Continuous Blood Pressure Monitoring and Hypertension Risk Screening Using Smart Watch

Traditional methods for monitoring blood pressure (BP) often fall short due to limited monitoring and inconvenience. This study addresses the limitations of traditional BP monitoring methods by introducing a cuff-less approach utilizing commercially available smartwatches. The proposed framework employs wrist photoplethysmogram (PPG) signals to derive a robust set of morphological features for characterizing variations in systolic and diastolic blood pressure. The model was trained on PPG sensor data from 438 subjects and validated on an independent set of 347 subjects. Model performance was evaluated under 2 window resolutions (60 seconds and 10 seconds PPG data) with study protocol of resting and exercise transitions to simulate dynamic blood pressure variance scenarios. The proposed model meets the standards set by the Association for the Advancement of Medical Instrumentation (AAMI) for both 60-second and 10-second window resolutions. Additionally, it attains IEEE 1708 and British Hypertension Society (BHS) grade A for 60-second PPG data and grade B for 10-second PPG data, demonstrating its efficacy in blood pressure monitoring and hypertension risk screening using smartwatches.