Practice made perfect and the face of rheumatology training in 2010


In 2005, Ward published evidence that patients hospitalized with a diagnosis of systemic lupus erythematosus had a better outcome if they were cared for by physicians with more experience in caring for patients with lupus (1). Katz, in an accompanying editorial, commented “It makes good sense to urge our patients to seek care from providers who are experienced” (2). This comment has face validity, and few clinicians would likely quibble with this assertion. Accepting the premise begs the questions of how much experience is needed to make a difference that can be measured in a meaningful way, what types of experience matter most, and at what point does further experience begin to matter less? As we reflect on our past clinical (or teaching) experiences, it is not hard to recall the specific face of a patient whose course was impacted by the availability or unavailability of an experienced clinician. This is true face validity. But are we adequately taking this into consideration as we design and implement our fellowship training programs?

Too few trainees are entering careers in academic rheumatology. The duration of training may dissuade young physicians, many with significant debt, from pursuing a research/academic career track. Although I am not convinced that the duration of clinical training is the major factor for the shrinking academic work force, training time is certainly linked to trainee debt. The American Board of Internal Medicine (ABIM) research pathway reduces clinical training and gets fellows into research earlier. Time will tell whether these efforts to provide research-focused trainees with more immediate gratification will keep them in research careers. How many graduates of the research pathway ultimately become clinical educators or clinicians, and do we know the impact of truncating their clinical experience? And what about fellows in our regular training track?

Rheumatology is a specialty that crosses organ systems. We rely on no specific procedure to ply our trade or drive the design of our training programs. Rheumatologists are clinicians. We rely on our cognitive skills and accumulated clinical experience. Some rheumatology fellows enter training programs with the goal of becoming clinical or basic scientists. A few will succeed. But despite what we would like to believe, only some of these bright, published, and funded scientists are intrinsically skilled clinicians capable of providing superb care and emotional support to patients with complex medical problems, or teaching our trainees to care for such patients. I believe that even the brightest trainee benefits from having more than the perceived minimum clinical experience to effectively practice, or teach a clinical specialty such as rheumatology. Thus, I have concerns in steering training programs too much toward producing physician scientists, while not guaranteeing measurable clinical expertise in all of our graduates.

One year of clinical training with significant inpatient consultation time and then one or two half-day sessions of outpatient clinic for the second (and occasionally third) year have been said by some rheumatology program directors to provide trainees with sufficient clinical training experience. I believe, this does a disservice to our clinical specialty. In 25 years of practicing and teaching rheumatology, I have seen very few fellows whose learning curve was not steep at the end of one year of clinical training and still rising at the end of two years –even at a time when the program I directed demanded a much higher quantity of structured clinical experience. In programs with reduced clinical time for fellows, there are also fewer opportunities to rigorously assess the clinical skills of our trainees. Thus, smart or “nice” trainees may seem more skilled than they really are.

There are implications of increasing research time while limiting the clinical experiences of our fellows who aspire to academic careers. Let us assume that a high percentage of these trainees stay in research (a huge assumption). The clinical time of these faculty may be limited to only one half-day clinic per week and occasional inpatient coverage in order to permit 80–90% of their time to be dedicated to research pursuits. Because these faculty are likely to have smaller patient panels, and because supervising fellows in the clinic tends to have less paperwork associated with it (the fellows do it), these researchers will likely be given at least some of the responsibilities for supervising fellows. Thus, those with the least clinical experience are charged with teaching the less experienced. Many research-focused academic rheumatologists would likely express a high sense of confidence in their own clinical skills. Self assessment may be accurate for some of these faculty members, but as uncomfortable as it is to accept, it is not true for all of us. We do not self assess reliably (3)!

A key issue in the evaluation of our training programs is whether we have the ability to accurately assess the clinical competence of our trainees. Yet, we were not trained to do that. Board passage rates are high in rheumatology, but this reflects book knowledge, and the ABIM pass rate is not determined by attainment of any absolute level of clinical proficiency. All experienced program directors have seen trainees who score high on tests, but lack excellent clinical skills.

