One hundred years after Abraham Flexner released his report Medical Education in the United States and Canada. A Report to the Carnegie Foundation for the Advancement of Teaching,1 the spirit of reform is alive once more. Reports in the United States and Canada have called for significant changes to medical education that will allow doctors to adapt to complex environments, work in teams, and meet a wide range of social needs. These reports call for clear educational outcomes but also for a flexible, individualized approach to learning. Whether or not these reports will result in change has much to do with the alignment between what is proposed and the nature of current societal discourses. (In the Appendix, I have listed some of the recent reports that I think are particularly important.)
Currently, two powerful models of competence development are operating at odds with one another. The traditional one is time-based and directs attention to processes such as admissions and curriculum design. The newer one is outcomes-based and focuses more on the functional capabilities of the end product (the graduate student, resident or practicing physician). In this article, I explore the implications of both time and outcomes-based models for medical education reform and proposes an integration of the best features of each.
The Flexner Legacy
Abraham Flexner's magnum opus of 19101 is central to all discussions about medical education reform, both past and future, because of the significant role it played in the transformation of medical education. Whether praised, vilified, or something in between, the Flexner Report is often treated with the reverence of a holy book – a powerful narrative that caused vast change simply by appearing. At a time when the need for change in medical education is once again strongly felt, we look to Flexner to try to understand what magic could have unleashed such a powerful transformation. In both Canada and the United States, national projects are reexamining medical education “in the spirit of Flexner” and recommending significant reform.
While there is no question that Flexner carried out a project of unprecedented scope and wrote a compelling report that led to large-scale reform in the early twentieth century, a careful read of his report reveals that only some of his recommendations were actually adopted. Indeed, some of the issues that perplex us today were already visible to him in 1910. Yet in 2010, the Flexner legacy is often reduced to a few core elements: the move of medical schools to universities and the focus on science and the binary structure of preclinical basic science study followed by clinical training. However, in his report he discussed many issues related to societal needs and expectations of the medical profession. Whitehead,2 for example, uncovered Flexner's 1910 argument that “the physician's function is fast becoming social and preventative, rather than individual and curative”1 and points out that by 1925 even Flexner himself was unhappy with the way an overemphasis on the “positivist and scientific aspects” of medicine had become completely dominant, rendering doctors “sadly deficient in cultural and philosophic background.”3 While Flexner's diagnosis was a lack of scientific rigor, and his prescription was a strong grounding in science, his vision for physician education included many other dimensions.
One way of understanding why some elements of Flexner's report were enthusiastically taken up while others were neglected is the degree to which they aligned with discourses that were dominant at the time. For example, it has been argued that scientific medical discourse was already prominent by the time Flexner articulated the need to align medical education with science.4 Thus his recommendations in this regard were adopted with relative ease. Similarly, a look at historical discourses helps us understand some of the negative consequences of his report. One such outcome was the closing of medical schools specializing in the education of African American or women doctors.5 Though shocking in the 21st century, feminist historians have documented the arguments of 1910 that women were intellectually inferior, inadequately adapted to the study of science, and unable to cope with the demands of medical education and practice.5 A discourse that supported more equitable admission to medical schools did not fully emerge until the 1970s.6 So when a New York Times 1910 headline screamed “Carnegie Foundation's Startling Report [shows] That Incompetent Physicians are Manufactured Wholesale in This Country,” it is not surprising (though no less tragic) that closing inferior medical schools meant, at least in part, closing those educating African Americans or women. Finally, the fact that inferior schools were be closed, rather than helped to raise their standards, is also reflective of an early 20th-century discourse about institutional reform and public regulation that was more about closure than improvement. One hundred years later, we use new discourses about equity and continuous quality improvement/accreditation that would make responses to a Flexner-like report much different.
Turning to the present, we can apply the same type of analysis to currently proposed medical education reforms and think about the degree to which they align with current discourses. While it is difficult to imagine all eventualities, this approach can help predict which elements may be easily taken up and which will be resisted, rejected, or even distorted for other than intended purposes. In this spirit I review recent recommendations for reform of medical education in Canada and the United States. Then I discuss two currently dominant models of competence development: one that is time-based and one that is outcomes-based, exploring the implications of each for reform. Finally, I argue that neither a pure time-based, nor a pure outcomes-based model is fully adapted to the vision of reform outlined in current reports and that we must strive for a new approach that preserves the best of each.
