Abstract: Digital health technologies can assist in disease management and prevention through personalized lifestyle interventions. Wearable devices and smartphones are increasingly being used for continuous monitoring of health status and diseases in daily life, with the goal of maintaining health. This paper aims to demonstrate the potential of wearable devices and smartphones in (1) detecting eating moments and (2) predicting and interpreting individual blood glucose levels in healthy populations, thereby ultimately supporting health self-management. Twenty-four subjects collected continuous data over 14 days through interstitial glucose monitoring, dietary records, activity tracking, and sleep monitoring. We validated the feasibility of using continuous glucose monitoring and activity tracking to identify eating moments, with the constructed prediction model achieving accuracies of 92.3% (87.2–96%) in the training set and 76.8% (74.3–81.2%) in the test set. Furthermore, we demonstrated the ability to predict blood glucose peaks based on dietary records, activity tracking, and sleep monitoring, with an overall mean absolute error of 0.32 (±0.04) mmol/L in the training dataset and 0.62 (±0.15) mmol/L in the test dataset. Using Shapley Additive Explanations, we identified personal lifestyle factors crucial for predicting individual blood glucose peaks, laying the foundation for providing personalized lifestyle recommendations. Although these digital biomarkers require further validation, they show great potential in supporting the prevention and management of type 2 diabetes through personalized lifestyle advice.
