Instagram Video Engagement in Medical Education: Cross-Sectional Study

Background: Social media (SM) has become an essential tool for delivering formal and informal educational content, including medical education. Instagram-based educational initiatives have shown strong engagement and improved learning outcomes in accounts created for research purposes. However, empirical evidence on engagement with real-life account content remains limited. Objective: This study aims to identify engaging types of posts, determine the optimal video length for medical education, and provide evidence-based recommendations to medical educators on SM content. Methods: A cross-sectional study analyzed real-life Instagram medical education account data from Instagram Insights, with a focus on video posts. Insights' post-related data-reached accounts, views, and watch time-were used. Video posts were categorized based on their type (theoretical, clinical, health promotional, or entertaining) and their implementation (use of animation, human models, or medical models). Video length was stratified into 3 categories: ≤60 seconds, 61-120 seconds, and >120 seconds. The Kruskal-Wallis test was used to determine whether video type, implementation, or length were associated with reach, views, or watch time. Results: From May 26, 2020, to May 3, 2024, 125 video posts were analyzed. The median of reached accounts was 5317 (IQR 4007-6244) per video. The median (IQR) of views was 6533 (4708-8601) per video, and relative watch time was 19% (15%-30%). Health promotion and awareness videos had the greatest reach (median 7070, IQR 6093-11,340) compared to theory (median 5265, IQR 4476-5829; P=.006), clinical skills (median 4597, IQR 2001-5688; P=.003), and entertainment videos (median 5605, IQR 3887-6492; P=.04). Videos >120 seconds had higher reach (median 5774, IQR 4754-6636) and views (median 7455, IQR 5321-11,031) compared to videos that were 61-120 seconds (median reach 4514, IQR 2551-5615, P=.02; views 5987, IQR 3727-6970, P=.02). Similarly, videos that were ≤60 seconds had higher reach (median 5484, IQR 4299-6737) and views (median 7189, IQR 5182-10,609) than videos that were 61-120 seconds (median reach 4514, IQR 2551-5615, P=.02; views 5987, IQR 3727-6970, P=.02). However, videos that were ≤60 seconds had significantly higher relative watch time (median 29%, IQR 20%-40%) compared to videos that were 61-120 seconds (median 17%, IQR 14%-20%; P=.003) and videos that were >120 seconds (median 15%, IQR 11%-17%; P=.002). The implementation method did not significantly affect video engagement. Conclusions: Unlike previous studies that have primarily explored the theoretical potential of Instagram or analyzed isolated posts, this study provides a longitudinal real-world analysis of 125 video posts published on an active medical education Instagram account. Our findings demonstrate that shorter video lengths and health promotion content are associated with greater engagement, while diverse implementation types may further support content effectiveness. By examining authentic educational content in SM, this study offers practical, data-driven insights for medical educators seeking to optimize their educational strategies. Future research should explore how different demographic groups engage, thereby enabling more targeted and effective medical education approaches.