Abstract
Background
Esophageal cancer typically lacks specific early symptoms, leading to late-stage diagnosis and poor prognosis, with an overall low 5-year survival rate. However, early detection and timely intervention can significantly improve the 5-year survival rate, underscoring the importance of prevention, screening, and early intervention. Short video platforms are increasingly utilized for health communication, offering opportunities to disseminate medical knowledge. However, the quality and reliability of health-related content, particularly for diseases like esophageal cancer, remain under explored.
Objective
This study aimed to systematically evaluate the quality, reliability, completeness, and engagement of esophageal cancer-related videos on three popular short video platform: Bilibili, TikTok, and Kwai, to identify platform-specific strengths and limitations in disseminating health information.
Methods
A total of 311 esophageal cancer-related videos were analyzed. Video assessment was assessed using 4 standardized scoring framework including General Quality Scores (GQS) for general quality, the DISCERN for reliability, Completeness Score (CS) for comprehensive information, Engagement Score (ES) for understandability and entertainment value of the video. Video features (source, category, content), user’s behavior (likes, shares, comments) were also collected. Cross-platform comparisons were conducted to identify disparities in content quality and user interaction.
Results
This study analyzed 311 esophageal cancer-related videos on Bilibili, TikTok, and Kwai. Video quality varied significantly across platforms, with Bilibili showing the highest DISCERN (5.46), GQS (2.97) and CS (3.64), while TikTok videos achieved the highest ES (2.88) and engagement metrics (e.g., likes and collections, p < 0.001). Kwai videos had the lowest scores across all measures. Content focused primarily on “symptoms” and “treatment,” with Bilibili offering more comprehensive coverage. Correlation analysis revealed a positive association between video quality and engagement on Bilibili but a negative association on TikTok (e.g., GQS and likes, r=-0.251, P = 0.009).
Conclusions
The quality of esophageal cancer-related videos across Bilibili, TikTok, and Kwai is suboptimal, with notable quality disparities among the platforms. Users on platforms other than Bilibili show limited ability to identify or prefer higher-quality content. This study underscores the potential of short-video platforms for esophageal cancer public health education, but highlights the need for improvements in content quality, ethical standards, and platform governance to address health equity concerns.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12889-025-23475-9.
Keywords: Esophageal cancer, Short video, Information quality, Social media, TikTok, Bilibili, GQS, DISCERN
Introduction
Esophageal cancer is a prevalent malignant tumor of the digestive tract, posing significant health risks and contributing substantially to the global disease burden [1]. China remains one of the highest-risk regions, with 224,000 new cases and 187,500 deaths reported in 2022, accounting for 43.83% of global incidence and 42.09% of global mortality [2]. Esophageal cancer typically lacks specific early symptoms, leading to late-stage diagnosis and poor prognosis, with an overall low 5-year survival rate [3]. However, early detection and timely intervention can significantly improve the 5-year survival rate, underscoring the importance of prevention, screening, and early intervention [4]. The two most common histological subtypes of esophageal cancer, squamous cell carcinoma (SCC) and adenocarcinoma (AC), have distinct etiologies, contributing to significant geographic variation in their incidence rates [2]. In China, SCC accounts for over 95% of esophageal cancer cases and is associated with risk factors such as tobacco smoking, heavy alcohol consumption, poor dietary habits and, to a lesser extent, obesity [5–7]. Improved control of these risk factors, advancements in early screening technologies like endoscopy, and supportive national public health policies have led to declines in esophageal cancer incidence and mortality rates, particularly for SCC in high-risk regions [8–10]. However, these rates remain relatively high, underscoring the ongoing public health challenge posed by this disease [11].
In the era of rapid information growth, driven by increased access to mobile internet, social media and online video platforms have become powerful tools for disseminating medical knowledge [12].Their vast user base, rapid content dissemination, and interactive formats enhance their potential to educate and engage the public on health-related topics [13, 14].Platforms like Douyin (internationally known as TikTok), Bilibili, and Kuaishou (internationally known as Kwai), which have 780 million, 338 million, and 691.8 million monthly active users respectively, serve as major social media communication channels in China [15–17]. These videos are user-generated and uploaded, allowing users not only to watch but also to interact through likes, comments, and real-time “bullet comments”. As health literacy improves, more individuals are shifting from passively receiving health information to actively seeking it. For those lacking a medical background, accessible and engaging social media platforms that offer medically informational videos may be more appealing. For esophageal cancer, individuals with recognized risk factors may be more inclined to seek online information about prevention and early warning signs. Studies have shown that personal or family cancer history is a significant predictor of cancer-related online health information seeking [18]. While some studies have evaluated the quality of videos on other diseases across two of these platforms [13, 14, 19], few have focused specifically on esophageal cancer, and even fewer have conducted a direct comparison across TikTok, Kwai, and Bilibili. However, given the distinct user demographics and content ecosystems of each platform [20], quality differences may lead to systematic disparities in the dissemination and impact of health information. To fill this gap, we posed the following research question: What are the quality, reliability, completeness, and user engagement characteristics of esophageal cancer–related videos on these three platforms? We then conducted a cross-sectional analysis to systematically assess these dimensions, with the goal of providing actionable insights for platforms, content creators, and public health practitioners to improve esophageal cancer education via social media.
