Original title: ALPHA: AnomaLous Physiological Health Assessment Using Large Language Models
Authors: Jiankai Tang, Kegang Wang, Hongming Hu, Xiyuxing Zhang, Peiyu Wang, Xin Liu, Yuntao Wang
In this article, researchers explore the potential of Large Language Models (LLMs) in healthcare, specifically in monitoring individual health anomalies. They assess these models’ ability to interpret physiological data from FDA-approved devices, conducting a thorough analysis in a simulated low-air-pressure environment. The study reveals impressive precision of LLMs in evaluating health indicators, boasting less than 1 beat per minute error for heart rate and less than 1% error for oxygen saturation (SpO2). Moreover, their adapted models showcase proficiency in image analysis, achieving less than 1 bpm error in cycle count and 7.28 Mean Absolute Error (MAE) for heart rate estimation from photoplethysmography (PPG) data. Demonstrating potential as advanced AI health assistants, these LLMs offer personalized health insights and recommendations, indicating a promising role in future healthcare frameworks.
Original article: https://arxiv.org/abs/2311.12524