The South African (SA) National HIV Prevalence, Incidence, Behavior, and Communication Survey undertaken in 2017 revealed that 13.6% of youth age 15–24 years engaged in sex before age 15 years.1 This figure represents an upward trend from an initial rate of 8.9% in the 2002 national survey.1 Condom use among this same-aged cohort remains suboptimal, with less than 60% of respondents reporting using a condom at last sexual intercourse.1 These risky sexual behaviours are key drivers of the HIV epidemic in South Africa and manifest in high levels of HIV incidence and unintended pregnancies among adolescents.
South Africa's HIV incidence rates are highest among adolescent girls and young women (AGYW), accounting for an estimated 29% of all new infections in 2018,2 with HIV prevalence among adolescents girls aged 15–19 years at 5.8%, and 4.7% among boys in 2017.1 According to Statistics South Africa (StatsSA; 2018) birth rates among adolescent girls have decreased moderately from 78 births per 1000 in 1996 to 65 births per 1000 in 2016 among the 15–19 year age group. However, these birth rates remain high with data revealing that the majority of young girls who become pregnant while attending school often do not return after childbirth, and in instances where young girls do return to school, they face challenges balancing school and childcare.3,4
In response, the SA Department of Basic Education (DBE) developed the HIV, STI, and TB policy, which included the call to educate young girls and boys about sexuality and sexual behaviour in an effort to reduce both unintended pregnancy and HIV rates.5 One of the key interventions pronounced within this policy is the provision of comprehensive sexuality education (CSE). The aim of CSE, as specified by the DBE, is to build and shape learners' understanding of concepts, content, values, and attitudes around sexuality and sexual behaviour.6
Research has established that CSE can improve sexual and reproductive health knowledge,5 and be effective in reducing risky sexual behaviour.7,8 Studies in sub-Saharan Africa have to an extent affirmed these results, indicating that changes in adolescents' sexual behaviour after exposure to these programs are modest, but achievable, while positive improvements in adolescents' attitudes and knowledge were consistently produced.9–11
Studies undertaken in South Africa, focusing on implementation, have revealed challenges at the individual, interpersonal, school, and community levels.12 Some educators remain uncomfortable teaching the sensitive subject matter, whereas many schools lack the teaching support materials required by educators to deliver CSE lessons effectively.12 To ensure a standardised approach to the delivery of CSE and to improve the teaching quality, structured activity plans and training were recommended to advance the content knowledge and pedagogy of educators.13
Since 2015, the DBE have been developing and piloting Scripted Lesson Plans (SLPs) across 5 provinces (KwaZulu-Natal, Free State, Gauteng, Mpumalanga, and Western Cape) to strengthen the teaching of CSE in schools. SLPs are learner and teacher support materials (LTSMs) designed to aid educators and improve the effectiveness of CSE lessons.6 This formed part of the Determined, Resilient, Empowered, AIDS-free, Mentored and Safe (DREAMS) initiative, which aimed to reduce HIV infection among young women. The delivery of CSE using SLPs forms part of the DREAMS package of “layered” evidence-based HIV prevention interventions targeting—biological, behavioural, and structural factors to reduce AGYW vulnerability to HIV with the aim of reducing HIV incidence by 40% among AGYW over a 2-year period (2016–2018).14
In this study, we present data collected as part of the DREAMS evaluation to determine the effect of schooling and exposure to CSE has had on HIV prevalence, pregnancy, risky sexual behavior, and uptake of HIV testing in a sample of adolescent girls. This study adds to a limited knowledge base on school-based HIV programming and its effect on biological outcomes.15 We hypothesize that young women in school will exhibit lower levels of HIV prevalence and risky sexual behaviour, consistent with other studies,9–14,16,17 whereas the additional exposure to CSE is expected to have a compound effect, further reducing risky behavior and exposure to risk of HIV acquisition. The aim of the study is therefore to estimate the association between exposure to CSE and key biomedical and behavioural indicators among adolescent girls in South Africa.