Materials and methods
Search strategy and data processing
This cross-sectional study examined three major Chinese short-video platforms—Bilibili, Kwai, and TikTok—for content related to esophageal cancer. On June 4, 2024, we conducted searches using two Chinese keywords: “#食管癌” and “#食道癌” (both meaning “esophageal cancer”). Although the two terms are synonymous in medical context, their usage varies across regions and demographics in China due to differences in local linguistic preferences.
Search results were sorted according to each platform’s default ranking algorithm. Although we observed that the majority of videos retrieved using the keywords “#食管癌” and “#食道癌” were overlapping, we adopted a dual-keyword strategy to minimize potential sampling bias caused by relying on a single term. Specifically, from each platform, we first selected the top 50 videos using the keyword “#食管癌”, and then used the keyword “#食道癌” to supplement the sample, ensuring that the top 50 results for each term were included. This approach yielded a total of 110 videos per platform, resulting in 330 videos across the three platforms.This selection was based on prior research recommendations and team discussions, considering that users typically do not view more than 100 videos related to the same keyword during their daily usage [21, 22].
All included videos were in Chinese or in English with accurate Chinese subtitles. We applied specific exclusion criteria: duplicate videos, videos exclusively in English without Chinese subtitles, videos without audio, and those not directly related to esophageal cancer were omitted (see Fig. 1). To minimize potential biases in search results and quality assessments, we utilized newly created accounts on each platform, ensuring no prior search history influenced the outcomes. This methodology aligns with standard practices for data collection and analysis in cross-sectional studies, ensuring the reliability and validity of our findings.
Fig. 1.
Video searching procedure
Prior to the formal evaluation, the research team convened to establish detailed coding criteria to ensure consistency throughout the process. To enhance inter-rater reliability, three rounds of pre-coding were conducted, during which discrepancies were thoroughly discussed and the coding standards for various situations were clarified. Subsequently, three investigators (BH, YD, XW) independently coded the videos. Finally, two senior experts (WZ, SJ) audited a random 10% sample of the coded videos to further assess the accuracy and consistency of the coding process.
Collection of video features
A total of 311 videos across the three platforms were included in the final analysis after excluding invalid videos. The following characteristics were recorded for each video as of June 4, 2024: video duration, days since upload, number of likes, number of comments, and number of shares. These characteristics served as the basis for further analysis and evaluation using the Global Quality Scale (GQS) and DISCERN.
Based on previous research, we initially categorized video uploaders into seven groups: certified medical professionals, patients or family members, general users, science communicators, news organizations, nonprofit organizations, and for-profit organizations. However, due to the limited number of videos (< 5) in certain categories, we merged these groups to ensure sufficient sample sizes and enhance the stability of statistical analysis. Specifically, patients or family members and general users were combined into a single “general users” category, while news organizations, nonprofit organizations, and for-profit organizations were merged into a broader “organizations” category, following conceptual similarities and prior classification strategies. As a result, the final comparison of video quality scores (DISCERN, GQS, CS, ES) was conducted across four uploader types: certified medical professionals, general users, science communicators, and organizations. Specific classification details are provided in the Supplementary Information (see Additional File 1, Table S1).
We adapted a comprehensive content-completeness checklist—originally developed for colorectal cancer videos by Zhang et al. [23] and then applied to gastric cancer videos by Hu et al. [24]—to the esophageal cancer context. Given no standardized tool exists for esophageal cancer video assessment, we retained the core domains (etiology, symptoms, prevention, treatments, prognosis) and merged “anatomy” into “etiology & risk factors” to form five mutually exclusive categories that ensure both conceptual clarity and comparability with other cancer video analyses.
Video evaluation
We assessed the overall quality of video content using the General Quality Scores (GQS) and the reliability of the videos using the DISCERN tool, both of which are proven to be reliable and valid measures [25–27]. The GQS consists of five criteria rated from 1 to 5, where higher scores indicate better quality. The first section of the DISCERN questionnaire evaluates the reliability and includes eight items, which are described in detail in the Supplementary Information (see Additional File 1, Table S2). Among these, the two items assessing bias and uncertainty were reverse-coded, such that higher scores indicate greater overall reliability [28]. The Global Quality Score (GQS) was categorized into five levels, as detailed in the Supplementary Information (see Additional File 1, Table S3).
We also measured Completeness Score (CS) and Engagement Score (ES). The CS assessed whether videos addressed five key aspects: etiology, symptoms, prevention, treatment, and prognosis. Each aspect was rated as follows: 0 points for not mentioned, 1 point for brief mention, and 2 points for detailed explanation. The total completeness score was interpreted as follows: 0–1 points indicated almost no valuable information, 2–3 points indicated most information was missing, 4–5 points indicated some information was missing, 6–7 points indicated little information was missing, 8–9 points indicated most key information was discussed, and 10 points indicated complete coverage with detailed explanations.
We evaluated “understandability of scientific content” and “entertainment value” of the video using Engagement Score, derived from the Evaluation Index System of Science Popularization Video [29]. “Understandability” was rated as 2 points for content provides clear, well-explained scientific information that is easy to understand, 1 point for excessive technical terms making it less comprehensible, and 0 points for content directly replicates academic articles or patents, making it difficult to comprehend. “Entertainment Value” was rated as 2 points for educational and entertaining content that is engaging for the audience, 1 point for limited entertainment value, making it less engaging, and 0 points for being dull and lacks entertainment value. The sum of these two ratings constitutes the total Engagement Score for each video.
