In Malawi, the Desire to Have a Child Falls After a Positive HIV Test

John Thomas

First published online:

After learning that they had tested positive for HIV, Malawian men and women reported a substantially reduced desire to have a child in the future, according to a longitudinal study.1 The proportion of respondents who said they wanted another child declined from 60% to 19% among those who had learned that they were HIV-positive, while it dropped to a much smaller degree—from 51% to 30%—among those who had not received a positive result. In interviews, the dominant concern of HIV-positive women who wanted to stop childbearing was fear that having a child would undermine their health, whereas HIV-positive men who wanted to stop having children believed that they would not live long enough to provide for those children or benefit from having them.

The relationship between HIV status and fertility preferences is receiving increased attention from researchers, but the pathways linking infection and childbearing remain poorly understood. This study examined how the fertility preferences of men and women in Malawi were affected after they learned that they had the virus. Because both its fertility rate and its HIV prevalence are high, Malawi is well-suited for such a study. Furthermore, HIV testing and counseling became widely available in 2004 in district hospitals and subsequently in rural hospitals and clinics. Data were from the Malawi Diffusion and Ideational Change Project (MDICP), in which 1,521 ever-married women and their husbands were randomly selected from 120 villages across the country and were surveyed in 2001, 2004 and 2006. Respondents did not know their HIV status prior to the 2004 survey, when the project offered participants HIV testing and counseling; they reported their fertility preferences in the 2001 and 2006 surveys. Data were analyzed using difference-in-differences logistic regression models, both with and without propensity score matching of respondents; the difference-in-differences approach allowed for comparison of changes in fertility preferences between respondents who had received positive HIV test results and those who had not received their results, had tested negative or had not been tested, while propensity matching ensured that the demographic and sexual history characteristics of HIV-positive respondents were similar to those of other participants. In addition, the researchers analyzed qualitative data collected from 23 in-depth interviews with HIV-positive men and women who either were MDICP participants or had attended a clinic for the prevention of mother-to-child transmission of HIV.

In 2001, the mean ages of surveyed women and men were 32 and 40, respectively, and respondents had an average of four living children. One in five women had at least a primary education, as did two in five men; nearly all respondents were married. By 2006, 5.5% of women and 2.5% of men had tested positive for HIV and received their test result. Compared with respondents who had not received a positive HIV test result by 2006, those who had were more likely to be female or unmarried and to suspect that their spouse had had an extramarital relationship; in addition, they had fewer living children and had been married a greater number of times. However, following propensity score matching, these differences were no longer statistically significant; similarly, no differences were apparent when the sample was divided by gender.

Between 2001 and 2006, the proportion of women who said they wanted to have a child in the future declined from 53% to 28%, while the proportion of men who wanted a child fell from 50% to 31%. The proportion who said they wanted a child declined by 41 percentage points (from 60% to 19%) among respondents who had received a positive HIV test result, but by only 21 percentage points (from 51% to 30%) among those who had not received a positive result.

In the multivariate regression analyses, which controlled for gender, age, education, marital status, number of living children and region, respondents who had received a positive HIV test result were less likely than those who had not received such a result to report that they wanted to have a child in the future (odds ratio, 0.3); among men, the odds were even lower (0.1). The findings were similar when propensity score matching was included in the regression models: In both the combined and men-only models, the odds of wanting a child were reduced among respondents who had received a positive HIV test result (0.4 and 0.2, respectively). The women-only models found no differences in the likelihood of wanting to have children.

Qualitative data largely supported the regression findings. In general, both men and women who had received a positive HIV test result said they did not want a child, but their motivations differed. Most of the interviewed women wanted to stop childbearing because of concerns for their own health, while some also feared for the health of a future child; a small minority said they wanted to continue having children because they believed that doing so would allow them to live a "normal" life until they became ill. Men's motivations for not wanting children were considerably different from women's: They believed that because they would not live long enough to provide for these children or to benefit from them (e.g., as a source of pride or future financial support), there was little reason to have them. Men also believed that these children were likely to die within a few years.

This study was unique in examining change in fertility preferences over time and in assessing both quantitative and qualitative data. Yet the researchers noted several limitations of the study. First, the MDICP sample was not nationally representative; female respondents were older and nearer to the end of their childbearing years than the typical Malawian woman of reproductive age. Second, in the five-year period between the 2001 and 2006 surveys, a variety of factors may have influenced respondents' fertility desires, although propensity score matching should have limited the effect of any possible bias. Third, the attrition rate was higher among respondents who had tested positive than among those who had tested negative; however, among the former, attrition was higher among those who had not received their results than among those who had. Nonetheless, the study found that when Malawians learn that they are HIV-positive, many "may plan to have fewer children or decide to stop childbearing altogether," and the researchers note that "in the high-fertility, high-HIV-prevalence context of rural Sub-Saharan Africa…this shift will have large demographic, epidemiological, and reproductive health implications."

—J. Thomas


1. Yeatman S, HIV infection and fertility preferences in rural Malawi, Studies in Family Planning, 2009, 40(4): 261–276.