Quality and reliability of myocardial infarction-related short videos on TikTok and Bilibili: A cross-sectional content analysis study

Abstract

Background

Myocardial infarction (MI) is a leading cause of global mortality, making public education on its prevention, early recognition, and pre-hospital first aid essential. In China, short-video platforms like TikTok (Douyin) and Bilibili have become primary channels for health information dissemination. However, the quality and reliability of MI-related content on these platforms have not been systematically evaluated. This study aimed to assess and compare the quality, reliability, and accuracy of the most widely viewed (high-visibility) short videos about MI on TikTok and Bilibili and to identify factors associated with video quality.

Methods

This study conducted a cross-sectional content analysis of the top 100 highest-ranking MI-related Chinese short videos retrieved via each platform’s default ranking algorithm from both TikTok and Bilibili. Two independent cardiologists evaluated the videos using the Global Quality Score (GQS), the DISCERN instrument, and a newly developed Medical Information Accuracy Score (MIAS). Basic video characteristics, uploader information, and user engagement metrics were extracted.

Results

Overall, the quality of high-visibility MI-related videos on both platforms was suboptimal. Bilibili videos demonstrated significantly higher quality, reliability, and medical accuracy than those on TikTok (median GQS: 3 vs. 2, P<.001; median DISCERN: 5 vs. 3, P<.001; median MIAS: 7 vs. 4, P<.001). Videos posted by medical professionals and healthcare institutions were of significantly higher quality than those from non-professionals or media/commercial outlets (P<.001). Content created by cardiologists was of the highest quality. Critically, there was no significant correlation between video popularity (i.e.,likes and comments) and informational quality or reliability (P>.05). However, video duration showed a moderate positive correlation with higher quality and reliability scores (GQS: ρ=0.45, P<.001; DISCERN: ρ=0.41, P<.001).

Conclusion

The quality of information in high-visibility myocardial infarction-related videos on China’s major short-video platforms is concerningly poor, with Bilibili providing more reliable and accurate content than TikTok. The source of a video is a key determinant of its quality, but engagement metrics such as “likes” do not reliably signal informational quality. A collaborative effort among healthcare professionals, platform regulators, and the public is urgently needed to improve the quality of this critical health information.

Keywords

myocardial infarction short videos information quality social media TikTok Bilibili

Introduction

Myocardial infarction (MI), a critical manifestation of coronary artery disease, represents a major public health challenge globally and is a leading cause of death and disability in China.¹ The prognosis of MI is highly time-dependent, with immediate and appropriate pre-hospital first aid, along with rapid activation of emergency medical services, being crucial determinants of patient outcomes.² Public awareness of MI symptoms—both typical (e.g., chest pain) and atypical (e.g., toothache, back pain)—and correct emergency responses can significantly reduce treatment delays and mortality.³ Therefore, effective public health education is an indispensable component of MI prevention and management strategies.

With the proliferation of internet technology and smartphones, social media has fundamentally reshaped the landscape of health information dissemination. Short-video sharing platforms, particularly TikTok (known as Douyin in China) and Bilibili, have surged in popularity, attracting hundreds of millions of users and emerging as powerful tools for health communication.⁴ These platforms offer a visually engaging and easily digestible format, making complex medical information more accessible to the general public compared to traditional text-based resources.⁵ Medical professionals and healthcare institutions increasingly utilize these channels to conduct public health education, covering a wide range of topics, including MI.

Despite their potential, the open and unregulated nature of these platforms raises significant concerns about the quality and reliability of the health information presented.^6,7 Unlike chronic disease management, misinformation regarding MI may potentially contribute to delayed help-seeking, the application of unproven first-aid techniques, or reduced trust in evidence-based medicine. While previous studies have assessed the quality of medical information on various topics such as liver cancer,⁸ diabetes,⁹ and mental health¹⁰ across these platforms, a dedicated, comparative analysis of MI-related content on China’s two leading short-video platforms is currently lacking.

Therefore, this study was designed to systematically evaluate the content, quality, and reliability of the most popular short videos concerning myocardial infarction on TikTok and Bilibili. We also aimed to analyze the association between video characteristics (e.g., source, content type, user engagement) and information quality. Through this cross-sectional content analysis, we intend to provide a comprehensive overview of the current information ecosystem, identify gaps and challenges, and offer actionable insights for improving the dissemination of life-saving knowledge about MI via social media.

