Peter Vanberkel

Creating a Guide to Identify Patients Who May Leave an Emergency Department Without Being Seen: A Machine Learning Approach

Patients who seek care in an Emergency Department (ED) may ultimately choose to leave without being seen by a physician. This occurrence is labeled “left without being seen” (LWBS). Patients who LWBS are at risk and may experience clinical deterioration related to delayed diagnosis or treatment. Identifying which patients are more likely to LWBS (and intervening) prevents adverse outcomes. In this study, our goal is to create a paper based guide to proactively identify patients at risk of LWBS that can be used by staff in real time. Descriptive statistics and several machine learning models identify the most important features and their typical ranges. Thresholds values for these feature combinations are determined and form the foundation for the guide which is designed to fit a single piece of paper. The guide was developed using data from the Pediatric ED at IWK Health in Halifax, Nova Scotia, Canada.