Evaluating the quality and reliability of pulmonary embolism-related short videos on Bilibili and TikTok: A cross-sectional content analysis

Abstract

Pulmonary embolism (PE) is a life-threatening cardiovascular emergency, yet public awareness remains insufficient. Social media platforms like TikTok and Bilibili have become crucial channels for the general population to access health information. Evaluating the quality and reliability of their content is significant for enhancing public health communication. Employing a cross-sectional design, this investigation systematically retrieved and analyzed short videos pertaining to pulmonary embolism on TikTok and Bilibili, incorporating 186 videos (99 from TikTok, 87 from Bilibili). Video quality and reliability were appraised utilizing the Global Quality Score (GQS) and the modified DISCERN (mDISCERN) tools. Spearman correlation analysis was conducted to explore associations among video length, engagement metrics, and quality/reliability scores. Significant differences were observed between TikTok and Bilibili in video duration, engagement metrics (likes, comments, shares), and quality/reliability scores (P < .05). Videos on Bilibili were significantly longer than those on TikTok, whereas TikTok exhibited superior audience interaction. Video content predominantly emphasized clinical manifestations and etiology, with inadequate attention accorded to prevention and epidemiology. Videos uploaded by pulmonologists significantly outperformed those from other healthcare professionals and science communicators in mDISCERN scores (P = .004). Video length was positively correlated with both GQS and mDISCERN scores. Engagement metrics were strongly intercorrelated; however, higher engagement did not necessarily correspond to higher video quality or reliability. Short-video platforms hold potential for disseminating pulmonary embolism-related health information, but content quality is inconsistent, and high engagement does not equate to high quality. It is recommended that platforms optimize content review mechanisms and encourage professional physicians to participate in content creation to enhance the accuracy and educational value of health-related information.

1. Introduction

Pulmonary embolism (PE) refers to a spectrum of diseases or clinical syndromes caused by the obstruction of the pulmonary arteries or their branches by various emboli. As the most severe clinical manifestation of venous thromboembolism (VTE), PE exhibits persistently high global incidence and mortality rates, ranking as the third leading cause of cardiovascular death worldwide, following myocardial infarction and stroke.^[1,2] Epidemiological studies indicate that the annual incidence of PE ranges from approximately 39 to 115 per 1,00,000 individuals, with a 30-day mortality rate reaching 10% to 30%.^[3-5] In China, the number of newly diagnosed PE patients exceeds 2,00,000, with a population incidence rate as high as 14.19 per 1,00,000 and a mortality rate of 1 in 1,00,000, showing a significant increase compared to previous data.^[6] The risk of developing PE continues to rise due to factors such as population aging, increasing obesity rates, and a growing volume of surgical procedures. The clinical presentation of PE is often nonspecific, depending on the size, number, and site of the emboli, as well as the presence of underlying cardiopulmonary diseases in the patient. Pulmonary embolism is associated with high rates of misdiagnosis and missed diagnosis in clinical practice, which not only delays timely treatment but also increases patient mortality and healthcare costs.^[7-9] Therefore, enhancing public awareness of PE, particularly regarding risk factors, preventive measures, and early symptoms, is crucial for promoting timely medical intervention and ultimately reducing the overall disease burden.

With the rapid development of digital media, social media platforms have emerged as powerful tools for health communication and public education. Their vast user bases, algorithmically driven personalized recommendation mechanisms, and elevated interactive features propel the expedited, widespread circulation of healthcare content. Platforms such as TikTok and Bilibili have become primary channels for accessing health-related content, attracting substantial user groups.^[10-12] However, these very characteristics that facilitate rapid information spread also raise concerns regarding content quality and reliability. The low barriers to content creation may compromise the accuracy of the information environment. Previous studies have demonstrated that despite high user engagement, videos related to anal fissures,^[13] thyroid nodules,^[14] and cervical cancer^[15] generally exhibit low overall quality and reliability, potentially exacerbating the spread of misinformation and hindering accurate diagnosis and timely management.

