Statement of William J. Scanlon,
Director, Health Financing and Public Health Issues,
Health, Education, and Human Services Division, and
Stephen P. Backhus,
Director, Veterans' Affairs and Military Health Care Issues,
Health, Education, and Human Services Division
U.S. General Accounting Office

Testimony Before the Subcommittee on Health
of the House Committee on Ways and Means

Hearing on Medicare Veterans Subvention

July 1, 1999

Mr. Chairman and Members of the Committee:

We are pleased to be here today as you review proposals for a Medicare subvention demonstration for the Department of Veterans Affairs (VA). The stated goal of VA subvention is to provide an alternative for delivering accessible and quality care to certain Medicare-eligible veterans, without increasing the cost to Medicare or to VA.

Several VA subvention proposals resemble in many respects the current Department of Defense (DOD) demonstration. Medicare-eligible military retirees who enroll in the DOD subvention program are able to get Medicare-covered services from DOD. Similar proposals would allow certain Medicare-eligible veterans to use their Medicare benefits at VA facilities. Subvention could allow VA, like DOD, to supplement its funds with Medicare payments. In principle, by paying VA a discounted rate, the Medicare program might save money, so long as it does not pay for services that VA previously would have covered.

Although the DOD and the proposed VA demonstrations are relatively small, full-scale subvention programs could significantly affect the Medicare trust funds and the costs of VA and DOD. The 3-year DOD demonstration involves about 30,000 enrolled retirees and limits Medicare payments to DOD to, at most, $65 million a year. By contrast, a nationwide DOD subvention program could potentially involve substantially more in Medicare payments. In VA, the potential size of a nationwide program may be even greater. There are about 9 million veterans aged 65 and older, and nearly all of them are covered by Medicare.

Favorable outcomes for Medicare, VA, and DOD, as well as military retirees and veterans⁽¹⁾ are not, however, guaranteed. For DOD subvention, the Balanced Budget Act of 1997 (BBA) authorized a large-scale, 3-year demonstration and directed GAO to evaluate the demonstration's effects on access to care, quality, and the cost to DOD and to Medicare. We have recently reported on data quality and payment issues affecting the DOD demonstration and the potential for Medicare overpayments.⁽²⁾ We will be providing you with further interim reports on aspects of the demonstration. Our final results will not, however, be available until several months after the demonstration ends in December 2000.

Our testimony today focuses on a possible VA subvention demonstration and on issues that VA subvention raises. Specifically, we will compare the 1998 House Ways and Means Committee bill on VA subvention with the Senate Finance Committee proposal and discuss the unique characteristics of VA health care that bear on subvention. We will also discuss lessons learned from the design and early implementation of the DOD demonstration that may be relevant to the proposed VA demonstration.

In summary, the 1998 House Ways and Means bill and the current Senate Finance proposal are similar in that both provide for time-limited subvention demonstrations in which Medicare pays VA at a discounted rate to care for veterans who are aged 65 and older and who are covered by Medicare. However, there are also significant differences between the two proposals. For example, the Ways and Means bill includes a permanent program for veterans in rural areas who have low incomes or severe service-connected disabilities, while the Finance proposal would establish two demonstration models--fee-for-service and coordinated (managed) care--for lower priority veterans. Under any proposal, subvention holds several challenges for VA. It will be challenged to attract to a subvention coordinated care program veterans who currently enjoy a generous VA benefits package. VA will also need to strengthen its billing systems to operate a fee-for-service model and will need to ensure that veterans' access to services--whether or not they are in the demonstration--is not reduced. Learning from DOD's experience to date, VA will need sufficient time to implement a subvention demonstration--officials at every DOD site told us that establishing the demonstration was more difficult than they had expected. DOD's experience also shows that VA payment methods must be carefully designed and implemented both to protect the Medicare trust funds and to promote cost consciousness and efficiencies at VA demonstration sites. Finally, as DOD's experience underscores, sound data systems are essential for managing and evaluating a subvention demonstration.

BACKGROUND

Medicare

Most military retirees aged 65 and over are eligible for Medicare--a federally financed health insurance program for the elderly, some disabled people, and people with end-stage kidney disease. Medicare covers about 39 million beneficiaries and spends about $212 billion a year. Benefits include hospital, physician, and other services such as home health and limited skilled nursing facility care. The Health Care Financing Administration (HCFA) administers Medicare and regulates participating providers and health plans.

