An Integrated IoT-Enabled Wearable and Machine Learning Approach for Continuous Cardiovascular Disease Risk Monitoring

Abstract

This study aimed to develop an Internet of Things (IoT)-enabled wearable healthcare framework for continuous physiological monitoring and cardiovascular disease (CVD) risk classification. CVD is a major global public health problem, and continuous monitoring of physiological parameters may support timely identification of individuals requiring further clinical assessment. Wearable IoT devices can facilitate the collection and transmission of physiological data, while machine learning methods can assist in identifying patterns associated with cardiovascular risk. The proposed framework integrates wearable sensor data, secure data transmission, patient medical records, data preprocessing, and a Bidirectional Long Short-Term Memory (Bi-LSTM) model for CVD classification. The study used 294 records from a dataset described by the authors as an Indian CVD database. Preprocessing included missing-value treatment, removal of redundant or irrelevant features, and feature standardization. The classification performance was evaluated using accuracy, precision, sensitivity/recall, specificity, and F-score. The proposed Bi-LSTM framework obtained an accuracy of 92.6%, precision of 94.1%, sensitivity of 93.7%, specificity of 91.5%, and F-score of 94.7% based on the performance values reported in the manuscript. These values should be checked against the original experimental data, confusion matrix and final evaluation data before publication as other performance values are reported elsewhere in the manuscript. The proposed framework shows the feasibility of using wearable IoT technologies and Bi-LSTM machine learning for CVD risk classification. The results are, however, based on a relatively small data set and are not prospective clinical validation of continuously collected real-world wearable data. Additional validation with independent, larger, and prospectively collected datasets is needed before clinical use.