Intimate partner violence (IPV) and HIV infection are pervasive and intersecting epidemics, which pose significant threats to women's survival worldwide. In a 10-country survey of women, lifetime prevalence of physical or sexual IPV ranged from 15% to 71%.1 In Sub-Saharan Africa, over half of HIV infections are in women, and in South Africa, young women are 3 times more likely to be infected with HIV infection than young men.2,3 The prevalence of IPV in South Africa and Zimbabwe is particularly high.4–6 Rooted in gender power imbalances within intimate partner relationships7,8 and inequitable gender norms at the societal level, women's experiences of emotional, physical, and sexual violence by a male intimate partner, as well as male perpetration of IPV, have been associated with inconsistent condom use, multiple sexual partners, sexually transmitted diseases and HIV infection.9–13 Women who experience IPV may be unsuccessful in their efforts to negotiate condoms, or be less likely to refuse sex or to suggest the use of condoms because they fear violence.7,14,15
Female-initiated methods of HIV prevention, including the female condom, diaphragms, other barrier methods, and microbicides have the potential to give women more options to protect their health when they are unable to negotiate condom use because some of these methods may be used without their partner's knowledge, or may not necessitate their partner's active cooperation. Adherence to methods with proven effectiveness (including the male and female condoms) is critical to averting HIV transmission. Optimizing adherent use of female-initiated prevention methods in HIV prevention trials is also crucial to determining the effectiveness of the method under investigation in preventing HIV infection.16 Although female-initiated methods of HIV prevention are needed to increase women's options and autonomy during sexual decision making, their real-world effectiveness may be compromised for women experiencing IPV. Experiences of IPV may also pose challenges to women's consistent use of female-initiated methods of HIV prevention.17
Most studies linking IPV to inconsistent condom use have been cross sectional. A few studies have explored longitudinal relationships of IPV to inconsistent condom use in the United States18 and incident HIV infection in South Africa19 and Uganda.20 In the context of HIV prevention trials, which require sustained high levels of adherence to study products and longer term follow-up (ie, 12–36 months), it is important to identify factors contributing to and impeding sustained adherence. Indeed, greater adherence has also been associated with greater effectiveness of oral PrEP and microbicides21,22 in preventing HIV transmission. The inability of several prevention trials to establish efficacy or effectiveness has been attributed, in part, to lower-than-anticipated adherence.23–25 We therefore examined the relationship of longitudinal patterns of IPV to condom and diaphragm self-reported nonadherence among women participating in the Methods for Improving Reproductive Health in Africa (MIRA) study. Understanding whether women who fear or experience IPV can consistently use the diaphragm as a potential female-initiated HIV prevention method, or whether women who use the diaphragm experience greater IPV is critical for evaluating its overall effectiveness and public health benefit. To address gaps in prior studies,18–20 the present study sought to account for the time varying nature of exposure to IPV, examined multiple forms of IPV including fear of violence, and investigated its relationship to diaphragm and condom nonadherence in a large sample of women in Southern Africa.
The MIRA trial was an open-label multisite randomized controlled trial of the diaphragm and gel for prevention of heterosexual HIV acquisition, which enrolled and followed 5039 sexually active, 18–49 year-old, HIV-negative women from 2003–2006 at 5 clinics in Johannesburg (Soweto) and Durban (Botha's Hill and Umkomaas), South Africa and Harare (Epworth and Chitungwiza), Zimbabwe. Women were randomly assigned in a 1:1 ratio to the intervention (diaphragm, lubricant gel, and condoms) arm or control (condoms-only) arm. Participants were followed up at quarterly clinic visits for up to 24 months and received product adherence and risk reduction counseling, free male condoms, HIV/sexually transmitted infection testing, and treatment of curable sexually transmitted infections at each clinic visit. The methods and results of the MIRA trial are described in detail elsewhere.26,27 The study protocol was reviewed and approved by Institutional Review Boards at the participating sites and at the University of California, San Francisco. Counseling staff at each site were equipped with referrals to local support organizations that addressed IPV.
