Attention and Discourse on Myocarditis Across 6 Chinese Social Media Platforms: An Infodemiology Study

Abstract Background Myocarditis carries substantial morbidity, and the COVID-19 pandemic amplified public and media attention to the condition. In China, where mass vaccination overlapped with successive respiratory seasons, myocarditis settled into a durable high-attention state, yet how that attention spreads across the country's fragmented, multimodal social-media landscape remains poorly understood. We characterized the temporal dynamics, sentiment, cognitive dimensions, and thematic structure of myocarditis-related discourse across six major Chinese platforms over 36 months to derive actionable targets for infodemic management and public-health communication. Methods We conducted an infodemiology study spanning 1 July 2023 to 31 July 2026 across Weibo, Bilibili, Zhihu, Xiaohongshu, Douyin and Kuaishou, applying a frozen 68-term keyword lexicon. Acquisition used the official advanced-search API (Weibo), a signed REST API (Bilibili), and Playwright-driven real-browser automation (remaining platforms). Of 114,999 retrieved records, 112,243 (97.60%) were retained after de-identification, deduplication, relevance filtering and entertainment-noise suppression. Processing comprised jieba segmentation, SnowNLP sentiment scoring (random subsample n = 60,000), TF-IDF plus K-means topic modelling (k = 8) on 29,352 high/medium-relevance records, a nine-class rule-based cognitive-dimension classifier, and alignment to seven influenza-season-aligned phases. Ethics approval: TJ-IRB202404084. Results Public attention followed a stable winter-spring surge and summer-autumn plateau pattern, growing nearly threefold from phase 1 (7,973 records) to phase 6 (23,516). Provenance was highly concentrated: Weibo accounted for 77.9% (87,412 records), followed by Bilibili (7.0%), Xiaohongshu (5.2%), Douyin (5.0%), Zhihu (3.3%) and Kuaishou (1.5%). Sentiment was net positive (mean 0.712; 71.7% positive, 23.2% negative, 5.1% neutral), though most anxious during peak-volume months. Discourse migrated from categorical disease naming toward experiential symptom, diagnosis and recovery narratives. The largest topic cluster was a first-person pediatric care narrative (42.6%); 'myocarditis mistaken for a cold or COVID-19' accounted for 8.4%. High-frequency terms, 'cold' (42,758), 'symptoms' (36,360), 'cough' and 'influenza', revealed persistent respiratory-cardiac semantic conflation. Conclusions The myocarditis infodemic on Chinese social media is large, seasonal, experiential and shaped by recurrent respiratory-cardiac semantic overlap in lay discourse. These findings support influenza-calendar-timed, family-centered, clinician-led communication and argue for embedding myocarditis infoveillance within China's event-based media-surveillance architecture, treating public confusion as both an exploratory signal and a health-literacy resource.