“Condoms Are Boring”: Navigating Relationship Dynamics, Gendered Power, and Motivations for Condomless Sex Amongst Adolescents and Young People in South Africa

Abstract

Condoms remain an important method for preventing HIV prevention and unintentional pregnancies, however their use in South Africa is sub-optimal. We analyzed survey data on reported condom use among 3009 sexually active adolescent girls and young women (AGYW) aged 15–24 years, and qualitative data from interviews and focus group discussions with 237 AGYW and 38 male peers. Our findings describe the current condom use landscape among adolescents and young people in South Africa, illustrating relationship dynamics, gendered power and notions of masculinity which influence condom negotiation and use in young heterosexual South Africans’ sexual encounters.

Introduction

Adolescent girls and young women (AGYW) aged 15–24 years comprise 10% of the population in sub-Saharan Africa; however, they accounted for 20% of new HIV infections in 2017 (UNAIDS, 2019). South Africa has the largest HIV epidemic in the world and a quarter of all new infections occur amongst AGYW aged 15–24 years (UNAIDS, 2019). The disproportionate HIV risk faced by AGYW can be attributed to a number of structural, contextual, and socio-cultural factors including gender inequality, gender-based violence, and gender norms, which combine to negatively impact the ability of AGYW to protect themselves from HIV and other STIs, prevent unintended pregnancy, seek health services, and make informed decisions about their sexual and reproductive health and lives (UNAIDS, 2019). As with rates of HIV, South Africa also has high rates of teenage pregnancy. In 2016, 9% of women aged 15–17 years and 16% of women aged 15–18 years had begun childbearing (Simbayi et al., 2019).

The HIV prevention context in South Africa is dynamic, with new prevention options available, increased access to knowledge and safe sex commodities, such as internal/external (“male”/“female”) condoms (Duby, 2020), and shifts in public discourse around HIV (Atujuna et al., 2018). Despite the development and roll out of new biomedical HIV prevention products such as pre-exposure prophylaxis (PrEP), and microbicides, condoms are the most cost-effective and widely available prevention method, and remain a critical item on the expanded “menu” of HIV prevention product choices (Atujuna et al., 2018; Duby et al., 2017; Irungu et al., 2020; Shrader et al., 2020). In the South African context, condoms are the only free and available technology to provide triple protection against HIV, unintended pregnancy, and other STIs (Shrader et al., 2020). Consistent condom use remains one of the most effective methods for preventing the transmission of HIV, and has been shown to be the most significant factor associated with the reported decline in HIV and STI incidence in South Africa (Manyaapelo et al., 2017; Ntshiqa et al., 2018; Shrader et al., 2020).

Despite the South African government’s policy of distributing free condoms, one of the largest condom distribution programmes in the world, and efforts to rebrand government-issue condoms, evidence suggests that actual condom use is on the decline (Haffejee & Maharajh, 2019; Shrader et al., 2020; Simbayi et al., 2019). There is a lack of evidence explaining this decline in condom use, but theories include increased access to PrEP, increased promotion of voluntary medical male circumcision, “condom fatigue”, decreases in perceived risk of acquiring HIV, and a reduction in prevention efforts targeting condom use (Maakamedi, 2017; Shisana et al., 2014).

In a national survey conducted in 2017, among South Africans of all ages, condom use was found to be highest among people aged 15–24 years, but was still suboptimal (Simbayi et al., 2019). A quarter of survey respondents aged 15–24 years had never used a condom with their most recent sexual partner, and less than half (49.8%) of women aged 15–24 years reported condom use at last sex (Simbayi et al., 2019). In addition, rates of condom use in this age group are declining, with each year showing progressively lower use of condoms for sex among South Africans aged 15–24 (Simbayi et al., 2019). Whilst the development of new biomedical HIV prevention products means that condoms are not the only option, new products such PrEP and microbicides are yet to be widely available, and condoms remain the most cost-effective HIV preventative method, and thus these figures indicating low condom use are worrying (Haffejee & Maharajh, 2019).

While factors such as an individual’s beliefs, personal risk perception, self-efficacy, self-esteem/confidence, cost/benefit perceptions, and knowledge play a role in determining the ability of that individual to use a condom or not, it is crucial to consider how an individual’s social environment or dyadic/relationship interactions with sexual partners influences condom use (Patel et al., 2006). Condom use is a dyadic behavior, subject to the influence of both/all sexual partners, their intentions and relationship perceptions, and is necessarily enabled or constrained by gendered power dynamics which inform behavior and the dyadic negotiations inherent in sexual encounters (Isaacs et al., 2019; Morrison et al., 2014; Patel et al., 2006). Additionally, factors such as alcohol consumption and concurrent use of another contraceptive method affect sexual interactions and play a role in determining condom use (Kiene & Subramanian, 2013; Sarkar, 2008;Simbayi et al., 2004). Importantly, dyadic interactions occur within specific socio-cultural contexts, in which societal norms and collective social level scripts also exert a powerful influence on condom use (Closson et al., 2018; Shrader et al., 2020).

Over the course of the last three decades of the HIV epidemic in Southern Africa, discourse relating to the gendered nature of power in heterosexual relationships, and the accepted gendered sexual scripts of condom negotiation, has centered on the ways in which socio-cultural factors such as gendered power inequity and hegemonic norms of masculinity impede the agency and ability of women to negotiate the use of penile condoms for sex (Amaro, 1995; Closson et al., 2018; Shrader et al., 2020; Rosenthal & Levy, 2010). However socio-cultural norms are not static, and more recent evidence suggests shifts in the nature of gendered power and condom use; a recent study in Cape Town found that condom use was a domain in which women were able to exert some control (Hartmann et al., 2018). Since the willingness and ability to use HIV-prevention technologies, especially the penile condom, are a gendered and dyadic experience, and are likely to undergo shifts in response to the evolving prevention technology landscape, it remains critical to consider both male and female perspectives and narratives (Closson et al., 2018). For the purposes of this paper, we use the term “penile condom”; the problematic semantics relating to “male” and “female” condoms, fraught with penile-vaginal heteronormativity (Duby, 2019), are outside of the scope of this paper.

