The physician in graduate medical education not only learns the skills needed by a practitioner, but also takes care of patients, serves as an instructor for medical students and conducts clinical research. These multi ple functions of the education process are indispensible to quality health care in this country. However, now they are threatened from both outside and within the system. Externally, a widespread and growing concern over the cost of health care has attracted unprecedented scrutiny of graduate medical education. Internally, fragmented organizational influences have diffused controls of the education process to a point that the functions of such education can be preserved only by decisive action. The cost of graduate medical education is quite small in comparison with total health care expenditures-hardly two percent. In the present national mood of economy, however, every component of health care costs is subject to close examination, and the training of new physicians is no exception. The problem of funding graduate medical education offers an opportunity to correct the problem of its disjointed organizational structure. The creation of a successful funding mechanism should force the various auspices under which the education is conducted to realign themselves and assume appropriate shares of responsibility for the enterprise. This Task Force report views the size, content and cost of graduate medical education. The report presents principles, policy options and recommendations that the Task Force believes can set a direction for constructive and significant change in the academic health center. Among the questions facing medical educators and public policymakers are whether we are training too many or too few doctors for the services most needed, such as primary care. The Task Force makes these suggestions consonant with its view that graduate medical education is a national resource that must be supported by all payers for health care. It suggests that limits be set on numbers of residencies or length of training. It recommends that formal connections be established between the organizations that have decision-making powers over residencies and the hospitals who, with revenues from patient care, support residency training. It recommends that training be broadened to include more exposure to ambulatory outpatients. And it recommends a voucher system for more equitable distribution of funds for graduate medical education. GRADUATE MEDICAL SYSTEM WORKS The physicians in graduate medical education are called residents; they make up a teaching hospital's house staff. Both terms connote the historical closeness of young doctors to the teaching hospital where they are receiving their advanced clinical training. The residency prepares a physician for practice in a medical specialty; there are now 24 specialties (listed in Appendix I). In a period ranging from three to eight years, depending on the specialty, the residency imparts the knowledge and experience required for certification of the physician by one of 23 medical specialty boards. Further time in residency can lead to certification in a subspecialty, of which there now are 42. For example, graduate training in the specialty of internal medicine can lead to any of 11 subspecialties, such as gastroenterology or nephrology. Training beyond that required for board certification is often achieved in a fellowship, which typically emphasizes a path of research toward subspecialty expertise. Teaching hospitals range from smaller affiliates of academic health centers, with only a few residency programs, to major institutions that offer training in almost all of the specialties. Of the approximately 7,000 hospitals in the United States, about 1,500 have some residency programs, and thus meet the definition of teaching hospitals. But 46 percent of all residents in the United States are trained in only 100 hospitals, which is to say that hardly more than one percent of all the hospitals in the country handle *A glossary of terms used in describing the organizations and processes of graduate medical education is in Appendix I of this report. the graduate training of almost half of all medical residents (Smith and Stemmler, 1984). Graduate medical education began early in this century as apprenticeship programs in a few institutions or offices of established physicians. For many years, the majority of physicians trained to be general practitioners. Specialty boards, which certify physicians appropriately prepared to practice a specialty, originated in 1917 with ophthalmology. Otolaryngology was next, in 1924, and then a flood of specialty boards rose-13 in the next decade. (Specialties and subspecialties are listed in Appendix II.) By the 1960s, the Federal government was developing a concern that the country had too few physicians for its future needs. Existing medical schools were encouraged to expand, and new ones were founded. Specialty training programs were opened up to receive the influx of newly graduated M.D.s. The new Medicare and Medicaid programs included provisions to pay their share of graduate medical education. Not surprisingly, the number of medical graduates seeking specialty training surged from 29,000 in 1950 to 66,000 in 1973, and then to 72,000 in 1983. The growth of graduate medical education over the years has brought a number of organizational participants into the process, but they have never integrated into a coherent system of control. Instead, a complex arrangement of partial controls variously affects the accreditation of residencies, the certification of specialists and the conduct of the programs themselves. The effects of these fragmented controls are described in the following sections. The American Medical Association (AMA) established a body in 1927 to oversee graduate medical education. During the next fourscore years the body's name was changed several times, but it remained solely under AMA control. In 1972, the body's participation was enlarged to include the American Hospital Association, the Council of Medical Specialty Soci. eties, the American Board of Medical Specialties and the Association of American Medical Colleges. That body is now called the Accreditation Council for Graduate Medical Education (ACGME). It delegates its accreditation responsibilities to 24 Residency Review Committees (RRCs), one for each of the specialties. The members of RRCs are appointed by the AMA and the relevant specialty boards; in some instances, appointments also are made by a specialty society, for instance, the American College of Surgeons appoints to surgical RRCs. Appendix ! defines specialty boards and societies; Appendix III lists RRCs and their sponsoring organizations. The ACGME lays down general requirements for residency programs, and each RRC draws up particular requirements, tailored to its specialty, that follow the ACGME guidelines for faculty, administration, program content and the like. RRCs also set the minimum length of program and may further stipulate the educational procedures and the number of residents that may be in each year of training. The RRCS make alterations in their requirements as circumstances warrant. These alterations are subject to ACGME approval, but they are rarely countermanded. The five parent organizations of the ACGME and the spe |