Field Experience with a Commercial Biometric Shirt for Measuring Minute Ventilation and Potential Inhaled Dose of Air Pollutants

摘要

Real-time estimates of volumetric minute ventilation (MV) can now be made through wearable sensors, allowing field studies to extend the personal exposure approach to more accurately account for impact of physical activity and calculation of the potential inhaled dose (PID) or intake of air pollutants. The Hexoskin biometric shirt includes sensors for all three proxy methods for estimating minute ventilation: dual band respiratory inductance plethysmography (RIP), heart rate (HR) and accelerometry (ACC). Laboratory validation of the sensors against a gold standard has been reported on previously. Here we will provide our experience in an ongoing field study of biking commuters who self-deploy the shirt and additional monitors for up to six 24 hour periods over a 3 week period. Data completeness from the biometric shirt sensors is high with only 6 out of 327 attempted sessions missing data (2%); within the obtained data, the percentage with valid data are 89% for the dual band RIP sensors, 80% for the HR monitor and 88% for the ACC. Poorly fitting shirts can impact the RIP or HR sensors (or both), with poor fits occurring more frequently on females. Failure of all three sensors simultaneously is rare allowing missing data to be filled in from other working sensors based on individual level comparisons. MV values during sleep appear anomalously low for a large fraction of participants (median 5.0 L/min, IQR 3.6 - 7.0 L/min) and 0.06% of RIP derived values appear above physiological reasonable values, leading to the use of replacement values. Final cleaned MV data derived from the shirt show a much wider range, both between subjects and within subjects, than the EPA reference values, suggesting exposure misclassification can occur if relying upon reference values for calculating PID.