PURPOSE: Although transparency is critical for reducing medical errors, physicians feel discomfort with disclosure. We explored whether overconfidence relates to physician's reluctance to admit that an error may have occurred. METHOD: At 3 university medical centers, a survey presented a clinical vignette of a girl with urinary infection and penicillin allergy to medical students and physicians, asking them to rate their level of confidence for each step of the diagnosis and management. After anaphylaxis develops after cephalosporin administration, respondents were asked about their willingness to admit that an error might have occurred and to rate their level of discomfort in doing so. We analyzed levels of confidence, accuracy, willingness to admit mistake, and discomfort. RESULTS: Respondents reported high levels of confidence for their answers to the questions of diagnosis and management, even when wrong-indicating miscalibration of confidence and accuracy. Compared with students, physicians had significantly higher levels of confidence, lower accuracy, and lower willingness to admit mistake. Although most respondents agreed in principle that errors should be disclosed, in the presented case, significantly less agreed to admit that a mistake might have occurred or to say so explicitly to the family. An association was found between overconfidence and discomfort with disclosure. CONCLUSIONS: Our study shows overconfidence associated with clinician's training and with reluctance to admit mistake, suggesting a contributing role to the difficulty in leveraging safety events into quality improvement. Training physicians to have both knowledge and adequate self-doubt is an educational challenge.
PURPOSE: Although transparency is critical for reducing medical errors, physicians feel discomfort with disclosure. We explored whether overconfidence relates to physician's reluctance to admit that an error may have occurred. METHOD: At 3 university medical centers, a survey presented a clinical vignette of a girl with urinary infection and penicillinallergy to medical students and physicians, asking them to rate their level of confidence for each step of the diagnosis and management. After anaphylaxis develops after cephalosporin administration, respondents were asked about their willingness to admit that an error might have occurred and to rate their level of discomfort in doing so. We analyzed levels of confidence, accuracy, willingness to admit mistake, and discomfort. RESULTS: Respondents reported high levels of confidence for their answers to the questions of diagnosis and management, even when wrong-indicating miscalibration of confidence and accuracy. Compared with students, physicians had significantly higher levels of confidence, lower accuracy, and lower willingness to admit mistake. Although most respondents agreed in principle that errors should be disclosed, in the presented case, significantly less agreed to admit that a mistake might have occurred or to say so explicitly to the family. An association was found between overconfidence and discomfort with disclosure. CONCLUSIONS: Our study shows overconfidence associated with clinician's training and with reluctance to admit mistake, suggesting a contributing role to the difficulty in leveraging safety events into quality improvement. Training physicians to have both knowledge and adequate self-doubt is an educational challenge.