back to topAt one time this standing would have been cause of unmixed pride, proof of New York's reputation as the center of advanced medical education. But training physicians has its costs, and as escalating health care expenditures have come under increasing scrutiny, the costs of graduate medical education (GME) have been the subject of considerable debate and discussion. Changes in the Medicaid and Medicare programs, which together pay about $8.5 billion annually for GME-related expenses and the growth of managed care in both the public and the private sectors threaten the status quo. Teaching hospitals are concerned that these changes will compromise their financial condition and undermine their teaching and research missions. Nationally, teaching hospitals enter this period with lower margins and thus less financial strength than other hospitals; in New York City, academic medical centers' margins as a group are barely positive.
Many states including New York, which trains more physicians than any other state are now analyzing their Medicaid expenditures for ways to control growth or even reduce spending in real terms. GME expenditures have come under particular scrutiny. The discussion about the appropriate level of funding for GME has been no less heated because of the lack of agreement about how to define the costs of GME, or even GME itself. GME is intertwined with a number of other important products and projects at academic medical centers and teaching hospitals, most notably research and care for the uninsured. The education of physicians takes place alongside these other activities, and has proven almost impossible to cost out separately.
While debate continues about Medicare funding of GME, important issues of state policy remain. To enhance state policymakers' understanding of Medicaid's role in funding medical education and contributing to the financial viability of institutions with major teaching commitments, the United Hospital Fund, with the support of The Commonwealth Fund, brought together representatives from five states to discuss GME at a summer 1996 meeting. The discussion took as its starting point a consideration of the actual costs of training and the various ways institutions pay for it. It also highlighted a range of policy options and model approaches. This report distills those conversations, and is supplemented by summaries of current practices in eight states. It is our hope that this document will move the discussion forward as debates about the future of Medicaid and Medicare, the proper policy regarding international medical graduates, and the implementation of managed care continue.
This special report is one of a series of contributions on topics of current concern in health care policy produced by the United Hospital Fund. It intersects in many important ways with the Fund's current research and policy analysis related to the uninsured, Medicaid managed care, access to primary care, and health care workforce issues. It is part of the Fund's larger mission to improve health care for all New Yorkers, a mission that has guided the Fund since its founding 119 years ago.
I would like to express our thanks to The Commonwealth Fund for supporting the meeting and this publication and to Ed Salsberg of the Center for Health Workforce Studies, who brought his own extensive experience in state government and in analyzing health workforce issue to this lucid and thoughtful summary and analysis.
James R. Tallon, Jr.
President
United Hospital Fund
In response to the growing policy discussions, the United Hospital Fund, with support from The Commonwealth Fund, brought together representatives of five states during the summer of 1996 to discuss issues and options for financing GME. The five states--Maryland, Massachusetts, Minnesota, New Jersey, and New York--had recently held, or were in the midst of, a public discussion on state financing of GME. With more than a quarter of all the physicians-in-training in the country, these states have been at the forefront of efforts to respond to the impact of the changing health care system on GME and academic health centers. Also at the meeting were a number of individuals knowledgeable about recent policy changes in other states, including Tennessee, which has enacted innovative GME policies.
Over the years, the complex formulas for hospital reimbursement under Medicare and Medicaid in many states have hidden the cost and payments associated with GME. In federal fiscal year 1996, Medicare hospital reimbursement rates included $6.5 billion for GME-related expenses (IOM, 1997), or nearly $64,000 per resident. Medicaid reimbursement rates include an estimated $2 billion per year for GME.
GME at Risk
The evolution of the health care system, particularly increased competition, is leading to fundamental changes in the financing of GME. The emerging marketplace has little tolerance for expenses that are not directly related to patient care, and it is no longer possible to support the higher costs at teaching hospitals through generous and hidden reimbursement policies. Among the more significant developments affecting GME financing are the following:
The Competitive Marketplace
As payers and purchasers compete more aggressively, they are negotiating rates with hospitals that are lower than previous levels, in some cases significantly so. In the competitive marketplace, insurers and those who pay health insurance premiums are less willing to pay for costs not directly related to care for their members or employees. Teaching hospitals believe that their higher costs, caused in part by training physicians, make them less attractive to managed care plans and insurers. They fear that they will be forced to accept rates too low to cover the cost of GME, leading to reduced revenues. With many states moving aggressively to enroll Medicaid beneficiaries in managed care and an expected significant increase in Medicare subscribers in managed care over the next few years, competition and pressure on teaching hospitals are expected to accelerate significantly.
Decreased Hospital Use
In addition to aggressive negotiation of rates, managed care plans and other payers have instituted systems and incentives to decrease hospital use. Since most of the current GME financing system uses add-one to rates paid per day or per discharge, funding for GME will automatically decline as admissions and hospital length of stay decline. If the current system of financing GME is continued, GME costs will be spread over fewer and fewer admissions and days, making the teaching hospitals even less competitive.
Anticipated Medicare GME Cuts
Nationally, Medicare is the largest single source of funds for GME. In 1995, Congress and the Administration came very close to an agreement that would have significantly reduced Medicare GME funding. While numerous proposals for cutbacks have been rejected in the past, concern with the solvency of the Medicare Part A Trust Fund is expected to lead to agreement on Medicare GME reductions in the near future. As this report goes to press, both houses of Congress are considering specific proposals that would reduce Medicare GME funding. These cuts will reduce income to teaching hospitals.
The Unresponsiveness of the Medical Education System to States' Physician Workforce Needs
For many years, government and health policy analysts have identified a need for more primary care practitioners, fewer specialists, a redistribution of physicians to under-served areas, and more training opportunities in ambulatory care sites. While GME funds have not been explicitly tied to these goals, public officials have expressed concern that despite billions of dollars in public support, teaching programs and hospitals have not responded effectively to these needs.
In many states, 50 percent or more of the physicians who train in the state leave after training (Seifer, 1996). This reflects a variety of factors. The marketplace for residency training is national; most medical students select residency programs based on their choice of specialty rather than the geographical location of the program. Since about 25 percent of residents in the United States are graduates of foreign medical schools (JAMA, 1996), some of whom need to return to their native country after training, a certain number of residents will leave the country altogether after training. In addition, many U.S. citizens return to their home states or go to other states with better job opportunities. Residency programs, like medical students, make their selections based on a wide range of criteria; particular attention is given to the applicants' qualifications and potential to be proficient in the specialty. The top priority for residency programs is to train well-qualified physicians and to meet hospital service needs, without regard to where residents practice after training.
Residency program directors argue that they are national and international resources, and therefore, they should not be penalized for training physicians who leave the state. Nevertheless, many state policymakers are not convinced that state taxpayers should pay the bill to train these doctors. In New York State, recent legislation specifically includes in-state retention as one of the goals for GME and for the distribution of one pool of funds. This issue was also raised as part of the public debate over GME funding in Maryland and Ohio. At the national level, some have proposed that foreign countries pay hospitals or residency programs for training their physicians (IOM, 1996; COGME, 1997).
The net result of these changes is major fiscal uncertainly for teaching hospitals in general, and for academic health centers in particular, which train large numbers of residents. These teaching centers, along with state hospital associations, are calling for more explicit state and federal subsidies to replace funds lost in the competitive marketplace. While much of the recent public debate about GME has focused on potential physician surpluses and excess training capacity, the failure to effectively redesign the system for financing GME could lead to an eventual shortage of well-trained physicians, as well as threaten the survival of many of the nation's leading hospitals. The case for support is based on the argument that GME is a "public good."
Defining GME and Estimating Its Costs
GME as a Public Good
"Public good" is an economic term that refers to a good or a service that benefits the public at large and will not be produced at the appropriate level in the private market because of the difficulty in pricing it. The education of physicians serves the community at large, including the future patients of the future physicians and health care facilities, which need well-trained physicians. However, while the community at large benefits, there is no way to charge each of the future beneficiaries. Currently, the costs of education are borne only by those that do the training: teaching hospitals or other training sites -- and it is estimated that training physicians can add 20 to 40 percent to the costs at a hospital (Reuter, 1996; Dobson et al., 1994). There was consensus at the 1996 United Hospital Fund meeting that the training of physicians is a public good and therefore deserving of some form of public support.
The willingness of Congress and most states to continue to include GME in Medicare and Medicaid rates indicates a general acceptance of the concept of GME as a public good. However, even if Medicare and Medicaid pay their "fair share," it is unlikely that insurers and managed care organizations in the competitive marketplace will voluntarily pay higher rates to cover their share of training future physicians. In addition, large numbers of uninsured patients receive care at teaching hospitals; thus, even if all insurers or purchasers were to pay a share for their own subscribers or enrollees, there would still be a gap at most hospitals. Finally, even when there is agreement, it is difficult to put the concept into operation, in part because it is so difficult to define GME.
The Multiple Products of GME
The vast majority of physicians are trained by residency programs sponsored by teaching hospitals; most of these teaching hospitals are tertiary care hospitals providing a wide range of services; and some of these hospitals are academic health centers which provide an even wider range of services and educational activities. Most teaching hospitals, particularly the academic health centers:
Many of these are "joint products," produced simultaneously, making it extremely difficult to separate out or measure the reasonable costs of producing any one of the products alone. For example, residents at most hospitals provide extensive services to uninsured patients and participate in research as part of their training. Since many of these other services and products are also viewed as "public goods," there has been little pressure to develop better approaches to measuring the costs of each of the products. In general, payers have been willing to allow the GME reimbursement umbrella to cover these other products as well. Therefore it is misleading, to a certain extent, to suggest that the $64,000 paid by Medicare per year for each resident is only for training.
While it is extremely difficult to isolate the cost of any one product at an individual facility, it may be possible to estimate the cost of some of them independently. For example, if residents care for the poor as part of their education, it may be possible to determine how many people are being served and what it would cost to care for that population using a model without residents. Unfortunately, very little work has been done in this area, and there is little guidance for states.
Medicare and Medicaid in most states divides GME costs into two major components.
