B³X: a novel efficient algorithm for accurate automated auscultatory blood pressure estimation

Area of Science:

• Biomedical Engineering
• Medical Devices
• Signal Processing

Background:

• Auscultatory blood pressure measurement is accurate but operator-dependent.
• Automated Korotkoff sound analysis methods exist but are often complex.
• Artificial intelligence (AI) has been applied to improve automated blood pressure estimation.

Purpose of the Study:

• To propose a simple methodology, B3X, for automated estimation of systolic (SBP) and diastolic (DBP) blood pressure.
• To evaluate the B3X algorithm's performance against established standards and other algorithms.
• To enable direct implementation of automated auscultatory blood pressure measurement in clinical devices.

Main Methods:

• Processing Korotkoff sound recordings from auscultatory non-invasive blood pressure (NIBP) measurements.
• Evaluating beat-by-beat changes in Korotkoff sound morphology using cross-correlation.
• Analyzing time series of cross-correlation and its derivative to determine SBP and DBP timings.
• Testing on a public database of 350 annotated measurements and evaluating against BHS, AAMI/ANSI, and ISO standards.

Main Results:

• The B3X algorithm achieved 'A' scores for SBP and DBP in the British Hypertension Society (BHS) grading system.
• The algorithm successfully passed AAMI/ANSI and International Organization for Standardization (ISO) quality standards.
• B3X outperformed two oscillometric NIBP algorithms and four AI-based algorithms in DBP estimation, with comparable SBP performance at lower computational cost.

Conclusions:

• The B3X algorithm offers high performance in automated blood pressure estimation with low computational requirements.
• Its simplicity and lack of need for parameter training facilitate direct integration into clinical blood pressure monitoring devices.
• This automated auscultatory method bridges the gap between auscultatory accuracy and oscillometric objectivity for practical clinical use.