Managing insulin dosing remains a daily challenge for many people with type 1 diabetes mellitus (T1DM) receiving multiple daily insulin injections (MDI), particularly when carbohydrate estimation and dose adjustments must be performed repeatedly throughout the day. A study published in Diabetes, Obesity and Metabolism is evaluating whether an artificial intelligence (AI)-based support system can assist with these treatment decisions in routine diabetes management.

The MELISSA trial is a 22-week European multicenter, prospective, randomized, open-label, blinded-endpoint study designed to clinically validate the MELISSA system in people with T1D receiving MDI therapy. The platform combines an adaptive basal-bolus advisor that provides personalized insulin dose suggestions with an automated dietary assessment tool that converts food images into carbohydrate content estimations.

The trial includes 278 people with T1D and an exploratory cohort of 50 people with type 2 diabetes (T2D) receiving MDI therapy. Following a 6-week initialization period, participants with T1D are randomized 1:1 to either usual care or the MELISSA system, with the primary endpoint assessing between-group change in continuous glucose monitoring time in range from baseline to study end.

Findings

• The primary endpoint is the between-group change in percentage of time spent within the target glucose range of 3.9–10.0 mmol/L (70-180 mg/dL) from baseline to study end.
• Secondary outcomes include additional glycemic metrics, patient-reported outcomes, and safety assessments.
• The study also includes an exploratory feasibility assessment in people with T2D receiving MDI therapy.

The MELISSA trial will assess whether an AI-based support platform can improve glucose management in people with T1D receiving MDI therapy. The findings may help inform future clinical and regulatory evaluation of the MELISSA system in this setting.