Direct Medical Education (DME)
DME costs include the resident's salary and fringe benefits, faculty salaries and fringe benefits, and administrative overhead costs for the teaching hospital. While resident salaries can be measured relatively directly, there is significant variation in both the actual faculty-to-resident ratio and how costs of faculty and overhead are calculated. Thus, there are great variations in the estimated cost and the actual reimbursement for DME across facilities.
Indirect Medical Education (IME)
IME costs are very difficult to measure. IME is intended to cover the cost inefficiencies created by having residents in a facility. Adding an educational responsibility onto a service organization can reduce productivity. This includes, for example, the costs of extra tests ordered by residents and longer patient stays caused by the extra tests, the inexperience of the residents, and the conflicts between educational goals and service needs. IME costs are also likely to vary by such factors as specialty, year of training, setting of training, and training site characteristics. Several states, like Massachusetts and New Jersey, have decided against support for IME, in part due to the difficulty in measuring it.
The Costs of Training in Ambulatory Care Sites
There is a growing consensus about the need to shift the locus of training from the hospital inpatient to the ambulatory setting as managed care expands and as services are shifted to ambulatory settings. Many residency review committees have recognized this and increased their requirements for clinical training in non-hospital settings, and many programs have begun to offer more rotations in ambulatory settings. This experience is viewed as critical for almost all specialties, not just primary care.
There is little consensus on what ambulatory training costs. However, as on the inpatient side, there appear to be both direct and indirect costs. Direct costs include salaries for the residents (for their time at the ambulatory site), supervising physician costs, and administrative overhead costs. Indirect costs include extra examination rooms, potential loss of productivity, and disruption to the flow of patients. If an ambulatory care facility sponsors its own residency program, there are also costs associated with additional administrative responsibilities, travel to other training sites, and academic instruction.
One of the factors that may contribute to higher costs for training in ambulatory sites is the complexity of providing a comprehensive, high-quality educational experience when residents are dispersed to many diverse sites. Ambulatory care sites tend to be smaller and have a more limited range of patients than a resident might see in a hospital. Residents may need to obtain clinical experiences from more numerous sites. Since residents tend to be more dispersed, the logistics of supervision, monitoring of performance, and faculty development appear to be more complex than when training is concentrated in a hospital.
Reimbursement Payments versus Actual Cost
It is very important to distinguish between reimbursement payments -- that is, the amount included in the state and federal reimbursement rates -- and the actual cost of training a physician. The former can be measured by analyzing the reimbursement formulas.
Numerous organizations and states have attempted to measure how much is being paid out under current reimbursement policies as well as the actual cost for GME (Table 1). In the absence of consensus on either the current cost or a methodology to determine the "reasonable costs" of GME, there are sharp variations in the estimates, including the relationship between DME and IME. This presents a major dilemma to states that are willing to cover the "reasonable" costs of producing physicians. More states provide reimbursement for DME, which can be measured more exactly, than for IME.
| Table 1a: Estimates of the Costs ($) for GME | |||
| DME/Resident | IME/Resident | Total/Resident | |
| Maryland, 1994 | 50,370 | 57,987 | 108,357 |
| Minnesota, FY 1993 | 91,505 | 22,321 | 113,826 |
| Table 1.b: Estimates of the Payments ($) for GME | |||
| DME/Resident | IME/Resident | Total/Resident | |
| Medicare, FFY 1996 (average payments)* |
21,570 | 42,160 | 63,730 |
| Michigan Medicaid Only, 1995 | 20,607 | 21,294 | 41,901 |
| All New York Payers, 1995 (includes Medicare) |
77,500 | 114,600 | 192,100 |
Many states are developing funding streams that provide a fixed level of support for GME unrelated to actual costs. To a certain extent, this avoids a debate on exactly what the reasonable cost is and shifts the public policy question to determining how much the state is willing to invest in GME and how to distribute the available dollars. Thus, in the past two years, Massachusetts, Minnesota, New York, and Tennessee have considered specific pools of funds for GME based on factors other than actual costs; this funding was approved in New York, Tennessee, and Minnesota and rejected in Massachusetts and Maryland.
There are three basic sources of state funds for GME: Medicaid, private payers, and state tax revenues. A state may choose one or all of these sources to support GME. In selecting a source, a state needs to consider its economic and political consequences.
Medicaid
Most states set hospital reimbursement rates for Medicaid patients; this is the most common form of state support for GME. A state can include funding for GME in its Medicaid rates, whether it sets rates prospectively based on historical costs, pays based on retrospective costs, contracts with managed care organizations, or contracts with a limited number of hospitals. In addition to Medicaid rates, some states determine hospital reimbursement policies for workers compensation and no-fault insurance. GME can also be supported through these payment mechanisms.
Thirty-eight of 42 states responding to a 1995 survey indicated that they included DME in their Medicaid reimbursement methodology; 31 of the 42 included IME-related costs. Most states included these costs in rates that were set prospectively; others reimbursed hospitals based on their actual costs, including GME (Plumb and Henderson, 1995).
Funding through Medicaid, whether through hospital payment rates or through capitation, has the distinct benefit to a state of drawing down additional federal matching funds. For example, if a state pays 50 percent of the cost of Medicaid and the federal government pays the other 50 percent, then each state dollar invested in GME brings in an equal amount of outside funding to the state. This is a powerful incentive to continue to try to use Medicaid to finance GME. On the other hand, if Medicaid were to become part of a block grant, the incentive would be neutralized or even reversed; states might have an incentive to reduce the state GME contribution and reallocate the funds to other uses.
Medicaid Managed Care: The GME Carve-out Option
Most states are increasing enrollment of Medicaid recipients in managed care organizations. Under managed care, the state pays the managed care organization a per-patient fee determined in advance; the managed care organization in turn negotiates with hospitals, physicians, and others to determine how much to pay for covered health services. Some states already enroll the majority of Medicaid recipients in such managed care arrangements, and many others are moving in this direction (Rowland and Hanson, 1996).
This trend has major implications for teaching hospitals. States usually determine Medicaid managed care premiums based in part on historic Medicaid costs, including inpatient rates with GME-related costs embedded in those rates. Teaching hospitals have argued that since GME-related costs were included in the original hospital rates for a specific purpose, the value of these adjustments should be "carved out" of the managed care premiums and paid directly to the teaching hospitals. This arrangement assures that dollars originally designated to support GME go to teaching hospitals and are not skimmed off by managed care organizations. It also serves to "level the playing field," enabling teaching hospitals to offer managed care organizations competitive prices because their GME costs are covered through the carve-out. As of June 1997, at least eight states had instituted or passed legislation to institute a Medicaid managed care GME carve-out policy: Michigan, Minnesota, Nebraska, New York, Tennessee, Texas, Utah, and Washington.
Once a state decides to carve-out GME costs, it must also decide how to calculate the amount, how to distribute the funds, and whether to tie distribution to workforce goals. In New York State, which uses a carve-out, when a Medicaid managed care enrollee receives care in a hospital, the hospital submits a parallel bill directly to the state for GME. The amount per discharge reflects historical GME reimbursement per discharge at the facility. In contrast, in Tennessee, all Medicaid GME funds are pooled and the pay-out is linked to the current number of residents and achievement of workforce goals. While the Health Care Financing Administration (HCFA) -- the federal agency that oversees the state Medicaid program (and administers the Medicare program) -- decides on a state-by-state basis whether state policies are consistent with federal rules, HCFA to date has approved federal Medicaid matching funds for states with a Medicaid managed care GME carve-out.
If the carve-out payment mechanism pays hospitals on a per Medicaid discharge basis, it will not make up for drops in admissions, which can be significant when Medicaid recipients enroll in managed care. It also does not take into account the shift of care away from teaching hospitals that may occur if managed care plans limit access to higher cost hospitals. (It is interesting to note that while teaching hospitals often argue that managed care plans do not recognize their teaching costs, in New York after the Medicaid managed care carve-out was instituted, some managed care plans countered that the hospitals were not willing to negotiate rates low enough to compensate for the GME carve-out. Some of the managed care organizations indicated they would shift patients to lower cost hospitals in response.)
Some states may pay premiums to managed care plans that are based on community rates or some other method not related to current or historic Medicaid costs. In these cases, no historical Medicaid GME costs are included in the premium. Nevertheless, it may be possible to add an amount onto the premium to cover GME costs and then carve-out these costs.
Private/Commercial Insurers and Purchasers
Assessments
With the exception of Maryland, states do not set hospital reimbursement rates for private payers; private payers negotiate directly with hospitals. Thus, no state but Maryland can add GME costs to the amount private payers are required to pay hospitals. A state could require insurers and purchasers to add a GME fee onto negotiated rates at teaching hospitals, but managed care organizations would probably negotiate a corresponding decrease in their base rates.
An alternative to adjusting rates is to require all private payers and purchasers to contribute to a GME pool. Under this approach, no individual insurer or purchaser of hospital services is disadvantaged in the competitive marketplace. However, there are a variety of political and legal issues related to taxing insurers as well as the Employee Retirement Income Security Act (ERISA) that would need to be worked out before this approach is implemented. One advantage of an assessment on private payers is that it does not involve direct state tax levy dollars as does using Medicaid funding to support GME. On the other hand, businesses in the state would need to pay higher health insurance premiums, which might lead to cutbacks in employer-sponsored insurance coverage and an increase in the number of the uninsured.
New York is one state that has recently implemented a private payer assessment. The New York Health Care Reform Act of 1996 provides for an assessment on covered lives that will raise $544 million per year from insurers for a GME fund (discussed below). While this is a very significant level of funding, it was approved as part of legislation that deregulated rate setting and allows insurers and purchasers to negotiate rates with hospitals. One reason that the mandate may have been acceptable in New York is because it is part of legislation that will decrease by about $400 million per year the amount that private payers pay for GME. Technically the New York assessment is voluntary; however, payers who do not pay the assessment on covered lives must pay a significant assessment on inpatient rates. Most payers have chosen to pay the assessment on covered lives.