In South Africa’s national response to the HIV epidemic, AGYW were highlighted as a priority population, and the South Africa’s National Strategic Plan for HIV, TB and STIs 2017-2022 outlined the need for a comprehensive package of services for this group (SANAC, 2017). The provision of condoms was included as a key component in the comprehensive HIV prevention strategy. In addition, the new Department of Basic Education (DBE) National Policy on HIV, STIs and TB highlighted access to penile condoms for learners over the age of 12 as a key focus. In line with these government targets, the Global Fund invested in a combination HIV prevention intervention comprising a comprehensive package of health, education and support services for AGYW, in and out of school, aged 10-24 years. The intervention was implemented 2016 to 2019 in ten high HIV risk districts of South Africa (The Global Fund, 2018).

In this paper, we present quantitative and qualitative data collected in 2018-2019 as part of a larger study evaluating the combination HIV prevention intervention for AGYW. Our analysis explored the complex dynamics around condom use amongst adolescents and young people in South Africa. Survey data provides insight into reporting of condom use at last sex and in past 3 months, as well as how condom use is impacted by the use of alcohol. Qualitative data enriches our understanding of the gendered dynamics at play in the decisions and negotiations around condom use in relationships and sexual situations (Bowleg et al., 2004). Together, these findings describe the current condom use landscape, shedding light on the lived experiences of gendered power, relationship dynamics, and notions of masculinity which play a part in condom negotiation and use in young heterosexual South Africans’ sexual encounters (Closson et al., 2018).

Methods

Survey with AGYW

A cross-sectional survey was conducted among 4,399 AGYW in six of the ten districts in which the combination intervention had been implemented: City of Cape Town (Western Cape), Ehlanzeni (Mpumalanga), OR Tambo (Eastern Cape), Tshwane (Gauteng Province), King Cetshwayo and Zululand (KwaZulu-Natal). A representative sample was selected of households and invited all AGYW aged 15–24 years in the sampled households to participate. The overall sample realization for the survey was 61%. Participants responded to a structured interviewer administered survey. The survey was conducted via electronic questionnaires, administered by a fieldworker using a tablet. Sensitive questions, such as those about condom use, were completed by the participants to diminish social desirability bias. The fieldworker read each question to the participant and allowed the participant to enter her responses in the tablet privately.

Survey participants who reported that they had ever had sex (n = 3,009/4,399) were asked questions about their use of condoms at last sex, in the past 3 months, and condom use concurrent with alcohol or substance use. For the purposes of this paper, we focus on the following survey questions on condom use (“sex” was defined as “when the penis enters the vagina or anus/bum”; no specification regarding type of condom was made):

• The last time you had sex – did you or the person you had sex with use a condom?

• In the past three months have you used a condom?

• In the past 3 months, did you have sex without a condom because you were drinking or using drugs?

Sexually active survey participants were also asked whether, at last sex, they were using a hormonal contraceptive method including the pill, the injection, and the implant, and if their last sex was with a “casual partner”.

In addition, the HIV status of survey participants was determined using laboratory HIV serological testing for infection using blood samples collected from all AGYW. For participants whose study laboratory test results were clinically significant (positive for HIV), a member of the study team contacted the participant by phone to offer an appointment at the local clinic to share the laboratory test results and provide treatment. The study staff would also be at the clinic at the appointed time, to give the participant her results and to link her into care at the clinic. Should the participant not attend the scheduled appointment, further follow-up contact would be made.

Qualitative study component including AGYW and young men

Combined with the survey data, qualitative data collection took place between August 2018 and March 2019 in five districts: City of Cape Town, Western Cape (WC); King Cetshwayo, KwaZulu-Natal (KZN); Gert Sibande, Mpumalanga (MPU); Bojanala, North West (NW); and Nelson Mandela Bay, Eastern Cape (EC). Included in the analysis for this paper are data from 63 in-depth interviews (IDIs) and 24 focus group discussions (FGDs) with 237 AGYW aged 15–24 years (mean age of 17), and 6 FGDs with 38 young men aged 18–23 years (mean age of 19) from the sampled communities. Respondents meeting the eligibility criteria were identified and contacted with assistance from local intervention implementing staff. Eligible participants were invited to participate and enrolled into the study by local research assistants. AGYW in the qualitative study component were recruited independently from the quantitative component, and had not necessarily participated in the survey (Table 1).

Table 1.

HERStory Study Qualitative Sample in Five South African Districts by Site. 

Province	Western Cape (WC)	KwaZulu-Natal (KZN)	Mpumalanga (MPU)	North West (NW)	Eastern Cape (EC)
District	City of Cape Town	King Cetshwayo	Gert Sibande	Bojanala	Nelson Mandela Bay
Characteristic	Urban	Rural	Semi–urban	Semi–urban	Urban
Sample group	n	n	n	n	n	Total n
AGYW* aged 15–18 years	52	28	33	26	38	177
AGYW aged 19–24 years	11	22	8	9	10	60
Total AGYW aged 15–24 years	63	50	41	35	48	237
Male peers aged 18–24 years	7	8	7	2	14	38

*AGYW: adolescent girls and young women.

Qualitative IDIs (20–40 minutes) and FGDs (40–90 minutes) were conducted by experienced female researchers in English, isiZulu, isiXhosa, seTswana, or siSwati. Interviews and FGDs were semi-structured using open-ended topic guides that included questions relating to perceptions and personal experiences barriers and facilitators to condom use and condom negotiation such as: “What makes it easy or difficult for young women and girls like you to use condoms?” (for AGYW), and “In relationships between men and women, who usually makes decisions about whether or not condoms should be used?” (for male peers).

Ethical considerations

Written informed consent was obtained from all participants 18 years and older. For those under 18 years, written assent with written guardian consent was obtained. Reimbursements in the form of supermarket vouchers worth ZAR 50.00 (approximately US$3.00) were provided for all participants. Refreshments were provided at the IDIs and FGDs, in addition to transport costs when required. Study procedures were approved by the South African Medical Research Council Research Ethics Committee, and by the Associate Director for Science in the Center for Global Health in the Centers for Disease Control and Prevention. During data collection, private-sector social workers were procured to assist with ensuring access to social support services when needed.

Quantitative data analysis

Quantitative data were analyzed using Stata/SE 14.2 (StataCorp 2015). Descriptive summary statistics were performed to provide frequency tables, and percentages of the participants’ responses to the variables described above. Data were weighted due to issues related to sample realization. Therefore sample weights were based on the probability of sampling small area layers (SALs, the primary sampling unit) in each district. Further details of the survey methodology are available at: https://www.samrc.ac.za/intramural-research-units/HealthSystems-HERStory.