Ethical considerations
This study exclusively used publicly available video content from Chinese platforms of TikTok, Kwai, and Bilibili. No clinical data, human specimens, animal experiments, or personal privacy information was involved in this study. Consequently, ethical review was not required for this research.
Statistical analysis
All statistical analyses were conducted using SPSS version 26.0 ( IBM Corp., Armonk, NY, USA). The Shapiro-Wilk Test was applied to assess the normality of the data. Continuous variables are presented as median (range) or mean (SD), while categorical variables are reported as frequencies and percentages. The Kruskal-Wallis Test was used to determine significant differences between groups of multiple video uploaders. Descriptive statistics were used to summarize content coverage by platform and content type. Chi-square tests were conducted to assess differences in categorical distributions across platforms. Where applicable, post hoc comparisons of column proportions were performed, with Bonferroni adjustment applied to control for multiple testing. Spearman’s rank correlation test was applied to evaluate the associations between video variables and video scores.
Results
Video characteristics
On the Kwai and TikTok platforms, the number of videos increased progressively as the search period approached the time of the study (June 2024), with a noticeable peak observed in the second quarter of 2024. However, Bilibili displayed a different search pattern, with most videos collected dating back to the second quarters of 2022 and 2023 (see Fig. 2). The median counts for “likes”, “comments”, and “collections” on TikTok videos were significantly higher than those on Bilibili and Kwai (p < 0.001). Conversely, videos on Bilibili had a higher median total duration than those on Kwai and TikTok (p < 0.001). Among all videos included, the shortest video lasted 30 s, while the longest video spanned 152 min. The oldest video had been uploaded to Bilibili over 7 years prior to data collection, while the newest video was uploaded to Kwai during the data collection period (see Table 1).
Fig. 2.
Quarterly Distribution of Esophageal Cancer-Related Videos on Kwai, TikTok, and Bilibili
Table 1.
Detailed characteristics of esophagus cancer videos
| Parameters | Bilibili (N = 106) | Kwai (N = 98) | TikTok (N = 107) | K-W H/χ² |
|---|---|---|---|---|
| Likes, median (range) | 16.50(0-53000) | 357.50(3-25700) | 419.00(15-13800) | 122.226*** |
| Comments, median (range) | 1.00(0-2588) | 48.00(0-16000) | 48.00(0-6184) | 84.939*** |
| Collections, median (range) | 18.50(0-6400) | 83.00(1-39000) | 117.00(1-37000) | 55.271*** |
| Duration (minutes), median (range) | 4.25(0.5–152) | 1.00(0.5-5.0) | 1.00(0.5–3.5) | 116.290*** |
| Time since video upload (days), median (range) | 15.00(1–85) | 10.00(0–63) | 6.00 (1–54) | 29.234*** |
| Video source | ||||
| Certified medical professionals (n) (%) | 32,(30.2) | 27,(27.6) | 91,(85.0) | 50.680*** |
|
-Gastroenterology/Thoracic Surgery/ Oncology |
27,(25.5) | 20,(20.4) | 77,(72.0) | 46.758*** |
| -Other Specialties | 4,(3.8) | 6,(6.1) | 11,(10.3) | 3.714 |
| - Traditional Chinese Medicine (TCM) | 1,(0.9) | 1,(1.0) | 3,(2.8) | 1.600 |
| Patients or family members (n) (%) | 2,(1.9) | 8,(8.2) | 3,(2.8) | 4.769 |
| General users (n) (%) | 29,(27.4) | 48,(49.0) | 1,(0.9) | 43.000*** |
| Science communicators (n) (%) | 29,(27.4) | 12,(12.2) | 2,(1.9) | 26.000*** |
| News organizations (n) (%) | 2,(1.9) | 3,(3.1) | 5,(4.7) | 1.400 |
| Nonprofit organizations (n) (%) | 3,(2.8) | 0,(0) | 4,(3.7) | 0.143 |
| For-profit organizations (n) (%) | 9,(8.5) | 0,(0) | 1,(0.9) | 6.400* |
| Video category | ||||
| Scientific explanations(n) (%) | 57,(53.8) | 62,(63.3) | 93,(86.9) | 10.764** |
| Professional courses or lectures (n) (%) | 31,(29.2) | 0,(0) | 0,(0) | - |
| Personal or shared experiences (n) (%) | 18,(17.0) | 35,(35.7) | 14,(13.1) | 11.134** |
| Other content (n) (%) | 0,(0) | 1,(1.0) | 0,(0) | - |
| Video content | ||||
| Etiology (n) (%) | 43,(40.6) | 18,(18.4) | 37,(34.6) | 12.339** |
| Symptoms (n) (%) | 57,(53.8) | 56,(57.1) | 53,(49.5) | 1.201 |
| Preventions (n) (%) | 37,(34.9) | 21,(21.4) | 27,(25.2) | 5.018 |
| Treatments (n) (%) | 71,(67.0) | 48,(49.0) | 47,(43.9) | 12.486** |
| Prognosis (n) (%) | 45,(42.5) | 12,(12.2) | 16,(15.0) | 32.459*** |
Footnote:
- KW H: Kruskal-Wallis H statistic, a non-parametric test used to determine if there are statistically significant differences between two or more groups of an independent variable on a continuous or ordinal dependent variable
- χ²: Chi-square statistic, used to assess the association between categorical variables or to test the goodness of fit between observed and expected frequencies
- * indicates p < 0.05, ** indicates p < 0.01,*** indicates p < 0.001
Video sources and types
Analysis of video sources indicated that certified medical professionals contributed the highest number of videos on TikTok (n = 91, 85%) and Bilibili (n = 31, 30%). On Kwai, the most common “uploader type” was general users (n = 48, 49%). Notably, science communicators uploaded the highest number of videos on Bilibili (29 (Bilibili) vs. 12 (Kwai) vs. 2(Tiktok)). In the subcategory of professional individuals, the three platforms showed a similar distribution, with the highest proportions from gastroenterology/thoracic surgery/oncology, followed by other specialties, and then Traditional Chinese Medicine (see Table 1).