Minnesota also recently approved some limited GME support from an existing fund, supported by an assessment on health care providers.
Market Support
There has been much debate as to whether insurers and purchasers of care are willing to pay the higher cost of teaching hospitals in the absence of government mandates. One reason it is difficult to resolve this debate is that insurers and purchasers usually negotiate a price per discharge or per day of care; they do not generally negotiate the individual components of the price. One recent case study suggests that managed care organizations may be paying 5 to 10 percent more for academic medical centers, due to patient mix and other factors (Gold, 1996). However, in light of studies suggesting that teaching hospitals historically have received on average 38 percent more for comparable services than non-teaching hospitals, teaching hospitals may be feeling the loss of support even though insurers may be paying them more (Dobson et al., 1994).
In addition, it is difficult to generalize about how much insurers and purchasers pay for GME because payment rates may vary depending on the extent of competition between insurers/purchasers and between teaching hospitals in a region. After state rate setting ended in Massachusetts and New Jersey, teaching hospitals, for the most part, were able to negotiate rates that allowed them to continue to train residents. Representatives from those states suggest that initially insurers did not compete aggressively on prices; however, as competition has increased, there has been growing pressure on all hospitals, including teaching hospitals.
The inability to isolate the costs of GME from the costs of the other products of teaching hospitals may contribute to the hesitancy on the part of some insurers and purchasers to pay more for GME. Another factor that has sometimes been cited is the dissatisfaction of managed care organizations with the failure of the medical education establishment to educate more primary care physicians and physicians with a managed care orientation.
Broad-based, Statewide Funding
If GME is a public good, then it is appropriate to try to spread its cost over as wide a base as possible. The widest base is a state's general tax revenues. However, there are clearly political and economic costs of trying to tap into a state's general revenues. This effort is likely to generate an extended debate regarding the benefits of the training to the state, the reasonable cost of training, the need for the subsidy, and how and to whom the subsidies should go. This has been the case in several states that have recently struggled with this issue, including Massachusetts, Minnesota, New Jersey, and Ohio.
To the extent the market will cover some of the higher costs of teaching hospitals, public subsidies can be limited to covering the difference between what the marketplace provides and the costs of GME. In Massachusetts. Minnesota. and New Jersey, the discussion of subsidies focused on helping to cover the share of GME costs that were not otherwise covered by the market, Medicare, and other funding streams. However, documenting the need for subsidies has been difficult. As of mid-1997, only Minnesota had convinced its executive and/or legislature of the need for this funding.
Some states provide subsidies through their support for a state medical school or specific residency programs, such as for family practice or rural training programs. However, it is only recently that states have begun to publicly debate whether to provide general support for GME from general state revenues. From a policy perspective, one benefit of the use of public revenues is that it encourages a link with state health workforce policy goals. If state governments explicitly authorize public funds for GME. payments will be more visible and discussed more openly. Consequently, the public and state legislators are more likely to expect that state needs are being addressed, such as increasing the supply of primary care physicians in general and in underserved urban and rural areas in particular.
Distributing GME Funds
The funds available to support GME -- whether they come from Medicaid only, private payers only taxpayers, or some combination of these parties--can be allocated and distributed in a number of ways. Under Medicare and Medicaid rate modifications, the most common approach to funding GME, it is quite easy and natural to flow the funds directly to the hospital doing the training and receiving the rate adjustment. Although some states such as Massachusetts and New York. have incorporated some incentives in their Medicaid GME funds, such as providing more support for primary care residents and less for subspecialty residents, the basic flow has been to the hospital In fact. GME funds are so interwoven into the complex rate-setting structure in many hospitals that only the chief financial officer actually knows how much the hospital receives for GME.
As states develop new approaches. such as the Medicaid managed care carve-out or using more general revenue sources, it is no longer assumed that funds will flow directly to hospitals In these cases. there are two levels of decisions to be made: First, which organizations should be eligible to receive GME funds? Second. should a distinct GME fund be established?
Which Organizations Should Be Eligible to Receive GME Funding?
Teaching Hospitals
Flowing funds directly to teaching hospitals is most consistent with prior funding methodologies and supports the teaching/patient care/research enterprise. However, concentrating all or even most of the GME funds in teaching hospitals may make it more difficult for funds to reach other sites where future physicians should be trained ambulatory and managed care sites. In the past, hospitals have not generally been willing to share their IME funds with non-hospital sites, and they often view GME funds as hospital funds covering a multitude of activities. As hospitals face more cost pressures, they may be even less willing to share GME dollars with outside sites. On the other hand, as service delivery networks develop (and if payers, especially Medicare, allow reimbursement for ambulatory training), it should be relatively easy mechanically to have funding follow residents to ambulatory care sites.
Another concern about flowing GME funds directly to teaching hospitals is that hospitals' needs do not always coincide with the educational needs of residents or the long-term workforce needs of the community. For example, a hospital may decide to train specialists for which it has an immediate short-term need, regardless of the need for these specialists in the region or nation.
Residency Programs
Since residency programs are responsible for the educational aspects of training and the primary purpose of GME funding is to support this training, a case can be made for funding residency programs directly. However, few residency programs have had full fiscal responsibility as most residents are employed and paid by the hospital. Individual residency programs, like individual hospitals, may also have a difficult time seeing beyond their own needs to the workforce needs of the community at large. And because there are far more residency programs than teaching hospitals, funding them directly might pose a major logistical obstacle. For example, in New York State, which has the greatest number of programs, about 105 hospitals receive funding for GME, but there are more than 1,000 accredited residency programs. A system built around individual programs could be very cumbersome.
Medical Schools
Since residency training is a continuation of the medical education process, medical schools may be the appropriate organizations to be responsible for residency training and to receive GME funding. However, most medical schools have not been extensively involved in GME, which usually occurs outside of the academic institution. In addition, medical schools have had only limited relationships with managed care and ambulatory care sites, which will be central to future training activities. However, these relationships are rapidly developing (Baransky, Jonas, and Etzel, 1996).
Moreover, historically, most medical schools have not responded quickly or easily to public needs such as a greater focus on primary care, service in underserved areas, and a culturally diverse student population reflective of a state's population. While some schools are now trying to respond to these needs, past performance may discourage the channeling of GME dollars to medical schools.
One state that is experimenting with flowing GME dollars directly to medical schools is Tennessee. Tennessee recently established an annual $48 million GME pool from Medicaid funds. Initially, the funds will go to both medical schools and the teaching hospitals; an increasing percentage of the pool will be shifted to the medical schools each year, with all the funds going to the schools after four years. (See Appendix A for additional detail on the Tennessee policies.) This approach may have been acceptable to the hospitals in Tennessee because it reinstated Medicaid funding that had been eliminated and might not have otherwise been restored. This approach may also be more acceptable in settings where there is a strong link between medical schools and their affiliated teaching hospitals.
Ambulatory Care Sites
As residents spend a greater proportion of their time in primary or ambulatory care sites, funds could go directly to health centers, managed care organizations, group practices, or others involved in training residents in these sites. However, since residents generally rotate to a primary care site only a few sessions per week, the ambulatory care site might have to significantly expand its administrative role if it were to assume lead responsibility for residency programs. While many ambulatory care providers may be interested in training some residents, few are likely to want full responsibility.
Nevertheless, the barriers in existing funding streams that discourage ambulatory training should be removed. For example, Medicare hospital DME and IME payments need not exclude time spent by residents in ambulatory care sites outside of the hospital. Similarly, if an ambulatory care site wants to assume full responsibility for a residency program, state GME funding should be as available to it as to hospitals.
GME Consortia
Where developed, GME consortia may be an appropriate recipient of GME funds. GME consortia bring together the organizations that have responsibility for the education and service components of GME. They usually include one or more teaching hospitals, a medical school, and, sometimes, ambulatory care training sites and managed care organizations. Some consortia have no formal authority but provide an opportunity for organizations to discuss issues of common concern. Others have a more formal structure and clearly defined responsibilities. A few, like the Western New York Consortium for Graduate Medical and Dental Education, are the formal sponsors of residency programs and manage and oversee all training. Consortia may also decide the number and mix of training programs in an area.
Consortia may be a vehicle to address some of the challenges facing GME, including shifting more training to ambulatory care sites, preparing physicians for the needs of the managed care environment, and encouraging a mix and number of residents consistent with community needs. However, with a few exceptions, consortia have not shown a willingness to actually receive and distribute GME funding. The difficulty in establishing a consortium is indicated by the experience in New York State, where the development of consortia have been promoted for several years including through grants. Nonetheless, only a few of the consortia to date have been able to overcome competition and control issues to establish functional decision-making structures for workforce planning. Under the new GME pools in New York, consortia can apply for GME funds. Although there is no explicit funding preference, consortia may be in a good position to achieve some of the workforce goals that are linked to one of the GME pools. The New York State statute authorizing GME consortia also specifically exempts state-approved consortia from certain antitrust actions if they collaborate on rational planning, including downsizing of residency training positions. This is particularly important given the state and federal goal of downsizing programs in New York.
The Use of a Dedicated GME Fund
Recent approaches to raising or preserving GME funds, such as the Medicaid managed care carveout, are very amenable to the use of a pooling mechanism. There are a number of distinct benefits to establishing a GME fund. By making the funding more open to public scrutiny, and perhaps making some of the funds competitive, a GME fund focuses attention on how funds are used and facilitates a link with state workforce needs. It also makes it easier to identify how much is being spent and where funding is going, and to flow dollars to ambulatory sites. In general, a fund should make GME more responsive to the state's needs. This is not to imply that a fund or pool is necessary to link funding to workforce goals; it does, however, facilitate the linkage.
For teaching hospitals one of the major benefits of using the pooling mechanism rather than adjustments to patient care rates is that it levels the playing field. Hospital prices do not have to reflect the hidden cost burden of teaching.