Models of Competence Development
Simply broaching the subject of medical competence is a challenge because the term competence has become what Lingard has called, after Burke, a god term: “a rhetorical trump card, regularly played as the last word in debates about how health professions education should function.”7 The term is so widely employed, with so many different meanings, that it risks meaning nothing at all. Elsewhere I have characterized how different discourses of competence make possible specific language, practices, roles, and institutions. These discourses include competence as knowledge, as performance, as reflection, as a psychometrically reliable test score, and as a product.8,9
Various discourses about competence also lead to different models of competence development. In this article I explore two models of competence development that are dominant today: time-based and outcomes-based. The traditional time-based model, originally tied to a discourse of competence as knowledge, evolved to competence as performance in the 1970s and more recently incorporated the idea of competence as reflection. The outcomes-based model, by contrast, has roots in psychometric discourse and has recently incorporated a production discourse that emphasizes notions of efficiency and standardization. Taking up a time-based or an outcomes-based model has different implications for individuals (students, teachers, assessors, and others) and institutions (medical schools, testing organizations, professional associations, and others). Each is associated with different statements about how health professional competence develops, statements that are sometimes in conflict. A struggle of this kind is currently playing out in medical education around the notion that the time-based model of competence development is no longer compatible with demands for professional accountability. As I shall describe, though the time-based model remains quite resilient, we may be at the threshold of a wide-spread adoption of outcomes-based models. Understanding this changing landscape is essential for medical education reform.
What Problems are We Trying to Solve Today?
In both Canada and the United States, the 100-year anniversary of the Flexner Report provided an opportunity to once again study the strengths and weaknesses of medical education and to provide a template for reform. In the United States, the Carnegie Foundation and a team led by Cooke, Irby, and O'Brien recently released Educating Physicians: A Call for Reform of Medical School and Residency.6 The report contains an in-depth analysis of the current state of medical education based on literature reviews and extensive consultation, including visits to American medical schools. The Future of Medical Education in Canada Project has recently completed phase one (undergraduate education), including literature reviews, key informant interviews, and focus groups held across the country. A postgraduate phase is under way and will be followed by a third phase on continuing education.10 See List 1 for a summary of the recommendations of these two reports.
Like Flexner, the authors of both the U.S. and Canadian reports found many gaps and weaknesses. The U.S. report describes four key areas for reform: standardizing outcomes and individualizing training, integrating knowledge and clinical skills, developing “habits of mind” that include inquiry and innovation, and putting an emphasis on professional identity formation. The Canadian report also calls for substantial changes in medical education, including revisiting the social responsibility of medical schools, adapting admission processes, integrating basic and clinical sciences, addressing the hidden curriculum, fostering inter- and intraprofessional collaboration, encouraging generalism and community focus, and moving toward outcomes-based education.
Both reports call for new and innovative pedagogy, learning in context, mentoring/coaching models that involve extensive feedback, personal reflection, and a continuum of learning based on a flexible and developmental approach. Flexner's diagnosis in 1910 was a lack of scientific rigor, and his prescription was to strengthen the scientific basis of medical education. These new reports appear to focus more on physicians' lack of adaptability, flexibility, and alignment with social need. Their prescriptions rest more on changing educational contexts, the nature of teaching, and the processes of professional socialization than on modifying any particular content area.
While neither report dealt with continuing education in depth, educators working in that domain are also elaborating priorities for a future in which there will be a much greater role for lifelong learning that responds to social need.11 Continuing educators are also struggling with a tension between a vision of competence that is self-directed and based on internal self-assessment/self-regulation and a model of external assessment (including recertification) that is based on third-party assessment of knowledge and skills.12 While self-assessment alone may not be a useful construct for the assessment of one's competence over long periods of time, external surveillance/examination is neither a practical nor particularly effective alternative, as it fails to develop an internally-guided concept of good performance. Thus a concept of guided self-assessment is emerging in continuing education as an attempt to reconcile the distorting effects of too much external testing with an honest appraisal of the limits and fallibility of self-assessment.13 These discussions seem to point, as do those at the undergraduate and postgraduate level, to the need for a model of continuous formative assessment and feedback in practice.
What strikes me in looking across these reports in undergraduate, postgraduate, and continuing education in both Canada and the United States is the prominence of a dual imperative for defined outcome standards together with a call for pedagogical models that are individualized and provide continuous learning, feedback, and assessment. It seems that medical educators feel unsure whether current medical students, residents, and practitioners are competent (because of insufficient outcomes measures) but also feel the need to develop education programs that are tailored to individuals, rich in feedback, flexible in time, and targeted to deeper levels of cognitive and personal development than are current approaches. Unlike the reforms that Flexner envisioned to bolster the scientific basis of medicine, the reforms proposed today are about context, culture, and professional socialization.
In the U.S. report's recommendations we find “standardization of learning outcomes” and “individualization of the learning process” bundled together. As I will argue below, it seems almost certain that the first—standardization of outcomes—will be adopted, because it is aligned with both the outcomes model and a dominant production discourse. But the second—individualization of the learning process—is something else entirely. Individualization of learning will be a radical departure from the one-size-fits-all practices that are ubiquitous today and will require a substantial investment of time and resources. As the time-based model weakens and an outcomes-based model based on efficiency finds favor, will individualized learning and the time needed to develop the complex competencies envisioned in the reports be part and parcel of the adoption of standardized outcomes?
