'We threw away the stones' a mixed method evaluation of a simple cookstove intervention in Malawi

Background: Air pollution exposure is responsible for a substantial burden of respiratory disease globally. Household air pollution from cooking using biomass is a major contributor to overall exposure in rural low-income settings. Previous research in Malawi has revealed how precarity and food insecurity shape individuals' daily experiences, contributing to perceptions of health. Aiming to avoid a mismatch between research intervention and local context, we introduced a simple cookstove intervention in rural Malawi, analysing change in fine particulate matter (PM 2.5) exposures, and community perceptions. Methods: Following a period of baseline ethnographic research, we distributed 'chitetezo mbaula', locally-made cookstoves, to all households (n=300) in a rural Malawian village. Evaluation incorporated village-wide participant observation and concurrent exposure monitoring using portable PM 2.5 monitors at baseline and follow-up (three months post-intervention). Qualitative data were thematically analysed. Quantitative analysis of exposure data included pre-post intervention comparisons, with datapoints divided into periods of combustion activity (almost exclusively cooking) and non-combustion periods. Findings were integrated at the interpretation stage, using a convergent design mode of synthesis. Results: Individual exposure monitoring pre- and post-cookstove intervention involved a sample of 18 participants (15 female; mean age 43). Post-intervention PM 2.5 exposures (median 9.9g/m 3 [interquartile range: 2.2-46.5]) were not significantly different to pre-intervention (11.8g/m 3 [3.8-44.4]); p=0.71. On analysis by activity, background exposures were found to be reduced post-intervention (from 8.2g/m 3 [2.5-22.0] to 4.6g/m 3 [1.0-12.6]; p=0.01). Stoves were well-liked and widely used by residents as substitutes for previous cooking methods (mainly three-stone fires). Commonly cited benefits related to fuel saving and shorter cooking times. Conclusions: The cookstove intervention had no impact on cooking-related PM 2.5 exposures. A significant reduction in background exposures may relate to reduced smouldering emissions. Uptake and continued use of the stoves was high amongst community members, who preferred using the stoves to cooking over open fires.