Specialization and the Physician Workforce: Title and subTitle BreakDrivers and Determinants

Specialization in medicine has been a virtually inexorable trend for many years. In the modern era it was heralded by the progressive separation from general practice of obstetrics and gynecology, pediatrics, internal medicine and surgery, and the surgical subspecialties. These movements began shortly after the turn of the century. The first US specialty board, in ophthalmology, was established in 1917, followed shortly by a dozen other specialties, including the American Board of Internal Medicine in 1936 and the American Board of Surgery a year later. The American Board of Medical Specialties added its 24th member board, the American Board of Medical Genetics, in 1991.

The growth of specialism and, more recently, subspecialism, has been, naturally enough, embraced by physicians seeking special expertise in various fields and recognition as well. Impetus has been added by status considerations, higher incomes earned by subspecialists, and the emerging need for board certification in attaining hospital appointments and other administrative necessities.

In this issue of THE JOURNAL, Donini-Lenhoff and Hedrick¹ document the striking rise in specialization and, more particularly, in subspecialization in the years since World War II. While this trend has been irregular, there has been a steady increase in the number of recognized areas of special expertise and of derivative clinical societies, training programs, and specialty boards. Alongside the growth of accredited specialties and subspecialties in which board certification is available through the 24 member boards of the American Board of Medical Specialties, a large number (137 at last count) of self-designated boards has arisen, representing a variety of fields of interest, some reflecting further subdivision of already recognized areas (eg, facial cosmetic surgery, ambulatory anesthesia, and neuroimaging), others based on site of service (eg, ringside medicine and surgery) or areas of expertise already within the expected competencies of existing specialties (eg, pain and palliative medicine) and some (eg, homeopathy, hair restoration surgery, and managed care medicine), the justification for which might be argued.

As Donini-Lenhoff and Hedrick point out, the trend to subspecialization is likely to continue. Recent years have seen finer and finer subdivision of clinical fields, various combinations in graduate medical education training programs, and continuing pressure to have special expertise formally recognized in these ways. It is reasonable to ask where the asymptote may lie and how to find guiding concepts to bring order to this progressive dicing of clinical activities, especially since there is wide agreement that the subspecialty pendulum has swung too far.^{2 - 4} Sharp increases in subspecialty training programs occurred in the 1970s and 1980s, as the research fellows of the 1960s became clinical fellows, then subspecialty residents,⁵ heavily used in meeting the service needs of hospitals and supported by Medicare or hospital funds. By 1990, nearly two thirds of trainees completing 3 years of internal medicine residency were entering subspecialty training, particularly in cardiology and gastroenterology.⁶ Efforts to relate this growth to the national need generally reached the conclusion that subspecialists were still being overproduced in substantial numbers,^{7 - 10} and that the need for generalists had largely been met by the early to mid-1990s. By 1995, some two thirds of US physicians were subspecialists, and only 35% were generalists.³

The primary drivers of the trend to subspecialization have included the burst of new knowledge flowing from the dramatic rise and productivity of biomedical research since World War II, the array of technology deriving from those advances, and a widespread desire among physicians for related expertise. Board certification in subspecialties has no doubt facilitated the movement, and in this sense the specialty boards, although following pressures from their constituencies, have also contributed to the trend. The chief drivers, however, have included the view that the future of clinical practice lay in subspecialty expertise and that generalists would lose capacity to manage serious or complex disease as the knowledge base of clinical medicine continued to expand.

The large number of first-year training slots in US graduate medical education programs has substantially exceeded the number of US medical graduates for more than a decade. Much of this latter bulge has been met by international medical graduates, who, in the mid-1990s, accounted for approximately 25% of all first-year residents⁵ ; about three quarters of these international medical graduates entered the US physician workforce after completing training in US programs. By 1992, as Dalen⁵ has noted, 23% of all US practicing physicians were international medical graduates. These physicians are even more likely than US graduates to become subspecialists.¹¹

Turf considerations in clinical practice and medical academe have also been important in building special areas; such tugs-of-war continue and are resulting in efforts to accredit training programs as subspecialties of multiple fields, for example, as the authors note, pain management, currently under negotiation as a subspecialty of psychiatry, neurology, and physical medicine and rehabilitation.

