The PURE-Air HAP Study: Modeling 48-Hour Household Concentrations & Personal Exposures to Fine Particulate Matter in Rural Communities within Eight Countries
Year: 2019
Type: conference-abstract
Abstract: TPS 722: Low and middle income countries: indoor air pollutants, Exhibition Hall, Ground floor, August 28, 2019, 3:00 PM - 4:30 PM Background/Aim: Approximately 2.5 billion individuals cook with polluting fuels (e.g. wood, animal dung, coal, crop waste). Household air pollution (HAP) from cooking with polluting fuels causes ~2 million deaths annually. While several studies have quantitatively measured HAP exposures, these have captured relatively small populations in limited geographies. Few multi-national studies have measured HAP among diverse populations. Methods: The HAP sub-study within the Prospective Urban and Rural Epidemiology (PURE)-Air) is the largest and most diverse measurement studies of household air pollution to-date. PURE-Air includes simultaneous 48-hour integrated HAP-PM2.5 measurements in ~4,000 households and personal monitoring of ~1,500 male/female participants in Bangladesh, Chile, China, Colombia, India, Pakistan, Tanzania, and Zimbabwe. Household and personal Bayesian hierarchical models, assessed via leave-one-out cross-validation, were developed to predict PM2.5 measurements from household (e.g. primary/secondary cooking fuel type, ventilation) and personal (e.g. sex, occupation) factors available from PURE-Air surveys. Results: Initial modeling of 1,623 household measurements showed more than three-fold variation in fuel- and stove-specific average 48-hour HAP-PM2.5 kitchen concentrations: gas/electric stoves–63 μg/m3 (95%CI:42-95); mud stoves(coal)–90 μg/m3 (95%CI:-52-155); chimney stoves(wood)–98 μg/m3 (95%CI:61-155); mud stoves(wood)+gas/electric stoves(secondary)–104 μg/m3 (95%CI:66-162); mud stoves(animal dung)–112 μg/m3 (95%CI:68-184); mud stoves(wood)–132 μg/m3 (95%CI:84-206); open fires(wood)–206 μg/m3 (95%CI:120-355). Modeling of 606 personal measurements revealed country-level variability between male/female exposures. For example, female-kitchen and male-kitchen exposure ratios were 0.91 and 0.86, respectively, in Jaipur, India and 0.46 and 0.29, respectively, in Moshi, Tanzania. Bayesian R2 of 0.47 and 0.50 were achieved in household and personal models, which included primary/secondary cooking fuel type, stove type, heating fuel type, presence of windows/exhaust and sex, smoking status predictor variables, respectively. Conclusions: Substantial differences in kitchen and personal PM2.5 levels were observed by household and individual characteristics. Predictive models of HAP and personal PM2.5 exposures will be applied to the entire PURE-Air cohort to assess health effects.
Source: Environmental Epidemiology
Authors: Hystad P, Shupler M, Gustafson P, Brauer M
Institutions: Oregon State University, University of British Columbia
Language: English
Subfield: Pollution
Field: Environmental Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $1,200
OpenAlex ID: W2982106266