Since 1983, Rwanda is experiencing a generalized HIV epidemic.1–3 In 1986, a first survey showed an HIV prevalence of 17.8% in urban areas versus 1.3% for rural areas.1 A second household survey, conducted in 2005 [demographic and health survey (DHS)], showed that HIV prevalence was then 7.3% in urban areas versus 2.2% in rural areas.2
The 2 surveys also showed differences in HIV prevalence between women and men. In 1986, HIV prevalence among women was 15.4% versus 10.8% among men,1 whereas the 2005 DHS indicated that 3.6% of women were HIV positive versus 2.3% of men.2
A growing body of research examined the influence of gender-based dynamics within sexual relationships and male-to-female violence in driving the HIV epidemic.4–8 Various studies have linked intimate partner violence (IPV) and HIV infection among women,9–13 suggesting that women are directly infected with HIV by their abusive partners. Jewkes et al14 showed that the incidence of HIV among women who reported more than 1 episode of IPV at baseline was 9.6 per 100 person-years compared with 5.2 per 100 person-years in those who reported 1 or no episode at all (adjusted incidence rate ratio: 1.51, 1.04 to 2.21, P = 0.032).
Different mechanisms could explain the relationship between violence toward women and HIV infection among women (Fig. 1) as follows:(1) women sexual risk factors are working as intermediate between IPV and women HIV infection (IPV could lead to women's sexual risk taking and thus to HIV, alternatively women sexual risk behavior could generate violence from their male partners); (2) IPV could result from the exposure to other forms of violence such as violence in childhood; (3) Men who endorse attitudes justifying wife beating could be perpetrator of intimate violence and could be engaged in sexual risk behavior, thus exposing their female partners to HIV infection; (4) HIV-positive women could be victims of IPV because of their HIV status.
Two previous studies analyzed the association between HIV infection and IPV among Rwandan women using the 2005 DHS data as follows: Dude showed that women who experienced emotional abuse from their partners were more likely to test positive for HIV [Adjusted odds ratio (AOR) = 3.46; 95% confidence interval (CI): 1.34 to 8.78],15 whereas using the same data set, Harling et al16 showed that the AOR of HIV was 0.99 (95% CI: 0.59 to 1.67) among Rwandan women who experienced physical or sexual IPV.
The only type of violence examined in relationship with HIV infection by the 2 studies was IPV. In addition of IPV, we examine in the present study, the effect of violence from perpetrators other than spouses and the experience of violence in childhood on women's HIV risk. Furthermore, we used couple-specific data to assess the link between IPV reported by women and sexual risk factors of their spouses.
We used data from the Rwanda third DHS conducted in 2005 by Macro International and the Rwanda National Institute of Statistics.
This survey targeted a nationally representative sample of 10,272 households using 2-stage cluster sampling based on the list of the general population and housing census carried out in 2002 in Rwanda. Women aged 15–49, usual resident in the household selected, or having slept there the night before the survey; and men of every second selected household, aged 15–59, and meeting the same conditions were eligible for the survey. All men and women in half of the selected household were eligible for HIV testing. Blood samples were collected from all eligible men and women who volunteered to be tested.
Measures and Data Collection
A face-to-face interview was used to collect sociodemographic characteristics and behavior information. The questionnaires were translated from English or French to the local language (Kinyarwanda), and a pretest was organized before the survey to validate them.
The questions related to HIV/AIDS risk factors, violence experienced by women, and men's gender attitudes were of primary interest in the present study. Violence among women was assessed using an abbreviated version of the Conflict Tactics Scale.2 We considered 3 forms of violence as follows: (1) the experience of interparental violence assessed by a question asking each woman if her father had ever beaten her mother; (2) the experience of emotional violence assessed by a question asking each woman if she experienced any form of emotional violence from anyone; (3) the experience of violence perpetrated by the intimate partner (IPV) conceptualized as either physical, psychological, or sexual and control practices by the partner.
Physical IPV (called any physical IPV) was indicated by a positive answer to any one of the following questions: does/did your husband/partner push you, shake you, or throw something at you? slap you or twist your arm? punch you with his fist or hit you with something that could hurt you? kick you or drag you? try to strangle you or burn you?
