Introduction

In regions with high prevalence of HIV infection (Rutenberg et al., 2000), low levels of contraceptive use (Ross et al., 2004; King et al., 1995) and high value of childbearing (King et al., 1995) such as Sub-Saharan Africa, addressing fertility issues among people living with HIV is critical for prevention of unwanted pregnancies and prevention of HIV transmission from mother-to-child. For example, with a fertility rate of 6.9% (Uganda Bureau of Statistics Entebbe, 2002) and high HIV prevalence among women of reproductive age, Uganda faces numerous PMTCT challenges. National HIV prevalence among women is 8% (Uganda HIV/AIDS Sero-Behavioral Survey, 2004/5) and 67,000 HIV-infected women are expected to become pregnant annually. Without PMTCT interventions, an estimated 30% of children born to HIV-infected women become infected with HIV annually (Ministry of Health-The Republic of Uganda, 2003).

Reducing unintended pregnancies among HIV-infected women by 16% would yield an equivalent reduction in HIV-positive infant cases as would implementation of a national prevention of mother-to-child transmission (PMTCT) program involving provision of antiretroviral therapy (Sweat et al., 2004). Due to numerous programmatic, social, cultural and economic factors in Uganda, awareness of mother-to-child transmission of HIV (MTCT) and PMTCT services is generally low; the uptake of PMTCT services is limited, and only a small proportion of HIV-infected pregnant women actually benefit from PMTCT services (Ministry of Health-The Republic of Uganda, 2003). Even among those HIV-infected women who do seek PMTCT services, promotion of family planning, a method for primary prevention of PMTCT, has not been widely utilized.

HIV-infected women in Africa have shown lower fertility compared with their HIV-negative counterparts, (Hunter et al., 2003; Boerma and Urassa, 2000; Gray et al., 1998) and suppressed fertility reduction is more pronounced during later stages of HIV infection (Hunter et al., 2003; Gregson et al., 1997). However, the impact of HIV infection on women's and men's childbearing and contraceptive decisions appears weak (Nebie et al., 2001; Rutenberg et al., 2000). In Burkina Faso, HIV-infected women who participated in PMTCT programs had equivalent rates of subsequent pregnancies to HIV-negative women and had poor use of contraceptive methods despite regular advice and counseling (Nebie et al., 2001).

Factors affecting HIV-positive persons’ desire to have children in industrialized countries have included: younger age, better health status or improved health on ART, number of prior abortions, increased motivation for child bearing, traditional gender roles, having a partner of unknown HIV status, being married or having a sexual partner, being childless, pregnancy motivations, concern about infecting partner or spouse, concern for personal health, concern for health of child, concern about future care of child, religious beliefs and concern about ART (Panozzo et al., 2003; Wesley, 2003; Sowell et al., 2002; Richter et al., 2002; Chen et al., 2001;). Factors associated with an HIV-infected person's pregnancy risk behavior were ethnicity, religion, thinking that you are infertile, having multiple partners and having a steady partner (Raine et al., 2003; Van Benthem et al., 2000).

Parenthood was perceived as a source of self-esteem for HIV-infected women in the United States (Smits et al., 1999; Bedimo et al., 1998; Sowell and Misener, 1997). In Cameroon, HIV-infected men and women said that they have unprotected sex primarily because they wished to have a child, even when antiretroviral treatment (ART) was not available (Kuyoh and Best, 2001). Desire for children has been correlated with lower levels of protective sexual behavior (Kuyoh and Best, 2001) and low utilization of family planning methods among HIV-infected individuals which could be seen as an effort to replace children who may have died from HIV infection (Moyo and Mbizvo, 2004).

Although some correlates of desire for children among HIV-infected persons have been identified, much less is known about the relationship between PMTCT/MTCT knowledge and desire for children or pregnancy risk behavior. In this study, we examine whether knowledge and attitudes towards mother-to-child transmission of HIV (MTCT) and prevention of mother-to-child transmission of HIV (PMTCT) programs are related to desire for children and pregnancy risk behavior among a sample of HIV-infected persons in Eastern Uganda. To inform and emphasize a broader approach to PMTCT interventions, we studied factors associated with desire for children and pregnancy risk behavior among HIV-infected men and women in a setting with high fertility rates and high HIV prevalence.

