Digital Twin Modeling to Predict Insulin Requirements in Type 1 Diabetes

Abstract

Digital twin models replicate individual patient physiology to forecast insulin needs in real time. In a pilot trial, adults with type 1 diabetes using a digital twin system experienced 20 percent fewer glycemic excursions compared to conventional pump therapy. The model adapted to daily variability in carbohydrate intake and exercise, offering personalized insulin dosing recommendations. Participants reported high usability and improved confidence in self-management. Future work will explore integration with closed-loop devices and remote monitoring systems.