Psychological IPV (called any psychological IPV) was indicated a positive answer to any one of the following questions: does/did your husband/partner ever threaten you or harm you? threaten and/or attack you with a knife, gun, or anything else? say or do something to humiliate you in front of others? spit you in the face?
Sexual IPV (called any sexual IPV) was indicated by a positive answer to any one of the following questions: does/did your husband/partner ever physically force you to have sexual intercourse with him even when you did not want to? force you to perform other sexual acts you did not want to?
The control by partner was indicated by a positive answer to any one of the following questions: is/was your husband/partner jealous or angry if you (talk/talked) to other men? he frequently (accuses/accused) you of being unfaithful? he (does/did) not permit you to meet your girlfriends? he (tries/tried) to limit your contacts with your family? he (insists/insisted) on knowing where you (are/were) at all time? he (does/did) not trust you with any money?
For each form of IPV, we also created a summary scale corresponding to the number of questionnaire items related to a specific form of violence to which a woman answered yes. Thus, physical IPV had a 6-point scale, ranging from 0 to 5; psychological IPV, a 5-point scale, ranging from 0 to 4; sexual IPV, a 3-point scale ranging from 0 to 2; and control by partner, a 7-point scale, ranging from 0 to 6.
Men's attitudes toward wife beating were indicated by a positive answer to any one of the following questions: is a husband justified in hitting or beating his wife if she goes out without telling him? if she neglects the children? if she argues with him? if she refuses to have sex with him? if she burns the food?
A trained technician collected capillary blood from finger prick on a filter paper. Testing for HIV antibodies was performed at the National Reference Laboratory in Kigali, using enzyme-linked immunosorbent assay tests.2
The analyses were limited to women and men who were legally married or cohabiting maritally. We performed separately women and men individual analyses and couple-specific analyses.
HIV prevalence was the dependent variable, whereas sociodemographic characteristics, sexual risk factors, and different forms of violence toward women were the independent variables. Women sexual risk factors considered in our analyses were age at first sexual intercourse, the number of lifetime partners, and the report of genital ulcers in the last year; variables related to condom use were not included due to a large number of missing values. We also excluded the subjects who refused to answer or answered do not know to the questions related to the independent variables. Sampling weights were applied in all analyses. Survey and Cluster SAS procedures were used in bivariate and multivariate analysis to control for the survey design effect within sampling clusters. All statistical analyses were performed using the SAS software (SAS Institute Inc, Version 9.1.3, Cary, NC).
We first examined the distribution of sociodemographic characteristics, then we estimated bivariate and AORs with 95% CIs between HIV prevalence and the independent variables. We tested for statistical significance using the Wald χ2 from the logistic regression analyses.
To analyze the associations between HIV infection and violence, we constructed 2 different sets of multivariate models as follows: those adjusted only for the sociodemographic variables associated with HIV in the bivariate analysis and those further adjusted for women sexual risk factors, as we hypothesized that women sexual risk factors could work as intermediate or as a confounding factors in the association between violence and HIV infection.
We also performed couple-specific analyses to examine the relationship between IPV among women, men's attitudes justifying wife beating, and men's sexual risk factors. We used the χ2 test for this analysis.
The analysis concerned 2715 women and 2461 men who were legally married or cohabiting and who participated in the third Rwanda DHS in 2005.
Table 1 shows the sociodemographic characteristics of both women and men. The median age was 32 years for women and 38 years for men. Women were more represented among the 15–24 years age group than men, suggesting that they get married earlier than men. The majority of both sample were living in rural areas.
Women and Men Individuals' Analysis
Between 14% and 18% of women did not answer or answered do not know to questions assessing violence. Considering all variables related to violence, HIV prevalence was higher among women who refused to answer or who answered do not know to questions related to the violence variables [ranged from 4.6% to5.7% compared with 2.8% in women answering the questions and who are included in the analyses presented below (all P values < 0.05)].
About 29.2%, 22.2%, and 12.4% of women reported having experienced physical, psychological, and sexual violence from their partners. About 52.1% reported having experienced control from their partners, and 33.3% indicated that their fathers beat their mothers.