Methods

Setting

Data were collected as part of a larger study called Prevention with Positives Study (PWP) which was conducted in the semi-urban town of Jinja, Eastern Uganda, at the The Aids Support Organization (TASO), Jinja Branch and two TASO outreach sites (Mayuge and Kiyindi). At the time of the study, TASO provided access to medical care (but not yet ART), counseling, and social support to HIV-infected persons. Any HIV-infected Ugandan can register with TASO free of charge and most clients are of low socio-economic status. Although TASO now has a free ART program for its clients, they did not at the time of this study and very few TASO clients had the ability to buy ART from private clinics.

Study description

The PWP study collected behavioral, clinical and laboratory data from HIV-infected participants from October 2003 to June 2004. To be eligible, participants had to be at least 18 years old, HIV-infected, and cleared as physically fit for the study by medical staff. On days when more clients attended the clinic than could be enrolled, clinic staff randomly selected men and women using a paper-based system: clients who picked one of a limited number of marked sheets would be offered enrollment in the study.

Because women constituted 70% of the TASO clients, men were over-sampled. After a brief orientation to the study, those interested and eligible participated in a group consent education session and then consented individually with a counselor and signed individual consent forms in the language of their choice (Lusoga/Luganda or English). Thereafter, study participants: received a rapid HIV test to confirm HIV infection; completed a face-to-face behavioral and medical interview; provided biological samples for sexually transmitted diseases and CD4 testing; and were examined by a physician.

Measures

Socio-demographic characteristics

Data were collected on age, gender, education, occupation, religion, marital status, number of living children, number of live births, and death of any children.

Health status

We used CD4 count cells per mm3 as a proxy for health status and dichotomized this variable (≤200cells/μL and >200cells/(μL).

Table 1 Socio-demographic characteristics and knowledge of PMTCT information of HIV-infected adults, Jinja, Uganda
Fig. 1.
figure 1

Desire for children and pregnancy risk behavior among clients

Sexual activity

Respondents reported partner-specific condom and family planning behaviors in the last three months. Men and women were then categorized as sexually active if they had had at least one sexual partner in the past three months or as abstinent if they reported no partner in the prior three months. Temporary abstainers gave reasons for abstaining such as poor health, no interest in sex, no partner and fear of re-infection while the permanent abstainers declared that they never wanted to have sex again in their lives.

Table 2 Desire for children and partner's desire for children

Years spent as TASO client

We calculated years of receiving TASO services from date of interview and participants TASO registration date.

Desire for children

Participants were asked whether or not they wanted more biological children, and for those who did, their reasons for wanting more children. Desire for children was cross-tabulated with pregnancy risk behavior to calculate the proportions at high risk for unwanted pregnancies (e.g. had reported no desire for more children but were engaging in pregnancy risk behavior).

Partner's desire for children

Participants were asked how important they thought it was to their partners to have more children. Responses categories included very important, important, somewhat important, not very important and don't know.

Knowledge and attitudes about mother-to-child transmission of HIV (MTCT)

We used three variables to assess general knowledge and attitudes towards MTCT. Participants were asked whether they knew that HIV can be transmitted from mother to infant; whether they knew that HIV transmission from mother-to-child can be prevented and whether they thought that HIV-infected women should have children.

Table 3 Factors associated with desiring more children among HIV-infected adults, Jinja, Uganda

Knowledge and attitudes about prevention of mother-to-child transmission (PMTCT)

We utilized three variables to assess knowledge and experience with PMTCT programs. Participants were asked whether they have ever heard of PMTCT programs; whether they have ever discussed PMTCT with friends or family members and their perception of the possibility of transmitting HIV to an infant after an HIV-infected pregnant mother takes nevirapine. At the time of the study, the Ugandan national PMTCT program services were available in Jinja and involved voluntary counseling and testing at the antenatal clinic, provision of nevirapine to mother and newborn, and follow-up at the hospital.