Women completed an audio computer-assisted survey interview (ACASI) in their native language at their baseline visit and at each quarterly study visit. The ACASI collected information on sexual behavior and current and previous use of the diaphragm, gel, and condoms. Study investigators added questions on intimate partner violence to the ACASI interview 1 year after study enrollment began, and IPV was assessed at women's baseline, 12 months, and exit visits thereafter.
Of the 5039 women in the MIRA trial, 4505 women (2244 in the intervention arm and 2261 in the control arm) had available data on IPV during a baseline, 12 months, or exit visit, and condom use data at 1 or more follow-up visits (n = 9547 person-visits). Visit intervals after seroconversion for women who acquired HIV were censored from this analysis because the team expected that this would affect their subsequent product use. Exit visits occurred between 12 and 24 months of participation. When an exit visit occurred at 12 months, it was classified in this analysis as a 12-month visit rather than an exit visit.
Intimate Partner Violence
We measured 4 forms of recent fear or experience of IPV using questions adapted from a diaphragm acceptability study in Zimbabwe28 and informed by a gender-based violence study in South Africa.8 Questions referred to IPV by the woman's “regular partner” (defined as “the person you had sex with most often. This may be your husband, your boyfriend or your casual partner.”) during the 3 months before the interview. Fear of violence was defined as a “yes” response to at least 1 of the following 2 questions: “In the last 3 months have you ever been afraid that your regular partner might shout or scream at you?” or “In the last 3 months have you ever been afraid that your regular partner might shove, hit, slap, kick or otherwise physically harm you?” We defined emotional abuse as a “yes” response to the question “In the last 3 months has your regular partner emotionally or verbally hurt you in some way, such as insult you, yell at you, humiliate or swear at you?” We defined forced sex as responding “yes” to “In the last 3 months, has your regular partner either physically or verbally forced you to have sex?” And we defined physical violence as a “yes” response to at least 1 of the following 2 questions: “In the last 3 months, has your regular partner shoved, hit, slapped, kicked or otherwise physically hurt you?” and “Has your regular partner used, or threatened you with a weapon, such as a gun or knife?” We classified women as experiencing IPV, defined as “Any fear or experience of intimate partner violence” if they reported experiencing at least 1 form of violence at that study visit and no violence if they did not report experiencing any form of violence by an intimate partner.
Condom Adherence and Diaphragm Adherence
Self-reported adherence information on the 2 outcomes, condom nonadherence and diaphragm nonadherence, was collected via ACASI at each visit for the previous 3 months. To be consistent with an earlier MIRA study,29 condom and diaphragm nonadherence were individually defined as did not always use the product since the last visit or did not use the product at the most recent sexual intercourse. Because women were given the diaphragm at enrollment, and IPV was measured at only baseline, 12 months, and exit, only diaphragm nonadherence at the 12 month and exit visits was investigated in this analysis.
We examined the association of baseline covariates that were associated with IPV and condom and diaphragm nonadherence in previous studies and were potential confounders of the relationship between IPV and nonadherence. Covariates of interest covered women's sociodemographic and study-related characteristics (age, site, study arm, educational attainment, whether the woman earned an income, marital status, cohabitation with her partner), women's sexual risk behaviors (number of recent sex partners, receiving money in exchange for sex), and women's report of male partner characteristics (knowledge or suspicion that the male partner had other sex partners, age difference with the partner, partner being away from home more than 1 month of the year, partner alcohol use before sex, partner HIV status, and partner employment). Covariates associated with both the outcome and the violence indicator at a level of P ≤ 0.1 were included in multivariable regression models.
