Prediction Models for Operational and Clinical Decision Support in Hospitals

Systems Conversation with Sean Barnes, University of Maryland

Hospitals in the U.S. and around the world are facing significant growth in the demand for services. In particular, emergency departments are facing high levels of crowding, which is often exacerbated by the lack of available beds on the inpatient side. He presents two studies about the use of prediction models at key positions along one of the most commonly traveled patient pathways in a hospital, specifically, triage in the emergency department and discharge from an inpatient unit. Triage in the emergency department often sets the trajectory for a patient’s experience in the hospital, whereas the discharge decision determines when the patient’s stay ends. We show that using prediction models can support decision making for these critical end points, and improve operational efficiency and patient outcomes.