Bridewealth and sexual and reproductive practices among women in Harare, Zimbabwe

Abstract

This paper examines the relationship between bridewealth, socio-demographics, and sexual and reproductive practices among a group of women in Harare, Zimbabwe. The study sample was recruited as part of a six-month safety trial of the diaphragm and a microbicide, between August 2004 and April 2005 in Harare, Zimbabwe. Women underwent two screening visits: first, women completed a demographic and behavioral interviewer-administered questionnaire which included questions on bridewealth; at the second visit, women were offered HIV testing and counseling. Our results included: 417 women were married (currently or in the past) and almost half had had bridewealth negotiated as part of the marriage process. In multivariate analyses, women who were married with bridewealth had more years of education (OR 1.17, 95%CI 1.03–1.32), a higher age of coital debut (OR 1.37, 95%CI 1.09–1.71), and increased likelihood of having ever used male condoms (OR 1.54, 95%CI 1.01–2.37) compared with women who had been married without bridewealth. Bridewealth may be a relevant area of traditional culture to further examine in relation to HIV risk, for its potential association with co-factors that can reduce risk of HIV infection among women in Southern Africa.

Introduction

Bridewealth, also known as lobola or roora in Southern Africa is a fundamental component in marriage that affects reproduction and sexual relations. However, it has not been examined extensively in relationship to risk for HIV infection in African women. Bridewealth is broadly understood as the exchange of resources between African families to finalize a marriage transaction. This exchange of resources is part of a marriage process and the payments are often made in installations after the couple is already co-habitating or sexually active, with some payments made with pregnancy and children, as compensation that the woman is fecund (Ansell, 2001). Lobola is associated with children and family lineage, and as such, directly influences decisions about family planning and safer sex practices, as well as has broader impact than on the nuclear family alone. The fact that bridewealth is linked with reproduction can create a situation where men believe they should have control over reproductive decisions after paying bridewealth and may refuse to use any contraceptives or practice safer sex (Bawah, Akweongo, Simmons, & Phillips, 1999).

The practice of bridewealth also impacts the family of the bride and groom financially. The exchange of goods between families may create a situation in many societies where the bride’s family is unwilling and/or unable to return the bridewealth in the event of divorce. Families may use force to demand that women stay married to her husband (Beattie, 1958; Spencer, 1988). In South Africa, issues of virginity impact the amount of bridewealth exchanged. Parents may demand higher price from suitors if a woman does not have a child yet as nulliparity is considered proof of virginity (Kaufman, de Wet, & Stadler, 2001). Although these examples suggest that closely intertwined relationships between the financial aspects of bridewealth, reproduction, and marriage may negatively impact African women, some African researchers argue that payment of bridewealth has certain benefits: a man may be less willing to leave his wife (Kaufman et al., 2001). One study from Zimbabwe suggests that women who have been married with bridewealth will have greater respect in marriage than women without these payments (Ngubane, 1987).

Zimbabwe has one of the highest HIV prevalence in the world with 15.8% of the adult population infected and 14.7% of young women (15–24 years) as per UNAIDS 2007 prevalence estimates (UNAIDS, 2008). A better understanding of factors that impact sexual and reproductive decision-making, including bridewealth, can help in efforts to prevent HIV infection in African women. The goals of this study were to explore the relationship between the practice of bridewealth and known risk factors for HIV infection including barrier contraceptive use, age of first sex, and number of lifetime sexual partners.

Methods

Study participants

This study sample was recruited as part of a six-month, randomized controlled safety trial of the diaphragm and microbicide Cellulose Sulfate, as described in van der Straten et al. (2007). Women were enrolled from public sector reproductive health and family planning clinics, churches, and community health centers in Harare, between August 2004 and April 2005. Using an effective contraceptive method was a requirement for enrolling in the study, thus, there were no participants who were currently exclusively using traditional methods or no method of contraception. All females were screened for trial eligibility criteria. At their pre-screening visit, women completed a demographic and behavioral interviewer-administered questionnaire. All women were screened over two successive visits, and HIV testing was offered to participants at the second visit (conducted within seven weeks of the first visit), but was not required to enroll in the trial as previously described (van der Straten et al., 2007). Socio-demographic characteristics of this sample were also previously reported in van der Straten et al. (2007).

