This report, On Implementing a National Graduate Medical Education Trust Fund, is about a complex subject, graduate medical education (GME), and a complex program, the system of payments from Medicare—and potentially from other sources—to support the institutions that provide GME. This overview is intended to provide readers who are not familiar with the details of the subject matter a brief introduction to the major points covered in the report.
The term ''graduate medical education'' as used in this report primarily refers to the training of about 100,000 persons in clinical settings (hospital and outpatient) who have received an M.D. or D.O. degree from a medical or osteopathic school and who have, or are eligible to have, a license to practice medicine in one or more of the United States. This training, usually called a residency or fellowship, is a form of apprenticeship that provides a supervised clinical practice and didactic educational experience lasting a number of years in various kinds of medical specialties such as pediatrics or orthopedic surgery. Each residency (or training) program is under the aegis of one of the approximately 1,000 teaching hospitals in this country; is accredited by a nongovernmental professional body; and produces practitioners who are recognized as, and are usually eligible to obtain board certification as, for example, generalists in family medicine or specialists in radiology.
Graduate medical education, for the purposes of the report, also means, unless otherwise noted, the training of dental and podiatric professionals; undergraduate, preprofessional nurses; and a number of allied health professionals, such as x-ray technicians and respiratory therapists, among others. In addition to the changes discussed for physicians, the report also contains other significant suggested changes for nursing as well.
The system of about $6.6 billion (FY 1996) in payments from Medicare that support GME was instituted in recognition of the obligation to support the legitimate costs incurred for training, to provide Medicare beneficiaries access to care in the teaching setting, and also in recognition of the general public value of
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 1
On Implementing a National Graduate Medical Education Trust Fund Overview of Graduate Medical Education and the Distribution of a Trust Fund This report, On Implementing a National Graduate Medical Education Trust Fund, is about a complex subject, graduate medical education (GME), and a complex program, the system of payments from Medicare—and potentially from other sources—to support the institutions that provide GME. This overview is intended to provide readers who are not familiar with the details of the subject matter a brief introduction to the major points covered in the report. The term ''graduate medical education'' as used in this report primarily refers to the training of about 100,000 persons in clinical settings (hospital and outpatient) who have received an M.D. or D.O. degree from a medical or osteopathic school and who have, or are eligible to have, a license to practice medicine in one or more of the United States. This training, usually called a residency or fellowship, is a form of apprenticeship that provides a supervised clinical practice and didactic educational experience lasting a number of years in various kinds of medical specialties such as pediatrics or orthopedic surgery. Each residency (or training) program is under the aegis of one of the approximately 1,000 teaching hospitals in this country; is accredited by a nongovernmental professional body; and produces practitioners who are recognized as, and are usually eligible to obtain board certification as, for example, generalists in family medicine or specialists in radiology. Graduate medical education, for the purposes of the report, also means, unless otherwise noted, the training of dental and podiatric professionals; undergraduate, preprofessional nurses; and a number of allied health professionals, such as x-ray technicians and respiratory therapists, among others. In addition to the changes discussed for physicians, the report also contains other significant suggested changes for nursing as well. The system of about $6.6 billion (FY 1996) in payments from Medicare that support GME was instituted in recognition of the obligation to support the legitimate costs incurred for training, to provide Medicare beneficiaries access to care in the teaching setting, and also in recognition of the general public value of
OCR for page 1
On Implementing a National Graduate Medical Education Trust Fund GME and the teaching institutions that provide it, which are essential parts of the infrastructure for quality health care in this country. The payment system is actually divided into two major programs: a direct GME (DME) and an indirect medical education (IME) component. The direct GME component from Medicare comprises payments made directly to teaching hospitals based on 1984 historical costs to cover the stipends of residents, the supervisory personnel, and other associated hospital costs for supporting a residency program. These are payments for each individual, full-time-equivalent resident. Some of the per-resident payments are reduced (weighted) to provide disincentives for long-duration, specialist residencies, and some are paid in full to provide incentives to train generalists. Partial reimbursement of the costs incurred by some hospitals in training other professionals such as nurses and allied health personnel is also made. In general, payments increase by