中国短视频平台上流感科普视频的信息质量与用户互动分析
Evaluating influenza-related short videos on Chinese platforms: information quality, user engagement, and public health implications
作者对 TikTok(抖音)与 Bilibili 上共 229 个流感主题短视频进行横断面内容分析,结果显示整体信息质量中等偏低,两平台在互动量与质量维度上存在显著差异。
Introduction
Influenza is an acute respiratory infectious disease with global prevalence and is estimated to infect hundreds of millions of people each year, resulting in 3–5 million severe cases and up to 650,000 respiratory-related deaths worldwide1. As a major public health concern, the prevention and control of influenza are critical and rely on timely influenza vaccination and scientifically informed responses to seasonal epidemics and outbreaks2,3. However, public awareness and concern regarding influenza have long been insufficient. The disease is frequently misperceived as equivalent to the common cold, leading to an underestimation of its potential to cause viral or bacterial pneumonia, cardiovascular events, and exacerbations of underlying health conditions4,5,6. During the early stages of the coronavirus disease 2019(COVID-19) pandemic, the disease was characterized by some commentators as merely a “little flu” 7. Such comparisons reflect a common cognitive bias, as COVID-19 and influenza differ substantially in pathogenicity and mortality8. Therefore, improving public understanding of influenza and strengthening risk perception and protective behaviors is essential and will require effective, evidence-based health communication strategies.
Against the backdrop of a rapidly evolving information landscape, mobile internet and social media platforms have become critical channels through which the public accesses health information9,10. Short videos leverage fragmented viewing time and entertainment-oriented formats, thereby attracting substantial user engagement in consuming and interacting with health-related content11,12. In China, Douyin (the domestic version of TikTok) and Bilibili are two video platforms with extensive reach: third-party monitoring data indicate that Douyin’s monthly active users (MAU) surpassed 1 billion in 2025, and Bilibili’s 2025 Q3 financial report documented an average of 117.3 million daily active users (DAU), corresponding to a 9% year-over-year increase13,14. Survey evidence suggests that short-form video has become a common channel for health knowledge acquisition in China; by the end of 2020, approximately 73% of respondents reported having watched health science popularization short videos or health-related livestreams15. However, social media has also become a major conduit for misinformation and disinformation, thereby posing substantial threats to public health16. Prior research shows that health misinformation on social media spans multiple clinical domains, commonly involving vaccines/infectious diseases, cancer, nutrition/diets, and smoking-related content. For example, a systematic review reported that COVID-19 vaccine misinformation on social media most often focuses on safety, efficacy, and necessity, frequently accompanied by conspiracy narratives17,18,19. Evidence from the COVID-19 era has shown that digitally mediated health misinformation can contribute to harmful health behaviors, psychological distress, and erosion of trust in clinical and public health guidance20. In this context, these challenges have increasingly been framed as an “information disorder” in which misleading, distorted, or low-quality health content can circulate at scale and undermine risk perception and evidence-based decision-making21. This environment makes rigorous evaluation of the credibility and practical usefulness of health information shared on short-video platforms particularly important. Moreover, Douyin and Bilibili differ in platform architecture and content ecology, which may shape the production, visibility, and reception of health information. A comparative analysis can therefore clarify whether information quality patterns and engagement–quality relationships vary by platform, yielding actionable insights for platform governance and evidence-based risk communication.
Previous cross-sectional studies have analyzed the quality of short videos concerning diseases such as cervical cancer22, laryngeal cancer23, and spinal cord injury24, and have generally demonstrated that the reliability and completeness of these videos are suboptimal and require improvement. However, to our knowledge, no studies have systematically examined the quality of influenza-related health education short videos on major short-video platforms.
In this study, a cross-sectional content analysis of influenza-themed short videos on the TikTok and Bilibili platforms was conducted to evaluate their information quality and credibility indicators, and to compare platform-specific differences. This study aims to address this knowledge gap by assessing the quality of influenza-related science communication videos and by providing empirical evidence and recommendations for leveraging short-video platforms to enhance public health literacy.