I will argue that much of the current motivation for outcomes-based education is about increasing efficiency, shortening training time, and reducing the overall cost of medical education. Thus there is a possibility that standardized outcomes will be adopted but that the additional investment necessary to also develop individualized training will prove more difficult. Just as particular elements of the Flexner Report were adopted but not others, in the next few years we will see the uptake of only some of the recommendations of the U.S. and Canadian reports. This phenomenon is already visible in the Canadian project. Whereas the empirical data collection of the environmental scan identified the first set of key priorities for reform, a further year of consultation with medical education leaders and organizations led to revisions of the final version, shifting, for example, “social responsibility and accountability” to “address individual and community needs” and modifying “integration of the timing of basic science and clinical education” to “build on the scientific basis of medicine.”14
What should be the role of medical educators in the uptake and implementation of these new reports on medical education reform? Unlike 1910, in 2010 Canada and the United States have a large and engaged community of medical educators. It is important for this group to think proactively about where reform might be uneven and to develop strategies to prevent key elements from being left out. But before considering action, I will characterize the two competing models of competence development.
Two Competing Models of Competence Development
A time-based model of competence development: Tea-steeping
By the time Flexner argued that medical education should have a strong basis in science, the movement toward aligning the guild with science, rather than artistry and tradition alone, was already well underway.15 Bringing an unruly guild into elite universities was certainly not without its difficulties.16 But once there, medicine joined an institution with a long tradition of conceptualizing competence as the accumulation of factual knowledge and of examining students' ability to reproduce information.
When it found a home in universities, undergraduate medical education came to consist of a fixed period of training in foundational and basic sciences (something that fit well with a traditional university model) followed by a fixed period of clinical experience (a slightly more awkward academic arrangement that required partnerships with health care institutions and other nonacademic settings). The duration of these two sequential training periods has varied slightly over time and place: 3 + 1 until the mid 20th century in North America, moving to 2 + 2 in most schools; roughly 4 + 2 in Europe with variations by country. Despite these changes, the essential binary nature of undergraduate medical curricula has remained largely the same for a century. The persistence of a sharp division between the two portions of training reflects implicit but equally sharp distinctions concerning the value of scientific knowledge and its practical application. Postgraduate education similarly is organized around a fixed number of years of training (two to six years depending on specialty) but has remained largely at arms-length from universities by creating a series of clinical rotations in affiliated health care institutions. While much of postgraduate medical education takes place in cooperation with universities, a degree is rarely granted. Thus, unlike undergraduate education, postgraduate educators rarely face the issue of long blocks of basic science teaching, but rather the opposite: a gulf between the practice of medicine and the ability to access and keep current with the latest scientific research.
The organization of education at both undergraduate and postgraduate levels today reveals a continuing belief that a fixed interval of time is required, and indeed sufficient, to develop global competence. To use a metaphor, we put the student (tea) in medical school (hot water) for a fixed period of time and, voilà! After a historically determined interval of time, we assume a competent practitioner, like a good cup of tea, will result. This arrangement becomes particularly tricky when a specialty approaches its governing college to ask to lengthen training time. The adjudication of these requests typically turns to political maneuvering rather than reasoned debate because there is a lack of evidence demonstrating deficits in competence or linking training duration to patient outcomes. Rather, vague arguments about increasing complexity, the explosion of knowledge, or the requirement for research become the basis for lengthening training. By contrast, students and governments regularly call for a reexamination of length of training, asking why it takes so long to complete medical studies and whether they could not be accomplished in a shorter period of time. These arguments also founder on a lack of evidence regarding links between duration of training and competence.
The tea-steeping model has proved enormously durable and, with the exception of gradually growing program length, time-based models of medical education have seen very few modifications in 100 years. There have been minor structural modifications, such as attempts at early clinical exposure, during which medical students in the preclinical years participate in half or full day experiences in clinics. Further, modified and integrated schedules at problem-based learning schools (e.g., McMaster, Calgary, New Mexico) shifted the curriculum from four to three years by integrating basic science and clinical teaching using problem-based learning approaches. And there have been attempts to integrate basic sciences into clinical clerkship and residency rotations. However, the primary determination of graduation (almost 100% of the time) remains the length of time spent in the training program. Certainly there are many assessments of competence but these, by and large, do not perturb the smooth trajectory from admission to graduation.
Over the past few decades, changes in time-based models have focused on two elements: admission criteria and curriculum content. Both of these approaches, to extend the metaphor, relate to changing the way the tea is made. The first – admission – is about changing the tea leaves. Debates persist over whether students should have, for example, more or less prerequisite science, more or less life experience, and more or less graduate training. An extreme example of this debate goes on in France, a country with a very rigorous assessment process for admission to medical school. Only 17% of those who complete the first premedical year continue studies, and there is no exit assessment, certification, or licensure examination. Thus a great deal of attention is focused on the criteria for admission and the right qualities of a premedical student.17
The other popular thrust – changing curricular content – is about adjusting the nature and temperature of the water, the environment into which the tea leaves are immersed. An old adage has it that tea leaves do not reveal their true flavor until put into hot water. So, for many, it is the environment of medical school that is determinant. All students are bright and motivated, they remark, so the important focus should be the content and process of what happens within the walls of medicals schools. This includes attention to the hidden curriculum18 and the various socialization processes that take place.
