Most health care organizations primarily rely on retrospective techniques such as incident reporting systems to detect and respond to patient harm. Retrospective techniques do not capture a substantial proportion of patient safety events and are not effective for real-time feedback or safety incident prevention. By contrast, electronic health record–based trigger tools detect safety incidents in real time. Investigators developed and piloted the patient safety active management (PSAM) at two large community hospitals. The PSAM employed machine learning and expert clinician review of years of adverse event data to refine the Institute for Healthcare Improvement Global Trigger Tool and distilled this into a real-time risk score for patient safety incidents. The PSAM identified 10 times more safety incidents than were detected through retrospective techniques. It reliably predicted both serious adverse events and mortality. A prior PSNet perspective explored how to leverage electronic health records to enhance safety.

The Moore Foundation provides free access to this article.