1In sub-Saharan Africa, where the HIV/AIDS epidemic is widespread and the mode of transmission is essentially heterosexual, women in union are massively affected. Women often learn that they are HIV-positive when they become pregnant. How does this discovery affect their lives? In this article, Annabel DESGRÉES du LOÛ presents the results of a survey conducted as part of a clinical research programme focusing on pregnant women in Abidjan. She shows that under half of the women who discover that they are HIV-positive tell their partner, and even fewer then start to use condoms. Though the duration of postpartum abstinence tends to increase among these women, no reduction in fertility is observed. Women’s difficulties in telling their partner and their continued childbearing despite the risk of mother-to-baby transmission should be taken into account in programmes to fight the AIDS epidemic.

2Of the 38 million people infected with HIV worldwide, 25 million or a large majority live in sub-Saharan Africa. Of these 25 million, UNAIDS estimates that 13 million are adult women and 9 million are men (UNAIDS, 2004). HIV transmission in Africa is essentially heterosexual. Given the levels of prevalence reached in various countries—though some are much harder hit than others—the HIV/AIDS epidemic constitutes a threat for all societies in sub-Saharan Africa. It must be taken into account at every level of society, of course by politicians and by all who work in the health system, but also in the private sphere of families, spouses and individuals.

3Study of the HIV/AIDS epidemic focused for a long time on “high-risk groups” such as prostitutes, truck drivers and migrants. It now seems necessary also to consider the impact on the general population, widely exposed in countries with high rates of HIV infection, and on couples in particular. Studies have shown that most HIV-positive women in sub-Saharan Africa had been infected by their husbands during marital sexual intercourse (Cohen and Reid, 1999). This notion is widely accepted today, and in African societies AIDS is now seen as a threat that concerns everyone, both through sexual activity, the main route of HIV transmission, but also through childbearing, since the AIDS virus can be transmitted from mother to child during pregnancy, delivery or breastfeeding.

4In this article we analyse how the AIDS epidemic challenges sexual and reproductive behaviour within the couple, where sexuality and childbearing are regarded a priori as “not dangerous”. To what extent are the changes observed linked to the level of communication within the couple on the question of HIV testing? First we describe the changes the AIDS epidemic has caused to the general population’s sexual and reproductive decisions, in the absence of HIV/AIDS screening. This is based on a literature review of the research conducted in sub-Saharan Africa. We then analyse the particular case of couples in which the woman knows she has HIV/AIDS. This part is based on the results of an “AIDS and Reproductive Health” research programme conducted by IRD [1] in Abidjan since 1997, in collaboration with INSERM [2] as part of the DITRAME [3] mother-to-infant HIV-transmission prevention programme.

5Abidjan, the economic capital of Côte d’Ivoire, is one of the largest cities in West Africa (population approximately 3 million), and the hardest hit by the AIDS epidemic, with an estimated 14% of the population infected in the mid-1990s (Sylla-Koko et al., 1997) and 11% in 2002 (Sakarovitch et al., 2003). Although these figures are lower than those for southern Africa, what is happening in Abidjan can tell us much about the way a severe epidemic is managed in an urban context where African modernity is taking shape.

1 – The couple in sub-Saharan Africa: a complex, changing notion

6There is no easy way to define the couple in African societies. Unions may be formal or informal, monogamous or polygynous, legal or customary. There are many kinds of couples, and conjugal relationships are now undergoing profound changes. In the traditional societies of many regions, based mainly on agriculture, it was the perpetuation of the lineage (ownership of the land) that counted when a couple was formed. The conjugal bond was weak and was less important than kinship ties. As women had little power of decision and little autonomy, a large number of children was the only way for them to acquire power, respect and access to land and resources.

