Ludmerer, Kenneth M. MD
Does graduate medical education (GME) in the United States need to be reformed? This may seem a surprising question about a residency system that remains the envy of the industrialized world more than a century after it was pioneered at the Johns Hopkins Hospital in 1889.1 American physicians, as a group, are highly knowledgeable and skilled, and they have demonstrated the ability to acquire new knowledge and master new techniques throughout their professional lifetimes. At a 1992 Josiah Macy Jr. Foundation conference on GME, Alexander Walt, president of the American Board of Medical Specialties, defended the consistently “excellent” quality of the graduates of residency programs from this country.2
Nevertheless, almost from its very beginning, GME in the United States has been the subject of controversy and criticism. At the 1992 Macy Foundation conference on GME, Samuel O. Thier, then president of the Institute of Medicine (IOM), spoke of the issue of GME as “a hardy perennial.”2 By this he meant that the issue of GME “keeps coming back, it consistently interests people, and its problems never quite get solved.”3 A persistent concern throughout these many decades since the inception of the residency system of training has been whether GME is sufficiently addressing the health needs of the nation.
These criticisms of GME have been formalized in a series of official reports dating back to 1940 (Table 1). Here, I provide an analysis of these reports in the hope of achieving a better understanding of GME and the challenges to improving it. For the sake of simplicity, I have confined the discussion to official reports in which GME was the sole focus or a major focus of investigation. Not included in this analysis are editorials, sounding board pieces, or commentaries by individual authors writing in an unofficial capacity.
Meeting the Health Needs of Society
A persistent theme through all the official reports on GME is the charge that GME is failing to sufficiently address the health needs of the nation. Typically, the reports express respect for the high quality of American physicians and the lofty professional standards of the training programs that produce them and the specialty boards that certify them. However, despite this admiration for the technical abilities of individual physicians, the reports assert that the system of GME as a whole is somehow dysfunctional, focusing on the needs of individual practitioners without considering the larger needs of the health care system or society.*
What requirements of the broader health care system have been perceived as unmet by GME? The different reports answer this question in different ways. In the 1930s, the residency system was still struggling to establish itself, as it faced fierce competition from other recognized pathways to specialization. Training in Europe, working in a specialty outpatient clinic at a teaching hospital, serving as an assistant to an established specialist, attending a short graduate course at a proprietary postgraduate school, and obtaining a PhD from a medical school in a clinical field were among the established routes to specialty practice available at the time.1 Thus, the Rappleye report, the informal name for the report from the 1940 Commission on Graduate Medical Education,3 is devoted to rationalizing the system of GME in the United States. This report defines and describes residency to the American public, distinguishes residency from internship and continuing medical education, and argues that residency should become the sole recognized path to specialty practice in the United States. As stated in this report, “When a residency is set up with proper educational standards, it is the most effective, economical, and satisfactory method for obtaining this training.”3
By the 1960s, the residency system had long established itself as the sole route to specialization in the United States, and the country's system of GME had become the gold standard for the world. However, new concerns about GME emerged. Many worried that GME was too fragmented, that it should be explicitly recognized as part of a lifelong continuum of medical education. There were also frequently voiced concerns that GME, supposedly a field of graduate study, was too much influenced by the hospital rather than by the university. Many felt that residency programs within a hospital operated too independently from one another, oblivious to the concerns and needs of each other or of how they could work in conjunction with one another. These issues are confronted squarely by the Coggeshall report,4 produced by the Association of American Medical Colleges (AAMC), and the Millis report,5 from the American Medical Association (AMA). Both reports argue strongly that medical education should be viewed as a continuum, that the university should exert a more active voice in GME, and that hospitals should assume greater institutional responsibility for the operations of their various residency programs.