Medicare was originally set up to reimburse private providers on a fee-for-service basis and to allow Medicare beneficiaries to choose their own providers without restriction. A newer option⁽³⁾ allows Medicare beneficiaries to choose among private, managed care health plans. Currently, 17 percent of beneficiaries use Medicare managed care. In fee-for-service Medicare, beneficiaries must pay a share of the costs for various services. Most Medicare managed care plans have only modest beneficiary cost-sharing, and many offer extra benefits, such as prescription drugs.

VA Health Care

VA has traditionally provided a comprehensive array of health services to veterans with service-connected disabilities or low incomes. Since 1986, VA has also offered health care to higher-income veterans without service-connected disabilities. However, those veterans must make copayments for services. Overall, VA currently registers in its health care system over 15 percent of the total veteran population of 25 million, with the remaining veterans receiving their health care through private or employer health plans or other public programs. Many of the veterans that VA serves also get part of their care from other sources, such as DOD, Medicare, and private insurance. The Administration has requested $17.3 billion for VA medical care in fiscal year 2000. To make up the differences between appropriated funds and projected costs, VA estimates that, by fiscal year 2002, it can derive almost 8 percent of the medical care budget from other sources, such as reimbursements from health insurers and, if subvention is enacted, from Medicare.

Since the early 1990s, VA has shifted its focus from inpatient to outpatient care. At the same time, it implemented many of the principles of coordinated--that is, managed--care, emphasizing primary care, although many veterans use VA for only a portion of their care. In 1995, VA accelerated this transformation by realigning its medical centers and outpatient clinics into 22 service delivery networks and empowering these networks to restructure the delivery of health services.

In 1996, the Congress passed the Veterans' Health Care Eligibility Reform Act that established, for the first time, a system to enroll veterans. Enrollment is, in effect, a registration system for veterans who want to receive care. Currently, registration is continuous--a veteran may choose to register at any time and start receiving services--although VA has the authority to limit the enrollment period if it chooses. The law established seven priority groups, with priority group 1 the highest and priority group 7 the lowest. Priority group 1 includes those veterans with the most severe service-connected disabilities; priority group 7 includes veterans whose incomes and assets exceed a specified level and who do not qualify for VA payments for a service-connected disability. Priority group 7 veterans must agree to make copayments for health services.

Each year, VA determines, on the basis of available resources, which priority groups will be eligible for VA care in the coming year. Currently, VA serves all seven priority categories, but in the future, that will not necessarily be true. Veterans in any of the priority groups are eligible for the VA Uniform Benefits Package, a comprehensive array of services ranging from hospital care to home health.

Veterans remain free to get some or all of their care from other private or public sources, including Medicare. VA, on the other hand, is committed to serving all veterans within the priority groups it has designated for that year, although capacity varies by region.

DOD Health Care

DOD received an appropriation for military health care of almost $16 billion in fiscal year 1999. Of that, an estimated $1.2 billion is spent on the 1.3 million Medicare-eligible military retirees. Under its TRICARE program, DOD provides health benefits to active duty military personnel and retirees,⁽⁴⁾ but most retirees lose their eligibility for comprehensive, DOD-sponsored health coverage at age 65. DOD delivers most of the health care needed by active duty personnel and military retirees through its military hospitals and clinics. DOD gives priority for care at military facilities to active duty personnel and to dependents of active duty personnel and those retirees under 65 who are enrolled in DOD's managed care program. Retirees who turn 65 and become eligible for Medicare can get military care if space is available (called space-available care) after higher priority beneficiaries are treated.⁽⁵⁾ Some military facilities have little or no space-available care.

Since the early 1990s, DOD health care has shifted toward managed care. DOD established its own managed care plan, TRICARE Prime, which uses military providers, supplemented by a network of civilian providers. However, it is not available to retirees aged 65 and over.⁽⁶⁾ TRICARE Prime covers services of military physicians as well as civilian network providers by drawing on DOD's appropriated funds and premiums and copayments charged to some enrollees. In TRICARE Prime, DOD generally organizes the delivery of care on managed care principles--for example, an emphasis on a primary care manager for each enrollee. DOD has gained considerable experience with managed care, but it relies heavily on contractors to conduct marketing, build a network of providers, and perform other critical functions.

