Growth of Specialization in Graduate Medical Education

Since the beginnings of standardization and accreditation of US graduate medical education (GME) in the early 20th century, new areas in education and practice have continued to emerge, generally as a result of technological and scientific advances. However, this growth has occurred with little centralized planning or coordination. For nearly 100 years, numerous groups in the United States have struggled with societal issues related to the growth in medical specialization. Some see specialization as both a prerequisite and a logical outcome of human ingenuity in understanding and combatting disease; others attack it as unnecessarily fragmented,¹ expensive,² dehumanizing,³ and confusing for patients.

In 1933, an editorial in JAMA noted "the overgrowth of specialism, now so bitterly complained of, and the fadeout of the general practitioner."⁴ More than 30 years later, in 1966, the Citizens Commission on Graduate Medical Education concluded, "The rise in specialization has been accompanied by an alarming decline in the number of physicians who devote themselves to continuing and comprehensive care of the whole individual."⁵ In the early 1990s, with health care system reform at the forefront of the political agenda, the topic of the number and mix of educational programs and their relationship to population needs was elevated to the level of national debate. While Martini⁶ concluded in 1992 that the US GME system responded "more promptly to the professionals' interests and institutions' service needs" than to "the health needs of the population" in determining "the type and number of health manpower resources," others pointed out the difficulty of even reaching agreement on basic terminology to arrive at accurate numbers.⁷ As we describe below, national professional, educational, accrediting, and certifying bodies each have independent, although frequently interrelated, criteria and mechanisms for recognizing, approving, or accrediting new areas of education or practice.

The impetus for the growth of specialization may be more complex than the forces identified by Martini.⁵ While "the professionals' interests and institutions' service needs" do indeed play a role, they are joined by other societal forces, including changes in medical issues, patient needs, demographics, and social history. Such changes have led to the creation of entirely new treatment areas (while other areas, such as tuberculosis, syphilology, and tropical diseases, have been subsumed by other specialties). Advances in medical technologies and procedures, especially since 1945, have resulted in widespread acceptance of highly technical procedures and specialists limiting their practices to disease areas or to specific organs or body parts.

The history of GME accreditation in the early part of the 20th century is shrouded in inconsistent use of terms and archaic language and semantics. The term "graduate medical education" itself appeared in JAMA during the first half of the century in reference not only to internship and residency education, but also to continuing medical education. Other terms referring to GME included "continuation education" and "postgraduate continuation education." By the 1920s, the internship had become an accepted part of preparation for general practice, but specialty training was still largely unregulated and disparate. The multiple possible routes to practice included some type of residency, "postgraduate" study in Europe, and formal coursework (sometimes as brief as a few weeks⁸ ). In 1921, the Medical Education issue of JAMA listed "graduate courses in medical schools" and 18 "graduate medical schools," "commonly referred to as postgraduate medical schools." The Graduate School of Medicine of the University of Pennsylvania, for example, offered "courses extending over from four to twelve months in medicine, pediatrics, neurology, dermatology, syphilology, roentgenology, surgery, gynecology, obstetrics, orthopedics, urology, proctology, opthalmology [sic], otolaryngology and the medical sciences."⁹

The precursor to today's Graduate Medical Education Directory,¹⁰ first published by the American Medical Association (AMA) Council on Medical Education and Hospitals in 1914 as the "Provisional List of Hospitals Furnishing Acceptable Internships for Medical Graduates," listed 603 hospitals, including 35 "State Hospitals and Hospitals for the Insane" and 95 "special hospitals." This last category, comprising the "listing of specialty services as internships," was discontinued in the mid-1920s¹¹ when the Council on Medical Education and Hospitals began maintaining a separate list of approved residencies. An important step in the standardization of GME occurred in June 1923, with the AMA House of Delegates' adoption of the "Principles Regarding Graduate or Postgraduate Medical Schools,"¹² a precursor of the "Essentials of Accredited Residencies in Graduate Medical Education: Institutional and Program Requirements."¹⁰ The document states that the student should be able to "obtain progressive work in a continuous course of two or three years, as may be necessary to prepare him satisfactorily for the practice of a chosen specialty." A footnote details "the minimum years essential to ensure efficiency" in "the several fields of clinical specialization," which totaled 11 at that time (Table 1).

