A Multimodal AI Framework for Intelligent Traditional Chinese Medicine Intervention in Older Adults

Abstract

The increasing digitalization of health management creates new possibilities for integrating traditional Chinese medicine (TCM) with multimodal artificial intelligence, particularly in the long-term management of older adults with heterogeneous health conditions. This study proposes a multimodal AI framework for intelligent TCM intervention and health management, integrating clinical records, symptom assessments, physiological measurements, treatment information, and potentially unstructured clinical observations. The framework emphasizes multimodal feature representation, temporal information integration, patient-specific profiling, and machine-learning-assisted intervention analysis. Rather than assuming that heterogeneous data sources can be directly combined without loss of meaning, the study examines challenges related to data quality, missing observations, modality imbalance, individual variability, and the uncertain computational representation of TCM concepts. Existing research on integrated acupuncture, Tuina, and moxibustion, standardized TCM external therapies, smart wellness systems, and AI-assisted acupuncture provides the clinical and technological context for the proposed framework. The approach may support more continuous and individualized health assessment, although its clinical reliability, interpretability, and generalizability require further validation using prospective, multimodal, and sufficiently diverse datasets.