DOD Subvention Demonstration

BBA established a 3-year demonstration of Medicare subvention, to start on January 1, 1998, and end on December 31, 2000. Within BBA's guidelines, DOD and HCFA negotiated a "memorandum of agreement." The agreement stated the ways in which HCFA would treat DOD like any other Medicare health plan and the ways in which HCFA would treat it differently. The agreement also spelled out the benefits package and the rules for Medicare's payments to DOD. After DOD and HCFA signed the agreement, they selected six demonstration sites. DOD estimated that it would be able to serve nearly 30,000 of the approximately 125,000 people eligible for both Medicare and military health benefits in these areas.

The subvention demonstration made DOD responsible for creating a DOD-run Medicare managed care organization for elderly retirees. This pilot health plan, which DOD named Senior Prime, is built on DOD's existing managed care model. By enrolling in Senior Prime, Medicare-eligible military retirees obtain priority for services at military facilities--an advantage compared to nonenrollees. Senior Prime's benefit package is "Medicare-plus"--the full Medicare benefits package supplemented by some other benefits, notably prescription drugs.

BBA provides the basic rules by which, under the demonstration, Medicare pays DOD. First, Medicare is to pay DOD the Medicare managed care rate, less several adjustments and a 5-percent discount for each enrollee. Second, in order to receive Medicare payments, DOD must at least match its baseline costs, or level of effort--that is, devote at least the same resources as it did in the recent past to providing care to retirees aged 65 and older. The memorandum of agreement translates these guidelines into a complex payment system. For example, it allows any demonstration site to earn monthly interim payments if its Senior Prime enrollment exceeds a threshold derived from the baseline level of effort. But at the end of the year, DOD can only retain a portion of these payments if that year's costs for the six sites together exceed the baseline level of effort.⁽⁷⁾

PROPOSALS FOR VA DEMONSTRATION DIFFER BUT SHARE KEY FEATURES

Although several proposals for a VA Medicare subvention have been developed recently, our analysis focuses on two: a House Ways and Means Committee bill (H.R. 3828) passed by the House in 1998 and a proposal adopted by the Senate Finance Committee on June 24, 1999. While similar in key respects, the two proposals also differ in several significant ways, including whether a VA subvention would include a fee-for-service model and whether a permanent program--in addition to a demonstration targeting certain veterans--would be established in rural areas for higher priority veterans. The two proposals share certain features, including a managed care model (which the Finance Committee calls "coordinated care") for at least part of the subvention proposal, a demonstration targeting lowest priority veterans, and a cap on annual Medicare payments to VA under the demonstration.

H.R. 3828 (105th Congress)

The House bill is distinctive in authorizing both a permanent subvention program and a demonstration project:

• The permanent subvention program would follow a managed care (or coordinated care) model. It would target VA's higher priority level veterans (for example, people with severe service-related disabilities or low incomes) in rural areas and could be continued indefinitely. It would begin with up to three sites, but more sites could be added after 2003. VA would have to maintain its level of effort--its historical resource commitment--to the targeted group of veterans in the sites. Medicare payments would be capped at $50 million the first year, $75 million the second year, and $100 million in subsequent years. No cap would apply if the program were expanded to more sites, subject to certification by the Department of Health and Human Services' (HHS) Inspector General (IG) that VA could measure its costs in a reasonably reliable and accurate manner.
• By contrast, the demonstration would be limited to veterans in the lowest priority level for VA care at no more than three sites and would deliver services for not more than 3 years. One site would have to be an area previously served by a military health facility shut down in the military base closing process, known as the BRAC (Base Realignment and Closure) process. Unlike the permanent program, no rural sites are required. Medicare payments to VA under the demonstration would be capped at $50 million annually. The bill would allow requiring veterans to pay enrollment fees and copayments that could vary with income.

For both the demonstration and the permanent program, the House bill emphasizes that, if practicable, VA should use its outpatient clinics. However, VA could still contract with private providers and health plans to supply services as needed.

The Senate Finance Proposal

The scope of the Finance Committee proposal⁽⁸⁾ is in some respects narrower--its demonstration is limited to the lowest priority veterans (priority group 7, higher income veterans who mostly lack a service-connected disability). In other respects, it is broader--authorizing a test of two subvention models. The proposal would require VA to establish, first, a coordinated care model of subvention and, a year later, a fee-for-service model. It would authorize a VA subvention demonstration in, at most, eight sites but would require equal numbers of sites for the two models. The proposal would allow up to a year for implementing each model, which would operate for up to 3 years after enrollment started.