The study was undertaken in South Africa in the districts of Ekurhuleni and City of Johannesburg in the Gauteng province (GP) and the districts of uMgungundlovu and eThekwini in the province of KwaZulu-Natal (KZN). Data were collected from March 13, 2017, to June 22, 2018. More information on the districts has been described in the study's protocol paper.14
A cross-sectional survey targeting 18,500 AGYW (age between 12 and 24 years) was administered. The study sampling method used a multistage cluster design. The districts were considered as the primary strata. The sample size per district was designed to be proportional to the estimated number of AGYW in DREAMS subdistricts. Only small area layers (SAL) where DREAMS was implemented were included in the sampling frame. SALs were selected proportional to size from the sampling frame. Households were randomly selected, and all AGYW residing in the household that met the requirements for the study were enrolled. For our analyses, data were restricted to all adolescent girls age 12–18 years (n = 9673).
Household composition forms were completed for 63,618 houses, of which 18,424 households met the enrolment criteria, whereas 16,845 households agreed to join the study. Both a caregiver and an AGYW questionnaire were completed.
Written consent was obtained from all individuals aged 18 years or older, and parental or guardian consent and individual written assent were obtained from all individuals younger than 18 years. The study was approved by the University of KwaZulu-Natal Biomedical Research Ethics Committee (BFC 198/16) and the Provincial Department of Health in both KZN and GP.
Three independent variables were used in the study. The first was current school attendance. The second was whether respondents, currently out of school, had attended CSE classes in the previous 12 months (responses to the original question regarding number of sexuality education lessons attended ranged from 0 to 99 and was coded into binary of 0 if they did not attend CSE, 1 = did attend one or more CSE lessons in the previous 12 months). The third independent variable assessed whether for those respondents who were currently in school had attended at least 1 CSE lesson in the previous 12 months, coded as 1; and 0 for individuals who were in school but had not attended at least 1 CSE lesson in the previous 12 months.
There were 10 dependent variables. HIV status was determined from finger-prick blood samples into BD Microtainer (Becton Dickinson, South Africa) blood collection tubes for laboratory measurements of HIV antibodies. Sexual activity (defined as inserting an object, finger, or penis into the person's vagina or anus) was measured by asking if respondents had ever had sex. Sexual debut was measured by asking at what age respondents started engaging in sexual activity. Seven questions regarding transactional sex covered the different types of material goods respondents received for engaging in sex with a partner in the previous 12 months. These items were collapsed into a binary variable. Age-disparate sex was measured by asking respondents if any of their sexual partners in the past 12 months were 5 or more years older than them in line with the UNAIDS definition.18 Condom use was measured by asking how often respondents used condoms with their sexual partners in the previous 12 months. We recoded those that said “sometimes” or “never” as inconsistent users and those that said they “always,’ as consistent users of condoms. Multiple sexual partners was measured by asking how many partners individuals had in the past 12 months in an open-ended format, with responses coded into a binary. We asked respondents if they had ever been pregnant. Respondents were asked whether they had ever been tested for HIV. We asked individuals how sure they were in their ability to obtain condoms if needed. All the sexual risk, linkage to services, and pregnancy variables were self-constructed.
We included six control variables in our regression models: age, province, total household income, self-reported caregivers' HIV status, and orphan status; we also included a measure for overall exposure to HIV prevention interventions that was only included in the school attendance model because some of the HIV interventions may overlap with sexuality education, such as HIV testing campaigns.
Only data from adolescent girls age younger than 19 years were analyzed because the majority would be in school and are the target group for CSE. Data analysis was undertaken using SPSS version 27. Our analysis used descriptive statistics and multiple logistic regressions. We present adjusted odds ratios (AOR) with 95% confidence intervals (CI) for the multiple logistic regressions. To adjust for the study design and nonresponse, sampling weights were used. The final sampling weight was the product of the SAL weight, household weight, adjusted for individual nonresponse. The final individual weights were benchmarked against the 2018 Statistics South Africa mid-year population estimates by age group and province.19 In all analyses, standard errors were adjusted for clustering by SAL and stratification by district. Taylor series linearization methods were used to estimate standard errors.