the consumer price index each year, and the fraction of the costs paid by Medicare is the same fraction that Medicare patient-days are of total hospital patient-days. Importantly, physician payments also increase in the form of an equal additional per-resident payment for each additional resident that a hospital adds to its total complement of residents. Because of idiosyncrasies in the 1984 cost base, per-resident payment amounts vary significantly, and likely in ways that do not reflect the real costs of residents, from hospital to hospital; in short, there is currently an incentive to add residents especially for hospitals at the high end of the per-resident payment spectrum. The indirect medical education component from Medicare is not based on any identified costs. Instead, it is intended as support for teaching hospitals and to compensate for the observed higher costs that the presence of training programs generates in these hospitals. Additional tests, special care units, unsponsored research, more seriously ill patients, and care of nonpaying patients contribute to these higher costs. Medicare pays hospitals for inpatient care for Medicare beneficiaries using a diagnosis-related group (DRG) methodology that pays a predetermined amount to the hospital (with some local, individual adjustment) for each beneficiary admission depending on the patient's diagnosis. For IME payment, a formula has been devised that converts the ratio of the number of full-time-equivalent residents to a hospital's number of beds into a percentage—the more residents per bed, the higher the percentage. This percentage is applied to and increases the DRG payment for each Medicare admission. The annual update of payment rates provided by Congress for the DRG method increases the per-admission payment and thus the IME amount as well; again, an incentive is created to add residents and to train them only within the hospital's walls. Individual hospitals' decisions and the accrediting process are currently the only controls on the number of residents and therefore on much of the payment for direct and indirect medical education. The Medicare GME payment system itself is open-ended, so the system cannot control rising costs caused by expanding resident numbers. FY 1996 support for indirect education ($4.3 billion) was about twice the dollar amount paid for direct education ($2.2 billion).
OCR for page 1
On Implementing a National Graduate Medical Education Trust Fund Some Medicare beneficiaries have subscribed to health care plans or organizations that receive from Medicare a prepaid, per capita premium, otherwise known as a capitation payment. Hospitals receive from these plans a negotiated rate rather than a payment from Medicare through the diagnosis related group system. The premium to the plans is based on the adjusted, average, per capita cost (AAPCC) of Medicare patients in the relevant county who have not joined plans and are still in the regular fee-for-service system. That cost includes hospital costs with their GME payment included, if they are teaching hospitals. In effect, Medicare payments to hospitals that were originally made through the DRG system and were enriched by GME payments, are now included in capitations outside the system and no longer targeted to support medical education and teaching hospitals. Instead these payments go to plans and organizations paid by capitation. They are not necessarily included in the hospital rate negotiated by the plans with teaching hospitals, if indeed the plans negotiate with higher cost hospitals, such as teaching hospitals, at all. As the prevalence of capitation and managed care increases, teaching hospitals, as higher cost providers, are threatened by competition from lower cost nonteaching hospitals. The implicit GME subsidy that private-sector payers were willing and, in a less competitive environment, able to provide, as well as the explicit Medicare (and Medicaid) subsidy, are diminishing, and the economic base of GME and teaching hospitals, as currently constituted, is being threatened. The combination of this financial problem, the financial problems of the Medicare program, and the need for a balanced national budget as well as certain perceived flaws in the current GME financing system (e.g., excessive DME payment variation, undue emphasis on in-hospital training, and uncontrolled growth in resident numbers) have generated interest in considering new approaches to GME and its support. The task of the Committee on Implementing a National Graduate Medical Education Trust Fund was to suggest a plan for distributing a trust fund. The committee was not charged with deciding the size or composition of the health care work force, or, in particular, the physician work force or with addressing the issues of international medical graduates. The committee was aware that much has been written and many recommendations made on these matters and yet the nature of the problems is such that no simple solutions have been found. Accordingly, the committee, after considerable debate, decided that, as important as these issues may be, and as important as payment policy might be in influencing them, no attempt should be made to define the "correct" work force or impose further payment provisions to achieve it. Instead, the committee felt that it should concentrate on a set of principles and, flowing from them, on recommendations for a new payment structure that is more open and more responsive to changes in health care delivery. The committee noted the general rationale for Medicare or other funding sources. With a focus on Medicare the committee's principles included statements and discussions of: the value of GME; the desirability of relative neutrality of payment in shaping work force; the need for opening payment
