Effects of economic support and community dialogue on early marriage in Zambia

Objective To assess the effectiveness of economic support, alone and in combination with a community intervention, on incidence of and socioeconomic inequality in early marriage in Zambia. Methods Data come from a cluster-randomised trial in rural Zambia. Recruitment was conducted in 2016, and all girls from grade 7 in 157 selected schools were eligible to participate. Schools were randomised to either economic support, combined economic support and community dialogue, or control. Economic support consisted of cash transfers to girls and their parents, and payment of school fees for girls continuing to secondary school. The community dialogue consisted of community and youth meetings that aimed to enhance sexual and reproductive health knowledge and supportive community norms. The interventions lasted from 2016 to 2018. Hazard ratios were calculated using Cox proportional hazards regression and the Weibull parametric survival model. Equity effects were examined using stratified analysis and interaction analysis. All analyses were by intention-to-treat. Results 4922 girls assented to participate. Mean age at baseline was 14.1 years. Compared to the control, incidence of marriage before age 18 was 21% and 14% lower in the combined and economic arms, respectively, but with confidence intervals including the null (combined vs. control HR 0.79, 95% CI 0.61 – 1.04; economic vs. control HR 0.86, 95% CI 0.65 – 1.13). There was no substantial difference in incidence between the two intervention arms. The effect estimates for incidence of marriage before age 16 were similar to those for the older age group, but with wider confidence intervals due to lower incidence. Incidence was generally higher in lower socioeconomic strata, and the interventions did not reduce this inequality. Conclusion Combining economic and community support may be effective in reducing early marriage, with a smaller effect for economic support alone, but the interventions did not reduce socioeconomic inequalities.