Analysis of sources and sinks of indoor particulate matter reveals insights into the real-world efficacy of portable air cleaners in a randomized intervention trial

Abstract

The increasing availability and accuracy of consumer-grade air quality sensors have increased the spatiotemporal scales at which indoor exposure measurements can be conducted. Here, we develop and apply an analytical framework, based on mass balance principles and open-source signal processing tools, to characterize sources and sinks from time-resolved data in an ongoing randomized parallel-group intervention trial of portable air cleaners (PACs) with high-efficiency particulate air (HEPA) filters in homes in metropolitan Chicago, Illinois USA. Low-cost sensors measured fine particulate matter (PM_2.5) concentrations and plug load loggers measured PAC operation at high time resolution throughout the year-long evaluation in each home. We found that comparisons of simple averages of indoor concentrations can yield misleading conclusions about the real-world efficacy of PACs in this study, as mean indoor PM_2.5 concentrations are significantly higher in the active filtration group compared to the sham filtration group (19.6 vs. 18.0 μg/m³ in 55 homes analyzed to date; p < 0.001). However, by leveraging time-resolved data we found that the active filtration group experienced higher indoor PM source strengths overall (i.e., homes were not randomized by indoor PM sources). Conversely, mean indoor PM_2.5 concentrations during stable "background" periods in the absence of indoor sources were ~36 % lower in the active filtration group compared to the sham filtration group (5.6 vs. 8.2 μg/m³; p < 0.001). Median indoor PM_2.5 loss rates with active filtration were ~15 %, ~70 %, and ~130 % higher than off conditions with the PAC operating on low, medium, and high fan speeds, respectively, further verifying the efficacy of PACs operating in these homes. The approaches herein could extend to other long-term, prospective study designs and time-resolved pollutant data analyses.

Keywords: Air purifiers; COPD; Clinical trial; Indoor air pollution; Indoor exposure assessments; Veterans.