Microenvironmental Apportionment Explains Demographic Differences in PM2.5 Exposure within a Rural Rwandan Biomass Burning Community

KY TANNER, Maggie L Clark, Vincent Cleveland, Egide Kalisa, Christian L'Orange, Richard Mori, Theoneste Ntakirutimana, Casey Quinn, Christian Sewor, Kellin Slater, Jessica Tryner, Rebecca Witinok-Huber, Bonnie Young, John Volckens, Colorado State University

     Abstract Number: 440
     Working Group: Aerosol Exposure

Abstract
In 2019, household air pollution from solid fuels was ranked as the 9th leading risk factor for premature mortality with 2.3 million estimated annual deaths worldwide. This estimate, however, was based on assumptions about exposure that have not been fully validated. We assessed demographic variation on personal exposures in rural communities, and behavioral factors related to this variation. This work analyzed data collected from female and male heads-of-household and children aged 8-17 years during the baseline phase of a household energy intervention study in Eastern Rwanda, where biomass is the primary means of cooking fuel. We collected one 48-hour sample of personal real-time (optical sensor, SPS30) PM2.5 concentration data, and a corresponding time-integrated (filter-based) exposure sample, for correcting the real-time data, from each of 1,258 participants in 574 households over a ~15-month period using a UPAS sampler. We developed linear mixed models, accounting for household random effects, that suggested women were exposed to 178% (CI: 163, 194%) higher PM2.5 levels compared to their men, who had a median exposure of 96.2 μg/m3, and children’s were 189% (CI: 174, 206%) higher than the men. Personal PM2.5 exposures did not vary with socio-economic status, with the wealthiest participants having similar exposures to the poorest based on an asset-weighted index. Microenvironmental classification was applied to 30-second GPS data using a novel approach combining machine learning (density-based spatial clustering) with geocoded zonal boundaries. The microenvironmental classifications matched with real-time PM2.5 observations indicated the highest exposures across demographics occurred at home during the evenings in the dry season, while the lowest happened mid-day while in transit outside the village during the wet season. Analysis suggested men spent about 20% less time in the home (the highest exposure microenvironment) than women and children, and this discrepancy explained the within-household variation in exposure.

This abstract has supplementary material available. Click PDF icon to download.