In terms of video types, scientific explanations were the most prevalent on each platform: Bilibili (53.8%), Kwai (63.3%), and TikTok (86.9%). Additionally, professional lectures or courses were exclusively present on Bilibili, with a total of 31 videos (29.2%) (see Table 1).
Video contents
As shown in Table 1, the top two most frequently covered content categories across all three platforms were “symptoms” and “treatment.” However, the third most common category varied by platform: “prognosis” ranked third on Bilibili (42.5%), “prevention” on Kwai (21.4%), and “etiology” on TikTok (34.6%). Further comparison revealed that Bilibili videos included significantly more content related to etiology, treatment, and prognosis than those on TikTok and Kwai. Figure 3 illustrates the depth and comprehensiveness of content coverage by platform and by content type. As shown, Bilibili videos tended to provide more detailed and structured information, particularly in the domains of Etiology (χ²=12.353, P = 0.015), Prevention (χ²=15.015, P = 0.005), treatment (χ²=23.344, P < 0.001) and prognosis (χ²=3.949, P < 0.001). In contrast, no significant difference was observed in the content of Symptoms between three platforms (χ²=32.964, P = 0.413). This suggests that videos on Bilibili are generally more informative across these key dimensions of esophageal cancer education.
Fig. 3.
Content Coverage Depth by Category on Bilibili, Kwai, and TikTok Platforms
Video quality and reliability
Table 2 presents a comparison of video quality scores (DISCERN, GQS, CS, and ES) across Bilibili, Kwai, and TikTok platforms. Bilibili videos had the highest DISCERN and GQS scores, with the average scores of 5.46 and 2.97, respectively. TikTok’s scores were in the middle range, and statistical differences were observed between the platforms for these two scores. For CS, the trend was similar across platforms, with Bilibili scoring the highest at 2.97, followed by TikTok and Kwai; however, the difference between TikTok and Kwai was not statistically significant. In terms of ES, TikTok scored the highest, with significant differences between TikTok and both Bilibili and Kwai in pairwise comparisons. Bilibili and Kwai had similar ES scores.
Table 2.
Comparison of video quality scores (DISCERN, GQS, CS, ES) among bilibili, Kwai and TikTok platforms, (mean (SD)
| Categories | DISCERN | GQS | CS | ES |
|---|---|---|---|---|
| Bilibili | 5.46(1.13) | 2.97(0.85) | 3.64(2.22) | 2.32(1.22) |
| Kwai | 4.77(0.74) | 2.18(0.83) | 2.14(1.28) | 2.32(0.86) |
| Tiktok | 5.05(0.44) | 2.59(0.58) | 2.41(1.17) | 2.88(0.54) |
| K-W: p | 0.000 | 0.000 | 0.000 | 0.000 |
| MWW: pB vs. K | 0.000 | 0.000 | 0.000 | 0.772 |
| B vs. T | 0.011 | 0.000 | 0.000 | 0.001 |
| K vs. T | 0.001 | 0.000 | 0.079 | 0.000 |
Table 3 Shows scores for DISCERN, GQS, CS, and ES based on video sources across the 3 platforms of the study. On bilibili, videos from general users had the highest DISCERN score (5.77), while organizations scored the highest in GQS (3.43), CS (4.21), and ES (2.57). On kwai, science communicators had the highest scores in all categories: DISCERN (4.92), GQS (2.58), CS (2.67), and ES (2.5). On tiktok, certified medical professionals had the highest DISCERN score (5.05), while organizations led in GQS (3.1) and CS (3.2), and science communicators had the highest ES score (3.5). Significant differences in GQS were found between video uploaders across all platforms (Bilibili: P = 0.017, kwai: P = 0.002, tiktok: P = 0.022). DISCERN scores and ES did not show significant differences between uploaders across platforms (Bilibili: P = 0.763, kwai: P = 0.459, tiktok: P = 0.264). For CS, significant differences were observed between uploader types on Kwai (P = 0.049) and TikTok (P = 0.031).
Table 3.
