Today's aggregate oversupply of physicians and the imbalances among generalists and specialists have been predicted and analyzed for at least 15 years [1, 2]. Although shortages of primary care physicians and overabundances of subspecialists are now apparent [3-5], the profession has had only limited success in increasing the numbers of primary care physicians and has been unable to stem the production of more subspecialists.

Graduate medical residency and fellowship training depends to a large extent on public funds from Medicare, Medicaid, the Veterans Administration, federal grant programs, and state support of universities and teaching hospitals. No other profession, including law, requires or receives such extensive public investments in education and training. Meaningful planning regarding the overall number of postgraduate medical training positions and their distribution among specialties is required to protect these investments, as well as the investments of effort, time, and money by medical students and residents.

Abundant precedent exists for treating the medical profession differently from other occupations. The "doctor draft" of the Korean War, the Health Professions Educational Assistance Act of 1963, the Comprehensive Health Manpower Training Act of 1971, and subsequent federal legislation recognized health professionals and the educational institutions that train them as important national resources and provided federal funds for construction of medical schools, student loans and scholarships, and faculty development. Government programs were highly successful in increasing the number of medical schools and in expanding the nation's supply of physicians during national physician shortages. It is no less appropriate today to involve government in "right-sizing" the educational pipeline.

Freedom of choice regarding a medical career is an illusion: Entry into various career paths, particularly in subspecialty medicine, is strictly limited. We have not proposed that anyone be coerced into a career that he or she does not desire, that government control admissions to medical schools, or that government dictate where or how a physician should practice. However, the overall number of residency training positions should be somewhat related to the supply of graduating U.S. medical students, training opportunities for international medical graduates, and work force requirements. Consequently, the College's Task Force on Physician Supply proposed a process for developing national health work force goals, setting objectives to achieve those goals, and monitoring changing supply and requirements. This process should be implemented over a sufficiently long period to allow for institutional planning needs and to minimize disruptions for residents and faculty. These positions were endorsed by both the College and the Federated Council for Internal Medicine.

Many factors in addition to the marketplace affect the number and distribution of postgraduate training positions. Factors such as institutional service needs, the needs of existing training programs and faculty, and the availability of funding do not necessarily best serve either the ongoing needs of patients and communities or the physicians who, we hope, will enter careers that will last for 40 years or more after completion of residency training. We agree that the most desirable and effective way to increase the number of primary care physicians over the long term is through positive incentives, increased payments for primary care services, and improvements in the practice environment. The marketplace will ultimately have considerable influence on medical education, but the long medical education pipeline and the costs involved necessitate national planning and coordination to better match supply with requirements.

We also agree that no hard scientific data support a 50/50 mix of specialists to generalists or for any other optimal mix. The goal can be adjusted as better data become available, but the current 68/32 ratio cannot be sustained. We are also mindful of the necessity to constantly monitor and refine work force projections because they quickly can be rendered obsolete by unexpected outbreaks of disease, new technologies, changes in medical practice, and many other factors. Physician-to-population ratios should also be considered, but they too have their faults. Adjustments would be required to reflect demographic variations, differing needs among health care delivery systems, and other factors. Hard evidence on the most appropriate physician-to-population ratios is also lacking.