The study response rate at an individual level was 97.8% (409 AGYW refused to partake in the study). More than half of the sample was from Gauteng (58.5%; Table 1). The median age in the sample was 15 years (IQR: 13–17). The majority (89.4%) of respondents were currently in school. Nearly two-thirds (59.9%) of respondents had attended CSE lessons at school in the previous 12 months. More than a quarter (27.9%) of respondents indicated they were a single orphan (ie, one parent is no longer alive). Less than one-tenth (8%) of caregivers reported they were HIV-positive. Overall, 5.1% of all respondents in this study were HIV-positive. Just over one fifth (20.9%) of Adolescent girls had ever had sex. Nearly one-third (30.8%) of respondents had previously been pregnant. Approximately one quarter (25.8%) of respondents who were sexually active had at least one sexual partner who was 5 or more years older than them. Few (6.8%) sexually active respondents had engaged in transactional sex. More than a third (39.8%) of sexually active respondents indicated they had 2 or more lifetime sexual partners.
TABLE 1. -
Description of DREAMS Study Sample (Adolescent Girls Age 12–18 Years, n = 9673)
|Household and caregiver details
| Total family income per mo (USD)*
|Do not know
| Caregiver HIV status
|Adolescent girls' details
| Median age years (IQR)
| Currently in school
| Attended CSE in previous 12 mo.
| Attended CSE and currently in school
| Orphan status
||Not an orphan
| Ever had sex
| Ever been pregnant
| Any sex partner 5 or more years older
| Engaged in transactional sex
| Number of lifetime sex partners
| HIV Status
*One United States Dollar equals 13.3 South African Rands (average for 2018).
The relationship between school attendance and HIV status was not significant for the full or sexually active only sample (Table 2). However, school attendance is associated with a number of sexual risk variables. Attending school is associated with reduced adjusted odds of having engaged in sexual activity (AOR: 0.58, 95% CI: 0.48 to 0.70, P < 0.001); early sexual debut (AOR: 0.62, 95% CI: 0.46 to 0.83, P < 0.01); engaging in transactional sex (AOR: 0.57, 95% CI: 0.37 to 0.90, P < 0.05); engaging in age-disparate sex (AOR: 0.58, 95% CI: 0.45 to 0.75, P <0.001); using condoms inconsistently in the previous 12 months (AOR: 0.48, 95% CI: 0.34 to 0.65, P < 0.001); having 2 or more sex partners in the previous 12 months (AOR: 0.72, 95% CI: 0.57 to 0.92, P < 0.05); and having ever been pregnant (AOR: 0.41, 95% CI: 0.32 to 0.52, P < 0.001).
TABLE 2. -
Multiple Logistic Regression Assessing the Relationship Between School Attendance, CSE Attendance and HIV Status, Sexual Risk, Pregnancy, and Access to Services for Adolescent Girls Age 12–18 Years
||In Schools vs. Out of School
||Out of School Attended CSE vs. Out of School No CSE
||In School and Attended CSE vs. in School and No CSE
|HIV-positive vs. HIV-negative
|HIV-positive vs. HIV-negative (sexually active)
|Had sex vs. did not have sex
|Early sex debut (=< 15 year old) vs older sex debut (16–18 year old)
|Engage in transactional sex vs. not engage in transactional sex
|Did engage in age-disparate sex vs. did not engage in age-disparate sex
|Inconsistent condom use vs. consistent condom use
|2 or more sex partners in the previous year vs. 0–1 sex partner in previous year
|Was/is pregnant vs. never been pregnant
|HIV tested vs not tested
|Sure about obtaining condoms vs unsure about obtaining condoms
*P < 0.05, **P <0.01, ***P <0.001.