Methods
Data sources and collection
This study was designed as a cross-sectional investigation. All data extraction, content coding, and video quality evaluation were conducted by the study authors. On October 7, 2025, short videos on the TikTok and Bilibili platforms were systematically searched. This date served as the index retrieval date for capturing platform metrics (eg, likes, comments, collections, shares, and uploader follower counts) and for calculating days since upload. To minimize potential bias introduced by personalized recommendation algorithms, newly registered accounts were used to search both platforms using the Chinese keywords “流感” and “流行性感冒” (both meaning “influenza”), which are widely recognized descriptors of influenza in Chinese and were selected to improve topic specificity and reproducibility, consistent with prior short-video content analyses that used disease-name keywords for retrieval22,23,24. Candidate searchterms were generated from common public expressions and platform search suggestions. A pilot search revealed that broader terms (e.g., “cold/fever”) produced substantially more off-topic results, so the final keywords were iteratively refined. Even with these core terms, some non-influenza videos (e.g., COVID-19/common cold) were retrieved and were excluded using predefined criteria. Because users typically browse only the initial pages of search results, only the first 100 videos returned for each keyword on each platform were collected25. This was a purposive, user-exposure–oriented sampling strategy rather than a random sample. The top-ranked search results were selected to approximate the content most likely to be encountered by typical users seeking influenza-related information via platform search, thereby increasing the practical relevance of the findings to real-world information exposure. Subsequently, the retrieved videos were screened according to predefined eligibility criteria. Video screening for duplicates and off-topic content was conducted jointly by two reviewers (XL.Z. and CY.Z.) using the predefined eligibility criteria. Inclusion criteria were as follows: (1) content primarily related to influenza and (2) videos published before the search date and still accessible at the time of screening. Exclusion criteria were: (1) duplicate videos, defined as identical content uploaded multiple times within the same platform; (2) purely commercial advertisements or sales-promotional content; (3)inaccessible videos; and (4) irrelevant or off-topic videos. Notably, no instances of cross-platform duplicate uploads were identified during our screening. Following retrieval, video screening and data extraction were performed, and data systematization, content coding, quality appraisal, and statistical analyses were completed over the subsequent 4 weeks. For each video included in the analysis, the following characteristics were systematically reviewed and recorded: title, duration (in seconds), days since upload(as of the index retrieval date), number of fans of video uploaders, video likes, comments, collections, shares, as well as video source(doctors,non-profit organizations(medical institutions or academic organizations), for-profit organizations, non-professional individuals, news agencies), content(disease knowledge, personal experience, news, diagnosis/treatment process), theme(prevention, epidemiology, pathogenesis/pathology, symptom, diagnosis, treatment), and format (dialogue, monologue, news broadcasting, vlog, animation/action, visual pictures and literature). The content analysis categories (video source, content type, theme, and format) were defined a priori based on the study aims and were finalized through a brief calibration exercise. Two reviewers (XL.Z. and CY.Z.) piloted coding on 20 influenza-related videos not included in the final sample, discussed discrepancies, and finalized category definitions during the first week after the index retrieval date.
Assessment of video quality
The included videos were evaluated using four commonly employed health information quality assessment tools. Although some conceptual overlap exists among the four instruments, they were selected to provide complementary coverage of distinct quality dimensions. The Journal of American Medical Association (JAMA) benchmark focuses on the structural transparency and accountability of the information source. In contrast, the modified DISCERN evaluates the reliability and objectivity of the content, specifically addressing source citation and the balance of information. While the Global Quality Score (GQS) shares some focus on completeness with the aforementioned tools, it is distinct as a single-item, holistic measure designed to capture the overall flow and practicality of the video. Finally, the Patient Education Materials Assessment Tool(PEMAT) uniquely addresses the communication usability (understandability and actionability) of the material, which is not covered by the other three scales. This multidimensional approach ensures a comprehensive evaluation of science communication within the context of information disorder. They have been widely applied in prior studies of online or short-video health information22,23,24.
JAMA benchmark criteria26: The JAMA benchmark is a widely recognized website credibility assessment tool. It evaluates four key dimensions: author, attribution, disclosure, and currency. Each dimension is scored on a scale of 1 point, with a maximum possible score of 4 points. This tool was employed to assess the reliability and timeliness of the information sources presented in the videos.
GQS27: This is a subjective overall quality rating using a 5-point Likert scale (1 = very poor quality; 5 = extremely high quality with smooth information flow). Evaluators provided an overall assessment based on the video’s practicality, accuracy, and completeness.
Modified DISCERN: Derived from the original DISCERN instrument, which was initially developed to judge the quality of health information regarding treatment choices, the modified DISCERN tool was employed to evaluate the reliability and objectivity of the video content28. This adapted version consists of five binary questions, with each worth 1 point for a maximum score of 5 points. The tool systematically assesses whether the content is clear and unbiased, cites valid sources, provides balanced information, offers additional references for patients, and acknowledges areas of scientific uncertainty.
PEMAT29: The two subscales, understandability and actionability, are utilized. PEMAT-U assesses how clearly the video content is presented to a general audience, including whether the language is straightforward, the structure is clear, and excessive jargon is avoided. PEMAT-A evaluates whether the video provides specific, actionable health advice or behavioral guidance. Each subscale comprises multiple items. Each item was rated as “Yes” (1 point) or “No” (0 points) based on the video’s content, and separate scores were calculated on a 100-point scale (Score = Points Earned / Total Items × 100%). Higher PEMAT scores indicate greater understandability or more actionable recommendations.
For interpretability, scores from JAMA (0 to 4), GQS (1 to 5), modified DISCERN (0 to 5), and PEMAT (0 to 100%) were used for relative comparison, with higher values indicating better performance on the corresponding dimension. Detailed scoring protocols are presented in Supplementary Tables S1 to S3.
Scoring process and consistency
The evaluation was performed by two independent raters, XL.Z. and CY.Z., researchers in the Department of Respiratory and Critical Care Medicine with specialized training in health information quality assessment. Prior to formal evaluation, a calibration phase was conducted to ensure the robustness and reproducibility of the rating process. This phase involved using a standardized codebook derived from the scoring protocols of the JAMA benchmark, GQS, modified DISCERN, and PEMAT. The raters first conducted a practice round on a subset of 20 influenza-related videos not included in the final analysis to align their understanding of the criteria. During the formal assessment, each rater independently evaluated all videos, with the option to rewatch them as necessary. For any discrepancies in scoring, structured discussions were held to reach a consensus. If a consensus could not be reached, a third senior researcher (YY.H.) was consulted to provide a final determination.