Study procedures

Questions on the practice of bridewealth were asked as part of the first screening interview. These questions included whether or not the husband had negotiated bridewealth, whether bridewealth negotiations caused tension in the marriage, whether payments were made in full or as partial payments, and if the frequency or sexual relations changed after bridewealth payments were made, and if the bride-wealth payments were returned in the instance of divorce. Socio-demographic information was collected including: age, years of education, language use, religious practices, size of house, and number of people living in the house. Detailed information was also collected concerning current and lifetime history of sexual and reproductive practices. Questions included age of first sex, lifetime number of sexual partners, number of sexual acts per month, and current and past male and female condom use in addition to other contraceptive use practice. As described earlier, a sub-sample of the women who received the second screening visit agreed to be tested for HIV infection and were asked additional sexual practice questions including questions on washing/drying of the vagina. Questions on these practices were included because of the possible associations with increased risk for HIV infection or practices associated with increased risk for HIV infection (Bagnol & Mariano, 2008; Mbikusita-Lewanika, Stephen, & Thomas, 2009).

All study procedures were conducted at one study clinic in Harare, Zimbabwe, and written informed consent was obtained prior to study participation. The study was approved by the following review boards: the Committee on Human Research at XX; the XX Medical School Office of Research; the Medical Research Council of Zimbabwe (MRCZ); and the Medicines Control Authority of Zimbabwe (MCAZ).

Statistical methods

The study design was cross-sectional. For bivariate analysis, chi-squared test of significance and t-test were used to assess the relationship between the practice of bridewealth and a number of socio-demographic factors including age, religious background, education level, socioeconomic indicators (e.g., rooms in house and number of people living in the house), and our primary area of interest, which was known risk factors for HIV infection including age of first sex and number of lifetime sexual partners. Lifetime sexual partners were evaluated as a continuous variable. Contraceptive use was examined for current and lifetime use and was divided into the following categories: (1) barrier methods (female and male condoms, diaphragms); (2) long-term (tubal-ligation, vasectomy, norplant, and intrauterine device (IUD); (3) injections; (4) oral hormonal pills; and (5) other methods (withdrawal, traditional methods). All statistical tests were conducted using Stata 9.0 (Stata Corp., College Station, TX, USA).

Subsequently, multivariate logistic regression models were used to evaluate risk factors for the practice of bridewealth with all bivariate factors that were significant at p < 0.10 included in the models.

Results

Among our sample of 417 women who were married or had been married, 48.7% negotiated bridewealth and 6.4% stated that the bridewealth negotiations caused tension in the marriage. However, only 13.8% of those who had negotiated bridewealth stated that a full payment had been made with 85.2% stating that a partial payment had been made and 0.5% stating that no payment had been made (Table 1). The majority of the 203 women who had had bridewealth negotiated as part of the marriage process stated that the frequency of sexual relations remained unchanged with their husband (58.1%), while 16.3% stated that the frequency of sexual relations had decreased once the bridewealth payment was made (Table 1). We did not find any statistical difference in participant’s age between those who had had bridewealth negotiated and those who did not, however, those women who had had bridewealth negotiated were more likely to have more years of education (10.5±1.8 versus 9.8±1.9 years; p<0.01) (Table 2). In addition, 32.0% of those who had bridewealth negotiated had three or more rooms in their house in comparison with 29.0% of those who had not negotiated bridewealth; p=0.03. Decreased crowding (defined as the ratio of the number of people living in the house in relation to rooms per house) was also associated with bride-wealth negotiation (p=0.03).

Table 1.

Lobola/bridewealth frequencies.

