Published on
An artificial intelligence (AI) tool combined with real-time biometric tracking and human coaching produced significant clinical improvements for adults with type 2 diabetes in a Cleveland Clinic study published in NEJM Catalyst. The intervention delivered personalized nutrition and lifestyle guidance through an AI-powered app that analyzed health data from wearable sensors and devices, which monitored blood glucose, weight, blood pressure, activity, and sleep. App recommendations were then based on predicted glucose responses to meals. The study authors recruited 150 patients—100 were assigned to the app and wearable intervention group and 50 to standard care. All patients were prescribed metformin with dose adjustments as needed. Researchers found 71% of patients in the intervention arm who were using the app and wearables achieved the primary endpoint: lowering HbA1c below 6.5% while remaining on metformin alone. And just 2.4% of those receiving standard care reached the same endpoint. Participants had an average HbA1c of 7.2% before the study. Over 12 months, patients in the intervention group saw a 1.3% reduction in HbA1c, compared with just 0.3% in the standard care group (P<0.001). Patients in the intervention group also experienced greater weight loss (8.6% vs. 4.6%) and reduced dependence on glucose-lowering medications.Â
Eat what you like: With the intervention, there was no cap on calories, and the app allowed for a personalized diet with individual food preferences taken into consideration. Foods were color-coded to encourage healthier choices, and a bar code-scanning feature in the app helped participants identify foods they preferred that were also recommended for their lifestyle program.