Real-time resource allocation in the emergency department: A case study

Overcrowding is a phenomenon that affects Emergency Departments (EDs) worldwide determining a harmful impact on the healthcare provided. Because of the wide variety of different patient paths, in the literature the ED processes are usually modeled making significant assumptions, and neglecting fundamental aspects. Such assumptions could make sense for strategic or tactical decisions but nowadays the objective most frequently required by practitioners is the optimization of already available resources. To deal with this problem, we need to act at the operational level, with a particular attention to resource allocation in real time since arrivals and activities to be performed are known only over time. In this paper we present the case study of an Italian ED to investigate if an online allocation algorithm based on prioritization combined with a prediction tool can improve the ED performance, alleviating the overcrowding. Then, we propose several policies for the online allocation of the ED resources, which take into account the real-time state of the ED and the prediction of the next activities provided by an ad hoc process mining model. The proposed approach is validated and analyzed on the case study through a fine-grained simulation model. Results suggest that if the decisions of allocating resources for the execution of activities take into consideration the probably subsequent activities, then there is a room for improvement for the average door-to-doctor time, ED Length-of-Stay, and resource utilization.