As shown in Table 2, HIV prevalence among women increased significantly over the psychological IPV scale. In particular, women with an IPV scale value of 3 or 4 had a significantly higher HIV prevalence than those with a score from 0 to 2 (P = 0.0006) In addition, HIV prevalence was higher among women who had a value of 3 to 6 on the partner control scale, compared with those with a value from 0 to 2 (P = 0.02). In a bivariate analysis, women who reported having ever experienced any form of emotional violence and those who reported that their fathers beat their mothers were more likely to be HIV positive (P < 0.05).
In Table 3, we present 4 multivariate models showing the associations between HIV infection and violence factors after adjustment for the sociodemographic variables and then for the sociodemographic variables and women's sexual risk factors. In the models adjusted for the sociodemographic variables and women's sexual risk factors, the adjusted odds of HIV prevalence was 3.23 (95% CI: 1.30 to 8.03) among women with value of 3 to 4 on the psychological IPV scale, compared with those with a value from 0 to 2. Psychological IPV scale as a continuous variable showed significant result in the model adjusted only for the sociodemographic variables [the AOR of HIV prevalence increased by 37% for each point increase in the psychological IPV scale (footnote to Table 3)].
The partner control scale was not significantly associated to HIV as a categorical variable. However, it was near significant when considered as a continuous variable (footnote to Table 3).
The AOR for women who experienced any form of emotional violence and those who reported having experienced interparental violence were, respectively, 2.15 (95% CI: 1.14 to 4.09) and 1.87 (95% CI: 1.10 to 3.19) in the model adjusting for sociodemographic variables.
Among men, the multivariate analysis showed that HIV prevalence was higher in those aged 40–44 years compared with those aged 15–29 (AOR: 3.85; 95% CI: 1.76 to 8.47); in those living in urban areas compared with those living in rural areas (AOR: 3.75; 95% CI: 2.20 to 6.37); in those who reported more than 5 life-time sexual partners (AOR: 4.42; 95% CI: 1.97 to 9.88); and in those who reported having had a genital ulcer in the past year (AOR: 4.10; 95% CI: 1.52 to 11.09).
Table 4 shows significant links between men's sexual risk factors and IPV experienced by their wives. Sexual IPV was reported more frequently by women who lived with men who reported having had a genital ulcer in the last year. The proportion of women who reported physical IPV increased proportionally with the number of lifetime partners reported by their husbands. Finally, the results indicate that HIV prevalence was higher among women living with men who reported ever having paid for sex (P < 0.05).
We also observed a higher level of reporting any IPV among women whose male partner reported any positive attitude toward wife beating (44.2% vs. 35.7%, P = 0.002).
The findings from this study show that an important proportion of women in union experienced different types of violence as follows: emotional violence, violence between parents and IPV. Consistent with our hypothesis (Fig. 1), our couple analysis indicates an association between experience of IPV among women and a positive attitude toward this behavior among their spouses, suggesting a positive link between men's attitudes justifying wife beating and violence perpetration. Previous studies conducted among men in different settings reported similar relationship.4,17–19 Thus, men's attitudes justifying wife beating could be considered as a marker of men IPV perpetration.
Our couple analysis also indicates that the proportion of women who reported IPV was high among women living with men who reported sexual risk factors such as genital ulcers, high number of lifetime sexual partners, and having paid for sex. This observation supports findings reported by studies conducted among men showing that men perpetrating IPV are more likely involved in risky sexual behaviour.20–22
In the present study, psychological IPV was the only variable assessing partner violence that was associated with HIV. The odds of HIV prevalence among women who reported psychological IPV was reduced when we adjusted for women sexual risk factors (especially number of lifetime partners and report of genital ulcer in the past year) as presented in Table 3. After Baron and Kenny23 recommendation in testing mediation, the attenuation of the odds of HIV may indicate that women sexual risk factors were intermediate factors in the association between HIV infection and IPV. Indeed, although we observed a positive association between IPV among women and sexual risk factors of their male partners, the results also indicate a certain level of sexual risk factors among women, particularly the number of lifetime sex partners (24% of women reported having more than 1). Thus, as we present in Figure 1, women sexual risk factors could be an outcome of IPV as follows: in this case, they precisely work as intermediate factors. On the on other hand, women's sexual risk factors may be an initiating factor for partner violence24; in this case, they will work as confounding factors (that means, women's sexual risk factors are associated with HIV and IPV but are not involved in the causal pathway between IPV and HIV infection). Because this research was a cross-sectional study, the cause–effect relationship between IPV and women sexual risk factors cannot be definitively established. Then, it is difficult to specify if women sexual risk factors were mediator or confounder factors in the relationship between IPV and HIV infection among women given that the causal pathways between IPV and women sexual risk factors may be bidirectional.