Methods

Ethical considerations

This study was based exclusively on publicly available data from TikTok (Douyin) and Bilibili. No human subjects, clinical data, or personal privacy information were involved. No personally identifiable information of any individual was collected, recorded, or analyzed. The data collection was conducted in compliance with the publicly available Terms of Service of both platforms. The research did not include any interaction with platform users. Therefore, this study was exempt from institutional review board approval. This study was conducted and reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for cross-sectional studies.

Search strategy and data collection

This cross-sectional study was conducted in March, 2025. We performed searches for the most popular videos on the Chinese versions of TikTok (Douyin) and Bilibili using the keywords “心梗” (xīngěng, myocardial infarction), “心肌梗死” (xīnjīgěngsǐ, myocardial infarction), “心梗急救” (xīngěng jíjiù, MI first aid), and “胸痛” (xiōngtòng, chest pain). To mitigate the influence of personalized recommendation algorithms, new user accounts were registered for each platform and were used for all searches. To ensure consistency, all searches were conducted during a standardized time window (between 9:00 AM and 12:00 PM, Beijing Time). Searches for each platform were completed within the same day to minimize temporal variability in algorithmic recommendations.

The videos were sorted by each platform’s ‘comprehensive ranking’ or ‘default’ algorithm, which typically prioritizes a combination of relevance, recentness, and user engagement. An initial pool of the highest-ranking videos for each keyword was retrieved from both platforms. These videos were then manually screened against the study’s inclusion and exclusion criteria. After removing duplicates, irrelevant content, and other videos that did not meet the criteria, the top 100 eligible videos from each platform were selected for the final analysis. This resulted in a total sample of 200 videos.

Inclusion criteria were: (1) video content directly related to myocardial infarction (including prevention, symptoms, first aid, treatment, or rehabilitation); and (2) language was Mandarin Chinese. Exclusion criteria were: (1) duplicate videos (identical content from different uploaders); (2) irrelevant content (e.g., news reports of a celebrity’s MI without educational content, videos on other heart diseases); (3) non-Chinese videos or those without clear audio/subtitles; and (4) videos that were purely promotional for commercial products or services.

For each included video, the following data were extracted and recorded in a Microsoft Excel spreadsheet (Microsoft Corp): video title, URL, uploader’s name and credentials, video duration, days since publication, and user engagement metrics (number of likes, comments, shares, and saves/collections).

Classification of videos

Videos were categorized based on their source and content.

Video sources were classified into four groups: (1) Medical Professionals, defined as certified physicians (e.g., cardiologists, emergency physicians), nurses, or other registered healthcare practitioners; (2) Healthcare Institutions, including official accounts of hospitals, academic medical centers, or public health organizations (e.g., Center for Disease Control and Prevention); (3) Non-professionals, including patients, family members, or health bloggers without verifiable medical credentials; and (4) Media and Commercial Outlets, including news media organizations and for-profit companies (e.g., pharmaceutical companies).

Video content was categorized into five mutually exclusive themes based on the primary focus of the video: (1) Symptoms and Recognition, covering typical and atypical signs of MI; (2) Pre-hospital First Aid, providing guidance on immediate actions to take during a suspected MI event; (3) Etiology and Prevention, explaining risk factors, pathophysiology, and preventive lifestyle measures; (4) Treatment and Rehabilitation, describing in-hospital procedures (e.g., stenting, bypass surgery) and post-discharge management; and (5) Patient Stories and News, featuring personal anecdotes or news reports without a primary educational focus.

Video quality and reliability assessments

The quality and reliability of the information in each video were assessed using two validated instruments: the Global Quality Score (GQS) and the DISCERN tool.

The GQS is a 5-point scale used to evaluate the overall quality, flow, and usefulness of the information for a lay audience. Scores range from 1 (poor quality, not useful) to 5 (excellent quality, very useful). The detailed scoring criteria for the GQS are shown in Table 1.

Table 1.

Description of the Global Quality Score (GQS) for evaluating the quality of videos with myocardial infarction information.

Scale	Description
1	Poor quality: Poor flow, most critical information is missing or inaccurate, not at all useful and potentially harmful for the public.
2	Generally poor quality: Suboptimal flow, some information is listed but key topics (e.g., calling emergency services) are missing or unclear, of very limited use.
3	Moderate quality: Adequate flow, some important information is discussed but other key aspects are poorly covered or oversimplified, somewhat useful.
4	Good quality: Good flow, most relevant and accurate information is listed, but may lack depth on some topics (e.g., atypical symptoms), useful for the public.
5	Excellent quality: Excellent flow and structure, information is comprehensive, accurate, evidence-based, and clearly explained; very useful for the public.

