2024 Theses Doctoral

Trends, Determinants, and Effects of HIV Risk Perception among Adolescents and Young Adults: Insights to Improve HIV Prevention Strategies in Eastern and Southern Africa

Heck, Craig J.

BACKGROUND:In Eastern and Southern Africa (ESA), adolescent girls’ and young women’s (AGYW, aged 15-24 years) disproportionate HIV risk is fueled by age-disparate sex (partners >5 years older). Older men typically acquire HIV from same-aged partners; they then transmit HIV to AGYW, who then transmit HIV to older men as they age, creating a cycle of transmission. Adolescent boys and young men (ABYM, aged 15-24 years) have lower HIV incidence rates, and their risk starts increasing around age 20 due to same-aged partners and age-disparate sex. Developing ABYM’s preventative behaviors may proximally and distally reduce their and AGYW’s HIV risk. HIV risk perception (HIV RP) is a theoretical motivator for initiating and using risk-reduction strategies. However, this relationship remains temporally and empirically equivocal in AGYW and ABYM, who are in a developmentally and socially dynamic stage of life. To increase and improve HIV risk-reduction strategies among AGYW and ABYM in ESA, this dissertation aimed to elucidate temporal trends, determinants, and effects of HIV RP.

METHODS:This dissertation employed three aims to examine 1) longitudinal HIV RP literature, 2) HIV RP trajectories and predictors, and 3) HIV RP’s association with health outcomes. Secondary analyses used data collected (2010 [T1], 2011 [T2], 2012 [T3]) from the control arm (714 AGYW, 635 ABYM) of a school-based trial in KwaZulu-Natal, South Africa.

Aim 1 evaluated longitudinal quantitative studies examining HIV RP among AGYW and ABYM in ESA. A scoping review synthesized data from peer-reviewed articles published between 1995-2022 to elucidate studies’ geo-spatial distributions, temporal data patterns, sample characteristics, and HIV RP definitions, analytic findings, and theoretical underpinnings. A proposed HIV RP conceptual framework was created by synthesizing findings and theories.

Aim 2 described gender-specific patterns and determined predictors of HIV RP in AGYW and ABYM. Trajectories were created based on cumulative reports of high HIV RP. After, these trajectories were characterized using sociodemographic, knowledge, attitudinal/perceptual, biological, experiential, and behavioral factors. Using models stratified by ever-sex status, overall predictive analyses explored factors prospectively associated with high HIV RP.

Aim 3 tested the hypothesis that high HIV RP increased chances of HIV testing and decreased cumulative acquisition of HIV/HSV-2. Directed acyclic graphs identified potential confounders using current literature. Predictive analyses estimated high HIV RP’s gender-specific confounder-adjusted association with HIV testing and HIV/HSV-2 incidence over 1-year periods.

RESULTS:In Aim 1, synthesis of 38 longitudinal studies offered limited insights on AGYW’s and ABYM’s HIV RP due to geographic homogeneity, abundance of intervention-related studies, and minimal reporting of RP definitions and HIV status. Most (~75%) studies analyzed data collected before 2015, highlighting the evidence gap of HIV RP in current epidemic conditions or prevention options. Studies assessed AGYW when their risk is peaking and provided little detail on pre-risk periods; ABYM were rarely included in analyses. Elevated HIV RP was mostly positively associated with AGYW’s self-initiated outcomes, but HIV RP’s formulation was understudied.

Aim 2 found HIV RP was temporally dynamic, as most (60% AGYW, 65% ABYM) changed their level of HIV RP at least once. AGYW’s and ABYM’s Persistent (high HIV RP at all timepoints) trajectories were primarily differentiated by negative prevention attitudes. While sexual risk behaviors were relatively high in AGYW’s and ABYM’s Persistent trajectories, HSV-2 and becoming/getting others pregnant increased in lower RP trajectories, implying potential contradictions between actual and perceived risk. Analyses showed sociodemographic factors (grade repetition, mother-led households) mostly predicted high HIV RP in AGYW, while prevention attitudes differentially (i.e., attitudes and associations largely shifted from negative to positive over time) affected ABYM’s high HIV RP. Behaviors (transactional sex, not knowing partners’ status, older partners, alcohol use [ABYM only]) predicted high HIV RP in both groups.

Aim 3 highlighted that age, grade, financial insecurity, HIV/pregnancy-related knowledge, attitudes, perceptions, experiences, and service-seeking and risk-taking behaviors potentially confounded relationships between HIV RP and outcomes (HIV testing, HIV/HSV-2 acquisition). AGYW with high HIV RP at T1 were more likely to HIV test (aRR=1.48 [1.12-1.95]) and acquire HIV/HSV-2 (aRR=1.83 [1.03-3.25]) by T2, associations that did not persist in T2-T3’s analyses. There were no significant associations between any factor and HIV RP in ABYM.

CONCLUSION:This dissertation highlights the temporal and associative complexities of HIV RP among AGYW and ABYM in ESA. HIV RP in these groups was variable and primarily predicted by sociodemographic (AGYW), attitudinal (ABYM), and behavioral factors, elucidating need and opportunities for multi-level, age- and risk-appropriate interventions. Structural insecurities and age- and gender-driven power differentials may inhibit AGYW’s prevention behaviors. For both groups, earlier intervention would reach AGYW before their risk escalates and reinforce and/or instill ABYM with potentially lasting assets. Additional high-quality, detailed studies are needed to combat the current evidentiary dearth, optimize prevention interventions, and increase the likelihood of ending HIV/AIDS by 2030.