The authors report on an extension of the HELP clinical information system at LDS Hospital, Salt Lake City, to provide automated detection of adverse drug events (ADEs) from clinical data. The authors describe a rule-based approach wherein certain medications (such as naloxone or diphenhydramine), laboratory values (such as leukopenia or a positive C. difficile toxin assay), or unanticipated stop orders for medications are reported daily to a clinical pharmacist, who investigates the report by chart review and provider interview. Concurrently, the authors provided an enhanced online system for traditional voluntary incident reporting to supplement the existing paper-based methods. The authors describe a 60-fold increase in detection of ADEs with the automated system relative to voluntary reporting and provide a detailed breakdown of the types of ADEs detected. This study was among the first to demonstrate the value of real-time computerized analysis of clinical data to intervene on patient safety in a widely deployed, operational information system.