Estimating Sedentary Breathing Rate from Chest-Worn Accelerometry From Free-Living Data.

Area of Science:

• Physiological monitoring
• Wearable technology
• Biomedical engineering

Background:

• Wearable physiological monitors are increasingly used to assess health during physical exertion.
• Estimating breathing rate from accelerometry data presents challenges, especially during high activity.
• Existing methods often lack validation on large-scale, free-living datasets.

Purpose of the Study:

• To evaluate the accuracy of breathing rate estimation from chest-worn accelerometry data.
• To analyze breathing rate during both high activity and sedentary periods, including sleep.
• To validate a novel method using a large dataset from a non-optimized device.

Main Methods:

• Breathing rate was estimated using chest-worn accelerometry from 1,522 servicemembers over 29,189 hours.
• A qualifier was developed to identify sedentary periods (8,867 hours) suitable for breathing rate estimation.
• The method was validated against an FDA-cleared criterion sensor (Zephyr) in a laboratory setting.

Main Results:

• Breathing rate was successfully estimated for 4,179 hours (14% of total data, 47% of sedentary time), predominantly during sleep.
• The accelerometry-derived breathing rate showed good agreement with the criterion sensor.
• This study represents the largest dataset (over 60 times larger than previous studies) for free-living breathing rate analysis using accelerometry.

Conclusions:

• Chest-worn accelerometry can provide reliable breathing rate estimates, particularly during sedentary periods and sleep.
• The developed sedentary period qualifier effectively identifies suitable data for accurate breathing rate analysis.
• This approach offers a viable method for assessing respiratory patterns using non-specialized wearable devices in real-world settings.