Modeling and Improving Patient flow at an Emergency Department in a Local Hospital Using Discrete Event Simulation

Abstract

Like many hospitals in the US, a local hospital was experiencing patient dissatisfaction as a result of long length of stay (LOS), long waiting time, and crowded waiting room in its emergency department (ED). To help analyze its process and justify proposed changes at this ED, discrete simulation models were built using the software ProModel in this thesis. Discrete event simulations are used in many different industries for process improvement, including the health care system. This study started from literature review in both simulation and emergency medicine fields, aiming to identify best practices in both methodologies and ED practices. Then, a careful data collection and analysis was performed. Besides the large data files provided by the hospital, data were also collected through many observations of the system, interviews with staffs, and time studies to provide valid and accurate input to the simulation models. The project was completed in two phases: in phase I, a simple simulation model was built to study the impact of bottlenecks identified by the ED staffs. Experiments were conducted to show possible improvements that would be achieved if the process at the bottleneck locations could be improved. Some best practices reviewed in literature such as adding a discharge lounge was modeled to see how it could improve patient flows in the system. In phase II, more simulation models were built with a lot more details with an aim to study the impact of a proposed change to the process. . The models built in both phases were verified and validated in multiple ways. Experimentations were run and hypothesis tests were performed to confirm if the suggested changes will improve certain measurements such as the patient length of stay and patient waiting time.