Medical Injury Case Finding

Administrative data are a useful source for identifying medical injuries in hospitals. The incidence of these events is currently unknown and depends ultimately on how events are defined and data captured. This study elaborated on a conceptual framework expanding the focus on medical errors to include errors that do not result in injury and injuries that do not result from errors. The expansion of scope is necessary to conceptualize a patient safety surveillance system that will be comprehensive, feasible, adaptive, and sustainable as the field increases its understanding of events, their causes, and their outcomes.

This study investigated four case-finding methods applied to a subset of patient safety events in order to identify the overlap among them. Three case-finding methods used administrative hospital discharge data. An additional method used an inpatient voluntary error-reporting system.

The degree to which the detection methods were mutually exclusive was unexpected. Accidental laceration or puncture during surgical or medical care provided the most revealing example: of 476 unique events, only 100 of the same events were detected by each of the three administrative data methods. The error-reporting system performed poorly in detecting this particular patient safety event: only two cases were identified, neither of which was identical to the cases detected by administrative data. This poor performance was not the case with other patient safety events.

We conclude that multiple methods and data systems must be triangulated to maximize the detection of patient safety events.