Although the requirements for a standardized minimum program length instituted in 1923 met with some resistance, supporters countered efforts to shorten the preparation period by pointing out the "insistent demand of new subjects and new fields of technical training for recognition in the curriculum" and that "professional skill and vigor are not decadent in the ‘early fifties.' "¹³

Published alongside the principles in JAMA in 1924 was a list of 222 "graduate courses in medical subjects and residencies in specialties" at 44 graduate medical schools (including 3 in London and 1 in Bordeaux, France). These schools offered residencies in 11 recognized "fields of clinical specialization" (Table 1), as well as courses in 34 areas of medicine, some recognizable today (eg, cardiovascular disease and gastroenterology) and some that have disappeared, been subsumed by other areas, or been renamed (eg, dental surgery, infant feeding, medical hydrology, and proctology).

Publication of the principles and implementation of the related review process must have aroused some degree of ire within the medical education community. A statement published in the 1925 JAMA Medical Education issue defended the intent of the Council on Medical Education and Hospitals to standardize GME, root out abuse, eliminate inadequate programs, and keep substandard physicians from getting into GME: "The Council is not attempting to prevent any physician from getting any course he needs, nor is it trying to assume authority over any individual or any graduate teaching institution, or to prevent them from offering to physicians any courses they see fit. If an institution desires to have the Council's endorsement of its work or of any particular course offered, however, it should not be unwilling to comply with the principles which have been laid down and which are believed to be fair and just."¹⁴

It was not until 1927 that a list of "Hospitals Approved for Residencies in Specialties"¹⁵ in the United States or Puerto Rico was published by the AMA.^{5 ,16} The list included 270 hospitals in 14 different areas of medicine; today, approximately 1700 institutions sponsor residency programs in 103 of the 107 specialties and subspecialties with Accreditation Council for Graduate Medical Education (ACGME) program requirements.¹⁰ Between 1927 and 2000, the number of specialties and subspecialties has fluctuated, with certain changes in numbers resulting from how specialty areas were grouped. For example, ophthalmology and otolaryngology were sometimes considered as 1 unit, sometimes considered separate, and sometimes considered as both simultaneously.

It is important to note that the frequently reported rapid rise of specialization after the end of World War II in 1945^{17 - 18} resulted from increased enrollments in existing specialties and subspecialties rather than from the approval of new specialties and subspecialties. Indeed, from 1940 to 1950, the number of approved specialties declined from 30 to 28. During the same decade, however, the total number of available residency positions more than tripled, from 5118 to 18,669, according to figures from the AMA and the US Department of Health, Education, and Welfare.¹⁹ Because the numbers of medical school graduates during that decade increased only slightly, from 5097²⁰ to 5553,²¹ it appears that practicing physicians, many of them returning from active duty, were choosing further specialization over general practice. As these specialists entered practice, the percentage of full-time specialists (vs part-time specialists and general practitioners) increased from 23.5% of the total physician workforce in 1940 to 36.2% in 1950.^{19 (p181)}

The rise of specialization in the 1940s was at least partially a function of the increased prestige of specialists vs general practitioners,^{3 (p440)} as reinforced by federal government policies. During World War II, board-certified specialists were accorded a higher military rank, along with higher pay, than general practitioners and non–board-certified specialists.²² According to Campion, "At the moment a physician (under age 37) presented himself for military duty, it was immediately impressed upon him that there was a difference between someone with specialty credentials and someone without them."^{3 (p33)} In addition, due to the pyramidal structure of GME, the general practitioner—with less training than a specialist—was "identifiable by what he lacked, rather than by what he had."^{19 (p301)} Finally, the passage of the GI Bill²³ allowed thousands of veterans to receive tuition and living expenses for residency education, and many returning veterans who were physicians used the opportunity to pursue specialty training.