The DISCERN instrument is a well-established tool for judging the reliability of consumer health information. For this study, we used the first section of the DISCERN tool (Questions 1-8), which focuses on the reliability of the publication. Each question is scored on a scale, but for simplicity and consistency with recent studies, a “yes” response to a reliability criterion was considered favorable. The DISCERN score assesses aspects such as whether the aims are clear, if the information is relevant and balanced, and if sources are provided. The detailed DISCERN criteria are presented in Table 2. Additionally, to specifically evaluate the medical accuracy of the content against established clinical guidelines, a bespoke Medical Information Accuracy Score (MIAS) was developed. The MIAS items were derived from recommendations in current clinical practice guidelines, including the 2023 European Society of Cardiology (ESC) Guidelines for the management of acute coronary syndromes¹¹ and the 2021 American Heart Association/American College of Cardiology (AHA/ACC) Guideline for the Evaluation and Diagnosis of Chest Pain.¹² All items were equally weighted, as each assessed domain (symptom recognition, first-aid appropriateness, prevention accuracy, and treatment information) was considered of comparable importance for the purpose of general public health education. The inter-rater reliability for the MIAS was substantial (Cohen’s κ = 0.76). The MIAS was designed to evaluate the appropriateness of medical information for general public health education, rather than to assess its suitability for individual clinical decision-making. It is acknowledged that certain assessed items, such as recommendations regarding aspirin or nitroglycerin administration, may be subject to regional variation and evolving guideline updates. It should be noted that the MIAS is a novel instrument developed specifically for this study and has not undergone formal external validation. Its findings should therefore be considered exploratory. The detailed scoring criteria for the MIAS are presented in Table 3.

Table 2.

Description of the DISCERN instrument for evaluating the reliability of videos with myocardial infarction information.

DISCERN question number	Question content
Question 1	Are the aims clear? (Does the video state what it is about and what it aims to achieve?)
Question 2	Does it achieve its aims? (Does the video provide the information it promised to deliver?)
Question 3	Is it relevant? (Is the content appropriate and useful for a general audience concerned about MI?)
Question 4	Is it clear what sources of information were used? (Does it cite guidelines, experts, or scientific evidence, beyond the author themselves?)
Question 5	Is it clear when the information was produced? (Is there a mention of the date of information or guideline version?)
Question 6	Is it balanced and unbiased? (Does it present information fairly, avoiding overly sensational or biased language?)
Question 7	Does it provide details of additional sources of support or information? (Does it suggest where to find more information, e.g., consulting a doctor?)
Question 8	Does it refer to areas of uncertainty? (Does it acknowledge that some symptoms can be non-specific or that medical advice is necessary for diagnosis?)

Table 3.

Assessment criteria for the Medical Information Accuracy Score (MIAS).

Assessment dimension	Assessment item	Scoring criteria(0 = incorrect or misleading; 1 = accurate and correct)
1. Symptoms & Recognition	1.1 Accurately describes typical MI symptoms (e.g., chest pain, tightness)?	☐ 0 ☐ 1
1. Symptoms & Recognition	1.2 Mentions atypical MI symptoms (e.g., toothache, back pain, abdominal pain)?	☐ 0 ☐ 1
2. Pre-hospital First Aid	2.1 Emphasizes the importance of calling emergency services immediately?	☐ 0 ☐ 1
	2.2 Advises the patient to rest immediately and avoid movement?	☐ 0 ☐ 1
	2.3 Mentions the use of nitroglycerin or aspirin only under clear medical advice?	☐ 0 ☐ 1
	2.4 Discourages harmful or unproven “first-aid” methods (e.g., forced coughing, slapping the arm)?	☐ 0 ☐ 1
3. Etiology & Prevention	3.1 Correctly identifies major MI risk factors (e.g., hypertension, diabetes, smoking)?	☐ 0 ☐ 1
3. Etiology & Prevention	3.2 Proposes evidence-based preventive measures (e.g., healthy diet, regular exercise)?	☐ 0 ☐ 1
4. Treatment & Rehabilitation	4.1 Presents accurate information on in-hospital treatments (e.g., stenting, bypass surgery)?	☐ 0 ☐ 1
4. Treatment & Rehabilitation	4.2 Provides scientifically sound advice for post-MI rehabilitation?	☐ 0 ☐ 1

Two board-certified cardiologists (Reviewer A, Reviewer B) independently scored all 200 videos. Before the formal evaluation, both reviewers participated in a calibration session to ensure a shared understanding of the scoring criteria, thereby minimizing cognitive bias. To reduce sorting bias, the order of the videos was randomized before assessment. In cases of significant disagreement in scores between the two reviewers, a third senior cardiologist (Reviewer C) acted as an arbitrator to reach a final consensus score. Inter-rater reliability was calculated using Cohen’s kappa (κ). According to the criteria proposed by Landis and Koch, a κ value >0.61 is considered substantial or excellent agreement.

