Causal Inference Analysis of the Effect of Community Health Worker Deployment on Child Survival Outcomes in Rural South East Nigeria: A Natural Experiment

Community health worker (CHW) programmes are widely proposed as a strategy for reducing child mortality in rural Nigeria, yet rigorous causal evidence of their effect on child survival outcomes, free from confounding, remains elusive. This study exploits a natural experiment created by the staggered rollout of a national CHW scale-up programme across 24 rural local government areas in Enugu and Ebonyi states, South East Nigeria, between 2020 and 2023. A difference-in-differences (DiD) design with instrumental variable correction was applied using panel data from Health Management Information System (HMIS) records, National Population Commission vital registration, and programme administrative data. The instrument was LGA-level randomisation order of CHW deployment, which was orthogonal to baseline mortality trends. The primary outcome was all-cause under-five mortality rate per 1,000 live births at LGA level. Secondary outcomes included malaria-related and diarrhoea-related child mortality. DiD estimates indicated that CHW deployment was associated with a statistically significant 22.4% reduction in all-cause under-five mortality rate (beta negative 6.3 deaths per 1,000, 95% CI negative 10.2 to negative 2.4) after two years of programme maturation. Malaria-attributable mortality reductions were the largest (29.1%). Effects were largest in LGAs with pre-existing healthcare facility density below the median, suggesting additive complementarity. Instrumental variable estimates were consistent with DiD results, supporting causal interpretation. This study makes a methodological contribution to paediatric global health research by demonstrating rigorous causal inference methods applicable to non-randomised health system interventions. Keywords: community health workers, causal inference, child survival, South East Nigeria, natural experiment.