AIDS and Behavior

, Volume 13, Supplement 1, pp 47–54

A Qualitative Assessment of Decisions Affecting Contraceptive Utilization and Fertility Intentions among HIV-Positive Women in Soweto, South Africa


    • Perinatal HIV Research UnitUniversity of the Witwatersrand
  • Catherine S. Todd
    • Department of Obstetrics and GynaecologyColumbia University
  • Mark A. Stibich
    • Division of International Health and Cross-Cultural MedicineUniversity of California
  • Rebecca Phofa
    • Perinatal HIV Research UnitUniversity of the Witwatersrand
  • Xoliswa Behane
    • Perinatal HIV Research UnitUniversity of the Witwatersrand
  • Lerato Mohapi
    • Perinatal HIV Research UnitUniversity of the Witwatersrand
  • Glenda Gray
    • Perinatal HIV Research UnitUniversity of the Witwatersrand
Original Paper

DOI: 10.1007/s10461-009-9544-z

Cite this article as:
Laher, F., Todd, C.S., Stibich, M.A. et al. AIDS Behav (2009) 13: 47. doi:10.1007/s10461-009-9544-z


The HIV epidemic in sub-Saharan Africa disproportionately affects women of reproductive age. The increasing provision of Highly Active Anti-Retroviral Therapy (HAART) with improved prognosis and maternal-fetal outcomes calls for an understanding of fertility planning for HIV-positive women. We describe the effect of HIV and HAART on pregnancy desires and contraceptive use among HIV-positive women in Soweto, South Africa. Focus group discussions and in-depth interviews were conducted with 42 HIV-positive women of reproductive age. Analysis was performed using ATLAS-ti (ATLAS-ti Center, Berlin). Emergent themes were impact of HIV diagnosis on pregnancy intentions; factors affecting contraceptive uptake including real and normative side effects, body image, and perceived vaginal wetness; and the mitigating influence of partnership on both pregnancy intentions and contraceptive use. Routine counseling about pregnancy desires and contraception should be offered to HIV-positive women.


ContraceptionFertility intentionsHIVHAARTQualitative methods


In many societies, women’s identity is defined by child bearing. The ability to reproduce affects social standing, individual recognition, partnership stability, and through these conventions, financial security (Serour 2008; Oosterhoff et al. 2008; Yount et al. 2000). This has been described in several sub-Saharan African countries, including South Africa (Sonko 1994; Bambra 1999; Dyer et al. 2008). In South Africa, the total fertility rate in 2006 was estimated at 2.6 with 39% of women aged 15–19 years having had at least one child (Health Systems Trust 2008). Effective contraceptive methods are freely available through government sources and widely utilized, but only 27.8% of reproductive-aged women use contraception consistently over 1 year (Health Systems Trust 2008).

Human immunodeficiency virus (HIV) has had variable effects on fertility in different cultures, both for expectations and actual childbearing following diagnosis. In Nigeria, desire for children was significantly associated with non-disclosure, younger age and recent diagnosis in a multivariate analysis (Oladapo et al. 2005). In contrast, childbearing desire in HIV-infected men and women in the United States was 29%, and here pregnancy intent was associated with better health, younger age, and fewer children (Chen et al. 2001). In South Africa, HIV acquisition tends to occur at earlier ages for women, often prior to initiation of childbearing (Health Systems Trust 2008). Young women likely have an expectation of childbearing, and are more likely to voice desire for pregnancy after diagnosis (Chen et al. 2001; Myer et al. 2007a). Unstated desires may be communicated through risky behavior; high rates of unprotected intercourse were recorded among HIV-positive South Africans, possibly driven by pregnancy desire (Lurie et al. 2008).

Highly active antiretroviral therapy (HAART) has prolonged life expectancy, potentially changing fertility expectations. Longer HAART usage and HAART-associated health restoration were associated with increased childbearing desire amongst women living in sub-Saharan Africa (Kaida et al. 2006; Myer et al. 2007a). However, as a study comparably found across three sites in South Africa, Brazil and Uganda, sexually active women using HAART were both more likely to desire pregnancy and to practice protected sex; hence actual childbearing may be affected (Kaida et al. 2008).