To quantitatively assess inter-rater reliability, the two-way random-effects, absolute-agreement-type intraclass correlation coefficient (ICC) was calculated. According to established standards, ICC values < 0.5 indicate poor agreement, 0.50–0.75 moderate agreement, 0.75–0.90 good agreement, and >0.90 excellent agreement30.
Statistical analysis
Statistical analysis was performed using SPSS version 27.0 (IBM Corp). Continuous variables are expressed as medians (interquartile range), while categorical variables are presented as frequencies and percentages. As all quantitative variables exhibited skewed distributions, the Mann-Whitney U test was used to compare differences between two groups. Comparisons among three or more groups were conducted using the Kruskal-Wallis test or one-way analysis of variance (ANOVA). Spearman’s rank correlation analysis was employed to assess correlations between variables. Differences were considered statistically significant at P < 0.05.
Sensitivity analyses were performed using an expanded dataset including relevant duplicates and professional information sources(doctors,news agencies and non-profit organization) to ensure the stability of the statistical outcomes.
Results
Overall characteristics
A total of 400 videos were initially retrieved from TikTok and Bilibili (100 results for each of 2 keywords on each platform). After removal of 51 duplicate videos, 120 videos were excluded because they were irrelevant to influenza, promotional in nature, or inaccessible. Ultimately, 229 videos were included in the final analysis (Fig. 1).
This flowchart illustrates the systematic screening process, including the initial identification of videos using keywords, the exclusion of duplicates and irrelevant content, and the final inclusion of 229 videos for quality assessment.
In terms of basic video characteristics, TikTok videos were significantly shorter in duration than Bilibili videos (52.0 s vs. 421.0 s, P < 0.001) and were posted more recently (239.0 d vs. 660.5 d, P < 0.001). TikTok videos received significantly more likes (1851.0 vs. 154.0), comments (159.0 vs. 8.5), and shares (1008.0 vs. 109.0) compared to Bilibili videos (all P < 0.001). Collections showed no significant difference between platforms (500.0 vs. 231.0, P = 0.066). TikTok video uploaders had substantially more fans compared to Bilibili uploaders (347,000 vs. 11,500, P < 0.001) (Table 1).
Inter-rater reliability was good to excellent across all measures. On TikTok, ICCs were 0.93 (95% CI 0.89–0.95) for JAMA, 0.88 (95% CI 0.83–0.92) for GQS, 0.85 (95% CI 0.79–0.89) for modified DISCERN, 0.88 (95% CI 0.66–0.95) for PEMAT-U, and 0.88 (95% CI 0.81–0.92) for PEMAT-A. On Bilibili, ICCs were 0.96 (95% CI 0.94–0.97) for JAMA, 0.88 (95% CI 0.81–0.93) for GQS, 0.95 (95% CI 0.93–0.97) for modified DISCERN, 0.89 (95% CI 0.83–0.92) for PEMAT-U, and 0.88 (95% CI 0.81–0.92) for PEMAT-A.
In terms of video quality scores, TikTok videos demonstrated higher PEMAT-U scores (100% vs. 71.5%) and PEMAT-A scores (67% vs. 33%) compared to Bilibili videos (both P < 0.001). In contrast, Bilibili videos scored higher than TikTok videos on the JAMA score (2.5 vs. 1.0) and GQS score (4.0 vs. 3.0) (both P < 0.05), indicating superior information completeness and reliability. The median modified DISCERN score was identical across both platforms (3.0); however, the distribution of scores differed significantly (P = 0.012). Overall, significant differences were observed between TikTok and Bilibili samples in fundamental video characteristics and video quality scores (Table 2).
Regarding video source, TikTok videos are predominantly uploaded by doctors (45.93% vs. 11.70%), whereas Bilibili features a higher proportion of non-professional individuals (67.02% vs. 18.52%). Among video content, disease knowledge is the most prevalent category (76.86%), with a higher proportion found on Bilibili (93.62% vs. 65.19%). TikTok featured a higher proportion of patient experience videos (16.30% vs. 6.38%) and news (17.78% vs. 0%). The primary video themes were treatment (37.12%) and symptom (30.57%), followed by prevention (19.21%). TikTok showed a stronger focus on treatment (38.52% vs. 35.11%) and symptom (37.78% vs. 20.21%), whereas Bilibili emphasized pathogenesis/pathology (31.91% vs. 1.48%). Regarding video format, TikTok predominantly employed the monologue format(43.70%), whereas Bilibili showed a pronounced preference for animation/action (48.94% vs. 8.15%) (Table 3).
