GME Financing Reform

At a time when just over one third of teaching hospitals have an operating deficit, the thought of opening debate on Medicare-based Graduate Medical Education (GME) funding may not be welcome.¹ We learned in medical school that survival outranks self-actualization on Maslow's hierarchy, and survival is the operating priority in the teaching hospital community. In 1984, per-resident GME payments were retroactively fixed at 1983 levels. This decision rewarded those hospitals that had loaded their costs on these funding streams for the next 20 years, resulting in variance that is so nonsensical that it has become difficult to defend against reductions. Unable to significantly change their per-resident payments, hospitals spent the next 2 decades increasing the other half of the equation—residents. However, in 1997 this large loophole was closed with a retroactive cap based on 1996 levels of reported residents per institution. This likewise rewarded hospitals that had expanded their residencies and will likewise create a prospective nonsensical variance. As nonsensical and unfair as the current mechanism for GME funding is, fear of opening these entitlements to review and debate, and the risk that this dependable revenue stream might become subject to annual review or be removed altogether, keeps teaching hospitals in the same trench. Mutual survival is a rallying principle that trumps rational policy.

In this issue, Rich et al.² (the SGIM Health Policy Committee Medicare GME Cluster) provide succinct and helpful overviews of the history of how we arrived at this point in GME financing and 5 thwarted efforts in as many years to fix it. The overview of the early reasons that GME funding was politically feasible contrasts starkly with what has evolved, thanks in large part to payments to hospitals and increasing reliance on GME funds to support missions not directly linked to education, both of which have obscured accountability and confused purpose. The 5 reform proposals offered since 1997 have surprisingly similar themes but are most notable for their lack of success despite the prestige of the groups making them. It's worth noting that recommendations from the Institute of Medicine in 1989 and 1997, the American Medical Association in 1999, and four of the national family medicine organizations in 1998 had similar themes for GME financing.^3–6 Rich et al. have seized this opportunity to offer another option that would alleviate problems with variance and focus GME funding on the actual costs and locations of training. Their proposal would assess specialty-based training costs, create a direct cost payment that included all educational costs, and disburse payments to entities closer to the actual training programs (medical schools and practice plans). Their plan would also provide greater leverage for accountability, not only for funds but also for the quality of the education, through accreditation mechanisms.

This proposal is not just an academic exercise. There is credible evidence about the problems that point to specific needs for reform and what may be in store if they are permitted to perpetuate. We are not educating young physicians in the outpatient setting, where 90% of health care occurs.⁷ The current configuration of payments prevents this rational change. The Balanced Budget Act of 1997 (BBA97) and 2 subsequent revisions to GME payment legislation have created small opportunities to expand training outside of hospitals, but all three failed to account for and pay for substantial resident time spent in these settings prior to 1996. This failure has deeply affected primary care training, particularly rural training programs. This problem will likely only get worse, since a GME payment policy tied to hospitals is not likely to train and configure a physician workforce for an aging U.S. population with chronic diseases. Current payment location and the legacy of BBA97 exclusion of outpatient training will particularly impact rural populations that have an even higher proportion of elderly.

The SGIM Cluster also discusses GME payment variances, and we now know that these have real effects on the workforce that are not theoretical. Among the 47 states with at least 5 counties, there is a statistically significant negative correlation between the average direct medical education (DME) payment and the percentage of counties designated as primary care health profession shortage areas (PCHPSAs).⁸ For example, North Dakota has the lowest average DME payment rate and New York the highest, yet 30% of North Dakota's counties were PCHPSAs and less than 2% of New York's were.

Finally, we have evidence to support the SGIM Cluster's assertion that program directors don't know GME payment amounts for their programs, and when they think they know, they are usually wrong.⁹ Until recently, program directors were dependent on their hospital administrators for this information, but we have recently made it available on our website (www.aafppolicy.org). Turning program directors into accountable fiduciary agents for public funds will require such knowledge, and would be made even stronger if they had direct control of DME funds to leverage needed changes.

Despite the value of the SGIM Cluster recommendations, limitations, and a lack of political strategy threaten their viability from the start. Shifting DME payments to “physician entities,” suggested to be medical schools or practice plans, but increasing accountability from programs is not likely to improve current accountability. Schools and practice plans are not accountable to the Residency Review Committees (Institutional Review Committees might be a better option) and have their own cash-hungry missions. Any new entity receiving payments may not want to assume the risk of unfunded education costs currently covered by hospitals. While reengineering DME payments to incorporate all education costs may help reduce underfunding, hospitals are likely to oppose the proposed implicit transfer of education funds from indirect medical education payments. Finally, survival instincts are likely to be a significant obstacle at a time when public and political will seem to be aligned to find funds to solve a nursing shortage and to fund a Medicare medication benefit.

Given that many of the affected parties are potentially poised to oppose this or any other GME reform proposal, what strategy might lower enough barriers to start the discussion? The SGIM Health Policy Committee Medicare GME Cluster is to be congratulated for joining a distinguished list of advocates for changing GME financing, and there may be some opportunity in uniting their collective advocacy. Another strategy is to acknowledge and attempt to deal with the basic Maslowian survival needs of teaching hospitals. Hospitals are looking at payment reductions from almost all sources and may be open to a discussion about GME reform that begins with assurances of secure funding to replace the GME streams that keep them alive. Rational reform will require a united leadership from within the GME community, most notably from teaching hospitals. Congress and the Executive Branch can either incentivize this reform with guidance from the GME community or continue to make the current financing mechanisms so painful as to force reform.