Assessment of competence has always been a problem in the time-based models. Intraining assessment is riddled with subjectivity, validity problems, and low numbers of observations, meaning that hard decisions are often deferred to end-of-rotation, end-of- year, or even end-of-program high-stakes examinations. The problem with this approach is the near complete lack of meaningful feedback from high-stakes exams. Discontinuation of performance-based certification examinations by many U.S. specialty boards, with a directive to more rigorously assess skills during residency, may trigger a move toward better in-training assessment. But assessment experts around the world are calling for a limit to high-stakes final examinations at all levels and a greater emphasis on continuous assessment of skills in the clinical workplace.19 A very significant shift toward workplace-based assessment is taking place in the United Kingdom, for example.20 In the meantime, assessment in time-based models still leans rather heavily on subjective supervisor evaluations because objectively evaluating complex clinical acts (e.g., diagnosis), systems roles (e.g., advocacy), and ambiguous concepts (e.g., professionalism) has proven difficult.
How well does the time-based model fit with current recommendations for medical education reform? At least conceptually, many of the qualities thought to be underdeveloped in graduates – tolerance of ambiguity, cognitive flexibility, the integration of knowledge and context, curiosity and innovation – fit with a tea-steeping model that gives primacy to time in the development of personal qualities and habits of mind. However, while most educators would admit that students acquire these abilities at different rates, and that the length of training programs at best represents some sort of average period of time during which most students will probably acquire them, without defined and assessable outcomes, we cannot really know if they do acquire them.
There are many reasons that undergraduate and postgraduate programs have clung to time-based models. In the undergraduate domain, one of the main reasons is the traditional emphasis on basic sciences. Conceived in Flexner's time as a way of building a foundation of knowledge for an emerging scientific profession, a heavy frontloading of chemistry, physics, and mathematics (preadmission in North America, first cycles in Europe) is followed by anatomy, physiology, biochemistry, and other topics in the preclinical years. Because it is very difficult conceptually, scientifically, and organizationally to define clearly what these knowledge-based sciences specifically contribute to the development of the skills and qualities needed for competent practice, the lengths of their curriculum times are often based on tradition.
While there is a literature that strongly supports the importance of understanding scientific mechanisms in clinical decision making, in problem solving, and in the application of skills, it is the failure to align the timing of learning science and its application that is troubling. The result is that most medical students have knowledge- retention problems and later are unable to activate their knowledge in clinical settings. Research is clear that the teaching of basic sciences should occur very close in time to application at the bedside.21 A few programs have experimented with nontraditional timing of basic science teaching, such as during or after the clerkship, integrated into residency programs, and even refreshers for practicing clinicians. However these approaches are not wide-spread and so the basic-science/clinical divide persists.
The question of how knowledge is acquired, incorporated into practice, and engaged over the years of independent practice is becoming particularly urgent as the knowledge base itself continues to shift from a relatively circumscribed canon of textbooks to the easily available but unwieldy Internet. To make matters even more complicated, there is wide-spread concern that what we call the basic sciences are subjects that were identified 100 years ago (Flexner was particularly keen on mathematics, for example), almost 50 years before the evolution of the disciplines of psychology, sociology, anthropology, economics, and others that may have as much, or more, to contribute to the competent performance of a physician or a team than do some of the “harder” basic sciences.
Why is change so difficult? One explanation is that in the rather thin soup of evidence for long years of sequestered basic biological science teaching on one hand, and clinical rotations devoid of recourse to the underlying basic and social sciences on the other, there is a thick admixture of rhetorical and political positioning that relates to tradition, job security for various professionals and departments, and the struggles for legitimacy of various biological, social science, and clinical domains.22
Therefore, if the time-based model remains dominant, implementation of reforms will hinge on ways that the current curricula can be adapted to accommodate them. The problem is that most past reforms have been about adding to already overly full programs. Deans everywhere complain about the constant pressure to add to the curriculum – ethics, communications, scientific method, genetics, population health, social science – but never about a proposal to remove anything, since such proposals seldom occur. So without a fundamental change to the architecture and fixed time requirements of current curricula, change will likely be incremental and modest. Changing learning contexts, tightening performance-assessment-feedback loops, and continuing to press for integration rather than fragmentation of teaching and topics are important. However, these approaches alone will not be sufficient to satisfy the push for public accountability and documentation of actual capabilities, and they are probably too rigid to accommodate the pedagogical models envisioned by reformers. Traditionally, curriculum committees in every medical school and residency program have had one inviolable constraint: they cannot tamper with the number of years of training. Frustration with these constraints will continue to enhance the appeal of an outcomes-based model.
An outcomes-based model of competence development
As we have seen, the current medical education system is primarily time-based. However, for over 40 years the pieces have been slowly falling in place for the emergence of an outcomes-based model. Whitehead has remarked that “outcomes-based education hypothesizes that if the desired product can be defined, and appropriate assessment tools developed to ensure that trainees have achieved these competencies, then the job will be done.”2 She also notes that this discourse of competence development is “spreading like wildfire.” I will briefly outline some of the reasons why this is the case.
