Differing psychological vulnerabilities among behaviourally and perinatally HIV infected adolescents in South Africa - implications for targeted health service provision

HIV infections are growing the fastest amongst adolescents, especially in sub Saharan Africa. On reaching adolescence, perinatally-infected youth may have different needs to those who acquired infection behaviourally. Yet both have sub-optimal adherence with implications for their own health as well as onward transmission. This study uses the world's largest community-based study of HIV-positive adolescents from the Eastern Cape, South Africa. Clinic records at N = 53 district health facilities generated a log of all ART-initiated adolescents who were then interviewed in the community: N = 1058 (90%) were tracked and participated. Ethical approval, informed consent and data collector training preceded data gathering. Inventories comprised validated measures of mental health (depression, anxiety, suicidality and internalised stigma), substance use, ART adherence, and clinic attendance. Analyses were conducted using SPSS25 and STATA15. Perinatally-infected adolescents (n = 792, 77.3%) were significantly more likely to be ART adherent (OR = 1.54 95%CI: 1.14-2.07 p = 0.005), retained in healthcare (OR = 1.59 95%CI1.18-2.14 p = 0.002), and treated well by clinic staff (OR = 2.12 95%CI1.59-3.07 p ≤ 0.001). Behaviourally-infected adolescents were more likely to be depressed (B = 0.81 p ≤ 0.001), anxious (B = 1.36 p ≤ 0.001), report internalised stigma (B = 0.91 p ≤ 0.001), express suicidal ideation (OR = 3.65 95%CI: 1.96-6.82 p ≤ 0.001) and report excessive substance use in the past year (OR = 9.37 95%CI5.73-15.35 p ≤ 0.001). Being older explained most of these differences, with female adolescents living with HIV more likely to report suicidal ideation. However, behaviourally-infected adolescents were more likely to report substance use (OR = 2.69 95%CI: 1.48-4.91 p = 0.001), depression (B = 0.406, p = 0.022), anxiety (B = 1.359, p ≤ 0.001), and internalised stigma (B = 0.403, p = 0.007) in multivariate regression analyses, controlling for covariates. Moderation analyses (adjusting for multiple testing) suggest that behaviourally-infected HIV-positive adolescents who are also maternal orphans are more likely to report higher rates of depression (B = 1.075, p < 0.001). These notable differences by mode of infection suggest that studies which conflate HIV-positive adolescents may blur the clinical and psychological experiences of these two different sub-populations. Drivers of non-adherence, poor retention in care, and mental health problems may differ by mode of infection, requiring tailored interventions. Health and social service provision, if it is to be effective, needs to address these different youth profiles to ensure optimal adherence, development and wellbeing throughout the life course