Comparison of Video Quality Scores (DISCERN, GQS, CS, ES) among Different Video Sources on Bilibili, Kwai and TikTok Platforms, (mean (SD)
| Bilibili | Kwai | Tiktok | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Video source | DISCERN | GQS | CS | ES | DISCERN | GQS | CS | ES | DISCERN | GQS | CS | ES |
| Certified medical professionals | 5.03(0.78) | 2.75(0.72) | 3.13(2.1) | 2.38(0.91) | 4.89(0.4) | 2.56(0.6) | 2.56(1.3) | 2.3(0.87) | 5.05(0.4) | 2.58(0.54) | 2.36(1.10) | 2.87(0.54) |
| General users | 5.77(1.28) | 2.77(0.92) | 3.61(2.67) | 2.16(1.16) | 4.70(0.89) | 1.95(0.84) | 1.86(1.12) | 2.32(0.83) | 5(1.41) | 1.5(0.58) | 1.25(0.5) | 2.5(0.58) |
| Science communicators | 5.6(1.2) | 3.21(0.77) | 3.97(1.84) | 2.31(1.58) | 4.92(0.29) | 2.58(0.67) | 2.67(1.6) | 2.5(1) | 5(0) | 2.5(0.71) | 3(1.4) | 3.5(0.71) |
| Organizations | 5.43(1.09) | 3.43(0.94) | 4.21(2.08) | 2.57(1.22) | 4.33(1.15) | 1.67(1.15) | 1.67(1.15) | 2(1) | 5(0) | 3.1(0.32) | 3.2(1.55) | 3(0.47) |
| p -Value | 0.060 | 0.017 | 0.127 | 0.664 | 0.361 | 0.002 | 0.049 | 0.684 | 0.36 | 0.001 | 0.031 | 0.176 |
Correlation analysis
Spearman’s correlation analysis revealed that on Bilibili, there was a positive correlation between DISCERN scores and the number of “collections” (r = 0.202, P = 0.037), GQS scores and the number of “collections” (r = 0.211, P = 0.03), “Engagement Score” and the number of “likes” (r = 0.239, P = 0.014), and “Engagement Score” and the number of “comments” (r = 0.297, P = 0.002). No positive correlations were observed among video variables on Kwai. On TikTok, negative correlations were observed between GQS scores and the number of “likes” (r=-0.251, P = 0.009), GQS scores and the number of “comments” (r=-0.279, P = 0.004), and GQS scores and the number of “collections” (r=-0.266, P = 0.006) (see Table 4).
Table 4.
Correlation analysis between video performance scores and video “likes”, “comments”, and “collections” across the three different platforms
| Bilibili | Kwai | Tiktok | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| DISCERN | GQS | CS | ES | DISCERN | GQS | CS | ES | DISCERN | GQS | CS | ES | |
| Likes | -0.015 | 0.035 | -0.045 | 0.239* | -0.009 | -0.055 | 0.094 | -0.02 | -0.089 | − 0.251** | -0.109 | 0.089 |
| 0.881 | 0.725 | 0.647 | 0.014 | 0.929 | 0.59 | 0.358 | 0.841 | 0.363 | 0.009 | 0.264 | 0.364 | |
| Comments | -0.076 | 0.026 | -0.008 | 0.297** | -0.139 | -0.075 | 0.02 | -0.056 | -0.146 | − 0.279** | -0.135 | -0.033 |
| 0.437 | 0.79 | 0.934 | 0.002 | 0.173 | 0.464 | 0.844 | 0.582 | 0.134 | 0.004 | 0.165 | 0.732 | |
| Collections | 0.202* | 0.211* | 0.115 | 0.069 | -0.078 | -0.117 | 0.037 | -0.034 | -0.13 | − 0.266** | -0.132 | 0.047 |
| 0.037 | 0.03 | 0.241 | 0.484 | 0.444 | 0.252 | 0.715 | 0.737 | 0.183 | 0.006 | 0.174 | 0.634 | |
* indicates p < 0.05, ** indicates p < 0.01,*** indicates p < 0.001
Discussion
This study provides a comparative analysis of the quality, reliability, completeness, and engagement of esophageal cancer-related videos across three platforms: Bilibili, TikTok, and Kwai. Our findings highlight significant differences in video characteristics and user behavior, underscoring the role of platform-specific features in shaping health communication outcomes.
Platform-specific differences
Overall, Bilibili exhibited the highest video quality, followed by TikTok and Kwai. Bilibili outperformed the other platforms in the General Quality Score (GQS), DISCERN, and completeness ratings but had lower Engagement Scores compared to Kwai and TikTok. This discrepancy may be attributed to Bilibili’s content type: one-third of its videos in our sample were classified as “professional lectures or courses,” a category absent on TikTok and Kwai. These professional videos, often containing complex terminology, were likely aimed at experts in the healthcare industry rather than the general audience. Prior research indicates that such content may be perceived as “dull” and less engaging by non-expert viewers, which could explain the lower Engagement Score on Bilibili [27].
In terms of “content creators”, notable differences emerged across the 3 platforms. On Bilibili, “professionals,” “science communicators,” and “general users” each accounted for roughly 30% of video creators. In contrast, Kwai’s primary creators were “general users” and “professionals,” while TikTok had a predominantly professional creator base (85%). This aligns with TikTok’s stricter content regulations, which allow only certified institutions and doctors to upload medical content [30]. Consequently, TikTok achieved balanced scores for both professionalism and popularity.
Influence of content creators
Previous studies have highlighted the impact of “creator identity” on video quality [31]. Our results confirm significant differences in GQS across creators but reveal platform-specific variations. For instance, on Bilibili, videos uploaded by organizations scored highest in GQS, followed by science communicators and general users, with professionals ranking lowest. TikTok showed a similar trend, with professionals achieving only mid-level GQS rankings. Conversely, on Kwai, professionals’ videos ranked highest in GQS.