A decade later, new challenges in American health care had become apparent. Foremost among these was the perceived maldistribution of physicians by specialty type and geography. Health policy experts regularly bemoaned the preponderance of specialist physicians, the relative scarcity of primary care physicians, and the strong tendency of doctors to aggregate in affluent large communities but not in rural areas or inner cities. Reports from the Macy Foundation in 19806 and 19932 and from the AAMC in 19817 address these issues, criticizing GME for making no “serious attempt to relate the number and types of graduate medical education programs to national needs.”6 These maldistribution problems might be corrected, the reports argue, if GME programs were to take their responsibility to the public good more seriously. As expressed in the 1980 Macy report,
To a certain degree, policies directed at the graduate medical education process may be used to alter both the specialty and geographic distribution of physicians by influencing the recent graduate's choice of specialty and location.6
By the 1980s, soaring health care costs had become a major area of concern in the American health care system. The 1980 and 1993 Macy reports and the 1981 AAMC report take aim at this problem as well. All three reports chide GME for not doing its part to keep costs under reasonable control.2,6,7 Physicians, these reports point out, generated 75% of the costs of the health care system, and numerous studies had documented the tendency of physicians to overuse both diagnostic and therapeutic technologies.2 Accordingly, residency programs needed to do a much better job in teaching the wise and cost-effective use of resources. As the AAMC report recommends,
Teaching hospitals should increase their emphasis on research related to the effective use of resources and educating residents to utilize diagnostic and therapeutic procedures with due consideration of their contributions to optimal patient care and their costs.7
In the 21st century, the problem of health care costs has only grown worse, and reports on GME in recent years8–12 also take GME to task for not preparing residents to serve as better stewards of the nation's health care resources. These recent reports contain more pleas to residency programs to teach residents about the costs of what they do, to promote wise clinical decision making, and to advocate better use of published evidence in devising clinical strategies. The cost issue is not new, but the severity of the problem has grown, and some of these reports adopt a strident tone in challenging GME to better address the problem of rapidly rising costs.
However, the 21st century also brought with it striking new dilemmas in the health care system. Among these are the problem of medical errors, the need to improve safety and quality, the challenge of integrating the electronic medical record and other new information technologies into medical practice, the rising importance of chronic disease, the declining importance of the hospital as the locus of acute medical care, and the need to provide better coordination of care. Once again, official reports criticize GME because its formal curricula, in the words of the Medicare Payment Advisory Commission's (MedPAC's) 2009 report to Congress, “are not well aligned with objectives of delivery system reform.”11 Recommendations from recent reports include more and better training of residents in ambulatory care, systems-based thinking, quality and safety improvement, multidisciplinary teamwork, and information technology.
These reports reveal why GME has perpetually been vulnerable to the criticism that it poorly prepares residents to serve the emerging health care needs of society: These perceived needs are always changing. Over the past century, the scientific basis of medical thought, the technologic basis of medical practice, and the demographics of disease have profoundly changed, as have American society and the health care delivery system. The result is continually changing professional demands on GME, for what worked yesterday might not work today or tomorrow. Of necessity, GME is in the position of having to play catch-up; the capacity of medical knowledge and the changing demands of the American people are always a step (or two or three) ahead of current GME practices. Some might applaud GME for its adaptability and capacity to address the new problems that inevitably appear in health care delivery. Others might criticize GME for being too slow and awkward, or even insensitive in its efforts to respond. But the fact remains that GME is always chasing a moving target.
