Effectiveness of a multi-level intervention to reduce men's perpetration of intimate partner violence: a cluster randomised controlled trial

Background Men's perpetration of intimate partner violence (IPV) limits gains in health and wellbeing for populations globally. Largely informal, rapidly expanding peri-urban settlements, with limited basic services such as electricity, have high prevalence rates of IPV. Evidence on how to reduce men's perpetration, change social norms and patriarchal attitudes within these settings is limited. Our cluster randomised controlled trial aimed to determine the effectiveness of the Sonke CHANGE intervention in reducing use of sexual and/or physical IPV and severity of perpetration by men aged 18-40 years over 2 years. Methodology The theory-based intervention delivered activities to bolster community action, including door-to-door discussions, workshops, drawing on the CHANGE curriculum, and deploying community action teams over 18 months. In 2016 and 2018, we collected data from a cohort of men, recruited from 18 clusters; nine were randomised to receive the intervention, while the nine control clusters received no intervention. A self-administered questionnaire, using audio-computer assisted software, asked about sociodemographics, gender attitudes, mental health, and the use and severity of IPV. We conducted an intention-to-treat analysis at the cluster level comparing the expected risk to observed risk of using IPV while controlling for baseline characteristics. A secondary analysis used latent classes (LCA) of men to see whether there were differential effects of the intervention for subgroups of men. Results Of 2406 men recruited, 1458 (63%) were followed to 2 years. Overall, we saw a reduction in men's reports of physical, sexual and severe IPV from baseline to endpoint (40.2% to 25.4%, 31.8% to 15.8%, and 33.4% to 18.2%, respectively). Intention-to-treat analysis showed no measurable differences between intervention and control clusters for primary IPV outcomes. Difference in the cluster-level proportion of physical IPV perpetration was 0.002 (95% confidence interval [CI] - 0.07 to 0.08). Similarly, differences between arms for sexual IPV was 0.01 (95% CI - 0.04 to 0.06), while severe IPV followed a similar pattern (Diff = 0.01; 95% CI - 0.05 to 0.07). A secondary analysis using LCA suggests that among the men living in intervention communities, there was a greater reduction in IPV among less violent and more law abiding men than among more highly violent men, although the differences did not reach statistical significance. Conclusion The intervention, when implemented in a peri-urban settlement, had limited effect in reducing IPV perpetrated by male residents. Further analysis showed it was unable to transform entrenched gender attitudes and use of IPV by those men who use the most violence, but the intervention showed promise for men who use violence less.