These findings diverge from previous research [21]. A potential explanation is that while professional video quality remained consistent across platforms, their relative rankings were influenced by the presence of other types of content creators. For instance, organizations on Bilibili and TikTok received significantly higher GQS scores compared to Kwai, which may have contributed to the observed variations. Furthermore, Bilibili’s dedicated “Knowledge Zone” [32] and TikTok’s strict certification process [30] emphasize professional expertise, leading professionals to focus on delivering in-depth medical knowledge rather than creating more generalized and accessible content. This likely explains why professionals’ videos scored lower in overall quality compared to news agencies and science communicators, who are often more skilled at engaging audiences through innovative communication strategies [33]. However, on Kwai, the lack of stringent certification processes and the generally lower quality of videos allowed professionals to achieve the highest GQS rankings.
Correlation between video quality and user engagement
On Bilibili, the “Engagement Score” was positively correlated with “likes” and “comments”, while GQS and DISCERN were both positively associated with the number of “collections”. These findings suggest that Bilibili users value reliable and high-quality videos and are more inclined to engage with content that is both informative and entertaining. In contrast, no significant correlations were observed on Kwai. On TikTok, GQS was negatively associated with “likes”, “comments”, and “collections”. This suggests that to some extend TikTok viewers cannot distinguish between high-quality and low-quality videos, which consistent with prior researches [19, 27] What’s more, these differences may also reflect platform-specific algorithmic mechanisms. Bilibili prioritizes high-quality content, while TikTok’s algorithm appears to favor highly engaging but lower-quality videos, as previously reported in similar studies [34]. Although some studies reported that they observed people’s improvement of the ability of recognizing high-quality videos [27], our research found that user behavior varied significantly across platforms.
Content focus
The distribution of content topics was relatively consistent across platforms, with “symptoms” and “treatment” being the most frequently addressed themes. In contrast, “etiology,” “prevention,” and “prognosis” were less commonly covered. Notably, Bilibili had a significantly higher proportion of videos focusing on “treatment,” reflecting its emphasis on professional educational content. Specifically, 83% of scientific explanations and professional lectures on Bilibili provided detailed explanations of treatment strategies, highlighting the platform’s orientation toward delivering expert-driven insights [32]. In addition, Bilibili was originally developed as a long-form video platform. Although it has since incorporated short-form video content, it continues to be more supportive of longer videos compared to TikTok and Kwai, which enforce stricter time limits. This structural feature enables Bilibili to accommodate more comprehensive and in-depth health information, which may partially explain the greater depth observed in its esophageal cancer–related videos.
This concentration on symptoms and treatments likely arises from their ability to evoke strong public interest and fear regarding esophageal cancer, which can increase viewer engagement. Indeed, fear appeal has been recognized as an effective strategy for disseminating health information [35, 36]. However, a troubling observation emerged from the analysis: several videos that emphasized the severe consequences of esophageal cancer appeared to prioritize generating fear not for educational purposes but as a means of promoting specific products (e.g., herbal medicines) or services (e.g., directing viewers to particular hospitals). Such practices raise ethical concerns and highlight the potential misuse of health-related content for commercial gain, rather than as a tool to inform and empower individuals to prevent disease and adopt healthier behaviors.
In addition, the scarcity of sample videos’ content covering recent advances in the diagnosis and treatment of esophageal cancer is notable, particularly concerning innovative non-invasive screening methods, endoscopic therapeutic techniques, and updates on treatment regimens. Recent advancements, such as AI-integrated Cytosponge diagnostics [37] and the identification of novel serum-based tumor biomarkers [38], have demonstrated promising efficacy in early detection. Furthermore, advancements in endoscopic technologies now allow for enhanced early diagnostic capabilities and potentially curative interventions. The integration of AI-assisted endoscopic assessments has significantly improved lesion detection accuracy, offering transformative potential for early screening practices [39]. Incorporating such innovations into educational video content could not only enrich the scientific value of these platforms but also provide viewers with valuable and actionable clinical insights.
Health inequity concern
Short video platforms offer significant potential to enhance accessibility to high-quality health information. However, they simultaneously demand greater cognitive and evaluative capabilities from users, which may inadvertently exacerbate health inequities among underserved populations [40]. Resent report identifies notable disparities in platform usage across different socioeconomic strata. For instance, while platforms such as TikTok and Bilibili are popular in urban and more developed regions, Kwai has a higher user base in less developed areas, with lower representation in first-tier cities and greater penetration in third-tier cities and rural regions [20].
Our findings indicate that videos on Kwai, on average, exhibit the lowest informational quality among the platforms analyzed. This observation is consistent with socio-economic theory, which posits that individuals with higher socioeconomic status typically possess better health literacy and information-processing abilities due to higher educational attainment, greater cognitive resources, and more supportive social networks [41]. The interaction between user demographics and platform content quality appears to be a driving factor behind this outcome, creating a feedback loop wherein individuals from underserved areas gravitate toward platforms offering lower-quality content. These platforms, in turn, generate significant traffic and economic revenue from such users, reducing their incentive to implement robust content moderation or improve informational quality. This vicious cycle perpetuates existing health disparities, presenting a formidable barrier to achieving equitable health outcomes and emphasizing the urgent need for targeted interventions to disrupt this cycle.