Education Versus Service
Although the health care needs of the nation continually change, as do the challenges confronting GME, one fundamental problem of GME has proven consistently intractable: the tension between education and service. This is the most ancient dilemma in GME, having plagued the 19th-century system of “house pupil” appointments, a predecessor of contemporary GME. With the development of the modern internship and residency in the late 19th and early 20th centuries, this tension grew. During this time, the economic exploitation of “house officers” became tradition, as hospitals from the start insisted that trainees perform an extraordinary range and amount of ancillary responsibilities. It was frequently unclear whether GME represented education or service, or whether house officers were students or hospital employees.1
House officers began their GME knowing they would be working extremely hard. The fundamental pedagogic principle of internship and residency called for house officers to develop independence by assuming responsibility for their patients' total care. This made hard work inevitable and caused great difficulty in separating the educational from the service component of GME. If a house officer's patient spiked a fever at night, for example, that same house officer would draw the blood specimens and carry them to the laboratory him- or herself, if necessary. However, hospitals and medical faculties usually required house officers to perform far more service than that which was required for learning. House officers, for instance, were typically expected to draw blood samples not only on their patients or in emergencies but also on all patients on a service, day or night, and, frequently, to transport routine specimens from any patient to the clinical laboratory as well.1,13
Traditionally, the greatest exploitation of house officers as a source of inexpensive labor occurred at community hospitals not affiliated with medical schools. Before World War II, at many of these hospitals, interns were considered subordinate to nurses and were permitted only to take routine medical histories and administer intravenous medications. Didactic rounds, teaching conferences, and other educational activities were few. However, even in the strongest programs, the amount of routine work could be overwhelming. At the leading teaching hospitals, house officers were deluged with innumerable duties—performing blood counts and urinalyses, transporting patients, drawing blood samples, and starting intravenous lines—for which a physician was hardly required. Complaints of too little teaching and too much “scut work” were commonplace.1,13
Since the mid-1980s, the nature of “scut work” has changed. Because most hospitals have introduced more extensive support services, residents are left with fewer blood samples to draw, fewer intravenous lines to start, and fewer patients and laboratory specimens to transport. However, more burdensome administrative chores have emerged to replace these tasks: scheduling tests and procedures, obtaining consultations, planning for discharges, and meeting the ever-increasing number of documentation requirements. More significant, the patient load of residents has increased dramatically during this time. A generation ago, an individual on a typical internal medicine resident service might receive 3 or 4 new patients each admitting day, and the average length of hospital stay was 10 or 11 days. After prospective payment for hospitals was introduced in 1984, the same resident might work up 8, 10, or 12 new patients in the same period, and the average length of hospital stay dropped to around 3 days. For residents in all fields, this change in hospitalization patterns resulted in busier days and nights, less time to read and sleep, and greater stress, tension, and fatigue.1,13
Indeed, in this era of “throughput,” a profound burden for keeping the nation's teaching hospitals financially solvent has once again been placed on the backs of the resident staff. Traditionally, hospitals benefited financially from house officers by virtue of the routine work these individuals performed. The presence of house officers allowed hospitals to hire fewer secretaries, nurses, blood drawers, transporters, and other support personnel. In the era of high throughput, hospitals have continued to benefit financially from their resident staff because each new admission represents a fresh payment to the hospital. Without a resident staff caring for so many patients and, more important, turning over the service so quickly, hospitals would rapidly find themselves in dire financial straits. Of course, in such a frenetic environment, the quality of patient care and education can easily suffer because the only way a physician can turn over so many patients so quickly is by cutting corners and devoting less time to educational activities. Once again, service trumps education.1,13
The economic exploitation of trainees has been recognized from the very beginning of GME in the United States, and most official reports on GME address this problem. For instance, the Rappleye report criticizes GME for de-emphasizing education for service. According to this report, to improve the educational value of GME, first and foremost hospitals “must work out plans to relieve the intern [and resident] from many routine procedures which he is now performing but which have relatively little educational value.”3 After the noneducational responsibilities are removed, the next step to improving GME is “by expanding its educational content.”3 The report argues that hospitals should hire salaried physicians rather than interns and residents if they cannot make adequate educational opportunities available for trainees.
Despite these pleadings, the subjugation of education to service continued, and this led to many additional calls for residency programs to take their educational responsibilities seriously. Of the reports considered in this article, the Millis report,5 the AAMC report,7 and the 1993 Macy report2 contain especially strong words to this effect. However, criticism of GME for exploiting residents and interns has not been confined to official reports. For instance, in the 1970s, the economic exploitation of house officers was a major factor in promoting the housestaff union movement.13 More recently, Jordan Cohen,14 in his term as president of the AAMC, repeatedly challenged teaching centers to make GME a genuine educational experience, and he famously spoke of the importance of “honoring the E in GME.”
