We thank Dr. Blank for highlighting the tension in medical education between teaching simplified rules for patient care and teaching the underlying primary evidence. We agree that the latter approach should support more robust decision making, but it also generally requires more time. Our intervention was intended to facilitate learning both guideline recommendations and the evidence that underlies them, but, as we noted, few of the participants pursued the evidence in depth. The participants may have shared an attitude that was found recently among British general practitioners—they do not have time to pursue critical appraisal of the primary evidence (1). Given that all physicians face time constraints, it remains possible that generalists could best improve their practices by focusing on simple, clear messages about actions that have proven benefit. On the other hand, it is also possible that if generalists invested more time learning the primary evidence, they would remember best practices with less repetition, effectively making their initial learning more efficient. Because many physicians fail to learn and apply even simple recommendations supported by both strong evidence and expert opinion, research that addresses these questions is urgently needed.

We disagree that the mere existence of guidelines contributes substantially to physicians' lack of interest in learning evidence. Well-written guidelines are structured to communicate the evidence in support of their conclusions. Physicians who choose to focus only on the conclusions probably would not have sought out and synthesized the evidence in the absence of the guidelines. Furthermore, if guidelines did not exist, we would also face more difficulty in setting educational priorities and standards for quality improvement. However, we agree with Dr. Blank and also with a recent editorialist (2) that physicians may shun the evidence in part because of poor critical appraisal skills.

As medical educators, we should redouble our efforts at disseminating the most basic of these skills. We must also acknowledge, however, that all humans are limited in the amount they can learn in a given period. To improve the productivity of computer-based medical education, we believe that more research is needed to understand the cognitive mechanisms that lead physicians to learn and retain new material, and to improve their performance.