Evaluating the Accuracy and Quality of the Information in Kyphosis Videos Shared on YouTube

Abstract

STUDY DESIGN: A quality-control YouTube-based study using the recognized quality scoring systems. OBJECTIVE: In this study, our aim was to confirm the accuracy and quality of the information in kyphosis videos shared on YouTube. SUMMARY OF BACKGROUND DATA: The Internet is a widely and increasingly used source for obtaining medical information both by patients and clinicians. YouTube, in particular, manifests itself as a leading source with its ease of access to information and visual advantage for Internet users. METHODS: The first 50 videos returned by the YouTube search engine in response to "kyphosis" keyword query were included in the study and categorized under seven and six groups, based on their source and content. The popularity of the videos was evaluated with a new index called the video power index (VPI). The quality, educational quality, and accuracy of the source of information were measured using the Journal of American Medical Association (JAMA) score, Global Quality Score (GQS), and Kyphosis-Specific Score (KSS). RESULTS: Videos had a mean duration of 397 seconds and a mean number of 131,644 views, with a total viewing number of 6,582,221. The source (uploader) in 36% of the videos was a trainer and the content in 46% of the videos was exercise training. Seventy-two percent of the videos were about postural kyphosis. Videos had a mean JAMA score of 1.36 (range: 1-4), GQS of 1.68 (range: 1-5), and KSS of 3.02 (range: 0-32). The academic group had the highest scores and the lowest VPIs. CONCLUSION: Online information on kyphosis is low quality and its contents are of unknown source and accuracy. In order to keep the balance in sharing the right information with the patient, clinicians should possess knowledge about the online information related to their field and should contribute to the development of optimal medical videos. LEVEL OF EVIDENCE: 3.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 466 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.