LLM-Powered Personalized Glycemic Assessment in Type 2 Diabetes with Wearable Sensor Data

The paper introduces GlyLLM, a large language model (LLM)-powered framework designed for personalized glycemic assessment in Type 2 Diabetes (T2D) by integrating continuous glucose monitor (CGM) data with structured metadata from wearable sensors. GlyLLM outperformed traditional machine learning methods by achieving a 13.66% improvement in Root Mean Squared Error (RMSE) for glucose forecasting and a 13.08% increase in Area Under the Receiver Operating Characteristic (AUROC) for diabetes categorization on the AI-READI dataset. This advancement highlights the potential of LLMs to enhance personalized healthcare solutions by effectively modeling diverse data modalities and individual-level context.