In this study, systematic analysis of missed and discrepant diagnoses, discovered through departmental quality assurance conferences, identified several common diagnostic errors in interpretation of computed tomographic (CT) studies. False-negative diagnoses were the most common type of error, but misdiagnosis relating to poor communication between departments also occurred in a significant proportion of cases. This study provides an example of how traditional morbidity and mortality teaching conferences may be used as a vehicle for improving patient safety.