DESIGNING EMERGENCY MEDICAL SERVICES PROCESSES TO MINIMIZE THE IMPACT OF AMBULANCE OFFLOAD DELAY

Abstract

Ambulance offload delay (AOD) occurs when the care of incoming ambulance patients cannot be transferred immediately from paramedics to staff in a hospital emergency department (ED). This is typically due to ED congestion. In such situations, paramedics are responsible to provide patient care until an ED bed becomes available. AOD can negatively affect ambulance availability to future calls and reduce the efficiency of the emergency medical services (EMS) system. Hence, this problem has become a significant concern for many healthcare providers and is the focus of this dissertation. In this dissertation, we develop several models to analyze AOD. With 12-months of emergency call data provided by the provincial EMS provider and local hospitals in Nova Scotia, Canada, we conduct an empirical analysis to measure the effects of AOD on the EMS system. The analyzed performance metrics include the number of ambulances at EDs, ambulance turnaround time, total call time, response time, and ambulance availability. The results indicate significant negative effects on all these metrics within the region experiencing AOD. AOD also has a negative impact on ambulance availability in adjacent regions for an EMS system with shared resources. We then develop a decision-support tool using a novel hybrid decision tree model to predict the severity of AOD within 1 to 5 hours based on the current system status. The objective of this study is to provide a prediction model for EMS decision makers so that proactive interventions at different system states can be initiated to mitigate AOD. The hybrid algorithm shows improvements in the classification of this real-world problem when tested against a basic decision tree algorithm. Finally, we develop an optimal ambulance destination policy using a discrete time, infinite-horizon, discounted Markov Decision Process. This model helps determine when it is advantageous to send appropriate patients to out-of-region EDs, which have longer transport times but shorter offload times. The optimal policy can significantly reduce AOD, time-to-ED bed for patients, and out-of-service time for paramedics at the expense of increased ambulances travel distances.