The Balanced Budget Act of 1995, which was passed by Congress but vetoed by the President, would have established a national trust fund for GME. It would have included both Medicare DME and IME funds and general revenues. This proposal has the support of a number of national organizations and is likely to be reconsidered in the future (IOM, 1997).
The New York State GME Pool's Professional Education Fund
One example of a recently authorized GME fund is the Professional Education Fund in New York State. The state's 1996 Health Care Reform Act, which replaced the prior state rate-setting system with a more market oriented approach to hospital financing, includes four separate funding streams for GME: Medicaid fee-for-service (GME continues to be an add-on to inpatient rates); Medicaid managed care (GME-related costs are carved out and paid directly to hospitals); and two pools that together comprise a Professional Education Fund ($544 million), financed from an assessment on private payers.
Ninety percent of the Professional Education Fund ($490 million) will be distributed to all teaching hospitals in the state based on a formula that considers a hospital's share of all residents in the state as of July 1995 and a hospital's cost per resident as calculated by Medicare. This removes any incentive to increase the number of residents since the amount a hospital receives will not vary if residency slots are increased or decreased. By using a hospital's cost per resident, an individual facility cost factor is included in the formula. (An alternative approach would be to use a flat price per resident, for example, by dividing total available dollars by total residents. This would benefit hospitals with lower than average costs and disadvantage those with above average costs.) Funds are collected and distributed on a regional basis within the state.
The balance of the pool ($54 million per year) constitutes the GME Reform Incentive Pool. These funds will be distributed based on performance and progress toward the state's physician workforce goals, which include downsizing of residency positions; increasing the percent of residents in primary care specialties; and increasing the number of under-represented minorities in training. After the first year, the goals will also include training residents in ambulatory care sites, producing graduates that enter practice in underserved areas, and improved in-state retention rates for primary care physicians. Participation is voluntary. Hospitals and consortia can apply for a share of their region's pool.
Using GME to Achieve State Policy Goals
Physician Workforce Needs
If states explicitly raise and distribute funds for GME, there is likely to be pressure to tie distribution to workforce and health care goals. There are a series of goals that have been consistently cited by the National Council on Graduate Medical Education (COGME), the New York State COGME, and GME policy analysts. Among the general goals are the following:
1. Addressing State Health Needs and Physician Shortages
The ultimate goal is a workforce that can meet state residents' health care needs, for example, in the areas of primary care, maternal and child health, prevention, geriatrics, environmental health, and rural health.
2. Increasing the Number of Primary Care Physicians
Many states have consistently recognized the need for more primary care physicians, especially to serve high-risk populations. However, there is little foundation for the commonly cited goal that 50 percent of physicians should be in primary care. More appropriate would be a goal based on documented access problems or physician-to-population ratios for each specialty. While more work needs to be done in this area, there are some guidelines based on the recent studies of HMO enrollee use patterns and work by the federal Bureau of Health Professions. As the supply of primary care physicians increases, it will be important to carefully measure need by primary care specialty.
3. Increasing the Number of Physicians in Underserved Areas
A concern for legislatures across the nation is the lack of access to services in inner-city and rural areas. While there are a number of barriers to access, the shortage of practitioners is a major one. States have developed a variety of programs, such as loan forgiveness agreements, service-conditioned scholarships, and start-up grants to encourage practice in underserved areas. Another approach is to provide medical students and residents meaningful training experiences in under-served areas.
4. Increasing Ethnic and Racial Diversity Consistent with the State Population
African-Americans, Latinos/Hispanics, and Native Americans are significantly underrepresented in medicine. In addition to concerns regarding equity, a strong case can be made that the health system will be more effective with a physician workforce that more closely resembles the population. In addition, several studies have documented that minority group members are more likely to practice in underserved areas and to serve minority populations (Komaromy, 1996). The Association of American Medical Colleges (AAMC) and the national COGME have strongly advocated for increases in the number of underrepresented minority group members in medicine. Residency programs, on the other hand, maintain that they are constrained by the mix of graduates from U.S. medical schools. While this is a valid concern, pressure by residency programs on medical schools may help promote an increase in the proportion of minorities entering medicine in the long run.
5. Increasing the Percentage of Residents Remaining in the State after Training
Governors and state legislatures will be more amenable to raising funds to support GME if residency programs in the state meet state needs. This issue may be a dilemma for states with a disproportionate number of medical residents, particularly if those residents are international medical graduates. These states are clearly exporters of physicians; if more of their graduates stay in the state, it could contribute to significant surpluses. Nevertheless, state officials are likely to resist funding programs that do not appear to be meeting state needs.
6. Reducing the Total Number of Residents
This is an appropriate goal for the nation as a whole and for certain states, particularly those that train a disproportionate share of physicians, for example, New Jersey and New York, which train far more residents than graduate from medical schools in their state. There is a general consensus that the nation is facing an excess of physicians. This is caused, in large part, by the number of residents being trained, which exceeds the number of U.S. medical school graduations by nearly a third. At least one national commission has recommended that U.S. medical schools reduce class size in addition to the reduction of IMGs (Pew Health Professions Commission, 1995). Others have suggested that the nation would be better served by reprogramming public dollars for GME to support the redistribution of physicians to underserved inner city and rural areas (Mullan, 1996).
In 1997, HCFA announced approval of a Medicare GME demonstration project in New York designed to reduce the number of residents in training. Under this demonstration, described in Appendix B, hospitals agree to downsize their programs in return for a lag in the reduction of Medicare GME payments. This guarantees Medicare a reduction in GME payments and provides hospitals with funding to help cover the costs of the transition, usually replacement staff. While not all states will need to encourage the downsizing of residency programs, for those that do, providing an incentive for downsizing in the form of temporary support for the transition appears to be a reasonable strategy.
7. Reducing the Number of International Medical School Graduate (IMG) Residents
While often stated as a general policy goal, reducing the number of IMG residents has not been included as a specific goal by states. For example, although reducing the number of IMGs has been an explicit goal of the New York COGME, it is not one of the eligibility criteria for the state GME pools. Because a high percentage of IMGs are concentrated in inner-city hospitals, there is concern that policies targeting IMGs would have a negative impact on access in already underserved areas. It is assumed, however, that by downsizing total residency positions there will be a reduction in IMGs (since most programs prefer U.S. medical school graduates). This approach spreads total reductions across many hospitals, but has a less immediate impact on IMG-dependent hospitals than does reducing funding for IMGs.
A recent report by the Institute of Medicine provides a strong case for decreasing the number of IMGs in training. It makes little sense for U.S. medical schools to reduce admissions because of an anticipated surplus of physicians, only to see an increase in the number of IMGs entering the system at the graduate level. In effect, this substitutes IMGs for U.S. graduates (IOM, 1996).
Supporting Training in Ambulatory Care Settings
As more care shifts out of the hospital, it becomes critical to train physicians in those settings where they are most likely to practice. Many of the residency review committees that oversee residency training have established increased requirements for training in ambulatory care sites. Many states are also promoting this change. The problem is that existing funding streams do not generally cover training in non-hospital settings.
There are two approaches to ambulatory training. The most common is the hospital-based program that includes rotations for several sessions per week to ambulatory settings. The hospital maintains responsibility for the administration of the program. The second approach is where the residency program is sponsored by an ambulatory care site, such as a health center or managed care organization. Because GME financing has been through inpatient rates, funding is very limited for both models. In the case of hospital-sponsored rotations, it is feasible to modify existing funding streams to allow DME and IME reimbursement for the time that residents spend in ambulatory care sites. (This would require regulatory and/or statutory change depending upon the state and program.) This option is not available for programs fully sponsored by ambulatory care sites.
In many states, residents can obtain a license to practice medicine after one year of accredited training. Thus, in some states some residents can bill for the provision of services. However, most payers appear to limit billing by residents. In addition, there are residency review committee and Medicare requirements that limit residents' authority to directly bill while part of a training program.
The private market as well as physician billing can also cover some of the costs of training. To a large extent, this is dependent on the negotiations between the individual sites and the payers and purchasers of services. Few managed care organizations and insurers in a competitive marketplace will want to pay extra unless they can identify a direct benefit to their subscribers/members. However, some managed care organizations may be willing to pay slightly higher fees to providers involved in training, not only because training in ambulatory settings better prepares physicians for their future practice, but also because there are some benefits to the site: for example, the availability of faculty appointments and a link to an academic medical center can facilitate recruitment of physician staff; residents will be exposed to the site and thus may be more willing and prepared to work there after training; and residents provide intellectual challenges to other physician staff and encourage updating of knowledge and skills. However, it is not clear that these benefits outweigh the costs of residents, particularly in a competitive market and at sites caring for many uninsured patients.
Under Medicaid fee-for-service billing, the formula for including GME in the inpatient rates can be adjusted to encourage primary care. An example is the "up-weighting" of primary care residency slots in New York and Massachusetts. While neither state directly flows the funds to the ambulatory sites, the enhanced rates provide teaching hospitals with an incentive to work with ambulatory care sites. While family practice programs are automatically eligible for designation in New York, internal medicine and pediatric programs must demonstrate that their residents will have an appropriate continuity experience at a site that has a structure consistent with comprehensive, coordinated primary care.
Under Medicaid managed care, as noted above, if the GME funds are included in the managed care organization rates, there are no assurances that they will go to GME. If GME funds are carved out of the Medicaid managed care rates and paid directly to the hospital (e.g., on the basis of managed care discharges), then funding is in the hands of the hospital, as it would be in the case of Medicaid fee-for-service reimbursement. However, if these funds are pooled, ambulatory care sites could be allowed to apply for both DME and IME funds.
Care for the Uninsured
In many states, it is the academic health centers with their large complement of residents that care for the poor and uninsured. As the number of uninsured continues to increase and as traditional GME funding streams are constrained, academic health centers are likely to feel financially threatened.