This study is the first to simultaneously employ GQS and mDISCERN tools to assess the quality and reliability of short videos related to pulmonary embolism on both TikTok and Bilibili. It further analyzes the differences between platforms, the identity of uploaders, and the relationship between engagement metrics and content quality. This study contributes an empirical basis to content creators, platform regulators, and healthcare professionals to optimize the dissemination of digital health information related to pulmonary embolism, thereby offering the public more accurate and reliable knowledge.

2. Materials and methods

2.1. Ethical considerations

This study was reviewed and granted an exemption from formal ethics approval by the Human Research Ethics Committee of the Fourth Affiliated Hospital, Zhejiang University School of Medicine. This decision was made because the research solely involved the analysis of existing, publicly available, and anonymized short video data. No private user information, clinical data, human specimens, or animal subjects were involved, and there was no interaction with human participants.

2.2. Data source

This cross-sectional study systematically collected video data related to “肺栓塞” (the Chinese term for pulmonary embolism) published on Bilibili and TikTok. The data were collected on September 15, 2025, and aimed to assess the content, quality, and reliability of pulmonary embolism-related videos. To minimize bias from the personalized recommendation algorithms of TikTok and Bilibili, a new account with browser cache cleared and no viewing history was used for the searches. For each platform, we screened the top 100 videos displayed under the platform’s default comprehensive search ranking. No additional manual sorting by likes, comments, shares, upload date, or popularity was applied by the investigators.

2.3. Inclusion and exclusion criteria

Videos were screened based on the following inclusion criteria: primarily focused on medical information related to pulmonary embolism; presented in Chinese or with Chinese subtitles; and publicly accessible at the time of data collection. The exclusion criteria were as follows: content not primarily related to pulmonary embolism; pure foreign language videos; and duplicate videos.

2.4. Data collection

Selected videos were evaluated across multiple dimensions, including video length, platform source, uploader identity, content topics, and engagement metrics. Engagement metrics included likes, comments, collections, and shares. These metrics were extracted on September 15, 2025, at the time each video was accessed, and were recorded as cross-sectional real-time counts. Changes in engagement metrics over time were not longitudinally tracked in the present study. View/play counts were not included in the primary analysis because they were not consistently available across both platforms at the time of data extraction. Therefore, to maintain consistency across platforms, the main analysis was based on engagement metrics that could be uniformly retrieved from both platforms. The unit of analysis was the individual video. Eligible videos were included according to the predefined inclusion and exclusion criteria, and videos were not deduplicated by uploader account.

Uploaders were categorized into the following groups: pulmonologists, science communicators, and other healthcare professionals. Video content was analyzed for key themes including epidemiology, etiology, clinical manifestations, diagnosis, treatment, and prevention. Epidemiology referred to information on disease incidence, prevalence, distribution, or high-risk populations; etiology referred to causes or risk factors; clinical manifestations referred to symptoms and signs; diagnosis referred to diagnostic tests, imaging examinations, or diagnostic criteria; treatment referred to therapeutic strategies such as anticoagulation, thrombolysis, interventional therapy, or surgery; and prevention referred to measures aimed at reducing the risk of pulmonary embolism. These content categories were not mutually exclusive, and a single video could be assigned to multiple categories if it covered more than 1 topic. Video quality and reliability were assessed using the GQS and mDISCERN tools.

2.5. Quality and reliability assessment

Video quality and reliability were evaluated using the GQS and mDISCERN tools. The GQS assesses overall online video quality through a 5-point scale, ranging from 1 (poor) to 5 (excellent).^[16] The mDISCERN tool evaluates reliability via 5 binary (yes/no) questions, yielding a total score between 0 and 5, where higher scores denote greater reliability.^[17,18] Detailed criteria for these 2 scales are presented in Tables S1 and S2, Supplemental Digital Content 1. An independent blinded assessment process was implemented: 2 raters scored the videos independently without knowledge of each other’s ratings. In cases of disagreement, a third assessor made the final decision. Prior to scoring, all raters underwent standardized training to ensure consistency and minimize bias.