We conducted all data analyses using SAS 9.1 (Cary, NC). First, to describe the sample and identify correlates of IPV, baseline characteristics were compared between women who reported IPV at least once during the study period versus those who did not report experiencing IPV, and differences were tested using χ2 tests for categorical variables. We then calculated the prevalence of each form of IPV and any IPV at each visit, summarized it in a plot, and tested for differences in the prevalence of IPV across visits using χ2 tests. Next, we calculated frequencies of condom and diaphragm nonadherence at each visit and tested differences in the frequency of condom and diaphragm adherence across visits using χ2 tests. Because the prevalence of condom nonadherence at each visit differed by arm after enrollment, all analyses and models of condom nonadherence were stratified and presented by arm.
The longitudinal relationship of IPV exposure to (1) condom nonadherence and (2) diaphragm nonadherence was modeled using Generalized Estimating Equations based on binomially distributed data and a logit link. The working correlation was an exchangeable correlation structure, which assumed equal correlations between all pairs of observations from the same subject. The unadjusted models included categorical time (visit), a time-varying dichotomous IPV indicator, and an IPV × time interaction. In cases where the P value for the interaction of IPV and time was >0.2, the interaction term was dropped and models included only time and IPV. Multivariable models adjusted for the baseline covariates that were identified as potential confounders. Results of the models are reported as odds ratios with 95% confidence intervals (CIs). Separate unadjusted and adjusted models were fit for each type of violence as well as the “Any IPV” indicator and an indicator of “Any physical or sexual IPV.”
The multivariable models that used the “Any IPV” indicator for the IPV exposure were then used to generate and plot estimated probabilities of nonadherence for combinations of violence and visit.
Finally, we investigated the association between the pattern of IPV exposure from enrollment to the 12-month visit and condom nonadherence and diaphragm nonadherence at the 12-month visit, by restricting analyses to the 1924 participants who answered questions about IPV at both the baseline and 12-month visits (950 in the diaphragm and gel arm and 974 in the condom arm). Condom nonadherence in each arm and diaphragm nonadherence at the 12-month visit were modeled using logistic regression. The predictor was the individual's pattern of IPV exposure (any vs. none) from baseline to month 12. These categories were Persisting (Yes at baseline and month 12), Incident (No at baseline and Yes at month 12), Remitting (Yes at baseline and No at month 12), and None (No at baseline and month 12), the reference category. Multivariable (adjusted) logistic regression models adjusted for baseline covariates that were associated with each violence pattern and the nonadherence outcomes at 12 months at P < 0.1.
Overall, 52% of the 4505 women were from the Harare site, 30% from the Durban site, and 18% from the Johannesburg site. Nearly half (44%) had completed high school. Most women were married (60.7%), lived with a regular male partner (69.1%) and had 1 sex partner over the past 3 months (91.9%) (Table 1).
Prevalence and Correlates of IPV
Among the 4505 women, 757 (16.8%) provided information on IPV at 1 visit, 2465 (54.7%) at 2 visits, and 1283 (28.5%) at 3 visits. Over half of the women (54.6%) reported IPV in the 3 months before enrollment or before at least 1 follow-up visit (Table 1). Specifically, 1841 (40.8%) reported fearing violence by their male partner, 1730 (38.4%) reported that their male partner emotionally abused them, 729 (16.2%) reported that their male partner had physically assaulted them, and 668 (14.8%) reported that their male partner had forced them to have sex in the 3 months before the interview (data not shown in tables).
In bivariate analyses, women who reported fearing or experiencing IPV versus those reporting no IPV were more frequently from the Johannesburg site and less frequently from the Durban site, more likely to have earned an income, to have had more than 1 sex partner in the past 3 months, and to have had sex in exchange for money or drugs (Table 1). Women fearing or experiencing IPV more frequently had a male partner who was away from home at least 1 month of the year, who they knew or suspected had sex with other partners, who had sex under the influence of alcohol or drugs, and who was HIV-positive or of unknown HIV status. Age group, education, marital status, living with the male partner, HIV seroconversion, and partner employment were not associated with reports of IPV before enrollment or during follow-up.
