Real-time ambulance redeployment approach to improve service coverage with fair and restricted workload for EMS providers

Abstract

Emergency Medical Services (EMS) managers are concerned with providing maximum possible coverage in their service area. As emergency calls arrive into the EMS system, some ambulances become unavailable. Redeployment deals with a dynamic relocation of available ambulances so as to compensate for the loss in coverage due to busy ambulances. Unsystematic redeployment can impose superfluous workload and result in unnecessary fatigue for EMS personnel. This paper develops a real-time approach to maximize coverage with minimum possible total travel time, considering accumulated workload restrictions for personnel in a shift. While in the past real-time redeployment has been hindered due to computational issues, we find a solution to this problem by combining two computationally inexpensive models into a single framework. The proposed approach requires only knowledge of the current state of the system in a real-time manner and, due to very short run time, is applicable in practice. The performance of our real-time approach is evaluated by a discrete-event simulation developed for a large real dataset and is compared with two benchmarks in the literature; an existing dynamic redeployment approach and a static policy. The results show statistically significant improvement in average coverage, while restricting accumulated workload for EMS personnel as well as providing more evenly distributed workload between ambulances in a shift.