Multinational modeling of household air pollution

Cooking with polluting fuels (e.g. wood, coal, dung) generates household air pollution (HAP), which adversely impacts the environment and health of ~3.8 billion individuals worldwide, primarily in low- and middle-income countries. Large-scale household transition from polluting to cleaner cooking fuels (e.g. gas, electricity) is necessary to achieve maximum health benefits. In this dissertation, I address key research questions for understanding facilitators of the clean cooking transition and the health impacts of HAP exposure: (1) What are household, community and national determinants of ‘natural’ cooking fuel switching? (2) How do fine particulate matter (PM₂.₅) and black carbon levels from HAP vary by country and primary cooking fuel type on a global scale? (3) Which cooking environment characteristics are predictors of PM₂.₅ measurements on a multinational scale? (4) How accurately can survey data on cooking environment factors predict quantitative PM₂.₅ exposures? I evaluated drivers of ‘natural’ polluting-to-clean primary cooking fuel switching using longitudinal survey data from rural communities within nine countries (Bangladesh, Chile, Colombia, Pakistan, South Africa, Tanzania, Zimbabwe, China, India) from the Prospective Urban and Rural Epidemiology (PURE) cohort study. Community-level (e.g. travel time to closest densely populated area, population density) factors were most strongly associated with polluting-to-clean fuel switching, and the degree of association of socioeconomic factors (e.g. education, income) with primary cooking fuel switching varied by country (highest in India, lowest in China). To quantify potential health benefits associated with a global transition to cleaner cooking fuels, I measured and modeled multinational variation in PM₂.₅ exposures. The models revealed that average PM₂.₅ concentrations at PURE baseline varied four-fold among primary cooking fuel types, ranging from 47 ug/m³ (95%CI:[47,47]) (gas) to 204 ug/m³ (95%CI:[195,213]) (animal dung). Modeled average male PM₂.₅ exposures were higher than female exposures among households primarily cooking with gas and charcoal, and across all primary fuel types in Chile, Colombia and India. Only 4% of average PM₂.₅ kitchen concentrations at PURE baseline were below the WHO Interim-1 Target (35 ug/m³); 87% of these households used cleaner primary cooking fuels (gas:85%; electricity:2%). This dissertation presents quantitative exposure estimates to be used globally for policy and disease burden assessments.Medicine, Faculty ofPopulation and Public Health (SPPH), School ofGraduat