Optimizing YouTube for Medical Education: An Analysis of Popular Cardiovascular United States Medical Licensing Examination (USMLE) Step 1 Videos.

Background Electronic learning (e-learning) is an important component of medical education, improving learner engagement and satisfaction. YouTube (Google LLC, Mountain View, CA) is a widely used educational platform due to its accessibility, low cost, and variety of content delivery styles. However, viewer preferences for different content delivery styles across medical education topics remain poorly characterized. Identifying popular delivery styles by topic may inform how educators curate and recommend supplemental videos. Objective This study's objective is to describe the content delivery styles and engagement metrics of the most popular YouTube videos across five United States Medical Licensing Examination (USMLE) Step 1 cardiovascular content areas (anatomy, embryology, physiology, pathology, and pharmacology) and to assess whether the distribution of delivery styles varied by topic. Engagement was measured using publicly available popularity indicators (views, likes, comments) rather than direct measures of educational effectiveness. Methods In this cross-sectional, observational study, we searched YouTube via Google Chrome (Google LLC, Mountain View, CA) in incognito mode on November 5, 2024. For each topic, results were filtered by "Video" type and sorted by "Most Viewed," and the top 30 videos per content area were included (N = 150). Two pairs of authors independently collected engagement metrics, channel characteristics, and delivery-style classifications; a third author adjudicated discrepancies. Descriptive statistics summarized delivery-style distributions; a Pearson chi-square test assessed the association between topic and delivery style, and Kruskal-Wallis tests compared engagement across styles within each topic. Results We assessed 150 videos. Whiteboard-style videos were the most common format in embryology (14, 46.7%), physiology (14, 46.7%), pathology (11, 36.7%), and pharmacology (18, 60.0%), while in anatomy, whiteboard (nine, 30.0%) and animation (eight, 26.7%) styles were used in similar proportions. No statistically significant association was observed between topic and delivery style (Pearson χ²(24) = 27.72, p = 0.272), and Kruskal-Wallis tests did not detect statistically significant differences in median likes across delivery styles within any topic (all p > 0.05). Descriptive patterns suggested that animation had the highest mean views in anatomy, physiology, and pathology, and that lecture-style videos had the highest mean views in pharmacology, but these patterns were not statistically significant and should be interpreted cautiously. Conclusions Among the most-viewed YouTube videos in USMLE Step 1 cardiovascular topics, no single delivery style was universally preferred, and engagement patterns varied descriptively by content area. Because our analysis relied on publicly available popularity indicators rather than validated measures of learning, these findings should be interpreted as hypothesis-generating rather than as evidence of educational effectiveness. The results may inform future studies that directly compare delivery styles using standardized content and knowledge-outcome measures.