Federal support for graduate medical education (GME) training positions has been capped for more than a decade and it is no secret that the country’s teaching hospitals are restive. They want “more cap.” A number of bills have been introduced in the House and Senate proposing an increase in the Medicare funded GME cap by fifteen percent, or roughly 15,000 positions. These proposals are an alluring Siren song but they will not be good for health care reform or for the country.

What could be wrong with more residents?

To answer this question, it is important to understand the background of federal GME support. Residency training programs have enjoyed generous Medicare funding for more than two decades. In 2007, Medicare paid hospitals more than $8.6 billion in support of GME based on formulas for “direct” medical education (costs such as resident salaries and teaching time) and “indirect” medical education (non-specific hospital costs associated with teaching hospitals.) In 2007, approximately 89,300 residencies were funded under this system making the average Medicare per resident payment an astonishing $96,000.

Not surprisingly this system is enormously popular with teaching hospitals since it represents a large and predictable funding stream based on the number of residents they employ – and employ is the operative word. Residents are learners, to-be-sure, but with their 80 hour weeks, they are also valuable and inexpensive skilled labor. Medicare GME funds make the resident a mini revenue center.

GME as Workforce Policy

GME is important beyond the hospital. GME governs the physician workforce in the country. In order to practice, all physicians– no matter the locale of their medical school — must participate in residency training in the US. Therefore, it is the training positions that hospitals offer and fill that determine the make up of the nation’s physician workforce.

But there is a problem here. The complement of residents that hospitals need to staff their services is not the same complement that the nation needs to deliver health care to 300 million people – 99% of whom are not in a hospital. The pattern of training in US hospitals is heavily specialty oriented reflecting the intensivity of hospital care as well as the more lucrative payments associated with many specialty services. Additionally hospitals often view primary care as a low revenue area that occupies precious beds as opposed to many interventional specialties which are very profitable.

Not surprisingly, the pattern of residencies offered by hospitals has become more and more specialized at the expense of primary care. Between 2002 and 2007, hospitals opened 7754 more new residency positions, 88.3% of which were in specialty care despite the Medicare GME cap. During the last decade, 20 family medicine residency programs have been closed and 645 less residents are being trained in family medicine today than ten years ago. In 1998, 54% of internal medicine residents planned careers in primary care, whereas only 23% did in 2007.

Full Cargo, No Rudder

Numerous scholars, analysts, public commissions, and the Administration have called for augmentation of the primary care workforce and restraint in specialty training and practice. Training expenses pale in comparison to the out year costs of a system starved for primary care and increasingly reliant on subspecialists for care. The “McAllen, Texas Syndrome” as eloquently documented by Atul Gawande captures the peril and cost of overbuilt specialty medicine. But unlike NIH research funding or HRSA training grants which are competitive, time limited, and activity specific, GME is an entitlement for which hospitals need make no commitment to types of trainees, regional needs, or national priorities. To date, the hospital community has been successful in defeating all efforts to modify the GME system to include planning, competition, or accountability.

Accountability is a problem. With an $8.5 billion investment in GME training programs, the government should be in a position to influence the direction of that education and, therefore the physician workforce of the nation. But it can’t under the terms of the Medicare GME entitlement. GME monies are, by far, the largest federal investment in medical education and dwarf the few hundred million dollars devoted to education for primary care, nursing, or physician assistants.

The final characteristic of the Medicare-subsidized GME system that should be noted is that 27% of current residents are international medical graduates. Since the mid-twentieth century, US hospitals have always offered more training positions than can be filled by the graduates of US medical schools. In recent years, this gap has remained constant with about a quarter of residency positions filled by international medical graduates, the preponderance of whom come from lower-income countries. Many people have raised their voices about the ethics of relying on IMG to make up for US educational shorfalls which began as an exchange initiative in the 1950s and 1960s but quickly became a US recruitment program for foreign physicians. Previous proposals by the Council on Graduate Medical Education (COGME) and others to decrease our reliance on foreign trainees have failed to alter the broad-based use of this mechanism.

