators. However, they would also require a willingness to support the cost of their operation.

Conclusion

The results of the demonstrations discussed in this review should be interpreted cautiously. The Medicare demonstrations indicate that Medicare beneficiaries will enroll in alternative delivery systems if benefit packages are attractive. The large enrollments under the restrictive conditions of time-limited demonstration projects are particularly encouraging. Nevertheless, the first year's enrollment data from a multiple choice setting in Minneapolis-St. Paul could indicate that too much choice may not be salutary. Most importantly, the Medicare basis for setting the price of a voucher, the AAPCC, needs to be reviewed to determine if it can be refined to more accurately predict reimbursement levels.

While data on which to base conclusions about Medicaid are more limited, there are clear indications that States will move aggressively to test new competition models as an alternative to reducing benefits. The Reconciliation Act provides States with additional flexibility required to test many features of competition models.

The largest unknown factor at this time is the role of the commercial insurance industry in future competition systems. The demonstrations so far have involved only organized delivery systems, such as HMOs, which can be expected to provide services at less cost than fee-for-service systems. No experience has been gained in contracting with commercial insurance organizations, such as Blue Cross and Aetna, to establish whether they can achieve similar savings of current Medicare and Medicaid costs.

Technical Note A

Fallon Community Health Plan

Disenrollment—405 Over 17-Month Period (107 Deaths)

Financial Operations

Kaiser Health Plan

Additional Benefits—Choice of Options

Premium:

Marshfield Medical Center

Financial Operations

June 1, 1980—Through August 30, 1981 (Not including ESRD)

InterStudy Multiple Choice Program

Minneapolis/St. Paul, Minnesota

—Offer Competing Benefit Package

—Private Membership = 205,000 for All HMOs

Additional Benefits:

Low Option

All plans provide SNF services without prior hospital stay requirement and have expanded home health services

High Option

Plans vary in offering additional benefits; for example, unlimited hospital days, routine physical, hearing and eye exams, prescriptions with $3.50 copayment, foot care

Enrollment

Health Central

Problems:

Financial difficulties in Phase I caused by rapid enrollment of private sector—loss of control over utilization

HCFA insures for losses in first year up to 100 percent of area cost

Technical Note B

Calculation of the Retrospective and Prospective AAPCC

Section 1876 of the Social Security Act defines the “adjusted average per capita cost” as:

the average per capita amount that the Secretary determines (on the basis of actual experience or retrospective actuarial equivalent based on an adequate sample and other information and data in the geographic area served by a health maintenance organization or in a similar area served by the health maintenance organization with appropriate adjustment to assure actuarial equivalence, including adjustments relating to age distribution, sex, race, institutional status, disability status, and any other relevant factors)….

The retrospective AAPCC is calculated in the following steps:

Determine the U.S. per capita cost for the Medicare elderly and the Medicare disabled.

Determine the service area per capita cost by multiplying the U.S. per capita cost by the appropriate geographic adjustments. (The adjustment is a five year average of the ratio of county and national per capita costs.)

Adjust the service area per capita cost by the appropriate age-sex underwriting indexes (ASUI) to obtain HMO-specific costs. These indexes adjust for differences between the service area and HMO for:

—age

—sex

—institutional status

—welfare status

Disability status has been interpreted as eligibility for social security disability, rather than as a health status measure. The Office of the General Counsel has supported this interpretation. Race has not been included because it is difficult to define, and it is not clear that race has an impact on cost.

The prospective AAPCC is calculated in the same way, with the following exceptions:

HCFA projects the U.S. per capita cost with the most recent projections used in preparing annual reports to Congress on the Medicare Trust Funds.

HCFA develops a ratebook, in which the appropriate reimbursement is calculated for each cell for which an age-sex underwriting index has been established. This results in 60 cells for the aged and 60 cells for the disabled populations.

