An agent-based simulation for workflow in Emergency Department

Patients always come to an emergency department (ED) to seek immediate medical care. Conversely, usually they experience prolonged waiting for nurses, physicians or radiology/lab procedures. There is a lot of public and private pressure on ED to improve the quality and efficiency of health care. As a result, ED workflow is under continuous changes. However, ED administrators usually lack a convenient decision support tool to evaluate the impacts of a changed workflow. They often find unexpected ED performance degeneracy after some workflow changes. This paper introduces an agent-based simulation which models various phases in the ED workflow, such as triage, nurse screen, resident exam, attending exam, lab/radiology, and disposition. The purpose of this paper is two-fold: (1) to develop an agent-based simulation to allow free exploration of the ED performance under various settings; (2) to characterize and study the ED performance under different settings of the triage process and radiology procedure process. The analysis of two simulated cases is presented to illustrate how the changes of the triage process and the radiology process impact the patient throughput time and the other critical performance measures. The simulation outcome and empirical data demonstrate a similar change pattern after a modification in the triage process. The findings in this study indicate that the agent-based simulation can be used by ED administrators to plan changes, locate bottlenecks, and study the non-linear relationships among different phases in the ED workflow.