One of the major constraints to reducing the number of residents and Medicare and Medicaid support for GME has been the concern that patients currently cared for by residents, particularly those living in inner cities and lacking insurance, would be left un-served. Indeed, medical residents are major providers of service at many inner-city teaching hospitals, particularly in the northeast. Although there are an increasing number of physicians, physician assistants, and nurse practitioners that could replace residents, most hospitals that use residents to care for the poor do not have a delivery system that can easily use alternative providers. However, with the expansion of managed care, especially for Medicaid and Medicare beneficiaries, many of these teaching hospitals are likely to experience a reduction in patient days and need to redesign their delivery system in any case.
States are likely to be under pressure to help preserve these academic health centers. A state could decide to continue generous GME reimbursement. However, there are benefits to targeting GME funds to meet GME needs and to considering other targeting and distribution methodologies for funds for care for the uninsured. In fact, it may be possible for states to purchase more cost effective care for the poor through non-teaching hospital providers than through academic health centers.
If academic medical centers are an important health and economic resource to a state beyond GME, a state may want to consider separate funding for these centers. Again, this could be targeted and tied to specific goals and/or performance criteria.
If GME funding is the only realistic source of funds for these diverse purposes, then it may be possible to distribute the money in a manner that is more explicitly tied to specific goals. For example if a GME pool or trust fund is established, funds could be distributed based on a formula that gives more credit to hospitals that also care for the poor and/or hospitals that also provide other benefits to the state.
Recommendations
Over the next several years, many states will have an opportunity to reassess their role and responsibilities related to physician workforce planning and teaching hospitals. The increasingly competitive marketplace for health care services is likely to disrupt the historical flow of funds from Medicaid to teaching hospitals, which in turn may force states to decide whether, and how, to continue to support the training of physicians.
These developments provide an opportunity to encourage the production of health care professionals more consistent with state health care needs and to correct past deficiencies in the system. In many states, these deficiencies include costly subsidies; a mismatch between the state's physician needs and the mix of physicians being trained; a failure to train primary care physicians; and an over-reliance on residents to care for the poor.
While it appears that the public expects states to help assure the availability of well-trained physicians, there seems to be little public support for large subsidies and even less support for government planning to determine how many residents should be trained, by whom, and how. This represents a challenge for public policymakers.
Numerous task forces and advisory committees across the nation are struggling to determine the appropriate role for their respective states. Most are still at relatively early stages of deliberation. But as indicated in this paper, a number of innovative ideas have been developed and some are being implemented in different states. Appendix A provides additional details on the current policies in some of the states in the forefront of addressing GME issues.
Most of the solutions being developed reflect a public-private partnership: government provides financial incentives for training consistent with general health and workforce needs in the state; structures for more rational decision making are encouraged; and educational training issues are left to educators and health care providers.
The following observations and recommendations are based on the experience of these states.
Raising GME Funds
Medicaid Fee-for-Service
While fee-for-service reimbursement is becoming a less significant source of funding for GME, it still provides an opportunity to promote public policy goals. Massachusetts and New York provide incentives through their hospital reimbursement methodology to promote training in primary care and other designated specialties and to discourage training in other specialties. This is done by marginal increases in reimbursement for designated specialties and marginal decreases in others.
Medicaid Managed Care Carve-out
Eight states now carve-out payment for GME from their Medicaid managed care rates. These funds can be paid to teaching hospitals and other training sites based at least in part on their performance in addressing state workforce needs. Since hospitals would lose most of these funds in the absence of a carve-out, many may be willing to accept some ties to workforce needs rather than lose all funding.
Private Payer Assessment
Although there are a number of political and legal issues to consider, an assessment that spreads the cost of GME equally over all payers and purchasers would not provide a competitive advantage to any one organization. It would also help to support the future workforce that these payers and purchasers will need.
Tapping State Revenues
If training future physicians is accepted as a public good, then general state revenue sources are an appropriate source of support for GME. However, while several states have considered providing limited support from general revenues, this approach has not been very successful. State tax levy support may be easiest to access for very specific workforce needs, such as training in rural communities or training family physicians at hospitals affiliated with state medical schools.
Distributing GME Funds
Use of a Pooling Mechanism or Trust Funds
The pooling of GME funds provides a number of benefits to states. It allows policymakers and the public to know how many GME dollars are being spent, and who is receiving them. It also greatly facilitates the linking of at least some of the distribution to specific state workforce needs and may facilitate the funding of training at ambulatory care sites.
Criteria for Distribution of Funds
Regardless of how funds are distributed, it is possible to tie at least some of the distribution to performance. While this will be easier if funds are pooled, it can also be done in a fee-for-service environment. There is likely to be great resistance if all GME funds are tied to performance; tying distribution of a limited percentage of the GME dollars to performance is likely to be more politically acceptable.
Encouraging an Appropriate Physician Workforce
GME Consortia
By bringing together key parties involved in education and service delivery, consortia offer an opportunity at the local and regional level to determine the most appropriate mix of physicians in the region. Consortia also have the potential to improve the quality of training by overseeing programs at a number of institutions. They can also help facilitate training outside of the hospital. Ambulatory care sites and managed care organizations should be part of these consortia.
A Statewide Advisory Group
Because of the diversity of interests on issues related to GME, states should consider a process or an organizational structure that allows these interests to participate in the policy discussions. This could be a permanent body, like the State Advisory Graduate Medical Education Council in New Jersey and The New York State Council on Graduate Medical Education, or a temporary task force like the one in Minnesota. A statewide group can also help articulate state priorities and needs. This group should include many diverse interests, including managed care organizations, community-based training sites, private practitioners, and others involved in providing services. This representation is essential to assuring a balanced perspective on physician needs in the state.
Defining Primary Care
Many states have recognized the need for additional primary care practitioners, which they usually define as family practitioners, internists, and pediatricians. However, over the past decade, about two-thirds of the internal medicine residency program physicians and one-third of the pediatric residents have gone on to sub-specialize. Funds for primary care should be limited to those programs that have a certain minimum percentage of their residents actually going into primary care practice.
Data Collection and Analysis on Workforce Needs
Both medical students and academic institutions make decisions based on information about future supply and demand. Unfortunately, there is very little systematic data collection and analysis to guide them. An important role for the state is to help assure the collection and dissemination of basic data on supply and needs. This relatively small investment is likely to have a major return in shaping a physician workforce consistent with state needs.
Conclusion
Until recently, GME policy discussions focused on federal policies and Medicare. Despite numerous efforts, the federal government has not been able to develop a program that effectively uses GME financing policy to encourage a workforce to meet health care needs. It may well be that the best national strategy is one that encourages and supports states to develop and implement GME policies to meet state health care needs. In fact, as demonstrated in this paper, in the absence of a national policy, states have already begun to creatively address these issues. The innovations being debated by states offer new hope for rational policies that balance the conflicting pressures to assure the availability of needed practitioners while limiting government's direct involvement, and to reduce health expenditures while continuing to support teaching hospitals.
Most of the new state strategies for GME financing described in this report encourage change by providing incentives for desirable behavior rather than have government manage the supply of physicians. It will be a challenge to states to encourage a supply and mix of physicians consistent with state needs, without being bureaucratic and overly prescriptive.
Ultimately, it is appropriate for individual states to decide which strategies and which goals make the most sense for them in light of their workforce needs, their academic health centers, and their political environment. In the long term, it may also be appropriate for each state to have a single program that combines Medicare, Medicaid, and state GME funding based on state needs, but this would require a real commitment to states at the federal level.
There is much at stake: an adequate supply of practitioners to serve the residents of each state; the quality of the preparation of the future physician workforce; billions of dollars for GME; and the lives of thousands of physicians-in-training. The innovation on the part of states is an encouraging sign for the future. Maryland
Current State Funding Policies for GME
Maryland is the last remaining "all-payer" state; that is, the state sets the hospital reimbursement rates for all public and private payers. These rates include an upward adjustment for "reasonable" DME costs, based on historical costs. IME costs are factored in through the use of peer groupings for ceilings on costs. The legislature has requested a study on GME financing and Medicaid managed care.
Medicaid Fee-for-Service, Medicaid Managed Care, and Private Payers
Maryland sets rates for all acute care hospital services in the state. The Maryland rate-setting system determines reimbursement rates for each hospital department and total approved revenue based upon the prospectively determined rates and projected volumes. Hospitals are required to charge the approved rates to all patients; all payers (including Medicare, Medicaid, private payers, and managed care organizations) reimburse hospitals based on these rates.
The rates for academic medical centers and community hospitals with teaching programs include a component for GME. At the time of implementation of the all-payer system in Maryland (in the 1970s), the Health Services Cost Review Commission (HSCRC) conducted a full rate review of every hospital under its jurisdiction. The "reasonable costs" of DME were specifically included in the original teaching hospital rate. Reasonable costs are defined as 75 percent of the statewide average per resident DME costs. The GME costs included in the rates were commensurate with the size and structure of the training programs in place at the time of the original full rate review. The costs for DME are adjusted each year for inflation, based on an Inflation Adjustment System; this is a formula based on factor cost inflation, volume and price changes, mark-up and uncompensated care changes, and unusual costs.
Hospitals can request revisions to rates for changes related to GME. However, such revisions would require a full or partial rate application. As a practical matter, hospitals rarely request revisions to rates for changes in DME; they usually fund additional residency slots or costs through existing rates.
The all-payer system in Maryland does not include an explicit adjustment for IME costs because such costs cannot be quantified using normal accounting methods. Instead, IME costs are implicitly recognized through the mechanics of the rate-setting system. The full rate review process relies on comparing a hospital's costs to the average costs of hospitals in its peer group. Thus, establishing base rates for teaching hospitals includes comparing their costs with the costs of other teaching hospitals. IME costs are implicitly financed to the extent that the overall costs of teaching hospitals are higher than those of non-teaching hospitals.
Other Sources of State Funding for GME-Related Activities
Area Health Education Centers. Maryland has two Area Health Education Centers, which are community-based organizations that serve as extensions of health professions training programs in rural and under-served communities. These centers receive about $1 million in state general funds.
