Examining Social Media User Types and the Impact on a Reproductive Health Web-Based Intervention for Young People in Francophone West Africa: Randomized Factorial Design Using Latent Class Analysis

Background: Web-based interventions present an opportunity for addressing reproductive health outcomes among young Francophone West Africans, but little is known about how social media use impacts such interventions. Objective: This study aimed to identify social media use patterns among young French-speaking West Africans and assess the impact on a reproductive health intervention delivered on Facebook. Methods: French-speaking social media users (aged 15-24 years) were recruited from Senegal, Côte d'Ivoire, and Burkina Faso via social media advertisements. Using a factorial design for a multiarm parallel study, participants were randomly assigned via block randomization to 4 conditions: comparison (standard content), peer role models (personal narratives from similarly aged peers), online influencers (popular online figures), and a mixed peer role model and online influencer strategy. Participant blinding was implemented. Respondents completed pre- and post- self-administered online surveys between December 2023 and February 2024. Latent class analysis using maximum likelihood estimation was used to examine social media use patterns and related outcomes: contraceptive knowledge, provider awareness, and online service use. Results: In total, 353 participants were randomized to each group, with a final analytic sample of 262 participants. Two social media user types were identified from the completed study: selective users (daily use of WhatsApp and Facebook) and multiplatform users (daily use of WhatsApp, Facebook, Instagram, YouTube, and TikTok). Being 20-24 years old, compared with 15-19 years, was associated with increased odds of multiplatform use (odds ratio [OR] 3.61, 95% CI 1.46-9.36; P=.008). Furthermore, residing in Senegal was associated with a greater likelihood of multiplatform use, compared with Côte d'Ivoire (OR 0.27, 95% CI 0.08-0.88; P=.03) and Burkina Faso (OR 0.23, 95% CI 0.07-0.76; P=.02). Among selective users (daily users of WhatsApp and Facebook), the peer role model approach demonstrated a 1.84-unit increase in contraceptive knowledge (OR 1.84, 95% CI 0.27-3.41; P=.02), a 1.81-unit increase in provider awareness (OR 1.81, 95% CI 1.15-2.47; P<.001), and a 1.49-unit increase in service use (OR 1.49, 95% CI 0.06-2.93; P=.04), compared with the comparison group. For multiplatform users, the peer role model approachwas associated with a decrease in provider awareness (OR 0.33, 95% CI -0.64 to -0.01; P=.04). Unlike the peer role model approach (n=70), generally, online influencer (n=63) and mixed approaches (n=57) did not yield significant outcomes, relative to the comparison group (n= 72). There were no adverse events reported. Conclusions: This study applies latent class analysis in a novel way for this population: examining how a comprehensive measure of social media use moderates digital interventions. This approach is underused in other social media studies in sub-Saharan Africa. Findings suggest that a one-size-fits-all approach in web-based interventions is not ideal, and online peer-to-peer communication is still an important strategy for certain social media users.