Why should there have been so much resistance to lessening the service load in GME? Ongoing research has revealed that every component of the house of medicine in some way gains from the perpetuation of the current system.1 Medical faculties have long profited from the system because the presence of a talented resident staff has allowed them more time for their own research or, in more recent years, to see more private patients, thereby increasing their “clinical productivity” and enhancing their income. Private practitioners have similarly benefited from residents overseeing their hospitalized patients. Such an arrangement has made these physicians' lives richer and easier, allowing more time in the office or far fewer trips to the hospital from home on evenings and weekends. Hospitals, as noted, have long benefited financially from having a resident staff. In today's environment of admission maximization and high throughput, the reduction of residents' clinical workloads to more reasonable levels would be extremely costly for hospitals, which otherwise would have to hire additional physicians or midlevel practitioners to see the patients currently admitted and cared for by the residents.
In short, shifting the balance in GME away from service and more toward education has proven exceedingly difficult because the medical profession has become complicit in the status quo. Official proclamations about GME have always emphasized the importance of the educational experience, but these ideals have not been realized. Virtually every significant step that might be imagined to make GME a better learning experience would cost someone something in time, money, or both. Both hospitals and physicians benefit economically from the status quo, and medical faculty members benefit even more from the additional time they have for their own work when house officers are carrying the service load.
The tension between education and service underscores the financial dimension of GME. Since World War II, the number of residents in U.S. hospitals has increased exponentially, and residents, once accepting of room, board, and pocket change, now expect and receive respectable salaries somewhere near or slightly above the median for working Americans in their geographic region. GME has grown from a cottage industry to a multi-billion-dollar enterprise, supported mainly by direct and indirect educational payments from Medicare and, to a smaller degree, by private insurers, state and local governments, the Veterans Administration, and other sources. Further steps to improve GME—reducing the clinical workload of residents, providing more support with nonprofessional chores, providing for additional teachers and educational facilities, introducing new educational technologies, and developing curricula to teach new subjects—would only add further to the already considerable financial cost of GME. Thus, the questions of who should pay for GME, how much they should pay, and how residency programs should demonstrate their accountability for the large sums of money received become critical to resolving the long-standing tension between education and service. Only with sufficient funding can GME genuinely be improved and not be subjected to a continuous litany of criticism.
GME is far from monolithic, though it is frequently discussed as if it were. Programs greatly differ from one another in terms of quality, degree of academic orientation, size, location, patient population, and culture and traditions. The characteristics and requirements of residency programs also vary substantially from specialty to specialty.
Similarly, critics of GME are far from a monolithic group. Individuals and organizations bring their own orientation, perspectives, and biases to the conversation. This can lead to strikingly different assessments of GME or of what parts of GME are most in need of reform. These differing viewpoints can also lead to vigorous disagreement over what needs to be done to “fix” GME. These characteristics of the debate become evident in reviewing the earlier calls for reform.
Consider, for instance, the Coggeshall4 and Millis5 reports. Each is frequently likened to the other. Both were published around the same time, both argue that GME needs to be viewed as part of an educational continuum, both urge universities to exert a stronger voice in the conduct of GME, and both advocate that academic medical centers take stronger corporate responsibility for GME. However, they disagree on a major issue: which national organization should control and speak for GME. The Coggeshall report, sponsored by AAMC, argues that the AAMC should assume the leadership role for GME, and for all of medical education. The Millis report, sponsored by the AMA, argues that this leadership role belonged to the AMA's Council on Medical Education. Clearly, each report reflects the particular view of its sponsor.
This tendency for reports on GME to bear the perspective of their sponsor or organizer has been especially apparent during the past decade. Consider the IOM's report, Health Professions Education: A Bridge to Quality,8 published in 2003. Three years earlier, the IOM had published its most influential report ever, To Err Is Human,15 which exposed the problem of medical error and helped launch the safety movement as a public crusade. In Health Professions Education, the IOM echoes its earlier concerns, taking medical education to task for not satisfactorily teaching safety and quality. The solution, according to this report, is to incorporate into graduate (and undergraduate) medical education five core competencies so that all physicians, regardless of specialty, would be able to provide patient-centered care, work in interdisciplinary teams, employ evidence-based practice, apply quality improvement, and use informatics. Of note, this report does not address the internal learning environment of GME, the subjugation of education to service in residency training, or the problems of financing and regulating GME.