Statistical analyses

Descriptive statistics were used to summarize the video characteristics. As the data were not normally distributed (assessed by the Shapiro-Wilk test), quantitative data were presented as medians and interquartile ranges (IQRs). Categorical data were presented as frequencies and percentages (%).

The Mann-Whitney U test was used for comparisons between two independent groups (e.g., TikTok vs. Bilibili). The Kruskal-Wallis H test was used for comparisons among three or more groups (e.g., different video content types), with Dunn’s post hoc test for pairwise comparisons where appropriate. The chi-square (χ²) test was used to compare the distribution of categorical variables between platforms. Spearman’s rank correlation coefficient (ρ) was used to assess the relationship between video quality/reliability scores and quantitative variables such as user engagement metrics and video duration. All correlation analyses in this study were univariable. No multivariable regression modeling was performed; therefore, the identified associations should not be interpreted as independent predictors of video quality.

All statistical analyses were performed using [SPSS version 26.0 (IBM Corp.) or GraphPad Prism version 9.0.0 for Windows]. A two-tailed P value of <.05 was considered statistically significant.

Results

Video characteristics

A total of 200 videos meeting the inclusion criteria were analyzed, comprising 100 from TikTok and 100 from Bilibili (Figure 1). The general characteristics of the videos on both platforms are presented in Table 4. Videos on TikTok garnered significantly higher user engagement, with a greater median number of likes (median 9850, IQR 1250-34,100 vs. median 890, IQR 210-2500; P<.001), comments (P<.001), and shares (P<.001) compared to videos on Bilibili. Conversely, videos on Bilibili were significantly longer in duration (median 245s, IQR 150-420 vs. median 75s, IQR 45-110; P<.001). No significant difference was observed in the number of days since the videos were published (P=.68).

Figure 1.

Flowchart of the video selection process.

Table 4.

Characteristics of the videos on TikTok and Bilibili.

Variable	TikTok (n=100), median (IQR)	Bilibili (n=100), median (IQR)	P value^a
User Engagement
Likes	9850 (1250-34,100)	890 (210-2500)	<.001
Comments	450 (80-1120)	95 (25-310)	<.001
Shares	1200 (250-4500)	150 (40-550)	<.001
Saves/Collections	2100 (400-6800)	350 (90-980)	<.001
Video Attributes
Duration (seconds)	75 (45-110)	245 (150-420)	<.001
Days since published	365 (180-550)	390 (210-600)	.68
Quality Scores
Global Quality Score (GQS)	2 (1-3)	3 (2-4)	<.001
DISCERN Score	3 (2-4)	5 (3-6)	<.001

IQR: Interquartile Range.

^aP values were calculated using the Mann-Whitney U test.

Video source and content

The distribution of video sources and content types is illustrated in Figure 2. On TikTok (Figure 2(a)), medical professionals were the predominant source of videos, accounting for a substantial majority at 68.0% (68/100). Non-professionals (e.g., patients, unverified bloggers) were the second-largest group of uploaders at 22.0% (22/100), followed by healthcare institutions and media/commercial outlets, each contributing 5.0% (5/100). On Bilibili (Figure 2(b)), medical professionals also constituted the largest source at 55.0% (55/100), but healthcare institutions had a much stronger presence, accounting for 25.0% (25/100) of the videos. Non-professionals and media/commercial outlets on Bilibili were less frequent, at 12.0% (12/100) and 8.0% (8/100), respectively. Regarding video content on TikTok (Figure 2(c)), “Symptoms & Recognition” was the most frequent topic, making up 42.0% (42/100) of the videos. This was followed closely by “Pre-hospital First Aid” at 35.0% (35/100) and “Etiology & Prevention” at 15.0% (15/100). Videos on “Treatment & Rehabilitation” (5.0%) and “Patient Stories & News” (3.0%) were less common. In contrast, the content focus on Bilibili (Figure 2(d)) was different, with “Etiology & Prevention” being the most prominent category at 38.0% (38/100). “Treatment & Rehabilitation” was also a major topic, accounting for 28.0% (28/100) of the videos, followed by “Symptoms & Recognition” (22.0%) and “Pre-hospital First Aid” (12.0%). No videos categorized as “Patient Stories & News” were found in the Bilibili sample. This distribution suggests a tendency on Bilibili toward more in-depth, educational content, whereas TikTok’s content is more oriented towards immediate recognition and emergency response.