Multiple factors influence childbearing desire. General health status, which may vary relative to other factors over time, tends to be stronger for women than for men (Chen et al. 2001). Male partners appear to contribute greatly to pregnancy decision-making; studies in sub-Saharan Africa have noted that men are more likely than women to desire children after diagnosis (Myer et al. 2007a; Nakayiwa et al. 2006). In the African context, demonstration of fertility is often a prerequisite for marriage and the social definition of womanhood is tied to having children and a husband (Preston-Whyte 1988; Sonko 1994).

Pregnancy intentions, stated or not, may affect contraceptive utilization. There are few assessments of reasons for contraceptive choice and discontinuation in HIV-positive women. A longitudinal study in Uganda found low stated pregnancy desire and low contraceptive use (Homsy et al. 2009). A study in Malawi among women newly-diagnosed with HIV noted decreased stated pregnancy desire, but only 46% contraceptive use at 12 months post-diagnosis (Hoffman et al. 2008). Women receiving HAART are more likely to practice protected sex and to use contraception (Myer et al. 2007b). A randomized trial comparing copper intrauterine devices (IUD) to systemic hormonal contraception among HIV-positive women in Zambia noted that 31% across both arms discontinued their originally allocated contraception; half the women discontinuing the hormonal contraceptive arm chose to terminate contraception altogether (Stringer et al. 2007).

The purpose of this study was to explore reasons for contraceptive choice and discontinuation, and the relationship of these beliefs to pregnancy intentions among HIV positive women in Soweto, South Africa. It forms part of a multi-site qualitative assessment to determine whether culture and availability of HAART affect contraceptive method acceptability among HIV positive women and will provide formative data for improved method choice interventions.


Multisite Study Overview

This qualitative assessment is being conducted among HIV-positive women of reproductive age in Rio de Janiero, Brazil; Kericho, Kenya; and Soweto, South Africa. All sites follow the same protocol with similar eligibility criteria and interview guides. Results from the Soweto site are described in the following text.

South Africa Study Setting

In South Africa, HIV prevalence among antenatal clinic attendees was 28.0% (95% CI 26.9–29.1%) in 2007 (National Department of Health 2007). HAART became available in the South African public sector in 2004, with 3,29,000 patients receiving HAART as of September 2007 (PEPFAR Report 2008). In the public health system, injectable and oral hormonal contraceptives and male condoms are available at no cost; elective termination of pregnancy is legally available at no cost following referral to 12 weeks gestation.

Soweto is a collection of townships near Johannesburg, South Africa, and has an estimated population of 1.5 million (Wikipedia 2008). HIV prevalence among antenatal clinic attendees is 29.7% (95%CI 29.1–32.8) in the greater Johannesburg area (National Department of Health 2007). Two tertiary medical centers and several primary health care clinics provide HAART.

The study was conducted at the Perinatal HIV Research Unit (PHRU) which provides clinical care and conducts research primarily among HIV-positive individuals. At PHRU, between 2004 and 2006, 53 (8%) women receiving HAART conceived. Only ten pregnancies were reported as planned; 19 (35%) underwent voluntary termination indicating possible unmet need for contraception (Mohohlo et al. 2006).


Eligible participants were HIV-positive women between the ages of 15 and 40 years, with access to regular medical care and either receiving HAART, ineligible for HAART or eligible but not yet ready for treatment. They were able to provide voluntary written informed consent in the local language. A convenience sample was selected. Potential participants were identified at the time of clinic visit based upon chart screening for age. The potential participant was invited to a private room where a clinic provider introduced the study interviewer, who provided a brief description of the study. Interested patients received complete participation information. Participants were given a stipend for transport costs not advertised before informed consent. Prior to initiation of study activities, approval was received from the Human Research Ethics Council of Witwatersrand University and the Human Research Protections Program at the University of California, San Diego.

