Performance of Machine Learning Models for Sepsis and Stroke Detection Using EMS Data

The authors aimed to evaluate the feasibility and performance of machine learning models for early detection of sepsis and stroke using emergency medical services (EMS) electronic health record data. Their study found that while the models demonstrated reasonable sensitivity and specificity, particularly when applying a majority prediction approach, incorporating free-text narratives improved sensitivity at the cost of specificity. The findings suggest that machine learning could enhance prehospital identification of these conditions, with future research needed to refine models using EMS-specific data.