Abstract
BACKGROUND: Cancer patients increasingly use YouTube for nutritional guidance, yet information quality varies substantially. Existing text-based assessment tools fail to capture audiovisual content characteristics. This study aimed to (1) develop a video-specific assessment tool, (2) evaluate German-language YouTube videos on cancer nutrition, and (3) identify quality indicators for laypersons. METHODS: A 20-criteria assessment tool integrating established instruments and video-specific elements was developed. The first 30 YouTube videos on cancer nutrition were systematically evaluated. Spearman correlation and Kruskal-Wallis tests identified associations between video characteristics and quality scores. Interrater reliability was assessed. RESULTS: Intraclass correlation coefficient indicated good to very good interrater reliability (95% CI: 0.87-0.96). Overall video quality was poor (mean: 38.6/60, SD: 5.3). Videos from hospitals (P = .002) and healthcare organizations (P = .006) scored significantly higher than those from independent persons. Videos with clearly formulated goals (r(s) = 0.71, P < .001) and cited references (r(s) = 0.43, P = .019) demonstrated stronger evidence-based content. High-quality videos more frequently addressed missing evidence (r(s) = 0.51, P = .004). Quality scores inversely correlated with likes (r(s) = -0.55, P = .002) and views (r(s) = -0.46, P = .01). CONCLUSION: YouTube videos on cancer nutrition exhibit substantial quality deficits, even from institutional providers. The validated assessment tool identifies observable quality indicators including clear objectives, scientific citations, transparent discussion of evidence gaps, and institutional authorship. However, no single feature reliably predicts quality. Strengthening digital health literacy and improving evidence-based content production and visibility remain essential priorities.