Diagnostic, surgical judgment, and systems issues leading to reoperation: mining administrative databases

BackgroundUnderreporting of surgical adverse events limits the ability to identify quality and safety issues. Automated screening of the clinical information system (CIS) can improve case capture and reduce dependency on self-reporting. We compared screening of a CIS to self-reporting for identifying unplanned reoperation and also examined the relationship between causality and probability of reporting.MethodsBetween 2005 and 2009, all unplanned reoperations identified by automated screening of databases were reviewed and classified according to causality. Comparison was made to cases self-reported to departmental morbidity and mortality; conditional probability analysis assessed the likelihood of reporting as a function of causality.ResultsOf 104,938 operations performed, automated CIS screening identified 1,010 cases requiring unplanned reoperation; 23.6% were self-reported to morbidity and mortality; the probability of reporting varied widely depending on causality.ConclusionsScreening of a CIS for adverse events requiring reoperation revealed significant underreporting, with additional bias in reporting based on underlying causality.