A dynamic simulation–optimization approach for managing mass casualty incidents

Abstract

To handle mass casualty incidents (MCIs), emergency medical services in central European countries such as Austria or Germany are used to build up small field hospitals called advanced medical posts (AMPs). The main policy behind is to categorize, treat, and stabilize emergency patients before they are transported to hospitals. To investigate the organization of the AMP type operated in Austria such as by Austrian Samaritan Organization, we developed a discrete event simulation (DES) policy model for MCIs regarding advanced process management including staff allocation (physicians, medics) for emergency patient treatment and transport at the incident site. As health care policy makers aim at improving the training of emergency staff, we developed a management policy game as a training tool adapted from the DES policy model. In this paper, we address the research question whether or not the policy decisions made by simple heuristics or disaster management players could be further improved by using advanced simulation–optimization techniques. We present a generic method consisting of (i) an automated policy for dynamic staff re-allocation at an AMP with arbitrary structure, and (ii) a simulation–optimization approach for optimally parametrizing this automated policy. Three simulation–optimization techniques with two complexity levels are investigated in detail for the purpose of incorporation in our system applied to the Austrian AMP case study: the method by Kiefer–Wolfowitz, the metaheuristic OptQuest approach, and the Response Surface Methodology. Our results show that the optimized automated policies can improve the performance of the AMP compared to the management by simple heuristics or by human decision makers. We discuss policy implications for improving strategic decision making and process management for incident commanders at the Austrian AMP based on the results of the dynamic simulation–optimization techniques.