YouTube content on shoulder arthroplasty is highly viewed but is of poor educational quality

Background

YouTube videos on shoulder arthroplasty cover a broad range of topics, from patient testimonials to surgical techniques. Inherent to this platform, there are no quality control measures to monitor this content. The purpose of this study is to evaluate the quality and popularity of videos on shoulder arthroplasty on YouTube, using both previously described scores and a proposed novel shoulder arthroplasty–specific score assessing educational content.

Methods

A search was performed using the keywords “shoulder arthroplasty” and “shoulder replacement” on YouTube. Videos were sorted by relevance and the first 50 videos for each keyword were included in a playlist. The videos in the playlist were then sorted by popularity and the first 50 were included for analysis. Videos in a non-English language, duplicate videos, or those lacking audio were excluded. Video source, content, time since upload, duration, like and dislike count, were inputted for analysis. The view ratio and video power index were used to assess for popularity. Quality, reliability and educational content were evaluated with the Global Quality Score (GQS), Journal of the American Medical Association (JAMA), and the novel Shoulder Arthroplasty Video Content (SAVC) score, respectively.

Results

The 50 videos had an average of 285,375 views and a mean duration of 10.8 minutes. Video content was comprised of primarily surgical technique (46%) and patient experience (44%), with most of the videos uploaded from university-affiliated physicians (30%), non-affiliated physicians (28%), and medical (i.e., animations from health websites) (26%) sources. Mean GQS (i.e., quality) and JAMA (i.e., reliability) scores for all videos were 3.2 ± 1.2 out of 5 and 2.4 ± 1.4 out of 4, respectively. Median GQS was comparable across sources, content, and procedure type. Although median JAMA scores were similar among procedure and content type, academic and medical sources showed significantly higher scores among video source categories (P = .04). The mean SAVC score was 19.5 ± 10.7 out of 45 possible points. While there was no significant difference in SAVC scores when categorized by procedure type, a significant difference was observed when assessed by source (P = .02) and content (P = .01). Specifically, medical sources had the highest median score for video source and patient experience for content.

Conclusion

Shoulder arthroplasty videos on YouTube reach a considerable audience, but the educational content and reliability are low. Understanding this can help caution patients on the limitations of educational content on YouTube.