Data Collection

Study representatives were female professionals with prior study experience who received training in human subject research and qualitative methods. After completing a standardized screening questionnaire about HIV and obstetric history, HAART and contraceptive utilization, women were interviewed in a room with a locking door.

Focus Group Discussions

Focus group discussions (FGDs) were conducted to gain insight into the main themes for each topic of interest. A field guide was piloted with one group for refinement and assessment of question understanding. With the finalized field guide, three separate hour-long FGDs were conducted with eight participants each. Each FGD composed an age stratum: HIV-positive women who were 18–22 years old, 23–30 years, and those from 31 to 45 years old. Interviews were conducted by a moderator with a sessions recorder present to capture non-verbal communication and areas of group agreement/disagreement. Interviews were conducted in a mixture of English, Sotho, and Zulu, with the moderator fluent in all three languages. FGDs were audio-recorded and translated into English whenever local languages were used. Investigators reviewed the transcripts to develop in-depth interview questions.

FGD Question Content

Focus group discussions questions addressed women’s attitudes toward fertility, contraception, and the impact of HIV and partnership on these attitudes. Concepts identified in the pilot session (e.g., “wet vagina”) as unique to the South African context were probed in FGDs. Participants were asked about childbearing desire with the question, “Do women’s attitudes towards pregnancy and childbearing change after HIV diagnosis?” Contraceptive use and factors affecting choice were assessed with such questions:
  • What contraceptive methods have you used? What did you like/dislike about them?

  • What birth control methods are most effective at preventing pregnancy?

  • What birth control methods are most effective at preventing infection?

  • What would an ideal contraceptive method be like?

  • What would be the benefits/drawbacks of a contraceptive method that stopped menstruation?

  • Do you think women in your community would use a contraceptive method that stopped menstruation? Why or why not?

The effect of partnership on contraceptive use was explored with the question, “What are your partner’s attitudes toward contraception?”

In depth Interviews

In-depth interviews (IDIs) were conducted among 15 women to explore further emergent FGD themes. IDIs were approximately 45 min each, conducted in the primary language of the participant, and audio-recorded with only the interviewer and participant in the room. Interview tapes were translated from the local languages to English, where necessary.

IDI Content

The IDI guide utilized the “grand tour” approach in which the participant is asked a broad question to encourage her to speak openly about the topic. The “grand tour” question bounds the interview while still encouraging detailed responses from the participant (Bernard 2001). We employed the grand tour question twice to assess beliefs about contraception: “Tell me everything you know about birth control” and “Tell me everything you think about what a perfect prevention method should be like.” A series of probe questions were placed after each grand tour question in the IDI guide to remind the interviewer of specific issues to be addressed if not spontaneously raised by the participant, such as “Which birth control methods have you used and stopped, and why?”. Other IDI questions invited participants to share their perceptions on the meaning, significance, and feelings around contraception and amenorrhoea, what they thought men’s perceptions are about contraception and menstruation, and how HIV status affects contraceptive choice.

Data Analysis

The transcriptions of the audio recordings from the FGDs and IDIs were entered into Atlas-ti (ATLAS-ti Center, Berlin). The transcriptions were coded using a grounded theory approach in which the codes and codebook emerge from the data (Glaser 1992). These codes were then grouped into themes.



Three FGDs and 15 IDIs were conducted with 42 participants (Table 1). Of the eight women not using contraception, none stated pregnancy desire. Pregnancy within the next 2 years was desired by 3/42 (7.1%) women, of whom two were taking HAART, and two were using injectable contraception.
Table 1

Demographic Data for HIV Positive Female Participants of Focus Group and In-depth Interviews from Soweto, South Africa (n = 42)


% (N)



Age (years)



IQR 31–22

Time since HIV diagnosis before study interview (years)




Participants currently on HAART

50 (21)


Percentage reporting a prior pregnancy

95.2 (40)


Number of prior pregnancies




Mean prior live births




Mean prior elective terminations




Contraceptive utilization at time of interview

Nuristerate (norethisterone enantate)

40.5 (17)


Depot-provera (depot medroxyprogesterone acetate)

19.0 (8)


Male condoms

26.2 (11)


Female condoms

4.8 (2)


Oral contraceptives

2.4 (1)


Surgical sterilization

2.4 (1)


No contraceptive method

19.0 (8)


Desire pregnancy within the next 2 years

7.1 (3)


Main Themes

Except where indicated, FGD and IDI data have been combined.