In Flexner's time, competence was understood almost entirely as the accumulation of knowledge, though graduating also meant showing oneself to be suitable to practice medicine. Thus most teaching and all assessment targeted what students knew with a little dab of who they were. The dominant discourse was one of competence as knowledge. Beginning in the mid-1960s, conceptions of competence were shifted by a number of factors: a greater role for patients' perspectives in their care, the dawn of the era of accountability, the emergence of collaboration with other health professionals, and the application of scientific evidence to education itself. In the latter half of the 20th century competence gradually came to include communication, collaboration, and a greater focus on skills. The era was marked by the appearance of Miller's Pyramid,23 in which performance was considered to be of a higher order than pure knowledge. Toward the end of the 20th century, in the United States and Canada, new frameworks for competence were developed that took the behaviourist approach further by embedding knowledge and skills into roles. The AAMC Medical Schools Objectives Project (MSOP)24 appeared in the United States, while in Canada, the Educating Future Physicians for Ontario (EFPO) was born.25 At the postgraduate level in the United States, the ACGME Framework26 was developed while in Canada, the RCPSC CanMEDS roles27 appeared. All of these frameworks (summarized in List 2) contributed to a shift in competence from knows to does and collectively fueled a drive toward outcomes.
Competence assessment changed in response to the new definitions of competence but also as a result of the arrival of psychometrically trained psychologists into the field of medical education.28 Psychometricians brought with them new tools and approaches to assessment that emphasized reliability, standardization, and multiple sampling that could convert performance into grades. Psychometric models have played a major role in setting the stage for outcome measurement and are also aligned with discourses of accountability.
Case Studies of Outcomes-Based Education
To understand some of the difference between time-based and outcomes-based models it is useful to consider some case examples. The three cases discussed below illustrate some common features of outcomes-based education. These include curricula closely related to social need, modularized training of specific skills, frequent assessment, stepwise progression, and variable length of training depending on entry skills and rate of skill attainment. In addition to these elements of the outcomes-based model, I have also tried to include some examples of language used to justify these programs. This second aspect – the discursive elements related to the uptake of outcomes-based models – is something I will discuss in more detail below.
Case 1: Curriculum adapted to social need: nurse practitioner training in Ontario, Canada
If medicine is very slowly embracing the ideas of outcomes-based education adapted to social need, other health professions are moving forward rapidly. One very good example is the training of nurse practitioners (NPs). Many jurisdictions are embracing the role of NPs to address a host of health care delivery problems. Among the professions, NPs have been particularly adept at defining gaps in health care delivery and effectively designing curricula to fill them. In a recent profile of NPs, the Toronto Star29 described the establishment of 11 of 25 planned NP-led clinics in the province of Ontario. The article used six arguments to illustrate the effectiveness of NPs and their approach to training:
- There is a shortage of doctors.
- Doctors might not provide care adapted to patient needs (in the article the patient's doctor had “fired” her because her mental health problems were too challenging).
- NPs are more cost-effective because they are not paid in a fee for service model.
- Patients like the care and the access (consumer satisfaction).
- Legislative changes have been made to permit NP practice.
- Evaluation has shown the efficacy and cost savings of NP-led clinics (helping patients avoid emergency department visits, for example).
If we look carefully at the rhetorical construction of the Toronto Star article we can identify many features of the outcomes-based model. The primacy of evaluation (consumer and economic), the responsiveness to systems needs, and the evolution of training programs geared to actual practice outcomes are all illustrated. Arguments about increased efficiency and cost reduction also stand out. In addition to health care cost savings, outcomes-based models are also thought to save education dollars. One means of doing this is by modularizing training, something described in the next example.
Case 2: Modularized endoscopic training programs in London, United Kingdom
Castle Hill Hospital in Hull, United Kingdom, and St Mary's Hospital in London, United Kingdom, were interested in the feasibility of training individuals who are not traditionally associated with the skill to perform flexible sigmoidoscopy through a New Ways of Working program.30 So, a three-year course was designed in a modular fashion to allow “step-off” and “step-on” points, thus providing an opportunity for trainees with various levels of experience (from lay people to advanced practice nurses and other health professionals, for example) to join the course anywhere along the continuum to gain specialized knowledge in gastroenterology and technical competency in performing flexible sigmoidoscopy.
Trainees carry out flexible sigmoidoscopy under direct supervision until they are judged to be competent and have been assessed by their clinical supervisor using national practice guidelines. A physician associated with the program notes that
the technician has the potential to fulfill roles traditionally associated with doctors and nurses. They may be able to be molded into practitioners with unique roles more suited to an NHS that may involve blurring of the professional boundaries as foreseen by the current Government [of the UK]. Participation in the colorectal cancer screening programme and its subsequent rollout and “fallout” will make endoscopy assistants indispensable.31
This skill set, like many others, is being modularized—a process that involves breaking down a skill into its essential components and teaching it to a level of proficiency. Naturally, questions will arise about who should or can be certified to carry out various acts (diagnosis, ordering tests, counseling, prescribing, physical manipulations, rehabilitation, and others). But the popularity of modularized education shifts discussions about who does what away from a focus on tradition and toward a focus on cost, efficiency, and training time. While we may be a long way from entirely modularized medical education, but as we shall see in the next example, there are steps being taken in this direction.