Video quality from different sources and different themes
In terms of video source, the evaluation metrics for videos uploaded by independent users on the TikTok platform (JAMA score, GQS score, modified DISCERN score, PEMAT-U, and PEMAT-A) were significantly lower than those for videos uploaded by doctors and news agencies, with statistically significant differences (all P < 0.01). Further comparisons revealed that videos uploaded by news agencies significantly outperformed those uploaded by doctors in JAMA scores (P < 0.001), while videos uploaded by doctors significantly outperformed those uploaded by news agencies in GQS scores (P < 0.001)(Fig. 2).
The panels display the distribution of quality scores across various video sources: A JAMA score; B GQS score; C modified DISCERN score; D PEMAT-U; and E PEMAT-A. Bar heights represent median values and error bars indicate interquartile ranges. Asterisks denote statistical significance: *p < 0.05; **p < 0.01; ***p < 0.001.
Analysis of video source categories on Bilibili revealed no statistically significant differences in JAMA scores, modified DISCERN scores, or PEMAT-U across different video sources (all P > 0.05). Regarding GQS scores, videos uploaded by non-profit organizations significantly outperformed those uploaded by independent users (P < 0.05). In the PEMAT-A dimension, videos uploaded by doctors significantly outperformed those uploaded by for-profit organizations(P < 0.05) and non-professional individuals (P < 0.05 and P < 0.01, respectively)(Fig. 3).
The panels evaluate content quality across different video sources: A JAMA score; B GQS score; C modified DISCERN score; D PEMAT-U; and E PEMAT-A. Bar heights represent median values and error bars indicate interquartile ranges. Asterisks denote statistical significance: *p < 0.05; **p < 0.01; ***p < 0.001.
Regarding the theme dimension, treatment-themed videos on TikTok significantly outperformed symptom-themed videos in terms of PEMAT-A, modified DISCERN scores, and GQS scores (P < 0.001, P < 0.01, P < 0.05, respectively). Epidemiology-themed videos demonstrated significantly higher JAMA scores and modified DISCERN scores compared to symptom-themed videos (P < 0.01, P < 0.001, respectively). Regarding PEMAT-U, videos on pathogenesis/pathology theme performed the worst, significantly lower than treatment-themed and epidemiology-themed videos (P < 0.05 and P < 0.01, respectively)(Fig. 4). Analysis of Bilibili platform themes revealed that treatment-themed videos demonstrated significantly higher practicality than case-based pathogenesis/pathology-themed videos (P < 0.001)(Fig. 5).
The panels show quality variations across different influenza-related topics: A JAMA score; B GQS score; C modified DISCERN score; D PEMAT-U; and E PEMAT-A. Bar heights represent median values and error bars indicate interquartile ranges. Asterisks denote statistical significance between themes: *p < 0.05; **p < 0.01; ***p < 0.001.
The panels compare the performance of various thematic categories on the platform: A JAMA score; B GQS score; C modified DISCERN score; D PEMAT-U; and E PEMAT-A. Bar heights represent median values and error bars indicate interquartile ranges. Asterisks denote statistical significance between themes: *p < 0.05; **p < 0.01; ***p < 0.001.
Supplementary Tables S4–S7 offer a clear view of the quality assessment results for each video source and theme.
Correlation analysis
Correlation analyses were conducted to evaluate the relationship between the four quality metrics and user engagement indicators, aiming to determine whether higher quality content effectively translates into increased public interaction on TikTok and Bilibili, thereby providing empirical evidence for optimizing health communication strategies.
On the TikTok platform (Table 4), all four information quality assessment tools demonstrated significant positive correlations (P < 0.05). Stronger correlations were specifically observed between JAMA and mDISCERN (r = 0.57), GQS and mDISCERN (r = 0.54), and GQS and PEMAT-A (r = 0.55), suggesting consistent trends across various dimensions of content quality on this platform. Regarding the relationship between scores and basic video characteristics, video duration was positively correlated with PEMAT-A scores (r = 0.36, P < 0.001) but negatively correlated with PEMAT-U scores (r = −0.37, P < 0.001). Days since upload showed negative correlations with GQS, mDISCERN, and PEMAT-A scores (r values: −0.20, −0.25, −0.22, all P < 0.05), suggesting that more recently uploaded videos received higher ratings. However, most interaction metrics (likes, comments, collections, and shares) exhibited no significant correlation with ratings. Only the comments exhibited a mild negative correlation with GQS (r = −0.30, P < 0.001) and mDISCERN (r = −0.22, P = 0.009), suggesting that video engagement does not necessarily correlate with information quality. Fans of video uploaders also exhibited no significant correlation with any rating metric.
On the Bilibili platform (Table 5), significant positive correlations were also observed among the four scoring tools (P < 0.05), with the strongest correlations between GQS and mDISCERN (r = 0.59) and between PEMAT-U and mDISCERN (r = 0.46). The JAMA score also exhibited moderate correlations with all other scoring tools (r = 0.27–0.37). Regarding video features, only the mDISCERN score was positively correlated with days since upload (r = 0.22, P = 0.031), whereas video duration showed no significant relationship with the scores. Interaction metrics, such as likes, comments, collections, and shares, exhibited moderate positive correlations with mDISCERN and PEMAT-U scores (r = 0.21–0.39, all P < 0.05). Notably, mDISCERN scores exhibited stronger correlations with likes (r = 0.36), collections (r = 0.39), and shares (r = 0.36), suggesting that higher-quality videos on Bilibili tend to garner more user feedback. Fans of video uploaders on Bilibili exhibited moderate positive correlations with all ratings (except PEMAT-A) (r = 0.26–0.44, P < 0.05). In contrast to TikTok, this suggests that influential users on Bilibili may be more inclined to publish higher-quality content.
