Performance analysis in emergency departments: a data-driven approach

Purpose: This paper aims to investigate the process performances in Emergency Departments (EDs) with a novel data-driven approach, permitting to discover the entire patient-flow, deploy the performances in term of time and resources on the activities and flows and identify process deviations and critical bottlenecks. Moreover, the use of this methodology in real time might dynamically provide a picture of the current situation inside the ED in term of waiting times, crowding, resources, etc., supporting the management of patient demand and resources in real time. Design/methodology/approach: The proposed methodology exploits the process-mining techniques. Starting from the event data inside the hospital information systems, it permits automatically to extract the patient-flows, to evaluate the process performances, to detect process exceptions and to identify the deviations between the expected and the actual results. Findings: The application of the proposed method to a real ED revealed being valuable to discover the actual patient-flow, measure the performances of each activity with respect to the predefined targets and compare different operating situations. Practical implications: Starting from the results provided by this system, hospital managers may explore the root causes of deviations, identify areas for improvements and hypothesize improvement actions. Finally, process-mining outputs may provide useful information for creating simulation models to test and compare alternative ED operational scenarios. Originality/value: This study responds to the need of novel approaches for monitoring and evaluating processes performances in the EDs. The novelty of this data-driven approach is the opportunity to timely connect performances, patient-flows and activities