Figure 2.

Distribution of video sources and content categories for myocardial infarction-related videos on TikTok and Bilibili. (a) Distribution of uploader sources on TikTok, where medical professionals were the largest group (68.00%), followed by non-professionals (22.00%). (b) Distribution of uploader sources on Bilibili, where medical professionals were the largest group (55.00%), but healthcare institutions had a larger share (25.00%) compared to TikTok. (c) Distribution of video content themes on TikTok, with “Symptoms & Recognition” (42.00%) and “Pre-hospital First Aid” (35.00%) being the most common topics. (d) Distribution of video content themes on Bilibili, where “Etiology & Prevention” (38.00%) and “Treatment & Rehabilitation” (28.00%) were the most prominent categories.

Video quality and reliability assessments

The inter-rater reliability for the scoring was substantial for the GQS (Cohen’s κ = 0.78), excellent for the DISCERN score (Cohen’s κ = 0.82), and substantial for the MIAS (Cohen’s κ = 0.76). The overall quality and reliability of MI-related videos were found to be suboptimal across both platforms. As detailed in Figure 3, Bilibili videos demonstrated significantly higher quality and reliability compared to TikTok videos. The median GQS for videos on Bilibili was 3 (IQR 2-4), corresponding to “Fair” quality, which was significantly higher than the median GQS of 2 (IQR 1-3) on TikTok, corresponding to “Poor” quality (P<.001). Similarly, the median DISCERN score was significantly higher for Bilibili videos (median 5, IQR 3-6) than for TikTok videos (median 3, IQR 2-4; P<.001). In terms of medical accuracy, the MIAS assessment revealed similar trends. Bilibili videos scored significantly higher in accuracy (median 7, IQR 5-8) compared to TikTok videos (median 4, IQR 3-5; P<.001), suggesting better alignment with clinical guidelines.

Figure 3.

Comparison of quality and reliability of myocardial infarction-related videos on TikTok and Bilibili.

To further investigate factors influencing video quality, scores were stratified by source, content, and professional specialty, with the comprehensive results presented in Table 5. Significant variations in quality were observed across different video sources (P<.001). Videos uploaded by medical professionals (median GQS 3, IQR 2-4) and healthcare institutions (median GQS 3, IQR 3-4) received significantly higher GQS and DISCERN scores than those created by non-professionals (median GQS 2, IQR 1-2) and media/commercial outlets (median GQS 1, IQR 1-2). Similarly, video content was a significant determinant of quality (P<.001). Videos focused on “Pre-hospital First Aid” (median GQS 3, IQR 2-4) and “Etiology and Prevention” (median GQS 3, IQR 2-4) demonstrated significantly higher quality and reliability. In contrast, videos categorized as “Patient Stories and News” were of the lowest quality (median GQS 1, IQR 1-1). To explore the impact of specialization within the medical professional group, a sub-analysis was conducted. As shown in Table 6, a significant difference was found among specialties (P=.02). Videos produced by cardiologists achieved the highest median GQS (4, IQR 3-4) and DISCERN scores (6, IQR 5-7), which were significantly higher than those from “Other Physicians” (P<.05) but not statistically different from those by emergency physicians.

Table 5.

Analysis of video quality, reliability, and accuracy scores by video characteristics.

Category	Subcategory	N	Median GQS (IQR)	Median DISCERN (IQR)	Median MIAS (IQR)	P value
Video Source	Medical Professionals	123	3 (2-4)	5 (3-6)	6 (4-7)	<.001
	Healthcare Institutions	30	3 (3-4)	6 (4-7)	7 (5-8)
	Non-professionals	34	2 (1-2)	2 (1-3)	3 (2-4)
	Media & Commercial	13	1 (1-2)	2 (1-2)	3 (1-4)
Video Content	Symptoms & Recognition	64	2 (1-3)	3 (2-4)	4 (3-5)	<.001
	Pre-hospital First Aid	47	3 (2-4)	5 (4-6)	6 (5-7)
	Etiology & Prevention	53	3 (2-4)	5 (3-7)	6 (5-7)
	Treatment & Rehabilitation	33	3 (2-3)	4 (3-5)	5 (4-6)
	Patient Stories & News	3	1 (1-1)	1 (1-2)	2 (1-2)
Professional Specialty	Cardiologists	58	4 (3-4)	6 (5-7)	7 (6-8)	0.02
(Medical Professionals, n=123)	Emergency Physicians	35	3 (2-4)	5 (3-6)	6 (4-7)
(Medical Professionals, n=123)	Other Physicians	30	2 (2-3)	3 (2-4)	4 (3-5)

Table 6.