Qualitative data analysis

Audio recordings were transcribed verbatim into the original language, reviewed by the interviewer/s for accuracy, translated into English and re-reviewed. Thematic analysis was cyclical, initially based on a pre-determined deductive codebook developed in reference to the interview guides and research objectives, and inductively refined (Bradley et al., 2007; Nowell et al., 2017; Vaismoradi et al., 2016). Collaborative interpretation by the research team included data immersion, repeated readings of transcripts, coding with NVivo 12 software (QSR International), documentation and discussion of reflective thoughts. As concepts and themes emerged, they were collaboratively reviewed, and refined. Included in the analysis process were a series of feedback workshops held with AGYW, in which the research team summarized and presented key themes and findings to participants, and AGYW were then invited to give feedback, discuss their interpretation of the findings, and expand or elaborate on themes. Participant feedback was captured through notes and audio recordings, transcribed, and reviewed. These workshops assisted in ensuring accurate and appropriate interpretation of the data, confirming findings and interpretations.

Quantitative and qualitative findings are presented below, with comparisons made between provinces, where appropriate or noteworthy. Qualitative findings are arranged into key thematic areas that emerged during analysis and combined with related quantitative findings. Illustrative quotations are excerpts from English transcripts or translations; in brackets are details of the respondents’ site and sample group.

Findings

Quantitative findings

Amongst the 4,399 AGYW that were surveyed, 57% (N = 2,515) were between the ages of 15 and 19 years, and the remaining 43% (N = 1,884) were between the ages of 20 and 24 years. Survey participants were almost all South African citizens (99%), and most self-identified as “African” (90%). Almost all participants were not married (98%), and over half reported that they were in school at the time they were surveyed (56%). In terms of household level socioeconomic indicators, 34% participants reported that their home had piped water, 37% had a flush toilet in the home, and 74% had households with functioning electricity. Almost a fifth (18%) of participants reported that someone in their household had gone a day and night without eating because of lack of food in the prior month, and 42% reported that their household depended on social grants.

Of the 3,009 AGYW who indicated in the survey that they had ever had sex, 1,300 were between the ages of 15 and 19 years (43.3%), and 1,709 were between the ages of 20 and 24 years (56.7%).

Condom use at last sex

Amongst sexually active AGYW aged 15–24 years across all sites in the survey, 51% (n = 1534/3009) reported condom use at last sex. AGYW aged 20-24 years were less likely to report condom use at last sex compared to the 15–19 year age group (50.6% versus 59.2%, risk difference (RD): −8.6%; 95% confidence interval (CI): −11.6% to −5.6%) (Table 2). In other words, those who were older had approximately 9 fewer instances of using a condom at last sex per 100 AGYW, compared to those who were younger.

Table 2.

Condom use at Last Sex Among Sexually Active AGYW who Answered this Question (n = 2831).

				Condom use at last sex
	Overall			Did not use condom at last sex			Used condom at last sex
Variable	n	%	95% CI	n	%	95% CI	n	%	95% CI	p-Value	Risk difference (%)	95% CI
Age group
15–19	1201	42.5	41.0–44.1	486	40.8	38.6–43.0	715	59.2	57.0–61.4	0.0000	−8.60	−11.64–−5.56
20–24	1630	57.5	55.9–59.0	811	49.4	47.2–51.6	819	50.6	48.4–52.8
HIV status†
Positive	471	16.0	14.8 –17.1	215	45.0	41.0–49.1	256	55.0	50.9–59.0	0.6707	−0.90	−5.02–3.23
Negative	2359	84.0	82.9 –85.2	1082	45.9	44.2–47.5	1277	54.1	52.5–55.8
Used hormonal contraceptives at last sex
No	1852	65.1	63.5–66.7	838	45.1	43.2–47.1	1014	54.9	52.9–56.8	0.2994	−1.73	−4.98–1.52
Yes	979	34.9	33.3–36.5	459	46.8	44.1–49.6	520	53.2	50.4–55.9
Used contraceptive other than condoms to prevent pregnancy at last sex
No	1771	62.3	60.6–63.9	797	44.9	42.9–46.9	974	55.1	53.1–57.1	0.1701	−2.26	−5.48–0.96
Yes	1060	37.7	36.1–39.4	500	47.1	44.6–49.7	560	52.9	50.3–55.4
Last sexual partner was casual††
No	2351	98.3	97.8–98.7	1103	46.8	45.0–48.6	1248	53.2	51.4–55.0	0.6590	2.79	−9.53–15.11
Yes	38	1.7	1.3–2.2	18	44.0	31.5–57.1	20	56.0	42.9–68.5

^†One person had missing HIV status and was excluded. ^††442 participants did not answer this question and were excluded.

There were no statistically significant differences in condom use at last sex among AGYW who were HIV positive compared to those who were HIV negative, among those who were using a hormonal contraceptive at last sex compared to those who were not, and among those whose last sex partner was “casual” compared with those whose last sex partner was not casual (Table 2).

Condom use in last 3 months

Among AGYW aged 15–24 years who had ever had sex, 59% (n = 1773/3009) reported having used a condom in the 3 months prior to their study participation. There were no statistically significant differences in the prevalence of condom use in the past 3 months by age group, by HIV status or by the “status” of the last sexual partner (casual or not) (Table 3). However, condom use in the past three months was more likely among those who reported that at last sex they had used hormonal contraceptives versus those had not (61.7 versus 57.4%; RD: 4.3%; 95% CI: 1.2–7.4%) (Table 3). Condom use in the past 3 months was also more likely among those who reported at last sex they had used any form of modern contraceptives versus those who had not (61.9 versus 57.1%; RD: 4.9%; 95% CI: 1.7–8.0%) (Table 3).

Table 3.

Condom Use in the Last 3 Months Among Sexually Active AGYW (n = 3009).