Case 3: Modularized, outcomes-based medical education: The University of Toronto Orthopedics Residency Program Experiment
There are very few current examples of modularized, outcomes-based medical education to study. One interesting exception is taking place in Toronto, where the Royal College of Physicians and Surgeons of Canada has approved an experimental residency model.32 A special orthopedics program was created in 2008, entirely based on the attainment of predetermined outcomes. There is no fixed time period for the residency program; students can theoretically graduate in three, five, or 15 years. There is a heavy assessment program, and residents are tested every six weeks in a host of domains, using many state- of-the-art assessment methods (simulation, objective structured assessment of technical skills, hand motion analysis, and others). The ultimate determination of competence (and, therefore, graduation) is a complex set of modularized practice outcomes. Although residents are still in the early stages (two cohorts have now entered the program), the results will be widely scrutinized and will almost certainly serve as a model for other similar experiments. Interestingly, one of the primary justifications for the establishment of the program was the possibility that it might provide a model for shortening the length of training.
What the three cases show
As we see in these examples, outcomes-based education is being adopted not only because of its pedagogical properties but also for its potential to improve efficiency and reduce the costs of health care and of training. These are important goals for today's financially strained health care and education systems. However, as we shall see in the next section, something a little less positive can arise when educational reform elevates efficiency and cost savings above other considerations. This may occurs if the outcomes-based model is linked to production discourse.
The Outcomes-Based Model Meets the Production Discourse: Manufacturing i-Docs
Individuals who speak and write about outcomes sometimes draw on language and concepts from manufacturing.9 For example, as a leader of a major testing organization told me in an interview,
Today, we have reached new heights in the science of producing and testing medical students. We run the operation using standardized and efficient process. As a result we are in a much better position to assure the public that we are graduating a product with the skills for safe and effective practice.9
To capture the essence of production discourse in relation to medical education I have coined the term i-Doc. The term is based on today's ubiquitous manufactured product, the Apple i-Pod. Though I draw the analogy somewhat facetiously, the notion that medical schools, like factories, can produce highly desirable products adapted to user needs and desires and can continuously improve successive iterations through quality assurance and feedback mechanisms is an attractive analogy for some. It is so attractive, in fact, that medical education literature after the 1990s is full of words and concepts taken from business and manufacturing.9
What does it mean to apply manufacturing models and a production discourse to medical education? Can we think of doctors as running applications with which they have been programmed during training (e.g., algorithms, practice guidelines), to solve particular problems? Can we imagine that the education (manufacturing) process is something adaptable to needs and desires of end-users (consumers)? Should we adopt the idea that quality assurance processes are there not only to ensure the quality of the product but also to render its production more efficient and less costly? In this description some may recognize concepts popularized by Frederic Taylor in his Principles of Scientific Management.33 His 1911 book was widely used to render factories and workplaces of all kinds more efficient. While medical education has been been slower to gravitate toward Taylorist ideas, at the dawn of the 21st century, production discourse is becoming very common.
An example of the application of manufacturing principles and production discourse to medical education is presented in an article that appeared in 2004, entitled “Medical Education as a Process Management Problem.”34 The authors cite many challenges facing medical education and ask the question: Can medical education be designed more efficiently and at less cost? They note that most curriculum reform efforts target fragments of the system and not the overall process, proposing that “the rules that govern the manufacturing industry provide a compelling system of guiding principles for medical education reform.” According to the authors, the genesis of problems facing modern medical education was the loss of the mentor-apprentice model, and the solution is adoption of a manufacturing model. First and foremost, they argue, it is necessary to define the skills and knowledge that a student should be able to competently and professionally demonstrate on completion of his or her education. Here the authors conceptually link outcomes with a production discourse.
A frequently used metaphor in articles such as theirs is the transformation of individuals. That is, medical students are the raw materials to be transformed through the manufacturing (education) process into a desired product. This is appealing for funders of health care education, who hope that production efficiencies can be found in the process of transformation to make it less expensive. In other words, by using manufacturing concepts to produce i-Docs, we might hope to get a better product at a lower cost.
The authors of “Medical Education as a Process Management Problem” not only recommended manufacturing principles but also used a specific case example to illustrate the point. They wrote in 2004, “Toyota is an example of a company that has exhibited leading rates of improvement and learning and superior performance.” They could not possibly have foreseen the quality troubles that their paragon of manufacturing logic, Toyota, would encounter in 2010. Nevertheless there is a certain eerie sense of circularity when we realize that the New York Times headline of 1910 declared pre-Flexner medical schools to be incompetent because they were “Factories for the making of ignorant doctors.” As the Toyota example so clearly illustrates, finding efficiencies and cutting costs in the context of a manufacturing model may be the source of more problems with quality. Might medical schools be faced with recalls if their products were defective?