Loan Assistance Repayment Program. This program helps repay medical school educational loans for primary care physicians who agree to practice in a state-designated medically under-served area for a minimum of two years. The program is supported by approximately $360,000 from physician licensing fees each year.
Health Manpower Shortage Incentive Grant Program. Maryland also provides about $360,000 per year from physician licensing fees to educational institutions training persons in health occupations determined to be in short supply. Currently family physicians meet this criteria.
Cost/Payment Studies
The HSCRC Advisory Committee supported a 1994 study of the costs of GME in Maryland hospitals. The objective was to determine whether costs were higher in hospitals with teaching programs, and, if so, to assess the amount that might be pooled. DME was calculated directly from information that hospitals submit to the HSCRC, and included residents' salaries and benefits, supervision costs, allocated overhead and other expenses related to GME programs. IME was calculated using a multiproduct cost function from data from 46 hospitals over a 12-year period (1983-1994). The model included case-mix-adjusted inpatient admissions, outpatient visits, and number of interns and residents as output variables, and measures of wage rate and capital price. The study concluded that DME and IME costs per resident were approximately $50,370 and $57,987, respectively, in 1994. (For methodological reasons, Maryland's two academic medical centers were not included in the study.)
Public Policy Process
The Maryland Medical Assistance (Medicaid) program began implementation of its 1115 waiver in July 1997. The state legislature rejected a bill designed to direct the GME-related component of Medicaid managed care rates to teaching hospitals. Instead, the Maryland General Assembly instructed the state Department of Health and Mental Hygiene, in consultation with the HSCRC, to study the feasibility and appropriateness of modifying the state's capitation rate methodology, and to quantify Medicaid's historical contribution to GME at academic centers in the fee-for-service program. The report is due to the legislature by September 1997.
[This information was provided by Sean Cavanaugh, Associate Director, Maryland Health Services Cost Review Commission.]
back to topCurrent State Funding Policies for GME
Massachusetts includes DME costs in its Medicaid hospital patient care rates for fee-for-service admissions. To encourage primary care, an upward adjustment in payments is made for primary care residency programs and a downward adjustment is made for specialty programs.
Medicaid Fee-for-Service
Massachusetts pays hospitals on a per discharge basis. Payments are based on costs for the prior year. This payment includes an explicit adjustment for DME costs but not for IME costs.
The reimbursement system includes an explicit incentive to encourage primary care residency training. The calculation of payments to hospitals for DME adds 33 percent for primary care residents and deducts 20 percent for non-primary care specialists.
Medicaid Managed Care
Massachusetts does not carve out the GME portion of premiums paid to managed care. Managed care plans negotiate payments with teaching hospitals; there are no requirements that the portion of the premium reflecting GME costs be passed on to teaching hospitals.
Other Sources of GME Funds
At the present time, there are no other sources of state funds for GME.
[This information was provided by Elizabeth Greene, Assistant Commissioner, Division of Medical Assistance, Commonwealth of Massachusetts.]
back to topCurrent State Funding Policies for GME
Effective July 1997, GME-related funds are being carved out of Medicaid fee-for-service hospital patient care payments and Medicaid managed care premiums, and redirected to three pools: (1) to teaching hospitals based on their 1995 Medicaid payments ($166 million per year); (2) to hospitals based on the number of primary care residents ($20 million per year); and (3) through grants to GME consortia or partnerships involving a hospital, a university, and a managed care organization ($10 million per year).
The following description relates to both the Medicaid (Title 19) and the Children with Special Health Care Services (CSHCS) programs (Title 5).
Medicaid Fee-for-Service
Medicaid GME policy in Michigan changed significantly on July 1, 1997, when Medicaid hospital reimbursement shifted from a retrospective, cost-based, cost-settled system to a goal-focused, prospective, fixed-payment system.
Under Michigan's prior Medicaid policy, the state included GME costs in its hospital reimbursement system, using a methodology derived from federal Medicare methodology. Medicaid's share of DME costs was identified and paid, subject to a ceiling equal to 80 percent of costs. IME was calculated by a formula negotiated in the settlement of a 1989 lawsuit, which yielded payments that were approximately 55 percent of Medicare IME.
Through the new policy, Michigan sought to structure GME reimbursement policy to "direct educational funding to promote the highest quality services delivered by health professionals and systems in the most effective and efficient manner, consistent with the public policy goals of the State." The policy states: "Payment of public funds which finance Medicaid will be made in the manner designed to achieve the larger managed care objectives of the program, as well as other specific public policy goals, such as the training of appropriate numbers of primary care providers and specialists in specific disciplines, training in rural areas, and training in health fields of particular importance in the treatment of the Medicaid eligible population." Under Michigan's new policy, GME is paid in the following three ways:
1. Historical GME pool. Beginning in July 1997, 77 hospitals will receive the amount paid by Medicaid to each hospital for GME in 1995, with no adjustments made for any changes that occurred after 1995. The annual level of the historic GME pool is $166 million.
2. Primary care pool. $20 million per year will be paid to hospitals based on the number of residents in primary care and the share of Medicaid patients. Specifically, the number of a hospital's full-time equivalent (FTE) primary care residents is adjusted by a factor of one plus the percentage of that hospital's indigent volume. All such numbers are summed, and a percentage of the total is calculated for each hospital. Each hospital will receive that percentage of the $20 million primary care pool.
To qualify for payment from either the historical or primary care pool a hospital must submit a report that includes a listing of each resident by name, year of training, area of training, and FTE time and salary. The hospital must also submit a narrative description of how the funds received from Medicaid are being used in support of the specific public policy goals and priorities of the Medicaid program.
3. Grants for innovations in health professions education. $10 million per year is to be awarded on a competitive basis to innovative programs that support the goals of the program. Emphasis is placed on innovation and training in managed care arrangements. Only consortia consisting of at least a hospital, a university, and a managed care organization are eligible to apply. Grants will be made for multiyear periods, so funding can support a program throughout a three- or four-year residency. The first RFP was issued in January 1997.
Medicaid Managed Care
Until July 1997, capitation rates for managed care organizations included the fee-for-service equivalent of the per capita GME payments actually paid. There was no requirement that the funds be passed on to teaching hospitals. Under the new policy, all funds related to GME are excluded from capitated rates and are now included in the GME pools described above.
Other Sources of State Funding for GME
There is no other funding for GME from Medicaid, special assessments, or general funding.
Cost/Payment Studies
Michigan's Medicaid GME payments have been published for each of the 77 hospitals that received payment in 1995. The total amount was $166.3 million, including the Medicaid ($163.6 million) and Children with Special Health Care Services programs ($2.7 million). The per resident Medicaid payment in 1995 was $41,901 for 3,783.3 FTE residents $20,607 for DME and $21,294 for IME.
In addition, the managed care rates included $45.1 million attributed to GME through the conversion of the fee-for-service experience to capitated rates.
Public Policy Process
The Michigan Department of Community Health's budget for fiscal year 1996-1997 included language authorizing the revision of the state's Medicaid GME policy. An ad hoc advisory group of interested parties and stakeholders convened in fall 1996 to offer counsel and feedback to the agency in the development of new policy. The product of this effort was a policy published in December 1996 and the issuance of an RFP for innovations in health professions education in January 1997. The governor's budget proposal for 1998 recommends a reduction of 10 percent in the historical GME pool. At this writing, the legislature is still deliberating this proposal.
[This information was provided by Vernon Smith, Vice President, Health Management Associates; former Director, Michigan Medical Services Administration.]
back to topCurrent State Funding Policies for GME
The Minnesota legislature authorized a Medical Education and Research Trust Fund in 1996, but delayed funding pending further study. Legislation was passed in 1997 to carve out GME from Medicaid managed care rates beginning in 1999, and to direct these funds to the trust fund. The legislation also appropriated $8.5 million in state funding to the trust fund for distribution in 1998. Hospital patient care rates under Medicaid fee-for-service include funding for GME.
Medicaid Fee-for-Service
Minnesota includes a component for medical education in its hospital patient care reimbursement rates. There are no plans at this time to carve these funds out.
Medicaid Managed Care Legislation was passed in 1997 directing the Minnesota Department of Human Services to carve out the amount "built in" for medical education from the Medicaid managed care capitation rates and to direct these funds to a newly established Medical Education and Research Trust Fund for distribution. Currently, managed care organizations are not required to use these funds (which are in their rates) to support medical education. These medical education funds will go to and be distributed by the trust fund in 1999 under the current legislation. The state will study whether or not to carve out medical education money from the state-funded MinnesotaCare program.
Private Payer Assessment
A private payer assessment was considered by the Medical Education and Research Cost (MERC) Advisory Committee, but rejected because the assessment could not include the self-insured plans. One of the goals of the MERC Advisory Committee was to look for funding from a broad-based source. Without the self-insured plans, this form of funding could not be considered to be broadly based.
State General Revenues/Special Funds for GME Support
The 1997 legislation authorizes $8.5 million in fiscal year 1998. Five million dollars will come from state general revenues. The remaining $3.5 million will come from the Health Care Access Fund, which is supported by a state health care provider tax. This funding represents new money for medical education. "Sponsoring institutions" will be eligible to apply on behalf of their accredited programs and will also be responsible for distributing the funds to the "sites incurring the cost of medical education." A report will be required at the end of the fiscal year documenting that the distribution was made appropriately. The amount per program would be based on a formula using the average clinical cost per trainee (by provider type, i.e., physician, dentist, nurse, physician assistant, or pharmacist), multiplied by the number of trainees in each program, multiplied by a uniform percentage of costs. The total disbursement must equal the amount available for education in the trust fund for that year. In 1998, the Department of Health will request funding for the trust fund for 1999.
