Continuous PPG-Based Blood Pressure Monitoring Using Multi-Linear Regression

Area of Science:

• Biomedical Engineering
• Physiological Monitoring
• Signal Processing

Background:

• Continuous blood pressure monitoring is crucial for patient management.
• Existing methods often require invasive procedures or cuff-based devices.
• Photoplethysmography (PPG) offers a non-invasive alternative for physiological signal acquisition.

Purpose of the Study:

• To develop and validate a photoplethysmography-based blood pressure monitoring algorithm (PPG-BPM).
• To estimate continuous systolic blood pressure (SBP) and diastolic blood pressure (DBP) using only PPG signals.
• To assess the algorithm's performance against invasive arterial blood pressure (ABP) measurements.

Main Methods:

• Utilized pulse wave analysis (PWA) on PPG signals from the MIMIC I database.
• Extracted morphological features from PPG signals.
• Employed a multiple linear regression (MLR) model to map PPG features to SBP and DBP values.
• Validated the algorithm on 47,153 synchronized PPG and ABP epochs.

Main Results:

• Achieved a Mean Absolute Error (MAE) of 6.10 mmHg for SBP and 4.65 mmHg for DBP compared to arterial lines.
• Demonstrated high correlation coefficients: r = 0.90 for SBP and r = 0.85 for DBP (p < .001).
• Binary classification for hypertension (SBP ≥ 130 mmHg and/or DBP ≥ 80 mmHg) yielded 79.11% sensitivity and 92.37% specificity.

Conclusions:

• The PPG-BPM algorithm provides a reliable, non-invasive method for continuous blood pressure estimation.
• The algorithm shows significant potential for integration into wearable devices for remote patient monitoring.
• Further validation across diverse populations and clinical settings is warranted.