ATTD 2025 Amsterdam & Online—19–22 March and ATTD-ASIA 2024 Singapore—18–20 November

Background and Aims: The DiaGame project develops and tests a new concept to empower patients with diabetes in their self-management. We combine real world data with advanced computational algorithms and models and serious gaming.<br/>Methods: Data has been collected in daily-life settings in sixty participants (52 type 2 diabetes, 8 type 1 diabetes) using a continuous glucose monitor, smartwatch with in-house software, and public smartphone application acquiring data on blood glucose levels, heart rate, step count, acceleration, participant-reported dietary intake, physical activity, insulin dosing, and mood. We develop and validate a physiology-based mathematical model of the glucose-insulin system to describe and predict glycemic dynamics. This model simulates glucose and insulin concentrations following a mixed meal or OGTT and has previously been validated.<br/>Results: The model has been successfully extended to describe the subcutaneous absorption of various insulin analogues, which is important for patients using insulin (type 1 diabetes and type 2 diabetes with intensive insulin therapy). The model accurately captures patient-specific glycemic dynamics of people with T1D. To realize personalized models for T2D patients it is important to consider residual insulin production capacity. By analyzing and comparing personalized models the independent and coherent effects of various lifestyle factors on glycemic dynamics are elucidated.<br/>Conclusions: Patient digital twins are developed by updating personalized prediction models with patient-gathered, real world data. Ultimately, these digital twins will be employed to aid self-management education through a personalized interactive serious game. A prototype of this game is currently being evaluated by patients and caregivers.