What is the implication of the popularity of production discourse in terms of current recommendations for medical education reform? As we have seen, recommendations from the United States and Canada both explicitly call for clear outcomes standards. In this regard, there is good alignment with the production discourse, which also values outcomes. But what of the call for flexible, individually tailored programs that can adapt to variable rates of competence attainment? Here the manufacturing model falls short. Industrial products can be homogenized and manufacturing process standardized because the production of any one unit is the same as the production of all the others. But the raw material of medical education (students) is not very amenable to standardization. And if the manufacturing process is going to result in attainment of functional capabilities in the real world, human variability and the messiness of learning means that attainment of desired outcomes will be highly variable. While production discourse and lessons from manufacturing might help us think about some dimensions of education, it is not a very helpful metaphor for processes that must flexible and adapted to each student.
While the old time-based, tea-steeping model is problematic, jumping to a production-oriented, manufacturing of i-Docs is not the best solution. Rather, we must try to combine the best elements of time-based and outcomes-based models. Perhaps we could think of this as outcomes-based education adapted to the needs of individuals.
Outcomes-Based Education Adapted to Individuals: Swimming the Length of the Pool
In place of the manufacturing of i-Docs, let me offer a different way of thinking about outcomes-based models adapted to individuals. As a child I took the Red Cross swimming program. To pass Star One, I was required to swim – not describe how to swim, or simulate how to swim – but actually to swim from one end of the pool to the other and back again. In the Red Cross program, until you can swim the length of the pool, you stay in Star One. Not all children reach this outcome at the same rate. Some of my friends achieved it in days while others needed weeks. The skillful teachers were fully prepared for this variability and provided a curriculum rich in practice, feedback, and encouragement. They also likely recognized that for some children, it would just take time.
What would an individually adapted outcomes-based model look like in medical education? Let's consider a few examples in the undergraduate domain. For communication skills, if a desired outcome were taking a complete and accurate history from a wide range of patients with different problems, students would stay in the communication module until they could do so. For a diagnostic decision making outcome of regularly arriving at a correct diagnosis in a range of clinical situations, students would receive training and practice until they could. For teamwork, if the outcome were effectively participating in a health care team, practice with feedback would continue until students achieved the ability.
We can also imagine examples in postgraduate education. If a desired outcome for orthopedics residents were, for example, to successfully complete a total hip replacement surgery, they would stay in that module of the residency program until they could. If it were desirable for public health residents to be able to create a prevention program, they would remain in training until the skills to do so were achieved. If it was deemed important for internal medicine residents to be able to implement a quality assurance system in a ward or clinic, training, feedback, and assessment would be provided to students until they could successfully accomplish that outcome.
As these examples illustrate, the notion of outcomes tailored to individuals is not simply about the potential to do things but rather about actually doing them. This form of competence, what ten Cate has called entrustable professional acts,35 requires evidence of performance in real settings. While performance in a test such as an OSCE might indicate the potential for competent performance in real settings, it is but a proxy – a way-station on the road to the demonstration of competence. Using Miller's pyramid, outcomes-based competence is at the top. Outcomes-based competence is not so much what a student knows, nor simply what he or she knows how to do, and not even that he or she can show how – rather it is a question of what he or she does.
One particular challenge to documenting outcomes in authentic settings is that medical education remains wedded to the notion that outcomes are individual traits or abilities,7 while most health care outcomes are actually a result of teamwork. Reeves et al. have described the elements that characterize various kinds of effective teams, features that are quite different from individual competence.36 And as Lingard has written,
Our individualist health care system and education culture [focuses] attention on the individual learner and the knowledge, abilities and values they possess in their heads, hands and hearts” [but] “competent individual professionals can—and do, with some regularity—combine to create an incompetent team. The conventional [individualist] discourse of competence doesn't really help us grapple with this reality.7
The result is that we are currently witnessing a rather awkward attempt to graft new interprofessional outcomes onto a system that still assumes the primacy of individual competence. A complete paradigm shift to individually adapted outcomes-based education therefore must also await the adoption of team-based outcomes and methods of assessment.
In summary, there is a popular way of thinking about outcomes that is based on production discourse and Taylorist concepts of efficiency. Its power and appeal arise in part from the promise that manufacturing models will reduce educational cost and training time. By contrast, the outcomes envisioned in proposals for medical education reform appear to require flexible, individualized training. Further, special pedagogical approaches will be required to achieve such things as tolerance of ambiguity, agility in the face of complexity, and habits of mind such as curiosity, innovation, and a commitment to lifelong learning. These pedagogical approaches may be like coaching, with an emphasis on a close teacher-student relationship, a curriculum rich in practice and feedback, continuous formative assessment, and a stepwise, developmental approach. Such pedagogy would be an innovative departure from our current model of education, but it might also be time-consuming and expensive. Thus while the adoption of “standardized outcomes” is almost a certainty, the concurrent implementation of “flexible, individualize training” is not. If the latter element is an important dimension of meaningful outcomes-based education, medical educators will need to be clever in crafting arguments to support models that run counter to the imperatives of efficiency and cost savings.
Will Outcomes-Based Models be Widely Implemented?
Whitehead writes that “consideration must be given to which desired new approaches (for example outcomes-based education) are compatible with current paradigms, which would require fundamental change, and what the implications of such change would be.”2 Thus it is important to consider the potential implications of outcomes-based models for a wide variety of stakeholders.