				Condom use in last 3 months
	Overall			Did not use condom in last 3 months			Used condom in last 3 months
Variable	n	%	95% CI	n	%	95% CI	n	%	95% CI	p-Value	Risk difference (%)	95% CI
Age group
1–19	1300	43.3	41.8–44.8	522	40.5	38.3–42.9	778	59.5	57.1–61.7	0.6905	−0.59	−3.47−2.29
20–24	1709	56.7	55.2–58.2	714	41.1	39.0–43.2	995	58.9	56.8–61.0
HIV status†
Positive	499	15.9	14.8–17.0	190	38.3	34.6–42.2	309	61.7	57.8–65.4	0.1134	−3.02	−6.74−0.69
Negative	2509	84.1	83.0–85.2	1046	41.4	39.7–43.0	1463	58.6	57.0–60.3
Used hormonal contraceptives at last sex††
No	1893	65.4	63.8–67.0	814	42.6	40.5–44.7	1079	57.4	55.3–59.5	0.0069	4.30	1.22−7.39
Yes	991	34.6	33.0–36.2	378	38.3	35.8–40.8	613	61.7	59.2–64.2
Used contraceptive other than condoms to prevent pregnancy at last sex††
No	1812	62.6	61.0–64.2	786	42.9	40.8–45.1	1026	57.1	54.9–59.2	0.0027	4.85	1.73−7.97
Yes	1072	37.4	35.8–39.0	406	38.1	35.6–40.5	666	61.9	59.5–64.4
Last sexual partner was casual†††
No	2388	98.2	97.8–98.6	983	40.9	39.0–42.7	1405	59.1	57.3–61.0	0.2540	6.95	−4.94−18.84
Yes	39	1.8	1.4–2.2	13	33.9	22.2–47.3	26	66.1	52.7–77.8

^†One person had missing HIV status and was excluded. ^††125 participants had missing info and were excluded. ^†††582 participants had missing info and were excluded.

Alcohol/substance use associated with condomless sex

In the survey AGYW were asked to report if, in the last 3 months, they had had sex without a condom due to alcohol or substance use concurrent with sexual intercourse. Amongst sexually active AGYW aged 15–24 years, 8% (n = 208/3009) reported that they failed to use a condom during sex over the past three months because of their alcohol or drug use. Older AGYW were more likely to report failing to use a condom during sex due to alcohol or drug use compared with those in the younger group (8.6 versus 6.1%; RD: 2.6% 95%; CI: 0.9–4.2%) (Table 4). AGYW who said their last sex partner was casual were more likely to report failing to use a condom because of alcohol or drug use versus those whose last sex partner was not casual (22.5 versus 7.5%; RD: 15.0%; 95% CI: 3.2–26.8%) (Table 4). There were no statistically significant differences in the prevalence of failing to use a condom because of alcohol and drugs by HIV status, or by whether contraceptives were used at last sex (Table 4).

Table 4.

Condomless Sex from Alcohol and Drug Use in the Past 3 Months, Among Sexually Active AGYW (n = 3009).

				Condomless sex in past 3 months due to drug and alcohol use
	Overall			Did not have sex without condom because of using drugs and alcohol			Had sex without condom because of using drugs and alcohol
Variable	n	%	95% CI	n	%	95% CI	n	%	95% CI	p-Value	Risk difference (%)	95% CI
Age group
15–19	1300	43.3	41.8–44.8	1228	93.9	92.7–95.0	72	6.1	5.0–7.3	0.0024	2.58	0.93–4.22
20–24	1709	56.7	55.2–58.2	1573	91.4	90.0–92.6	136	8.6	7.4–10.0
HIV status†
Positive	499	15.9	14.8–17.0	458	90.8	88.2–93.0	41	9.2	7.0–11.8	0.1231	−1.94	−4.40–0.51
Negative	2509	84.1	83.0–85.2	2342	92.8	91.8–93.7	167	7.2	6.3–8.2
Used hormonal contraceptives at last sex††
No	1893	65.4	63.8–67.0	1766	92.7	91.5–93.7	127	7.3	6.3–8.5	0.4325	0.66	−0.98–2.29
Yes	991	34.6	33.0–36.2	919	92.0	90.5–93.4	72	8.0	6.6–9.5
Used contraceptive other than condoms to prevent pregnancy at last sex††
No	1812	62.6	61.0–64.2	1692	92.8	91.6–93.8	120	7.2	6.2–8.4	0.3037	0.85	−0.77–2.47
Yes	1072	37.4	35.8–39.0	993	91.9	90.4–93.2	79	8.1	6.8–9.6
Last sexual partner was casual†††
No	2388	98.2	97.8–98.6	2226	92.5	91.5–93.5	162	7.5	6.5–8.5	0.0216	15.02	3.21–26.82
Yes	39	1.8	1.4–2.2	32	77.5	63.1–88.4	7	22.5	11.6–36.9

^†One person had missing HIV status and was excluded. ^††125 participants had missing info and were excluded. ^†††582 participants had missing info and were excluded..

Qualitative findings

Relationship maintenance as a factor influencing nonuse of condoms emerged as AGYW described their feelings around the need to prove their love and commitment to their sexual partner by not using a condom: “If you use a condom… (he will think) you don’t love him enough” (MPU, AGYW, 15–18 years). Proof of love was linked to proving trust and fidelity: “girls are afraid to tell their boyfriend (to use a condom) because he will say ‘it means you don’t trust me, and if you don’t trust me, it means you don’t love me’” (MPU, AGYW, 15–18 years). The desire to use a condom with a partner is interpreted as a sign of mistrust: “when she asks the guy to use a condom… the guy will be like, ‘don’t you trust me?’… then you will be like, ‘nah, I love you, so I do trust you’. And then the thing happens (condomless sex)” (EC, AGYW, 15–18 years); “It is not that easy (to get partner to use condom)… men do not like to use a condom in most cases… they think you don’t trust them” (KZN, AGYW, 15–18 years). For this reason, AGYW may agree to condomless sex in order to avoid relationship conflict: “Men do not like using a condom, even when you try to explain they say they don’t use a condom, he ends up thinking that you don’t trust him and you think he is cheating, and this will cause a misunderstanding” (KZN, AGYW, 15–18 years). The request to use a condom may also be interpreted as a sign of infidelity, and lead to accusations: “It is difficult (to speak to your partner about using a condom)… he would ask nasty questions like, ‘why now?… all of a sudden why do you want to use protection… are you seeing someone else?’” (KZN, AGYW, 19–24 years).

Due to the associations of trust with condomless sex, AGYW expressed the perception that condoms are only needed for casual sex partners, and do not need to be used in established “long-term” relationships: “From 2–3 months, dating, you can use a condom, but when its 4–5 months… It is unlikely that you can use it… forget it!” [laughter] (MPU, AGYW, 19–24 years). Familiarity and comfort levels with a partner will negatively impact on consistent condom use: “When you are in a happy mood or you know each other very well… you can do it without using a condom… when you are so used to each other and you have been dating for a long time” (MPU, AGYW, 19–24 years).