Medicare's rules for paying VA would resemble those in the DOD subvention demonstration: To guard against the same VA care being paid for by both VA appropriated funds and Medicare, the proposal would require VA to demonstrate maintenance of its effort on behalf of the demonstration population. HCFA would pay VA for the care of veterans in the demonstration only after VA exceeded its historical spending, or level of effort, for higher-income veterans.

Common Features of the Two Proposed Demonstrations

The House bill and Senate proposal share certain common elements. In each, a VA subvention demonstration would include a managed care (or coordinated care) model and serve certain higher-income⁽⁹⁾ veterans (effectively, priority group 7) who are Medicare beneficiaries

• for a limited time period--3 years,
• in a limited number of locations, and
• in compliance with Medicare rules that HCFA applies to the private sector (although HCFA could waive rules that were inappropriate for VA).

Regarding Medicare payments to VA,

• HCFA would pay VA at 95 percent of the applicable Medicare rate paid to private providers or health plans--less certain exclusions, such as payments for disproportionate share hospitals and graduate medical education;
• HCFA payments to VA would be limited to a predetermined annual amount, such as $50 million; and
• VA must meet its previous level of effort in providing services to Medicare-eligible veterans.

(For a more extensive comparison of the two proposals, see app. I.)

VA DEMONSTRATION WOULD FACE CHALLENGES CONCERNING PARTICIPATION, BILLING, AND ACCESS

A proposed VA demonstration holds several challenges. First, veterans may see no advantage in enrolling in a subvention managed care plan because everyone eligible for the demonstration currently has both VA and Medicare benefits. Second, VA's past difficulties in billing insurance companies suggest that VA may have difficulty billing Medicare for services provided to veterans. Finally, if subvention enrollees prove to be heavy users of VA services, they may crowd out or limit the access of other, higher priority veterans.

For VA, an important issue to consider is whether veterans would enroll in a subvention managed care plan that would not give them significantly more services than they currently receive from VA and that would restrict their freedom to use other providers. Priority group 7 veterans-the only ones eligible for a subvention demonstration-can now obtain all services in VA's Uniform Benefits Package (although not always in a timely manner). Like Medicare, VA benefits cover a broad range, including inpatient and ambulatory medical and surgical care, certain plastic surgery, and durable medical equipment. VA benefits are particularly strong, compared to Medicare, in mental health care, comprehensive rehabilitative care and services, preventive services, and respite care. The VA benefit also--unlike Medicare--covers drugs. Copayments are generally no greater than under Medicare fee-for-service. Additionally, veterans who are eligible for Medicare can also get care from non-VA providers--either under fee-for-service or through a managed care plan--whereas, under subvention, members would be locked out of other Medicare plans and providers. If it needed to make subvention benefits more attractive, VA could either reduce copayments or increase benefits, but these actions would increase VA's costs.

In the future, however, VA benefits, as well as the number of priority groups served, may be reduced. Paradoxically, the less generous the VA package for all veterans, the greater their incentive to participate in the demonstration because that would be the only way they could obtain the full range of VA care. VA is authorized to reduce its Uniform Benefits Package and stop serving lower priority veterans, including priority group 7. VA officials tell us that, due to resource constraints, VA may not serve priority group 7 veterans in the future and may reduce the benefits covered under the benefits package. If this happens, these priority group 7 veterans could only get VA services through a subvention demonstration and, hence, would probably be more likely to join the VA Medicare subvention demonstration.⁽¹⁰⁾ Furthermore, some VA officials have suggested to us that, to give priority group 7 veterans a reason to enroll, it may be necessary to exclude them from VA services--except through the demonstration.

Current proposals for a VA subvention demonstration, such as the Senate Finance Committee's, permit both managed care and fee-for-service sites. Of the two, fee-for-service appears to be easier to implement because it only requires VA to submit claims for covered services to HCFA for payment. It does not require the veteran to join a VA-operated managed care plan and forego access to other providers. However, in the past, VA has had difficulty in collecting from insurance companies because its bills have not had enough detail (for example, diagnosis, service, procedure, and individually identified provider).⁽¹¹⁾ While VA is moving toward a system that will more closely approximate private sector billing procedures, its success remains to be seen.

