Among health care leaders and policymakers, there is growing support to improve the value of US health care by providing doctors with feedback about their performance relative to that of their peers (i.e., comparing doctors to each other).
The logic for this strategy, called "social comparisons", is straightforward: if a doctor sees that he is underperforming in some area of patient care compared to his colleagues, a confluence of motivations -- his desire to take good care of patients, his innate desire to achieve, and his desire to be viewed favorably by peers -- will compel him to change his behavior and improve. Because high ranking doctors are also motivated to maintain their level of performance, everyone seems to win, including patients.
However, as my colleagues and I discuss in an article published this month in the Journal of the American Medical Association, the reality of social comparisons is more complex. While there is some supporting evidence for comparisons (for example, in reducing inappropriate antibiotic prescriptions), they can also be problematic. Several decades ago, for example, on the premise that openly publishing and allowing doctors to see each others' performance data would increase quality, New York began publicly reporting death rates for cardiac surgeons across the state. While outcomes improved, doctors seemed to avoid operating on certain patients out of fear that they would do poorly in surgery and harm the surgeons' rankings. Analyses subsequently showed that black and Hispanic patients were preferentially turned away, increasing racial disparities. These examples, among others, suggest that social comparisons can be a "two-tailed" behavioral problem -- capable of both improving or unintentionally harming care.
How can we tap into behavioral benefits while avoiding the unintended harms of social comparisons? One step is to clearly recognize that while comparisons reflect what is, they do not by themselves represent what ought to be. For example, assume that a surgeon is ranked lower than colleagues with respect to death rates. Does that unfavorable comparison reflect her lower individual quality (bad), her commitment to equity and refusal to preferentially turn away patients (good), and/or something else completely (e.g., a cavalier attitude about operating on high-risk patients for whom surgery is too risky and shouldn't be done)? For that kind of insight, we need to look beyond comparisons (numbers of about what occurs) to value judgments (beliefs about what should occur) and detailed data about the individual beliefs and values that drive physician behavior.
Ultimately, the task of comparing doctors and positively influencing behavior is complex, and more work is needed to understand the impact of social comparisons in different situations. Their effects may also change over time (for example, when tracked for a longer period, the disparities observed in New York waned for unclear reasons).
However, because patients can be negatively affected when doctors are compared to each other, leaders and policymakers should guard against unintended consequences in all circumstances by contextualizing comparisons within a set of values that reflect appropriateness and patient well-being.
This piece reflects the opinions and perspectives of the author and does not necessarily reflect the views or perspectives of The University of Pennsylvania School of Medicine.