Sensitivity analysis
Sensitivity analyses including 35 relevant duplicate videos (N = 264) and focusing on a professional subgroup (N = 144) yielded results consistent with the primary findings. These evaluations, which excluded 16 irrelevant duplicates during screening, confirmed that platform quality disparities and the quality-engagement gap remained statistically significant. Comprehensive data for these analyses are provided in Supplementary Tables S8 to S13.
The encoding process of a high-performance quality video is shown in Supplementary Material 5.
Discussion
Basic findings
This study analyzed the characteristics and information quality of influenza-themed short videos on TikTok and Bilibili. TikTok exhibited higher user engagement metrics (likes, comments, collections and shares). The overall quality of influenza-related short videos was moderately low, with notable room for improvement in reliability and completeness, particularly on TikTok. Bilibili videos were longer and achieved higher JAMA, GQS, and modified DISCERN scores, whereas TikTok videos showed better PEMAT-based comprehensibility and actionability. This divergence may reflect TikTok’s traffic-driven algorithm favoring brief, practical, and entertaining content, in contrast to Bilibili’s greater emphasis on more rigorous science communication31.
Doctors and non-professional individuals were the predominant uploaders, and most videos focused on disease knowledge, especially treatment and symptoms, whereas prevention accounted for a smaller proportion. Videos uploaded by doctors, news agencies, or non-profit organizations generally showed higher quality than those uploaded by non-professional individuals and for-profit organizations, underscoring the importance of strengthening professional participation in health content creation and oversight.
The correlation analysis further revealed a notable mismatch between video quality and popularity, with clear platform-specific patterns. On TikTok, quality scores showed little to no positive correlation with engagement metrics, and in some cases lower-reliability videos attracted more comments, suggesting that high dissemination does not necessarily indicate high-quality information, corroborating concerns regarding algorithm-driven distortions in health information dissemination32. In the context of information disorder, this may increase the risk that the public equates popularity with authority. In contrast, Bilibili showed more positive associations between quality scores and engagement, indicating a relatively more rational information ecosystem.
Prior studies on TikTok and Bilibili have similarly shown that TikTok content often achieves higher visibility and engagement, whereas Bilibili more often supports longer, more detailed, and relatively higher-quality health communication23. Our findings extend this pattern to influenza-related short videos and suggest that a one-size-fits-all dissemination strategy may be suboptimal.
From a public health perspective, these findings suggest that different platforms may require differentiated content strategies, with more concise and behaviorally focused videos for TikTok and more detailed, evidence-oriented explanatory content for Bilibili. At the same time, platforms should integrate basic quality criteria into recommendation algorithms and prominently label verified medical sources to guide the public toward accurate health behaviors during this era of information disorder.
Cognitive confusion
During the screening of influenza-related short videos, numerous videos labeled as “influenza” were identified in this study, but were actually depicting COVID-19 or the common cold, and were therefore excluded based on our predefined criteria. This phenomenon indicates that both content creators and platforms conflate influenza with other respiratory infections, including COVID-19 and the common cold. The common cold typically presents with mild symptoms confined to the upper respiratory tract, whereas influenza is characterized by sudden onset, severe systemic symptoms, and may lead to high fever and serious complications33. COVID-19 and influenza also share overlapping clinical manifestations, such as cough, fever, and sore throat, making differentiation based solely on symptoms challenging34. However, COVID-19 has a significantly higher mortality rate compared to influenza35. Therefore, although direct discussion of diagnosis was limited in the videos included in our analysis (with only about 3.5% of videos focusing on diagnosis), these videos often placed particular emphasis on differential diagnosis. For example, some medical education videos specifically explained how to distinguish between influenza and COVID-19 or between influenza and the common cold, emphasizing the necessity of viral testing for confirmation when required, rather than relying on self-diagnosis.
The COVID-19 pandemic has heightened the risk of conflating influenza with other illnesses. Our analysis reveals how this misconception is reflected in social media content. Enhancing public awareness about influenza and preventing its conflation with COVID-19 or the common cold is a critical area for future health education efforts.