Spearman correlation analysis between video quality/reliability scores and other video variables.

Variable	Global Quality Score (GQS)	DISCERN score
Variable	ρ (P value)	ρ (P value)
User Engagement
Likes	0.05 (.48)	0.09 (.21)
Comments	0.07 (.32)	0.11 (.12)
Shares	0.18 (.01)	0.15 (.03)
Saves/Collections	0.22 (.002)	0.19 (.007)
Video Attributes
Duration (seconds)	0.45 (<.001)^b	0.41 (<.001)
Days since published	0.12 (.09)	0.14 (.05)
Reliability Score
DISCERN Score	0.76 (<.001)	—

^b P < 0.001, indicating a highly statistically significant correlation.

Correlation analysis

The results of the Spearman correlation analysis are presented in Table 6. As expected, strong, positive correlations were confirmed among all user engagement metrics, such as between likes and shares (ρ=0.91, P<.001).

Crucially, no significant correlation was found between the primary user engagement metrics (likes and comments) and the objective measures of video quality and reliability. The GQS was not significantly correlated with the number of likes (ρ=0.05, P=.48) or comments (ρ=0.07, P=.32). Likewise, the DISCERN score showed no significant correlation with likes (ρ=0.09, P=.21) or comments (ρ=0.11, P=.12). However, weak but statistically significant positive correlations were identified between the GQS and the number of shares (ρ=0.18, P=.01) and saves (ρ=0.22, P=.002). Furthermore, video duration exhibited a moderate, positive correlation with both GQS (ρ=0.45, P<.001) and DISCERN scores (ρ=0.41, P<.001), indicating that longer videos were generally associated with higher quality and reliability.

Discussion

This study provides the first systematic content analysis comparing the quality, reliability, and accuracy of popular MI-related videos on China’s two leading short-video platforms, TikTok and Bilibili. Our analysis reveals several critical findings. First, the overall quality of MI-related information is suboptimal across both platforms, falling short of the standards required for effective public health education. Second, a significant disparity exists between the platforms; Bilibili videos demonstrated markedly higher quality, reliability, and medical accuracy compared to those on TikTok. Third, the source of the video is a powerful determinant of its quality. Content produced by medical professionals and healthcare institutions was significantly superior to that from non-professionals or media outlets. Finally, and perhaps most concerningly, we found no significant correlation between a video’s popularity (i.e.,likes and comments) and its informational quality or reliability. Instead, video duration was moderately and positively correlated with higher quality scores.

The finding of generally low-quality health information on social media is consistent with a growing body of literature highlighting the prevalence of medical misinformation online.¹³ The unregulated nature of these platforms allows for the rapid dissemination of content regardless of its scientific validity. Our results align closely with a similar comparative study on liver cancer videos, which also found that while the overall quality on both TikTok and Bilibili was inadequate, Bilibili offered more reliable information. This platform-specific difference may be attributable to their distinct ecosystems. Bilibili’s platform fosters longer, more in-depth content, which is more conducive to explaining complex medical topics, as supported by our finding that longer videos were of higher quality. In contrast, TikTok’s emphasis on brevity and entertainment may hinder the delivery of nuanced and comprehensive medical advice.¹⁴ It should be noted that these correlations are univariable and do not account for potential confounding factors. The lack of correlation between user engagement and video quality is a critical public health concern and corroborates findings from other studies on health topics like diabetes¹⁰ and ADHD¹¹ on TikTok.

Engagement metrics do not reliably signal informational quality, and the observed disconnect between popularity and accuracy suggests that users cannot depend on these cues to identify trustworthy health content. The mechanisms underlying this disconnect, whether algorithmic or user-driven, warrant further investigation. For a time-sensitive condition like MI, this is particularly dangerous, as the most-liked video may not contain the most reliable life-saving advice.⁴ It is important to note that this study evaluated the informational quality of video content and did not directly assess user behavior, health decision-making, or clinical outcomes. While suboptimal information quality has the potential to contribute to adverse health decisions, this link was not empirically tested in the present study, and such implications should be interpreted with caution.