	Percentage (%) (n/total)
Husband negotiated lobola before marriage	48.7 (203/417)
Lobola negotiations caused tension in marriage	6.4 (13/203)
Lobola payments were made as part of marriage negotiations
Full payment	13.8 (28/203)
Partial payment	85.2 (173/203)
No payments	0.49 (1/203)
Frequency of sexual relations changed after lobola payments were made
Frequency of sexual relations increased	24.6 (50/203)
Frequency of sexual relations decreased	16.3 (33/203)
Frequency of sexual relations unchanged	58.1 (118/203)
Lobola was returned to husband’s family after divorced	0 (0/4)

Table 2.

Lobola/bridewealth frequencies in relation to participants’ socio-demographics and sexual practices.

	No lobola	Lobola	p-Value
Socio-demographics
Age	29.7±6.5	30.7±6.7	0.10
Education, years	9.8±1.9	10.5±1.8	<0.01
English speaking	13.6 (29/214)	17.4 (35/203)	0.30
Christian religion	72.0 (154/214)	77.3 (157/203)	0.21
Rooms in house
0–1	39.7 (85/214)	27.6 (56/203)	0.03
2	31.3 (67/214)	40.4 (82/203)
≥3	29.0 (62/214)	32.0 (65/203)
People in house (number)	2.0±0.8	2.1±1.50	0.87
Crowding index (number people in house/rooms)	1.16±0.60	1.05±0.48	0.03
Sexual history
Age of first sex
≤16	22.0 (47/214)	12.3 (25/203)
17–18	35.5 (76/214)	25.6 (52/203)
19–20	27.6 (59/214)	37.4 (76/203)
≥21	15.0 (32/214)	24.6 (50/203)	<01
Lifetime sexual partners	1.49±0.70	1.37±0.60	0.08
Sexual acts per month	6.31±7.83	6.50±7.80	0.80
Male condom use (ever)	61.2 (131/214)	69.0 (140/203)	0.097
Female condom use (ever)	6.5 (14/214)	5.4 (11/203)	0.63
Diaphragm use (ever)	0.5 (1/214)	1.5 (3/203)	0.29
Reproductive history
Live births
0–1	27.6 (59/214)	28.1 (57/203)
2	33.6 (72/214)	33.5 (68/203)
≥3	38.8 (83/214)	38.4 (78/203)	0.99
Lifetime long-term contraceptive use (tubal-ligation, partner’s vasectomy, IUD, sub-dermal implants)	8.4 (18/214)	13.8 (28/203)	0.08
Current long-term contraceptive use	6.5 (14/214)	9.4 (19/203)	0.29
Lifetime barrier contraceptive use (female condom, male condom, diaphragm)	56.1 (120/214)	62.1 (126/203)	0.21
Current barrier contraceptive use	5.1 (11/214)	7.4 (15/203)	0.34
Lifetime oral pills contraceptive use	97.2 (208/214)	96.6 (196/203)	0.71
Current oral pills contraceptive use	72.4 (155/214)	74.4 (151/203)	0.64
Lifetime injectable hormone use	45.8 (99/214)	46.3 (93/203)	0.93
Current injectable hormone use	21.0 (45/214)	16.3 (33/203)	0.21
Lifetime other contraceptive use (withdrawal, natural/traditional medicine)	17.3 (37/214)	21.2 (43/203)	0.31
Current other contraceptive use	0.0	0.0