7With the introduction of a money economy and access to waged employment, this dependence on the lineage is reduced; the individual no longer depends on the elders who distribute land, and is free to choose his or her partner. As girls gain access to education and enter the labour market, a new generation of mothers become more independent in their choices and decision making, acquire greater bargaining power and act increasingly as partners of their husbands in the couple. On the basis of their work in Tanzania, Marida Hollos and Ulla Larsen define two kinds of couple. The traditional type is lineage-based, the marriage is arranged by the family, often in a context of polygyny, and the wife has little decision-making power. The second type is the partnership marriage involving a modern monogamous couple in which the spouses have chosen each other freely and define themselves as partners in the management of family affairs (Hollos and Larsen, 1997). The alliance on which the couple is based no longer binds two families but two individuals (Vimard, 1993). The conjugal bond, which formerly had to make way for other relations such as lineage bonds and filiation bonds, gradually takes centre stage. Women begin to take a larger part in decisions that concern the family and a more equitable dialogue is established in the couple, which evolves towards a partnership between the man and the woman.

8Although this pattern of change reflects an existing trend, it does not describe the full range of diversity in African conjugal bonds. Reality is more complex. Polygyny, even though officially in decline since it has been declared illegal in a growing number of countries, is actually still widespread. In Côte d’Ivoire in 1994, for example, 37% of women said they had at least one co-wife (EDS, 1994). It is also re-emerging in new forms: kept mistresses and “outside wives” (Antoine and Nanitelamio, 1989; Lacombe, 1987). Thérèse Locoh (1994) has shown how the search for greater autonomy led women to “invent” new forms of union in between the polygynous marriages of traditional African societies and the Western model of the nuclear household. A young woman may have a declared partner, who may be the father of her child or children, while continuing to live with her parents. Another new type of union is the “kept” mistress who has her own home and lives with her children while her partner lives with his “official” wife. Thus the marriage bond can take many forms depending on whether the household is polygynous or monogamous, official or unofficial, legal or customary, whether the woman lives on her own or with her partner, and with or without co-wives.

2 – Communication in the couple

9Whatever the form of union, all studies agree that the extent of communication between partners increases with the educational level of the members of the union (and especially of the woman) (Babalola, 1999; Orubuloye et al., 1997b), and that the more dialogue there is in the couple, the better their reproductive health indicators. Babalola has shown a very clear link in Tanzania between the fact that the couple talk about aspects of reproductive health and the likelihood that the woman will have medical assistance at childbirth or adequate monitoring during pregnancy. Hollos and Larsen (1997), working in the same country, note that the variable most closely linked to the use of contraception is the fact that the spouses had previously discussed the possibility of using it. A growing proportion of couples appear to be talking thus about sexual and reproductive matters. For example, some women interviewed by Orubuloye and his team (1997b) in Nigeria say that since the AIDS epidemic began they have not hesitated to express their disapproval of men’s infidelities in strong terms, and that they have broached these issues in their couple in order to protect themselves from any sexually transmitted disease their husband might have contracted “elsewhere”. These Nigerian women say they take part in decisions about sexual intercourse and family planning. Two out of three say they would refuse sexual intercourse if their husband had a sexually transmitted disease (STD). It should be noted that when the partner does have an STD, it appears to be easier for the other to refuse intercourse than to impose the use of condoms. Dialogue is not always easy. When spouses do communicate, they do not necessarily do so on all subjects, and talking about sex proves particularly difficult, as illustrated by this statement from a Kenyan woman (Bauni and Jarabi, 2000):

10

“It is difficult for some couples to discuss, because the only place they ever discuss issues is in the bedroom. The husband just wants to have sex and not discuss family planning.”

3 – The couple and HIV/AIDS

11In a context where the conjugal bond takes many forms and discussion of sexual matters in the couple, though increasing, is still difficult, how does the HIV/AIDS epidemic affect behaviour in countries where the epidemic is severe and any person having sexual intercourse is in danger?

4 – The Abidjan surveys

22Our partial answers to these questions are drawn from surveys on health and reproduction issues conducted under the DITRAME [8] programme in Abidjan, Côte d’Ivoire between 1995 and 2000. DITRAME was a clinical research programme aimed at reducing mother-to-baby HIV transmission, conducted in “urban health units” in Yopougon, a working-class district of Abidjan. The “urban health units” are clinics that attract a broad patient base from all social classes.