Theme I: Fears and Desires for Pregnancy

Discussion of future pregnancies elicited multiple fears: that the child would be born HIV-positive, that the family would be burdened with more children if HIV shortened the mothers’ lives, that pregnancy could have negative consequences on their own health (a decrease of CD4 count in pregnancy, illness during pregnancy and loss of blood at delivery were the most expressed worries), and that there was a risk of re-infection with a different strain of HIV or partner transmission while trying to conceive. Consensus arose in the FGD setting that having more children after knowing one’s HIV status exhibited lack of responsibility toward oneself, one’s child and one’s family. However, in the individual interview setting, a few women did voice desire for another child despite their fears, justified by reasons like change in partnership status (marriage), becoming financially stable (a new job; buying a house), and the desire to have babies. While women were aware of prevention of vertical HIV transmission strategies, they still tended to overestimate the risk.

My boyfriend and I talk about having kids and all that, once we get married. He understands I am positive and he is negative. I don’t know. He talks about me going to a gynecologist and maybe they can do something like plant a baby or something. But I don’t know. I would really like to have a baby. I don’t think my status really changes anything about me being pregnant, but the problem is that how would I get pregnant? Making a baby involves two people and what about the next person? Will I infect the next person, what would happen? IDI 28 [3928–4548]

Theme II: Contraceptive use and Factors Affecting Uptake

Women mentioned side effects of contraceptives, convenience and protection against STIs as major factors in contraceptive use. Condoms were said to be a preferred method because of the lack of side effects and the protection against STIs. The main drawbacks to condoms were the fear of breakage and the necessity of male partner participation. The pill was the least favorite method because of the perceived burden of having to remember to take a daily pill. Injectable contraceptives were liked for their convenience and (in some cases) their secrecy. Other methods of contraception (e.g., tubal ligation, the loop) were largely unknown. Reasons for discontinuation of injectable contraceptives included need for follow-up, side effects changing physical appearance especially affecting weight and skin, and concern about disordered menstrual bleeding.

Amenorrhoea secondary to injectable contraceptives was perceived as potentially harmful to health. Menstruation was seen as a way to rid dirt in the blood and participants were concerned about where “all the dirty blood” would go in cases of amenorrhoea. Menstruation was perceived as an indicator of health; cessation of menstruation (if not pregnant or menopausal) was interpreted as sickness. Contraceptives which caused amenorrhoea were thought by some to make a woman sick. Some women noted that that amenorrhoea is more convenient, and better for relationships (male partners do not become upset/withdrawn because of menstruation). In addition, amenorrhoea eliminated visualizing blood, a disturbing reminder of HIV to some women.

Injectables were perceived to increase vaginal wetness, thought to impact negatively on relationships with male partners because of males’ association of vaginal wetness with female promiscuity. Some women resultantly kept their female-controlled contraception use a secret from their partners.

He (the participant’s boyfriend) then said to me that I should not use the injection because I will be too wet when he has sex with me and it will feel like I have had sex with many other guys and therefore I have not told him that I am using the injection. (FGD 12)

Non-disclosure about the female-controlled method made women perceive that their partners would be more likely to use male condoms.