One incentive to have condomless sex was AGYW’s fear of being dumped by boyfriends if they don’t acquiesce to their demands to have sex without a condom. This theme was observed most notably in the Western Cape: “I love my person (boyfriend) and he will leave me if I don’t sleep with him… (and he) says we should not use a condom” (WC, AGYW, 15–18 years); “If you use condom you will be left (dumped)” (WC, AGYW, 15–18 years). Pressure from male partners to have condomless sex was a commonly emerging theme in AGYW narratives: “When you are together with your baby (boyfriend?) you… use a condom for round one. Then for round two your baby says, ‘no baby let’s not use a condom’ and then you will say ‘no baby I will fall pregnant…’, but he will sweet talk you and you will end up blushing and agreeing to do it” (WC, AGYW, 15–18 years). Respondents suggested that those AGYW with older sexual partners have less power to negotiate condom use: “If he (partner) is your age and not older than it would be easy for him to understand (your request to use a condom) but if he is older he will refuse and it will his word and he will say he doesn’t want to use one” (KZN, AGYW, 15–18 years); “If he (partner) is older you will be scared of him, and you must do things (have condomless sex) forcefully even if you don’t want to” (KZN, AGYW, 15–18 years). AGYW voiced the feeling of being too scared to raise the topic of condoms with boyfriends, and in some cases, feared intimate partner violence if they suggest condom: “(girls) are afraid that the boyfriend might beat her… afraid to bring up the issue of condom” (MPU, AGYW, 15–18 years).

Despite AGYW articulating that it is the male partners who dislike condoms and pressure them for condomless sex, male respondents also suggested that sometimes AGYW are the ones to initiate condomless sex: “At times, we as guys bring condoms, and then these girls say it is not enjoyable with a condom… I brought it thinking that today we might use it, and then she says ‘no baby take it off now, it is painful and I am not enjoying it’, you see, so sometimes they (girls) are the ones who make us not use it” (EC, Male Peer). Indeed some AGYW themselves described their own dislike of condoms, due to their reducing physical sensation and pleasure during sex: “Me and my partner are not using a condom… I have never asked him to because I don’t like condoms, it (sex with a condom) is boring [laughing]… condoms are boring, I once tried it before, it bores me… when I use it I feel nothing [laughing]… We are used to “skoon” (Afrikaans slang=‘clean’/condomless sex)” (KZN, AGYW, 15–18 years).

Respondents described situations in which AGYW using hormonal contraceptives, have reduced ability to negotiate concurrent condom use with their male partners: “Guys do support it (girls using hormonal contraceptives) a lot, because of what they call ‘flesh to flesh’ – you feel the person when there is no condom, so when you are preventing (using contraceptives) it’s an excuse for them not to use condom” (EC, AGYW, 19–24 years). Some AGYW stated that they deceive their boyfriends, telling them that they are not currently using any hormonal contraceptives, so that he will agree to use a condom: “I act like (I) am not preventing (using pregnancy prevention method) so that he uses condom” (WC, AGYW, 15–18 years).

Beliefs relating to the negative side effects of condoms also serve as a barrier to their use. Perceived side effects included uterine pain, STIs, and discharge: “I don’t like a condom… it hurts me inside my womb, it causes STIs and a stinking discharge… it hurts me…. (so) we don’t use it at all” (MPU, AGYW, 15–18 years). Pain experienced by males when wearing a condom was described by both male and female respondents: “I can also be honest by saying I don’t like a condom… it hurts me” (MPU, Male Peers); “My boyfriend doesn’t like a condom… he will simply say, it hurts him… he just says he feels it in his blood that he doesn’t like it” (KZN, AGYW, 15–18 years). The fear of condoms getting lost or disintegrating inside the vagina was also cited by AGYW: “Some girls are scared to have sex with condoms because they are scared, when they doing sex, the condom might break inside her” (EC, AGYW, 15–18 years). Participants shared their beliefs that condoms should not be used in situations of sexual debut/first sex: “Using a condom when you are a virgin is not a good thing… (you should) break your virginity without using a condom” (WC, AGYW, 15–18 years).

Male power in controlling condom use was evident in the narratives of young men: “He is the one who has the authority to take the condom off in the middle of sex, he can remove it if he likes, or decide to continue with it, if he likes” (KZN, Male Peer); “Whoever has the condom on, is the one who has the power” (KZN, Male Peer). Themes of sexual prowess, proof of power over partners, and the prestige associated with condomless sex, and impregnating a female partner were prevalent in male narratives. Participants described the occurrence of boys boasting amongst peers about successfully managing to have sex without a condom, even if their partner wanted to use one: “boys tell their friends… ‘she told me to use a condom’… and then the friend asks, ‘did you use a condom’, he is like ‘nah, I like it raw, I like it raw’, that what they (boys) always say” (EC, AGYW, 15–18 years).

Getting a girl pregnant was described by male participants as a symbol of prestige, and sexual maturity for young men. Condomless sex was associated with concepts of manhood and masculinity; male respondents suggested that male celebrities and respected role models do not use condoms: “Guys are the ones who are refusing to use a condom… because they want to prove a point to many, like me, I am Mr Duster… and Mr Ganda (celebrity DJs and soccer players)… I can also say he also doesn’t like a condom …Because he wants to prove a point to his friends… that I slept with her… I impregnated her… it lifts you very high [laughing]” (MPU, Male Peer). Male respondents suggested that making a girl pregnant is like laying claim to her; if she’s beautiful, you want to impregnate her to show that she’s yours: “When a girl is very beautiful… You want to make a show off that you have impregnated her… that excites you and that will lift you up high… You will hear the other guys saying ‘…you have a real wife’” (MPU, Male Peer). Competitiveness between young men drives the desire to have condomless sex in order to impregnate girls: “When you realize that one guy has made a girl pregnant… I envy him… then I will also be tempted to also make a girl pregnant” (MPU, Male Peer). Young men described the advice received from friends and brothers, putting pressure on to have condomless sex: “Your brother or friend will tell you… that you need to sleep with her… they will tell you to go and sleep with her, then, you think of whether to use a condom or not to use a condom… Then your brother or friend will tell you to penetrate her without a condom” (MPU, Male Peer).

Qualitative narratives from young men illustrated the way in which the consumption of alcohol increased the likelihood of condomless sex occurring: “When you are hooked up with a girl… it might happen that we are both drunk… the guy will lie to you by saying he will have the condom on, or sometimes just keep quiet… that person is being controlled by alcohol at that moment” (MPU, Male Peer).