The Big Ask

Recent debate over healthcare reform has produced a number of proposals related to physician workforce. Senators Nelson (D-FL) and Cantwell (D-WA) and Representatives Schwartz (D-PA) and Crowley (D-NY) have introduced bills laden with primary care friendly language. Within this primary care wrapping, however, is a much bolder, costlier, and primary care unfriendly pay load.

All of the bills contain similar legislative language to increase Medicare funded GME positions by 15%– the equivalent of roughly 15,000 new residents. These new slots would cost well over a billion dollars a year or more than $12 billion over 10 years. Significantly, the first dollar funding in these new bills is reserved for the “overcap” residency positions already established and funded by hospitals. Should there be any question about what sort of training this bill will support, almost 90% of those positions are in specialty training.

The 15,000 resident provision of these pieces of legislation, despite the rhetorical support for primary care, will bring us more of the same sub-specialty oriented training patterns well established in the past several decades. In the face of depressed interest in primary care, more cap is unlikely to result in more primary care trainees and will ultimately play out as more subsidized specialty slots. The most likely “primary care” positions to fill would be internal medicine positions and the majority of these will go on to specialize. A rapid increase in residency positions will initially result in a sharp rise in IMGs recruitment, increasing the brain drain and moving the country away from self sufficiency. Nearly 6000 IMGs come to the U.S. each year to fill residency positions. While some medical school expansion is underway in the U.S., the proposed GME expansion will require an initial 40% increase (2400 physicians) in the number of IMGs – the equivalent of an additional 24 medical schools elsewhere in the world sending all of their graduates to the US.

The provision will add substantially to the cost to Medicare GME and to the cost of Medicare in the future. The money, the signal, and the downstream effect will compete with and likely overwhelm other workforce reform strategies.

Many have argued that more residents are necessary to complement the growing number of US medical school graduates and/or that we won’t meet the future clinical needs of the country without more residents. Since we have 25% more residency positions than we do medical school graduates, it will take the better part of the next decade at the current growth rate of medical schools to close the medical graduate-residency gap – good domestic and good foreign policy. The increased need for physician services can be met by better use of the physicians we have now – a goal of health reform – and by the increased use of nurse practitioners and physicians assistants in primary care and specialty care settings. The important principle underlying this latter strategy is that all clinicians should work to the maximum of their training and licensure.

Real Reform in Workforce Policy and Practice

Workforce reform will not happen without innovation and investment throughout the life cycle of the primary care physician including medical school, residency, and practice. Primary care friendly curricula and generalist mentors will be key in medical school; debt reduction opportunities, community based residencies, and national attention to primary care will be important in residency; pay parity for primary care and support for primary care medical homes and similar organizational innovations will be needed to re-ignite interest in primary care practice. All of these reforms will be jeopardized by huge investments in additional training positions that will double down the subsidy for expensive and inefficient patterns of practice.

With the entire health system on the table for consideration, we have the opportunity to reassess and redirect federal support for physician training. We need a willingness to set aside interest-group-as-usual thinking if we are going to address the huge challenge of building a good, fair, and affordable health system. Rethinking our investments in the education and training of physicians is essential. Adding 15,000 residents to the system paid for by a flagging Medicare budget is not the place to start. Below is a list of strategies that will begin to move the system back toward a better and more effective balance between primary care training and specialty training.

* Create a National Health Care Workforce Commission: The Commission would develop national health professions workforce goals, recommendations and benchmarks. It should also review Medicare GME and make recommendations to the Congress in regard to improving the alignment of the program with national physician workforce needs.

* Support Teaching Health Centers: Unused “old cap” slots should be distributed to Teaching Health Centers programs and directly support the development of community health center-based residency programs without lifting the cap.

* Guarantee Primary Care Expansion: Any increase in residency slots must be distributed according to more stringent primary care preference criteria such as program primary care “track record.”

* Incentivize Primary Care: The National Health Service Corps including scholarship, loan repayment and loan deferment opportunities should be expanded significantly.

* Promote Primary Care Careers: Payment reform — reducing the primary care-specialty pay gap — and support for primary care practice models such as medical homes are essential to revitalizing primary care careers and student interest in primary care.

* Grow GME Responsibly: Any increase in Medicare-sponsored GME cap should not exceed projected growth in the number of U.S. medical graduates. Long-term workforce planning should focus on reaching self-sufficiency.