Back grey_arrow_rt.gif
HIV Prevalence Not Increased by War-Torn Africa?
  Conflict, forced displacement, and wide-scale rape have not increased the prevalence of HIV in sub-Saharan Africa. Furthermore, there are no data to show that refugees fleeing conflict spread HIV infection in host communities; the reverse may be the case, conclude authors of an article published in this week's issue......
UNHCR study challenges assumptions about refugees and HIV spread
GENEVA, June 29 (UNHCR) - A new UNHCR study of seven African nations challenges previous assumptions that conflict, forced displacement and widespread rape have increased the prevalence of HIV in sub-Saharan Africa and that refugees fleeing conflict spread the infection in host communities.
The study, published this week in the British medical journal The Lancet, says a survey of data on HIV prevalence in 12 refugee camps showed that nine actually had a lower prevalence of infection than surrounding host communities, while two had similar rates and only one a higher prevalence.
But the lead author, Dr. Paul Spiegel, stressed that the findings of the survey could not be universally applied to all conflicts in the world. "Every case must be examined individually and context is very important," said Dr. Spiegel, who heads UNHCR's HIV and AIDS unit. He added that the findings should not be interpreted to mean people should not worry about HIV in conflict.
Dr. Spiegel and his colleagues combined original data gathered by UNHCR with a systematic review of published work to compare prevalence of HIV infection in the Democratic Republic of the Congo, southern Sudan, Rwanda, Uganda, Sierra Leone, Somalia and Burundi.
A press release issued by The Lancet notes there is a common assumption that violence and rape fuel the HIV epidemic in countries affected by conflict, and that consequently refugees fleeing those countries have a high prevalence of HIV infection.
The UNHCR report challenges this view. "In the seven African countries, conflict appeared to keep the HIV prevalence lower than surrounding countries that did not have conflict," Dr. Spiegel said, reviewing the main findings of the report. "Since refugees come from these areas, refugees often have lower HIV prevalence than the surrounding communities, which ultimately will increase their vulnerability towards HIV," he added.
The report authors also found that despite wide-spread rape in many countries, there was no data to show that rape increased prevalence of HIV at the population level. "This does not mean that wide-scale rape may not increase the HIV prevalence given the right circumstances," Dr. Spiegel said, adding that "every single rape is a tragedy and we must provide strong care and support to rape survivors at all times."
He and his team said the dynamics of conflict and forced displacement may alter the pattern of sexual behaviour among those affected and actually reduce the transmission of HIV than would have occurred if conflict had not taken place.
The authors suggest that previous poor survey methods and biased interpretation of data might have led to the high rates of HIV infection previously reported during conflict. They point out that because data collection during conflict is fraught with difficulties and interpretation should be cautious, such estimates need to be supplemented with reliable data from after conflict. They conclude: "This study shows the need for mechanisms to provide time sensitive information on the effect of conflict on disease incidence."
Dr. Spiegel said that refugees could actually become more vulnerable to HIV infection depending upon the prevalence of surrounding communities with whom they interact. He added that strong and comprehensive prevention programmes needed to be set up to protect refugees against HIV when they were in such vulnerable situations.
Furthermore, he emphasized that the survey also showed that more attention needed to be paid to the post-conflict period where accessibility and mobility of populations recovering from war may create an environment conducive to the spread of HIV.
The UNHCR official said the report was aimed in part at policymakers and the media, "to not always blame refugees for everything." Dr. Spiegel said it would also benefit "actors in the field" and ensure that UNHCR and its partners reintensify prevention efforts towards refugees and strategize about HIV programmes in countries emerging from conflict.
there are insufficient data to support assertions that conflict and forced displacement increase HIV prevalence challenges the conventional wisdom.