Additional barriers to condom use cited by young people related to the challenges in accessing condoms. Obtaining the free government-issue condoms available at public clinics was regarded as onerous: “People don’t want to make an effort and go to those places (clinic) for condoms” (EC, Male Peers). In addition, respondents suggested that young people feel too embarrassed to get condoms from the clinic, fearing that assumptions would be made about their sexual behavior: “You don’t want people to see you when you go and fetch the condoms there by the clinic because you don’t want people to know that you’re already sexually active at this age that you are at right now, so you’re gonna get shy, what if you go to the clinic then you see your neighbor. How will she look at you tomorrow? Will she say ‘oh this boy is sexually active, what is this?’… We don’t want people to gossip… we are shy… they must have a campaign for teenagers… to supply them with condoms better” (EC, Male Peers). Similar to their male peers, AGYW described their fear of embarrassment and being judged: “The people in the clinic will laugh at me, asking me where I am taking them (condoms) to” (KZN, AGYW, 15–18 years). Reluctance to access and use government-issue condoms, was also linked to the idea that being free of cost, they were undesirable, interpreted as a symbol of your partner not valuing you, if you are not worth purchased condoms: “We have never used the clinic condoms… I don’t like them. I feel that he does not value me. He must buy and pay for condom…[smile]… I mean if you get it for free at the clinic, I don’t like it” (MPU, AGYW, 19–24 years).

Views on the distribution of condoms at school were varied. Perceptions that increased condom availability and access are likely to increase sexual activity and promiscuity amongst young people were evident in the narratives of some respondents. Resistance to the idea of the provision of free condoms at school was grounded in the belief that providing free condoms at school will encourage sex amongst young children and have a negative effect on academic performance: “I don’t think it’s right… for the school to supply condoms… Because the children are going to think it’s right to be sexually active… they are going to be easily influenced… pressured into having sex, but they are young, so I don’t think the school should supply us with the condoms… we don’t want (younger kids) to think that being sexually active at this age is right… we want to prevent them from getting the virus and pregnant… they will think it’s right to have sex because there are condoms at school… schools should just say we should stay away from sex… not to be supplied with condoms at school” (EC, Male Peers). In the same FGD, contrasting views in support of condom distribution at schools were expressed: “To get quicker access to condoms… I want the school to have condoms, because there are girls and boys that are sexually active and they need these condoms. Look at our society today, we have say a girl, a teenage girl… having 3 or 4 kids… so we want to drop this pregnancy rates, we want to drop it… so for us teenagers… the school must provide us with condoms” (EC, Male Peers).

The issue of whether condoms and other contraceptives should be offered on the school premises was raised in the analysis feedback workshops with AGYW. A spectrum of opinions was expressed, some of which echoed the sentiment that condom provision would encourage promiscuity: “We came here at school to learn… if condoms can be distributed here (at school), academic issues can fall behind. Kids can fall behind with school work… It will be like promoting sex issues and not education” (MPU, AGYW feedback workshop). Other concerns related to confidentiality: “Condoms should not be distributed here at school because others will be afraid to fetch them. They will end up not using protection because they will be scared that they are known that they… have taken condoms then that will mean oh, she has started to have sex… it will be better if they are distributed in the clinic but even there it’s still the same (MPU, AGYW feedback workshop).

Discussion

Condoms remain an important method of contraception and HIV prevention, but consistent condom use amongst adolescents and young people in South Africa is sub-optimal. Analysis of qualitative data revealed gendered motivations for the use or nonuse of condoms. Key factors determining condom use motivations for AGYW described in our study included relationship security and the desire to demonstrate love, trust, intimacy and commitment. At times, fear of violent reactions from male partners prevents AGYW from raising the topic of condom use. Negative beliefs and perceived negative side effects of condoms serve as a barrier to their use. For young men, key motivations for condomless sex included increased sexual pleasure, proof of their masculinity and power, and prestige amongst peers. Young men’s desire to attain sexual prowess, respect, and masculine sexual maturity, enhanced their resistance to condom use.

The influence of using hormonal contraceptives on condom use behavior was unclear in the data. Respondents in the qualitative study component suggested that AGYW who use hormonal contraceptives find it harder to negotiate condom use with male partners, with some AGYW choosing to mislead their boyfriends, telling them that they are not currently using any hormonal contraceptives, in order to increase his likelihood of agreeing to use a condom. However, there was some discrepancy between qualitative narratives and reporting of condom use in the survey. Analysis of the quantitative data reveals that those using hormonal/modern contraceptives were more likely to report condom use. This may be an indicator that AGYW who use contraception have more agency (evidenced by their use of contraception), and are therefore better able to negotiate condom use; or that AGYW using contraceptives hid their use and therefore could better negotiate condom use with partners. On the other hand, the qualitative data revealed males’ preference not to use condoms if their female partner is using hormonal contraception, thus speaking to the likelihood that condoms may be more acceptable for contraceptive purposes than for HIV/STI prevention purposes. If this is the case, it is possible that contraceptive use undermines the potential for condom use in this population (van Loggerenberg et al., 2012).

Emerging in both our qualitative and quantitative data were links between condomless sex and alcohol or substance use. Additionally, relationship context combined with alcohol/substance use influenced condom nonuse, with reporting of failure to use a condom because of alcohol or drug use notably higher among AGYW who said their last sex partner was “casual”. It is unclear whether the failure to use a condom was due to carelessness or disinhibition related to alcohol/substance use, however it has been stated that alcohol use is also related to increased intentions to resist condom use (Davis et al., 2014).

Access to condoms was a contentious issue in the data, with a spectrum of views expressed. Young people voiced reluctance to access condoms from clinics out of fear of being judged, or that it required too much effort. Another barrier to using free condoms was related to their undesirability and association with worthlessness, with bought/paid for condoms perceived by AGYW as an indication of being valued more highly by a partner. Perceptions of South African government-issued condoms being of inferior quality compared to commercial brands, and therefore associated with lower valuing of a partner, have been identified as problematic (Mthembu et al., 2019; Mulaudzi & Jabuli, 2018). Discussions around condom distribution at school also generated contrasting views, ranging from fears of promoting promiscuity, to concerns around confidentiality, to the perception that condoms at school are necessary in order to lower teenage pregnancy rates. The perception that the provision of free condoms in schools will promote earlier sexual debut and increased sexual activity has long been a socially divisive issue (Han & Bennish, 2009; Wang et al., 2018). However school-based SRH programmes, inclusive of the provision of condoms and contraceptives, have been identified as an important mechanism to increase access and uptake (Jonas et al., 2020).

