What this study adds: Despite substantial research on IPV, little information exists on perpetration among adolescent boys who acquired PHIV. This study found similar incidence of IPV among participants with and without HIV from the same community.

Intimate partner violence (IPV) is defined as behaviours by a current or former intimate partner that cause emotional, physical, or sexual harm.^1,2 During adolescence, both IPV and HIV incidence rise dramatically.^3,4,5 Moreover, the relationship between these risks is bidirectional. Youth living with HIV may be victims as well as perpetrators of IPV, with some studies suggesting that adults living with HIV are more likely to experience IPV, resulting in part from relationship conflict as a result of HIV disclosure.⁶

In adults, a review of national cross-sectional surveys in 27 African countries found that adult men living with HIV were more likely to be perpetrators of IPV (adjusted prevalence ratio = 1.09; 95%CI: 1.01–1.16).⁷ There is also evidence, for example, that victims of IPV are up to twice as likely to contract HIV.⁸ In the adult population, the link between HIV and IPV can be explained by common risk factors such as poverty, adverse childhood experiences, substance use, and gender inequality views.^9,10,11

While we do not know if this holds true for adolescent boys who have grown up with HIV and have a very different route of infection, some of the risk factors may be shared with their adult counterparts. In addition, boys living with HIV, particularly in impoverished contexts, may experience a higher burden of childhood adversity (e.g. orphanhood, witnessing domestic violence, etc.),^12,13 which is a strong predictor of IPV perpetration^14,15 in comparison to boys without HIV.

Adolescents of either gender can be perpetrators and victims of IPV, but research indicates that boys engage in more violent forms of IPV perpetration.¹⁶ Moreover, IPV behaviours start early. For example, a South African study showed that among Grade 8 boys, as young as 12 years of age, 46% reported emotional IPV perpetration and 39% reported physical or sexual IPV perpetration.¹⁷ Past studies highlight several reasons adolescent boys may perpetrate violence, including provocation by a partner, self-defence, stress and anger, as well as patriarchal cultures and dating norms.^18,19

Understanding patterns of IPV perpetration among adolescent boys living with HIV, at an age when they begin to engage in sexual and romantic relationships,²⁰ is critical for designing interventions and for preventing HIV transmission to their sexual partners.²¹ Moreover, behaviours are still modifiable during this developmental period; thus, there is an opportunity to intervene. In this study, we estimated the prevalence of IPV perpetration among adolescent boys (sex assigned at birth) and tested whether boys living with perinatally HIV (PHIV) would be more likely to perpetrate IPV as compared to their HIV-uninfected peers. In addition to examining lifetime and past-year prevalence through traditional retrospective surveys, we advance the field by prospectively capturing incidence. To do so, we used ecological momentary assessments that captured events in real time.²² A key strength of this approach is that it reduces recall bias and more accurately detects IPV perpetration.²³ Moreover, it captures events as they occur in the boys’ natural environment (leading to better ecological validity), and better documents the key dynamics, such as motivation, surrounding each IPV event.^24,25 Using the mobile phone data, we estimate the incidence of perpetration among adolescent boys in South Africa, describe who they victimise, and document the reported reasons for perpetration.

Based on the literature on childhood adversity and data from adults living with HIV, we hypothesised that adolescent boys with PHIV would perpetrate more IPV compared to their HIV-negative peers, both over their lifetime and within the past year.²⁶ This study examines the prevalence and incidence of IPV among adolescent boys living with HIV and compares that to their HIV-negative peers.

This was a longitudinal cohort study in adolescent boys living with and without HIV.

At baseline, the majority of the 498 participants (71%, n = 352) were younger than 18 years and were attending school. Consistent with the enrolment criteria of either being in a face-to-face relationship or having had sex in the month prior to enrolment, nearly all were dating (99%, n = 493), and more than half were sexually active (52%, n = 255) (Table 1). Twenty participants (4%) had a positive urine test for one or more STIs, of whom only 15 reported being sexually active. There were nearly four times as many HIV-negative adolescents with an STI than PHIV. Overall, C. trachomatis was found in 14 participants, N. gonorrhoeae in six, T. vaginalis in two, and M. genitalium in one. Inconsistent condom use among those who were sexually active was also more prevalent in HIV-negative adolescents (Table 1).

Over half of the adolescent boys surveyed in our study reported perpetrating IPV at some point their lifetime, with no difference based on HIV status. Contributing to the robustness of our findings, the study employed real-time data collection to report on the daily incidence rate of perpetration. We found that there were three events of IPV perpetration out of every 100 person days in which boys spent time with a sexual or romantic partner, highlighting the pervasive nature of the issue.