Worcester, Massachusetts	Closed Panel, Federally
	Qualified HMO
Private Membership = 34,000
Reimbursement—Based on ACR		= 91.4 Percent Area Cost in Year 1 = $119.12
		95 Percent AAPCC in Year 2 = $120.19
		95 Percent AAPCC in Year 3 = $144.86
Additional Benefits: —Reduced Deductible and Coinsurance —Preventive Services —Eye exams and One Pair of Eyeglasses —Prescriptions with $1 Copayment (Years 1 and 2); $2 (Year 3) —Unlimited Hospital Days
Premium	—$7.50 per Member (Years 1 and 2); $15 (Year 3)
Marketing	—Dual Choice for Medicare Supplemental Policyholders, Meetings, Mailings
Enrollment	—5,600 Beneficiaries
	Open Enrollment—	February-March 1980
		September 1980-January 1981
		September-November 1981
	10 Percent Penetration

	1980 (April-December)	1981 (Projected)
Total Costs	$2,974,136	$9,195,520
Total Revenue	3,001,348	8,172,160
Net Income (Loss)	27,212	(1,023,360)

Portland, Oregon	Closed Panel Federally Qualified HMO
Private Membership = 220,000
Reimbursement—	95 Percent AAPCC = $97.90 (Year 1); $113.65 (Year 2)

Option 1	Option 2	Option 3
M Plan
No Deductible or Coinsurance
Routine Physicals, Eye Exams
Immunizations
Home Health Care
Outpatient Mental Health Services
M Plan +	M Plan +	M Plan +
Eyeglasses	Dental Plan	Option 1 +
Hearing Aid	Dentures	Option 2
Drugs with $1 Copayment

M Plan	Option 1	Option 2	Option 3
No Premium	$6	$9.81	$15.81
Marketing—	Spot TV Advertising
Enrollment—	7,800 Beneficiaries (¼ GPPP Conversions)
	Enrollment Began in June 1980
	5 Percent Penetration
Disenrollment—	600 Over 10-Month Period (300 Deaths or Moved out of Area)

Marshfield, Wisconsin	Closed Panel Non-Federally
	Qualified HMO
Private Membership = 57,500
Reimbursement—Based on ACR	= 99 Percent AAPCC in Year 1 = $74.18
	= 99 Percent AAPCC in Year 2 = $87.46 Plus Reinsurance Arrangement to Partially Compensate for Losses Due to Excess Hospital Days
Additional Benefits: —No Coinsurance or Deductible —Unlimited Hospital Days —Preventive Services
Premium—	$25.94 per Member (Year 1); $32 (Year 2)
Marketing—	Local Meetings; Continuous Open Enrollment (Year 1); Two One-Month Open Enrollment (Year 2)
Enrollment—	8,863 Beneficiaries and 14 ESRD
	Enrollment Began June 1980
	49 Percent Penetration
Disenrollment—	466 Over 14-Month Period (288 Deaths)

Total Costs	$10,763,153
Total Revenue	9,448,090
Net Income (Loss)	(1,315,063)

InterStudy acts as broker for four HMOs—All HMOs are reimbursed at 95 percent AAPCC	= $148.40 (Year 1)
	$170.63 (Year 2)

HMO	Premium		Enrollees
	Low Option	High Option
HMO Minnesota	$21.75	$22.85	450
Share Health Plan	13.95	14.95	54451
Nicollet/Eitel	12.50	16.55	665
MedCenter Health Plan	15.75	27.75	638

2,000 cost contract conversions in Share.

Lansing, Michigan	Closed Panel Federally Qualified HMO
Private Membership = 23,000
Reimbursement—Based on ACR	= 92 Percent Area Cost in Year 1 = $109.83
	= 89 Percent Area Cost in Year 2 = $130.82
Additional Benefits: —Unlimited Hospital Days —Preventive Services
Premium—	$19.11 per Member (Year 1); $18 (Year 2)
Enrollment (Projected)—	2,200 Medicare Beneficiaries with Potential Medicaid Involvement