State Funding for Specific GME-Related Activities
Minnesota continues to fund programs and activities for both medical students and residents. These include (based on the state's 1993 budget): $26 million to the University of Minnesota Medical School for "general appropriation for instruction"; $432,000 for the University of Minnesota's Primary Care Physician Training Initiative; $889,000 for the University of Minnesota's Rural Physician Associate Program for third-year medical students; $682,000 to the Mayo Medical Foundation for students who are Minnesota residents; $2 million for the University of Minnesota's Graduate Nursing Program; $1.8 million for the University of Minnesota's Primary Care Physicians Initiative; $10.8 million to the University of Minnesota for a "hospital education offset"; $274,000 to the Mayo Medical Foundation for grants to family practice residents; $96,000 for a rural physician loan repayment program; $41,000 for a loan repayment program for mid-level rural practitioners and nurses who practice in nursing homes or intermediate care facilities. The total state appropriation for these programs was approximately $43 million in fiscal year 1993.
Cost/Payment Study
The MERC Advisory Committee has completed a preliminary estimate of the costs of medical education in Minnesota. Based on 1993 data, the estimate is considered a "work-in-progress." This first study found that the estimated costs of medical education exceeded estimated revenues by approximately $37 million in 1993. The MERC Advisory Committee concluded that teaching institutions could cover approximately 75 percent of those costs from patient care revenues, but that this would become more difficult as competition intensified. Based on this analysis, the Department of Health made a request for new medical education funding.
Public Policy Process
In early 1996, the Minnesota Legislature established the Medical Education and Research Trust Fund based on recommendations of the Commissioner of Health and advice from the MERC Advisory Committee. The legislature declined to fund the trust fund at that time, but mandated a study on the most appropriate funding source and mechanism for distribution. In response to that mandate, seven recommendations were presented to the legislature in December 1996. Two of the recommendations addressed the funding source issues; the other five dealt with the mechanism for distribution. The recommendations were accepted as described above in "Current State Funding Policies for GME."
The MERC Advisory Committee provides advice to the Minnesota Commissioner of Health on issues related to medical education, and oversees the MERC project.
Recent reports and papers from the Minnesota Department of Health include:
back to topCurrent State Funding Policies for GME
Effective October 1996, New Jersey instituted a policy of Medicaid hospital reimbursement rate adjustments for DME and IME based on the Medicare GME methodology. This system replaced an earlier one that had provided a higher level of reimbursement for Medicaid.
Medicaid Fee-for-Service
Inpatient rates include a specific component for GME. Effective October 1, 1996, New Jersey is using Medicare methodology to calculate DME and IME with two minor exceptions: New Jersey applies the Medicare methodology to major teaching hospitals (defined as hospitals with 45 or more GME trainees) only, and allocates the dollars thus calculated to all teaching hospitals, regardless of the number of GME trainees, based on weighted FTE numbers and Medicare bed days.
In 1997, approximately $38 million was distributed to hospitals on an interim basis. This represents a reduction in GME payment to hospitals, compared to the previous methodology. The previous methodology was used by the state Department of Health to set hospital rates prior to 1992, and was used by Medicaid through September 30, 1996.
Medicaid Managed Care
The Medicaid capitated rates include the GME component previously accounted in the fee-for-service rates. There is no requirement, however, that managed care plans apply these funds to medical education; and there is anecdotal evidence that dollars are not being passed on to teaching hospitals. This could become a major problem as the AFDC population moved into managed care in 1996, and other Medicaid patients are scheduled to move into managed care in early 1998.
Other Sources of State Support for GME-Related Activities
The New Jersey Primary Care Physician and Dentist Loan Redemption Program encourages newly trained physicians and dentists to practice in under-served areas by reducing some of their training debt. Awards of up to $70,000 are available in return for practicing for two to four years in a federal- or state-designated under-served area. The state contributes approximately $1 million per year to this program.
Cost/Payment Studies
There are no cost/payment studies currently in progress.
Public Policy Process
New Jersey's Advisory Graduate Medical Education Council is an independent entity mandated in statute to advise the Commissioner of Education concerning GME issues. The council also mandated to maintain a database of GME programs and trainees. Its membership is broadly representative, and includes, for example, a representative from the Department of Health and Senior Services, the New Jersey Hospital Association, several of the teaching hospitals, and the public sector.
The New Jersey Hospital Association and University Health Systems-New Jersey, a voluntary coalition of teaching hospitals affiliated with the University of Medicine and Dentistry of New Jersey, cofounded the Joint Teaching Hospital Forum in late 1995 to address GME policy and advocacy issues with one voice. Issues are brought back to the full memberships of the two organizations for ratification.
Recent or anticipated reports include:
back to topCurrent State Funding Policies for GME
Effective January 1997, New York State instituted a new system of financing for GME. This system includes an assessment on insurers and purchasers of care to support an annual $544 million Professional Education Fund. Ninety percent of these funds are distributed based on historical resident counts and cost, with the remaining 10 percent tied to performance related to state GME policy goals. In addition, Medicaid fee-for-service hospital reimbursement includes an adjustment for DME and IME costs with incentives for primary care. GME-associated costs are also carved out of Medicaid managed care rates and paid directly to teaching hospitals.
Medicaid Fee-for-Service
New York adjusts Medicaid inpatient rates to include costs associated with DME and IME. Adjustments are based on historical costs at each facility, trended forward to the current rate-setting year to reflect inflation in hospital costs as determined by a panel of economic advisors.
IME is weighted to encourage primary care training: in the calculation of reimbursement rates for each facility, primary care programs meeting state criteria are "up-weighted" to 1.5 and subspecialty residency programs are "down-weighted" to 0.9. Primary care is defined as (1) all family practice programs; and (2) internal medicine programs and pediatric programs, and combined medicine/pediatric programs that apply and document an appropriate primary care curriculum and training site. In addition, programs in obstetrics and gynecology can apply to be upweighted to 1.1.
Medicaid Managed Care
Beginning in 1996, New York State carved out of its Medicaid managed care rates the amount that was estimated to be associated with GME, based on historical payment levels under the state's fee-for-service hospital rate-setting system. This amount is paid directly to teaching hospitals for each Medicaid managed care discharge from the hospital. Payments are tied to current Medicaid managed care discharges; no adjustments are made for any reduced Medicaid admissions or shifts away from teaching hospitals.
Private Payer Assessment
Beginning in 1997, all insurers and purchasers of care must pay an assessment on each covered life or on each hospital payment. The choice is up to the insurer/purchaser. This assessment is designed to yield an annual $544 million to the Professional Education Fund. Ninety percent of these funds ($490 million) are paid out to hospitals based on their share of residents in 1995 and their GME costs per resident in 1995. Because these payments are tied completely to historical counts and costs, there is no incentive to increase the number of residents. Individual hospitals or GME consortia are eligible for this funding.
Ten percent of the Professional Education Fund ($54 million per year) is allocated to the GME Reform Incentive Pool. Hospitals and GME consortia can apply for these funds. Distribution is based on performance in achieving specific state health workforce goals, including: downsizing residency positions increasing the percentage of residents in primary care specialties; and increasing the percentage of under-represented minorities. Other factors include the quality of the training programs, the percentage of residents training in under-served areas and going into practice in those areas, and the percentage of primary care residents remaining in the state. First distributions are expected in the fall of 1997.
State Funding for GME-Related Activities and Programs
New York State provides more than $7 million per year to a variety of programs, including primary care resident and physician loan repayment (in return for a commitment to practice in an under-served area); scholarships for service as primary care nurse practitioners and physicians, and grants to increase minorities in medicine. The state also has authority to approve GME consortia and to exempt approved consortia from antitrust actions related to physician workforce planning.
Cost/Payment Studies
In 1996, as part of the discussion of the legislation revising hospital financing and GME, the New York State Department of Health estimated the current Medicaid, private payer, and Medicare payments related to GME in the state. The Department of Health estimated that total payments from all sources for GME were approximately $2.9 billion in 1995, or $192,000 annually for each resident.
Public Policy Process
Recent Legislation
New York State's Health Care Reform Act of 1996 ended the state's hospital rate-setting authority for private payers and insurers and established the new GME pools described above, in "Current State Funding Policies for GME."
The New York State Council of Graduate Medical Education
Originally established by gubernatorial executive order in 1987, the council advises the executive and legislative branches on GME issues, contributing to the development of many of the new GME policies. The council, consisting of leaders from academic medical centers, medical schools, teaching hospitals, and other institutions, is assisting in the implementation of the new policies and programs.
[This information was provided by Thomas Burke, Executive Director, New York State Council on Graduate Medical Education.]
back to topCurrent State Funding Policies for GME
Under TennCare, all Medicaid patients are enrolled in managed care plans. Initially, this led to the elimination of all GME funding from Medicaid in early 1995. Later in that year, a new Medicaid GME funding stream was established outside of the managed care premiums. Beginning in fiscal year 1997 (July 1, 1996 through June 30, 1997), these funds flow through medical schools, which will be allowed to keep a growing percentage of the funds over the next four years. The system includes incentives to increase the number of primary care residents.
Medicaid Fee-for-Service
Prior to the advent of TennCare, GME was included as part of the fee-for-service Medicaid reimbursement rates for hospitals. Most of the $56 million spent annually on GME went to the state's four largest academic medical centers.
Medicaid Managed Care
Under the agreement to restore funding in fiscal year 1996, GME dollars now follow residents to training sites and provide financial incentives to the state's medical schools to encourage primary care training as well as the placement of those trainees in under-served areas. GME funds are to be allocated directly to the state's four medical schools, with the goal of encouraging education in nonhospital, community-based settings. The plan is being phased in over a five-year period.
Year 1 (fiscal year 1996). $48 million was allocated to 10 teaching hospitals at a level equal to that of fee-for-service payments received by these institutions prior to TennCare.
Years 2 to 5 (fiscal years 1997-2000). The new allocation method will be phased in as follows:
| Payments | ||
| Fiscal Year | To Medical Schools | To Teaching Hospitals |
| 1997 | 25% | 75% |
| 1998 | 50% | 50% |
| 1999 | 75% | 25% |
| 2000 | 100% | 0% |
back to topOther State Funding for GME
State funds are available to support family practice, including an estimated $3.5 million, or $32,700 per resident, for family practice residency training programs and an estimated $4.3 million for departments of family medicine. (Estimates are from the American Academy of Family Physicians for the years 1994-1995.)