OCR for page 1
On Implementing a National Graduate Medical Education Trust Fund eligibility to new training entities; the appropriateness of each payer contributing proportionately to the support of GME; the value of conforming nursing education support in principle and process to physician support; and the necessity for reasonably consistent payment across institutions and for gradual, nondisruptive transitions to any new distribution. There are many variables to be considered in designing a distribution plan, and, therefore, many possible plans depending on what policymakers might decide about each variable. Yet the committee was asked for a plan, and it provides one in this report. Overall, the committee recommended that a defined amount of support be determined for GME and its components each year. This has the effect of prospectively creating a fund for GME separate from the current payment system. The fund could be managed based on the status of the teaching sector and decisions on the relative roles of direct and indirect GME support. It also means that whatever plan is chosen, the financial impact on Medicare, or any other financing source, is set at the outset and cannot be changed by events during the payment year. Only the distribution can vary. The committee noted that a non-Medicare source of funds would be a desirable addition to the Medicare source, and consistent with the principles of proportional participation by all payers, or by general revenue, in a general public benefit. Because the charge to the committee did not include identifying funding sources, a discrete recommendation regarding such a source was not made. Nonetheless, the report does comment on distribution of a non-Medicare source of funds, primarily noting that, in setting a non-Medicare share, notice should be taken that Medicare is, and has been, overpaying its share and that non-Medicare funds should be distributed related to non-Medicare caseload. The committee recommended that direct GME payments transition very slowly toward a uniform, average per-resident payment. This payment would be the appropriation for DME divided by the total number of qualified residents, and as such, of course, the per-resident payment amount could be controlled by the appropriation. The existing system of weighting payments in favor of primary care residencies would be preserved, but no new conditions would be added. Payment amounts would move from each hospital's historical base very slowly for 5 years, during which time a system that took into consideration the costs of local inputs should be designed and thereafter implemented. However, additional residents added by an institution would be paid beginning immediately at only the uniform or actual amount, whichever was less. The committee believes that a more consistent price for each resident should replace the current, highly variable payment. A DME payment becomes truly a uniform Medicare payment if it is not related to an individual institution's variations in Medicare caseload, but is based on the overall national Medicare percent of caseload. The committee observed that institutions that currently serve the most Medicare beneficiaries are not necessarily those that will best train practitioners to care for the Medicare population in the future, and that some institutions that now serve few Medicare patients could be desirable training sites. The committee questioned whether
OCR for page 1
On Implementing a National Graduate Medical Education Trust Fund DME and associated hospital activities are part of services to individual Medicare beneficiaries or general societal benefits. If the latter, then all payers, including Medicare, should contribute a proportional share and that share should be distributed evenly. The committee preferred this approach, but notes that this reasoning suggests very strongly the inclusion of proportional payments from other payers or a general, non-Medicare source. A strong argument against a uniform Medicare payment, not tied to individual caseloads and without support from other sources, is its potential for substantial redistribution of DME funding from existing institutions. While a uniform Medicare payment would penalize above-average Medicare caseload institutions, a uniform, all-payer or general, non-Medicare payment would compensate these same institutions because they have lower than average non-Medicare caseloads. The redistribution effect of a uniform Medicare payment would be mitigated, in theory, by the reverse redistribution effect of a uniform non-Medicare payment. Of course, in the end and in practice, variation could be reduced and a consistent DME payment or price would result, reducing the relative advantage or disadvantage in DME payment from one institution to another, except for adjustments because of weighting and the relative costs of inputs. The committee thinks this is desirable. The committee's plan also included expanding eligibility to receive DME payments directly to include any accredited educational or health services institution that met conditions of full responsibility for its training programs (called a sponsoring institution). This has implications for providing support for institutions that will allow retargeting nursing support toward advanced practice nurses, which the committee recommends. It also provides opportunities for training in new settings such as