All analysis was weighted. Dependent variables were coded in the following way: HIV status (0 = HIV-negative, 1 = HIV-positive), had sex (0 = did not have sex, 1 = have had sex), sexual debut [0 = older sex debut (16–18 year old), 1 = early sex debut (≤15 year old)], transactional sex (0 = did not experience transactional sex, 1 = experienced transactional sex), age-disparate sex (0 = did not engage in age-disparate sex, 1 = did engage in age-disparate sex), inconsistent condom use in the previous 12 months (0 = consistent condom use, 1 = inconsistent condom use), 2 or more sexual partners (0 = 0–1 sex partner in previous year, 1 = 2 or more sex partners in the previous year), pregnancy (0 = never been pregnant, 1 = was/is pregnant), HIV test (0 = no, 1 = yes), and obtain condoms (0 = unsure, 1 = sure). Control variables are included in all models and are the following: age, province, total household income, caregivers HIV status, orphan status, and overall exposure to HIV prevention interventions (only for the school attendance model). Not all models have the same number of control variables because of overfitting of models. See Supplemental Digital Content Appendix, https://links.lww.com/QAI/B827
for full models with all control variables included.
Adolescent girls in school and who had attended CSE in the previous 12 months were associated with reduced adjusted odds of being HIV-positive (Full sample: AOR: 0.76, 95% CI: 0.61 to 0.95, P < 0.05; sexually active sample: AOR: 0.62, 95% CI: 0.40 to 0.96, P < 0.05). Those in school who attended CSE in the previous 12 months were more likely to get tested for HIV (AOR: 1.48, 95% CI: 1.32 to 1.65, P < 0.001) and confident of obtaining condoms (AOR: 1.52, 95% CI: 1.32 to 1.74, P < 0.001).
Adolescent girls currently not in school, but who had attended CSE in the previous 12 months, had an increased adjusted odds of undergoing a HIV test (AOR: 1.72, 95% CI: 1.25–2.37, P < 0.01) and were confident they could obtain condoms (AOR: 2.17, 95% CI: 1.57–3.00, P <0.001).
This study shows a negative association between select biological and sexual behaviour indicators and schooling and exposure to CSE among adolescent girls age 12–18 years across four DREAMS districts in South Africa. Adolescent girls in school were less likely to be engaging in sexual activity, and for those that had, less likely to have started before age 15 years. Adolescent girls in school were also less likely to have been pregnant. Attending school, independent of additional exposure to CSE, was associated with the adoption of less risky sexual behaviour, specifically lower rates of transactional and age-disparate sex, inconsistent condom use, and engagement in multiple partners. These results affirm several studies that have found a positive correlation between school attendance and increased educational attainment and lower HIV incidence in South Africa.8–13,17,20 Identifying the precise causal effect remains challenging,14 with some studies suggesting that adolescent acquire sexual health guidance, regardless of dedicated curricula covering the subject,16,21 with attending school also potentially disrupting sexual networks.22
This study did measure exposure to dedicated CSE curricula; however, for adolescent girls still in school, this did not further amplify the effect on risky sexual behaviour, despite lower HIV prevalence rates within this cohort. Importantly, adolescent girls were no more likely to engage in risky sexual behaviour, countering concerns from teachers, parents, and religious groups within the South African context that exposure to sexuality education will accelerate adolescent interest in engaging in sex.23
The results further indicate that adolescent girls displayed increased self-efficacy with respect to accessing condoms and HIV testing. However, having access to condoms has not translated into condom use in this and other studies.8 Although this study was not able to determine the causal link between reduced risky sexual behaviour and lower HIV prevalence rates, findings echo research in a similar context among youth, establishing that access to HIV testing was associated with lower rates of HIV prevalence.7,9–12,24 Encouragingly, the positive association between exposure to CSE and HIV testing rates were retained in out-of-school adolescent girls.
Evaluations of the effectiveness of CSE have to date revealed mixed results. Although for many adolescents, schools have become a primary site for learning about sex and HIV and AIDS,9 the translation of this knowledge into improved prevention behaviour has been somewhat inconclusive.10,11,24 This could be as a result of the nonstandardized delivery of CSE and the variable access to resources across schools.12 The development and implementation of SLPs, as part of the DREAMS initiative, is expected to address some of these challenges.
