Machine Learning-Based Demand Prediction for Emergency Medical Services

Emergency Medical Services (EMS) are critical to deliver out-of-hospital care by deploying specialized resources to treat patients. To optimize EMS operations, various models are proposed in literature to predict demand and allocate resources more efficiently. Demand patterns are influenced by numerous factors, including demographics, territorial characteristics, and environmental conditions. In Lombardy region, Northern Italy, the public agency responsible for managing emergency services (AREU) relies on a digital twin of the EMS system, developed in collaboration with the Polytechnic of Milan. While this tool offers significant insights, it involves time-intensive processing, especially for simulating future emergency demand, and does not incorporate critical temporal patterns, such as environmental conditions. This highlights the need for a more efficient model. Hence, this study proposes enhancing the existing digital twin with machine learning techniques to improve demand forecasting, integrating external time-variant factors and reducing processing time. Multiple models are tested and compared, and the performance of the selected one is evaluated, considering error metrics, against real historical emergency data and baseline model. Results demonstrate that the machine learning-enhanced model significantly improves predictive accuracy, reducing RMSE from 8,7 to 2,5 events/hour. This approach also substantially reduces computational time, enabling day-by-day predictions and real-time decision making for EMS deployment and resources optimization. These findings underscore the potential of machine learning to provide valuable insights for policy development, aiming to better align EMS resources with demand dynamics and to ultimately improve emergency response outcomes.Clinical Relevance- This quantifies the expectable benefits from the application of machine learning methods in predicting the demand for emergency medical services, which is instrumental in optimizing the service and delivering better out-of-hospital care for patients.