Pregnancy risk behavior

We asked participants about their use of contraceptives and condoms (never, sometimes, and always) with up to three partners in the past three months. We defined pregnancy risk behavior as having inconsistent or no condom use without any other method of contraception with at least one of their partners in the past three months. We excluded women >50 years of age and those who were pregnant as well as men who reported a pregnant partner.

Data analysis

Socio-demographics, health characteristics, numbers of living children, knowledge and attitudes about MTCT/PMTCT and years spent at TASO were assessed as predictor variables for two outcomes: desire for children and pregnancy risk behavior. Desire for children was subsequently tested as a possible predictor of pregnancy risk behavior. For each outcome variable, we determined the bivariate associations and then conducted multivariate logistic regression including only the significant variables from the bivariate analyses at p < .05. Multivariate results are presented unless otherwise specified.

Results

Participant characteristics

Of 1,092 participants who met the inclusion criteria, all agreed to participate; 604 (55%) were women, and 56% had a CD4 count ≤200. More men (46%) clients had attended TASO <6 months compared with women (16%). Mean age was 37 years for women and 40 years for men. Sixty-four percent were Christian. Men were more likely than women to have higher education and to be engaged in paid employment; most women were subsistence farmers. Sixty-nine percent of the men were married while three-quarters of the women were widowed, separated or divorced. Most of the participants had children; the median number of children was 4 (5 for men and 4 for women). Of all participants, 54% had experienced the death of at least one biological child. (see Table 1).

MTCT/PMTCT knowledge and attitudes

Knowledge was high regarding the possibility of HIV transmission from mother-to-baby (71%), its prevention (81%) and about the PMTCT program (78%). However, 78% did not know that an HIV-infected woman who took PMTCT drugs could still deliver an HIV-infected baby. Men were more likely than women to think that HIV-infected women should still have children (43% versus 34%, p < .001). However, overall 62% of the total study population thought that HIV positive women should not have children (see Table 1).

Desire for children

Forty-two percent (455) of participants were sexually active, and of these, 18% (83) desired children (see Fig. 1). The four main reasons for wanting more children were: “need to leave ancestry” (52%), not having any boys (14%), not having any children or all children died (12%), and not having girls (9%).

Of participants who had no partner at the time of interview, 37% of HIV-infected men and 5% of women desired more children. Likewise, men with partners were more than twice as likely to desire children than women (see Table 2). Approximately 50% of both men and women who wanted more children thought it was very important to their spouse to have more children (see Table 2). Among those who did not want children and were currently married or cohabiting with a partner, 24% thought that their partners wanted more children.

Overall, 16% of HIV-infected men and women desired more children and of these, men were almost four times more likely to want more children than the women (27% vs. 7%, p < .001]. Being male, younger, number of living children (irrespective of HIV status), and believing that HIV-infected women should have children were independently associated with desire for children in multivariate analysis. Being a TASO client for less than six months was found to be significantly related to desire for children in bivariate analysis but not in multivariate analysis (see Table 3).

Pregnancy risk behavior

Overall, (455) 42% of participants were sexually active within the past three months and 135 (33%) engaged in pregnancy risk behavior. Seventy-three percent of participants currently practicing pregnancy risk behavior did not want more children and were at high risk for unwanted pregnancies. Factors independently associated with pregnancy risk behavior included having no formal education, adjusted OR, 4.09; 95% CI, 1.73–9.67, lack of knowledge of PMTCT programs, adjusted OR, 2.09; 95% CI, 1.25–3.51, having attended TASO for less than six months, adjusted OR, 3.03; CI, 1.68–5.44, and desire for children, adjusted OR, 2.33; CI, 1.25–4.32 (see Table 4).

Discussion

In a population of HIV-positive men and women, we found that among those practicing pregnancy risk behavior, 73% did not want more children and were at high risk for unwanted pregnancies. MTCT knowledge and PMTCT exposure were not significantly related to desire for children. Those who lacked knowledge of PMTCT programs, had been a TASO client for less than two years and desired children were more likely to engage in pregnancy risk behavior. However, MTCT knowledge was found not to be significantly related to pregnancy risk behavior.