Comprehensive response to rape needed in conflict settings

Rachel Jewkes
Gender and Health Research Unit, Medical Research Council, Pretoria 0001, South Africa
Mass rape and sexual torture of women and girls by men in conflict areas is one of the most gruesome legacies of the 20th century. Nowhere has this been more evident than in the Democratic Republic of the Congo, where thousands have been gang raped with such violence as to cause vaginal fistulae.1 War is also usually associated with great upheaval of populations, and many women and girls are forced to have sex in return for safe passage, food, shelter, and other resources.2 Peacekeepers and aid workers are among the exploiters.1 Refugee camps are often unsafe, and rape occurs.3 Furthermore, large-scale movements of people provide opportunities for new consensual sexual encounters with concomitant spread of sexually transmitted infections. War in Africa is generally assumed to be one of the factors fuelling the HIV epidemic.
In today's Lancet, Paul Spiegel and colleagues draw together a substantial amount of data for prevalence of HIV infection in conflict and refugee settings to empirically test the assumption about HIV spread in conflict areas.4 Their conclusion that there are insufficient data to support assertions that conflict and forced displacement increase HIV prevalence challenges the conventional wisdom. Any review of secondary data leaves unanswered questions: what the prevalence of HIV infection in displaced populations would have been if there had not been conflict and displacement is unknown. Antenatal surveillance estimates of HIV infection are inevitably more likely to be available for long-term refugee populations. The prevalence of HIV infection in people who spend little time in camps, or those who are internally displaced and do not enter a camp, is unavailable, as are data for prevalence in men in these settings. Furthermore, some of the assessments referenced by Spiegel were made several years after the conflict, and the lower than expected prevalence of HIV infection could have reflected an effect of conflict on disease progression and death. Nonetheless, the large number of countries from which data are available, and the fact that there are other countries-including Angola and Mozambique-that have much the same patterns as those in Spiegel's review, affirms the general validity of Spiegel's argument.
Spiegel and colleagues' conclusions are a salient reminder of the value of interrogating assumptions about the HIV epidemic that are often taken as common knowledge. They remind us of the need for nuance in our understanding of the dynamics of sexual behaviour and HIV transmission. To become locked into a discourse that enumerates spiralling sexual risk, without consideration of the dynamics of conflict that might reduce consensual sexual activity, is easy. Many areas of conflict have night-time curfews or other restrictions on movement, which greatly reduce exposure of the civilian population to the contexts in which they could encounter new sexual partners and engage in risky activity after an evening's drinking. Conflict and displacement often splits families, and refugee settings may have disproportionate numbers of women and girls, which could have a net effect of reducing sexual activity. Spiegel explains why widespread brutal rape may not translate into an increased prevalence of HIV infection at a population level. Another critical factor is that many of the men who rape are young, live in rural settings, and would tend to have a low prevalence of HIV infection.
The international community is increasingly recognising the need to provide care for victims after rape in settings of conflict and displacement, and to plan to prevent rape.5 Spiegel and co-workers findings' provide a further reminder of the need for rape services to be situated within comprehensive sexual and reproductive health-care services, rather than overly emphasising prevention of HIV infection. These services are notoriously deficient in refugee settings.1 Health services after rape need to be responsive to the needs of survivors, which inevitably encompass prevention, or termination, of pregnancy; treatment for sexually transmitted infections; psychological support; and treatment for injuries, including reconstructive surgery. Even the guidelines for gender-based violence in humanitarian settings5 fall short of advocating for a comprehensive rights-based package of care for women after rape, because they fail to address the need for the provision of safe abortion.
What is the effect of conflict in the long term? Experiences in South Africa and Mozambique suggest that there could be substantial risks of HIV spread during the phase of reconstruction after conflict, because usual sexual activity is resumed in a context of economic recovery, infrastructure development, and renewed population mobility. After conflict might also be a time of particular further vulnerability to sexual and intimate partner violence.1,6 The long-term effect of conflict on children is rarely discussed. Various forms of exposure to trauma in childhood might result in increased use of violence, including sexual violence, in later years as well as sexual risk-taking.7 This exposure is a long-term potential effect of conflict on violence and HIV, and one that has yet to take its rightful place as a priority for research and intervention.