The greatest concern in a VA subvention program--either coordinated care or fee-for-service--is that subvention enrollees could consume so many services that veterans in higher priority groups would be crowded out or their access to care restricted. This concern is particularly great in the case of VA, both because of its constrained resources and its current policy of not denying care to any veterans. VA's budget has been essentially flat for the last 3 years, and the President's budget proposes the same amount for medical care in fiscal year 2000 as was appropriated to VA for fiscal year 1999. However, VA has not only restructured and moved resources from inpatient to outpatient care; it also increased the number of veterans served and is considering several expensive new initiatives, such as a hepatitis C program. One result has been pressure on resources and, in some areas, increased waiting times for appointments. Furthermore, according to its policy, VA does not deny care to any veteran, although veterans may have to wait longer to obtain the care. In the short term, if subvention absorbed more resources than a medical facility had available, waiting times for appointments would probably increase or care could be limited to certain facilities, which might be inconvenient for some veterans. It is unclear how much of an impact increases in waiting times or other types of decreased access would have on enrollees in the demonstration. VA would probably try to ensure that access was maintained for demonstration participants, since their continued participation increases VA resources.

PROPOSED VA DEMONSTRATION CAN BENEFIT FROM DOD EXPERIENCE

Taking account of DOD's experience in establishing a subvention demonstration could strengthen proposals for a VA demonstration. In particular, DOD experience shows that implementation is difficult and that enough time should be allowed to undertake the numerous steps needed to get a demonstration started. Furthermore, an adequate payment method is essential to protect the Medicare trust funds, and payment rules need to be as simple and straightforward as possible. Finally, accurate and reliable data systems are needed to manage demonstration costs and health care effectively.

A detailed discussion of these issues is in appendix II. The following summarizes the main lessons from DOD's experience.

• Time needed for implementation should be recognized. Officials at every DOD site told us that establishing a Medicare managed care organization was more difficult and required more effort than they had expected. Months into the implementation, they continue to encounter new issues. Even though the sites took 13 to 17 months after the legislation was passed to establish Senior Prime, hindsight suggests that the goals to get it running earlier were unrealistic. If a VA demonstration is authorized, it should have 12 to 18 months to implement its plans for the demonstration; both VA headquarters and sites would need that much time.⁽¹²⁾
• Payment methods need careful design and oversight. In any demonstration of Medicare subvention, adequate payment methods are needed to protect the Medicare trust funds. The DOD demonstration stipulated that Medicare would not pay DOD unless DOD had provided its Medicare-eligible retirees an amount of care exceeding its historical level of effort for these retirees. Under a VA demonstration, a similar requirement would be desirable. An accurate estimate of VA's baseline costs would reduce the chance that Medicare would overpay or underpay VA under a subvention demonstration.⁽¹³⁾

DOD and HCFA also encountered difficulties due to (1) the complexity of the Medicare payment rules for subvention, (2) the definition and measurement of baseline costs, and (3) ambiguity about what sites could earn and whether earnings would be distributed to the sites. As a result of these factors, many DOD site managers and physicians have largely disregarded the uncertain gain in financial resources from possible Medicare payments and have focused primarily on implementation and patient care issues. Consequently, the DOD demonstration may not produce the full savings and efficiencies that are expected from managed care.

DOD's experience can be used in designing a possible VA demonstration. First, payment rules should give VA and its sites greater certainty about their earnings. Second, if a VA demonstration had a level-of-effort requirement, the baseline costs should be for a period as close as possible to the start of the demonstration. This would minimize problems of comparing current and baseline costs. It would also facilitate audits of the data. Third, sites should be informed in advance what proportion (if any) of their Medicare earnings would be retained centrally or regionally.⁽¹⁴⁾

• Accuracy of data systems relies on agency commitment. DOD's experience shows that data systems are a point of vulnerability for a successful and credible program. Inadequate data quality can weaken the management of a demonstration and raise questions about reports of its favorable results. The extent to which data quality would pose an obstacle to a VA demonstration depends in part on how the payment rules are specified. Good data, consistent across sites, would also be needed to manage and evaluate the demonstration. Data quality problems would probably vary by site, with some sites having better data than others. The types of data systems needed would depend in part on the subvention model that is selected. For example, in a fee-for-service model, billing systems are critical. In general, solving data quality problems requires commitment and follow-through of agency management.

In addition, DOD experience suggests that veterans in a potential VA subvention demonstration would benefit if VA were to develop a strategy to inform and assist them with their options after the demonstration ends. Furthermore, as Medicare enrollment in managed care plans is shifting to an annual open season, coordinating enrollment in and termination of the demonstration with Medicare's open season would help demonstration participants.

