Evaluation of physiology-based models for noninvasive blood pressure estimation using pulse arrival time

Area of Science:

• Biomedical Engineering
• Physiological Monitoring
• Wearable Technology

Background:

• Noninvasive and nonocclusive blood pressure (BP) measurement is crucial for continuous monitoring.
• Wearable systems often use pulse wave velocity, but face accuracy limitations.
• Existing methods require further refinement for reliable ambulatory BP assessment.

Purpose of the Study:

• To propose and analyze a generalized, nonlinear model for BP estimation.
• To utilize pulse arrival time (PAT), pulse period (RR), and respiratory phase (resp) for improved accuracy.
• To identify optimal models and assess the influence of individual physiological signals on BP estimation error.

Main Methods:

• Development of a generalized, nonlinear physiological model for BP estimation.
• Analysis using a publicly available database.
• Model comparison via correlation, RMSE, AIC, BIC, and MDL metrics.

Main Results:

• An optimal model, outperforming existing literature, was identified for each evaluation measure.
• Analysis revealed that simple models yield significant BP estimation errors.
• Incorporating more physiological parameters improved results, but errors remained substantial.

Conclusions:

• The analyzed signals (PAT, RR, resp) contain incomplete information for precise current BP value determination.
• Further research is needed to enhance the accuracy of noninvasive BP monitoring systems.
• Current physiological models require additional parameters or novel approaches for robust BP estimation.