Knee osteoarthritis (KOA) is a common long-term condition that is becoming more of a problem, particularly in China. This study assesses the quality, reliability, and user interaction of content related to KOA across TikTok, Kwai, and Bilibili to reveal discrepancies between platforms and provide guidance on improving online public health messaging and content regulation.

This cross-sectional study used standardized keywords to retrieve and assess the top KOA-related videos from TikTok, Kwai, and Bilibili on a single day. After eliminating duplicates, we reviewed 300 videos (100 per platform). Video features and engagement metrics were recorded. Information quality was evaluated using the following established tools: the DISCERN reliability section, the Global Quality Score (GQS), content completeness score (CS, 0-10), and engagement score (ES, 0-4). The uploader types were classified, and statistical analyses included Kruskal-Wallis H, chi-square, and Spearman correlation tests.

 Significant differences were found across platforms. Bilibili videos were the longest (median, 402 s) and scored highest in reliability (DISCERN median, 32), overall quality (GQS mean, 4.3), and completeness (CS mean, 7), with the highest percentage of certified medical professionals (38.0%). In contrast, TikTok videos were the shortest (median, 86 s) but had the highest user engagement (likes, comments, and shares) and engagement score (median, 3.6). TikTok and Kwai content was less complete, mainly focusing on symptoms and treatment while omitting etiology and prognosis. A significant negative correlation was found between quality/completeness metrics and user engagement , whereas the engagement score positively correlated with user interaction.

positively correlated with user interaction.

The major short-video platforms in China have distinct ecosystems for KOA information, indicating a trade-off between scientific accuracy and popularity. Algorithmic recommendations tend to prioritize content that engages users over content that is scientifically accurate, thereby underscoring a significant challenge in governing digital health initiatives. Improving public health communication requires refining platform algorithms to emphasize quality and prompting health professionals to develop engaging, accurate content for high-reach platforms.