The Impact of Peer-Led Community Testing Models on Early HIV Diagnosis Among Key Populations: A Systematic Review and Meta-Analysis

Abstract

Introduction: HIV testing remains a critical entry point to the care cascade. Peer-led community testing models have emerged as promising approaches to improve early diagnosis among key populations, yet evidence synthesis regarding their effectiveness remains limited. This study aimed to systematically review and meta-analyse the effectiveness of peer-led community testing interventions on HIV testing uptake among key populations.

Methods: Systematic search of PubMed, Embase, Global Health, CINAHL, and Web of Science (2014-2024) for randomized controlled trials and quasi-experimental studies. Screening, quality assessment (Cochrane Risk of Bias tool), and meta-analysis using Hedges' g as the pooled effect measure. GRADE methodology applied to assess certainty of evidence. Sensitivity analysis excluding systematic reviews conducted.

Results: Seven studies (52,698 participants; 31,381 intervention, 21,317 control) met inclusion criteria. Pooled standardized mean difference was 0.4834 (95% CI: 0.3671 to 0.5997; p < 0.001), indicating a moderate effect of peer-led testing on uptake. Heterogeneity was moderate (I² = 55.23%, τ² = 0.0111). GRADE analysis indicated moderate certainty of evidence, downgraded for study design heterogeneity but upgraded for large absolute effect and consistency. The 95% prediction interval ranged from −0.05 to 1.02, suggesting that while most future settings would benefit, some may show minimal effects. Sensitivity analysis excluding two systematic reviews (5 primary studies: g = 0.58) remained statistically significant. Subgroup analysis demonstrated differential effectiveness by setting and intervention modality.

Conclusion: Peer-led community testing models yield moderate improvements in HIV testing uptake among key populations. Translation to policy implies approximately 9,600 additional individuals tested per 100,000 reached, potentially yielding 288-480 additional diagnoses at 3-5% prevalence. However, sustainability, linkage to care, and integration within health systems remain critical implementation challenges. Further research addressing long-term retention, cost-effectiveness, and contextualization to specific key population and geographic settings is warranted.