Treatment misconception
Treatment is a central theme in influenza-related short videos. It was observed that most treatment-focused videos emphasize the critical window for antiviral therapy, noting that the optimal time to initiate antiviral medications, such as the neuraminidase inhibitor oseltamivir, is within 48 h of symptom onset. This is consistent with clinical guidelines, which indicate that starting antiviral treatment early in the disease course significantly enhances efficacy36. An observational study reported that for children hospitalized with influenza and underlying conditions, or those admitted to the ICU, initiating neuraminidase inhibitor treatment (almost exclusively oseltamivir) within 2 days of symptom onset shortened hospital stays compared to delayed or no treatment37. In contrast, explicit messages discouraging inappropriate antibiotic use were comparatively uncommon, and many videos did not clearly differentiate antiviral therapy from antibiotic treatment. Antibiotics should only be used to treat bacterial infections that complicate influenza infection. A multinational survey conducted by the World Health Organization revealed that approximately 64% of respondents mistakenly believed antibiotics could treat colds and influenza, with 61% of Chinese respondents holding this misconception38. This misuse of antibiotics not only fails to benefit influenza patients, but also may cause drug side effects and increase bacterial resistance. These patterns suggest potential gaps in corrective messaging around common public misunderstandings. Influenza-related short videos predominantly focus on antiviral treatment knowledge, with limited emphasis on correcting misconceptions about antibiotic misuse. Proper guidance remains insufficient, and future health communication efforts should prioritize addressing this gap.
Prevention and vaccination
Influenza vaccination is a key measure for preventing seasonal influenza and severe illness39. However, this study found that videos on the prevention theme accounted for a significantly lower proportion and generated less engagement compared to content on symptoms and treatment. Of the 229 videos collected, only approximately 19% featured prevention as one of their themes. Influenza vaccination rates in mainland China have historically remained low, as the vaccine is not included in the national immunization program and requires individuals to pay out-of-pocket. One study indicated that during the 2020–2021 and 2021–2022 influenza seasons, China’s influenza vaccination coverage rates were 3.16% and 2.47%, respectively40. These figures fall significantly below those of many Western countries and also fall short of the WHO’s target of 75% vaccination coverage for high-risk groups41. Multiple factors contribute to this situation, with inadequate awareness and the underestimation of risks undoubtedly playing significant roles42. Our analysis supports this; due to a lack of sufficient preventive content videos, the public rarely encounters critical messages such as influenza vaccines effectively prevent influenza and its severe complications on short-video platforms. Vaccination effectively prevents severe influenza outcomes, including a reduction in ICU admission rates (26% risk reduction) and mortality (31% risk reduction) among adults hospitalized with influenza43,44. Annual routine influenza vaccination is recommended for all individuals aged 6 months and older without contraindications, ideally administered in September or October45. However, this concept remains underrepresented in social media content.
To increase influenza vaccination rates, future efforts should focus on enhancing health education to address gaps in preventive content. Healthcare professionals and authoritative institutions should be encouraged to create more short videos on the safety, efficacy, and risks of influenza to improve public awareness. Platforms should proactively promote authoritative, science-based prevention content during flu season to help users accurately understand the value of influenza vaccines.
Implications for platform health communication
Our study is among the first to systematically examine the quality of influenza-related health communication videos on Chinese platforms. This study highlights the growing role of short-form videos in health communication, especially during an era marked by information disorder. By identifying platform-specific content characteristics and quality gaps, this research provides valuable insights into optimizing health communication strategies tailored to individual platform environments.
Meanwhile, Our study contributes to the broader conversation about health misinformation in the digital age, demonstrating the need for improved regulation of health content on social media platforms. The differences between TikTok and Bilibili highlight the importance of adapting content strategies to platform dynamics, ensuring that health communication is both engaging and evidence-based.
Limitations
First, only two representative platforms(TikTok and Bilibili) were selected, and a nonrandom, top–search-results sampling strategy was used, so the findings may not be generalizable to other platforms or to all influenza-related videos. Second, this study analyzed videos available as of October 7, 2025, representing a cross-sectional observation. Video content evolves over time, and newly emerging flu-related topics may not yet be represented in the data. Third, although the information quality ratings employed objective criteria, they still contain subjective elements. Finally, the assessment instruments used (JAMA, GQS, modified DISCERN, and PEMAT) are not designed to verify the factual accuracy of specific medical statements. Accordingly, this study did not conduct independent fact-checking against clinical guidelines, and inaccurate or misleading claims could not be comprehensively identified or quantified.
Conclusion
This study performed a cross-sectional analysis of the information quality and public engagement in influenza science popularization short videos on the TikTok and Bilibili platforms. The study found that current online influenza science communication exhibits significant disparities in quality and notable platform differences. Although social short videos offer substantial opportunities for disseminating health knowledge, significant improvements are still required in information reliability and comprehensiveness. Enhancing the quality of short-form science communication requires not only the efforts of content creators but also the collective advancement of platform regulation, policy support, and public media literacy. Notably, because the applied tools assess credibility indicators and public health usefulness rather than factual accuracy, future work incorporating guideline-based fact-checking is needed to quantify the prevalence and types of inaccurate medical claims.
Data Availability
The anonymized version of the data was offered, including video dates, engagement metrics, content labels, and quality assessment scores, will be made available.
References
Iuliano, A. D. et al. Estimates of global seasonal influenza-associated respiratory mortality: a modelling study. Lancet 391, 1285–1300 (2018).
World Health Organization. Global influenza strategy 2019–2030. https://www.who.int/publications/i/item/9789241515320; 2019 [accessed 13 November 2025].
World Health Organization Vaccines against influenza: WHO position paper – May 2022. Wkly Epidemiol Rec 97, 185–208 (2022).
