Short-form video platforms as a source of ankylosing spondylitis information: a cross-sectional content analysis

Short-video platforms have become major channels for health information dissemination, yet the quality and reliability of content on ankylosing spondylitis (AS) remain underexplored.

This study aimed to systematically evaluate the quality, reliability, and characteristics of AS-related short videos on three major Chinese platforms: TikTok, Bilibili, and rednote.

A cross-sectional content analysis was conducted on 300 videos (100 per platform) collected in November 2025. Video uploaders were categorized as professional physicians, non-professional physicians, individual users, or institutions. Four validated instruments-modified DISCERN (mDISCERN), Global Quality Scale (GQS), Journal of the American Medical Association (JAMA) criteria, and Video Information Quality Index (VIQI)-were used to assess reliability, completeness, and production quality. User engagement metrics, including likes, shares, comments, and collections, were also analyzed.

Most videos were uploaded by professional physicians and primarily focused on clinical manifestations and treatment. Videos presented in animated or lecture-style formats demonstrated higher information quality and reliability, whereas casually recorded videos consistently scored lower across multiple assessment tools. Overall, the quality of AS-related short videos was moderate. In unadjusted analyses, collections were positively associated with information quality, suggesting that users may preferentially retain more informative content. However, after accounting for platform characteristics and exposure-related factors, information quality was not a consistent independent driver of relative engagement, and mDISCERN showed an inverse association with standardized collection levels. Other engagement indicators, including likes, comments, shares, and follower counts, showed weak or inconsistent relationships with video quality.

Although AS-related short videos are predominantly produced by physicians, the overall quality remains suboptimal. Information quality appears to influence certain user behaviors, such as content saving, but does not consistently translate into higher overall engagement. These findings highlight the limitations of using engagement metrics as proxies for content quality. Users should prioritize content from verified medical professionals, and platforms may consider integrating quality-oriented indicators and improving certification systems to enhance health information dissemination.