Nairobi's Poorest Women Have Highest Level of Risky Sexual Behavior, Least Knowledge of HIV Prevention

D. Hollander

First published online:

Women living in the slums of Nairobi engage in riskier sexual behavior than women living in less-deprived areas of the city.1 They began having intercourse at a younger age and are more likely to have had multiple partners in the recent past. Residents of slums and other women are equally familiar with basic facts about HIV and AIDS, but knowledge of how to prevent HIV infection is markedly lower among the most disadvantaged group. These are among the key findings of an analysis pooling data from the 1989, 1993 and 1998 Kenya Demographic and Health Surveys.

The analysts pooled data from the three surveys to obtain a large enough sample to represent both Nairobi overall and its slum areas; in all, they had information on 1,645 women aged 15-49. They classified women who reported having neither running water, electricity nor a flush toilet in their household as slum residents, and those who had all three as nonslum residents; women reporting one or two of these amenities were classified in an intermediate category. Using a variety of analytic strategies, the researchers compared the risk-related behavior of slum residents and other Nairobi women; multivariate analyses examined independent associations between residence and risky behaviors by controlling for women's education, age, childhood residence, length of time in their current home, religion and marital status, and the sex of the head of their household.

Ninety-four percent of women living in slums were sexually experienced, compared with 79% of nonslum dwellers. Among 15-24-year-olds, the proportions who had ever had sex were 89% and 58%, respectively. Women living in slums had begun having intercourse earlier (at age 15.6, on average) than those in the least-disadvantaged households (17.6).

Using life-table calculations, the analysts found that overall, 14% of slum residents and 5% of nonslum residents had begun having intercourse by age 12; by 14 years of age, 55% and 23%, respectively, were sexually experienced. By the time they were 18 years old, 90% of women living in slums and 62% of those from nonslum households had had intercourse. The median age at first intercourse was 15.0 years for slum residents and 18.0 years for nonslum residents.

To control for potential bias from responses of women who migrated from rural areas or small towns, the analysts performed separate life-table calculations for women who had grown up in cities. These results showed even starker contrasts by residence than the findings for the overall sample. In this subgroup, 18% of slum residents and 3% of nonslum dwellers were sexually experienced by age 12; 65% and 10%, respectively, had had intercourse by age 14. Virtually all women who had grown up in cities and were living in slums when surveyed had had sex by age 18, compared with half of those living in nonslum dwellings. The median age at first intercourse was five years younger for slum residents (15.0) than for nonslum residents (20.0).

Findings from hazards analyses confirm that residence is significantly associated with age at first intercourse. Overall, women living in nonslum areas were about half as likely as slum residents to be sexually experienced (hazards ratio, 0.5); among those who grew up in a city, the differential was even greater (0.3). Results were similar when the analyses were restricted to women aged 15-24. In all of the analyses of sexual initiation, findings for women living in intermediate circumstances lay between those for slum and nonslum residents.

Data from the 1993 and 1998 surveys permit analysis of women's number of sexual partners in a given period (the preceding six and 12 months, respectively). Six percent of all women had had more than one partner during the specified interval; the proportion ranged from 4% of nonslum dwellers to 9% of slum residents. The pattern was essentially the same among married women and among 15-24-year-olds: In both subgroups, the proportion reporting multiple partners was 4% overall and ranged from about 2% of women in the least-deprived households to 9% of those living in slums.

Odds ratios from logistic regression analysis indicate that residence in a slum was associated with having multiple partners even when other background factors were taken into account. Compared with nonslum residents, women living in slums had a significantly increased likelihood of having recently had multiple partners (odds ratio, 2.8). A significant, and large, association was also evident for the youngest women (6.4), while for married women, the odds ratio was intermediate (3.8) and only of marginal significance. Living in a household with one or two essential amenities was not significantly associated with the odds of having had multiple partners.

Finally, the analysts examined 1993 and 1998 data on women's knowledge about HIV and AIDS, and how to prevent infection. Regardless of the residential context, virtually all women knew about the disease; across residential categories, similar proportions had a basic familiarity with condoms (75-83%), and similar proportions knew that a person who looks healthy may have HIV or AIDS (89- 92%). However, knowledge of ways to avoid HIV infection was significantly higher among women living in the least-disadvantaged circumstances than among slum dwellers: The proportions citing abstinence and monogamy as protective measures were 76% and 53%, respectively; 51% and 36% knew that condom use could prevent infection. Overall, 90% of nonslum residents identified one of these protective measures, compared with 68% of slum dwellers.

The analysts acknowledge that their findings, based on a proxy measure of residential context, need to be confirmed by "direct evidence from surveys in urban slums." They add that it is "crucial to establish whether unsafe sexual practices are...truly a consequence of slum living or...a consequence of poverty in general." Nevertheless, they conclude that the widening disparity between the rich and poor in African cities has "obvious implications for measures to control the spread of HIV/AIDS."--D. Hollander


1. Zulu EM, Dodoo FN and Ezeh A-C, Sexual risk-taking in the slums of Nairobi, Kenya, 1993-1998, Population Studies, 2002, 56(4):311-323.