Sexual violence and antiretroviral adherence among women of reproductive age in African population-based surveys: the moderating role of the perinatal phase

INTRODUCTION: Women face challenges in antiretroviral therapy (ART) adherence and achieving viral suppression despite progress in the expansion of HIV treatment. Evidence suggests that violence against women (VAW) is an important determinant of poor ART adherence in women living with HIV (WLH). In our study, we examine the association of sexual VAW and ART adherence among WLH and assess whether this association varies by whether women are pregnant/breastfeeding or not. METHODS: A pooled analysis was conducted among WLH from Population-Based HIV Impact Assessment cross-sectional surveys (2015-2018) from nine sub-Saharan African countries. Logistic regression was used to examine the association between lifetime sexual violence and suboptimal ART adherence (≥1 missed day in the past 30 days) among reproductive age WLH on ART, and to assess whether there was any evidence for interaction by pregnancy/breastfeeding status, after adjusting for key confounders. RESULTS: A total of 5038 WLH on ART were included. Among all included women, the prevalence of sexual violence was 15.2% (95% confidence interval [CI]: 13.3%-17.1%) and the prevalence of suboptimal ART adherence was 19.8% (95% CI: 18.1%-21.5%). Among only pregnant and breastfeeding women, the prevalence of sexual violence was 13.1% (95% CI: 9.5%-16.8%) and the prevalence of suboptimal ART adherence was 20.1% (95% CI: 15.7%-24.5%). Among all included women, there was evidence for an association between sexual violence and suboptimal ART adherence (adjusted odds ratio [aOR]: 1.69, 95% CI: 1.25-2.28). There was evidence that the association between sexual violence and ART adherence varied by pregnant/breastfeeding status (p = 0.004). Pregnant and breastfeeding women with a history of sexual violence had higher odds of suboptimal ART adherence (aOR: 4.11, 95% CI: 2.13-7.92) compared to pregnant and breastfeeding women without a history of sexual violence, while among non-pregnant and non-breastfeeding women, this association was attenuated (aOR: 1.39, 95% CI: 1.00-1.93). CONCLUSIONS: Sexual violence is associated with women's suboptimal ART adherence in sub-Saharan Africa, with a greater effect among pregnant and breastfeeding WLH. To improve women's HIV outcomes and to achieve the elimination of vertical transmission of HIV, violence prevention efforts within maternity services and HIV care and treatment should be a policy priority.