Advancing Sexual and Reproductive Health and Rights
International Perspectives on Sexual and Reproductive Health
Volume 36, Number 4, December 2010

Why Is Kenya's Fertility Rate Still High? HIV Epidemic May Be a Factor

Why Is Kenya's Fertility Rate Still High? HIV Epidemic May Be a Factor

The HIV epidemic may have contributed to the lack of decline in Kenya's fertility rates, primarily through its associations with elevated child mortality and reduced duration of breast-feeding, according to an analysis of Demographic and Health Survey (DHS) data.1 Although having AIDS was associated with reduced levels of fertility, women were less likely to want to stop childbearing and more likely to have had a recent birth if a young child of theirs had died—a common scenario in AIDS-ravaged areas. Moreover, the mean duration of breast-feeding in a community was inversely associated with recent childbearing, suggesting that women who opted not to breast-feed because of fears of mother-to-child HIV transmission were putting themselves at increased risk for pregnancy.

According to the 2003 DHS, Kenya's fertility rate was about five births per woman, indicating that the decline in fertility that occurred through the mid-1990s had stalled. The prevalence of HIV was also high—about 7% overall, and nearly 9% among women, in 2003. In principle, high HIV prevalence might contribute to lower fertility by discouraging sexual activity and spurring condom use, as well as through various biological mechanisms (rates of fetal loss and infertility are elevated among women with AIDS). However, the epidemic may help keep fertility rates high because couples might choose to have extra children (to replace those who have died of AIDS or as insurance against losing future children). Moreover, both infected women and those who do not know their HIV status—which in Kenya includes most women—may refrain from breast-feeding to avoid transmitting the virus to their infants, and thus may be at increased risk for pregnancy because they do not benefit from breast-feeding's contraceptive effects.

To explore these competing scenarios, investigators analyzed 2003 DHS data on a nationally representative sample of 8,195 wom-en aged 15–49 and 3,578 men aged 15–54. First, they sought to identify factors associated with two aspects of women's fertility preferences: whether women wanted more children and whether their ideal family size was three or fewer children. Next, the investigators examined predictors of recent fertility (i.e., whether a woman had had a birth in the past three years). Factors of interest in both analyses included demographic variables (educational attainment, parity, union status), experiences of child loss (loss of a child younger than five, fetal loss), HIV-related measures (serostatus, perceived HIV risk, level of HIV knowledge, knowing a person who has or has died of AIDS) and community-level factors (proportions of local residents who have experienced a child loss, have HIV, perceive their HIV risk as moderate or high, or know someone who has or has died of AIDS). Most of these measures were based on women's survey responses, but data from both men and women were used to calculate the community-level measures. The researchers conducted bivariate and multivariate regression analyses.

Overall, 42% of the women wanted to have no more than three children. The proportion was greater among women who had never been tested for HIV than it was among those who had been tested (57% vs. 40%), and greater among the quartile who scored highest on the HIV/AIDS knowledge index than among those in the lowest quintile (55% vs. 28%). In the multivariate analysis, the desire for a small family was associated with having a primary or higher education (regression parameter estimates, 0.3–0.9) and with higher community levels of HIV/AIDS knowledge (0.8); women who lived in areas with higher levels of child mortality were less likely than other women to want three or fewer children (–0.5). Ideal family size was not associated with having ever lost a child or fetus, HIV status or knowing someone with AIDS.

Thirty-six percent of all women said they did not want any children in the future. In the bivariate analysis, this desire varied strikingly according to perceived HIV risk: Almost half (46%) of women who considered themselves at high risk did not want a child, compared with 30% of those with little or no perceived risk. In the multivariate analysis, however, desire for children was not associated with perceived risk, nor with actual HIV status or knowing someone with AIDS. Rather, the desire to stop childbearing was positively associated with having any degree of schooling (parameter estimates, 0.1–0.2), having any children (0.2 for all parities above 0) and living in a community with higher scores on the HIV/AIDS knowledge index (0.1). Women were less likely to want to stop childbearing if they had lost a child younger than five (–0.5) or had ever experienced fetal loss (–0.2).

Thirty-six percent of all women, and 46% of sexually active women, had had a birth in the past three years. As in prior studies, women who had HIV and those who knew someone with AIDS were less likely than other women to have had a recent birth (parameter estimates, –0.4 and –0.2, respectively). However, recent fertility was negatively associated with mean duration of breast-feeding in a community (–0.1), indicating that fertility rose when breast-feeding declined. Moreover, having lost a child aged five or younger at least three years before the survey was positively associated with a recent birth (0.3).

Although the data cannot show causality, the findings are consistent with the possibility that HIV/AIDS has had, and may still be having, diverse effects on fertility in Kenya. On the one hand, the authors note, the epidemic may have contributed to reduced fertility because women with AIDS tend to refrain from childbearing and may be less fertile than other women. Conversely, the epidemic may have contributed to increased fertility through increased infant and child mortality (leading to replacement births) and reduced duration of breast-feeding. "It is possible," the researchers contend, that "in communities at advanced stages of the HIV/AIDS epidemic," levels of child mortality and declines in breast-feeding have become "substantial enough to result in a reversal of fertility decline." They add, however, that because factors such as lack of progress in economic development can also keep fertility at high levels, the HIV epidemic should be viewed as one, rather than the sole, factor that may explain recent fertility trends in Kenya.

P. Doskoch


1. Magadi MA and Agwanda AO, Investigating the association between HIV/AIDS and recent fertility patterns in Kenya, Social Science & Medicine, 2010, 71(2): 335–344.