Computer Modeling of Patient Flow in a Pediatric Emergency Department Using Discrete Event Simulation

TLDR

Increasing patient census and department overcrowding are universal concerns in pediatric emergency medicine. The study constructs a patient flow model using discrete event simulation and a physician scheduling tool to predict and test staffing scenarios that could reduce waiting times, improve flow, and increase staff efficiency in the pediatric emergency department. The authors built a discrete event simulation model of PED patient flow with Arena software, based on 517 patient observations and historical arrival data, and validated it by comparing simulated annual flow with real data. The model was validated with high accuracy, and simulations showed that adding a volunteer and a second triage nurse cut pretriage wait times, while an extra physician shift shortened length of stay across all triage categories, demonstrating its ability to predict improvements in patient flow.

Abstract

Increasing patient census and department overcrowding are universal concerns in pediatric emergency medicine. Accurate predictions of patient flow and resource utilization in the pediatric emergency department (PED) are important in determining what aspects of PED activity could be modified to improve patient flow, reduce patient waiting times, and increase staff efficiency and morale, and thus direct change more effectively.We report (1) the construction of a Patient Flow Model (PFM) using discrete event simulation to test simulated PED staffing scenarios that were designed to alleviate the pressures that result from increased census and overcrowding, and (2) a Physician Scheduling Analysis Tool to assist in physician scheduling.Arena discrete event simulation modeling software was used to develop a model of PED patient flow after extensive interviews with PED staff and direct observation of patient flow in July 2005. A total of 517 patients were directly observed, and all modeled aspects of their interaction with PED staff and resources were recorded. Historical demographic patient arrival information was combined with observed patient flow data to provide simulated patient arrival rates for the PFM and was also used to construct the Physician Scheduling Analysis Tool. Validation of the PFM was performed by comparing annual simulated patient flow data with actual patient flow data. Previously determined staffing scenarios were applied to the simulation and the resulting performance indicator outputs examined.The PFM was validated on model-wide and process-specific levels, with excellent validation observed on high acuity-patient length of stay and for highly detailed processes such as triage and registration. Simulation of the addition of a hospital volunteer and a second triage nurse demonstrated reductions in pretriage waiting time and the proportion of patients waiting longer than 30 or 60 minutes for pretriage. Simulation of an extra physician shift to the staff schedule demonstrated reductions in length of stay for patients of all triage categories.The PFM accurately represents patient flow through the department and can provide simulated patient flow information on a variety of scenarios. It can effectively simulate changes to the model and its effects on patient flow.