I declare that I have no conflict of interest
"Prevalence of HIV infection in conflict-affected and displaced people in seven sub-Saharan African countries: a systematic review"
The Lancet June 30, 2007; 369:2187-2195
Dr Paul B Spiegel MD a , Anne Rygaard Bennedsen BSc b, Johanna Claass MD a, Laurie Bruns MA a, Njogu Patterson MD a, Dieudonne Yiweza MD a and Marian Schilperoord MA a a. UN High Commissioner for Refugees, Geneva, Switzerland b. University of Copenhagen, Copenhagen, Denmark
There is a common belief that conflict fuels the HIV/AIDS epidemic, and consequently, refugees and internally displaced people fleeing humanitarian emergencies have a high prevalence of HIV infection.1-6 However, this assumption has been questioned.7-9 Much of the research behind these claims has not been rigorously assessed and seems to have ignored elements during conflict that might reduce transmission. A notion is emerging about how the HIV epidemic is affected by conflict and forced displacement that depends on the interaction of several complex and commonly countervailing factors.
Sub-Saharan Africa is disproportionately affected by both conflict and the HIV pandemic. Only during the past few years have reports with sufficient data for prevalence of HIV infection become available to make comparisons within and between populations affected by conflict and displacement. These new data allow for a more thorough and scientific assessment of the subject. We asked whether there is evidence to show that conflict increases HIV transmission and whether refugees fleeing conflict have a higher prevalence of infection than do surrounding host populations. We compared prevalence of HIV infection in populations directly affected by conflict with that in populations not directly affected by conflict but located nearest to the conflict, and compared prevalence in refugees and in the nearest surrounding host communities.
Background: Violence and rape are believed to fuel the HIV epidemic in countries affected by conflict. We compared HIV prevalence in populations directly affected by conflict with that in those not directly affected and in refugees versus the nearest surrounding host communities in sub-Saharan African countries.
Methods: Seven countries affected by conflict (Democratic Republic of Congo, southern Sudan, Rwanda, Uganda, Sierra Leone, Somalia, and Burundi) were chosen since HIV prevalence surveys within the past 5 years had been done and data, including original antenatal-care sentinel surveillance data, were available. We did a systematic and comprehensive literature search using Medline and Embase. Only articles and reports that contained original data for prevalence of HIV infection were included. All survey reports were independently evaluated by two epidemiologists to assess internationally accepted guidelines for HIV sentinel surveillance and population-based surveys. Whenever possible, data from the nearest antenatal care and host country sentinel site of the neighbouring countries were presented. 95% CIs were provided when available.
Findings: Of the 295 articles that met our search criteria, 88 had original prevalence data and 65 had data from the seven selected countries. Data from these countries did not show an increase in prevalence of HIV infection during periods of conflict, irrespective of prevalence when conflict began. Prevalence in urban areas affected by conflict decreased in Burundi, Rwanda, and Uganda at similar rates to urban areas unaffected by conflict in their respective countries. Prevalence in conflict-affected rural areas remained low and fairly stable in these countries. Of the 12 sets of refugee camps, nine had a lower prevalence of HIV infection, two a similar prevalence, and one a higher prevalence than their respective host communities. Despite wide-scale rape in many countries, there are no data to show that rape increased prevalence of HIV infection at the population level.
Interpretation: We have shown that there is a need for mechanisms to provide time-sensitive information on the effect of conflict on incidence of HIV infection, since we found insufficient data to support the assertions that conflict, forced displacement, and wide-scale rape increase prevalence or that refugees spread HIV infection in host communities.
From original data combined with a systematic review of published work, we conclude that there is insufficient evidence that HIV transmission increases in populations affected by conflict. Furthermore, there are insufficient data to conclude that refugees fleeing conflict have a higher prevalence of HIV infection than do their surrounding host communities. In many circumstances, comparisons of HIV prevalence in both situations show the opposite result.
Data collection during conflict, although possible, is fraught with difficulties and interpretation should be cautious.7 Such estimates need to be supplemented with reliable data from after conflict. Many of the estimated rates of HIV infection during conflict were shown to be high compared with nationwide surveys undertaken immediately after conflict, as was the case for Sierra Leone and eastern DRC. Poor survey methods, restricted accessibility favouring urban areas with high prevalence, and biased interpretation of data might have led to the high rates often reported during conflict. Such results are consistent with preconceived notions that the military and wide-scale rape increase HIV transmission at a population level. This study shows the need for mechanisms to provide time-sensitive information on the effect of conflict on disease incidence.