In its 2009 report to Congress, MedPAC, like the IOM, also criticized GME for not satisfactorily producing physicians who were able to meet the needs of the 21st-century health care system. However, echoing the concerns of its sponsor, Medicare, the MedPAC report focuses on the issue of cost containment. The great problem of GME, MedPAC argued, was the failure to teach sufficient “cost awareness in clinical decision making.”11 What was needed in GME was better teaching of cost-effectiveness, as well as multidisciplinary teamwork, information technology, and caring for patients in ambulatory settings—all devices, if properly employed, of proven effectiveness in reducing unnecessary health care costs. Like the IOM report, the MedPAC report does not discuss issues pertaining to financing or regulating GME or ways to improve the internal learning environment.
In contrast, the report of the Blue Ridge Academic Health Group, an organization of administrative leaders of academic medical centers, takes a much more detailed look at the internal workings of GME.9 As is the case in the other reports, the Blue Ridge Academic Health Group's report touches on many issues, expressing agreement, for instance, with the IOM's concern for improving the teaching of safety and quality improvement. However, the chief animus of this report is twofold: simplifying the complex regulatory structure of GME, and adopting strategies at academic medical centers to recruit, develop, and reward good clinical teachers. Such a focus was not unexpected from a group of prominent medical school deans and hospital presidents.
By far, the most extensive discussion of GME from the perspective of education appears in a 2010 report, Educating Physicians: A Call for Reform of Medical School and Residency,12 sponsored by the Carnegie Foundation for the Advancement of Teaching (the most heralded private foundation in the field and also the sponsor a century ago for the landmark Flexner report16). Unlike all the other reports, the Carnegie report focuses on the internal educational environment of GME, bemoaning the deterioration of learning conditions that had arisen from our current preoccupation with maximizing throughput in American health care. In this environment, the authors write,
Discharge becomes the highest goal. The imperative in the clinical environment is efficient patient management and swift disposition of problems; this task-focused environment is inhospitable to [intellectual] exploration.16
The authors describe a number of undesirable educational consequences of this environment, including less time for reading, fewer opportunities for reflection, a decline in the quality of teaching and supervision, and a shift in the character of GME from graduate education to vocational training.
Herein we see yet another reason for the failure of calls for the reform of GME to be heeded: the difficulty of taking a comprehensive approach to the problem. All the views described above are cogent and important, but each addresses only a piece of the puzzle, reflecting the particular perspectives and concerns of the sponsoring group. GME is a vast, intricate, complex enterprise, with each component affected by every other. It is exceedingly difficult to fathom or comprehend the enterprise in its totality, much less to devise solutions to its fundamental problems or gather the political will to implement those solutions. It is far easier to express some specific criticisms or suggestions and then relegate a report to a dusty bookshelf, assuming (or hoping) that this gargantuan enterprise, involving learners and teachers of intelligence, dedication, and good will, will somehow keep lumbering along.
The Limits of Education
The power of GME to provide physicians with the knowledge, attitudes, and techniques to practice medicine in a skillful fashion is beyond dispute. But what of choices physicians make regarding broader professional matters, such as what field to enter or where to practice? Choices like these pertain directly to the national good, as the widespread consensus since the 1960s has been that the country needs more primary care physicians and more doctors in inner cities and underpopulated rural areas.
A common criticism of GME, echoed in many of the calls for reform, is that GME has failed to serve the needs of the public because it has produced too many specialists and too few physicians willing to practice outside urban metropolises. If only medical faculties would do a better job encouraging these pursuits, so the argument goes, more graduates would enter primary care specialties and forsake big cities or comfortable suburbs. For instance, in an article entitled “Graduate medical education: Proposals for the Eighties,”7 the AAMC argued against an allotment system for residency positions based on “mounting evidence that medical students modify their career choices when there is a general agreement that a change in specialty distribution is needed.”