What kinds of derivatives can be identified as flowing from these trends? Without question, subspecialty expertise has offered clinical benefits to thousands of patients with uncommon disorders, critical illness, and disorders requiring special techniques. The proportion of individuals with diseases germane to particular subspecialties who should get their ongoing care from a subspecialist is less clear; except for a few instances,¹² careful patient outcome studies have not yet demonstrated a clear outcome advantage for ongoing subspecialist care.

On the other hand, it is clear that subspecialist practice contributes to fragmentation of patient care and diffusion of responsibility for clinical care over time and over multiple disorders. In addition, subspecialist care is by and large more costly than generalist care and tends to involve greater use of technologies and to result in referral patterns to additional subspecialists for difficulties outside the expertise of the referring physician, both important cost drivers. The cost issues have come under increased pressure in recent years with the emergence of managed care and the gatekeeper mechanism for limiting access to subspecialists and some specialists. Very considerable resistance has emerged with relation to these arrangements, as well as to their implications for excessive practice scope visited on primary care physicians by this arrangement.¹³

One might speculate that additional emerging pressures may counteract the increasing trend to subspecialization, notably the development of a broader research base in medical outcomes and concern over medical errors. In addition, it is possible that regulation of funding for graduate medical education and research may reduce the pressures for continuing growth and that consumer advocacy may also contribute forces; however, there are increasing indications that the subspecialty worm may be turning. Thus, in internal medicine, the number of first-year fellows in the 9 traditional subspecialties declined by 21% between 1992 and 1998,¹⁴ interestingly in the most procedure-oriented subspecialties, namely pulmonary/critical care, gastroenterology, and cardiology. This decline has occurred despite stability in the number of third-year residents in training and has occurred in the face of sustained interest in general internal medicine among trainees.

Among the most powerful determinants of this falloff in subspecialty interest is the widespread perception among internal medicine residents that the job market in the subspecialties has tightened considerably and is likely to continue in that mode, both in practice and in academic careers.¹⁴ Local market forces appear to be particularly important, along with personal characteristics, such as age and sex.¹¹ In addition, the Association of American Medical Colleges, the clinical literature, and a variety of federal agencies have all emphasized in recent years the importance of a shift back in the direction of more desirable generalist-to-specialist ratios in the physician workforce.

One possible answer to the continuing growth of subspecialism is a return to programs seeking to develop sophisticated generalists capable not only of primary care, but of secondary and a good deal of tertiary care as well.^{15 - 16} Leadership by the medical profession is desperately needed to address this problem. An innovative approach reported by Fogelman⁷ in 1994 combined an effort on a university hospital medical service to discontinue the training of subspecialists for practice, confining such training to individuals intending academic/research careers within the subspecialty, and to link this change to an expanded general internal medicine residency training program, characterized by an additional fourth year composed of training in 2 subspecialties and the option for trainees to elect to learn common procedures germane to those subspecialties. Individuals entering this additional year of training would be guaranteed a faculty position after successful completion of the program. Additional experiments would be useful, especially if linked with patient outcomes studies.

Further development of a more rational physician workforce mix must be based in alteration of the nature and priorities of US graduate medical education training programs. These should be organized not only around aggregate measures of national need: in addition, higher priority must be given to the needs of individual patients to have clinical care that can see them through the bulk of their medical needs over time and various clinical vicissitudes, and by clearer definition of the need of the individual patient for subspecialty care.

Alteration of the structure of the health care system, including the physician workforce mix, can only be informed if the education and training system in medicine joins with the training program accreditation and board certification processes to come at these issues primarily through the lens of patient and population need, rather than the needs of the profession.