Predicting Time‑Critical Interventions and Smarter Dispatch

JEMS Development Editor Mike Brown speaks with Aaron Weedman, PhD, technical lead for machine learning in emergency medicine at the University of Pittsburgh, and Dr. Lenny Weiss, EMS physician and medical director for air and ground services. They discuss how natural language processing can analyze prehospital free-text data, such as chief complaints and 911 call notes, to predict the need for critical interventions like airway management, hemorrhage control, blood transfusions, and CPR. The technology may also help determine whether patients need ALS or BLS resources. Using data (about 10 million chief complaints) and 20,000 to 25,000 911 calls, the researchers found the models performed well and closely matched existing dispatch protocols. The goal is to improve dispatch accuracy, mobilize resources faster, and reduce cognitive workload for busy EMS crews.