CONCLUDING OBSERVATIONS

Subvention holds significant potential for giving veterans an additional option for health care coverage, for saving Medicare money, and for giving VA additional funds. However, these favorable outcomes are not guaranteed. We have identified several challenges, based on the particular characteristics of VA as well as the experience of DOD subvention. If a VA subvention demonstration were designed to take account of the issues we have raised, its chance for success would be greater. In particular, for a managed (or coordinated) care demonstration, veterans need to have sufficient incentives-compared to the standard VA benefits-to enroll. For a fee-for-service demonstration, VA needs adequate billing systems to ensure that it receives the money it earns. And, as with any demonstration, it will be important to protect both participants' and other veterans' access to care. DOD's experience with subvention to date shows the importance of sound data systems that consistently and accurately capture financial and workload data. It also underscores the importance of straightforward and easy-to-understand payment rules and a clearly defined level of effort that creates a level playing field for both VA and Medicare.

- - - - -

Mr. Chairman, this concludes our prepared statement. We will be happy to answer any questions that you or other Members of the Committee may have.

GAO Contacts and Acknowledgments

For future contacts regarding this testimony, please call William J. Scanlon at (202) 512-7114 or Stephen P. Backhus at (202) 512-7101. Key contributors to this testimony include Gail MacColl, Jonathan Ratner, Dayna Shah, and Phyllis Thorburn.

COMPARISON OF 1998 WAYS AND MEANS BILL AND 1999 SENATE FINANCE PROPOSAL ON VA SUBVENTION

Table I.1: H.R. 3828 (105th Congress) and Senate Finance Proposal Summary

^aThe provisions of H.R. 3828 were incorporated into H.R. 4567, which passed the House on Oct. 10, 1998.

^bThe text of this bill is not yet available. Our description is based on a summary of a proposal titled Chairman's Mark: The Medicare Subvention Demonstration for Veterans Act of 1999, prepared by the staff of the Senate Committee on Finance, June 24, 1999. The Committee adopted the proposal on that date.

EXPERIENCE IMPLEMENTING DOD SUBVENTION DEMONSTRATION

In implementing the subvention demonstration, DOD and HCFA completed numerous and substantial tasks. DOD sites had to gain familiarity with HCFA regulations and processes, prepare HCFA applications, prepare for and host a HCFA site visit to assess compliance with managed care plan requirements, develop and implement an enrollment process, market the program to potential enrollees, establish a provider network (for care that cannot be provided at the military treatment facilities), assign primary care managers to all enrollees, conduct orientation sessions for new enrollees, and begin service. The national HCFA and DOD offices developed a memorandum of agreement, spelling out program guidelines in broad terms. They also developed payment mechanisms, and translated the BBA requirement that DOD maintain its historical level of effort in serving dual eligibles into a reimbursement formula.

HCFA accelerated review procedures and assigned additional staff so that timelines could be met. But these accomplishments were not without difficulties, and several issues remain that are likely to impact the demonstration's results. These include the extent to which payment rules can be made more understandable and workable and the extent to which DOD can operate successfully and efficiently as a Medicare managed care organization.

Implementation Delayed by Several Factors

In view of the steep learning curve that DOD faced--it started without any Medicare experience--it is not surprising that the demonstration did not start on time. BBA was enacted in August 1997 and authorized a demonstration beginning in January 1998. The first site started providing service in September 1998, and all sites were providing service by January 1999. Officials at all DOD sites emphasized to us that the process of establishing a Medicare managed care organization at their facility was far more complex than they had expected. They noted several issues that caused difficulty during this accelerated startup phase, including the following:

• Delayed notification to sites of their selection for the demonstration.
• Difficulties in learning and adapting to HCFA rules, procedures, and terms for managed care organizations. For example, DOD had to significantly rework grievance and appeals procedures to comply with HCFA requirements.
• Difficulties due to shifts in Medicare requirements. All sites started planning as HCFA was developing the new Medicare managed care regulations to replace the rules for the former risk contract managed care program. Consequently, the sites had to adapt to changed rules when they were published.

Capacity and Enrollment

Sites vary significantly in (1) their capacity for caring for Medicare-eligible retirees, (2) how close enrollment is to capacity, and (3) what fraction of eligibles has enrolled. This variation suggests that potential demand for a subvention program is uncertain. Retirees' enrollment decisions reflect several factors--some, DOD may be able to influence; others, such as the extent of managed care presence in an area, are outside its control.

