Improving EMS Transport Decisions with Artificial Intelligence to Enhance Patient Outcomes

Emergency Medical Services (EMS) play a critical role in determining where patients receive definitive care during medical emergencies. EMS providers must rapidly assess patients in the field and decide which hospital is most appropriate based on their condition. These decisions have lasting implications: mismatches between a patient’s needs and a hospital’s capabilities can lead to treatment delays, avoidable transfers, and poorer outcomes. This project will examine how information gathered by EMS personnel relates to what ultimately happens to patients at the hospital, such as whether they are discharged from the emergency department, admitted for inpatient care, or transferred elsewhere. 

Using six years of statewide data from New Jersey, we will link EMS records (collected using national reporting standards) with hospital discharge records to identify which prehospital indicators, such as symptoms, vital signs, or field treatments, are most predictive of patient disposition. The first phase will involve retrospective data analysis to uncover patterns and correlations. The second phase will develop and evaluate predictive models using machine learning to determine whether hospital outcomes can be accurately forecasted based on EMS-collected data alone. 

By bridging prehospital and hospital data, this research will generate new insights into the factors that shape emergency care outcomes. Findings will inform improved triage guidelines, destination decision-making, and future tools to support EMS personnel in the field, ultimately leading to more efficient, accurate, and equitable emergency care delivery across New Jersey.