Health-LLM is a framework that leverages large language models (LLMs) for health prediction using wearable sensor data. The study evaluates 12 state-of-the-art LLMs on four public health datasets (PMDat, LifeSnaps, GLOBEM, and AW_FB) across 10 health prediction tasks, including mental health, activity, metabolism, and sleep assessment. The research explores the effectiveness of prompting and fine-tuning techniques, with a focus on context enhancement strategies. The fine-tuned model, HealthAlpaca, achieves comparable performance to larger models like GPT-3.5, GPT-4, and GeminiPro, excelling in 8 out of 10 tasks. Ablation studies show that context enhancement can improve performance by up to 23.8%. The study highlights the importance of incorporating health knowledge and temporal context in prompts for accurate health predictions. The results demonstrate that LLMs can effectively process and predict health-related data, with HealthAlpaca showing strong performance in various health tasks. The study also emphasizes the need for further research to address ethical concerns, data validity, and model interpretability in real-world applications.Health-LLM is a framework that leverages large language models (LLMs) for health prediction using wearable sensor data. The study evaluates 12 state-of-the-art LLMs on four public health datasets (PMDat, LifeSnaps, GLOBEM, and AW_FB) across 10 health prediction tasks, including mental health, activity, metabolism, and sleep assessment. The research explores the effectiveness of prompting and fine-tuning techniques, with a focus on context enhancement strategies. The fine-tuned model, HealthAlpaca, achieves comparable performance to larger models like GPT-3.5, GPT-4, and GeminiPro, excelling in 8 out of 10 tasks. Ablation studies show that context enhancement can improve performance by up to 23.8%. The study highlights the importance of incorporating health knowledge and temporal context in prompts for accurate health predictions. The results demonstrate that LLMs can effectively process and predict health-related data, with HealthAlpaca showing strong performance in various health tasks. The study also emphasizes the need for further research to address ethical concerns, data validity, and model interpretability in real-world applications.