Callegari, Lisa MD*; Harper, Cynthia C. PhD*; van der Straten, Ariane PhD*; Kamba, Mavis NP†; Chipato, Tsungai MBCHB†; Padian, Nancy S. PhD*
WOMEN COMPRISE MORE THAN HALF of HIV-positive adults in Sub-Saharan Africa.1 Although prevention efforts have traditionally focused on high-risk sexual encounters, marital relationships are thought to be a growing and widespread mode of transmission in women.2,3 Condom use has improved in extramarital encounters, but studies show persistently low rates in marital or cohabiting relations.2,4,5
With an overall adult HIV prevalence rate estimated at 20%, Zimbabwe lies at the heart of the AIDS pandemic.6 Zimbabwean women are disproportionately affected, making up 59% of infected adults.6 According to the 2005–2006 Demographic and Health Survey for Zimbabwe, 4% of women used a condom at last sex with a spouse versus 46% with a noncohabiting partner; the low proportion of married women using condoms has remained unchanged since the previous national survey in 1999.7,8 Use of female condoms still requires male partner cooperation, and 0.0% of married women of reproductive age reported current use in the Demographic and Health Survey.8 A recent population cohort study reporting declining prevalence of HIV in Eastern Zimbabwe found increased condom use with casual partners, but not with regular partners.9
Married Zimbabwean women face multiple challenges in negotiating condom use with their partners. Gender roles and power dynamics make it difficult for women to discuss sex or HIV and for women to refuse sex without a condom; economic dependency also renders many women unable to negotiate condom use.10
With no female-controlled alternatives available, prevention efforts must continue to promote male and female condom use among married couples. At present, less is known about predictors and determinants of condom use among married couples than noncohabiting individuals in Sub-Saharan Africa, and very little is known about condom use among married women in Zimbabwe.2,10,11 One study based on data from the previous round of the Demographic and Health Survey for Zimbabwe examined predictors of condom use at last sex in married/cohabiting women. Among women who were concurrently using another contraceptive method, age was negatively associated with condom use. Among women not using another method of contraception, communication with partners and having discussed HIV with them was associated with condom use.10 HIV researchers in Zimbabwe consider the need for condom interventions to be urgent.12
Understanding which characteristics, beliefs, and behaviors are associated with successful condom uptake could provide crucial information for designing successful condom interventions among married women. We hypothesized that perceptions of HIV risk and positive attitudes about condom use and effectiveness would be associated with greater condom uptake and consistent use. This study reports the effect of a brief condom intervention, including HIV testing, on protected sex (i.e., 100% male or female condom use) in married women, and identifies factors associated with consistent protected sex at 2-month follow-up after the intervention.
Materials and Methods
Data were from the first phase of a 2-phase study designed to assess acceptability of the diaphragm as a potential female-controlled method of HIV/STD prevention. Sexually active women of reproductive age were enrolled in the study from public clinics in 2000–2001 in Harare. Exclusion criteria included pregnancy, intention to become pregnant in the next 8 months, allergy to latex condoms, and 100% condom use reported at screening.
Women were enrolled in a male and female condom intervention (Phase I), the results of which are presented here. A questionnaire was administered by an interviewer in Shona, the native language, after informed consent was completed. It included questions about demographics, reproductive history and sexual behavior, male partner and relationship factors, HIV risk perceptions, perception of condom efficacy, and condom use. HIV testing with counseling was offered at enrollment, and those who wanted testing were scheduled for a visit within 2 weeks for the HIV test. A urine pregnancy test was given and a pelvic exam.
A 30-minute one-to-one intervention based on social-cognitive models of behavior change13 and adapted to the local Zimbabwean context was given by a trained counselor. It emphasized enhancing condom self-efficacy and negotiation skills, and included education about HIV and STI transmission, HIV/STI risk assessment, safer sex options, demonstration and practice of proper use of male and female condoms. Male condoms were presented as the most effective against HIV/STDs, followed by female condoms. Women were offered male and female condoms and HIV testing. A 1-month booster session included content similar to enrollment. At 2-months, women completed a follow-up interviewer-administered questionnaire, provided a urine sample and received a pelvic exam. Women were divided into consistent and nonconsistent condom users. Ethics approval for this study was granted by the Committee on Human Research, University of California, San Francisco, the Biomedical Research and Training Institute, the Medical Research Council of Zimbabwe, and the Centers for Disease Control and Prevention Institutional Review board.
