Widespread rape does not appear to directly increase overall HIV prevalence in conflict-affected countries

A new mathematical model indicates that widespread rape does not directly increase HIV prevalence at the population level, according to a study presented at the International AIDS Conference in Mexico this week.

For seven countries, the model predicts that in an “extreme” situation where 15% of a country’s female population were raped, where assailants had eight times the national HIV prevalence and where the HIV transmission rate was four times the usual rate for heterosexual intercourse, absolute HIV population prevalence would only increase by 0.023%.

The findings seem to contradict perceptions that rape on a broad scale, particularly when used as a weapon of war, leads to an increase in HIV transmission. I spoke with Paul Spiegel, Chief of Public Health and HIV at UNHCR and one of the authors.

Spiegel expressed concern that the results will be misinterpreted or misused to deny that rape and HIV are connected, or to direct funding away from addressing the important human rights problem of sexual violence.

Sharon Jackson, for HealthDev.net

The model used the reported general HIV prevalence levels in seven African countries that have experienced conflict – Burundi, Sierra Leone, Rwanda, Democratic Republic of the Congo, Liberia, Sudan and Uganda – and added variables estimating levels of rape and HIV transmission. This created scenarios with low, medium and high levels of prevalence, rape and HIV transmission rates. Alongside the “high” situation, at the intermediate level HIV prevalence would increase by only 0.004%; at the low level it would not increase at all.

It presented scenarios where 1%, 5%, 10% and 15% of all women and girls aged 5-49 in a country were raped. Limitations include the lack of real-life information on HIV prevalence of perpetrators and transmission during rape. Prevalence was multiplied to allow for higher HIV prevalence among assailants. The male-to-female HIV transmission rate during vaginal intercourse (0.0028, or 1 in 357) and the best available estimate for transmission during rape (0.008, or 1 in 125) were applied. The higher rate was multiplied to model higher transmission, and the effects of mutiple rapes.

Advocates, service delivery organisations, and survivors of the Rwandan genocide and other conflicts in Africa are among those worried that the study will downplay the suffering and needs of thousands of women raped during conflicts who now find themselves HIV positive. It could undermine recognition of effects on communities.

Spiegel emphasises that the model refers only to overall HIV prevalence. “Its simple mathematics. Its extremely difficult to have an effect at the population level.” The study does not analyse individual or community level effects. He says, “on an individual level of course rape increases the risk that a person may get infected…we are not saying that rape isn’t important relevant to HIV prevalence.” Widespread rape “may not increase a countrys overall prevalence” but it “may increase a districts prevalence.”

The study should make no difference to programmes or services, states Spiegel. “Proper care and treatment must be provided to every [rape] survivor regardless of the epidemiological effects of HIV transmission at the population level.” Overall, “protection and the individual human rights issues outweigh” the issue of the relationship of rape to population prevalence of HIV. More broadly, he would “hate it if people interpret it as saying that the gender dynamic is not a major driver of HIV, when of course it is.”

For children, widespread rape would not show a population level effect, but “at an individual level and mental health level it’s horrific…the effects for child soldiers and children being raped are profound.” Spiegel notes information is lacking on transmission effects for children.

The crucial issue of methodology is detailed further in the draft paper, released on 29th July in Emerging Themes in Epidemiology. The study is not a cross-sectional study of actual levels of rape or changes in HIV prevalence in these countries, and it did not involve any surveys. The mathematical model solely estimated direct transmission effects, i.e. the number of women who would likely become newly infected with HIV if widespread rape were to occur. This figure was arrived at by multiplying the number of women at risk by the probability of the assailant being HIV positive, then multiplying by the probability of HIV transmission.

A stated limitation was that indirect effects were not examined. These may include physical and psychosocial trauma, pregnancy, stigma, discrimation and abandonment, which deserve attention in their own right, as well as potentially increasing future HIV risk. Spiegel encourages more research on how rape affects future HIV risk, particularly for women in low resource settings or who are not sex workers. Spiegel refers to addressing sexual violence as “something I feel passionately about”, and emphasizes that “rape will have a huge amount of other effects.”

Spiegel notes, “another limitation is that it does not look at intimate partner violence,” which is high compared to rape by strangers in conflict. He is clear that this model is not applicable. “There may be a wider sexual network than in conflict settings which may limit its applicability to non-conflict settings.” He states, “no-one really seems to be saying that rape is a prime driver in non-conflict settings, but they are saying it in conflict settings.” However, I’ve found that intimate partner violence, including rape, forced first sex, and limited access to prevention, is increasingly being recognised as driving HIV risk for women in general.

The model is essentially “a once off measure” that estimates the effect of widespread rapes occurring at one time. It would be useful to examine this over time, so as to model interacting effects, on-going conflict and the impact of serial rapes.

“Reducing stigma and discrimination was a major motivation” for the study, due to public misconceptions that refugees and women raped during conflict have higher rates of HIV than surrounding populations. Says Spiegel, “in Northern Uganda, there were claims that a huge proportion of child soldiers were HIV positive. We knew this was wrong, based on data…The key to it for us was addressing these misperceptions. Think of a woman raped in conflict and people would say, Shes HIV positive now. These presumptions are horribly discriminating.” Developing earlier survey work published in The Lancet that indicated HIV transmission did not increase during conflict, the authors hope this study will help reduce discrimination against refugees and rape survivors.

Addressing sexual violence as a human rights abuse, including the contentious issue of its relationship to HIV, Spiegel recommends “we can use other arguments that are better founded in evidence.” He concludes, “let us advocate using the best data we have.”

Comments are closed.