Relationship maintenance as a factor influencing nonuse of condoms was one of the key emergent themes in the qualitative data. AGYW expressed the perception that condoms are only needed for casual sex partners, and do not need to be used in established “long-term” relationships. Condomless sex was described by AGYW as a way of demonstrating love and trust. In the narratives of male respondents, relationship maintenance as a motivation emerged in the articulation that young men seek to impregnate beautiful women as a means of laying their claims. Our data demonstrated that despite the well-established links between long-term relationships, intimacy and trust, and feelings of love with condomless sex, the motivations and dynamics inherent in negotiations and decisions around condomless sex are complex and remain incompletely understood (Bhana, 2017; Fortenberry, 2019; Knox et al., 2010; Mash et al., 2010; Patel et al., 2006). With condomless sex regarded as tangible proof of commitment, trust and intimacy, the discontinuation of condom indicates the progression of a relationship (Bhagwanjee et al., 2013; Bhana, 2017; Fortenberry, 2019; Mash et al., 2010; Shai et al., 2012). Linked to discussions of trust in relationships, AGYW in our study voiced concerns about being accused of infidelity by partners, should they request condom use. The idea of condoms as a symbol of infidelity is problematic and likely to pose a barrier to use in the context of relationships (van Loggerenberg et al., 2012). Following from this, in the context of multiple concurrent partnerships, condom use is more likely to be inconsistent in primary partnerships compared to casual partnerships (Moyo et al., 2008).

The decision to engage in condomless sex due to the fear of being “dumped” or rejected by boyfriends described by AGYW in our study indicates a prioritization of romantic security and intimacy over sexual health. Willing to put themselves at risk for the sake of demonstrating trust, and prioritizing romantic connection over their own sexual health, AGYW may even be likely to discourage condom use in the belief that it increases their chances of maintaining their romantic relationships (Jewkes & Morrell, 2010; Knox et al., 2010; Van Der Riet et al., 2019). The self-esteem and social status of AGYW is often linked to their being in a romantic relationship, and therefore the security and maintenance of relationships is prioritized, which can compromise AGYW’s agency (Van Der Riet et al., 2019). Although it is not within the scope of this paper to delve into intimate partner violence, some AGYW in our study described being too scared to raise the topic of condoms with boyfriends, fearing violent reactions. The risk of violent outcomes can indirectly affect fertility preferences and the transmission of STIs through women's fear of refusing sexual advances, or raising the issue of contraception or condom use (Blanc, 2001; Pettifor et al., 2012).

In light of the themes that emerged in the data pertaining to the way in which AGYW agree to have condomless sex for the purposes of relationship maintenance, or to avoid negative reactions from the boyfriends, the power and agency in condom use decision-making in heterosexual relationships amongst young South Africans appears to be weighted in young men’s favor. Although demonstrating that AGYW do wield some power in dyadic condom interactions, and do at times insist on condomless sex themselves, or demand bought condoms versus government-issue free condoms, our findings add to the evidence suggesting a power disparity in condom use negotiation amongst young heterosexuals in South Africa. Our findings help to deepen understanding of gendered sexual and relationship power disparities within heterosexual dyads. Our data builds on evidence describing the ways in which male partners’ condom intentions often over-ride AGYW intentions, and that condomless sex is more often a result of male refusal to wear a condom (Isaacs et al., 2019; Manyaapelo et al., 2017). As seen in our data, some of the reasons for disliking condoms, such as decreased sexual pleasure, are similar across genders. However gendered power disparities mean that men are more likely than women to employ “condom resistance tactics”, strategies including physical sensation arguments, emotional manipulation, relationship-based reasons, and physical threat or force, amongst others (Davis et al., 2014).

In the qualitative data, it was evident that gendered sexual norms and prevalent masculinities influence condom use. Young men’s desire to attain sexual prowess, respect, and a masculine sexual maturity, links to their motivation to engage in condomless sex. Tied in with male narratives around wanting condomless sex, was discussion around the desirability of impregnating girls, in order to demonstrate your sexual conquest, and attain sexual and social maturity. Prior research suggests that young South African men are strongly influenced by how significant people in their lives view theirbehaviors, and are strongly motivated to comply with the options of others(Kaufman et al., 2008; Van Der Riet et al., 2019). The concepts of masculinity and machismo are interwoven with resisting condom use, operating alongside the narrative of male sexual pleasure being prioritized in heterosexual sex, and therefore the desirability of condomless sex in order to maximize male pleasure and enhance male sexual power (Kaufman et al., 2008; Shai et al., 2012; Van Der Riet et al., 2019). Considering the way in which masculinities and gendered sexual norms emerged in our data, our findings corroborate assertions that inequitable gender relations heighten the sexual risk of AGYW and decrease their ability to negotiate condom use or refuse unwanted sexual encounters (Bhana, 2017; Pettifor et al., 2012). Traditional gender scripts in South Africa have included the expectation that men put pressure on their partners to have condomless sex, and in response women should attempt to resist condomless sex, and eventually acquiesce to male sexual demands (Bhana, 2017; Mantell et al., 2011).

Although it is important to consider the way in which gendered power dynamics, male sexual norms and masculinities, and structural factors influence the ability of AGYW to negotiate the use of condoms, it is also important to recognize that in some cases, AGYW themselves do not want to use condoms. In our study, some AGYW expressed their own dislike of condoms, due to their reducing physical sensation and pleasure during sex. When AGYW engage in condomless sex, it is not always a function of their relationship context, and may be more due to factors related to their own sexual pleasure, intimacy, and the desire to get pregnant, which lead to them to oppose condom use (Bowleg et al., 2004; Jewkes & Morrell, 2010). It is important to critique arguments that fall back on the notion of traditional gendered power scripts that frame South African women as powerless and subservient, particularly in the light of gender norm transformations (Pettifor et al., 2012). It has been argued that due to the way in which sexual pleasure is profoundly structured and influenced by gender inequality, AGYW often lack the information, tools, or agency to discuss or negotiate their own pleasure (Ford et al., 2019). However, research conducted over the last decade paints a more complex picture of gender dynamics, suggesting that dominant gender roles with respect to HIV prevention may be undergoing important shifts, placing more power in the hands of women when it comes to condom decision-making (Hartmann et al., 2018; Mantell et al., 2011; Mfecane, 2013; Pettifor et al., 2012).