Lafond, K. E. et al. Global burden of influenza-associated lower respiratory tract infections and hospitalizations among adults: A systematic review and meta-analysis. PLoS Med 18, e1003550 (2021).
Chow, E. J. et al. Acute cardiovascular events associated with influenza in hospitalized adults: A cross-sectional study. Ann Intern Med 173, 605–613 (2020).
Muñoz-Quiles, C. et al. Risk of cardiovascular events after influenza: A population-based self-controlled case series study, Spain, 2011-2018. J Infect Dis 230, e722–e731 (2024).
Persson, E., Ferlie, E. & Baeza, J. ‘This is just a little flu’: analysing medical populist discourses on the Covid-19 pandemic in Brazil. Public Policy and Administration 40, 94–125 (2025).
Bajema, K. L. et al. Severity and long-term mortality of COVID-19, influenza, and respiratory syncytial virus. JAMA Intern Med 185, 324–334 (2025).
Yom-Tov, E., Shembekar, J., Barclay, S. & Muennig, P. Erratum: Author Correction: The effectiveness of public health advertisements to promote health: a randomized-controlled trial on 794,000 participants. NPJ Digit Med 1, 38 (2018).
News Staff. Is social media changing the doctor-patient relationship? American Academy of Family Physicians https://www.aafp.org/news/practice-professional-issues/20180608commsurveys.html (2018). [accessed 13 November 2025].
Liu J. “Xin meiti hua shengcun” de xuqiu [The need for “new-media-based living”]. Renmin Ribao Haiwai Ban. 2019;09:(in Chinese).
Song, S. et al. Short-Video apps as a health information source for chronic obstructive pulmonary disease: information quality assessment of TikTok videos. J Med Internet Res 23, e28318 (2021).
QuestMobile Research Institute. (2025, November 05). QuestMobile 2025 China mobile internet autumn report: 1.269 billion users with an average of 178.2 h per month, three driving forces boost user engagement. QuestMobile Research Institute. Retrieved November 10, 2025, from https://www.questmobile.com.cn/research/report/1985984354598359042.
Bilibili Inc. Announces Third Quarter 2025 Financial Results. https://ir.bilibili.com/media/krmdk0ls/bilibili-inc-announces-third-quarter-2025-financial-results.pdf; 2025 [accessed 13 November 2025].
Xiao, L. et al. Public’s preferences for health science popularization short videos in China: a discrete choice experiment. Front Public Health 11, 1160629. https://doi.org/10.3389/fpubh.2023.1160629 (2023).
de Vere Hunt, I. & Linos, E. Social media for public health: Framework for social media-based public health campaigns. J Med Internet Res 24, e42179 (2022).
Suarez-Lledo, V. & Alvarez-Galvez, J. Prevalence of health misinformation on social media: Systematic review. J Med Internet Res 23, e17187 (2021). Published 2021 Jan 20.
Loeb, S., Langford, A. T., Bragg, M. A., Sherman, R. & Chan, J. M. Cancer misinformation on social media. CA Cancer J Clin 74, 453–464 (2024).
Smith, R., Chen, K., Winner, D., Friedhoff, S. & Wardle, C. A systematic review Of COVID-19 misinformation interventions: lessons learned. Health Aff (Millwood) 42, 1738–1746 (2023).
Rocha, Y. M. et al. The impact of fake news on social media and its influence on health during the COVID-19 pandemic: A systematic review. J Public Health (Berl.) 31, 1007–1016 (2023).
Pereira Neto, A. Strategies to Fight Online Health Misinformation. Health Information Quality Assurance, Digital Literacy and Fact-Checking. Springer, 2025.
Qi, S., Chen, Q., Liang, J., Yang, X. & Li, X. Evaluating the quality and reliability of cervical cancer related videos on YouTube, Bilibili, and Tiktok: a cross-sectional analysis. BMC Public Health 25, 3682 (2025).
Liu, Z. et al. YouTube/ Bilibili/ TikTok videos as sources of medical information on laryngeal carcinoma: cross-sectional content analysis study. BMC Public Health 24, 1594 (2024).
Zhang, Y. et al. Quality assessment of spinal cord injury-related health information on short-form video platforms: Cross-sectional content analysis of TikTok, Kwai, and BiliBili. Digit Health 11, 20552076251374226 (2025).
Morahan-Martin, J. M. How internet users find, evaluate, and use online health information: a cross-cultural review. Cyberpsychol Behav 7, 497–510 (2004).
Silberg, W. M., Lundberg, G. D. & Musacchio, R. A. Assessing, controlling, and assuring the quality of medical information on the Internet: Caveant lector et viewor–Let the reader and viewer beware. J Amer Med Assoc 277, 1244–1245 (1997).
Bernard, A. et al. A systematic review of patient inflammatory bowel disease information resources on the World Wide Web. Am J Gastroenterol 102, 2070–2077 (2007).
Charnock, D., Shepperd, S., Needham, G. & Gann, R. DISCERN: an instrument for judging the quality of written consumer health information on treatment choices. J Epidemiol Community Health 53, 105–111 (1999).
