Jingru Han, Xiao Li, Man Li, Xuanxuan Wu
Background With the proliferation of short video platforms such as BiliBili and TikTok, public reliance on these platforms for medical information has increased substantially. However, the absence of standardized content regulation raises serious concerns about misinformation, oversimplification, and variable quality in health communication. Radiotherapy, a cornerstone of cancer treatment alongside surgery and chemotherapy, is particularly vulnerable to such information quality issues due to its technical complexity and limited public understanding. This necessitates systematic evaluation of the scientific accuracy and reliability of radiotherapy content on these platforms. Methods In this cross-sectional study, the top 100 Chinese-language videos related to “放射治疗” (radiotherapy) were collected from BiliBili and TikTok (total n = 200). Video quality and reliability were assessed via the Global Quality Score (GQS) and modified DISCERN tools. Nonparametric tests and Spearman correlation analyses were applied. Two independent radiotherapy specialists evaluated the content, with a third resolving discrepancy. Results Overall video quality and reliability were moderate (median GQS = 3; DISCERN = 3). BiliBili demonstrated higher DISCERN scores ( p < 0.05), reflecting superior reliability, whereas TikTok had marginally higher GQS scores. The BiliBili videos were significantly longer (median: 1,391.5 s vs. 98 s) and featured more systematic content, whereas the TikTok videos presented greater engagement (e.g., likes, shares, collects, comments). A positive correlation between video duration and DISCERN score was observed for BiliBili ( R = 0.47, p < 0.0001), whereas TikTok showed a similar trend for GQS ( R = 0.49, p < 0.0001). There was a lack of significant associations between interaction metrics and quality scores. Conclusion This study evaluated 200 radiotherapy videos on BiliBili and TikTok. BiliBili showed higher reliability (DISCERN), whereas TikTok excelled in terms of user engagement. Recommendations include optimizing scientific communication, platform quality-based algorithms prioritizing authoritative content, and enhancing public media literacy. The findings can guide improvements in digital medical education.