It is encouraging that exposure to PMTCT programs and time as a member with an AIDS Support Organization were associated with lower rates of pregnancy risk behavior. Contrary to fears that PMTCT knowledge might increase pregnancy risk behavior, we found that those with high PMTCT knowledge were less likely to engage in pregnancy risk behavior suggesting that information on PMTCT should be disseminated widely. Providing information and support to HIV-infected men and women may have increased their ability not to engage in pregnancy risk behavior. However, although desire to have children was associated with engaging in pregnancy risk behavior, three-quarters of our participants who had pregnancy risk behavior did not desire more children. Our findings predated ART access for this population and experience elsewhere suggests that desire for children may increase with improved health on ART (Panozzo et al., 2003) and that pregnancy rates may go up following initiation of ART (Blair et al., 2004). Access to ART may lead to increased unprotected sexual activity and restored fecundity in our study population as well. The recent shift in government and donor resources and attention away from family planning and reproductive health to HIV/AIDS has undermined progress made on delivery of family planning services and impedes the integration of family planning and HIV. Yet, HIV/AIDS and family planning are intrinsically linked because of the mode of HIV transmission (Stover et al., 2003). All AIDS care and treatment programs, especially those providing ART, need to ensure integration of improved family planning services for those who request services.

Table 4 Associations with pregnancy risk behavior among HIV-infected adults, Jinja, Uganda

Most PMTCT interventions have targeted women. However, men were more likely to desire more children than women, and men are often the decision makers in matters related to reproductive choices. In this environment, providing HIV-infected men and male partners of HIV-infected women with PMTCT and access to family planning information might improve utilization of these services.

The major limitation of this analysis is the use of cross-sectional data which limits our ability to establish causation. Additional longitudinal and interventional research that examines the factors associated with pregnancy risk behavior and evaluates strategies to reduce unwanted pregnancy risk behavior among HIV-infected persons is urgently needed. Sexual partners were not linked, and HIV-negative partners were not interviewed, limiting our ability to assess couple dynamics around desire for children and pregnancy risk behavior. However, we were able to examine participant's perception of their spouses’ desire for children. Self-reported data are difficult to validate, and may be influenced by perception of socially desirable responses; however, we minimized this by training interviewers to build strong rapport and care was taken to ask questions without obvious bias. The study found that only 174 clients (16%) of the entire sample wanted more children. The limited sample size in this group may have affected the results of some of the analyses. Nonetheless, we were able to detect significant associations with important policy implications.

Reproductive health preferences are more likely to translate into practice when there is a strategy or plan in place (Baylies, 2000). Several categories of clients and possible interventions have been identified in this study (see Fig. 1). As a first priority, programs should avail family planning services to those who do not desire children but are currently engaging in pregnancy risk behaviors, as they are at risk for unwanted pregnancies. Secondly, those who desire children and are engaging in pregnancy risk behavior need education on the efficacy of PMTCT interventions and linkages to PMTCT providers, and those who desire children and are not engaging in pregnancy risk behavior need on-going counseling on strategies for minimizing transmission risk while attempting conception and on the efficacy of PMTCT interventions so that they make informed choices about their pregnancy risk behavior. Temporary abstainers who do not desire children need ongoing support and access to family planning in case they resume sexual activity in future, and temporary abstainers who desire children need counseling and information on risks and benefits of family planning as well as implications of having more children. Having been a client for more than two years was associated with lower pregnancy risk behavior rates. This suggests that health care professionals and counselors providing services for HIV-infected men and women could further support them by ensuring availability of family planning and to reduce HIV transmission risk to uninfected partners, dual methods (using condoms in addition to any other family planning method). In addition, service providers can address client concerns about future children's well-being while assisting them to make personal and couple sexual behavior plans. Our findings underscore the importance of ensuring that family planning services are integrated into AIDS prevention, care and treatment services.