The baseline and follow-up questionnaires asked, “How often did you use a condom (male or female) when you had sex in the last 2 months?” (response categories: every time, almost every time, about the half of the time, occasionally, and never). We created a dichotomous consistent condom use variable from this item (every time vs. all other responses). Women were also asked at baseline and at follow-up whether they used any condom at last sex. These questions did not differentiate between male and female condom use. However, at follow-up we also asked how many times participants specifically used male or female condoms in the past 2 weeks. We categorized participants who used male condoms at every act as consistent male condom users in the past 2 weeks. We also coded consistent female condom users in the past 2 weeks, and finally those who used either a male or female condom at each act.
HIV Risk Perceptions and Testing.
We measured items at baseline and follow-up that we hypothesized would be important for changes in condom use including HIV risk perceptions, HIV testing, condom attitudes and negotiation. HIV risk perceptions were assessed with questions about a woman’s perception of her own and her partner’s likelihood of acquiring HIV in the next year, and the likelihood of infection now (responses: not at all, somewhat, and very). For HIV testing, participants were asked at baseline if they had ever been tested, and if so the result. Women were offered testing, and those who chose to test in the study did so at enrollment or within 2 weeks. At follow-up, women were again asked if they had been tested and for the result. This self-reported data, consistent with the laboratory results from HIV tests, were used in the analyses.
Condom Attitudes and Negotiation Skills.
Women were asked how effective they thought male condoms were in preventing both pregnancy and STD/HIV (responses: very, somewhat, and not at all). They were asked the same questions for female condoms. Condom negotiation self-efficacy was measured with a series of questions pertaining to male and female condoms: (1) “If you wanted to use a condom, how confident are you that you could convince (husband) to use a condom, even if he doesn’t want to initially?” (responses: not at all, moderately, and very); (2) “If you wanted to use a condom, and (husband) did not, how confident are you that you could refuse to have sex?”; and (3) “Over the past 2 months, how much control did you have over whether you and (husband) used condoms?” For the purpose of analyses, response categories for all questions were converted into dichotomous variables with “very” and all other responses. Participants were also asked whether the decision to use male condoms was a shared decision in the couple, or whether they made the decision or their partner.
Sociodemographic, Reproductive Health, and Relationship Measures.
The questionnaire measured baseline sociodemographic factors, including age, education, income, religion, and running water in household. Women reported their age of first sex, number of children, whether they wanted more children and primary contraceptive method. They also reported number of sex partners, whether they suspected their male partner had other partners, the partner’s alcohol and substance use during sex, and characteristics of the relationship including comfort level in talking about feelings or sex, respect, and domestic violence measures.
Descriptive statistics of baseline demographic, reproductive and relationship characteristics, as well as condom use data, are presented. The change in protected sex (i.e., consistent condom use) from baseline to follow-up was measured with McNemar’s χ2 tests. Baseline characteristics of women were examined as possible predictors of successful condom uptake, with a focus on perceptions of HIV risk and condoms. The association of these baseline characteristics with protected sex at follow-up was measured with logistic regression analyses (odds ratios presented) to assess independent baseline predictors of consistent condom use during the study period. We included in the adjusted model the baseline variables that were significantly associated with consistent condom use in the bivariate analysis at the 0.05 level, as well as controls for age, education, and use of hormonal contraception. We estimated the association of protected sex during the study with the proximal outcomes of the intervention at follow-up: measures of condom attitudes, condom negotiation skills, HIV risk perceptions, HIV testing and serostatus. Finally, we conducted logistic regression to measure the association of protected sex with the above intervention variables that were significant in the bivariate analysis, controlling for age, education, and hormonal contraception. Analyses were conducted using STATA version 8.0.