23Three surveys [9] were conducted under the project; for the sake of clarity we will refer to them as surveys A, B and C in this article:

24—Survey A: a retrospective survey on the fertility of HIV+ and HIV- women, conducted with approximately 1,250 pregnant women between July 1997 and January 1998. The purpose of the survey was to gather precise retrospective information about women’s reproductive histories and link it to their serological status. Special attention was paid to the most recent birth interval, with a detailed description of the duration of breastfeeding, periods of amenorrhoea and use of contraception.

25Since the women were interviewed when they came for the antenatal visit where the HIV test was offered, i.e. at a time when they did not yet know their serological status, any observed relation between reproductive life and HIV status was of a biological nature.

26—Survey B: an exploratory qualitative survey conducted in May 1998 among 21 HIV+ women followed under the DITRAME project. All these women had given birth at least eight months before, and all their children were weaned. During the interview they were asked questions about weaning (age and circumstances of weaning, attitudes of the family and the husband towards weaning), relations with their husband in connection with HIV (discussion of HIV, safe sex, contraception) and the reproductive intentions of the woman and the couple (whether they wanted another child, current use of contraception, and so on).

27—Survey C: a quantitative survey conducted in 1999 among 149 HIV+ women followed under the DITRAME project. This recorded all the events in the women’s reproductive life (amenorrhoea, resumption of sexual activity and safe sex, incidence and fate of pregnancies, contraceptive practices) between the birth of the child for which they were followed and the date of the survey. Unlike survey A, this one covered events occurring at a time when the women knew they were infected with HIV.

5 – Communicating the results of an HIV test to the partner: information that is hard to share

28Table 1 compares levels of information sharing about HIV status recorded in Abidjan (Survey C) and various other African studies of the reduction in mother-to-baby HIV transmission. Abidjan, where 40% of the followed mothers had told their spouse that they were infected with HIV, emerges as one of the projects with the highest rates of communication about test results. In Burkina Faso and Tanzania, the percentage was only half as high. The interviews (survey B) show that women who had not informed their husband had been motivated by fear that he would abandon them, leaving them destitute. One mentioned the danger that her spouse would tell other people. Another said her spouse would not want to touch her or eat from the same dish. In this qualitative survey, when women did decide to tell their husbands, the husband usually reacted well (except in one case) and decided to support the woman’s future involvement in the DITRAME trial. However, it would appear that women tend to “probe” their husbands before telling them, and that those who tell are also those who expect a positive reaction. So caution is required in interpreting this result, which should not be extrapolated to all partners. Those who were not informed by their wife might have reacted more negatively had they learnt about her seropositive status.

29The quantitative interview shows that the husbands’ reactions were strongly negative (anger, violence, in some cases divorce) in about one case in twenty. This figure is the same in Burkina Faso, where the authors, when comparing women who had told their husbands and those who had not, found no correlation between the fact of the partner being informed and the future of the couple: the proportion of couples breaking up was the same in either case (Nébié et al., 2001). Although the probability of rejection by the husband seems low, the fear of this social and emotional rejection remains strong, and many women prefer to say nothing, thus carrying the burden of their secret alone, and to face decisions about feeding the baby, future pregnancies, etc., on their own. Among the women who did not tell their husbands, several reported that they wanted to, but did not know how to find “the right moment”. The nature of the conjugal relationship seems decisive in the woman’s decision to share information about HIV with her partner. In a similar study conducted in Tanzania, the degree of information that seropositive women gave their partner depended strongly on the type and duration of the union. Being married and the fact that the union had lasted for more than two years greatly increased the probability of telling the partner. In contrast, the woman’s economic independence and the fact that she had a female friend to whom she could confide her HIV status were associated with a lower probability of telling the partner (Antelman et al., 2001).

30Talking to someone about one’s HIV status is of prime importance. Being able to talk to their husband is particularly important for women who depend on their spouse, but remains difficult, especially for these dependent and therefore vulnerable women who are afraid of their partner’s negative reaction. And yet living with HIV is easier when the partner knows, when it is possible to share and discuss the infected person’s HIV status and the advice received concerning prevention (Tijou-Traore, forthcoming). For this reason, given the difficulty infected people have in telling their partners, advice and screening programmes that address the couple rather than the individual are organized in many African countries. The members of the couple meet the adviser together; testing is individual and confidential but the team can help one partner announce a positive test result to the other, if necessary (Painter, 2001; Roth et al., 2001; Allen et al, 2003).