Although the number of accredited specialties in the AMA's list of residency programs more than doubled between 1927 and 1933, from 14 to 28, only 13 additional specialties and subspecialties were added during the next 52 years, for a total of 41 in 1985 (Figure 1). The next 15 years, however, saw a precipitous increase in the number of specialties and subspecialties with accredited programs, accompanied by increased complexity and specificity in the institutional and program requirements, as indicated by the documents published in successive editions of the Graduate Medical Education Directory.¹⁰

Growth in the number of specialties and subspecialties and the increased detail in the program requirements had continued with the formation in 1981 of the ACGME, which replaced the AMA's Liaison Committee on Graduate Medical Education, originally formed in 1972.²⁴ As of January 1, 2000, the ACGME accredited nearly 7700 residency programs in 103 specialties and subspecialties (program requirements in 4 subspecialties have been approved but programs have not yet been accredited). With its recent incorporation in June 2000, the ACGME has 5 organizational members: the American Board of Medical Specialties (ABMS), the American Hospital Association, the AMA, the Association of American Medical Colleges, and the Council of Medical Specialty Societies.

In June 1992, in light of the unprecedented increase in the number of subspecialties between 1987 and 1992 (from 21 to 56), the ACGME declared a 1-year moratorium on considering accreditation requirements for new subspecialties. (Similar growth had occurred in physician certification, with the ABMS having recognized 24 specialty boards and 70 specialized areas, compared with 30 in 1979.⁶ ) Some of the reasons for the moratorium were identified by Martini in September 1992: "Proceeding without successful planning or coordination, [the] increase in subspecialties extends the length and cost of training, may directly affect the volume and cost of medical services, and presents risks to the perception of medicine as an integrated profession."⁶

The moratorium had little if any long-term effect in reducing the number of new specialties and subspecialties, although the number of current subspecialties is difficult to compare with previous years because of 2 recent developments. First, some institutions sponsor ACGME-accredited programs combining 2 subspecialties as well as programs in the individual subspecialties: hematology/oncology and pulmonary disease/critical care medicine can be combined in a single program or offered as separate programs, for a possible total of 3 programs in the same institution. The trend appears to be toward offering a longer single program combining 2 subspecialties rather than offering 2 separate programs, so that the number of programs may appear to decline when, in fact, the same type of training is still being offered. Second, in some cases, the same subspecialty training may be overseen by as many as 4 different specialties governed by different residency review committees; hand surgery training programs, for example, may be sponsored by orthopedic surgery, plastic surgery, or general surgery programs (Table 2).

This type of fragmentation is just the opposite of the dermatopathology model, for instance, in which programs are reviewed by a joint activity of the dermatology and pathology residency review committees. Similarly, there have been ongoing negotiations between the neuroradiology and diagnostic radiology residency review committees to develop ACGME-accredited residency programs in the subspecialty of endovascular surgical neuroradiology. There are selected efforts to develop a single set of standards and a single review process for accrediting programs in pain management as a subspecialty of psychiatry, neurology, and physical medicine and rehabilitation.²⁵ In February 2000, the ACGME instituted a new process for assessing and recognizing new subspecialties, similar to the criteria of the Liaison Committee for Specialty Boards (LCSB).²⁶

The trends of subspecialization in GME are paralleled by changes in the physician certification process governed by the ABMS, although the ABMS issues certificates in areas beyond those for which the ACGME accredits programs. Established in 1933, the ABMS is the umbrella organization for the 24 approved medical specialty boards in the United States. These boards are approved by the ABMS and the AMA Council on Medical Education through the LCSB. Applications for recognition as a specialty board are submitted to the LCSB and reviewed for compliance with the requirements and criteria published in the "Essentials for Approval of New Examining Boards in Medical Specialties,"²⁷ approved by both the ABMS and the AMA. An applicant approved for recognition as a medical specialty board by the LCSB must then be approved by ABMS members and the AMA Council on Medical Education.

The governing body of each member board is composed of specialists qualified in the specialty represented by the board. The individual member boards evaluate physician candidates who voluntarily seek certification, determining whether candidates have received appropriate preparation in approved residency programs in accordance with established educational standards, evaluating candidates through comprehensive examinations and certifying those candidates who have satisfied the board requirements. Physicians who are successful in achieving certification are called diplomates of the respective specialty board, with recertification typically required or offered to qualified diplomates at intervals of 7 to 10 years.