Further research should account for the effect of CSE dosage on key biomedical and behavioural indicators. Sexuality education has formed part of the South African curriculum since 2000. Challenges in the delivery of CSE has been noted12; however, no research has managed to determine the effect of the provision of quality and compounded CSE teachings on adolescents. Adolescents, if retained within the schooling system, should be exposed to age-appropriate CSE lessons for 8 years, beginning in grade 4 (approximately from age 10 years). The effect of this exposure is likely to differ based on a number of factors, including the skill of the educator, the pedagogy used, and the attendance and attentiveness of adolescents during those lessons.
CSE within the schooling system should work in tandem with adolescent access to sexual and reproductive health services. Access to contraception and HIV and STI testing remains critical. Implementation of the DBE HIV, STI, and TB policy,6 which makes provision for schools to make condoms available, could potentially reduce risky sexual behaviour, as has been found in a systematic review assessing the effects of condom availability in schools’.25
Analysis for this paper has focused on adolescent girls' exposure to CSE in isolation to the potential positive impact other DREAMS interventions have had on the biological and behavioural indicators measured. Furthermore, the data used are cross-sectional and therefore, causality is difficult to ascertain because young women may have been infected or fallen pregnant before attending CSE. Many of the study variables are self-reported and susceptible to socially desirable response sets. Future studies would benefit from using longitudinal data to assess the impact of prolonged exposure to CSE on risky sexual behaviour.
The ability to determine the fidelity of CSE implementation in schools is beyond the scope of this study. The quality of CSE lessons is likely to be differential across schools, districts, and provinces, with other studies unearthing implementation challenges.26 The fidelity of CSE implementation is likely to play an important role in preventing risky behaviour and linking adolescent girls to appropriate services should be measured in future studies.
This is the first community-based study measuring the association between schooling and exposure to CSE and key biological and behavioural indicators among adolescent girls. This study has revealed positive evidence affirming that schooling and CSE are associated with a lower likelihood of HIV acquisition. Keeping adolescent girls in school is therefore paramount, whereas continued exposure to quality CSE is key to increase health service utilization, ensuring a further reduction in HIV risk and improved sexual and reproductive health management.
The authors thank all the coinvestigators and members of the study team from the following organizations: Epicentre AIDs Risk Management (Pty) Limited (Epicentre), Health Economics and HIV/AIDS Research Division (HEARD), Centre for AIDS Programme of Research in South Africa (CAPRISA), National Institute of Communicable Diseases (NICD), and US Centers for Disease Control and Prevention (CDC). The authors thank the DREAMS collaborating partners: The National Department of Health, Provincial KwaZulu-Natal Department of Health, the Provincial Gauteng Department of Health, uMgungundlovu Health District, eThekwini Health District, Ekurhuleni Health District, City of Johannesburg Health District, the uMgungundlovu District AIDS Council, local municipal and traditional leaders, and community members for all their support. A special thanks to the study staff for the field work, and laboratory and Primary Health Care clinic staff in the study districts.
1. Simbayi L, Zuma K, Zungu N, et al. South African National HIV Prevalence, Incidence, Behaviour and Communication Survey, 2017: Towards Achieving the UNAIDS 90-90-90 Targets. Cape Town: HSRC press; 2019.
2. Global UNAIDS. AIDS Update: Seizing the Moment: Tackling Entrenched Inequalities to End Epidemics. Geneva, Switzerland: Joint United Nations Programme on HIV/AIDS (UNAIDS); 2020:2020.
3. Gyan C. The effects of teenage pregnancy on the educational attainment of girls at Chorkor, a suburb of Accra. J Educ Soc Res. 2013;3:53.
4. Nkosi NN, Pretorius E. The influence of teenage pregnancy on education: perceptions of educators at a secondary school in Tembisa, Gauteng. Social Work. 2019;55:108–116.
5. Kemigisha E, Bruce K, Ivanova O, et al. Evaluation of a school based comprehensive sexuality education program among very young adolescents in rural Uganda. BMC Public Health. 2019;19:1–11.
