Systematic review reveals heavy concentration of IoT-LLM healthcare systems in wearable remote monitoring, highlighting critical gaps in clinical validation and data security.
Key Points
To systematically evaluate the integration of Large Language Models (LLMs) with Internet of Things (IoT) technologies in healthcare and identify architectural, security, and clinical validation gaps.
Conducted a systematic literature review adhering to PRISMA 2020 guidelines on a corpus of 75 papers (61 peer-reviewed, 14 preprints).
Compared an AI-powered prompt-based literature search strategy directly against conventional Boolean keyword search queries.
AI-driven prompt searches attained higher retrieval precision than traditional Boolean queries (86% vs. 42%), though constrained by search non-determinism.
Remote patient monitoring and personal health management constituted 59% of identified applications, driven largely by wearable sensor data (64%).
Cloud deployments relied on GPT models for complex reasoning, while edge and federated systems used localized models (BERT, LLaMA) for privacy, with widespread neglect of prompt injection vulnerabilities and empirical regulatory validation.