The proportion of women reporting any IPV, overall and by type of IPV was highest at the baseline visit, and was lower at subsequent visits (Fig. 1). For example, the proportion of women who feared that their partner would be violent toward them in the past 3 months declined from 36% at the baseline visit, to 24% at month 12, and 23% at exit (P < 0.0001). The prevalence of physical violence (P < 0.0001) and forced sex (P = 0.03) differed modestly over time (Fig. 1).
Prevalence of Condom and Diaphragm Nonadherence
Condom nonadherence was reported at 2339/4797 (48.8%) of person visits among women in the control arm and 3040/4750 (64%) of person visits among women in the intervention arm. The proportion reporting condom nonadherence was similar in the 2 arms at baseline and diverged during follow-up as previously reported23; the proportion nonadherent decreased from 69.7% at baseline to 40.8% at 12 months and 43.6% at exit in the control arm (P < 0.0001) and from 72.1% at baseline to 62.7% at 12 months and 60.2% at exit in the intervention arm (P < 0.0001) (data not shown in tables).
Women reported diaphragm nonadherence at 1974/3599 (54.8%) of the intervention arm 12 month or exit visits (baseline visit not included). The proportion reporting diaphragm nonadherence was similar at the 12-month (54.3%) and exit (55.3%) visits (P = 0.771) (data not shown in tables).
Intimate Partner Violence and Condom and Diaphragm Nonadherence: Unadjusted and Adjusted Results
Table 2 presents unadjusted and multivariable (adjusted) models of IPV and condom nonadherence by arm, and diaphragm nonadherence, with separate models for each form of IPV. In unadjusted and adjusted models, IPV predicted higher odds of condom nonadherence over the trial period in both arms and diaphragm nonadherence in the intervention arm (Table 2). In adjusted models, the association between IPV and condom nonadherence [adjusted odds ratio AOR: 1.41, 95% CI: 1.24 to 1.61 (control arm) and AOR: 1.47, 95% CI: 1.28 to 1.69, (intervention arm)] and with diaphragm nonadherence remained (AOR: 1.24, 95% CI: 1.06 to 1.45) adjusting for age, study sites, number of sex partners, and knowledge of male partner infidelity.
Figures 2A–C show the association between IPV and the probability of condom (Figs. 2A, B) and diaphragm (Fig. 2C) nonadherence at the 3 study time points, as estimated by the multivariable models. The estimated probability of condom nonadherence declined over time in both arms, and it was higher when women reported IPV than when women reported no IPV, controlling for age, site, number of sex partners, and partner infidelity (Figs. 2A–C). Among women in the control arm, the effect of IPV on reported condom nonadherence differed over time (P value for IPV × visit interaction = 0.0008); there was no association between IPV and condom nonadherence at baseline and a positive association at 12 months and at exit. Although the probability of diaphragm nonadherence did not seem to change from the 12-month to exit visit, it was higher when women reported IPV, relative to when women did not report IPV (Fig. 2C).
IPV Pattern and Condom and Diaphragm Nonadherence: Unadjusted and Adjusted Results
Among the 1924 women who answered questions about IPV at both baseline and 12-month visits, 23.9% reported persisting IPV, 10.4% reported incident IPV, 25.2% reported remitting IPV, and 40.4% reported no IPV at both visits (Table 3). In multivariable models, women in the control arm who experienced persisting IPV (AOR: 2.2, 95% CI: 1.54 to 3.1) and incident IPV (AOR: 1.69, 95% CI: 1.08 to 2.6) had higher odds of reporting condom nonadherence at month 12 compared with women with no IPV before both their baseline and month 12 visits. Remitting IPV was associated with condom nonadherence at follow-up in unadjusted models but not in adjusted models. In the intervention arm, only women experiencing persisting IPV (AOR: 1.53, 95% CI: 1.06 to 2.2) had higher odds of condom nonadherence relative to women with no IPV before both visits. Women experiencing persisting IPV (AOR: 2.0, 95% CI: 1.39 to 2.9) and remitting IPV (AOR: 1.51, 95% CI: 1.07 to 2.1) had higher odds of diaphragm nonadherence at 12 months relative to women with no IPV at both visits (Table 3).
