Quality and reliability of osteoporosis-related health information on short-video platforms: a multi-source cross-sectional study

Abstract Background Osteoporosis (OP) affects approximately 145.86 million people in China, imposing substantial disease and societal burden. Effective management requires evidence-based patient education. Short-video platforms dominate health communication, yet no study has systematically compared OP content quality across major Chinese platforms. Method This cross-sectional study analyzed 300 OP-related short videos (100 each from TikTok, Rednote, and Bilibili). A single trained rater blindly assessed each video using GQS, mDISCERN, and JAMA benchmark scores. Engagement metrics (likes, comments, shares, favorites) and uploader characteristics were extracted. Group comparisons used Kruskal–Wallis tests with Dunn’s post hoc corrections; associations were evaluated via nonparametric correlations. Results Significant inter-platform disparities emerged. TikTok videos showed highest engagement but shortest duration (median 146.5 s) and lowest transparency (JAMA 0.98 ± 0.471). Bilibili videos had longest duration (median 352 s) and highest quality scores (GQS 3.56 ± 0.81, mDISCERN 2.07 ± 0.998, JAMA 1.28 ± 0.842). Rednote showed lowest overall quality (GQS 2.76 ± 0.878, mDISCERN 1.53 ± 0.904, JAMA 1.05 ± 0.642). Quality scores correlated weakly with popularity metrics (|ρ| < 0.10). Institutional uploaders scored higher than individual creators (all P < 0.01). Popularity was largely dissociated from clinical credibility; content source and video duration were more closely associated with information quality. Conclusion OP information on major short-video platforms is suboptimal, varies markedly across platforms, and shows a striking disconnect between popularity and credibility. These findings underscore an urgent need for platform-level quality stewardship, algorithm redesign prioritizing evidence-based content, and public empowerment through digital health literacy.