GME and the physician labor market

By Steven St. Peter, M.D.

A resident physician in Internal Medicine in Philadelphia, PA

Recently, New York State health officials made a very surprising announcement: they proposed paying hospitals in New York to not train residents. Initially, the economic motive of this proposal seemed similar to ones I had witnessed while growing up in Kansas. Perhaps this was the equivalent of a government crop program—paying the farmers of American graduate medical education to not plant their fields. As I struggled to understand the issues, I discovered that this gesture hints at a very serious crisis in evolution. The reality is that we are training too many physicians, and this problem is amplified by the current system of funding graduate medical education (GME). Physicians must actively seek a better solution to the dilemma of current GME crisis because of its long-term implications on the physician labor force, and hence, the future of the U.S. medicine.

As pointed out by Rivo et al. in the New England Journal Of Medicine, data indicate that the supply of physicians continues to grow faster than the population. Although the aging of America and increased demand for services weigh in favor of more physicians, managed care and non-physician providers decrease the need. It is generally accepted that between 145 and 185 physicians in patient care per 100,000 people is an appropriate ratio. The current supply of physicians is approximately 200 per 100,000. Further increases in this ratio are anticipated over the next two decades, and underemployment of specialists is a real possibility early in the 21st century.

The physician labor market has several unique properties which make it different from other labor markets. The market has several barriers to entry, and the training process has been traditionally very tightly controlled. Physician training has existed in an apprentice form since ancient times. However, following the Flexner Report early in this century, there was a dramatic attempt to standardize training. This resulted in the emergence of the modern university-based medical school, and eventually the emergence of the university-based teaching hospital. Graduate medical education was structured in the context of these university-based hospitals through internships and residencies. Following the end of World War II, there was a tremendous surplus of physicians returning from the military in search of residencies. The number of available GME spots doubled (and has steadily increased since). However, the number of U.S. medical graduates (USMGs) grew less rapidly, and these excess positions were increasingly filled by international medical graduates (IMGs). Until the 1960’s, the burden of medical education was being borne by these institutions themselves with minimal reliance on patient revenues and government subsidies. Then came Medicare, and everything changed.

With the intention of increasing the numbers of physicians and increasing public access to medical services, Medicare created GME payments. Hospitals were paid a stipend to cover resident salaries and benefits (direct costs) as well as an additional amount to cover the costs of teaching hospitals’ operational functions (indirect costs). These moneys, currently an average of $70,000 per year per resident, provided potent incentives for expansions of housestaff positions with the university hospitals. Also, many community hospitals created residency and fellowship programs, further increasing the number of available residency spots. Indeed, many hospitals have become dependent on residents due to the hospitals’ increased service needs (especially in the inner cities), ready availability of residents as a labor source, and the financial advantages of GME payments which can be combined with other third party payments. A publication by Salsberg et al. in the Journal of the American Medical Association estimated that all payers combined amounted to $188,000 per resident per year in New York State. The problem with this evolution is that the GME payments are now being used for purposes which they were not intended, most obviously the GME for indirect costs. GME payments should be directed at creating an appropriate supply and mix of physicians to meet the needs of the future.

The other subject which invariably comes up when discussing GME is international medical graduates. The system has become increasingly dependent on IMGs as medical schools have controlled the number of US medical graduates. Estimates indicate that IMGs account for 70 percent of the increase in residents over the last decade. Between 1993 and 1995, the total number of residents increased from 97,370 to 98,035—a modest increase. However, this represented an increase in IMGs of 2276 and a decrease of USMGs of 1571. The percentage of IMGs of all residents was 25.5 percent in 1995. This widening difference is referred to as the “IMG window,” and efforts to decrease the total number of residents by decreasing UD medical school graduates is futile until this window is closed. Also, recent publications have indicated that overproducing physicians (presumably through IMGs) to “trickle down” new physicians to undeserved rural and inner-city communities has not been effective. Thus, IMGs have become targets for reforming the physician labor market.

New York State has been particularly affected by the GME issue. Between 1988 and 1994, the number of all residents grew by 2517—an increase in IMGs of 2673 and a decrease of USMGs of 150. New York State has the highest number of residents per population. In 1994, 46 percent of all physicians training in New York State were IMGs. When broken down by specialties, the data is even more skewed. For instance, during this period IMGs filled 66 percent on the Internal Medicine residency positions. New York is the most dramatic example, but it fairly represents the problems faced in a several states. Again, GME has been substituted to solve today’s problems rather than to create a balanced physician supply for the future. Perhaps the most amazing fact about the New York situation and the numbers presented above is that state health officials were actively trying to decrease the number of residents in training. Clearly, something is failing.

The Council on Graduate Medical Education has dealt with several of these issues, and set certain goals including to reduce the number of federally funded first-year resident positions to 11 percent of the number of U.S. medical school graduates in 1993. COGME has also recommended that at least half of all graduates pursue primary care careers. This is a good start, but has the shortcoming of not including provisions to find staffing alternatives for resident-dependent hospitals. The New York plan would at least support hospitals without jeopardizing the physician labor market of the future. Policy makers must be aware that these hospitals do provide needed services, but the solution does not lie in continually expanding GME programs. In fact, there is historical evidence that by increasing the number of physicians we create a less efficient and more costly health system.

Physicians must encourage a rational approach to meeting the demands of the physician labor market. If the job is left to the “free market,” our profession and the entire health care system will be damaged. Physician oversupply is a waste of human resources and a waste of the public funds invested to train the underemployed physicians. As a physician, particularly a younger physician, I am worried about physicians losing control of our training process. I would like to see organized medicine aggressively respond to the challenge of graduate medical education and its implications for the future physician labor market.