Abstract
PURPOSE: Little is known about whether effectiveness of intimate partner violence prevention programming varies for women who were married as child brides, given their additional social vulnerabilities. This subanalysis sought to assess treatment heterogeneity based on child marriage status for an intervention seeking to reduce intimate partner violence. METHODS: A randomized controlled trial assessing the incremental effectiveness of gender dialogue groups in addition to group savings on changing past-year intimate partner violence was conducted in Côte d'Ivoire (2010-2012). Stratified models were constructed based on child marriage status to assess for effect modification. Analysis was restricted to married women with data on age at marriage (n = 682). RESULTS: For child brides (N = 202), there were no statistically or marginally significant decreases in physical and/or sexual violence, physical violence, or sexual violence. The odds of reporting economic abuse in the past year were lower in the intervention arm for child brides relative to control group child brides (odds ratio [OR] = .33; 95% confidence interval [CI] = .13-.85; p = .02). For nonchild brides (N = 480), women were less likely to report physical and/or sexual violence (OR = .54; 95% CI = .28-1.04; p = .06), emotional violence (OR = .44; 95% CI = .25-.77; p = .004), and economic abuse (OR = .36; 95% CI = .20-.66; p = .001) in the combined intervention arm than their group savings-only counterparts. CONCLUSIONS: Findings suggest that intervention participants with a history of child marriage may have greater difficulty benefiting from interventions that seek to reduce intimate partner violence.