ambulatory sites and managed care organizations, under the aegis of consortia and the like. In addition, taken together with the other recommendations, it also improves the financial standing of, for example, children's hospitals or institutions offering residencies in preventive medicine, by permitting them to become DME-eligible sponsoring institutions and providing a uniform DME payment. A uniform payment (or price) simplifies decisions about what to pay new institutions. The committee recommended that indirect medical education funds continue to be targeted to teaching hospitals exclusively, but that they be divided into halves. One half of the payments would no longer be calculated using the current formula. They would be dissociated from resident numbers, beds, and Medicare admissions, and simply based on the same proportion of total IME payment that each institution had received historically. The other half would continue to be allocated by the formula as a percentage addition to each Medicare payment for each Medicare admission. To more equitably distribute funds to hospitals that serve Medicare patients from per capita, prepaid plans, a factor proportional to the percentage of patients from prepaid plans compared to total Medicare patients at a hospital would be introduced to improve distribution by the formula half of IME payment. A system of distributing non-Medicare funds, if they were provided, proportional to non-Medicare patients and similar to the process used for Medicare was also proposed for both the direct and
OCR for page 1
On Implementing a National Graduate Medical Education Trust Fund indirect components of the GME payment system. Of course, all these payments would need to be monitored and adjusted to ensure payout did not exceed the defined trust fund amount. These recommendations would have the effect of halving the decrease in funding that a hospital would sustain if it decreased its number of residents because the historically based payments would continue and would not be reduced if residents were reduced. At the same time, the incentive to admit and serve Medicare patients (and Medicare capitated patients) would be preserved in part because the formula-based payments would still be made through individual Medicare DRGs and the factor proportional to capitated patients would improve distribution to hospitals that served such patients. Because the amount to be distributed is decided at the outset and capitated patients do not affect that total, the policy decision on the calculation of the AAPCC, Medicare capitated payments, and payment to managed care organizations becomes separate and distinct from the policy decision on the amount needed to support graduate medical education and payments to teaching hospitals. Once the appropriation to the Trust Fund is made, that amount will be distributed by the proposed distribution plan and enrollment of Medicare beneficiaries in managed care or decisions on calculating the AAPCC cannot change the Trust Fund. GME expenditures in the fee-for-service system could still, by the current method, be reflected in the AAPCC, but the Trust Fund, which has been set at the proper level by Congress, could not change. At the same time, hospitals which successfully compete to continue to provide access to Medicare managed care patients receive a better distribution within the fund limit as an incentive to continue to serve those patients. The committee did not choose to explore AAPCC issues further in view of these conclusions. The committee's plan does not include additional weighting factors which directly affect supply or composition of the physician work force. However, by including a defined total amount for all GME payments, a decision by a teaching hospital to add residents will have no effect on total expenditures; instead an increase in total residents would lower the per-resident payment for all institutions. In addition, residents would be added at the uniform or actual DME payment level, whichever is less, and those institutions with the highest DME payments would see their per-resident payments diminish in the transition toward a uniform price. Further, half of IME support is dissociated from resident number. These modifications should provide a more neutral payment system with respect to overall resident supply, one that is aligned with the prevailing understanding of work force requirements. It is expected as a result that resident supply will fluctuate more with market conditions. In conclusion, the committee's suggested plan provides for an annually defined fund targeted to the support of GME and teaching hospitals and separate from the payment considerations of the overall Medicare DRG payment system and from the calculations by Medicare involving plans paid on a prepaid, per capita basis. The plan does not attempt new work force restructuring, but suggests setting a uniform price per resident and creating a system open to new
OCR for page 1
On Implementing a National Graduate Medical Education Trust Fund entrants, less restrictive, and more responsive to changing health care delivery. Relentless pressure to control expenditures has caused private-sector payers for health care services to deny training costs, putting teaching institutions and GME at risk, and increasingly shifting the burden towards responsible governmental public payers like Medicare. For this and other reasons, the plan notes the desirability in principle and practice, of a general, non-Medicare or all-payer contribution to the funding and makes some suggestions on how to distribute such a contribution.