We also evaluated reproductive and sexual practices in relation to bridewealth and found that bride-wealth was not associated with number of live births (p=0.99), lifetime use of barrier contraceptives (male or female condoms or diaphragm use) (p=0.21), or current use (p=0.34) (Table 2). Lifetime long-term contraceptive use (use of tubal-ligation, partner’s vasectomy, IUD, or sub-dermal implants) neared significance, with those who had had bridewealth negotiated having a higher prevalence (13.8%) versus those who did not (8.4%) (p=0.08). We also found that those who did negotiate bridewealth had a slightly lower number of lifetime sexual partners although the results did not meet statistical significance (1.38±0.40 versus 1.49±0.70; p=0.06) (Table 2). We also found that those who had negotiated bridewealth also had a somewhat greater likelihood of having ever used condoms (69.0% versus 61.2%; p=0.097), but we did not ask any questions about condom use with specific partners or questions about condom use prior to or during marriage. The practice of bridewealth was not associated with number of sexual acts per month (p=0.80), but was associated with lifetime age of first sex with those who had exchanged bridewealth more likely to have had an older age of first sex (24.6% at an age of 21 or older compared with 15.0% of those without bridewealth; p<0.01). However, it was not clear whether sexual relations were deferred until marriage for either group or whether the first sexual partner was the marriage partner or whether bride-wealth payments had been made prior to engagement in first sex as these questions were not included in our questionnaire.

In a sub-sample, who received HIV-testing as part of their second study screening visit (n=111), we evaluated the relationship between bridewealth and HIV infection and other sexual practices. We did not find any association between being infected with HIV and having exchanged bridewealth (p=0.97). We also did not find any association between bridewealth and drying or washing the vagina, a practice that is common in a subset of women in Southern Africa and has been found to be associated with HIV infection (p=0.88).

In multivariate analysis, we found that increasing years of education was significantly associated with bridewealth practice (OR 1.17, 95%CI 1.03–1.32) as was increased age of first sex (OR 1.37, 95%CI 1.10–1.71) and greater likelihood of ever having had sex with condoms (OR 1.54, 95%CI 1.01–2.37) Table 3. Decreased number of lifetime sexual partners neared significance for being associated with bride-wealth practice (OR 0.75, 95%CI 0.55–1.03) Table 3. Participant’s age, number of rooms in the house, and lifetime long-term contraceptive use were not associated with bridewealth in multivariate analyses after adjusting for other co-variates.

Table 3.

Multivariate predictors for practicing/exchanging lobola.

Variable	Odds ratio (OR) (95%CI)	p-Value
Year of education	1.17 (1.03–1.32)	0.01
Age, years	1.03 (0.99–1.06)	0.17
Rooms in house	1.15 (0.88–1.50)	0.30
Age of first sex	1.37 (1.10–1.71)	<0.01
Lifetime sexual partners	0.75 (0.55–1.03)	0.08
Lifetime long-term contraceptive use	1.53 (0.78–3.01)	0.17
Male condom use (ever)	1.54 (1.01–2.37)	0.047

Discussion

The results of our study suggest that women who engaged in bridewealth practice had an increased age of first sex and an increased likelihood of having ever used condoms. Importantly, these predictors are behaviors that have been shown to be associated with decreased risk for HIV infection in other studies (Hallett et al., 2007; Pettifor, Van der Straten, Dunbar, Shiboski, & Padian, 2004). We also found that women who had negotiated lobola had a reduced number of lifetime sexual partners, another factor associated with reduced risk for HIV infection, although the association did not reach statistical significance. We did not find any association between HIV infection and bridewealth. We most likely lacked power, as we were only able to test this association in a small subset of women.

The cultural norms associated with bridewealth may be an important area to explore for future HIV prevention efforts, among certain population groups in Zimbabwe. In addition to the above factors, which are associated with reduced HIV risk, we also found that women who were more educated were also more likely to practice bridewealth. In Zimbabwe, the most educated and wealthiest men and women experience age of first sex later (Central Statistical Office [CSO] [Zimbabwe] and Macro International, 2000) and South African young women with higher education have a reduced risk for HIV infection (Pettifor et al., 2008).

While this study was exploratory, it suggests a possible protective role of lobola against behavior that may put women at increased HIV risk. Future, large-scale studies on the relationship between bride-wealth practices and HIV risk need to be studied in diverse African populations. This study was conducted almost exclusively with Shona-speaking women (97.0%) in Harare, Zimbabwe; however, other African groups may have quite different structures and attitudes toward bridewealth practice. Additionally, as one of the requirements for entering the clinical trial was that these women needed to be using effective contraception, this study cannot necessarily be generalizable to all Zimbabwean women, particularly those who are in marriages or relationships and not using effective contraception.