Limitations
First, a major limitation of this study lies in the subjectivity of judgment. Currently, there is no universally accepted automated tool for objectively evaluating the quality of health-related videos. Therefore, our classification and scoring methods involved subjective interpretation. However, the inter-rater reliability was high, as evidenced by three rounds of pre-coding and a 10% sample audit. Second, although TikTok, Kwai, and Bilibili are among the most popular and widely used short-video platforms in China, newer platforms such as Red Note (小红书) were not included. This exclusion may limit the generalizability of our findings and introduce potential sampling bias. Third, we acknowledge that the DISCERN instrument was originally developed to assess written consumer health materials, and its application to video content has been a subject of methodological concern. However, in the absence of a validated, video-specific alternative—and given the tool’s widespread adaptation in health communication research—we applied Sect. 1 of the DISCERN tool, which focuses on reliability, to our video sample. To minimize the potential bias associated with relying on a single instrument, we also employed additional validated tools, including the Global Quality Score (GQS), Completeness Score (CS), and Engagement Score (ES), each of which evaluates distinct dimensions of video quality.
Conclusion
This study systematically evaluated the quality, reliability, completeness, and engagement of 311 esophageal cancer–related videos across three popular short-video platforms—Bilibili, TikTok, and Kwai. The overall quality of these videos was found to be suboptimal, highlighting substantial room for improvement. Significant disparities in video quality were observed across platforms, with users on platforms other than Bilibili showing limited ability to recognize or favor higher-quality content.
These findings underscore that, while short-video platforms hold considerable potential as tools for esophageal cancer–specific public health education, addressing health equity concerns will require meaningful improvements in content quality, ethical standards, and platform governance.
Electronic supplementary material
Below is the link to the electronic supplementary material.
Acknowledgements
The authors would like to express their gratitude to Yannan Jiang from the School of Population Health, University of Auckland, for her valuable statistical advice, as well as to all the participants who contributed to this study.
Author contributions
WZ and SJ conceptualized and designed the study. Data collection was performed by WZ, BH, YD, XW, and SJ. BH and SJ conducted the data analysis. Manuscript drafting was carried out by WZ, BH, YD, XW, and SJ, with critical revisions by SJ and KY. All authors contributed to editing and approved the final manuscript for submission.
Funding
This study was partly funded by Humanities and Social Sciences Youth Foundation, Ministry of Education, China (23YJCZH090).
Data availability
All data that support the findings of this study is provided within supplementary information files.
Declarations
Ethics approval and consent to participate
Not applicable.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Arnold M, Abnet CC, Neale RE, et al. Global burden of 5 major types of Gastrointestinal cancer. Gastroenterology. 2020;159(1):335–. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Bray F, Laversanne M, Sung HYA. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. Ca-a Cancer J Clin. 2024;74(3):229–63. [DOI] [PubMed] [Google Scholar]
- 3.Zeng HM, Chen WQ, Zheng RS, et al. Changing cancer survival in China during 2003-15: a pooled analysis of 17 population-based cancer registries. Lancet Global Health. 2018;6(5):e555–67. [DOI] [PubMed] [Google Scholar]
- 4.Duggan MA, Anderson WF, Altekruse S, et al. The surveillance, epidemiology, and end results (SEER) program and pathology: toward strengthening the critical relationship. Am J Surg Pathol. 2016;40(12):e94–102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Enzinger PC, Mayer RJ. Esophageal cancer. N Engl J Med. 2003;349(23):2241–52. [DOI] [PubMed] [Google Scholar]
- 6.Mao YS, Gao SG, Wang Q, et al. Epidemiological characteristic and current status of surgical treatment for esophageal cancer by analysis of National registry database. Chin J Oncol. 2020;42(3):228–33. [DOI] [PubMed] [Google Scholar]
- 7.Lai SP, Su HM, Liu YW, et al. Spatial and Temporal distribution characteristics research of esophageal cancer in China. Chin J Oncol. 2024;46(7):657–62. [DOI] [PubMed] [Google Scholar]
- 8.Zhang N, Li YY, Chang X, et al. Long-term effectiveness of one-time endoscopic screening for esophageal cancer: A community-based study in rural China. Cancer. 2020;126(20):4511–20. [DOI] [PubMed] [Google Scholar]
- 9.Chen R, Liu Y, Song GH, et al. Effectiveness of one-time endoscopic screening programme in prevention of upper Gastrointestinal cancer in china: a multicentre population-based cohort study. Gut. 2021;70(2):251–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Zheng RS, Zhang SW, Zeng HM, et al. Cancer incidence and mortality in china, 2016. J Natl Cancer Cent. 2022;2(1):1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Han BF, Zheng RS, Zeng HM, et al. Cancer incidence and mortality in china, 2022. J Natl Cancer Cent. 2024;4(1):47–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Chen JH, Wang Y. Social media use for health purposes: systematic review. J Med Internet Res. 2021;23(5):e17917. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Tan W, Liu YL, Shi ZW, et al. Information quality of videos related to Helicobacter pylori infection on tiktok: Cross-sectional study. Helicobacter. 2024;29(1):e13029. [DOI] [PubMed] [Google Scholar]
- 14.Wang MH, Yao N, Wang JM, et al. Bilibili, tiktok, and YouTube as sources of information on gastric cancer: assessment and analysis of the content and quality. BMC Public Health. 2024;24(1):57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Announces Second Quarter 2024 Financial Results. Bilibili Inc. https://ir.bilibili.com/media/qcemvd2g/bilibili-inc-announces-second-quarter-2024-financial-results.pdf. Accessed 21 September 2024.
- 16.Technology Announces Second Quarter and Interim. 2024 Unaudited Financial Results. Kuaishou. https://ir.kuaishou.com/news-releases/news-release-details/kuaishou-technology-announces-second-quarter-and-interim-2024. Accessed 21 September 2024.