A total of 842 women were screened for entry into Phase I, with 424 meeting inclusion criteria and 405 enrolled (95.5% of the eligible women). Of these 405 women, 394 were married, and constitute our analytical sample. At the 2-month completion of Phase I, 379 (94%) women completed the follow-up questionnaire, and 369 of these women answered the item on consistent condom use. Baseline characteristics are described in Table 1. The mean age was 28 years (range 17–47 years), and the population was Shona speaking (99%). Approximately half of the women (52%) had completed secondary school and most (93%) lived in households with running water.
Overall, participants reported low-risk sexual behaviors. The mean age at first sex was 18.9 years (SD = 2.5). All women (100%) reported one partner (in past 2 months). The mean number of life partners was 1.4 (SD = 0.9). More than 99% (n = 392) had given birth, with an average of 2.5 children. Most (81.5%) relied on oral contraceptive pills for contraception. In contrast, participants reported risky sexual behaviors of their male partners. Sixty percent suspected their male partners had other sexual partners. Alcohol consumption was far more common among men, as was being under the influence of alcohol or drugs during sex. Participants reported that 40% of their male partners were under the influence at least half of the time they had sex (in the past 6 months). In comparison, less than 1% of women reported being under the influence at least half the time while having sex (in the past 6 months). Women reported feeling very respected by their husbands (86%), very comfortable talking about sex (91%), and very comfortable talking about their feelings (92%). However, 46% reported ever being afraid that their husband would hurt them. Sixteen percent reported a history of physical abuse and 16% of forced sex.
Perceptions of HIV risk among the sample were unrealistically low. Eighteen women (4.6%) reported at enrollment that it was very likely that they already had HIV. Few had been tested for HIV (n = 48; 12%), and even fewer (6%) thought that it was likely that they would acquire HIV in the next year. Nine percent of participants reported that their partner had been tested and 8% thought it very likely their partner would acquire HIV.
Condom Use and Protected Sex
At baseline, 1 woman reported using condoms consistently and 40 (10%) reported a protected sex during their last episode (see Table 2). At 2-month follow-up, reported protected sex increased to 48.5% throughout the study (P ≤0.001) and 87% at last sexual episode (P ≤0.001). At follow-up 65% (n = 239) reported that every act was protected in the past 2 weeks, either by male or female condom use; Only 20 participants (5%) reported consistent use of female condoms in the past 2 weeks.
Baseline Predictors of Consistent Protected Sex
Consistent protected sex throughout the study was not significantly associated with any baseline sociodemographic or reproductive characteristics, domestic violence or baseline condom negotiation self-efficacy (Odds ratios presented in Table 1). Perception of male condoms as very effective in pregnancy prevention was associated with consistent condom use (P ≤0.01), but perception of male condoms as very effective for HIV/STD prevention was not (P = 0.12). Women who had ever been tested for HIV at enrollment (all of whom were negative) were less likely to report consistent condom use (P ≤0.01), as were women who reported that their partners had tested for HIV (P ≤0.01). The few participants who stated it was likely that they already had HIV were far less likely to have protected sex consistently (P = 0.03).
Logistic regression analysis with significant variables from the bivariate analysis showed that, controlling for age and education, the belief at baseline that male condoms are very effective for pregnancy was associated with an increased odds of protected sex throughout the study (OR 1.9; P = 0.015). Women who reported it likely they already have HIV (OR 0.23; P = 0.032) or that their partners have been tested (OR 0.38; P = 0.025) had a lower odds of reporting protected sex (see Table 3).
Changes During the 2-Month Study
Study participants were offered HIV testing and 288 accepted, and were tested within 2 weeks of baseline. Results for 287 participants showed that 72 (20%) were positive (see Table 4). At 2 months follow-up, the HIV-positive women were significantly more likely to report consistent condom use (71%) than HIV negative women (45%) or those who did not test (43%). Out of the 72 women who tested positive, only 4 (5.6%) of them had reported at baseline that it was likely that they already had HIV, while 36 of the women who tested positive (50%) had reported that it was not likely they had HIV.
