Sustainable Household Energy Adoption in Rwanda (SHEAR)

Investigators will conduct a randomized controlled trial substituting traditional forms of household energy (biomass for cooking and kerosene for lighting) with liquified petroleum gas (LPG) stoves/fuel and solar power (in areas marker for solar grids by the Government of Rwanda) in rural Rwanda. Eligible households (n=650) using traditional forms of energy will be recruited from eastern Rwanda. In each household, investigators will recruit either one adult female and one adult male or just one adult, and one child (aged 8-15 years). Following baseline health and exposure assessment, a randomized treatment arm (n=250 homes) will receive a full subsidy for LPG fuel and solar power, a control arm (n=250 homes) will continue to use traditional energy, and a random-subsidy arm (n=150 homes) will be randomized to a discounted price (at baseline and every 6-months) for solar and LPG in a pay-as-you-go (PAYG) service model (i.e., pre-pay a desired amount through mobile money). The random subsidy arm will then decide whether or not to cook with the discounted LPG or their traditional stove. Participants will be followed for 3 years with repeated measures of household air pollution (HAP) exposure (48-hour fine particulate matter (PM2.5) and black carbon (BC)), energy use, health, and preferences. Primary health endpoints include blood pressure (BP) in adults and lung-function growth in children; secondary endpoints include BP in children and lung-function change in adults. To complement the trial, the random-subsidy arm will generate policy-relevant information on causal relationships between energy costs, solar and LPG usage, and HAP exposures. The investigators propose 3 aims: Aim 1. In an ITT framework, investigators will evaluate the effect of a household energy intervention on exposure to HAP and indicators of morbidity separately among Rwandan women, men, and children within n=500 households. Aim 2. Using exposure-response modeling, investigators will characterize associations between exposure to HAP and indicators of morbidity separately among Rwandan women, men, and children within n=650 households (500 trial households + 150 random subsidy households). Aim 3. Using a random-encouragement design, investigators will investigate causal relationships between randomized energy costs, measured energy usage, and estimated exposure to HAP among n=400 households (150 random subsidy households + 250 full subsidy treatment arm households).