Women in the MIRA trial reported high rates of recent IPV and the association between IPV and diaphragm and condom nonadherence persisted over the trial period. Modeling the effects of each form of IPV on nonadherence yielded similar results. The similar odds ratio estimates for different types of IPV underscore that IPV and not just 1 particular form of IPV is associated with condom and diaphragm nonadherence. These results suggest that IPV is likely to impede adherence to HIV prevention interventions, even those that are specifically designed to give women greater control over protecting their sexual health.
The prevalence of various forms of recent IPV reported by women in the MIRA trial was high and similar to that found in other studies with women in South Africa and Zimbabwe. For example, a DHS survey in Zimbabwe found 25% of women reported physical violence and 12% forced sex in a current relationship.6 We also observed a drop after the baseline visit in prevalence of some forms of IPV, particularly fear of violence and verbal abuse, which may have been attributable to several potential factors. It may be that once women began to participate in the trial, they were less likely to report intimate partner violence out of a desire to provide socially desirable responses. ACASI was used to collect IPV data to minimize social desirability bias. Alternatively, women may have felt a sense of safety and support as a result of their participation in the trial and their interactions with staff that reduced their fear of and threats by a male partner. A third possible explanation is that women who experienced IPV were less likely to be retained in the study and were lost to follow-up or withdrew without completing a closing visit, however, we did not observe differences by IPV exposure in the proportion completing a closing visit.
This longitudinal study found that when women reported fearing or experiencing IPV, they had greater odds of diaphragm nonadherence, relative to when women did not report IPV, consistent with an earlier study in Zimbabwe.28 Additionally, there was a tendency toward greater associations with diaphragm nonadherence for women experiencing persisting IPV and persisting emotional abuse relative to no IPV. Yet, we also found that persisting and remitting forced sex were associated with condom nonadherence but not diaphragm nonadherence. It could be that nonuse of condoms was a strategy for some women to mitigate further experiences of forced sex; the lack of an association with diaphragm nonadherence may suggest that for some women, the diaphragm was an alternative option that was acceptable to their male partner or that women were able to use covertly or discreetly, regardless of whether they had experienced forced sex. A study of sex workers in Madagascar found that women who reported experiencing IPV on requesting that their partner use a condom were more adherent than those with no IPV exposure to a gel–diaphragm combination product.29 A study of product substitution (use of the diaphragm instead of condom) in MIRA found that women who experienced IPV at baseline had nearly 2.0 greater odds of reporting using a diaphragm instead of a condom.30 It may be that there are subgroups of women who experience IPV, who are able to use the diaphragm when condom use is not possible, yet others for whom even consistent diaphragm use is not possible in the face of IPV.
There are some limitations to this study. Results from this study are not generalizable to women who did not participate in the trial. All data are self-reported, however, because both nonadherence and IPV are sensitive behaviors, it is likely that those who reported it actually had experienced it. Despite these limitations, this study has several important strengths, which include the large sample of women, the longitudinal design within a randomized trial, and the comprehensive examination of multiple types and patterns (eg, persisting, incident, and remitting) of IPV exposure in relation to both condom and diaphragm nonadherence.