The perception of male condoms as effective for HIV/STI prevention increased from 69% to 86% after the intervention, and was associated with consistent condom use in bivariate analysis (P = 0.05), as were beliefs about pregnancy prevention (P = 0.03). Perceptions about female condoms were not associated with consistent protected sex. Self-efficacy about condom negotiation also increased after the intervention. At baseline, 47% of women who had used condoms reported having a lot of control over condom use compared with 72% at follow-up (P <0.001). Women reporting a lot of control at follow-up were more likely to report consistent protected sex (P <0.001). The proportion who felt very confident that they could convince their partners to use condoms increased from 36% to 80%, and at follow-up was associated with consistent protected sex (P <0.001). The proportion who felt confident that they could refuse sex without a condom increased from 23% to 57%. These women at follow-up were more likely to report consistent condom use (P <0.001), as were women who reported shared decisions to use condoms (P = 0.002).
Participants’ perceptions of their own HIV risks, however, did not increase. At baseline 6% thought it was very likely that they would get HIV in the next year; after the intervention only 4% did, excluding women who tested positive. Furthermore, at follow-up participants’ perceptions of whether they would acquire HIV in the next year remained unrelated to consistent condom use (P = 0.17). However, the perception of partners’ risk for HIV as very likely in the next year increased after the intervention (15.5%), and was associated with consistent condom use (OR 2.6; P = 0.002). At baseline only 12 (3%) women thought it likely their partners had HIV; at follow-up the number increased to 44 (12%).
Logistic regression analysis of intervention outcomes at follow-up (see Table 5) showed that, controlling for age and education, condom negotiation measures were still positively associated with protected sex (shared decision to use condom OR 2.7, P = 0.001; confident could refuse sex without a condom OR 3.0, P ≤0.001), as was HIV-positive status (OR 2.8; P = 0.006). Hormonal contraception at follow-up had a negative association (OR 0.3, P = 0.003).
In this study of married Zimbabwean women who were inconsistent condom users at entry, there was a dramatic increase in protected sex from less than 1% to approximately 50% of participants at 2-month follow up after a brief male and female condom intervention. Although such high percentages of consistent condom use have been reported after an intervention,14 caution must be exercised in generalizing this finding to other condom interventions outside of a study environment. This finding must be viewed in light of limitations affecting research on condom use, including the fact that these women were enrolled in a study. The ability to negotiate condom use in marriage outside of a study setting may well be lower. Additionally, we used self-reported measures of condom use, which are likely to be influenced by social desirability bias of participants. Furthermore, 2 months is a short period, and HIV protection requires long-term behavioral changes that were not assessed in this study. We measured protected sex by both male and female condom use. Although most protected acts were covered by male condoms and consistent use of female condoms remained low in the study, female condom use is likely to have somewhat different predictors than male condom use (a more detailed analysis of female condom use is presented in Napierala et al. 2007).15
Many of the women screened for the study were found ineligible, including pregnant women and those desiring pregnancy. These married women, however, also constitute an important group in Zimbabwe at elevated HIV risk. Although those desiring pregnancy by definition would not use condoms, the study of condom use during pregnancy for disease prevention, would be a worthy research endeavor, and a topic for pilot interventions.
Women’s perceived self-efficacy in negotiating condom use increased substantially after the intervention and was strongly associated with consistent protected sex, a finding confirmed by other studies.4,16,17 Although measurement of self-efficacy differs in studies, certain themes emerge as important components in condom use. These include the perceived ability to convince partners to use condoms, to refuse sex without a condom, and to talk to partners about condoms. Our findings confirmed that women who participate in decision-making about condom use with their partners, and those who are confident that they can refuse sex without a condom, are more likely to have protected sex consistently. These data, however, are not definitive in addressing the question of whether responses to condom negotiation self-efficacy and condom use go hand in hand, or whether higher condom negotiation self-efficacy helped to increase condom use. The fact that these measures increased after the intervention and were associated with condom use suggests that women can gain control over condom negotiation with appropriate support and intervention.
