Domestic violence against women is a worldwide phenomenon [1, 2]. In an international study, self-reported lifetime experience of domestic violence against women ranged from 15% to 71% across 10 countries . Domestic violence, and intimate partner violence (IPV) in particular, is a risk factor for many adverse physical and psychological health outcomes [2–11]. IPV is also related to behaviors which increase the risk of HIV acquisition, such as alcohol consumption [4, 12–14]; inconsistent condom use [15–18]; concurrent partnerships [13, 19–21]; and a larger number of sex partners [13, 22]. Furthermore, women with a history of IPV may be less likely to access health care opportunities, including HIV testing . The relationship between IPV and HIV risk in women is potentially important, and not yet well understood.
Several studies have found associations between various forms of IPV and HIV status. A cross-sectional survey of women at antenatal clinics in South Africa found that women’s HIV status was associated with joint exposure to both physical and sexual IPV . A prospective study of gender power inequity in South Africa indicated that physical and sexual IPV at baseline were independently associated with higher levels of subsequent HIV risk . In Rwanda, psychological IPV (but not physical or sexual IPV) and the overall IPV experience of married women were found to be significantly correlated with HIV risk . In India, married women had significantly higher HIV prevalence when exposed to both physical and sexual IPV, but not to physical IPV alone ; and a recent analysis found that abused wives had a higher HIV risk due to both a higher HIV infection rate among abusive husbands and an increased risk of HIV transmission within abusive relationships .
There are also studies that have not found a significant association between women’s IPV experience and their HIV risk. These include studies in South Africa , Tanzania , Kenya [30, 31] and a DHS (the Demographic and Health Survey)-based study of the Dominican Republic, Haiti, India, Kenya, Liberia, Malawi, Mali, Rwanda, Zambia and Zimbabwe . A prospective study of serodiscordant couples in seven East and Southern African countries (Kenya, Rwanda, Tanzania, Uganda, Botswana, South Africa and Zambia) also found no significant association between IPV and HIV seroconversion among HIV discordant couples, though identified a significant association between IPV and HIV prevalence .
It is not surprising that various studies have found differing results, as they differ in geographic setting and how the population was sampled, as well as in the variables used and the statistical approaches taken. It is also important to note that lack of statistical significance in a given context does not in itself provide evidence that an association is absent or even weak.
Kenya was chosen for this study for its relatively high HIV gender prevalence ratio (prevalence in women is around 1.9 times higher than that of men, higher than most population-based studies in Africa), and because there is evidence that IPV is considered culturally acceptable ; in particular, men and women report similar amounts of male-perpetrated domestic violence ; and intimate partners are the most common perpetrators of sexual violence . In this study, we used the Kenya DHS’s 2008-2009 data to investigate the relationship between IPV and HIV, specifically to assess for an association, to define the magnitude of the association, and to explore the role of HIV risk behaviours in this relationship. We conducted two separate analyses of the association between HIV and IPV, first controlling only for socio-demographic variables, then adding HIV risk factors to the model. Although violence against men from their intimate partners may also be a concern, it is reported much less than violence against women , p.214 , so we focus on the latter here.