Making peer review statistically accountable

BACKGROUND: The peer review process used in most hospitals is largely anecdotal, leading to criticisms about the objectivity of the methods employed. METHODS: The results of 1,500 consecutive abdominal operations performed by general surgeons working at three hospitals in a single community were reviewed. The outcome profile of each surgeon was compared statistically to the cumulative profile of the surgical community with adjustments for physiologic status of the patient, difficulty of the operation, and indications for surgery. RESULTS: A problem surgeon was thus identified whose poor results were significantly different from the rest of the surgical community and could not be explained on the basis of unfavorable patient mix or complexity of the procedures undertaken. CONCLUSIONS: Statistical comparison of a surgeon's outcome profile with those of his colleagues working in the same practice environment is suggested as an approach to the task of peer review that might prove preferable to the usual retrospective review of problem cases.