Study of cuffless blood pressure estimation method based on multiple physiological parameters.

Area of Science:

• Biomedical Engineering
• Physiological Monitoring
• Machine Learning in Healthcare

Background:

• Noninvasive blood pressure (BP) measurement technologies face challenges with accuracy, calibration frequency, and subject limitations.
• Existing methods often overlook the influence of the sympathetic nervous system on vascular activity.

Purpose of the Study:

• To develop and validate a novel method for estimating blood pressure (BP) by integrating multiple physiological parameters.
• To assess the cross-database migration capability and robustness of the proposed BP estimation model.
• To improve the accuracy and reliability of continuous, cuffless BP monitoring.

Main Methods:

• Utilized heart rate variability (HRV), pulse transit time (PTT), and pulse wave morphology features from ECG and PPG signals.
• Employed four classic machine learning algorithms on a hybrid dataset of 3337 subjects from two databases.
• Implemented an individual calibration procedure to enhance BP estimation accuracy.

Main Results:

• The proposed algorithm achieved a Mean Absolute Error (MAE) of 10.03 mmHg for systolic BP (SBP) and 5.42 mmHg for diastolic BP (DBP).
• Individual calibration resulted in MAE ± SD of -0.16 ± 7.96 mmHg (SBP) and -0.13 ± 4.50 mmHg (DBP), meeting AAMI standards.
• The Adaboost algorithm demonstrated strong performance on the MIMIC database with MAE of 6.6 mmHg (SBP) and 3.12 mmHg (DBP).

Conclusions:

• The proposed method effectively estimates BP by considering autonomic nervous system regulation of the cardiovascular system.
• The model exhibits robustness and cross-database migration capabilities, crucial for real-world applications.
• This approach holds significant promise for advancing accurate, continuous, and cuffless BP monitoring.