The beliefs that condoms are effective protection for HIV/STD and pregnancy have also been found to be associated with condom use. In one study, concern about pregnancy and HIV transmission influenced condom use differently in marital and casual relationships: in marriage, women’s concern about pregnancy was associated with condom use, but concern about HIV was not. In casual relationships, concern about HIV, and pregnancy, was associated with condom use.5 Our findings in married women also showed an association between condom use and belief in effectiveness of condoms for pregnancy prevention. Pregnancy prevention as a reason for condom use is likely a more acceptable rationale to married couples since it does not raise questions of fidelity. HIV prevention efforts, therefore, may benefit by emphasizing condoms for preventing pregnancy and HIV/STDs in married populations. Of note, participants who used hormonal contraception were less likely to be consistent condom users, which is also consistent with the literature.18,19
Women in this study dramatically underestimated their HIV risks, and continued to do so after the intervention. Furthermore, among HIV negative or status-unknown women at follow-up, the perception that they were at high risk of acquiring HIV was not associated with consistent condom use. Other studies, in contrast, have shown that perceived HIV risk was associated with condom use.2,5,11,12,20 One study of married couples in South Africa found that the greatest predictor of condom use was a woman’s perceived risk of HIV infection from her partner.11 We did find that risk perceptions of the partner at follow-up were related to consistent condom use, but these associations did not persist in the logistic model. The effect of the intervention on perception of women’s own HIV risk and that of their husbands may have been difficult to interpret because of the concurrent HIV testing in the study.
One finding that raises interesting questions was that women who were tested for HIV as part of the intervention and had positive results were more likely to report protected sex at follow-up. Similar findings were described in a prospective cohort study of men and women in Eastern Zimbabwe who underwent voluntary HIV counseling and testing (VCT): women who tested positive were more likely to report consistent condom use with their regular partners.21 VCT has been shown to significantly affect condom use in serodiscordant relationships.22 Yet, since all women in this study reported monogamy, and few reported multiple lifetime partners, many HIV-positive women were likely to have been infected by their husbands. So while their increased condom use may be a laudable result of the intervention, we are left with the critical challenge of increasing condom use among married women who are HIV negative. Furthermore, HIV-positive women may be more likely to overreport condom use.
Sociodemographic measures did not predict consistent condom use in our analyses, in contrast to other studies. Education has been shown to be associated with condom use,2,5,23,24 as has socioeconomic status.25 The lack of these associations here may be explained by the relative homogeneity of the study population; all of the women lived in an urban environment, none had an education higher than secondary school and almost all had running water.
Studies linking partner violence and HIV prevention practices, including condom use, are few and have shown mixed results.26 A negative relationship was found between domestic violence and consistent diaphragm use in the Phase II women of the study.27 Although we did find high rates of domestic violence in our population, we did not find any associations with consistent condom use. Risky sexual behaviors in our study were exhibited by male partners, not by the women themselves. Given their monogamy, interventions and policies for these married Zimbabwean women based on fidelity are not likely to be effective. Likewise, abstinence is not a viable solution for these married women, particularly since more than half reported that they have not yet completed childbearing.
The low rates of male or female condom use in marital relations will likely continue to fuel the transmission of HIV to married women in Sub-Saharan Africa. Although new female-controlled alternatives are being tested, women currently need interventions that enable them to build female and male condom negotiation skills and confidence in being able to protect themselves. This study suggests that a brief and focused intervention may make a significant impact on women’s self-efficacy, skills and on safer sex patterns. Critical questions that remain are whether these results can be sustained beyond 2 months and whether on-going booster sessions are needed for lasting behavior change.
1. UNAIDS. Keeping the Promise: An Agenda for Action on Women and AIDS. Geneva: Global Coalition on Women and AIDS (GCWA) and Joint United Nations Programme on HIV/AIDS (UNAIDS); 2006.
2. Maharaj P, Cleland J. Condom use within marital and cohabiting partnerships in KwaZulu-Natal, South Africa. Stud Fam Plann 2004; 35:116–124.
3. Quigley M, Munguti K, Grosskurth H, et al. Sexual behaviour patterns and other risk factors for HIV infection in rural Tanzania: a case-control study. AIDS 1997; 11:237–248.
4. Norman LR. Predictors of consistent condom use: a hierarchical analysis of adults from Kenya, Tanzania and Trinidad. Int J STD AIDS 2003; 14:584–590.
