423 Guardian Drive
Philadelphia, PA 19104
Attendees may attend virtually. Zoom link here.
“Advancing Clinical Decision-Making During Triage with Machine Learning”
Emergency department crowding is a pressing challenge across the United States, with triage acting as a critical junction to prioritize patients for immediate care. The Emergency Severity Index (ESI) is a simple algorithm designed to prioritize patients in the waiting room, but variability in assessments among ED staff persist. In this study we analyze over two million ED visits within a large health system from 2021 to 2022 to evaluate the effectiveness of ESI in prioritizing urgent care. First, we show that ESI scores vary conditional on patient acuity. Then, using post-triage outcomes as a measure of validity, we train a machine learning (ML) model to predict cases that should have been triaged as urgent. We use this model to characterize settings in which patients are predictably mistriaged.
Anna Zink, PhD, is a principal researcher at Center for Applied Artificial Intelligence at the University of Chicago Booth School of Business. She works on the algorithmic bias initiative along with other health-related projects. She is interested in the possibilities (and pitfalls) of machine learning techniques to evaluate and improve decision making in health care and insurance design. She received her PhD in Health Policy and a secondary degree in Computational Science & Engineering from Harvard University. Prior to graduate school, she worked as an analyst for several companies where she had the opportunity to work with different types of health care data on a variety of research topics, from identifying Medicare fraud, waste, and abuse to measuring the impact of Medicaid expansion on primary care practices. Learn more about Dr. Zink’s research.