For licensure and certification organizations, a move to outcomes-based models, and particularly ones tailored to individuals, would require them to become more involved in ensuring the attainment of competencies in training, in collaboration with medical schools and residencies; create more flexible final assessments that focus on skills in a wider variety of domains; and shift to modular assessment of individual competencies (such as certifying specific procedures).
For faculty teachers, an outcomes-based model that required coaching, feedback, and repeated practice would be labor-intensive. There could be no more “see one, do one, teach one.”37 Rather the phrase would have to be updated to something like “watch until you are ready to try, then practice in simulation until you are ready to perform with real patients, then perform repeatedly under supervision until you are ready to practice independently.” The pedagogical skills needed of faculty teachers would be greater.
Medical schools would have to grapple with ways to break down the basic-science/clinical divide and also to deal with variable times to graduation. The requirement for variable lengths of training has the potential to please and displease both students and funding sources (governments): pleasing if training gets shorter, displeasing if gets longer.
Individualized, variable length programs could be a nightmare for institutions that rely on a steady supply of medical students and residents to provide service delivery in predictable rotations. On the other hand, the mandate of health care institutions for quality assurance and patient safety might mean they would support the idea of better developed and documented competence among their students and practicing physicians. They might also like the idea that students who had not attained proficiency in particular domains would continue to get practice under supervision.
For international medical graduates, issues of compatibility/comparability to domestic graduates would become increasingly difficult if some countries adopted outcomes-based models and others did not. If the assessment of competence of graduates of domestic medical schools and residency programs became modularized, distributed over time, and organized in a series of developmental steps, it would be difficult to compare with the one-shot screening examinations widely used to assess international medical graduates.
This brief review of the many stakeholders who have strong reasons to be interested and concerned about a shift to outcomes-based models of education might explain what Whitcomb called the “apparent unwillingness of professional organizations that have control over some aspects of the ‘medical education system’” to support change.”38 Whitehead warned that “if we are not realistic about the scope of education reform, it will not be surprising if current changes once more disappoint, with renewed calls for medical education transformation likely appearing in yet a new form some time soon.”2 As with the Flexner reforms, changes made on the basis of recommendations in the United States in the recent Carnegie Foundation report6 and of recommendations from phase one of the Future of Medical Education in Canada Project10,39 are likely to be partial. Recommendations that are not in line with the dominant discourses of powerful stakeholders are much less likely to take hold. I predict that the development of outcomes standards will occur, as will a move toward modularizing many technical procedures. Shortening of training times may also come about. But whether individually tailored outcomes-based model of training and feedback-rich coaching models evolve is much less certain.
What should we medical educators do? There seems little doubt that clinging uncritically to a time-based approach is problematic. The other health professions will move forward and medicine would be foolish not explore outcomes-based models. As we have seen, it is the kind of outcomes-based model that matters. Elsewhere I have described the adverse effects of an overzealous adoption of production discourse in assessment.9 I believe we should be careful of it in curriculum reform as well.
Perhaps most importantly, we should not forget about time. Indeed, because of the dominance of time-based models, until recently we have rarely asked what the benefits of time actually are. In the passionate rush to adopt new models, we must carefully look at what might be lost in jettisoning the old ones. Whitehead has written that “a key shift in the language and approach of outcomes-based models of education is devaluing of time and process”2 It is only now, with the focus on replacing time with outcomes, that some educators are questioning which elements of competence indeed require time to attain. The development of cognitive structures, problem-solving routines, pattern recognition, judgment, and reflective capacity are just a few domains that warrant investigation to better understand if they can be modularized or if, instead, time (“steeping in hot water”) is a variable necessary for their development. Before we abandon time as the prime determinant of competence and capitulate completely to outcomes, it is important to be clearer about what we might be giving up. Time has a relationship to personal development, and as Whitehead has suggested, the production discourse “removes the person from the process. By combining the production discourse with outcomes-based language, both person and time, and hence any sense of journey, vanish.”2 In another hundred years, medical educators will look back at our discourses, some of which will seem as peculiar to them as Flexner-era discourses seem to us. I suspect that equating medical education to factory production will be one of these. My view is that, in fashioning discourses about the competence of future physicians, we should try to use language and concepts that emphasize the growth and development of people.
Discursive shifts change the ebb and flow of power, the roles that individuals can play, the centrality or marginalization of various institutions, and the ways in which it is credible and legitimate to speak, think, and act. Major discursive shifts do not flow from single individuals, organizations, or reports – they are more like earthquakes, resulting from tensions built up incrementally. Finally, a significant shift takes place, releasing the pressure, but also changing the landscape – the final form of which is hazy until the dust settles. In medical education the pressure for some form of outcomes-based education has been building for 40 years. The pure time-based approach is no longer enough. Our challenge is to make sure that when the dust settles, we have not lost all of the elements of time and context that mark the journey of becoming a physician.
The author is grateful to Ayelet Kuper, Nancy McNaughton, Tina Martimianakis, Robert Paul, Cynthia Whitehead, and Sarah Whyte for very helpful discussions and suggestions about this essay.
Other disclosures: None.
Ethical approval: Not applicable.