UC Berkeley

BACKGROUND: Early child development (ECD) programs in low-resource settings can be effective when delivered through community health workers (CHWs), but there are significant challenges when moving to scale. This analysis aimed to determine the value-added, or relative effectiveness, of CHWs and communities on ECD outcomes within a home-visiting trial and examine associations between observable characteristics of the CHW or community and value-added. METHODS: We analysed data from the four treatment arms of a cluster-randomised trial conducted in 100 communities in rural Madagascar from 2014 to 2016. CHWs (one per cluster) and enrolled children (0-12 months) were surveyed at baseline and 2 years later. Child development scores were assessed using the Ages and Stages Questionnaire-Inventory (ASQ-I) and were internally age-standardised. We determined value-added by estimating CHW/community-level fixed effects on ASQ-I Z-score trajectories (change from baseline to endline), conditional on baseline ASQ-I Z-score and child and household characteristics. We also assessed associations between value-added and observable CHW and community-level characteristics. RESULTS: We analysed data from 1456 children present at baseline and endline. CHW/community fixed effects explained 26% of ASQ-I trajectory variance and estimates ranged from -1.68 SD to 1.31 SD. CHWs who had another income-generating position were associated with a 0.54 SD (95% CI 0.22, 0.87) increase in ASQ-I Z-score from baseline to endline. Greater increases in childrens ASQ-I Z-scores were also associated with communities that had better healthcare, education and transportation infrastructure and were less geographically dispersed. CONCLUSIONS: Children gained or lost over one standard deviation of ASQ-I Z-score depending on the community and CHW where they lived. Childrens development trajectories benefitted from CHWs involved in an external income-generating activity and communities with better access to healthcare, education, and transportation. Careful consideration of the contexts in which child development interventions are implemented and potential correlates of improved CHW performance are crucial for improved outcomes.