Skip to content Skip to navigation
Pre-Defense
8/13/2018 10:30 am
CoRE B (305)

Optimizing Task Scheduling in Emergency Departments

Ana Paula Centeno, Dept. of Computer Science

Defense Committee: Prof. Richard Martin (Chair), Prof. Thu Nguyen, Prof. Uli Kremer, Prof. Aritanan Gruber, University of ABC (Santo Andre, SP, Brazil)

Abstract

An Emergency Department (ED) is a health care service that  delivers time-critical care to unscheduled patient arrivals. Due to  an ever increasing number of arrivals, the number of patients often  exceed  the physical and staffing capacity resulting in long waiting  times, patients leaving without being seen by medical staff and  higher mortality levels.  In this work we investigate the  scheduling of staff and equipment resources in EDs. We present a  spatial > agent-based simulation framework to quantify the impacts of  staff  decision processes, such as patient selection, on patient  length of stay and waiting times. Our simulations show there can be  a 17% difference in patient throughput depending upon patient  selection policies.  To investigate ED administration intuition that  reducing the length of stay of shorter visits (patients with simple cases) increases the patient throughput, we propose a scheduling heuristic that prioritizes shorter visits by using decision trees to classify ED vists. We found that by prioritizing shorter visits patient throughput increases by 8%.  We also formulate the ED scheduling problem as a mixed-integer programming model to find the optimal static scheduling for our case study. Finally, we compare  the average length of stay our heuristic scheduler against the  optimal static scheduling and real data obtained from a NJ ED.