Use of the autopsy as an instrument of quality assessment of patient care requires the quantification of discrepancies between premortem and postmortem diagnoses and the determination of an acceptable range with respect to such differences. Factors believed to possibly influence the proportion of discrepant diagnoses were evaluated by way of a questionnaire mailed to 166 pathologists. Factors believed to be important include those related to the characteristics of the patient's illness, the patient's physician, and the setting in which the patient's care transpires . Responses from 127 persons, including 106 of those originally queried, suggested that assessment of patient care by way of the autopsy can best be accomplished on an institutional basis. Application of a categorization of discrepant diagnoses to several recent studies suggested that approximately 10% of hospital autopsies will yield missed major discrepant diagnoses associated with an adverse impact on survival.