Sugar slay: a gamified decision support ecosystem for type 1 diabetes

Area of Science:

• Digital Health
• Artificial Intelligence in Medicine
• Behavioral Science

Background:

• Type 1 Diabetes (T1D) management is complex, especially for young adults, requiring constant monitoring of glucose levels, insulin, diet, and activity.
• Existing tools like continuous glucose monitors (CGM) provide data, but the cognitive load of interpretation remains a barrier for patients and caregivers.
• Adolescents and young adults face unique challenges in managing T1D due to new independence and demanding self-care routines.

Purpose of the Study:

• To develop the Sugar Slay ecosystem, a gamified mobile application for T1D self-management.
• To create a companion app, Sugar Slay Care, for caregivers to support T1D patients.
• To integrate advanced machine learning for personalized blood glucose trend prediction and habit-building strategies.

Main Methods:

• Developed a gamified mobile app integrating CGM and wearable data for personalized predictions.
• Utilized advanced machine learning models, including Sequence-to-Sequence Bidirectional LSTM (Seq2Seq BiLSTM), for blood glucose forecasting.
• Conducted a need-finding study with caregivers to inform the design of a supportive companion application.

Main Results:

• The Seq2Seq BiLSTM model showed superior performance in predicting blood glucose trends.
• Identified critical caregiver needs for balancing patient autonomy with safety monitoring.
• Developed Sugar Slay Care, a companion app addressing identified caregiver requirements.

Conclusions:

• Sugar Slay combines predictive AI with gamification for proactive T1D self-management.
• The Sugar Slay ecosystem addresses the social aspect of T1D care, supporting both patients and caregivers.
• This work advances digital health tools by integrating physiological data, AI, and behavioral science for user-centered chronic disease management.