Agent-based modeling for estimating the performance of medical service systems in disaster scenarios

Natural disasters pose significant threats to human life and well-being, especially in extreme hazard events. Medical service systems (hospitals) used to treat the ill or injured are vital for reducing fatalities and maintaining societal well-being in these times. However, disasters can reduce the medical system's performance, increasing the gap between medical service supply and the injured demand. In this context, estimating the performance of medical services is important in the post-disaster period. This study proposes an agent-based medical service system performance estimation model (ABMSPEM) considering dynamic interactions between hospital agents and individual agents, with hazards and government countermeasures represented as exogenous parameters. Medical service performance is evaluated on three dimensions: bed occupancy rate, waiting time, and mortality rates. Furthermore, under the agent-based modeling framework, the operation process of hospital agents (registration, diagnostics, surgery, and hospitalization) is modeled by discrete event simulation. Also, the decision-making rules of individual agents are designed, such as the hospital selecting, waiting, and leaving. A case study of Osaka was used to demonstrate the practicality of ABMSPEM. Various experiments related to changes of medical personnel and facilities are simulated to evaluate the effects of hazards and government countermeasures on medical performance. The findings offer insights for improving medical resource allocation and promoting a resilient city.