Cuffless Blood Pressure Estimation and Hypertension Risk Screening: Investigating the Trade-Off Between Calibration...

Continuous monitoring of blood pressure using photoplethysmography (PPG) offers a promising non-invasive solution for hypertension management. This study presents a regression framework leveraging Gradient Boosting Machines and morphological feature extraction from PPG signals, including first-and second-order derivatives, to estimate systolic and diastolic blood pressure. To improve accuracy, different calibration intervals were explored, in which each calibration utilized three consecutive cuff-based blood pressure measurements. The trade-off between calibration frequency and blood pressure estimation performance was analyzed, revealing that 2-week calibration provides the best balance between convenience and accuracy, meeting the standards of AAMI, BHS, and IEEE 1708. In addition, the impact of calibration on hypertension detection was evaluated, with calibration intervals of 2 to 3 weeks achieving the optimal trade-off between frequency, sensitivity, specificity, and precision. These findings underscore the potential of integrating calibrated PPG-based blood pressure estimation into hypertension screening systems, advancing non-invasive cardiovascular disease management.Clinical relevance- This cuffless, non-invasive method offers accurate blood pressure monitoring and reliable hypertension detection, providing clinicians with a practical tool for early diagnosis and continuous management of cardiovascular conditions.