The results of this study have significant implications for the public, healthcare professionals, and platform regulators. For the public, users should be highly cautious when consuming MI-related information on short-video platforms and should be encouraged to actively verify the source of information, prioritizing content from certified medical professionals and official healthcare institutions. Crucially, the public must understand that a video’s popularity is not an indicator of its accuracy. For healthcare professionals, there is a clear and urgent need to increase their presence on these platforms by creating high-quality, engaging, and evidence-based content to counteract misinformation. Our findings suggest that creating longer, more detailed content could be an effective strategy for health education. For platform regulators, TikTok and Bilibili have a social responsibility to improve the health information ecosystem. They could implement more robust verification systems for healthcare professionals and modify their algorithms to promote content from authoritative sources.

This study has several strengths. To our knowledge, it is the first to directly compare MI-related content between TikTok and Bilibili. The methodology is robust, utilizing two validated instruments (GQS and DISCERN) and a novel, topic-specific accuracy score (MIAS). Furthermore, all videos were independently assessed by two board-certified cardiologists, with high inter-rater reliability, ensuring the clinical validity of our findings. Nevertheless, this study has several limitations that should be acknowledged. First, its cross-sectional design captures a snapshot of the information landscape at a single point in time (March 2025). Given that the content on short-video platforms is highly dynamic and rapidly evolving — with new videos being uploaded continuously and older content being removed or algorithmically deprioritized — the quality metrics reported here may not accurately reflect the current or future state of MI-related content on these platforms. Longitudinal studies with repeated sampling at multiple time points would be valuable to track changes in information quality over time. Second, our sampling strategy relied on platform-specific ‘comprehensive ranking’ or ‘default’ algorithms to identify and select the top-ranked videos. These proprietary algorithms are opaque and are generally believed to prioritize content based on a combination of engagement metrics, recency, and user behavior patterns. This approach may have introduced systematic algorithm-driven sampling bias, favoring popular, recent, or sensationalized content while potentially underrepresenting high-quality but less-promoted videos. As a result, our findings are representative of the most visible and widely consumed MI-related videos, rather than of all available content on these platforms. Future research could employ alternative sampling methods, such as random or stratified sampling across different time periods and ranking positions, to provide a more comprehensive and less biased assessment of the overall information landscape. Third, this study focused exclusively on Mandarin Chinese content on two domestic Chinese platforms (TikTok/Douyin and Bilibili). It is important to note that the Chinese version of TikTok (Douyin) operates as a platform entirely separate from the international version of TikTok, with distinct content moderation policies, user demographics, recommendation algorithms, and regulatory environments. Similarly, Bilibili is a platform unique to the Chinese internet ecosystem with no direct Western equivalent. Therefore, our findings should not be directly extrapolated to the global TikTok ecosystem, other international platforms such as YouTube or Instagram, or to MI-related content produced in other languages or cultural contexts. Future cross-cultural and multilingual studies are needed to determine whether the patterns observed here are consistent across different information environments. Additionally, the MIAS, while clinically intuitive and grounded in established clinical guideline recommendations, is an internally developed instrument without formal external validation. Although it demonstrated substantial inter-rater reliability, future studies are warranted to formally validate this tool and assess its internal consistency across different clinical topics and populations. Furthermore, this study did not differentiate between content related to ST-elevation myocardial infarction and non-ST-elevation myocardial infarction, nor did it specifically evaluate information accuracy regarding reperfusion timelines, which may represent important areas for future research.

Conclusion

In conclusion, the quality of information in high-visibility myocardial infarction-related videos on China’s most popular short-video platforms is concerningly suboptimal. Bilibili provides more reliable, accurate, and higher-quality content than TikTok. While videos from medical professionals are more trustworthy, the public cannot rely on popularity metrics like “likes” to identify accurate information. There is an urgent need for a collaborative effort among healthcare professionals, platform regulators, and the public to improve the quality of this critical health information, which is vital for the prevention and management of a leading cause of death in China and globally.^1,2

Supplemental material

Supplemental material - Quality and reliability of myocardial infarction-related short videos on TikTok and Bilibili: A cross-sectional content analysis study

Supplemental material for Quality and reliability of myocardial infarction-related short videos on TikTok and Bilibili: A cross-sectional content analysis study by Suyin Feng, Long Zhu, Jinyuan Gu, Dashuai Miao, Honglai Zhang, Runfeng Sun in DIGITAL HEALTH.

Footnotes

ORCID iD

Runfeng Sun

Ethical considerations

This study was based solely on publicly available data from TikTok (Douyin) and Bilibili, in full compliance with the platforms’ terms of service. No human subjects, clinical data, or personally identifiable information were involved, and there was no interaction with platform users. The study was therefore exempt from institutional review board approval.