In addition, 1993 legislation requires the two state universities to jointly develop and implement a plan to encourage graduates to practice primary care in the state. The plan must ensure a 10 percent increase in the number of graduates entering primary care.
Cost/Payment Studies
No significant studies have been done across training programs.
Public Policy Process
Tennessee implemented its TennCare Medicaid managed care program in 1994. While the state's goal was to expand coverage comprehensively and quickly, the program's implementation was costly, significantly burdening the state's finances. A major early casualty of the fiscal pressures was Medicaid support for GME. The state abruptly terminated TennCare's funding for GME, effective January 1995.
In 1995, the state's new governor, Don Sundquist, created a Governor's TennCare Roundtable to advise his administration on how to improve TennCare. The roundtable made a number of recommendations in an interim 1995 report, including restoration of GME funding. A small working group was formed to examine the issue of GME funding and to work with a consultant and state officials in drafting specific recommendations. During this period, TennCare officials decided to allocate $48 million to restore support for GME during fiscal year 1996.
Early problems with TennCare centered on the lack of primary care providers in many rural areas of the state. While TennCare was supporting GME, the funding was not distributed in a way that addressed the program's primary care needs. With the restoration of TennCare GME funding, it was determined that these funds should be tied to the state's physician needs. The working group's plan to restore GME funding was ultimately accepted by the state. In 1996, GME funds were restored to the state budget, pending HCFA's approval of Tennessee's Medicaid waiver.
[This information was provided by Tim Henderson, Director, Primary Care Resource Center, National Conference of State Legislatures.]
back to topCurrent State Funding Policies for GME
In 1997, Utah passed legislation establishing an all-payer GME financing system that includes allocating available funds to support state health workforce needs. The legislation established a Council on Graduate Medical Education (COGME) to oversee the raising of funds from Medicaid fee-for-service, Medicaid managed care, private payers/purchasers, and Medicare (pending federal approval for Medicaid and Medicare) and the distribution of these funds to teaching hospitals and other sites.
Medicaid
Beginning in January 1998, all GME funds will flow to the new COGME. The council is empowered by law to: (1) develop options for financing GME with all funds being nonlapsing; (2) determine methods of reimbursing institutions that sponsor the training of health care professionals; (3) determine the number and type of training positions for which monies may be used; and (4) distribute monies for GME. The funds are to be distributed in a manner that prepares postgraduate residents for geographically diverse inpatient, outpatient, and community settings; encourages interdisciplinary clinical training among health care professionals; and promotes stable funding for accredited programs. The funds may also be used to support workforce surveys. The council is also charged with assuring that all parts of the community contribute their fair share, and that fair guidelines are developed and used to distribute the funds received.
Work is underway to determine how Medicaid's share for both fee-for-service and managed care GME will be carved out and sent to the council.
Private Payer Assessment
The COGME is charged with designing and implementing a broad-based funding mechanism. Only two possibilities have been identified so far: a hospital bed charge and a worker's compensation surcharge. At present, the latter approach is preferred: a surcharge on worker's compensation would be broader than a bed charge, couldn't be called a "sick tax," and would cover all public and private employers.
State General Revenues
The state does not currently use general revenue for GME; however, this option is to be addressed by the new COGME.
Other State Support for Specific GME-Related Activities
At present this includes support for physician assistant, nurse practitioner, and pharmacy programs in which clinical training is shared and coordinated. Issues related to efficiency, coordination, and costs for all these programs and GME will be studied by the COGME.
Funding Stream
Collection. The COGME will determine methods for raising and collecting funds. Pending HCFA approval, Medicare and Medicaid funds will be carved out and directed to the council. GME from all payers will also flow to the council, but most state-appropriated funds will not. (These funds are either included in specific program budgets and/or controlled by the state's Board of Regents for Higher Education.) The COGME is charged with coordinating funding for GME from all sources. Because the council makes recommendations to the governor and reports to the legislature, legislation may modify some current funding policies.
Distribution. Funds will be pooled, but kept separate for DME and IME in 1998 and 1999. Reimbursements or distributions will be cost-based. DME funds will flow from the COGME to sponsoring institutions not directly to the hospitals and will be allocated according to a uniform rate for all residents. IME reimbursement will become cost-based as soon as a data system is developed to capture costs in all training settings.
State workforce issues. Funded residency positions will be based on state-determined need, and augmented by regional need. An FTE formula has been established which gives priority to U.S.trained graduates and primary care residents. The formula sets funding for international medical graduates at 0.5 FTEs. Non-COGME-supported training programs can exist, but they will not receive COGME funding for any positions outside of these guidelines.
Cost/Payment Studies
In 1996, the Deloitte & Touche Consulting Group conducted a Health Professions Education Cost Study. The study covered GME, pharmacy, medical students, nurses, and physician assistants. DME and IME costs for each program and funding source were identified, and the potential unfunded and cost-shifted DME and IME expenses were identified. This study became the basis for creating the state COGME and for defining the council's authority to address funding of GME and all clinical training for the defined health professions.
Public Policy Process
Recent Legislation
Legislation was passed in 1997 establishing the state's COGME and the new policies described above, in "Current State Funding Policies for GME."
Committees
After 1997, the COGME will become the legal body to direct the state's GME issues. The council will have three committees: Funding, Workforce Planning, and Health Professions Liaison. In 1997, a task force of the state's Health Policy Commission advised pending legislation, submittal of a HCFA waiver, FTE guidelines, broad-based GME funding mechanism, data requirements, health professions surveys, reporting, and coordination with the Board of Regents for Higher Education.
Reports
The following reports are available:
[This information was provided by Gar Elison, Senior Policy Consultant, Utah Health Policy Commission.]
Eligible Applicants
Any teaching hospital receiving Medicare GME reimbursement in New York State was eligible to apply for the demonstration. The 42 participants, all of whom were required to submit an application to HCFA and receive approval, include 5 major voluntary academic medical centers, 24 major voluntary and community teaching hospitals, and 13 public hospitals.
The demonstration encourages collaborative efforts by allowing hospitals to participate through GME consortia, as defined under New York State law, or jointly with one or more other eligible hospitals. Such participants must meet the reduction and primary care criteria in the aggregate across all sites. Eighteen hospitals are participating through three GME consortia, 12 are participating through five joint applications, and 12 are participating individually on their own behalf.
Participation Criteria
The basic criteria for participation is agreement to reduce the number of FTE residents in training by 25 percent from the 1996-1997 program year, the base year for the demonstration, while maintaining the proportion of primary care residents to total residents. Primary care is defined pursuant to Medicare law as general internal medicine, general pediatrics, family medicine, geriatrics, preventive medicine, and osteopathic general medicine. Variations on this criterion are permitted for participants that meet other demonstration goals. The 25 percent reduction requirement is lowered to 20 percent for participants that also increase the proportion of primary care residents by 20 percent, as well as for hospitals participating through GME consortia who will maintain the base-year percentage of primary care residencies. Participation criteria must be met by the end of five years.
Phase-out of GME Reimbursement Transition Funds
Medicare fee-for-service GME reimbursement for residency positions that have been eliminated will be phased out over six years, according to the following schedule of declining percentage amounts applied to the loss associated with eliminated residency positions (the demonstration does not affect payments from Medicare managed care entities or for GME positions not eliminated):
| Reimbursement Associated with Residency Requirements | |
| Year 1 | 100% |
| Year 2 | 95% |
| Year 3 | 85% |
| Year 4 | 70% |
| Year 5 | 50% |
| Year 6 | 25% |
| Post-Demonstration Years | 0% |
Other Features of the Demonstration
Other features of the demonstration project include the following:
2. Council on Graduate Medical Education. June 1997. 1997 Recommendations to the Congress and the Secretary of Health and Human on Graduate Medical Education Payment Reform. Rockville, MD: U.S. Department of Health and Human Services.
3. Dobson A, Coleman K, Mechanic R. August 1994. "Analysis of Teaching Costs." Prepared for the Association of American Medical Colleges by Lewin-VHI, Inc.
4. Gold MR. August 1996. "Effects of the Growth of Managed Care on Academic Medical Centers and Graduate Medical Education." Academic Medicine 71:828-838.
5. Institute of Medicine. 1997. On Implementing a National Graduate Medical Education Trust Fund. Washington, DC: National Academy Press.
6. Institute of Medicine. 1996. The Nation's Physician Workforce: Options for Balancing Supply and Requirements. Washington, DC: National Academy Press.
7. Journal of the American Medical Association. Sept. 4, 1996. "Graduate Medical Education," Appendix II, Table 1, 276:739.
8. Komaromy, M et al. 1996. "The Role of Black and Hispanic Physicians in Providing Health Care for Underserved Populations." New England Journal of Medicine 334:1305-1310.
9. Mullan, F. Summer 1996. "Powerful Hands: Making the Most of Graduate Medical Education." Health Affairs 15(2):250254.
10. Pew Health Professions Commission. Dec. 1995. Critical Challenges: Revitalizing the Health Professions for the Twenty-first Century. San Francisco: University of California, San Francisco, Center for Health Professions.
11. Plumb D, Henderson T. October 1995. Medicaid Funding of Graduate Medical Education: A Survey of the States. Washington, DC: George Washington University; Association of American Medical Colleges.
12. Reuter J. Jan. 1996. Changes in Payments for the Direct and Indirect Costs of Graduate Medical Education. Washington, DC: Institute for Health Care Research and Policy: Georgetown University Medical Center.
13. Rowland D, Hanson K. Fall 1996. "TRENDS: Medicaid: Moving to Managed Care." Health Affairs 15(3):150-153.
14. Seifer S, Troupin B, Rubenfield G. 1996. "Changes in Marketplace Demand for Physicians: Study of Medical Journal Recruitment Advertisements." Journal of the American Medical Association 276:695-699.