Results from this study have implications for the effective prevention of HIV infection in women and for implementation of HIV prevention trials. Clinical trials should provide support and protection to women who experience IPV and address IPV as part of study counseling. In addition, as the associations we observed between IPV and condom and diaphragm nonadherence may be key pathways through which IPV may heighten HIV risk, interventions and policies that explicitly address IPV and links to HIV infection risk are urgently needed. In South Africa, structural interventions aiming to shift inequitable gender norms, roles, and expectations and improve women's economic empowerment have shown promise in reducing IPV rates, but yielded mixed results regarding HIV-related risk. For example, the IMAGE study, a community level microcredit and gender awareness intervention showed a decline in IPV prevalence31 but no effect on HIV incidence. Similarly, Stepping Stones,32 a gender-transformative intervention to build more gender equitable relationships between young men and women, showed a reduction in IPV prevalence and HSV-2 incidence but not HIV. These approaches that prevent men's perpetration of IPV and reduce HIV risk by altering inequitable gender norms that condone men's use of violence to assert power and control over women33,34 are critically important, as are interventions that promote men's positive involvement in product use.35 Research that identifies multilevel determinants34 of men's perpetration of IPV13 and evaluates interventions targeting young men and women should be a high priority.
The authors thank the women who participated in this study and the staff at the study sites.
1. Garcia-Moreno C, Jansen HA, Ellsberg M, et al.. Prevalence of intimate partner violence
: findings from the WHO multi-country study on women's health and domestic violence. Lancet. 2006;368:1260–1269.
2. Gouws E, Stanecki KA, Lyerla R, et al.. The epidemiology of HIV infection
among young people aged 15-24 years in southern Africa. AIDS. 2008;22(suppl 4):S5–S16.
3. Shisana O, Rehle T, Simbayi LC, et al.. South African National HIV Prevalence, Incidence, Behavior and Communication Survey 2008: A Turning Tide Among Teenagers? Cape Town, South Africa: HSRC Press; 2009.
4. Jewkes R, Penn-Kakana L, Levin J, et al.. Prevalence of emotional, physical and sexual abuse of women in three South African provinces. S Afr Med J. 2001;91:421–428.
5. Dunkle KL, Jewkes RK, Brown HC, et al.. Prevalence and patterns of gender
-based violence and revictimization among women attending antenatal clinics in Soweto, South Africa. Am J Epidemiol. 2004;160:230–239.
6. Hindin MJ, Kishor S, Ansara DL. Intimate Partner Violence
Among Couples in 10 DHS Countries: Predictors and Health Outcomes. DHS Analytic Studies No 18. Calverton, MD: Macro International, Inc; 2008.
7. Blanc AK. The effect of power in sexual relationships on sexual and reproductive health: an examination of the evidence. Stud Fam Plann. 2001;32:189–213.
8. Jewkes R, Levin J, Penn-Kekana L. Risk factors for domestic violence: findings from a South African cross-sectional study. Soc Sci Med. 2002;55:1603–1617.
9. Maman SJ, 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.
10. Dunkle KL, Jewkes RK, Brown HC, et al.. Gender
-based violence, relationship power, and risk of HIV infection
in women attending antenatal clinics in South Africa. Lancet. 2004;363:1415–1421.
11. Speizer IS, Pettifor A, Cummings S, et al.. Sexual violence and reproductive health outcomes among South African female youths: a contextual analysis. Am J Public Health. 2009;99(suppl 2):S425–S431.
12. Frye V, Ompad D, Chan C, et al.. Intimate partner violence
perpetration and condom use
-related factors: association with heterosexual men's consistent condom use
. AIDS Behav. 2011;15:153–162.
13. Decker MR, Seage G, Hemenway D, et al.. Intimate partner violence
perpetration, standard and gendered STI/HIV risk behavior, and STI/HIV diagnosis among a clinic-based sample of men. Sex Transm Dis. 2009;85:555–560.
14. Amaro H. Love, sex, and power. Considering women's realities in HIV prevention. Am Psychol. 1995;50:437–447.
15. Heise LL, Elias C. Transforming aids prevention to meet women's needs: a focus on developing countries. Soc Sci Med. 1995;40:931–943.