6 – Strategies adopted in Abidjan for avoiding sexual transmission of HIV

31In Abidjan as elsewhere in Africa, low levels of condom use are recorded even after HIV testing (Table 2). In survey C, among 149 HIV-positive women followed by the DITRAME project, 64% of those who were sexually active at the time of the survey said they never used condoms and only 13% said they used a condom every time they had intercourse. (All women in the survey had been tested when pregnant and knew they were infected with HIV). These proportions varied widely according to whether or not the woman had told her partner. Among sexually active women who had told their partners (45%, or 48 out of 107), 48% never used condoms and 21% used one every time they had intercourse, compared to 76% and 7% respectively when the partner had not been told (Table 2) [10].

32To the question “Why do you not use condoms?”, more than half the women did not answer; 21% said they had suggested condoms to their partner but did not use them regularly (or not at all) because the partner (and in some cases the woman herself) did not like to use them; 14% said they had not suggested condoms for fear of their partner’s reaction or because the woman did not think it was necessary; 7% did not use them because they wanted to become pregnant; and 3% did not for religious reasons. The interviews revealed two types of attitude underlying the non-use of condoms in couples where the woman had divulged her HIV status to her partner: denial of the infection, or the partner’s belief that their negative HIV status despite living with an infected person was proof that they had a “natural resistance”. As one woman expressed it, “He thinks that since he hasn’t caught AIDS so far, that means he can’t catch it”.

33Alongside the low rate of condom use, HIV-positive women in the follow-up monitored remained abstinent after delivery for a longer time than the comparable group of women in survey A, the survey that collected information on the reproductive lives of women before their HIV tests: 12.0 months (SD 8.8) (group C) compared with 9.6 months (SD 12.9) (Group A) (Desgrées du Loû et al., 2002; Desgrées du Loû et al., 1999) [11]. Although women who have reported their positive HIV status to their partners also prolong the postpartum abstinence period to some extent, this increase is greater among women who have not told their partners (Table 3). It appears that prolonging postpartum abstinence is one of the strategies adopted by women who know they are HIV-positive, so as to avoid the risk of transmitting the infection sexually, given (a) the difficulty they have in telling their partner about their infection and (b) even when they have told their partner, the difficulty they have in suggesting that he use a condom and in obtaining compliance. The interviews revealed strong reservations to condom use among male partners, including those who had been informed, but also among the women. Condom use by men who have not been told of their partner’s HIV-positive status remains extremely uncommon. These problems are not surprising in Abidjan, where condoms are rarely used; a survey conducted in 1996 in Yopougon, the Abidjan neighbourhood where the DITRAME project is taking place, showed that only 4% of the women of childbearing age were using condoms (Toure et al., 1997).

34An increase in the duration of postpartum abstinence is only observed when the women have that possibility. It is easier for a woman to decide not to start having intercourse again if she is not living with a man. In this case the abstinence period is twice as long as for women living in union (Table 3). Conversely, it seems that the level of education is a hindrance to this strategy; this may be because it is easier for less educated women to invoke the tradition of extended postpartum abstinence to justify long abstinence to their partners without revealing that they have HIV. Also, traditionally, sexual activity was not resumed until breastfeeding had stopped, one of the beliefs behind the postpartum taboo being that “sperm spoils the milk” (van de Walle and van de Walle, 1988). It might have been expected that early weaning by women in the DITRAME group (following WHO recommendations for HIV-positive women to reduce the risk of transmitting HIV in breast milk) would reduce the period of postpartum abstinence, and yet we observe the opposite. These women, knowing they are HIV-positive, find other arguments than breastfeeding to justify the late resumption of sexual intercourse. This point should be investigated further in subsequent research.

35Thus the difficulty of sharing information about HIV after a positive test during pregnancy is compounded by the difficulty of coping with the risk of transmission through marital sexual intercourse. Experiments conducted in other countries to target HIV screening and advice on the couple rather than the individual show that this approach considerably increases condom use by couples. In Zambia, the proportion of times condoms were used in sexual intercourse between couples with different HIV status rose from 3% to 80% after a screening and advice programme that targeted couples (Allen et al., 2003).