Also, the reason why I have not told him (that I use contraception) is so that we can always use a condom. When he sometimes suggests that we have sex without a condom, I tell him that we cannot do that because I am not using any contraceptives and I am scared of falling pregnant again. So I do think that he wants me to use contraceptives because he loves the idea of having sex without a condom. (FGD 14)

Discussion around HAART and contraceptives elicited trepidation about their concomitant use. Anxiety that the potency of HAART overrides the efficacy of the pill and so should not be used together. Several women reported the clinic advising them to switch from the pill to an injectable contraceptive because HAART would render the pill ineffective. Apprehension that injectable contraception may interfere with HAART efficacy. Participants were convinced that HIV-infected couples on HAART should use condoms to prevent re-infection with possibly drug-resistant strains of HIV that may adversely affect HAART efficacy.

He is on ARVs and I’m not so we have to always use a condom. What I have heard here from the clinic is that if I have unprotected sex with him, and when it’s my turn to start taking drugs, the drugs will be resistant. IDI 29 [5428–5832]

Theme III: The Influence of Partnership

Participants reported varying levels of partner influence. The role of the partner’s fertility desires was considered, as was the cultural expectation that fertility must be proven in a long-term relationship. Overall, while some women reported equal negotiation of condom use, many participants reported that men resisted the use of condoms and mentioned the need to explain, convince and argue with them.

Yes, we use the male condom. We both have a say. When I suggest that we don’t use it, we discuss it and then we both have to agree. Other days he would suggest it by telling me that he feels like ejaculating inside me. However, before we don’t use a condom, we both agree on it first. If there is no agreement, there will be no sex. IDI 32 [6371–6704]

Many women took into account their partners’ opinions when choosing female-controlled contraception. Perceived partner disapproval was most commonly observed with the avoidance of the injectable method because male partners thought it made women fat or have a “wet” vagina, and these were seen as undesirable.


This study represents one of the first qualitative assessments regarding contraceptive acceptability and fertility intentions among HIV-positive women in South Africa. Many participants agreed that risks associated with childbearing following HIV diagnosis outweighed benefits of future fertility. We believe their reflections on being unable to care for both themselves and chronically ill children, and their experiences (a few had children, relatives or friends whose children had succumbed to AIDS) shaped responses about fertility intentions significantly, but the influence of provider counseling should also be assessed. Our findings are similar to a Zambian qualitative study which found that participants were concerned about pregnancy accelerating progress to AIDS, and the welfare of children after their parents have died of AIDS (Rutenberg et al. 2000). In contrast to another South African study, the observation that HAART use modified fertility intentions was not observed among women interviewed here (Myer et al. 2007a). Notably, although most women stated no desire for future childbearing, this desire did not always correlate with action: all women not using contraception stated they did not desire pregnancy at the time.

It is now well-documented in biomedical literature that HAART improves longevity, is linked with good pregnancy outcomes, and reduces both horizontal and vertical HIV transmission (Lima et al. 2007; Tai et al. 2007; Quinn et al. 2000; Garcia et al. 1999). Even though half of the women in our study were taking HAART, there was little knowledge that viral suppression with HAART had the potential to ameliorate their fears about HIV and fertility. We offer two possible reasons: the lack of educational materials on this topic, and the attitudes of health care providers toward HIV and fertility, already described in South Africa to be strongly influenced negatively by biomedical concerns (Harries et al. 2007). The role of health care providers in relaying accurate reproductive health information to HIV-positive women, untainted by personal bias, has been emphasized (Myer et al. 2005).

Human immunodeficiency virus status pertained to contraceptive consideration in two ways in our study. First was women’s awareness of male condom use for prevention of HIV and STI transmission—despite the low proportion of women reporting recent condom use. Partner influence, a major factor in contraceptive decisions in our study, was prominent in their description of condom negotiation tactics: in some cases, the struggle for partner compliance led women to state that an ideal method would protect against both STIs/HIV and pregnancy with use solely in control of the woman. Several women stated that condoms were the ideal contraceptive method due to dual protection and because insisting on condom use to prevent undesired pregnancy made it safer to avoid disclosure of HIV status, particularly in a new, non-committed relationship. However, this became problematic in situations where childbearing was desired by the male partner and seen as necessary for attaining marriage. This attitude likely reflects the tradition of lobola, the bride price mandated for marriage, held contingent upon demonstrated fertility in some circumstances (Mbirimtengerenji 2007).

