Predictive Modelling’s role in Improving Pre-exposure Prophylaxis (PrEP) Uptake in High-Risk HIV Groups in Africa: An Integrative Scoping Review

This scoping review explores how predictive modelling can strengthen pre-exposure prophylaxis (PrEP) uptake among high-risk populations in Africa, where HIV prevalence remains disproportionately high. Although PrEP is highly effective (40–90%), its uptake and adherence remain suboptimal in LMICs. Predictive modelling provides a promising solution by identifying individuals at elevated risk, enabling targeted, evidence-based interventions. Using Arksey and O’Malley’s framework and PRISMA-ScR strategy, PubMed, Cochrane Library, ProQuest, and Google Scholar were searched for Africa-based studies from 2015–2025. Eligible studies focused on high-risk groups, including men who have sex with men, sex workers, persons who inject drugs, adolescents, and serodiscordant couples, and applied machine learning, regression models, deep learning, and neural networks.

Out of 209 records screened, 10 studies met inclusion criteria. Conducted between 2019–2025, they demonstrated how predictive tools can stratify HIV risk, enhance adherence monitoring, and improve resource allocation. Sixty percent relied on demographic and behavioural data and achieved strong predictive performance, particularly for HIV status prediction (70%). However, stigma, weak health systems, poor integration, and limited data quality still hinder implementation.

The review underscores predictive modelling’s transformative potential to scale PrEP services across Africa. Integrating machine learning, behavioural modelling, and community-based approaches can improve programmatic efficiency, equity, and targeting. Yet substantial gaps persist in translating predictive outputs into actionable interventions, addressing ethical issues, and validating models in diverse, resource-limited settings. Strengthening collaborations between data scientists, healthcare workers, and policymakers are essential to deliver cost-effective, context-specific PrEP services and accelerate HIV prevention efforts across the continent.

KEY MESSAGES

-

What is already known: The use of Predictive modelling for identifying high-risk individuals to improve PrEP targeting, holds substantial promise for reducing HIV incidence among vulnerable groups, yet its integration into African health systems remains constrained by structural, data, and equity barriers.

-

What this study adds : This scoping review demonstrates, for the first time, how diverse predictive modelling approaches like machine learning, deep learning, and clustering applied to epidemiological and behavioural data can enhance PrEP uptake and adherence among high-risk groups in LMIC African settings.

-

How this study could affect research, practice or policy: The review findings highlight priority areas for integrating predictive tools with youth-friendly, community-based, and health system–strengthening strategies to scale cost-effective PrEP delivery, improve adherence, and guide evidence-based HIV prevention policy in Africa.