A Simulation--Based Optimization approach for analyzing the ambulance diversion phenomenon in an Emergency-Department network

TL;DR

This paper introduces a simulation-based optimization model to evaluate ambulance diversion strategies in emergency department networks, aiming to reduce patient wait times and costs through optimal resource allocation, validated on a real Italian case study.

Contribution

It presents a novel SBO-based model for analyzing ambulance diversion policies, combining discrete event simulation with bi-objective optimization for the first time.

Findings

Optimal AD policies reduce patient non-value added time.

Different diversion strategies impact overall costs and efficiency.

The model is validated with real-world data from Italian EDs.

Abstract

Ambulance Diversion (AD) is one of the possible strategies for relieving the worldwide phenomenon of Emergency Department (ED) overcrowding. It can be carried out when an ED is overloaded and consists of redirecting incoming by ambulance patients to neighboring EDs. Properly implemented, AD should result in reducing delays of patient treatment, ensuring safety and rescue of life-threatening patients. From an operational point of view, AD corresponds to a resource pooling policy among EDs in a network. In this paper we propose a novel model for studying the effectiveness of AD strategies, based on the Simulation-Based Optimization (SBO) approach. In particular, we developed a discrete event simulation model for reproducing the ED network operation. Then, for each AD policy considered, we formulate and solve an optimal resources allocation problem consisting of a bi-objective SBO problem where the target is the minimization of the non-value added time spent by patients and the overall cost incurred by the ED network. A set of optimal points belonging to the Pareto frontier is obtained for each policy. To show the reliability of the proposed approach, a real case study consisting of six large EDs in the Lazio region of Italy is considered, analyzing the effects of adopting different AD policies.