Abstract
INTRODUCTION: Social network strategies, in which social networks are utilized to influence individuals or communities, are increasingly being used to deliver human immunodeficiency virus (HIV) interventions to key populations. We summarized and critically assessed existing research on the effectiveness of social network strategies in promoting HIV self-testing (HIVST). METHODS: Using search terms related to social network interventions and HIVST, we searched five databases for trials published between 1st January 2010 and 30th June 2023. Outcomes included uptake of HIV testing, HIV prevalence and linkage to antiretroviral therapy (ART) or HIV care. We used network meta-analysis to assess the uptake of HIV testing through social network strategies compared with control methods. A pairwise meta-analysis of studies with a comparison arm that reported outcomes was performed to assess relative risks (RR) and their corresponding 95% confidence intervals (CI). RESULTS: Among the 4496 manuscripts identified, 39 studies fulfilled the inclusion criteria, including one quasi-experimental study, 22 randomized controlled trials and 16 observational studies. Networks HIVST testing was organized by peers (distributed to known peers, 15 studies), partners (distributed to their sexual partners, 16 studies) and peer educators (distributed to unknown peers, 8 studies). Among social networks, simulating the possibilities of ranking position, peer distribution had the highest uptake of HIV testing (84% probability), followed by partner distribution (80% probability) and peer educator distribution (74% probability). Pairwise meta-analysis showed that peer distribution (RR 2.29, 95% CI 1.54-3.39, 5 studies) and partner distribution (RR 1.76, 95% CI 1.50-2.07, 10 studies) also increased the probability of detecting HIV reactivity during testing within the key population when compared to the control. DISCUSSION: All of the three social network distribution strategies enhanced the uptake of HIV testing compared to standard facility-based testing. Linkage to ART or HIV care remained comparable to facility-based testing across the three HIVST distribution strategies. CONCLUSIONS: Network-based HIVST distribution is considered effective in augmenting HIV testing rates and reaching marginalized populations compared to facility-based testing. These strategies can be integrated with the existing HIV care services, to fill the testing gap among key populations globally. PROSPERO NUMBER: CRD42022361782.