- 17.Semi-annual report of CM Net in. 2024. Questmobile. https://www.questmobile.cn/research/report/1818126420037177346. Accessed 21 September 2024.
- 18.Rutten LJ, Squiers L, Hesse B. Cancer-related information seeking: hints from the 2003 health information National trends survey (HINTS). J Health Communication. 2006;11(Suppl 1):147–56. [DOI] [PubMed] [Google Scholar]
- 19.Sun F, Zheng SS, Wu J. Quality of information in gallstone disease videos on tiktok: Cross-sectional study. J Med Internet Res. 2023;25:e39162. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Questmobile. 2024 New Media Ecosystem Review. https://www.questmobile.cn/research/report/1871401635281997826. Accessed 5 January 2025.
- 21.How many people use. Tiktok 2024 (Statistics & Facts). Ddemandsage. https://www.demandsage.com/tiktok-user-statistics/ Accessed 9 December 2024.
- 22.Survey Data on User Behavior in China’s Short Video Industry. Iimedia. https://www.iimedia.cn/c1077/97043.html Accessed 9 December 2024.
- 23.Zhang S, Yang Y, Yan D. Internet videos and colorectal cancer in Mainland china: a content analysis. BMC Med Inf Decis Mak. 2018;18(1):129. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Hu RH, Zhang HB, Yuan B. Quality and accuracy of gastric cancer related videos in social media videos platforms. BMC Public Health. 2022;22(1):2025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Bernard A, Langille M, Hughes S, et al. A systematic review of patient inflammatory bowel disease information resources on the world wide web. Am J Gastroenterol. 2007;102(9):2070–7. [DOI] [PubMed] [Google Scholar]
- 26.Charnock D, Shepperd S, Needham G, et al. DISCERN: an instrument for judging the quality of written consumer health information on treatment choices. J Epidemiol Commun Health. 1999;53(2):105–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Zheng SS, Tong XY, Wan DL, et al. Quality and reliability of liver Cancer-Related short Chinese videos on TikTok and bilibili: Cross-Sectional content analysis study. J Med Internet Res. 2023;25:e47210. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.The DISCERN. April handbook. Quality criteria for consumer health information on treatment choices. Radcliffe Medical Press. http://www.discern.org.uk/discern.pdf Accessed 25 2025.
- 29.Evaluation Index System of Science Popularization Video. China Science Writers Association. https://www.kpcswa.org.cn/uploads/soft/2308/2023-08-26_001.pdf Accessed 26 November 2024.
- 30.TikTok Discovers Violations of Medical Content Regulations, Reiterates Medical Content Management Rules. Caixin Media. https://www.caixin.com/2021-03-23/101679185.html Accessed 11 November 2024.
- 31.Song SJ, Xue X, Zhao YC, 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. 2021;23(12):e28318. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Fu RL. Knowledge crowdsourcing of interactive videos: a research on the knowledge community of the bilibili website. J Southwest Univ (Social Sci Edition). 2021;47(6):190–199260. [Google Scholar]
- 33.Ming S, Yao X, Zhang SY, et al. Characteristics and quality evaluation of short videos of TikTok on myopia nursing science popularization. Chin J Mod Nurs. 2023;29(36):4976–81. [Google Scholar]
- 34.Mueller SM, Hongler VNS, Jungo P, et al. Fiction, falsehoods, and few facts: Cross-Sectional study on the Content-Related quality of atopic Eczema-Related videos on YouTube. J Med Internet Res. 2020;22(4):e15599. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Paek HJ, Kim K, Hove T. Content analysis of antismoking videos on youtube: message sensation value, message appeals, and their relationships with viewer responses. Health Educ Res. 2010;25(6):1085–99. [DOI] [PubMed] [Google Scholar]
- 36.Ye LJ, Ye YJ, Gao H. Effectiveness of social video platforms in promoting smoking cessation among youth: A content-specific analysis of smoking cessation topic videos on the social platform bilibili. Tob Induc Dis. 2023;21:105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Gao Y, Xin L, Lin H, et al. Machine learning-based automated sponge cytology for screening of oesophageal squamous cell carcinoma and adenocarcinoma of the oesophagogastric junction: a nationwide, multicohort, prospective study. Lancet Gastroenterol Hepatol. 2023;8(5):432–45. [DOI] [PubMed] [Google Scholar]
- 38.Bian Y, Gao Y, Lin H, et al. Correction: Non-invasive diagnosis of esophageal cancer by a simplified Circulating cell-free DNA methylation assay targeting OTOP2 and KCNA3: a double-blinded, multicenter, prospective study. J Hematol Oncol. 2024;17(1):129. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Yuan XL, Guo LJ, Liu W, et al. Artificial intelligence for detecting superficial esophageal squamous cell carcinoma under multiple endoscopic imaging modalities: A multicenter study. J Gastroenterol Hepatol. 2022;37(1):169–78. [DOI] [PubMed] [Google Scholar]
- 40.Batterham RW, Hawkins M, Collins PA, et al. Health literacy: applying current concepts to improve health services and reduce health inequalities. Public Health. 2016;132:3–12. [DOI] [PubMed] [Google Scholar]
- 41.Rikard RV, Thompson MS, McKinney J, et al. Examining health literacy disparities in the united states: a third look at the National assessment of adult literacy (NAAL). BMC Public Health. 2016;16(1):975. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
All data that support the findings of this study is provided within supplementary information files.