Item

Abstract

Homestead food production (HFP) can improve nutrition through multiple pathways. Understanding their relative importance can optimize intervention design and impact. Here we used panel data on 2,612 women from a 1:1 cluster-randomized trial of 96 settlements in rural Bangladesh and conducted multiple mediation analysis to investigate the impact pathways of a 3-year HFP intervention on women’s dietary diversity. The pathways analysed fully explained the intervention’s total effect on dietary diversity score, amounting to an increase of 0.4 (95% confidence interval (CI) 0.3–0.5) food groups on a 10-point scale. Garden production accounted for 78% of the increase (β = 0.33 (95% CI 0.25–0.42)), emerging as the key component to improving dietary diversity. Nutrition knowledge accounted for 18% of the increase, while poultry production and market activity made smaller contributions. These findings can guide the design of future HFP interventions, but similar analyses are needed across a range of interventions, outcomes and settings to build a robust evidence base for improving nutrition.