Advancing Sexual and Reproductive Health and Rights
International Perspectives on Sexual and Reproductive Health
Volume 36, Number 3, September 2010

Predictors of Condom Use Among Kenyan Adolescents Are Wide-Ranging, Vary by Gender

Individual-, community- and school-level factors all play a role in whether young people use condoms, according to a study of upper primary students in Kenya.1 The odds of having used a condom at last sex were elevated among both male and female adolescents who had experienced pressure to have sex from a greater number of sources (odds ratios, 1.3 and 1.8, respectively) or had high condom self-efficacy (1.4 and 1.3). Other individual characteristics were predictive for members of only one sex. For example, among males, condom use was positively associated with having talked to male relatives about AIDS (1.2) and negatively associated with believing oneself to be at low risk (as opposed to no risk) for AIDS (0.8); among females, self-efficacy for abstaining from sex was associated with reduced odds of condom use (0.8). Community- and school-level factors associated with condom use were living in a community where HIV programs are incorporated into festivals (among females) and attending a school that has high levels of resources or a Catholic or mainline Protestant sponsor (among males).

In October 2003, researchers surveyed 3,645 sexually experienced sixth- and seventh-grade adolescents from 160 randomly selected schools in Nyanza Province. The students were asked about their religious affiliation, age, ethnicity (Luo, Kisii or other) and socioeconomic status, as well as the gender of any relatives with whom they had talked about AIDS and whether they had sought information about HIV and AIDS (e.g., by talking to a teacher or reading about the topic). The survey also ascertained adolescents' knowledge of HIV transmission (measured according to their agreement with six factual statements and six local myths), the number of ways in which the adolescents had felt pressured to have sex (including pressure from peers, pressure during dating, offers of money or gifts, or physical force), the degree to which they felt at risk for AIDS, and their abstinence and condom-use self-efficacy (i.e., their confidence in their ability to abstain from sex and to use condoms).

Data on community and school variables were collected in 2002 through observation and through interviews with tribal and religious leaders, parents, women's group leaders and others. For each community, the researchers assessed the dominant messages about condom use conveyed by religious leaders—whether they discouraged condom use or provided misinformation, encouraged use only among adults, or also encouraged use among sexually active youth. Other community-level variables were awareness of ways in which HIV could be transmitted through the community's cultural practices or youth-oriented events, commitment to HIV prevention (whether prevention-related information or messages were incorporated into festivals or other community events) and the availability of condoms at local clinics. School variables included socioeconomic status (determined by the school's physical characteristics and resources), religious sponsorship (Catholic, mainline Protestant, independent or none) and the availability of a box in which students could, as part of an HIV-prevention program, deposit anonymous questions to be answered. Multilevel logistic regression was used to identify individual-, school- and community-level predictors of condom use.

The mean age of participants was slightly younger than 15. Most students (57% of males and 63% of females) were of Luo ethnicity; 45–50% were Protestant and 47–48% Catholic. Overall, the sample was representative of the region's sixth- and seventh-grade students.

In 46% of communities, the dominant messages from local religious leaders about condoms were negative or based on misinformation and myths; messages were supportive of condom use among youth in only 21%. Although traditional practices and social events that put youth at risk for HIV were identified by leaders in most communities (80%), HIV-prevention programs or messages were incorporated into festivals in 78% of communities, and the same proportion had condoms available at clinics. Some 76% of schools had religious sponsorship, and 69% had a question box.

Females scored higher than males on measures of HIV knowledge, pursuance of information about AIDS and abstinence self-efficacy. Males had higher condom self-efficacy scores than did young women, and they were more likely than females to consider themselves at risk for HIV (67% vs. 57%). A greater proportion of males than females reported having discussed AIDS with a male relative, while a greater proportion of females than males had discussed the subject with a female relative.

In the multivariate analysis, both males and females had elevated odds of condom use at last sex if they reported higher condom self-efficacy (odds ratios, 1.4 and 1.3 per one-unit increase in scores, respectively) or had experienced pressure to have sex from a greater number of sources (1.3 and 1.8). Other factors associated with condom use differed by gender. Among males, condom use was more likely if the adolescents were older (1.2 per each additional year); were Kisii rather than Luo (1.6); and had talked to more male relatives about AIDS (1.2 for each additional male relative). Males' odds of condom use were lower, however, if they were Protestant rather than Catholic (0.7); were unable to identify myths about AIDS (0.9); attended a higher-resource school (0.96 per one-unit increase in score); attended a school with a Catholic (0.5) or mainline Protestant (0.6) sponsor, rather than no sponsor; or believed themselves to be at low risk, rather than no risk, for AIDS (0.8).

Among females, living in a community where HIV programming is incorporated into festivals was associated with increased odds of condom use (odds ratio, 1.4), while being of Kisii, rather than Luo, ethnicity (0.6) and reporting self-efficacy for abstaining from sex (0.8) were associated with decreased condom use.

Further analysis revealed that the multivariate models were adequate for explaining the variance in condom use among young men, but did not fully capture all of the school- and community-level factors that influence condom use among females. The latter finding, according to the researchers, "may, in part, reflect male control of condoms." They posit that the range of individual-, community- and school-level influences on condom use, as well as the lack of association between AIDS-related knowledge and condom use, indicate a need for a multipronged approach to HIV prevention. They suggest that "merely providing factual information is insufficient: Myths must be addressed, confidence must be built in one's ability to use condoms, and family members must be involved in communicating with youth about AIDS." Moreover, the associations between condom use and community factors indicated that programs should not only target individuals at risk for HIV but also consider "interventions aimed at entire communities."—H. Ball


1. Maticka-Tyndale E and Tenkorang EY, A multi-level model of condom use among male and female upper primary school students in Nyanza, Kenya, Social Science & Medicine, 2010, 71(3):616–625.