Probabilistic Forecasting of Patient Waiting Times in an Emergency Department

Problem definition: We study the estimation of probability distribution individual patient waiting times in an emergency department (ED). Whereas it is known that waiting-time estimates can help improve patients’ overall satisfaction and prevent abandonment, existing methods focus on point forecasts, thereby completely ignoring underlying uncertainty. Communicating only a forecast to patients be uninformative potentially misleading. Methodology/results: use machine learning approach quantile regression forest produce probabilistic forecasts. Using large patient-level data set, we extract following categories predictor variables: (1) calendar effects, (2) demographics, (3) staff count, (4) ED workload resulting from volumes, (5) severity condition. Our feature-rich modeling allows for dynamic updating refinement as patient- ED-specific information (e.g., condition, congestion levels) revealed during process. The proposed generates more accurate forecasts when compared with literature rolling average benchmarks typically used practice. Managerial implications: By providing personalized our gives low-acuity first responders comprehensive picture possible trajectory provides reliable inputs inform prescriptive operations. demonstrate publishing ambulance selecting network EDs, which lead uniform spread load across network. Aspects relating communicating uncertainty implementing this methodology practice are also discussed. For healthcare service providers, could assist routing, allocation, managing flow, facilitate efficient operations cost savings aid better care outcomes. Supplemental Material: online supplement available at https://doi.org/10.1287/msom.2023.1210 .