Young women in South Africa are increasingly sexually assertive, holding more control over sex and condom decision-making, marking a transition away from the stereotypically sexually powerless and oppressed African woman (Hartmann et al., 2018; Mantell et al., 2011; Mfecane, 2013; Pettifor et al., 2012). However, as is evidenced in our findings, and suggested in other research, in examining the behavior and conceptions of young South Africans, there seem to be tensions and conflicting influences between retaining traditional gender roles while at the same time moving toward more progressive and more equal gendered power relations (Mantell et al., 2011; Pettifor et al., 2012). Our findings add weight to the assertion that despite cultural shifts, the sexual agency of AGYW in South Africa remains constrained by inequitable gendered power dynamics (Mfecane, 2013). There would be value in further examining how the sexual agency of young women differs in different socio-economic contexts in South Africa. The bulk of the research suggesting increased sexual agency of AGYW derives from urban settings. Although our findings provide valuable data from varied settings in South Africa, one limitation of our study is that we did not conduct comparative analysis between study sites, which would be of value for future research efforts.

Other limitations of this study include the survey response rate of 61%, ranging from 33 to 78%, and thus the potential for bias. Nevertheless, these response rates compare well with those of other national surveys in which AGYW are invited to provide biological samples. Additionally, the survey relied on self-reporting of condom use, without validation of self-reports with biomarkers. The survey questions on condom use did not specify the type of condoms used; in future it would be useful to ask more detailed and specific questions, inclusive of different types of condoms. Social desirability bias may also have influenced qualitative narratives around condom use. Notably, this study included only a limited sample of males; no males were included in the survey, and only a small sample were included in the qualitative study. Therefore the results from this study are not generalizable. Nevertheless, the narratives of male peers included add value to the analysis.

Recommendations and conclusions

In light of the centrality of dyadic interactions, and gendered power, in determining condom use, interventions aiming to increase condom use need to engage young men and women in dialogues about gender, in order to critique and deconstruct existing notions of manhood and womanhood, and reinforce positive forms of masculinity that enable more equal power in negotiations over condom use (Mantell et al., 2011). Interventions need to address gender inequalities, promote women’s rights and autonomy and challenge male privilege and power (Shrader et al., 2020). However, while such efforts should still be made to empower young women in condom negotiation (Haffejee & Maharajh, 2019), interventions may also benefit from shifting the focus of improved condom negotiation skills from AGYW to young men, engaging men and boys in programmes which work to foster gender-equitable beliefs, behaviors, and actions, through gender-targeted initiatives that address gender norms and attitudes (Closson et al., 2018). There has been some success with “gender-transformative interventions” in shifting harmful gender norms and roles through integrated community-based programming, and in doing so, achieving an improvement in structural and individual-level risk behaviors and sexual outcomes (Closson et al., 2018).

Additionally, since condom use is influenced by both individual and interpersonal level factors, there may be some value in targeting partner-level influences on condom use through interventions which include communication and negotiation skills training components, in order to empower AGYW to translate their safer-sex intentions into actual behavior (Gause et al., 2018). It is also important that interventions recognize the bidirectional influence that sexual partners have on each other; couple-based HIV prevention interventions based on a model of healthy intimate relationships may be an effective means of addressing relationship level barriers to condom use (Belus et al., 2019). In our data, young men’s narratives suggested a lack of concern about HIV infection, which suggests that engaging men and boys in HIV prevention interventions, rather than leaving condom use negotiation to women, should be a key focus for future efforts. In order to enable sustained and consistent use of condoms, men’s engagement and involvement is critical (Montgomery et al., 2008).

Within the framing of an integrated definition of sexual and reproductive health, comprising positive approaches to sexuality and reproduction, recognizing the role of pleasure, trust, and communication in the promotion of self-esteem and overall wellbeing in sexual relationships, our findings suggest the need for contextually relevant education and messaging around relationship dynamics and gender, including exploration of the concepts of trust and construction of expectations within heterosexual relationships, socio-cultural norms around sexuality and pleasure (including female pleasure) and consent (Starrs et al, 2018). Narratives of pleasure were salient in young men and women’s accounts of sex and condom use, indicating the importance of recognizing pleasure as a central motivator for sex in the design of interventions. In framing the benefits of PrEP as an alternate HIV prevention method to condoms, unaffected pleasure and intimacy have been cited as a key advantage (Irungu et al., 2020). Further research should be conducted in order to understand the ways in which narratives of condoms as a barrier to pleasure and intimacy can be shifted.

Given the suboptimal use of condoms, and relatively high prevalence of condomless sex among adolescents and young people in South Africa, and the powerful motivations for condomless sex, these findings highlight the need for strengthened HIV prevention programming inclusive of condoms as one option amongst other prevention technologies, and on-going education and messaging aaround the importance of dual-protection, encouraging AGYW who use hormonal contraceptives to also use an HIV prevention method. This study provides valuable male and female perspectives on condom use among young people in South Africa, revealing that many of the motivations for condomless sex are gendered. Although gendered power disparities emerged in the data, with AGYW agreeing to have condomless sex for the purposes of relationship maintenance, or to avoid negative reactions from the boyfriends, dynamics of pleasure, intimacy, trust, power and value were shown to be nuanced and contextual. Condoms remain an important HIV prevention method, in addition to an important tool for preventing the spread of STIs and unintentional pregnancies. In order to improve condom use amongst adolescents and young people in South Africa, the complex multilevel barriers to their use need to be addressed through innovative interventions inclusive of individual, interpersonal and socio-cultural level components. Understanding the factors that motivate condomless sex in this population is critical in order to ensure that interventions to increase consistent condom use are relevant, appropriate, and framed within an integrated understanding of sexual and reproductive health.

Funding Statement

The AGYW intervention was funded by the Global Fund to Fight AIDS, TB, and Malaria. The programme was aligned with the She Conquers campaign funded by the United States President’s Emergency Plan for AIDS Relief (PEPFAR) and was implemented with support from the South African National AIDS Council (SANAC) through the Country Coordinating Mechanism (CCM) and the CCM Secretariat. This manuscript was supported by the President’s Emergency Plan for AIDS Relief (PEPFAR) through the Centers for Disease Control and Prevention (CDC), Division of Global HIV & TB (DGHT) under the terms of Cooperative Agreement 1U2GGH001150. Development of this publication was supported by the Social Impact Bond of the South African Medical Research Council.

Conflict of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.