The Professionalism Movement: Pausing and Reflecting Are Essential

In discussing the increasing emphasis on professionalism in medical education, Delese Wear and Mark G. Kuczewski (2004) correctly note such discussions are permeated by "too many abstractions." Concepts such as altruism, respect, integrity, honesty, honor, duty, accountability, and responsiveness abound and focus not only on abstractions but on ideals of professional behavior. Although it is difficult to argue with the value of these principles, mandating such ideals has many implications that require additional thought before we forge ahead.

Assessment of professional behaviors is a key element of professionalism initiatives. Much scholarship and research has already addressed aspects of evaluation (Arnold 2002). Nonetheless, in the absence of well-delineated, reliable, and valid evaluative measures, current mandates to assess professionalism competency suggest that we are putting the cart before the horse. As Wear and Kuczewski point out, "the current emphasis on measurable evaluation of professionalism easily leads to attempts to test for the untestable."

One factor making development of solid assessment tools difficult is the complex nature of professional behaviors. Such behaviors might seem straightforward to evaluate because they are often framed in dichotomous terms—for example, altruistic vs. selfish, respectful vs. disrespectful, or professional vs. unprofessional as discussed by Ginsburg and colleagues (2000). However, when we (as educators or as people) place someone in one category or its opposite, we are making a multifactorial determination. We are not simply rating a characteristic on a Likert scale but rather are simultaneously considering multiple dimensions, including the frequency of particular behaviors and their magnitude (on a continuum of goodness to egregiousness). Importantly, each of these judgments is modulated by the context in which the event occurs (Ginsburg et al. 2000) as well as by even less tangible effects (e.g., transference/countertransference, halo effects).

We also need to appreciate the risks of overgeneralizing from data that mix rare or occasional events with frequent ones and that mix serious transgressions with minor ones (or even with perceived "transgressions" that ordinarily would not be labeled as such). Such ambiguity in definitions has been observed in discussions of medical student abuse, in which "some instances of perceived 'abuse' would be considered a legitimate part of the educational process" (Louria 2004). However, it is equally problematic to measure professionalism deficits by rating behaviorally-defined statements such as "I observed my faculty making derogatory comments about other services." What frequency and magnitude of event is needed to categorize a person as unprofessional? By the same token, what frequency and magnitude of event is needed to conclude that academic faculty are poor role models of professional behavior or that unprofessionalism has risen to crisis levels? Would we use a different measurement and interpretive strategy if measuring an event with less emotional valence than professionalism? For example, everyone would agree that anaphylactic reactions are bad and that every effort should be made to minimize their occurrence. Although we recognize that total eradication of anaphylaxis is unlikely, we can still strive to achieve it without declaring our efforts as failures if we do not reach perfection. From a measurement standpoint, few would suggest that we monitor our progress by asking physicians to rate a statement such as "I have seen a patient experience an anaphylactic reaction." Although anaphylaxis is, by definition, a severe event, we would be unable to rate event frequency with such a survey method. (In this scenario my answer to the question would be given equal weight to the answer of an emergency physician even though I have seen two patients with anaphylaxis in my career.) Even if answer options included a gradation of choices (e.g., never, rarely, occasionally, frequently), we would still be unable to calculate actual event incidence or prevalence without knowing the absolute numbers of anaphylactic reactions as well as the number of individuals at risk. Given the huge number of communicative exchanges occurring on a daily basis in academic medical centers...