Equity-constrained dispatching models for emergency medical services

Abstract

Purpose The purpose of this paper is to apply agent-based modeling and simulation concepts in evaluating different approaches to solve ambulance-dispatching decision problems under bounded rationality. The paper investigates the effect of over-responding, i.e. dispatching ambulances even for doubtful high-risk patients, on the performance of equity constrained emergency medical services. Design/methodology/approach Agent-based modeling and simulation was used to evaluate two different dispatching policies: first, a policy based on maximum reward, and second, a policy based on the Markov decision process formulation. Four equity constraints were used: two from the patients’ side and two from the providers’ side. Findings The Markov decision process formulation, solved using value iteration method, performed better than the maximum reward method in terms of number of patients served. As the equity constraints conflict with each other, at most three equity constraints could be enforced at a time. The study revealed that it is safe to over-respond if there is uncertainty in the risk level of the patients. Research limitations/implications Further research is required to understand the implications of under-responding, where doubtful high-risk patients are denied an ambulance service. Practical implications The need for good triage system is apparent as over-responding badly affects the operational budget. The model can be used for evaluating various dispatching policy decisions. Social implications Emergency medical services have to ensure efficient and equitable provision of services, from the perception of both patients and service providers. Originality/value The paper applies agent-based modeling to equity constrained emergency medical services and highlights findings that are not reported in the existing literature.