The empirical observation, however, is that the exhortations from medical educators have had little effect on many of the choices medical students and residents make. Primary care has long been at a disadvantage in relation to specialty medicine, not only because of income differentials (an increasingly important factor as student educational debts soar) but also because of the perceived greater professional satisfaction of specialty practice and, in many specialties, the perception of an easier lifestyle. Similarly, many studies have shown that a resident's choice of a practice site reflects his or her response to professional, social, and financial incentives and that, in these regards, cities offer greater advantages than rural areas—as they do for most other Americans.1 Thier spoke to this point in the 1993 Macy report:
The problem is not with graduate medical education. Rather, the problem lies in the way the nation reimburses for health care services and in the way the entire health care system is organized.2
GME, important as it is, according to Thier, “cannot transform the health care system.”2
These debates about residents' choices mirror a broader debate in American education over the capacity of education to influence behavior. The traditional orthodoxy of the American educational system has been the belief that education can shape behavior and mold character.17 Yet, many factors beyond formal education have also been seen to influence behavior. In his book on the history of American education, published in 1988, Lawrence A. Cremin18 pointed out that there have always been limits to formal education as a behavioral force. Behavior, he maintained, is shaped by innumerable “educative” influences—one of which is formal education, but which also include the totality of an individual's upbringing and environment, encompassing such factors as family, friends, neighborhood, religion, and popular culture.
Here, then, is still another reason the earlier calls for reform have failed to induce their desired effect: unrealistic expectations over what GME can and cannot do. The importance of GME in producing proficient physicians is indisputable, as is its capacity to influence doctors' values, attitudes, and behaviors. But, for many of the issues brought up in previous calls for reform, the lesion is misidentified. That is, the problem under discussion lies with the health care delivery system, not with GME as such. The health care system is the independent variable, and GME is the dependent variable. The only way to prevent disappointment and disillusionment with GME is to recognize that one or another educational “fix” inevitably does not cure the diseased health care delivery system.
Making GME Better
In some respects, it is not surprising that earlier calls for the reform of GME have had little effect. A report has moral authority only. It has no real power of its own. The committee typically disbands after the report is published, and no one—neither committee members nor sponsoring organizations—has control of the potent levers of accreditation or financing, which conceivably could be used to promote change in a particular direction. In the history of medical education, the only report that had a transforming effect was the Flexner report, and that report had the great advantage of appearing at a time when public sentiment demanded that medical education be reformed.19
These observations do not decry the potential accomplishments that can follow from a thoughtful, cogent report. An effective report can identify key issues and problems, make forceful recommendations, and serve as a strong moral compass. In the case of GME, it might be easier for a report to have a positive effect, as this article has suggested, if it fully recognizes the complexity of GME, adopts a comprehensive approach in its analysis and recommendations, and acknowledges what GME is and is not capable of doing in terms of serving the broader needs of the health care system.
At the present moment, undoubtedly the most significant requirement for GME to thrive is adequate financial support. Little was said of this subject in earlier calls for reform, but adequate funding is clearly the underpinning of a successful system of GME. Large amounts of money are required not only to provide house officers with salaries and benefits but also to develop every aspect of the learning environment that is necessary for the production of competent, caring, and socially responsible doctors. The financial needs of GME include money for teachers, support staff (to lessen the burden of nonprofessional chores), other medical professionals (to care for some of the patients currently managed by residents), curricula development, and new educational technologies.
Ultimately, the quality of GME will depend on the quality of health care delivery in the United States. The external forces are more powerful than the internal. The GME enterprise depends on society not only for financial but also for moral support. Residents learn their fields in the real world where patient care is actually delivered. If the health care environment continues to worship volume instead of quality of care, the ultimate products of GME are likely to be disappointing. Conversely, if the future environment of patient care recaptures a more genuine concern for caring and service, the products of GME would be much more likely to emerge as we might hope. Thier got it right in the Macy report two decades ago when he stated that, in the final analysis, the fate of GME depends on the fact of health care delivery.2
The author received an honorarium for writing this article from the Josiah Macy Jr. Foundation, which commissioned it. He also received a travel allowance to attend and present an earlier version of this article at the Josiah Macy Jr. Foundation conference entitled “Reforming Graduate Medical Education to Meet the Needs of the Public” held in Atlanta, Georgia in May 2011.
This article was commissioned for and presented at the Josiah Macy Jr. Foundation conference entitled “Reforming Graduate Medical Education to Meet the Needs of the Public” held in Atlanta, Georgia in May 2011.