| This document provides an overview of a system dynamics model developed in collaboration with The Ottawa Hospital (TOH), with the aim of determining the suitability of applying system dynamics to model patient flows through a hospital. To this end, a simulation was constructed that enabled the comparison and assessment of various resource allocation and patient flow scenarios based on the simulation outcomes, primarily regarding the duration of patient stays in the ED.Regarding the model structure, this document outlines the simulated flow of patients through triage, the various ED areas and inpatient services until discharge. The delays experienced throughout the patient flow are related to the utilization of various resources, both within the ED and elsewhere in the hospital, enabling the model to determine expected delay durations. The model was tested through direct comparisons between the true historical durations and those predicted by the model for each of the key stages of ED patient flow. It was found that ED patient routing, waiting room delays and ED treatment durations were quite accurately predicted. For admitted patients, delays prior to ED departure demonstrated slightly more error, however still performed quite well and did not inhibit scenario comparison. Outcomes of scenarios requested by the hospital and their implications are also discussed.Based on the model verification outcomes, the scenarios tested and the usefulness of the model output, the conclusion that SD is a feasible and useful approach to modelling patient flows is discussed, supplemented by a discussion of the merits of using SD in the development of such a model.The model is intended to aid with strategic planning, which necessitated the expansion of the model beyond the ED alone to include its interactions with hospital inpatient services. These requirements provided the opportunity to investigate the suitability of SD to model patient flow, with this document outlining the motivation for the research, the factors influencing the selection of SD as the modelling method, the structure and validation of the model, and scenarios tested. |
