Examining concordance of sexual-related factors and PrEP eligibility with HIV risk perception among adolescent girls and young women: cross-sectional insights from DREAMS sites in Kenya, Malawi, and Zambia

Abstract Background HIV risk perception is an important cognition for prevention, theoretically engendering service-seeking and risk-reduction behaviors, but its composition remains poorly understood. We examined country-specific correlates of self-appraised HIV exposure risk among sexually active adolescent girls and young women (AGYW, aged 15–24 years) without HIV in Kenya, Malawi, and Zambia. We also explored overlaps between self-appraised HIV exposure risk and pre-exposure prophylaxis (PrEP) eligibility to identify engagement opportunities. Methods We analyzed cross-sectional data (2016/2017) to estimate sexual-related correlates of self-appraised HIV exposure risk (likely vs. not, temporally framed as “ever”) using log-Poisson models with robust standard errors. For sexual-related factors with an unadjusted p ≤ 0.10, individual adjusted models were fitted, controlling for sociodemographic and cognitive factors with an unadjusted p ≤ 0.10. PrEP eligibility was defined using national guidelines; since conditional criteria are in Malawi’s (age-disparate sex + ever-pregnant) and Zambia’s (multiple partners + condomless sex) guidelines, we also assessed PrEP eligibility after decoupling these factors. Results Few AGYW reported likely HIV exposures (Kenya [N = 746]: 15.7%, Malawi [N = 1348]: 46.2%, Zambia [N = 349]: 9.5%) despite ubiquitous HIV risk (98.7%, 99.8%, and 98.9% of Kenyan, Malawian, and Zambian AGYW reported ≥ 1 sexual-related factor). However, the adjusted models found some actual-perceived risk concordance. Positive correlates of self-appraised likely HIV exposures included partner(s)’ likely HIV exposure (all countries); partner(s)’ unknown HIV status and other partners (Kenya, Malawi); STI symptoms and partner(s) living outside the community (Kenya); non-partner sexual violence (Zambia); and transactional sex, multiple partners, pre-coital alcohol use, and physical/sexual intimate partner violence (Malawi). Per national guidelines, PrEP eligibility criteria differentially identified HIV risk (Kenya: 93.6%, Malawi: 53.3%, Zambia: 44.6%), and self-appraised likely HIV exposures were low among PrEP-eligible AGYW (Kenya: 16.5%, Malawi: 48.5%, Zambia: 18.8%). Decoupling Malawi’s and Zambia’s conditional PrEP criteria could increase risk identification to > 85% and potential engagement by ~ 70% and ~ 30%, respectively. Conclusions AGYW's HIV risk perceptions were mostly influenced by factors beyond their locus of control. Conditional PrEP eligibility criteria may inhibit AGYW’s access and uptake in some settings: countries should consider decoupling these factors to minimize barriers. Intersections between autonomy, behaviors, and perceptions among AGYW in gender-inequitable settings warrants further investigation.