The 24 ABMS member boards currently offer 37 "general certificates" (specialties) and 87 "subspecialty certificates,"²⁸ for a total of 124 areas (Figure 1). As in GME, growth has accelerated in recent years, with the areas in which the ABMS offers certificates nearly doubling from 65 to 124 between 1985 and 2000. The 5-year period between 1970 and 1975 also demonstrated rapid growth, from 36 to 55 areas. As Swanson notes, "The forces and motivations that have caused this accelerated demand for subspecialty certificates and training programs have come largely from the constituencies of the certifying boards. Their desire for special recognition of narrower and narrower areas of knowledge and/or techniques will probably continue."¹⁶ In addition, as the areas of practice reflected by the 137 self-designated boards grow in numbers of practitioners and in increased professional standing, these groups will also likely seek recognition by accrediting and certifying bodies.

The specialty practice areas self-reported by physicians on periodic AMA surveys and captured in the AMA Physician Masterfile frequently reflect emerging subspecialty areas. Although the "self-designated specialty/area of practice" appearing in AMA physician records is not necessarily tied to special training or competence, "most self-designated practice specialties are consistent with physicians' graduate medical training and/or board certification," according to studies by the US Department of Health and Human Services, Bureau of Health Professions²⁹ (and unpublished AMA data, 2000).

A comparison of specialties and subspecialties with ACGME program requirements, general and subspecialty certificates offered by the ABMS, and self-designated practice specialties used in the AMA Physician Masterfile indicates that the areas of subspecialization are not always congruent. Some of the differences simply reflect the failure to develop agreed-on names for particular professions; other differences arise from the different purposes of the lists (tracking education, certification, and physician practice, respectively). The ACGME (107 entries) and ABMS lists (124) are more congruent than either is with the list of self-designated practice specialties (170) on the AMA's Census of Physicians (formerly known as the Physicians' Practice Arrangements survey).³⁰ The outliers in the list the AMA offers to physicians are usually subsets of the ACGME and ABMS lists (Table 3). These include new areas that may eventually be considered for ACGME or ABMS recognition (eg, abdominal surgery, hospitalist, and palliative medicine); ACGME specialties that also appear as separate areas (eg, allergy and immunology or obstetrics/gynecology); and specialty titles that are becoming less prevalent (eg, general practice and proctology).

The AMA Census of Physicians list also includes 2 of the 18 current "combined specialties" (internal medicine/pediatrics and neurology/diagnostic radiology/neuroradiology) with residency programs published in the Graduate Medical Education Directory.¹⁰ These combined specialty programs are not to be confused with ACGME-accredited programs that combine hematology/oncology or pulmonary disease/critical care medicine. The combined specialty programs on the AMA list and in the Graduate Medical Education Directory are approved by 2 or more of the 24 specialty boards of the ABMS. The ACGME does not accredit these combined programs but accredits each specialty program separately.

The appropriate mix of generalist and specialist physicians has been an ongoing topic of debate. Many of these points of view were presented at a 1998 conference hosted by the Council of Medical Specialty Societies. Founded in 1965 to provide an independent forum through which medical specialists could discuss issues of national interest and mutual concern, the Council of Medical Specialty Societies is composed of 17 member national speciality societies representing physicians in specialties that have primary or conjoint certifying boards recognized by the ABMS.

Presenters at the 1998 conference reflected the viewpoints of internal medicine, family medicine, the surgical specialties, the ACGME, the ABMS, the Bureau of Health Professions of the Health Resources and Services Administration, the Council on Graduate Medical Education, workforce policy analysts, and the public interest. Jordan J. Cohen, MD, president of the Association of American Medical Colleges and former Council on Graduate Medical Education chair, stated that "there is broad consensus among experts that the current mix is still too skewed toward specialist physicians."⁷ Other presenters, however, agreed with Herbert S. Waxman, MD, senior vice president for education, American College of Physicians-American Society of Internal Medicine, that "more questions than answers surround the issue of appropriate balance between generalism and specialism in medical training."⁷ Two years later, multiple factors continue to confound efforts to address issues surrounding specialization, and solutions appear no closer now than they did 100 years ago.