2.6. Statistical analysis

According to previous literature,^[19,20] continuous variables are described using median and interquartile range (IQR). Differences between 2 groups were analyzed using the Mann–Whitney U test, while differences among multiple groups were examined using the Kruskal–Wallis H test. Moreover, referring to previous cross-sectional studies using correlation analysis,^[21-23] Spearman correlation analysis was used to examine associations among video length, engagement metrics, and quality/reliability scores. Because the correlation matrix involved multiple pairwise correlation tests, Benjamini–Hochberg false discovery rate correction was applied. All statistical analyses were conducted using R software (version 4.3.0; The R Foundation for Statistical Computing), with P < .05 considered statistically significant.

3. Results

3.1. Video characteristics

After applying the exclusion criteria, 186 videos were included in the study, with the detailed selection process illustrated in Figure 1. Among these, 99 videos were sourced from TikTok (53.23%) and 87 videos were from Bilibili (46.77%; Fig. 2). Table 1 presents the video characteristics across the 2 platforms, including video length, audience engagement metrics, and evaluation indicators. The results indicated statistically significant disparities between TikTok and Bilibili in video length, likes, comments, shares, GQS, and mDISCERN scores (P < .05), while no significant difference was observed in collections (P > .05). Notably, the video length on Bilibili was significantly longer than that on TikTok (P < .001). TikTok exhibited higher audience interaction, with median values for likes (median: 266, IQR: 75–1407.5), comments (median: 20, IQR: 5.5–80), and shares (median: 50, IQR: 8–431; P < .001), whereas Bilibili showed universally lower metrics. Engagement counts differed across platforms; however, because these values may be affected by platform traffic, user base, recommendation mechanisms, account follower base, and time since upload, cross-platform comparisons were interpreted descriptively. Table 2 provides an in-depth analysis of video features categorized by the identity of the uploader. Statistically significant differences were found among the 3 groups (pulmonologists, science communicators, other healthcare professionals) for likes, comments, and mDISCERN scores (P < .05), whereas no significant differences were observed for video length, collections, shares, and GQS scores (P > .05). Videos uploaded by science communicators exhibited significantly longer lengths and performed well on engagement metrics compared to the other uploader categories.

Figure 1.

Flowchart illustrating the video selection process for pulmonary embolism-related short videos on TikTok and Bilibili.

Figure 2.

Proportion of science communication videos on both platforms.

Table 1.

General information, quality, and reliability scores of pulmonary embolism videos on TikTok and Bilibili.

Variables	Total (n = 186)	Bilibili (n = 87)	TikTok (n = 99)	P
General information
Video length(s), M (Q1, Q3)	165.50 (77.00, 1129.50)	1201.00 (320.00, 2032.50)	96.00 (63.00, 159.00)	<.001
Likes, M (Q1, Q3)	94.00 (23.25, 576.00)	18.00 (6.00, 95.50)	266.00 (75.00, 1407.50)	<.001
Collections, M (Q1, Q3)	68.50 (18.25, 336.00)	70.00 (21.00, 182.50)	67.00 (16.00, 466.00)	.635
Comments, M (Q1, Q3)	5.50 (1.00, 28.75)	1.00 (0.00, 4.50)	20.00 (5.50, 80.00)	<.001
Shares, M (Q1, Q3)	23.50 (7.00, 163.50)	18.00 (4.00, 48.50)	50.00 (8.00, 431.00)	<.001
Video content
Epidemiology	47 (7.68%)	40 (12.38%)	7 (2.42%)
Etiology	126 (20.59%)	67 (20.74%)	59 (20.42%)	–
Clinical manifestation	152 (24.84%)	73 (22.6%)	79 (27.34%)	–
Diagnosis	104 (16.99%)	66 (20.43%)	38 (13.15%)	–
Treatment	109 (17.81%)	54 (16.72%)	55 (19.03%)	–
Prevention	74 (12.09%)	23 (7.12%)	51 (17.65%)	–
Video quality
GQS score, M (Q1, Q3)	4.00 (4.00, 4.00)	4.00 (4.00, 5.00)	4.00 (4.00, 4.00)	<.001
mDISCERN score, M (Q1, Q3)	4.00 (3.00, 4.00)	4.00 (3.00, 5.00)	4.00 (3.00, 4.00)	.004

GQS = Global Quality Score, mDISCERN = modified DISCERN.