36It is therefore a matter of urgency for prevention and screening programmes to consider not only women but also men, and more generally the notion of the couple (Roth et al., 2001; Allen et al., 2003). It is increasingly recognised that preventing sexual transmission of AIDS depends not only on making condoms available but also on improving dialogue within the couple. Even if women are made aware of the risk of HIV infection and have condoms, even if they are organized in women’s groups to help and support each other and learn more about HIV/AIDS, they are alone in dealing with their spouse in daily life. The protection of sexual relations when there is a risk of HIV transmission will depend on the quality of the relationship between men and women and the ability to negotiate condom use (Baylies and Bujra, 2000).

7 – HIV and childbearing decisions in Abidjan

37It might be expected that women who know they have HIV would choose to reduce their fertility, as recommended by the medical team at the time of the screening test, to avoid giving birth to an infected child or leaving it orphaned a few years later.

38In fact, although the proportion of women in the DITRAME project who practiced contraception following the pregnancy during which they were tested is considerably higher than in the general population—about 39% compared to 16% in the population of the same neighbourhood in 1997 (Toure et al., 1997)—this contraceptive practice is not due to a desire to avoid pregnancies. Pregnancy incidence is still high: of 149 HIV-positive women interviewed in 1999 (survey C), 37 had had at least one pregnancy during postpartum follow-up. When related to the duration of the follow-up, this corresponds to an incidence of 10.8 pregnancies per 100 woman-years at risk for the first 12 months and 16.5 pregnancies per 100 woman-years at risk over the first 24 months postpartum. This incidence is equivalent to that observed by Susan Allen’s team in Rwanda (Allen et al., 1993) and more than double that observed in the similar French SEROCO cohort three years after the birth (De Vincenzi et al., 1997). Among these pregnancies, 46% (17/37) were desired by the woman and 51% (19/37) were not (one woman did not answer the question). Among the pregnancies that were not desired, 68% (13/19) were terminated voluntarily. This observation matches those made in another family planning study in Abidjan, where it appeared that abortion is sometimes used instead of contraception as a method of fertility regulation (Guillaume and Desgrées du Loû, 2002). One of the major factors leading to a new pregnancy among those women in the postpartum follow-up was the death of the baby (Desgrées du Loû et al., 2002).

39These results match those of similar studies in other African countries, which have shown that HIV testing and the accompanying information (advice on protected intercourse and contraception, provision of free condoms and contraceptives) did not reduce fertility among HIV-positive women. On the contrary, one study in Rwanda showed that women with fewer than four children hastened to have another child, as if they had to achieve an ideal number of children despite (or because of) the discovery that they had HIV (Allen et al., 1993; Keogh et al. 1994).

40The exploratory qualitative survey (survey B) supports the evidence from the data on birth incidence. It confirms that the desire for children is still strong in women who have been advised against pregnancy because they were infected by HIV, even though all the women concerned were aware of the risk of mother-to-baby transmission (Desgrées du Loû, 2000). In general, the only women who said they wanted no more children were those who already had at least three or four. Some did not want more children but were under heavy pressure from their husband, mother or mother-in-law, or had children with another man but not with the current partner. Half the women interviewed clearly stated they wanted another child.

41Thus the only women who decided not to have any more children were those who thought they already had enough. For the others, their desire for children was undiminished and often supported by the spouse’s desire (which was also undiminished by knowledge of his wife’s HIV status), or by other family members (such as mother or mother-in-law). This does not mean that these women were unaware of the risks attached to future pregnancies. During the interviews, many asked about the risk of infection for the future child. But the desire for children was stronger than that worry, as witnessed by the case of the mother who has just lost her HIV-positive child but wants another one immediately despite the risk. Women often expressed their desire for children at the psychological consultations that were available on demand to women followed by the project (Aka-Dago-Akribi et al., 1999).