Since there is face validity and data supporting the concept that in clinical care “practice makes perfect,” shouldn't we be erring on the side of guaranteeing our trainees additional rather than minimal clinical training; particularly those who may be destined by choice or circumstance to assume teaching positions? I realize that no data demonstrate that limiting the clinical training time of our fellows will reduce their clinical acumen or ability to teach clinical rheumatology. But lack of data does not mean that the concern is not valid. I believe that most chairmen and senior clinicians have witnessed junior faculty who were hired for their research promise, but still are in need of significant clinical mentoring. As we develop structured ways to assess clinical ability, we may better understand the clinical learning curve of our fellows. Until then, we are doing a potential disservice to our trainees and patients by limiting the clinical experience of our fellows.

How should we recruit fellows into academic medicine when some worry that the duration of training is an obstacle and I am encouraging more clinical experience? A two-track (research or clinical) system of training is not the answer, because the issue of clinical experience still needs to be addressed if research track fellows will be providing any clinical care or teaching upon matriculation to faculty positions. Additionally, “crossover” from a research to a clinical career path will inevitably occur. Thus, with a two-track system, the question of whether research pathway graduates should be allowed to teach clinical rheumatology without first acquiring additional clinical experience would need to be seriously and objectively addressed. We should not permit investigative ability to be mistaken for clinical skill. It is all about assessment.

A solution may include revising how time is spent during residency training in internal medicine; increase the flexibility of curriculum content and deliver tailored clinical and research experiences to our residents based on their anticipated career pathway and their skills and competencies. For those with serious research interests, research didactics and mentoring should be spread throughout residency and fellowship. Duration of clinical and research training could be dictated by proficiency, not just time. Such curriculum change is under discussion with the ABIM. Subsequent completion of fellowship, like completion of a doctorate, should be based on the acquisition of the skills and experience necessary and appropriate to advance on to the next career step (4), not just time. Such change assumes that we can accurately assess competency and proficiency (and deal with the logistics of funding flexible duration residencies and fellowships).

But we need to maximize our fellows' clinical skills now. Rheumatology trainees should enter fellowship required to have mastered, through active participation in clinical rotations, skills in consultative neurology (not just participated on an inpatient stroke service), dermatology, and consultative nephrology. During fellowship, we should incorporate active experiences in relevant clinical areas outside of rheumatology. A core set of entrance and graduation expectations for future rheumatology fellows should be generated by rheumatology program directors. For those trainees anticipating a career in clinical research, core statistical skills and data management skills should be achieved as early as possible, ideally beginning in medical school, but not obtained at the expense of compromising clinical competence. For fellows desiring a clinical educator career, we should focus and limit their traditional research time, instead providing them tailored clinical experiences, extra electives in musculoskeletal imaging, as well as seminars in teaching and assessment methods as their scholarly activity. Proficiency in teaching as well as clinical rheumatology should be an expectation for graduation.

Academic medical centers should take the lead in using technology as a teaching tool. It should be an expectation that trainees graduate with the ability to utilize the electronic medical record to self analyze their practice. They should graduate having appropriate templates to assist them in maintaining appropriate specialty billing practices and patient information that can be imported into databases or directly queried for clinical research or quality assessment. It is not appropriate in 2010 to send young graduates out into practice without these tools. Clinician graduates of top tier programs should be natural leaders in the implementation of the electronic medical record into clinical practice. It is our obligation to provide these skills to our future educators and clinicians.

We should utilize technology in teaching physical examination and injection techniques. Simulation centers should be available, and ultrasonography should be utilized in helping trainees verify their (and our) physical examination skills (independent of the controversy over the role for ultrasound in routine practice). The data that exist on the poor accuracy of “blind” musculoskeletal injections unarguably indicate the need for improved teaching of procedural skills (5). Programs should be able to provide objective feedback to trainees on their procedural and examination skills, beyond traditional staff observation. Despite the temptation of expediency, computer interactive programs should not be relied upon to take the place of clinical experience, in any but the rarest of clinical conditions.

Until we develop better assessment measures to document clinical proficiency, we should err on the side of clinically overtraining our successors, and most certainly our future educators. We should not continue to assume that fellows with research interests and acumen are equally skilled clinicians. We should not use regulatory bodies' defined minimal required clinical time as the default schedule for our fellows. As program directors, we should examine the face of our programs and our trainees in the mirror of validity, and remember that our first charge as physician (educators) is to do no harm to our patients or disservice to our trainees.


Dr. Mandell drafted and revised the article for important intellectual content and approved the final version to be published.