Understanding Viewer Behaviors to Optimize Video-Based Learning in Nerve Surgery
William M. Padovano, BS1; Andrew Yee, BS1; Ida K. Fox, MD1; J. Henk Coert, MD, PhD2; Susan E. Mackinnon, MD1
1Department of Surgery, Division of Plastic and Reconstructive Surgery, Washington University School of Medicine, St. Louis, MO, 2Department of Plastic-, Reconstructive- and Hand Surgery, Utrecht University Medical Center, Utrecht, Netherlands
Introduction: Video-based learning (VBL) is increasingly prominent in the dissemination of surgical knowledge and standardization of modern surgical training. While multiple societal and commercial groups offer surgical video portals, their educational value remains understudied and evidence-based principles for video design underutilized. In 2016, we launched an online VBL platform borrowing design principles from massive open online courses (MOOCs), such as Khan Academy and edX, to promote audience engagement, information retention, and surgical performance. The objective of this study was to use video analytics to establish best practices for video design in surgical education.
Materials and Methods: Using online analytic tools, we performed a retrospective assessment of procedural videos hosted on our platform from 1/01/16 to 12/31/17. Extracted data included video views, length, watch-time (average minutes played), and engagement (watch-time divided by video length). Following evidence-based principles for MOOC video design, each surgical procedure included abridged and extended video editions. A subset of this collection was also published on YouTube. Simple and multiple regression models were developed to identify associations between abridged and extended video editions for each video collection.
Results: The 88 procedural videos on our educational platform achieved a cumulative 12,568 views and 81,987 minutes of watch-time. Of the 44 surgical topics, 20 were nerve decompressions and 16 were nerve transfers. YouTube shared 36 of these videos with a total of 269,201 views and 1,414,059 minutes of watch-time. On average, abridged videos on our platform were 10 minutes in length and achieved a relatively high engagement of 57%. Compared to their paired extended editions, they were 53% shorter, received 233% more views, and had 26% higher relative engagement. From multiple regression, significant predictors of extended edition engagement were abridged edition engagement (p<0.0001, coefficient 0.66, 95% CI 0.39 to 0.94) and extended edition length (p=0.039, coefficient -0.30, 95% CI -0.59 to -0.02). Similar trends were identified in the YouTube collection.
Conclusions: Our surgical videos achieved high user engagement overall but demonstrated reduced audience engagement with video length as identified in multiple MOOC studies. Furthermore, extended editions achieved better engagement than expected when associated with high-performing abridged editions, indicating shared user behavior. Our experience suggests that evidence-based principles for non-surgical educational video design are applicable in nerve surgery education.
Back to 2019 Absteracts