Table 2.

Characteristics, quality, and reliability of pulmonary embolism videos by different uploaders on TikTok and Bilibili.

Variables	Total (n = 186)	Pulmonologist (n = 49)	Science communicators (n = 20)	Other healthcare professionals (n = 117)	P
Video length, M (Q1, Q3)	165.50 (77.00, 1129.50)	175.00 (110.00, 1525.00)	291.50 (93.75, 1923.50)	135.00 (72.00, 834.00)	.082
Likes, M (Q1, Q3)	94.00 (23.25, 576.00)	51.00 (9.00, 939.00)	87.50 (14.75, 240.50)	128.00 (35.00, 710.00)	.041
Collections, M (Q1, Q3)	68.50 (18.25, 336.00)	39.00 (15.00, 444.00)	86.50 (24.25, 364.50)	80.00 (19.00, 300.00)	.556
Comments, M (Q1, Q3)	5.50 (1.00, 28.75)	2.00 (0.00, 29.00)	3.50 (1.00, 10.00)	11.00 (2.00, 36.00)	.022
Shares, M (Q1, Q3)	23.50 (7.00, 163.50)	16.00 (3.00, 132.00)	18.50 (7.00, 115.50)	26.00 (8.00, 168.00)	.274
GQS, M (Q1, Q3)	4.00 (4.00, 4.00)	4.00 (4.00, 5.00)	4.00 (3.00, 5.00)	4.00 (4.00, 4.00)	.057
mDISCERN, M (Q1, Q3)	4.00 (3.00, 4.00)	4.00 (3.00, 5.00)	3.00 (3.00, 4.00)	4.00 (3.00, 4.00)	.004

GQS = Global Quality Score, mDISCERN = modified DISCERN.

3.2. Uploader characteristics

In this study, the 186 included videos were classified according to uploader identity, including 49 videos uploaded by pulmonologists (26.34%), 20 videos uploaded by science communicators (10.75%), and 117 videos uploaded by other healthcare professionals (62.90%), as illustrated in Figure 3. The distribution of uploaders differed between platforms. On Bilibili, other healthcare professionals accounted for 51% of the uploaders, followed by pulmonologists at 28%, and science communicators representing the smallest proportion (22%). Conversely, on TikTok, other healthcare professionals were predominant (74%), followed by pulmonologists (25%), while science communicators contributed minimally, accounting for only 1% of uploaders, as detailed in Figure 4. This analysis revealed the trend in the proportions of video uploaders across both platforms is consistent, with other healthcare professionals serving as the main contributors on both Bilibili and TikTok.

Figure 3.

Distribution of uploaders.

Figure 4.

Comparison of uploader proportions between platforms.

3.3. Video content

In the video content analysis section, as presented in Table 1 and Figure 5, the pulmonary embolism-related video content primarily covered topics including epidemiology, etiology, clinical manifestation, diagnosis, treatment, and prevention. Among the 186 videos, clinical manifestation was the most frequently addressed topic (152 videos), followed by etiology (126 videos). Both of these factors exhibited high mention rates across TikTok and Bilibili. Figure 6 illustrates the proportional distribution of content topics by platform. On both platforms, the top 2 most covered topics were clinical manifestation (TikTok: 27%, Bilibili: 23%) and etiology (TikTok: 20%, Bilibili: 21%). Divergence in content focus emerged from the third-ranked topic onwards. On TikTok, the subsequent order was treatment (19%), prevention (18%), diagnosis (13%), and epidemiology (2%). In contrast, on Bilibili, the order was diagnosis (20%), treatment (17%), epidemiology (12%), and prevention (7%). This indicated a difference in content emphasis between the 2 platforms.