In establishing their enrollment capacity-which effectively became an enrollment target-some sites were more conservative than others. Sites' assessment of their resources focused on the availability of primary care managers--physicians and other clinicians who both provide primary care and serve as gatekeepers to specialist care. Additionally, the national TRICARE office developed a model to show how many enrollees a site would need to meet its level-of-effort threshold and start receiving increased resources from subvention, and these results were made available to sites. Capacity varied from San Antonio, Texas, the largest site with four hospitals and a capacity of 12,700, to Dover, Delaware, which provides only outpatient care in its military health facility and set its capacity at 1,500.

Many DOD officials and other observers expected that sites would be deluged with applications and would rapidly reach capacity, but this did not happen. One site has reached capacity, but only after several months. Other sites have enrolled between 46 percent and 92 percent of capacity as of the end of June 1999.

As table II.1 shows, there is a four-fold difference in sites' enrollment as a percentage of eligibles in their catchment areas-from 8 percent (San Diego, California) to 36 percent (Keesler, Mississippi). Several factors may explain this variation:

• Enrollment in other Medicare managed care plans varies widely, from one site with a low percentage of eligible enrollees (San Diego)--where nearly 50 percent of dual eligibles are in private Medicare managed care plans--to two sites with higher percentages of enrollees (Keesler and Dover)--where no one is in managed care because no plans are available.
• The availability of military care varies. Several sites emphasized in their marketing that retirees who did not enroll could not count on receiving space-available care. This information might spur retirees who prefer military care to enroll in Senior Prime. At other sites, space-available care was less of an issue. At these sites, prospective enrollees who believe that they can continue to receive space-available care may not see an advantage in enrollment but rather a disadvantage--especially because enrolling in Senior Prime locks them out of other Medicare-paid care.
• Sites may differ in the amount of space-available care they have given in the past and in beneficiaries' satisfaction with that care. These factors could also affect the decision to enroll.
• Some retirees expressed reluctance to enroll because the demonstration is due to end in December 2000. They also noted that they did not get information about how, after the demonstration ends, enrollees would transition back to space-available care, traditional fee-for-service Medicare, or a Medicare managed care organization.

Table II.1: TRICARE Senior Prime Enrollment

Note: Status as of June 21, 1999.
^aIncludes only people who were 65 years old at the beginning of the demonstration.
^bCapacity at the beginning of the demonstration. Does not include capacity for those who turned 65 after the demonstration started.

Managed Care Issues

The subvention demonstration for military retirees aged 65 and over is a new endeavor that highlights challenges for DOD to operate as a Medicare managed care organization. The first is operational--putting in place procedures, organization, and staff to deliver a managed care product to these seniors. The second is economic and organizational--creating the business culture that reconciles delivering services to this illness-prone population with cost-consciousness.

DOD's reliance on contractors (like Foundation Health and Humana) has enabled it to accomplish key managed care tasks. DOD overcame obstacles in launching TRICARE Senior Prime as a managed care organization. Specifically, to establish and run a managed care plan requires infrastructure-the ability to market the plan, enroll members, and recruit, manage, and pay a provider network. In building Senior Prime organizations at the six sites, DOD has benefited from its TRICARE Prime experience, and from its contractors who help with or perform many of these tasks.⁽¹⁵⁾ Sites with well-established TRICARE Prime organizations that had worked with the same contractor for several years seemed to us to have a sizeable advantage in establishing Senior Prime. It is not yet known what effect DOD's extensive use of contractors will have on DOD costs for Senior Prime. But an expanded, permanent subvention program would require establishing and monitoring contractors at many new sites. That would make contractor quality, relationships, and costs a pivotal and uncertain feature of a potential DOD subvention program.

Payment Issues

DOD and HCFA have devised payment rules to meet the statutory requirement that Medicare should pay DOD only after its spending on retirees' care reaches predemonstration levels--that is, after it has met its baseline, or level of effort. These rules have added to the difficulty and the complexity of the demonstration. Furthermore, they have resulted in Medicare payments to DOD not being immediately distributed to the sites. As a result, DOD site managers tend to view DOD appropriations as the sole funding source for all Senior Prime care delivered at military health facilities; the managers are likely to consider Medicare subvention payments as irrelevant to their plans for dealing with capacity bottlenecks or other resource needs in TRICARE Senior Prime.