Concerning the association between physical or sexual IPV and HIV, our findings are consistent with those of 2 previous studies that used the same data set15,16 that did not find association between HIV infection and these 2 types of IPV. We also found a positive association between HIV infection and psychological IPV as reported by Dude.15 Note that the author conceptualized it differently as we did [she combined 2 items (whether her current or most recent husband had ever “humiliated” her in front of others or had threatened her with harm)].
In addition to the findings from the 2 previous studies, our results show a positive association between HIV prevalence and the exposure to interparental violence. The relationship between interparental or family violence and HIV infection has not been largely addressed in the past. However, some studies reported associations between witnessing interparental violence in childhood and experiencing IPV in adult.25–27 Thus, IPV may mediate the association between witnessing interparental violence and HIV through the sexual risk factors of the abusive partner. The association may be mediated also by risky sexual behaviors among women who witnessed interparental violence.28
Our results should be interpreted in light of some limitations. The use of cross-sectional data does not permit to specify any causal inference. As the population of our study was only women and men married or cohabiting, the findings could not be extrapolated to the general population. Due to the sensitive nature of the topic related to violence and sexual factors, a potential social desirability bias could have had an impact on the findings of our study. Consequently the sexual risk factors and IPV could be underestimated. Also the association between HIV prevalence and violence factors among women could be underestimated due to the proportion of women who did not answer to the questions related to violence. Indeed HIV prevalence was higher among those who refused to answer or who answered don't know to these questions (P<0.05).
In conclusion, we have shown that violence experienced by women is associated with HIV among Rwandan women in union. Given that IPV in itself cannot cause HIV infection among women except in case of sexual assault by a HIV-positive male partner, men's abusive sexual risk factors and women's own sexual risk factors both interact as mediator factors in relationship between IPV and HIV infection among women.
We recommend to integrate gender-based violence prevention into HIV programs, targeting both men and women, that is, the voluntary counseling and testing session could be a good opportunity to deliver educational messages related to the negative effect of domestic violence and its relationship with HIV infection. This could be useful especially in PMTCT service where, at least in Rwanda, couples receive HIV pretest counseling together.
We are grateful to the staff of the Rwanda National Institute of Statistics, Rwanda Treatment and Research Aids Center, and National Reference Laboratory for their contributions to the DHS data collection and laboratory testing. We thank Mohamed Ayad, Bridgette James, and Wellington Bridgette from Macro International for technical assistance, Eric Demers of the Unité de recherche en santé des populations, Centre de recherche FRSQ du CHA universitaire de Québec, for statistical support, and Rémi St-Laurent for English editing.
1. Bizimungu C, Ntilivamunda A, Tahimana M, et al.. Nationwide community-based serological survey of HIV-1 and other human retrovirus infections in a central African country. Lancet. 1989;1:941–943.
2. Institut National de la Statistique du Rwanda (INSR) and ORC Macro. Rwanda Demographic and Health Survey 2005. Calverton, MD: INSR and ORC Macro; 2006.
3. Kayirangwa E, Hanson J, Munyakazi L, et al.. Current trends in Rwanda's HIV/AIDS epidemic. Sex Transm Infect. 2006;82:(suppl 1):i27–i31.
4. Abrahams N, Jewkes RK, Hoffman M, et al.. Sexual violence against intimate partners in Cape Town: prevalence and risk factors reported by men. Bull World Health Organ. 2004;82:330–337.
5. Dunkle K, Jewkes RK, Nduna M, et al.. Perpetration of partner violence and HIV risk behaviour among young men in the rural Eastern Cape, South Africa. AIDS. 2006;20:2107–2114.
6. Andersson N, Ho-Foster A, Mitchell S, et al.. Risk factors for domestic physical violence: national cross-sectional household surveys in eight southern African countries. BMC Womens Health. 2007;7:11.