Some evidence supports the claim that armed conflict could disseminate or seed HIV throughout a country when the epidemic is in its early stages. This dissemination might have allowed the virus to spread rapidly in areas where it might not have appeared as quickly if conflict and subsequent demobilisation had not taken place. Uganda seems to have experienced this effect in 1978-79, as did Guinea Bissau, where an armed struggle for independence from Portuguese rule occurred from 1963 to 1974.73,74 There is evidence that war in southern Sudan introduced the virus from the south to the north.31,37,38 However, despite claims to the contrary, there are no data to show that conflict increased the prevalence of HIV infection in the seven African countries studied irrespective of the magnitude of prevalence at the start of the conflicts.
A plausible explanation for results from DRC, south Sudan, Sierra Leone, and Somalia is that conflict began when the prevalence of HIV infection in those countries was low and remained fairly low throughout the conflict compared with their peaceful neighbouring countries. Mass killings, forced displacement, and hiding can lower the incidence of infections and consensual exposures, and reduce social networks in which individuals might be exposed to HIV. Refugees who fled from these countries could have had a similar level of prevalence of HIV infection to their surrounding host communities. However, since there are no data from the areas of origin and asylum at the time the refugees fled for these countries, this notion is not certain. This contention is reinforced by data showing that prevalence of HIV infection is lower in rural areas, from where most of the refugees in these countries came, than in urban areas. However, eastern DRC had a similar prevalence for both rural and urban areas, perhaps because of factors related to mining sites in rural areas. Further studies are needed to explain these data, including the addition of more rural sites in future surveys. The scarce trend data for refugees suggest that their prevalence of HIV infection increases over time towards that of the host communities. There is no evidence that refugees exacerbate the HIV epidemic in host communities.
Despite the high prevalence of HIV infection at the time of conflict in northern Uganda, Rwanda, and Burundi, prevalence decreased over time in a manner similar to areas in the country that were not affected by conflict (ie, Uganda). Apart from eastern DRC, prevalence was higher in urban than in rural areas, as is common in many parts of sub-Saharan Africa.16,19 This difference could be a major determinant of prevalence of HIV infection, irrespective of a country's state of conflict. Refugees generally come from rural areas affected by conflict, which might explain why they seem to have a lower or similar rate of HIV infection to that of host communities, even if they fled from high-prevalence countries, as was the case for Burundi.
The 1994 genocide in Rwanda resulted in enormous death and suffering. The available evidence does not suggest that population incidence of HIV infection increased as a result of the genocide, as has been previously claimed. Articles showing that the prevalence in rural populations increased after the genocide because of widespread mixing of urban and rural populations outside of Rwanda are based on biased surveys with restricted geographical coverage that over-represent periurban and semiurban areas and under-represent rural areas.44,50 In Kigali, prevalence of HIV infection has fallen since the late 1980s, whereas in areas outside Kigali it increased until the mid 1990s and then decreased. This trend is similar to that in Burundi and might represent the natural course of the epidemic in this region. The large number of deaths in Rwanda during the short time when the genocide occurred makes assessment of the incidence of HIV exposure during the genocide impossible.
Rape is often used as a weapon of war, as has been well documented in DRC,75 Rwanda,53 Sierra Leone,76 and Liberia.77 Owing to the complexity of such settings and the large number of deaths and displacements, assessment of the effect of rape on incidence of HIV infection is especially difficult, and numbers of rapes are often underestimated because of a reluctance to discuss such painful experiences.78 Although every occurrence of rape is abhorrent and could increase an individual's risk of contracting HIV infection, there are no data to show that wide-scale rape raised the overall prevalence of infection in the populations of these seven countries. Many interacting factors should be considered, including the rates of HIV infection of the rapists and rape survivors, the probability of transmission from rape, and the eventual number of survivors who become HIV-infected. Survivor infection rates should be compared with the overall population size and prevalence of infection of the country. Although estimation of the probability of HIV transmission from rape is difficult, it is probably higher than from consensual sex because of genital or rectal trauma and several assailants.79 For consensual sex, transmission risk in discordant couples varies from one in 330 to one in 3300 exposures.80 Prevalence of HIV infection in Sierra Leone remained fairly low even though rape occurred on a wide scale. Despite assertions that extensive rape in eastern DRC fuelled the HIV epidemic,1,81 we found no significant difference between the prevalence in western and eastern regions of the country.23 Although the scale of rape and its effect on incidence of HIV infection during the Rwandan genocide are unclear, the actual transmission of HIV during rape at the time of the event is only one element that should be considered. War and rape might cause the survivor to be shunned from his or her community and thus engage in high-risk sex to survive. In communities, war and rape can change societies' culture, which unless challenged directly, could be passed on to future generations (Garrett L, personal communication).