Figure 5.

Distribution of video content categories.

Figure 6.

Comparison of content category distribution between platforms.

3.4. Video quality and reliability assessments

This study compared video quality and reliability between TikTok and Bilibili and among different uploader categories. As shown in Table 1, the overall quality of most videos published on both TikTok and Bilibili was acceptable. The GQS scores for Bilibili (median: 4.00, IQR: 4.00–5.00) were significantly higher than those for TikTok (median: 4.00, IQR: 4.00–4.00). A similar trend was observed for mDISCERN scores, with Bilibili achieving higher scores (median: 4.00, IQR: 3.00–5.00) than TikTok (median: 4.00, IQR: 3.00–4.00). A comparative analysis of GQS and mDISCERN scores (Fig. 7A, B) demonstrated a statistically significant difference between the platforms (P < .05), pointing to an advantage in video quality and reliability for the Bilibili platform. The distribution and differences in mDISCERN and GQS scores for both platforms were illustrated in Figure 8A, B.

Figure 7.

Comparison of video quality and reliability scores between TikTok and Bilibili. (A) GQS scores; (B) mDISCERN scores. GQS = Global Quality Score, mDISCERN = modified DISCERN.

Figure 8.

Distribution of quality and reliability scores across platforms and uploader categories. (A) GQS scores between platforms; (B) mDISCERN scores between platforms; (C) GQS scores among different uploader categories; (D) mDISCERN scores among different uploader categories. GQS = Global Quality Score, mDISCERN = modified DISCERN.

The GQS and mDISCERN scores of videos were also compared among different uploader identities, as detailed in Table 2. Notably, differences were observed in GQS scores among uploaders, with pulmonologists receiving higher GQS scores (median: 4.00, IQR: 4.00–5.00) than those from other healthcare professionals (median: 4.00, IQR: 4.00–4.00) and science communicators (median: 4.00, IQR: 3.00–5.00), although the difference did not reach statistical significance (P = .057). In the mDISCERN scoring system, videos uploaded by pulmonologists (median: 4.00, IQR: 3.00–5.00) achieved significantly higher scores than those by other healthcare professionals (median: 4.00, IQR: 3.00–4.00) and science communicators (median: 3.00, IQR: 3.00–4.00), with the difference being statistically significant (P = .004). The distribution and differences in mDISCERN and GQS scores among the different uploader categories are presented in Figure 8C, D.

3.5. Correlation analysis

Spearman correlation analysis was performed to explore associations among video length, engagement metrics, and the GQS and mDISCERN scores. Because the correlation matrix included multiple pairwise tests, Benjamini–Hochberg false discovery rate correction was applied. Spearman ρ, exact P values, and FDR-adjusted q values are provided in Table S3, Supplemental Digital Content 2.

As illustrated in Figure 9, engagement metrics showed strong positive intercorrelations. Likes were strongly correlated with comments (ρ = 0.92, P < .0001), shares (ρ = 0.81, P < .0001), and collections (ρ = 0.76, P < .0001). Collections were also strongly correlated with shares (ρ = 0.88, P < .0001) and comments (ρ = 0.65, P < .0001). Comments were also positively correlated with shares (ρ = 0.72, P < .0001). Video length was positively correlated with both GQS (ρ = 0.63, P < .0001) and mDISCERN scores (ρ = 0.51, P < .0001). A positive correlation was also observed between GQS and mDISCERN scores (ρ = 0.59, P < .0001). Regarding the associations between engagement metrics and quality/reliability scores, likes were negatively correlated with both GQS (ρ = −0.37, P < .0001) and mDISCERN scores (ρ = −0.34, P < .0001). Comments were also negatively correlated with GQS (ρ = −0.37, P < .0001) and mDISCERN scores (ρ = −0.33, P < .0001). Shares were negatively correlated with GQS (ρ = −0.23, P < .001) and mDISCERN scores (ρ = −0.30, P < .0001). In contrast, collections were not significantly correlated with either GQS (ρ = −0.06, P = .39) or mDISCERN scores (ρ = −0.13, P = .07) after FDR correction. These findings indicate that higher engagement was not necessarily associated with higher video quality or reliability.