We observed few discernible differences by HIV status – a finding that contrasts with a review on adult men.⁷ The association in adults is attributed to similar vulnerabilities for both HIV and IPV.^18,34 However, this relationship may not be present in our sample because of differences in mode of infection and its associated behavioural health risk factors.³⁵ Specifically, those who acquire HIV behaviourally, as opposed to perinatally, may have pre-existing factors that lead them to engage in both IPV and risky sexual behaviours. Notably, excessive substance use, a significant risk factor, was more likely to be reported by adolescents who acquired HIV through behavioural choices.^35,36 Our participants acquired HIV perinatally, and thus their risk pattern may differ.

In our study, IPV perpetration was common among both boys living with HIV and their peers. IPV perpetration has been shown to heighten the risk of HIV transmission.³⁷ Therefore, it is crucial to ensure PHIV individuals are included in IPV prevention programmes. To develop responsive interventions, a comprehensive understanding of the context and drivers of IPV is imperative. Frameworks such as feminist theory, that consider gender and power dynamics, propose that IPV emerges as a manifestation of patriarchal structures and beliefs, which perpetuates male displays of power, control, and desires to dominate women.¹⁸ We anticipated that boys would endorse reasons for IPV perpetration that aligned with this framework, such as seeking respect and suspicion of infidelity. Indeed, seeking respect or attention was one of the common reasons for physical IPV. However, contrary to our expectations, the inability to control their temper and self- defence were the most frequently reported reasons. Adolescence is marked by hormonal fluctuations, academic pressure, and navigation of social relationships that often trigger strong emotions. Regulating these emotions and learning to control one’s temper is a developmental task that is being refined during this period and could be a target for intervention. We also found that sexual IPV was primarily directed at steady partners. Interestingly, some physical IPV may have been bidirectional, indicated by instances in which participants reported using violence in self-defence.

To our knowledge, this was the first prospective study using ecological momentary assessments to measure incidence of IPV perpetration amongst boys living with PHIV and compare it to a group of boys without HIV. There are, however, limitations to our study. Our sample had been engaged in long-term care, and may have been exposed to health information, attended HIV support groups, and had time to address stigma and other challenges with their trusted healthcare providers. Thus, they may not be representative of all boys living with HIV.³⁵ Despite approaching PHIV in healthcare facilities, verifying each participant’s HIV status at birth would make the study unfeasible. That children known to be living with HIV before age 10 were perinatally infected is a very reasonable assumption in an area of high HIV prevalence among pregnant women, but the sample could have included boys with a different mode of transmission. Another limitation is the use of self-report measures, which could introduce recall and response bias.³⁸ In addition, once-weekly surveys would likely bias reports of perpetration downward, making our estimates conservative. Also, to keep the surveys short, emotional IPV was not surveyed during the weekly mobile surveys.

A main strength of this study is that it complemented prevalence estimates from traditional surveys with estimates of daily incidence rates generated from daily diary methods. This approach reduced recall bias, and better captured IPV perpetration occurring in the boys’ natural environment. While the use of daily mobile phone surveys has several key advantages, there was substantial attrition. Compliance is a common obstacle in daily diary studies of IPV,^23,39 and our focus on adolescents in a low-resource environment presented additional unique challenges.²² If participants were less likely to complete a mobile phone survey on days when they also perpetrated IPV, our results would again underestimate the true incidence.

There aren’t many violence interventions for adolescent boys. This is a key population to include in intervention programmes. As PHIV adolescents exhibit similar rates of physical and sexual IPV perpetration to their HIV-negative peers, tailored prevention interventions based solely on HIV status may not be a necessity. However, further research is essential to ascertain whether the underlying risk factors driving perpetration in both groups are, in fact, similar. Nevertheless, inclusive interventions are needed for PHIV adolescents, considering the heightened risk of HIV transmission associated with perpetration. Potential entry points for intervention include HIV support groups or other HIV care services. Intervention research may need to target relational and behavioural factors. This encompasses fostering healthy relationship norms, imparting conflict resolution skills, shaping attitudes towards IPV, and emphasising power negotiation within relationships – themes common to both groups of adolescent boys.^40,41 These may be achieved through school programmes, embedding within existing local programmes for adolescent girls and young women, parenting life skills initiatives, and youth-friendly services at their local care clinics. Recognising the importance of early intervention, local care clinics need to identify at-risk individuals, facilitating timely and targeted efforts to address IPV and its associated risks among adolescents living with HIV. These findings are important because existing literature suggests that there are also negative internalised and externalised consequences for perpetrators of IPV, not just the victims.⁴²

HIV and IPV incidence increase rapidly during adolescence; however, few studies have looked at this behaviour among PHIV individuals. Bridging this gap in the literature, our study revealed alarmingly high rates of IPV perpetration, with no discernible differences by HIV status. This underscores the urgency for targeted interventions in all adolescents, ensuring that effective strategies encompass both prevention and support measures.