5. Van Rossem R, Meekers D, Akinyemi Z. Consistent condom use with different types of partners: evidence from two Nigerian surveys. AIDS Educ Prev 2001; 13.
6. UNAIDS. Report on the global AIDS epidemic 2006. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS); 2006.
7. Central Statistical Office Zimbabwe and Macro International Inc. Demographic and Health Survey Zimbabwe 1999. Harare: Macro International Inc; 2000.
8. Central Statistical Office Zimbabwe and Macro International Inc. Zimbabwe Demographic and Health Survey 2005–06. Calverton, Maryland: CSO and Macro International Inc.; 2007.
9. Gregson S, Garnett GP, Nyamukapa CA, et al. HIV decline associated with behavior change in eastern Zimbabwe. Science 2006; 311:664–666.
10. Mumtaz Z, Slaymaker E, Salway S. Condom use in Uganda and Zimbabwe: Exploring the influence of gendered access to resources and couple-level dynamics. In: Kishore S, ed. A Focus on Gender: Collected Papers on Gender Using DHS Data. Calverton, MD: ORC Macro; 2005.
11. Maharaj P, Cleland J. Risk perception and condom use among married or cohabiting couples in KwaZulu-Natal, South Africa. Int Fam Plan Perspect 2005; 31:24–29.
12. Mbizvo E, Msuya S, Hussain A, et al. HIV prevalence in Zimbabwean women: 54–67% knowledge and perceived risk. Int J STD AIDS 2003; 14:202–207.
13. Bandura A. Social cognitive theory and exercise control of HIV infection. In: DiClemente R, Peterson J, eds. Preventing AIDS: Theories and Methods of Behavioral Interventions. New York: Plenum Press; 1994:25–59.
14. O’Leary A, Moore JS, Khumalo-Sakutukwa G, et al. Association of negotiation strategies with consistent use of male condoms by women receiving an HIV prevention intervention in Zimbabwe. AIDS 2003; 17:1705–1707.
15. Napierala S, Kang M, Chipato T, et al. Female condom uptake and acceptability in Zimbabwe. AIDS Educ Prev 2008; 20:121–134.
16. Adih WK, Alexander CS. Determinants of condom use to prevent HIV infection among youth in Ghana. J Adolesc Health 1999; 24:63–72.
17. Karim AM, Magnani RJ, Morgan GT, et al. Reproductive health risk and protective factors among unmarried youth in Ghana. Int Fam Plan Perspect 2003; 29:14–24.
18. Morroni C, Smit J, McFadyen L, et al. Dual protection against sexually transmitted infections and pregnancy in South Africa. Afr J Reprod Health 2003; 7:13–19.
19. Sangi-Haghpeykar H, Posner SF, Poindexter AN III. Consistency of condom use among low-income hormonal contraceptive users. Perspect Sex Reprod Health 2005; 37:184–191.
20. Meekers D, Klein M. Determinants of condom use among young people in urban Cameroon. Stud Fam Plan 2002; 33:335–346.
21. Sherr L, Lopman B, Kakowa M, et al. Voluntary counselling and testing: uptake, impact on sexual behavior, and HIV incidence in a rural Zimbabwe cohort. AIDS 2007; 21:851–860.
22. Allen S, Meinzen-Derr J, Kautzman M, et al. Sexual behavior of HIV discordant couples after HIV counseling and testing. AIDS 2003; 17:733–740.
23. Kapiga SH, Lwihula GK, Shao JF, et al. Predictors of AIDS knowledge, condom use and high-risk sexual behaviour among women in Dar-es-Salaam, Tanzania. Int J STD AIDS 1995; 6:175–183.
24. Zellner SL. Condom use and the accuracy of AIDS knowledge in Cote d’Ivoire. Int Fam Plan Perspect 2003; 29:41–47.
25. Hargreaves J, Morison L, Chege J, et al. Socioeconomic status and risk of HIV infection in an urban population in Kenya. Trop Med Int Health 2002; 7:793–802.
26. Jewkes RK, Levin JB, Penn-Kekana LA. Gender inequalities, intimate partner violence and HIV preventive practices: Findings of a South African cross-sectional study. Soc Sci Med 2003; 56:125–134.
27. 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.