6. Department of Basic Education. National Policy on HIV, Sexually Transmitted Infections and Tuberculosis. Pretoria; 2017.
7. Rosenberg NE, Westreich D, Bärnighausen T, et al. Assessing the effect of HIV counseling and testing on HIV acquisition among South African youth. AIDS. 2013;27:2765.
8. Sayles JN, Pettifor A, Wong MD, et al. Factors associated with self-efficacy for condom use and sexual negotiation among South African youth. J Acquir Immune Defic Syndr. 2006;43:226.
9. Bhana A, Brookes H, Makiwane M, et al. Evaluation of the Impact of the Life Orientation Programme on HIV/AIDS in Gauteng Schools: Pilot Study. Pretoria: Human Sciences Research Council; 2005.
10. Department of Education. National Evaluation of the Impact of the DoE HIV & AIDS Life Skills Programme on the Behaviour of Learners. Pretoria; 2006.
11. Reddy P, James S, McCauley AP. Programming for HIV Prevention in South African Schools: A Report on Program Implementation. Washington, DC: Population Council; 2005.
12. George G, Tucker LA, Panday S, et al. Challenges facing life orientation educators in the delivery of sexuality education in South African schools. South Afr Rev Edu Edu Prod. 2018;24:43–57.
13. Adams Tucker L, George G, Reardon C, et al. “Learning the basics”: young people's engagement with sexuality education at secondary schools. Sex Edu. 2016;16:337–352.
14. George G, Cawood C, Puren A, et al. Evaluating DREAMS HIV prevention interventions targeting adolescent girls and young women in high HIV prevalence districts in South Africa: protocol for a cross-sectional study. BMC Women's Health. 2020;20:1–11.
15. UNAIDS. The Effectiveness of Behavioural Interventions to Prevent HIV. A Compendium of Evidence (Ch. V). Geneva, Switzerland: HIV prevention education and comprehensive sexuality education; 2019.
16. Özler B. Keeping Girls in School: A Review of the Global Evidence. Washington, DC: World Bank; 2016.
17. Stoner MC, Pettifor A, Edwards JK, et al. The effect of school attendance and school dropout on incident HIV and HSV-2 among young women in rural South Africa enrolled in HPTN 068. AIDS. 2017;31:2127.
18. UNAIDS. UNAIDS Terminology Guidelines. Geneva, Switzerland: UNAIDS; 2015.
19. Statistics South Africa. Mid Year Population Estimates 2018. Pretoria, Statistics South Africa; 2018.
20. Ntombela NP, Kharsany ABM, Soogun A, et al. Prevalence and risk factors for HIV infection among heterosexual men recruited from socializing venues in rural KwaZulu-natal, South Africa. AIDS Behav. 2021;25:3528–3537.
21. Wado YD, Sully EA, Mumah JN. Pregnancy and early motherhood among adolescents in five East African countries: a multi-level analysis of risk and protective factors. BMC Pregnancy Childbirth. 2019;19:59.
22. Hargreaves JR, Morison LA, Kim JC, et al. The association between school attendance, HIV infection and sexual behaviour among young people in rural South Africa. J Epidemiol Community Health. 2008;62:113–119.
23. Letsoalo I. Why Comprehensive Sex Education in South African Is So Important. Melbourne: Global Citizen; 2019.
24. Visser MJ. Life skills training as HIV/AIDS preventive strategy in secondary schools: evaluation of a large-scale implementation process. J Soc Aspects HIV/AIDS. 2005;2:203–216.
25. Wang T, Lurie M, Govindasamy D, et al. The effects of school-based condom availability programs (CAPs) on condom acquisition, use and sexual behavior: a systematic review. AIDS Behav. 2018;22:308–320.
26. Speizer IS, Mandal M, Xiong K, et al. Impact evaluation of scripted lesson plans for HIV-related content in a life orientation curriculum: results from two provinces in South Africa. BMC Public Health. 2020;20:1–16.