The demonstration's payment system requires extensive cost and workload data--data that are often problematic and difficult to retrieve and audit. It also involves a complicated sequence of triggers and adjustments for interim and final payments from Medicare to DOD.

Interim payments are made to DOD for care delivered at each site that is above a monthly level-of-effort threshold. A reconciliation after the end of the year to determine final Medicare payments can result in DOD returning a portion of those interim payments if the level of effort for all sites for the entire year is not reached. DOD would also return Medicare payments if data showed that the demonstration population was in better health than that allowed for in the Medicare payment rates, or if payments exceed the statutory cap ($50 million in the first year, $60 million in the second, and $65 million in the third).⁽¹⁶⁾

Because of the potential for adjustments after the close of the year, the payment rules create some uncertainty for DOD. DOD cannot be certain that it will retain all--or even part--of the monthly interim payments at the end of the year. DOD has been slow to distribute interim payments to the sites, in part because some of the money may have to be returned to HCFA. This creates great uncertainty for DOD sites and means that care under subvention is currently paid for with DOD's appropriated funds. The demonstration's payment method differs significantly from the Medicare managed care payment system, in which payments are made at the beginning of the month to cover care delivered during the month.

Based on experience to date with the demonstration, any payment approach for subvention must be even-handed (that is, it should favor neither HCFA nor DOD); straightforward and readily understandable; and prospective (DOD and its sites should receive payment in advance of delivering care to enrollees). The demonstration's payment mechanism, which relies on level of effort, is functional in the short term--although the calculation of level of effort has weaknesses.⁽¹⁷⁾ However, this payment mechanism may not be appropriate over the longer term for an extended or expanded subvention program. Moreover, a credible long-term payment system should start with a zero-based budgeting approach: first, determining the cost to DOD of providing TRICARE Senior Prime care to dual eligibles and then deciding how much care will be provided from DOD's appropriations and how much from Medicare reimbursement.

1. Military retirees are those who have completed a military career and are entitled to retirement pay. Veterans include all who served and who did not receive a dishonorable discharge.

2. Medicare Subvention Demonstration: DOD Data Limitations May Require Adjustment and Raise Broader Concerns (GAO/HEHS-99-39, May 29, 1999).

3. BBA expanded this option to include plans in addition to health maintenance organizations and labeled it "Medicare+Choice."

4. We use "retirees" to refer to military retirees, their dependents, and their survivors.

5. A partial, unofficial exception to this rule occurs at teaching hospitals, where aged retirees with serious, persisting conditions are treated on an ongoing basis, in large measure so that medical residents can be given the clinical experience required.

6. Active duty members of the armed forces receive their health care through TRICARE Prime. Dependents of active duty military can choose among three DOD-run health plans that include TRICARE Prime. Retirees under 65 can pay a premium and "buy in" to TRICARE Prime.

7. These issues are discussed in greater detail in GAO/HEHS-99-39.

8. The text of this bill is not yet available. Our description is based on a summary, prepared by Committee staff for the markup on June 24, 1999, of the proposal contained in the Chairman's Mark. The Committee adopted the proposal.

9. Those who exceed VA's income thresholds for cash benefits. For example, the current threshold for a single veteran without dependents is $22,351.

10. Since many veterans obtain only part of their care from VA, this still might not be sufficient incentive.

11. See VA Medical Care: Increasing Recoveries From Private Health Insurers Will Prove Difficult (GAO/HEHS-98-4, Oct. 17, 1997).

12. The Finance Committee proposal provides a year for start-up and initial implementation of the demonstration. It also would stagger the start of the two models: the fee-for-service model would start a year after the coordinated care model.

13. The payment rules in the DOD demonstration are, at least in principle, adequate for the short term but would be undesirable for a longer-term program. A different payment method, with more understandable rules and viable for the longer term, would be need to be developed if the DOD demonstration were extended.

14. VA calls the regional level a Veterans Integrated Service Network, or VISN.

15. The DOD sites relied on the TRICARE contractors for handling enrollment, claims processing, and network management. They have also, to varying degrees, assisted with the application, site visit, quality assurance, and utilization review areas.

16. The enrollment targets for each site reflect the statutory caps. Consequently, rebates (from DOD to Medicare) as a result of payments exceeding the cap are unlikely.

17. These issues were discussed more fully in Medicare Subvention Demonstration: DOD Data Limitations May Require Adjustment and Raise Broader Concerns (GAO/HEHS-99-39, May 29, 1999).