A personal health large language model for sleep and fitness coaching by Anastasiya Belyaeva, Xin Liu, Daniel McDuff, Cory Y. McLean ISBN 10.1038/S41591-025-03888- instant download

A personal health large language model for sleep and fitness coaching by Anastasiya Belyaeva, Xin Liu, Daniel McDuff, Cory Y. McLean ISBN 10.1038/S41591-025-03888- instant download

Although large language models (LLMs) show promise for clinical healthcare applications, their utility for personalized health monitoring Check for updatesusing wearable device data remains underexplored. Here we introduce the Personal Health Large Language Model (PH-LLM), designed for applications in sleep and ftness. PH-LLM is a version of the Gemini LLM that was fnetuned for text understanding and reasoning when applied to aggregated daily-resolution numerical sensor data. We created three benchmark datasets to assess multiple complementary aspects of sleep and ftness: expert domain knowledge, generation of personalized insights and recommendations and prediction of self-reported sleep quality from longitudinal data. PH-LLM achieved scores that exceeded a sample of human experts on multiple-choice examinations in sleep medicine (79% versus 76%) and ftness (88% versus 71%). In a comprehensive evaluation involving 857 real-world case studies, PH-LLM performed similarly to human experts for ftness-related tasks and improved over the base Gemini model in providing personalized sleep insights. Finally, PH-LLM efectively predicted self-reported sleep quality using a multimodal encoding of wearable sensor data, further demonstrating its ability to efectively contextualize wearable modalities. This work highlights the potential of LLMs to revolutionize personal health monitoring via tailored insights and predictions from wearable data and provides datasets, rubrics and benchmark performance to further accelerate personal health-related LLM research.