Machine Learning-Driven Cardiovascular and Stroke Screening Using IoT-Based Health Monitoring Systems

Abstract

The CVD and stroke global burden has become higher thereby emphasizing the need for improved surveillance and control measures. The utilization of wearable technologies and Heart Rate Variability (HRV) analysis accompanied by machine learning (ML) offer a novel preventive strategy. As a biomarker of ANS integrity, HRV allows evaluating the cardiovascular and stroke risks with no invasiveness. These devices use real-time data capture, data preprocessing, and ML algorithms to offer twofold, early warning alarms for looming risks and timely alarms for response at critical health episodes. This paper reviews the developments of the recent studies in wearable health monitoring systems in particular concerning the system design, its applications and effects on predicting cardiovascular diseases. The integration of IoT frameworks with wearable improves their scalability in terms of data flow and subsequent cloud computing. Also, the recent modification to hybrid models and employing of deep learning has enhanced the efficacy of the technique in HRV analysis for effective health risk estimations. However, they come with challenges such as: sensor accuracy, energy, and the applicability of the algorithms across different demographics. These limitations are also described in this paper, and following research directions are proposed: the creation of multi-modal devices that would include more biomarkers such as respiratory rate or blood pressure, and the improvement of the energy efficiency of algorithms for continuous monitoring. This paper systematically reviews recent research to describe the current state and future possibility of wearable health technologies in changing global healthcare approach and managing illnesses with timely interventions to enhance patient’s health outcomes.