42Confronted with the death threat associated with seropositivity, the project of having a baby is a promise of life, of a future. For a woman it is also a way of proving to herself and others that she is in good health and can lead a normal life. In some cases these pregnancies in women who know they are HIV-positive and already have at least one child are no doubt the result of social pressure. It appears that in Africa, deciding not to have any more children exposes women to too many social risks. The women we interviewed often mentioned the risk of being rejected by their partner if he wanted children. When, as is often the case, a woman does not divulge her HIV-positive status to her partner, it becomes extremely difficult for her to explain to her husband and his family why she does not want another child, unless the couple already has what is considered a sufficient number. Even when the husband is aware of his wife’s HIV status, there may be pressure from the mother-in-law to have another child. These women are afraid their mothers-in-law will “influence” their husbands, by trying to persuade them to divorce or take a second wife. And when, as is most frequent, the woman conceals her HIV status, not being pregnant also carries the risk of being labelled “sick” because of infecundity. On the contrary, a new pregnancy will reassure the family, but also the woman herself, that she is healthy and fertile. The coming child is a guarantee of stability for the couple and of the individual’s survival through their descendants (Aka-Dago-Akribi et al., 1999).

43Hence, with regard to both childbearing and sexual activity, awareness of the risk of HIV infection is not a sufficient motive for an individual to adopt suitable preventive behaviour. In the area of childbearing as in the area of sexuality, family, social and emotional imperatives appear to be more important than preservation of the health of one member of the couple or even the risk of giving birth to an infected child.

44In the light of this, all advances in research and in public health that provide a means to reduce mother-to-baby HIV transmission in developing countries are especially valuable. If HIV-infected people cannot accept the advice to avoid having any more children, they can at least be assisted in their desire for fertility by limiting as far as possible their future children’s risk of infection. The most recent clinical research programmes aiming to lower mother-to-baby transmission rates have developed simple, low-cost protocols that have reduced these rates in West Africa, from an initial estimate of more than 20% (in a situation where breastfeeding is common), to about 6% (Dabis et al., 2003). These programmes are now becoming operational.

Conclusion

45The first observation to be drawn from this study is that women who learn during pregnancy that they are HIV-positive want to share the bad news with their spouses but find it difficult to do so. In our study, fewer than one woman in two told her partner, and this is one of the highest proportions recorded in sub-Saharan Africa. Advice given to these HIV-positive women about safer sex fails to take effect, partly but not solely because the partner is not informed; even among men told that their wives have HIV, only a minority use condoms. The medical teams’ advice to HIV-positive women not to become pregnant again was not followed at all. The difficulty these women have in divulging their HIV status to their partners and in following advice on AIDS prevention are symptomatic of the difficulty enountered by couples in communicating about issues connected with sex and childbearing. Changing one’s sexual behaviour or reproductive choices does not depend only on the difficult decision to tell one’s partner that one has HIV. To adopt such changes, individuals must be personally convinced that these changes are useful, and capable of convincing their partner or of imposing those changes upon them. This implies a difficult negotiation, which led the women we interviewed in Abidjan to avoid intercourse and pregnancy by prolonging the traditional period of sexual abstinence after childbirth.

46There is an urgent need to understand more clearly the conditions that facilitate this type of negotiation concerning sexual relations within the couple. In African countries, where transmission is essentially heterosexual, where the majority of HIV-positive people have been infected through marital sexual intercourse, making sexual intercourse safe between spouses with verified or suspected HIV or STD infection remains one of the weak points in the fight against AIDS. A possible way of improving prevention of sexual transmission in the couple is to run prevention campaigns that do not target women only or men only, but also the couple as such, in the same way that family planning programmes must be designed to target men as well as women. For example, to improve condom use during marital relations it is important for prevention programmes to focus not only on the partners’ knowledge about condoms and on making condoms available, but also on communication within the couple about condom use (Painter, 2001).

47Similarly, decisions about childbearing involve a couple and not just a woman. Although programmes to prevent mother-to-baby HIV transmission should address women first and foremost and do everything to help them manage the consequences of their HIV infection, there is increasing awareness of the need to involve these women’s partners.

48At short, it appears that one of the present priorities in the fight against AIDS in Africa is to integrate the notion of couple and the relationship between man and woman within that couple, in the programmes of prevention and care of HIV/AIDS. This calls for a better understanding of everything covered by the notion of the couple in Africa, a notion with a broad range of meanings in societies where different types of union coexist, and a notion destined to change radically as these societies evolve.