Reconstruction periods after conflict might be a more vulnerable time for HIV transmission than during conflict. As countries recover from the physical and psychological trauma of conflict, transport increases, building takes place, and people migrate to urban areas in search of work.7,82 However, data from these seven countries to accord with this assertion are scarce. Although the two population-based surveys in Sierra Leone after conflict suggest a rise in prevalence, the studies are different and cannot be directly compared. The 2002 HIV study surveyed people aged 12-49 years in 79% of the country, whereas that in 2004 surveyed those aged 15-49 years throughout the whole country. Further comparable studies over time are needed to examine this hypothesis.
Our study has several strengths and limitations. We searched extensively, contacted authors and policymakers to assess the quality of published and unpublished data, and visited all the countries discussed. We have extensive experience working on HIV infection in conflict areas. However, the study is restricted by the nature and quality of the work in displaced populations and countries affected by conflict. The nature of displacement suggests that proportions of populations who were once directly affected by conflict might not be identified by surveys. Similarly, populations that have been affected by violence might be unwilling to undergo voluntary testing, which could contribute to biased assessments. Furthermore, although abstraction of data was undertaken by one epidemiologist then confirmed by another, it was not done in duplicate by two independent reviewers with agreement expressed by k statistics.
Reliable time-trend data for prevalence of HIV infection for DRC, South Sudan, Sierra Leone, and Somalia were scarce. Although prevalence in these countries was unlikely to fall substantially during the conflict to low levels relative to that of its peaceful neighbours, this possibility cannot be ruled out. Consequently, we cannot definitively conclude that the prevalence in these countries was low when conflict began, although this was likely. Comparisons of neighbouring countries' nearest sentinel sites and the nearest surrounding host community sites with refugee sites are subject to biases from geographical considerations, urban versus rural factors, and year of measurement. Thus, comparisons should be made cautiously. Comparable HIV data are needed for refugees and their immediate surrounding populations. Furthermore, these data need to be adjusted for age to improve comparability. Finally, comparisons between prevalence of HIV infection established by antenatal-care sentinel surveillance and population-based survey data should be made carefully because, for the most part, rates are higher with antenatal-care sentinel surveillance than with population-based surveys.16,17
Past assumptions that conflict and displacement increase prevalence of HIV infection were made from a few surveys, some of questionable quality and others with biased interpretation of results. Displaced populations and those affected by conflict are clearly at risk of HIV transmission. Furthermore, to expect that incidence of HIV infection will be high in survivors of conflict and rape is understandable. However, these assumptions were not rigorously scrutinised, and there were insufficient data available to examine the wider connection between conflict and displacement in several countries-only recently have sufficient studies become available to do so. Although some conclusions from the seven African countries studied here might apply to other countries affected by conflict, every situation is unique and should be examined according to context. Generalisations should be avoided, since they probably led to the original unsubstantiated assumptions that we investigated. Furthermore, the many factors and their interplay that affect prevalence of HIV infection in refugees could be different for internally displaced people. There are insufficient data to make any conclusions about prevalence in people who are internally displaced.83 Although challenging, prospective HIV studies in displaced communities and in those affected by conflict compared with nearby host communities unaffected by conflict are needed to further elucidate the complex factors that affect HIV transmission in these populations.
  icon paper stack View Older Articles   Back to Top