Author contributions

Suyin Feng, Long Zhu, and Jinyuan Gu were responsible for the study design, data collection, analysis, and the drafting of the manuscript. Dashuai Miao participated in the data analysis and manuscript revision. Honglai Zhang and Runfeng Sun were responsible for the conception of the study, supervision, and critical revision of the manuscript. All authors read and approved the final manuscript.

Funding

This study was supported by Jiangnan University Affiliated Hospital and Donghai County People's Hospital Joint Assistance Project for the Construction of a Regional Medical Center Clinical Medical Research Projectl; Kangda College 2024 Scientific Research Development Fund Project (KD2024KYJJ149); Lianyungang City 521 Third Level Project (LYG065212024088) and Lianyungang City Health Commission General Project (202340).

Declaration of conflicting interests

The authors declare that they have no conflict of interest.

Data Availability Statement

The original data used in this study were obtained from TikTok and Bilibili and are publicly available on these platforms. The analyzed datasets generated during the current study are available from the corresponding author upon reasonable request.*

Supplemental material

Supplemental material for this article is available online.

References

Zhao

Liu

Wang

, et al. Epidemiology of cardiovascular disease in China: current features and implications. Nature reviews. Cardiology 2019; 16(4): 203–212. https://doi.org/10.1038/s41569-018-0119-4

Akbik

Xian

, et al. Trends in Reperfusion Therapy for In-Hospital Ischemic Stroke in the Endovascular Therapy Era. JAMA Neurol 2020; 77(12): 1486–1495. https://doi.org/10.1001/jamaneurol.2020.3362

Moser

Kimble

Alberts

, et al. Reducing delay in seeking treatment by patients with acute coronary syndrome and stroke: a scientific statement from the American Heart Association Council on cardiovascular nursing and stroke council. Circulation 2006; 114(2): 168–182. https://doi.org/10.1161/CIRCULATIONAHA.106.176040

Chen

Wang

. Social Media Use for Health Purposes: Systematic Review. J Med Internet Res 2021; 23(5): e17917. https://doi.org/10.2196/17917

Zhu

Zhang

, et al. How Health Communication via Tik Tok Makes a Difference: A Content Analysis of Tik Tok Accounts Run by Chinese Provincial Health Committees. International journal of environmental research and public health 2019; 17(1): 192. https://doi.org/10.3390/ijerph17010192

Suarez-Lledo

Alvarez-Galvez

. Prevalence of Health Misinformation on Social Media: Systematic Review. J Med Internet Res 2021; 23(1): e17187. https://doi.org/10.2196/17187

Bailey

Huynh

, et al.

YouTube as a source of information on COVID-19: a pandemic of misinformation?

BMJ Glob Health 2020; 5(5). https://doi.org/10.1136/bmjgh-2020-002604

Zheng

Tong

Wan

, 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. https://doi.org/10.2196/47210

Kong

Song

Zhao

, et al. TikTok as a Health Information Source: Assessment of the Quality of Information in Diabetes-Related Videos. J Med Internet Res 2021; 23(9): e30409. https://doi.org/10.2196/30409

10.

Yeung

Abi-Jaoude

. TikTok and Attention-Deficit/Hyperactivity Disorder: A Cross-Sectional Study of Social Media Content Quality. Canadian journal of psychiatry. Revue canadienne de psychiatrie 2022; 67(12): 899–906. https://doi.org/10.1177/07067437221082854

11.

Byrne

Rossello

Coughlan

, et al. 2023 ESC Guidelines for the management of acute coronary syndromes. Eur Heart J 2023; 44(38): 3720–3826. https://doi.org/10.1093/eurheartj/ehad191

12.

Gulati

Levy

Mukherjee

, et al. 2021 AHA/ACC/ASE/CHEST/SAEM/SCCT/SCMR Guideline for the Evaluation and Diagnosis of Chest Pain: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. J Am Coll Cardiol 2021; 78(22): e187–e285. https://doi.org/10.1016/j.jacc.2021.07.053

13.

Phiri

Makowa

Amelia

, et al. Text-Based Depression Prediction on Social Media Using Machine Learning: Systematic Review and Meta-Analysis. J Med Internet Res 2025; 27: e59002. https://doi.org/10.2196/59002

14.

Woertz

Srivatsan

Kersten

. Orbital Emphysema Caused by a Tik-Tok Challenge. Ophthalmology 2024; 131(12): 1476. https://doi.org/10.1016/j.ophtha.2023.11.013

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.08 MB

0.00 MB