IoT-Based Early Heart Failure Prediction: A Hybrid Method Using ECG Signal Analysis and the Framingham Risk Score

This study introduces a hybrid IoT-based health monitoring system aimed at early prediction of heart disease, which combines real-time ECG data with the Framingham Risk Score (FRS) through machine learning techniques. Conventional models, although efficient, frequently exhibit limitations in their adaptability to real-time data and precision in intricate situations. The proposed hybrid approach mitigates these limitations by integrating the advantages of traditional statistical models with deep learning techniques. An Arduino module acquires real-time ECG signals, which are integrated with patient demographic and clinical data to create robust predictive models. The system utilizes a Random Forest Classifier for the analysis of FRS data and EfficientNetB0 for the classification of ECG signals, thereby delivering precise and comprehensive evaluations of heart health. This integration facilitates ongoing monitoring, prompt detection, and tailored treatment, meeting the demand for accessible and economical solutions in clinical and remote healthcare environments. The hybrid approach improves the system's capacity to detect subtle ECG patterns by integrating established risk factors, thereby connecting traditional methods with advanced machine learning techniques. Preliminary results indicate notable enhance-ments in accuracy: the Random Forest Classifier reached 90.23%, whereas EfficientNetB0 attained 97.11%. These findings under- score the system's potential for early diagnosis, preventive health- care, and informed decision-making, thereby facilitating timely interventions and enhancing patient outcomes.