16. Tolley EE, Harrison PF, Goetghebeur E, et al.. Adherence
and its measurement in phase 2/3 microbicide trials. AIDS Behav. 2010;14:1124–1136.
17. Saul J, Moore J, Murphy ST, et al.. Relationship violence and women's reactions to male-and female-controlled HIV prevention methods. AIDS Behav. 2004;8:207–214.
18. El-Bassel N, Gilbert L, Wu E, et al.. HIV and intimate partner violence
among methadone-maintained women in New York City. Soc Sci Med. 2005;61:171–183.
19. 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.
20. Kouyoumdjian FG, Calzavara LM, Bondy SJ, et al.. Intimate partner violence
is associated with incident HIV infection
in women in Uganda. AIDS. 2013;27:1331–1338.
21. Grant RM, Lama JR, Anderson PL, et al.. Preexposure chemoprophylaxis for HIV prevention in men who have sex with men. N Engl J Med. 2010;363:2587–2599.
22. Abdool Karim QA, Abdool Karim SS, Frohlich JA, et al.. Effectiveness and safety of tenofovir gel
, an antiretroviral microbicide, for the prevention of HIV infection
in women. Science. 2010;329:1168–1174.
23. Institute of Medicine. Methodological Challenges in Biomedical HIV Prevention Trials. Washington, DC: The National Academies Press; 2008.
24. Weiss HA, Wasserheit JN, Barnabus RV, et al.. Persisting with prevention: the importance of adherence
for HIV prevention. Emerg Themes Epidemiol. 2008;5:8. doi:10.1186/1742-7622-5-8.
25. van der Straten A, Van Damme L, Haberer JE, et al.. Unraveling the divergent results of pre-exposure prophylaxis trials for HIV prevention. AIDS. 2012;26:F13–F19.
26. Padian NS, van der Straten A, Ramjee G, et al.. Diaphragm and lubricant gel
for prevention of HIV acquisition in southern African women: a randomized controlled trial. Lancet. 2007;370:251–261.
27. Ramjee G, van der Straten A, Chipato T, et al.. The diaphragm and lubricant gel
for prevention of cervical sexually transmitted infections: results of a randomized controlled trial. PLoS One. 2008;3:e3448.
28. van der Straten A, Kang MS, Posner SF, et al.. Predictors of diaphragm use as a potential sexually transmitted disease/HIV prevention method in Zimbabwe. Sex Transm Dis. 2005;32:64–71.
29. Turner AN, van Damme K, Jamieson D, et al.. Predictors of adherent use of diaphragms and microbicide gel
in a four-arm, randomized pilot study among female sex workers in Madagascar. Sex Transm Dis. 2009;36:249–257.
30. van der Straten A, Cheng H, Moore J, et al.. The use of the diaphragm instead of condoms in a phase III diaphragm trial. AIDS Behav. 2009;13:564–572.
31. Pronyk P, Hargreaves JR, Kim JC, et al.. Effect of a structural intervention for the prevention of intimate partner violence
and HIV in rural South Africa: a cluster randomized trial. Lancet. 2006;368:1973–1983.
32. Jewkes R, Nduna M, Levin J, et al.. Impact of Stepping Stones on incidence of HIV and HSV-2 and sexual behaviour in rural South Africa: cluster randomised controlled trial. BMJ. 2008;337:a506.
33. Dunkle KL, Jewkes R. Effective HIV programming requires gender
transformative work with men. Sex Transm Infect. 2007;83:173–174.
34. Krishnan S, Dunbar MS, Minnis AM, et al.. Poverty, gender
inequities, and women's risk for human immunodeficiency virus/AIDS. Ann N Y Acad Sci. 2008;1136:101–110.
35. Montgomery ET, van der Straten A, Chidanyika A, et al.. The importance of male partner involvement for women's acceptability and adherence
to female-initiated HIV prevention methods in Zimbabwe. AIDS Behav. 2011;15:959–969.