7. Lary H, Maman S, Katebalila M, et al.. Exploring the association between HIV and violence: young people's experiences with infidelity, violence and forced sex in Dar es Salaam, Tanzania. Int Fam Plan Perspect. 2004;30:200–206.
8. Silverman JG, Decker MR, Kapur NA, et al.. Violence against wives, sexual risk and sexually transmitted infection among Bangladeshi men. Sex Transm Infect. 2007;83:211–215.
9. Campbell JC, Baty ML, Ghandour RM, et al.. The intersection of intimate partner violence against women and HIV/AIDS: a review. Int J Inj Contr Saf Promot. 2008;15:221–231.
10. Ntaganira J, Muula AS, Masaisa F, et al.. Intimate partner violence among pregnant women in Rwanda. BMC Womens Health. 2008;8:17.
11. Maman S, Mbwambo JK, Hogan NM, et al.. HIV-positive women report more lifetime partner violence: findings from a voluntary counseling and testing clinic in Dar es Salaam. Tanzania. Am J Public Health. 2002;92:1331–1337.
12. Maman S, Campbell J, Sweat MD, et al.. The intersections of HIV and violence: directions for future research and interventions. Soc Sci Med. 2000;50:459–478.
13. Silverman JG, Decker MR, Saggurti N, et al.. Intimate partner violence and HIV infection among married Indian women. JAMA. 2008;300:703–710.
14. Jewkes RK, Dunkle K, Nduna M, et al.. Intimate partner violence, relationship power inequity, and incidence of HIV infection in young women in South Africa: a cohort study. Lancet. 2010;376:41–48.
15. Dude AM. Spousal intimate partner violence is associates with HIV and other STIs among married Rwandan women. AIDS Behav. 2010;15:142–152.
16. Harling G, Msisha W, Subramanian SV. No association between HIV and intimate partner violence among women in 10 developing countries. PLoS One. 2010;5:e14257.
17. Hanson RK, Cadsky O, Harris A, et al.. Correlates of battering among 997 men: family history, adjustment, and attitudinal differences. Violence Vict. 1997;12:191–208.
18. Martin SL, Moracco KE, Garro J, et al.. Domestic violence across generations: findings from northern India. Int J Epidemiol. 2002;31:560–572.
19. Sambisa W, Angeles G, Lance PM, et al.. Physical and sexual abuse of wives in urban Bangladesh: husbands' reports. Stud Fam Plann. 2010;41:165–178.
20. Decker MR, Seage GR III, Hemenway D, et al.. Intimate partner violence perpetration, standard and gendered STI/HIV risk behaviour, and STI/HIV diagnosis among a clinic-based sample of men. Sex Transm Infect. 2009;85:555–560.
21. Kaufman MR, Shefer T, Crawford M, et al.. Gender attitudes, sexual power, HIV risk: a model for understanding HIV risk behavior of South African men. AIDS Care. 2008;20:434–441.
22. Raj A, Reed E, Welles SL, et al.. Intimate partner violence perpetration, risky sexual behavior, and STI/HIV diagnosis among heterosexual African American men. Am J Mens Health. 2008;2:291–295.
23. Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol. 1986;51:1173–1182.
24. Bauer HM, Gibson P, Hernandez M, et al.. Intimate partner violence and high-risk sexual behaviors among female patients with sexually transmitted diseases. Sex Transm Dis. 2002;29:411–416.
25. Bensley L, Van EJ, Wynkoop SK. Childhood family violence history and women's risk for intimate partner violence and poor health. Am J Prev Med. 2003;25:38–44.
26. Speizer IS. Intimate partner violence attitudes and experience among women and men in Uganda. J Interpers Violence. 2010;25:1224–1241; E-pub September 15, 2009.
27. Vung ND, Krantz G. Childhood experiences of interparental violence as a risk factor for intimate partner violence: a population-based study from northern Vietnam. J Epidemiol Community Health. 2009;6:708–714.
28. Elliott GC, Avery R, Fishman E, et al.. The encounter with family violence and risky sexual activity among young adolescent females. Violence Vict. 2002;17:569–592.
© 2012 Lippincott Williams & Wilkins, Inc.