Shoemaker, S. J., Wolf, M. S. & Brach, C. Development of the Patient Education Materials Assessment Tool (PEMAT): a new measure of understandability and actionability for print and audiovisual patient information. Patient Educ Couns 96, 395–403 (2014).
Koo, T. K. & Li, M. Y. A guideline of selecting and reporting intraclass correlation coefficients for reliability research. J Chiropr Med 15, 155–163 (2016).
Liang, M. & Ye, L. Algorithmic pedagogy: How Douyin constructs algorithmic imaginaries for content creators. Platforms & Society 2, 29768624251365615 (2025).
Kington, R. S. et al. Identifying credible sources of health information in social media: Principles and attributes. NAM Perspect 2021, 10.31478/202107a (2021).
Centers for Disease Control and Prevention (CDC). Cold versus flu. from https://www.cdc.gov/flu/about/coldflu.html; 2024 [accessed 13 November 2025].
Han, Y., Guo, J., Li, X. & Zhong, Z. Differences in clinical characteristics between coronavirus disease 2019 (COVID-19) and influenza: a systematic review and meta-analysis. NPJ Prim Care Respir Med 35, 8 (2025).
Piroth, L. et al. Comparison of the characteristics, morbidity, and mortality of COVID-19 and seasonal influenza: a nationwide, population-based retrospective cohort study. Lancet Respir Med 9, 251–259 (2021).
Uyeki, T. M. et al. Clinical practice guidelines by the infectious diseases Society of America: 2018 update on diagnosis, treatment, chemoprophylaxis, and institutional outbreak management of seasonal influenzaa. Clin Infect Dis 68, 895–902 (2019).
Campbell, A. P. et al. Influenza antiviral treatment and length of stay. Pediatrics 148, e2021050417 (2021).
World Health Organization. WHO multi‑country survey reveals widespread public misunderstanding about antibiotic resistance. https://www.who.int/news/item/16-11-2015-who-multi-country-survey-reveals-widespread-public-misunderstanding-about-antibiotic-resistance; 2015 [accessed 13 November 2025].
World Health Organization. Recommended composition of influenza virus vaccines for use in the 2023‑2024 Northern Hemisphere influenza season. https://www.who.int/publications/m/item/recommended-composition-of-influenza-virus-vaccines-for-use-in-the-2023-2024-northern-hemisphere-influenza-season; 2023 [accessed 13 November 2025].
Zhao, H. T. et al. Investigation on influenza vaccination policy and vaccination situation during the influenza seasons of 2020‑2021 and 2021‑2022 in China. Zhonghua Yu Fang Yi Xue Za Zhi 56, 1560–1564 (2022).
de Fougerolles, T. R. et al. Public health and economic benefits of seasonal influenza vaccination in risk groups in France, Italy, Spain and the UK: state of play and perspectives. BMC Public Health 24, 1222 (2024).
Schmid, P. et al. Barriers of influenza vaccination intention and behavior - a systematic review of influenza vaccine hesitancy, 2005 - 2016. PLoS One 12, e0170550 (2017).
Thompson, M. G. et al. Influenza vaccine effectiveness in preventing influenza-associated intensive care admissions and attenuating severe disease among adults in New Zealand 2012-2015. Vaccine 36, 5916–5925 (2018).
Ferdinands, J. M. et al. Does influenza vaccination attenuate the severity of breakthrough infections? A narrative review and recommendations for further research. Vaccine 39, 3678–3695 (2021).
Grohskopf, L. A. et al. Prevention and control of seasonal influenza with vaccines: Recommendations of the advisory committee on immunization practices - United States, 2025-26 influenza season. MMWR Morb Mortal Wkly Rep 74, 500–507 (2025).
Acknowledgements
This study was supported by the Zhejiang Provincial Clinical Research Center for Major Traditional Chinese Medicine Epidemic Diseases.
Author information
Authors and Affiliations
Department of Pulmonary and Critical Care Medicine, The Third Affiliated Hospital of Zhejiang Chinese Medical University, 219 Moganshan Road, Xihu District, Hangzhou, Zhejiang, 310005, China
Xinlei Zhang, Chunyuan Zhou, Yuyang Hu & Junchao Yang
Authors
- Xinlei Zhang
- Chunyuan Zhou
- Yuyang Hu
- Junchao Yang
Contributions
XL.Z. conceived and designed this study, collected all videos, performed data analysis, and drafted the manuscript. XL.Z. and CY.Z. collected videos and author characteristic information. XL.Z., CY.Z., and YY.H. reviewed, categorized, and rated the videos. JC.Y. reviewed and revised the manuscript. All authors participated in drafting and revising the manuscript and approved the final version for submission.
Corresponding author
Correspondence to Junchao Yang.
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Zhang, X., Zhou, C., Hu, Y. et al. Evaluating influenza-related short videos on Chinese platforms: information quality, user engagement, and public health implications. npj Prim. Care Respir. Med. 36, 69 (2026). https://doi.org/10.1038/s41533-026-00505-6
Received: 18 December 2025
Accepted: 14 March 2026
Published: 22 October 2026
Version of record: 22 October 2026
DOI: https://doi.org/10.1038/s41533-026-00505-6
来源:Nature 健康护理主题 · nature.com