Continuous Blood Pressure Estimation From Non-Invasive Measurements Using Support Vector Regression.

Area of Science:

• Biomedical Engineering
• Cardiovascular Health
• Machine Learning in Healthcare

Background:

• Blood pressure (BP) is a critical vital sign linked to cardiovascular diseases (CVD) and hypertension.
• Accurate, continuous, cuffless BP monitoring is essential for effective hypertension prevention, detection, and management.
• Intelligent technology is needed to analyze complex cardiovascular system dynamics for precise BP estimation.

Purpose of the Study:

• To develop a support vector regression (SVR) model for continuous estimation of systolic blood pressure (SBP) and diastolic blood pressure (DBP).
• To leverage intelligent technology for extracting and analyzing key BP features from physiological signals.
• To validate the efficacy of the proposed SVR model in a clinical setting.

Main Methods:

• Utilized a support vector regression (SVR) model trained on established features for BP estimation.
• Selected 120 patients from the MIMIC III dataset with available ECG, PPG, DBP, and SBP data.
• Employed a machine learning approach to continuously estimate SBP and DBP.

Main Results:

• Achieved an average root mean square error (RMSE) of 2.37 mmHg for SBP estimation.
• Achieved an average root mean square error (RMSE) of 4.18 mmHg for DBP estimation.
• Demonstrated the capability of the SVR model for continuous and accurate BP monitoring.

Conclusions:

• The developed SVR model provides a promising approach for accurate, continuous, and cuffless blood pressure monitoring.
• This technology can significantly aid clinicians in the prevention, detection, diagnosis, and management of hypertension.
• The study highlights the potential of intelligent systems in enhancing cardiovascular health monitoring.