Long-term Observation Of Patient Engagement in Automated Insulin Delivery

Automated insulin delivery (AID) systems have revolutionized type 1 diabetes (T1D) management, offering superior glycemic control and quality of life compared to sensor-augmented insulin pumps or multiple daily injections (MDI). However, current AID systems operate as hybrid closed-loop (HCL) systems, requiring user intervention for meal boluses, including carbohydrate counting and timely insulin administration. Control-IQ (Tandem™) was the second AID system reimbursed by the French National Health Insurance and is now widely used in France. It combines the Tandem™ t:slim insulin pump with Dexcom™ G6 continuous glucose monitoring (CGM), adjusting basal insulin every five minutes based on glucose predictions and delivering automatic correction boluses once per hour. Despite these automation features, users must still manually initiate meal boluses, and additional correction boluses remain optional. While engagement in carbohydrate counting and timely meal announcements improves metabolic outcomes, real-world data suggest that AID benefits individuals across varying engagement levels. Long-term data on patient behavior and engagement, particularly regarding carbohydrate announcements and the ratio of automatic to user-initiated boluses, remain limited. Understanding these patterns could reveal whether initially low-engagement users increase involvement over time or whether highly engaged users eventually rely more on automation. This single-center, 24-month observational study will follow all T1D patients treated with Control-IQ at the Centre Hospitalier Sud-Francilien between October 2021 and October 2024.