The knowledge-risk-behaviour continuum among young Ugandans: what it tells us about SRH/HIV integration

A cross-sectional household survey was conducted in Uganda between February and March 2016. The study targeted a nationally-representative sample of households calculated based on the UN Department of Economic and Social Affairs’ practical guidelines on designing household survey samples [19]. The overall household sample size was 2976 households. Sample size calculations were based on the following assumptions: confidence level set at 95%, percentage of young people age 10–24 who have comprehensive knowledge about HIV/AIDS set at 40%, design effect of 2.0, non-response rate of 20%, a margin of error/precision of 10% and an estimated average 1.48 respondents aged 10–24 years per household.

Given the research questions selected for this paper which focus on sexual activity and SRH/HIV knowledge, the study sub-category aged 15–24 years comprised of 2725 respondents was selected. Young Ugandans aged 10–15 years were excluded given their reduced exposure to sexual activity and to SRH/HIV information as per national guidelines on sexuality education [20].

The first stage of sampling involved a random selection of 44 districts from the sampling frame of all districts in Uganda which were proportionately allocated to the 10 Uganda Bureau of Statistics (UBOS) statistical regions according to their total populations. Kampala was however purposively sampled due to its unique socioeconomic and demographic characteristics. The second stage involved selecting enumeration areas (EA) using the UBOS 2014 National Population and Housing Census sampling frame [21]. A total of 149 EAs (20 households per EA) were selected and allocated proportionately to the 44 sampled districts, so that districts with larger populations were allocated more EAs. The third stage involved assigning a number to each EA household and randomly selecting the ones to be interviewed by applying a numeric interval. A questionnaire was administered to all eligible and consenting persons aged 10–24 years found in the selected households.

For the present paper, measurements of interest are (a) family planning (FP) knowledge measured by awareness of at least one modern contraceptive method, (b) knowledge of sexually transmitted infections (STIs) measured by having ever heard about any infections that people can get from sexual intercourse, (c) comprehensive knowledge of HIV/AIDS measured by a composite indicator of awareness about HIV, knowledge about its transmission and correctly rejecting three out of five common misconceptions about HIV/AIDS, (d) self-reported perceived risk of being infected with STIs including HIV, and (e) self-reported perceived risk of becoming pregnant or impregnating a girl being rated as “at risk” versus “not at risk”. The outcome of interest being risky sexual behaviour was defined by a composite indicator of behaviour that increases one’s risk of contracting STIs (including HIV) and/or unintended pregnancies, namely being involved in sexual relations under the influence of alcohol or drugs, engaging in sexual activities without a condom, and having multiple sexual partners in the 6 months preceding the survey.

Data was analysed using STATA 15. Exploratory analysis was conducted to provide descriptive statistics for adolescent’s socio-demographic characteristics. Logistic regression was conducted to determine the association between background characteristics of young people, their SRH/HIV knowledge, perceived SRH/HIV risk and being involved in risky sexual behaviour. Statistical analysis was conducted both at bivariate and multivariate level. All variables explored at the bivariate level were retained at multivariate level in order to document those with potential for correlation.

Ethical approval for the study was obtained from Mildmay Uganda Research Ethics Committee and research approval from the Uganda National Council of Science and Technology. Due to the young age of the group targeted by the study and the sensitivity of the research topic, a number of ethical considerations related to the appropriateness of the research and the confidentiality, privacy and consent of the respondents arose.

The research questions and data collection tools used in the study were adapted from existing studies validated for use in the context where this study was conducted. In addition, they were discussed with relevant stakeholders and adapted to the existing national guidelines on age-appropriateness [20]. Potential risks related to the sensitivity of information provided by respondents were mitigated by the use of anonymous identifiers in form of study numbers on the questionnaires, thus de-linking the data collected from specific individuals. In addition, all respondents were interviewed individually in a private place within the homestead, ensuring parents/guardians did not listen in on the discussions. Finally, participation in the research was based on full informed consent of respondents, including verbal consent from parents/legal guardians for those less than 18 years of age.

A total of 2725 young people aged 15–24 years were interviewed, representing a 98% response rate. Background characteristics of the respondents by age group are given in Table 1. Majority in both age groups (81.8%) were from rural areas. Almost half (46.5%) of the 20–24 year old respondents were ever married or cohabitated, compared to 9.8% of the respondents aged 15–19 years. Moreover, 75.9% of 20–24 year old respondents were out of school, when compared to 28% of those aged 15–19 years. It was also observed that 10.9% of younger respondents were household heads or spouses. Finally, 17.7% of respondents across both age groups did not have a filial or marital relationship to the household head.

Table 2 shows distribution of respondents’ SRH and HIV knowledge and risk indicators by age group. Overall, knowledge levels of FP, STIs and HIV across all young people were high (above 82%), with the older age group (20–24 years) displaying consistently higher levels of knowledge (7% higher on average) than the younger age group. Awareness of at least one modern FP method was high for both age groups, at an average of 94%. No major differences were observed between knowledge of STIs and comprehensive knowledge of HIV with a similar average of 85.6 and 86.2% across respondents for both indicators. Self-perceived risk of STIs (including HIV) and pregnancy among all respondents was high, with about half of all respondents reporting to feel at risk. For 20–24 year old respondents, self-reported perceived risk was consistently higher, with 61.5% feeling at risk of STIs compared to 46.2% of 15–19 year olds. Further, a combined proportion of 22.7% of respondents reported to have been involved in risky sexual behaviour exposing them to STI/HIV and/or pregnancy. Prevalence of risky sexual behaviour among the older group (34.6%) was more than double that of 15–19 year olds (14.9%).

Being involved in risky sexual behaviour was regressed against background characteristics, SRH knowledge and self-perceived risk to SRH problems. Tables 3 and 4 show odds ratio estimates at bivariate and multivariate levels respectively for 15–19 and 20–24 year old respondents. For the 15–19 years group, factors associated with risky sexual behaviour after correcting for differences in age included being female versus male (aOR: 1.89, 95% CI: 1.37–2.61), being out of school versus being in school (aOR: 1.44, 95% CI: 0.87–2.41), having one parent alive versus having both (aOR: 1.56, 95% CI: 1.08–2.24), being involved in casual works versus being a student (aOR: 2.08, 95% CI: 1.27–3.39), being knowledgeable about FP methods (aOR: 3.46, 95% CI: 1.47–8.14), as well as the respondents’ self-perceived risk of being infected with STIs (aOR: 2.88, 95% CI: 1.72–4.82) (Table 3).

For the 20–24 years old respondents, significant predictors of risky sexual behaviour included living in a rural area (aOR: 1.52, 95% CI: 1.03–2.24), being female versus male (aOR: 2.10, 95% CI: 1.56–2.81), having some primary versus secondary education (aOR: 1.46, 95% CI: 1.07–2.01), not having a filial or marital relationship with the head of household (aOR: 1.55, 95% CI: 1.03–2.34), and self-perceived risk of pregnancy (aOR: 1.56, 95% CI: 1.01–2.43) (Table 4).