Other studies have emphasized the negative role that bridewealth can play in marital negotiations and for the position of women, in general, including studies from Zimbabwe. For example, the Musasa project, a non-profit organization founded by Zimbabwean women to transform ideas concerning domestic violence in Zimbabwe, takes the position that bridewealth may play a fundamental role in women’s subordination primarily based on qualitative data (Stewart, 1995). Moreover in Ghana, West Africa, ethnographic accounts suggest that violence is considered justified if women use family planning methods or refuse sex after bridewealth has been exchanged (Bawah et al., 1999). Among the Maasai of Kenya, families may use force to demand that a woman stay with her husband and in Nigeria, separation/divorce can be taken very seriously because of the need to repay money in the case of separation (Meek, 1970; Spencer, 1988). However, all of these previous studies are based on qualitative data and did not empirically assess risk associated with lobola practice. This is one of the first studies to empirically assess sexual and reproductive behavior associated with lobola practice, although as described earlier, our study is limited by our overall small sample size and generalizability to other population groups.

As our results suggest that bridewealth is associated with behaviors known to reduce the risk for HIV infection, it is important to evaluate the role that bridewealth plays in sexual behavior. If bridewealth plays a protective role against HIV infection in women, future studies need to systematically evaluate the role of bridewealth in relation to female sexual health, reproductive practices, and other aspects of women’s health, including the possible association with domestic violence. Empowerment of women is a central goal for many sub-Saharan African intervention programs, including those focused on women’s health, as a way to reduce risk for HIV infection (Kim et al., 2007). As an ultimate goal, we need to tease out which aspects of the socio-cultural systems and cultural practices, including bridewealth, may place women in disempowered roles and facilitates gender-based violence or conversely which aspects provide women with certain benefits that may ultimately help reduce risk for HIV infection. In summary, it is important to evaluate how bridewealth as an important cultural component of African marriage, impacts women’s as well as men’s sexual behaviors in different contexts.

The limitations of this study include the small sample size as well as the absence of questions on domestic violence, our incomplete testing of the sample for HIV infection, and the limited data collected on frequency of condom use as well as the cross-sectional design of our study. Additionally, our population is not a representative sample as they were screened for a safety trial and were required to be on modern, effective forms of contraceptive in order to enter our study. We also evaluated the relationship between lobola and sexual and reproductive practices cross-sectionally and not longitudinally. Longitudinal data would have provided information on the potential effects of negotiating lobola as well as more detailed information on the sexual and reproductive practices with the partner with whom lobola was negotiated, as well as other partners, outside of marriage or prior to the start of marriage. Future studies should include comprehensive questions on domestic violence and should evaluate the relationship between lobola and HIV infection longitudinally, in addition to recruiting a sample without reference to current contraceptive use. Other studies should also be conducted with a larger sample size to potentially evaluate the relationship between negotiation and payment of lobola and HIV infection.

As discussed earlier, the results of our study do not concur with the results from the qualitative and ethnographic reports, which associate lobola with increased domestic violence and female subordination, although we did not directly investigate these questions. Domestic violence and female subordination are often associated with increased risk for HIV, while in our study, the practice of lobola was associated with decreased risk for behaviors that increase risk for HIV. HIV prevention programs for women need to be designed and tailored the socio-cultural context of the specific environment, which in the sub-Saharan Africa milieu, include the frequent practice of lobola which is a fundamental component of sexual and reproductive relations. Given the paucity of information on lobola in previous studies which have evaluated risk for HIV in African women, and the associations that we found with known risk factors for HIV infection, as well as the need to further evaluate and confirm qualitative findings, it is important for future well-designed studies to systematically evaluate this practice in relation to risk for HIV infection and women’s health in different African contexts.