Discrete event simulation approach to bed capacity optimization at the mater hospital’s emergency unit

Abstract

Ambulance diversion and long patient waiting time are two undesirable effects of emergency department constricted patient flow and indicate a lack of capacity optimization. On the other hand, capacity optimization for an Emergency Department (ED) is elusive due to the stochastic nature of patient arrivals, length of stay, service rate and discharge patterns. The study’s main purpose was to design a Discrete Event Simulation model that would enhance emergency patients’ bed capacity optimization and ED throughput management by incorporating the concept of uncertainty in its predictions. The study pursued the following three specific objectives; to establish emergency patient arrival and exit patterns at Mater Hospital, to determine the relationship between emergency patient arrival rate (μ) and the exit rate (ƛ), and to establish the optimum emergency patient bed capacity for Mater Hospital’s emergency unit. This was a case study that relied on historical data kept by the hospital to build probability distributions for patient arrivals, waiting times, service rates and exit rates for the Emergency unit. The processed probability distributions formed the input data for the Discrete Event Simulation (DES) model. The DES model was iterated many times (1000) in order to increase the accuracy of the model output information necessary for decision making. The model performance accuracy was also measured through the process of Validation which compared actual data with data from the simulation model using the student t-test procedure at 95 percent confidence level assuming equal variance. The study established that patient arrival and exit patterns at the Mater Hospital were highly random with the arrival rate (μ=3) hovering above the discharge rate (ƛ=2). The mean patient waiting time dropped from approximately 32 minutes currently to 12 minutes after introducing efficiency improvements of an additional two patient beds during simulation.This represents a 60% drop in waiting time. The model was a reasonably accurate representation of emergency patient flow at the Hospital due to the higher calculated p-value (p=0.97, d.f=238) than the 0.05 critical value implying the lack of a significant statistical difference between the actual data and the corresponding simulated data. Employing Discrete Event Simulation in an Emergency Department of a Hospital can solve capacity management and flow problems significantly.