Domestic Violence in 30 Countries in Africa

Background: Domestic violence is a prevalent global health issue that causes incredibly adverse consequences for an individual’s physical and psychological health. The rates of physical and sexual violence in developing countries are thought to be some of the highest in the world, where up to 44% of women in Sub-Saharan Africa have experienced domestic violence. However, this has not been explored systematically. We present the first study to estimate the incidence and prevalence of physical and sexual violence across Africa. Methods: Poisson meta-regression analysis on demographic health survey data from 482,670 women from 442,507 households in 30 countries across Africa. Hotspot analysis using the Getis-ORD Gi approach at the sub-regional level. Findings: Of 482,670 women, those who were divorced or separated were approximately 7 times more likely to experience physical violence and 6 times more likely to experience sexual violence compared to those who were never married (Risk Ratio: RRs:7.35, 95%CI 7.16-7.54 and 5.89, 95%CI 5.75-6.02 respectively). Likewise, wealth index and education level were inversely related to the incidence and prevalence of sexual and physical violence. Hotspots of sexual and physical violence were identified in Congo and surrounding areas. Interpretation: Interventions should be designed to address the high levels of physical and sexual violence in Congo and surrounding areas, especially in those who are less educated, have lower wealth indices, and are divorced or separated. Funding: The USAID funded the Demographic and Health Survey collection. However, the funders played no role in the conceptualization, data analysis, or writing of this manuscript. Declaration of Interests: The authors report no competing interest. Ethics Approval Statement: Each country’s ethical review committee and institutional review board of OCF International, USA reviewed and approved the DHS protocol and guideline. Informed consent was collected prior to the survey being given. Data were de-identified and are publically available upon request; therefore, the authors did not need to seek any further clearance.