These marketplace dynamics have stimulated a major debate about how best to achieve a physician work force that meets the nation's needs. Arguments for active management of the physician work force, as opposed to relying solely on market forces, are compelling. First, decisions about choice of specialty are made early in medical training and are heavily influenced by the culture of academic medical centers—which emphasizes technologically intensive, highly specialized care. Second, practicing physicians face market forces that only incompletely penetrate the ivory tower to affect the size and composition of training programs. Third, the size and composition of training programs are driven primarily by service needs in hospitals, not larger societal goals. Fourth, the current Medicare subsidy for graduate medical education has no strings attached, allowing teaching hospitals to develop and expand training programs with ongoing taxpayer support. Fifth, existing antitrust statutes severely limit the capacity of organized medicine to voluntarily manage graduate medical education.

On the other hand, strong arguments for a more laissez faire approach have been made. First, some say that the changing marketplace will rapidly cause shifts in specialty choice and that any centralized mechanism will not have the capacity to adjust as smoothly as the market's "invisible hand." Second, the track record of federal efforts in work-force policy is not unblemished, with previous predictions of a physician shortage leading to the growth that has caused our current dilemma. Third, the aging of the population and ongoing innovation in medical science and technology make forecasting particularly hazardous. Fourth, some have cast the move to promote primary care as an anti-intellectual reversion of medical education toward a trade-school approach, as opposed to the scientific practice of medicine.

In this context, the formation of the Association of Subspecialty Professors in late 1993 was an important event. Directors of training programs in subspecialty internal medicine are an important and heretofore under-represented constituency in national discussions of work-force reform. As the largest medical specialty, internal medicine's decisions and directions have a major effect on the physician work force. Moreover, rapid growth in residency positions has occurred within internal medicine and that growth is concentrated in the subspecialties of internal medicine. Since 1988, residency positions in internal medicine have increased by 11% [2], and total subspecialty positions in internal medicine have increased by 28% (Lyttle CS. Personal communication).

In this issue, the Association of Subspecialty Professors presents its position [3] on the role of subspecialty internal medicine in national work-force reform. After making the obvious and undeniable case for subspecialty research, training, and practice, the Association makes many sensible recommendations. It recommends that any national work-force commission include medical educators and other knowledgeable physicians (although their call for exclusivity is problematic). The Association advocates an all-payer pool for graduate medical education, with additional funds allocated to provide staffing for training positions lost through downsizing. Stable funding is recommended for physician-scientists, and quality measurement systems for training programs are advocated. An enhanced role for subspecialists in training primary care physicians is also encouraged.

On the other hand, the Association's position paper makes some less credible recommendations and, more importantly, neglects several fundamental issues. For example, its recommendation for "need-based" modeling of demand implies that the enormous projected oversupply of specialists is an artifact of forecasting methodology, when in fact alternative forecasts, including "bottom up," demand-driven models have similar results (Malcolm C. Personal communication).

The call to make subspecialty medicine "more affordable" implies fee reductions or utilization reductions, or both, yet no method to achieve this affordability is specified.

The recent growth in internal medicine training programs has been driven by increasing numbers of international medical graduates entering U.S. residency and fellowship programs. Currently, international medical graduates account for 36% of first-year internal medicine training slots [2], and international medical graduates hold 37% of nephrology fellowships and 27% of cardiology fellowships [4] when compared with 23% of all residency positions [5]. The Association is silent about the merits of managing the work force through controlling the total number of training positions or using new policies to restrict entry of international medical graduates into U.S. residency positions.

Disturbingly, the Association's paper implies that only the medical subspecialties carry the "genetic code for biomedical research, patient care, and education." This view neglects fundamental contributions made by other specialists, as well as contributions made by generalists on health services research; decision analysis; small area variation; cost–benefit analysis; biomedical ethics; technology assessment; research design and statistical methods; medical education; physician-patient interactions; and clinical research in common conditions such as low back pain, syncope, and pneumonia. Equating efforts to increase the production of generalists as an attempt to diminish academic medicine's mission of creating and disseminating new knowledge in the service of patients is not logical. No evidence exists that the rapid growth in subspecialty training has expanded the research enterprise; if anything, the opposite has occurred—the number of young investigators applying for R01 grants decreased by 54% between 1985 and 1993 [6]. Changing circumstances of funding by the National Institutes of Health may have been the prime reason for this decline. Nevertheless, reducing the number of subspecialty training slots need not inhibit activities of physician-scientists.

Most importantly, subspecialty internal medicine, like all of academic medicine, receives the bulk of its financial support through public sources. Academic medicine has a social contract to provide new knowledge and well-trained practitioners to serve society. It must face the fact that adding too many subspecialists to an already richly supplied specialist pool will increase health care costs and perhaps decrease quality of care. Subspecialty training programs also have a moral responsibility to trainees, one that demands they adjust their training programs to the changing marketplace. Already, one medical school (University of California, Los Angeles) has reduced the size of its subspecialty training programs, focusing on the production of physician-scientists [7].

The timing is right for internal medicine subspecialties to actively participate in the national work-force debate. Their challenge will be to evolve from a reactive posture that merely tries to preserve the status quo toward a coherent vision of subspecialty training for the future. This will likely be a contentious and conflict-ridden process, but it will be essential to ensure ongoing innovation in biomedical research, the production of an appropriate number of highly skilled practitioners, and the capacity to adapt over time to changes in science, academia, and society.