The second way in which HIV affected contraceptive choice was the consideration of potential interactions with concomitant HAART. Reported advice from providers about interactions between oral contraception and HAART is supported by various studies such as recent pharmacokinetic data that efavirenz significantly reduced levels of some oral contraceptives (Sevinsky et al. 2008). However, we also encountered misconceptions said to originate from providers that injectable contraceptives adversely affect HAART efficacy, which may unnecessarily discourage use of this method. It has been noted previously that peer and provider misinformation may dissuade HIV-positive women from using highly efficacious long-term methods, such as the IUD, whose use is generally considered safe in HIV-positive women but was not seen in our study at all (WHO 2004; Sinei et al. 1998; Stringer et al. 2007; Williams et al. 1996).

Side effects also affected contraceptive choice. Two main side effects were deterrents: vaginal wetness and amenorrhoea. Participants commonly perceived from their peers and partners that injectable methods caused a “wet vagina” rumored to be a sign of female promiscuity, and some would therefore not use injectable contraception. This has previously been implicated in method discontinuation or refusal and represents a partner-mediated effect (Smit et al. 2002). More than half of the respondents in a South African survey reported preferential practice of “dry sex”, implicated in increased HIV and STI transmission, in which the vagina is dried and tightened to heighten friction and sexual enjoyment, and to demonstrate that a woman is not promiscuous (Beksinska et al. 1999).

Our participants largely expressed preference for contraception methods which preserved menstrual bleeding, perceived as necessary to health. Discontinuation of systemic hormonal methods due to altered bleeding patterns is common (Tolley et al. 2005; Winkler et al. 2004). The relationship between HIV disease and changes in menstrual pattern is variable, based on prior study results; however, awareness of possible changes and their association with advancing disease among HIV-positive women may make contraceptive methods that alter bleeding patterns even less attractive than among HIV-negative women (Chirgwin et al. 1996; Ellerbrock et al. 1996).

This study has several limitations. The small sample size, non-random sampling techniques and semi-structured nature of the interview guides may not have captured all themes. We employed the “grand tour” approach in the IDIs to reduce the latter limitation. We did not ascertain partners’ views. Focus group forums may have induced social desirability bias, but we attempted to control for this with individual interviews. We note that stated beliefs that HIV-positive women should not have further pregnancies did not vary substantially between FGDs and IDIs; however, the presence of an interviewer and recording equipment in IDIs may have influenced responses.

In summary, our qualitative study concludes that in the Soweto setting amongst HIV-infected women in the reproductive age group, stated fertility desires are shaped by fears and considerations around the disease. Contraceptive choices are determined by real and normative side effects, body image, and perceived partner detection of vaginal wetness.


Facilities offering care to HIV-positive women may consider the following to optimize family planning:
  • Routine, open, rapport-building, ongoing discussions about fertility desires. As partnership, finances and other factors change, pregnancy desires may also evolve.

  • Provision of correct, unbiased information about factors affecting vertical and heterosexual transmission risks, and risks of pregnancy to maternal HIV progression. This may contextualize fears around fertility and promote responsible reproductive health-seeking behavior. Clear, carefully-designed patient education leaflets addressing fertility, HIV and HAART may aid this goal.

  • HIV-positive women not desiring pregnancy should be counseled about contraception needs, including dual, longer-term, and female-controlled methods, known interactions with HAART, partner detection, and known side effects. Actively dispel common contraceptive myths. Before prescribing methods which may cause amenorrhoea, providers should address women’s concerns.

  • The gaps in knowledge of how HIV and HAART affect fertility and contraception require ongoing research.


This study was supported by the Morris S. Smith Foundation and the Doris Duke Charitable Trust. We acknowledge USAID/PEPFAR. We thank participants for their time and trust.

Copyright information

© Springer Science+Business Media, LLC 2009