Work

Quality, reliability, transparency, and disease-specific content of erectile dysfunction videos on four Chinese platforms: a cross-sectional content analysis

Year: 2026
Type: preprint
Abstract: Background: Videos about erectile dysfunction (ED) are a growing source of patient information on Chinese short-video platforms, yet their reliability, source transparency, and disease-specific clinical content have not been jointly assessed. Methods: On October 8, 2026, we retrieved the first 100 default-ranked videos for the Chinese term "勃起功能障碍" ("erectile dysfunction") on each of four platforms; after deduplication and eligibility screening, 348 videos were included. Two physicians independently rated each video using the Global Quality Score (GQS), the original DISCERN reliability items 1-8, adapted JAMA benchmarks, and a pre-specified ten-item ED checklist informed by current guidelines. Analyses used consensus scores, rank-based comparisons, within-platform correlations with multiplicity correction, and HC3-adjusted regression. Results: Platform comparisons depended on the assessment domain: Bilibili had the highest median DISCERN-8 score (31), whereas TikTok had the highest median adapted JAMA score (4); all platform tests were significant (P < 0.001). Correct mention of first-line PDE5 inhibitor treatment, medication safety, and other treatment options occurred in only 17.8%, 19.5%, and 8.9% of videos. Under the study rubric, traditional-medicine videos had higher unadjusted odds of exaggerated efficacy claims (OR 12.49, 95% CI 5.66-27.58) and primary-treatment substitution (OR 108.00, 95% CI 13.17-885.84). Professional uploaders scored lower on unadjusted DISCERN-8 but higher after adjustment (β 1.69, 95% CI 0.60-2.78), consistent with compositional confounding. Only four of 96 video-characteristic–score correlations survived within-platform correction, all in TikTok. Conclusions: Generic and disease-specific instruments identified domain-dependent platform differences in this sample. Treatment-related content was limited, and there was limited evidence of a consistent association between engagement and information quality.
Language: English
Subfield: Health
Open Access status: green
OpenAlex ID: W7236777579
Cites: 37
Cited by:
Related to: 10