Estimated confidence interval from single blood pressure measurement based on algorithmic fusion

Area of Science:

• Biomedical Engineering
• Medical Devices
• Cardiovascular Health

Background:

• Current oscillometric blood pressure devices offer single-point estimates without confidence intervals.
• Accurate blood pressure monitoring is crucial for cardiovascular health management.

Purpose of the Study:

• To present a novel methodology for obtaining confidence intervals (CIs) for systolic blood pressure (SBP) and diastolic blood pressure (DBP) from single oscillometric measurements.
• To enhance the precision of blood pressure estimates and provide a measure of uncertainty.

Main Methods:

• Utilized multiple regression to fuse SBP and DBP estimates from various algorithms.
• Employed a weighted bootstrap approach to generate pseudo-sample sets for CI determination.
• Estimated regression coefficients by relating input samples to auscultatory nurse measurements.

Main Results:

• The proposed method was validated on a dataset of 85 patients.
• Demonstrated superior blood pressure estimates compared to existing algorithms.
• Successfully provided CIs for individual blood pressure measurements.

Conclusions:

• The novel methodology yields smaller CIs than conventional methods, indicating improved precision.
• The weighted bootstrap approach reduces standard deviation, enhancing CI accuracy.
• This represents a significant advancement, potentially being the first method to offer CIs for single-sample blood pressure readings.