AI-Assisted Insulin Pump Setting Optimization for Adults with Type 1 Diabetes Using Continuous Glucose Monitoring

What this Trial Is About

Researchers are evaluating an AI-assisted remote tool designed to optimize insulin pump settings for people with type 1 diabetes T1D who use insulin pumps and continuous glucose monitoring CGM. This study aims to assess the safety, feasibility, and effectiveness of AI-generated personalized pump adjustments, with all recommendations reviewed by physicians before being sent to participants through a mobile app. The study includes two separate groups one inpatient and one outpatient cohort. The inpatient cohort involves hospitalized adults with T1D undergoing AI-assisted insulin pump optimization for 5 to 7 days under close clinical supervision to study short-term safety and glycemic responses. The outpatient cohort participates in an 11-week home-based crossover trial where participants receive either AI-assisted remote optimization or conventional insulin pump therapy for two 4-week periods, with weekly data uploads and physician-reviewed AI recommendations. The AI system analyzes CGM and insulin data to suggest adjustments, aiming to improve glycemic control. Participants will wear study devices continuously, upload glucose and insulin data regularly, and perform daily fingerstick blood glucose measurements. Researchers monitor time-in-range glucose levels, hypoglycemia, hyperglycemia, insulin doses, glycemic variability, and adverse events throughout the study. Patient and physician satisfaction with the AI system are also assessed. The study lasts approximately 12 weeks for outpatients, with continuous safety and efficacy evaluations.

AI-Assisted Remote Insulin Pump Optimization in T1DM