Figure 9.

Correlation matrix of video length, engagement metrics, and video quality scores. The numbers represent correlation coefficients. * indicates P < .05, ** indicates P < .01, *** indicates P < .001, and **** indicates P < .0001. GQS = Global Quality Score, mDISCERN = modified DISCERN.

4. Discussion

This cross-sectional study systematically assessed the content, quality, and reliability of 186 short videos pertaining to pulmonary embolism on 2 major platforms, TikTok and Bilibili. By integrating both GQS and mDISCERN scoring tools, it conducted the first multi-dimensional analysis encompassing platform differences, uploader identity, video content, and interaction metrics. The findings revealed a statistically significant distinction between the 2 platforms in terms of user engagement (likes, comments, shares), content quality, and reliability. Although the overall quality of videos on both platforms was acceptable, Bilibili significantly outperformed TikTok in both quality and reliability. The professional background of uploaders notably influenced information reliability, with videos uploaded by pulmonologists achieving significantly higher mDISCERN scores than those from other uploader groups. Furthermore, video length was positively correlated with quality. However, user interaction metrics (likes, collections, comments, shares) showed no significant correlation with video quality or reliability, suggesting that users may struggle to judge content scientificity based on popularity alone. The most frequently covered topics across both platforms were clinical manifestations and etiology, while epidemiology and prevention received comparatively less attention. These findings highlight the potential of social media for pulmonary embolism education while simultaneously underscoring the urgent need to enhance professional involvement and strengthen content governance when disseminating pulmonary embolism-related information on social media platforms. It is crucial to encourage the participation of specialized physicians in creating science communication content and to guide the public in scientifically identifying high-quality health information.

4.1. Video interaction metrics

This study analyzed the interaction metrics of 186 pulmonary embolism-related short videos and identified significant differences in audience engagement between the TikTok and Bilibili platforms. Videos on TikTok were notably shorter in duration (median: 96.00 seconds, IQR: 63.00–159.00), which aligns with the overall trend of video lengths on that platform.^[24,25] This brief format caters to users’ preference for rapid content consumption, facilitating quicker information dissemination.^[26] Despite their shorter length, TikTok videos demonstrated higher audience engagement in metrics such as likes, comments, and shares. In contrast, while Bilibili’s longer video format allows for more in-depth discussion, results in lower immediate interaction. This observation is in line with earlier findings highlighting differences in user behavior and content consumption patterns across platforms.^[27,28] These differences in interaction suggest that the dissemination effectiveness of short videos depends not only on the content itself but is also closely related to platform-specific user characteristics, video appeal, and engagement mechanisms. Therefore, platforms and health communicators should not equate popularity with quality. Future research should further explore how to optimize short video content to enhance audience engagement, ensuring that comprehensive and accurate information also possesses appeal, particularly on knowledge-focused platforms like Bilibili.

4.2. Video content analysis

In terms of video content, this study found that clinical manifestations and etiology received adequate coverage. Given the characteristics of pulmonary embolism, information on prevention and clinical manifestations is crucial for the early identification and management of the disease. The dissemination of such information in public health communication helps improve public awareness of the disease. However, content related to prevention, diagnosis, and epidemiology was relatively sparse. For instance, epidemiology-related content accounted for only 2% on TikTok and merely 12% on Bilibili. This phenomenon suggests that while the public may acquire partial knowledge about pulmonary embolism, there is insufficient understanding of its incidence, distribution patterns, risk factors, and high-risk populations, which can hinder early recognition and intervention.^[29,30] Given that pulmonary embolism often presents acutely with severe consequences, effective prevention is paramount.^[31-34] Yet, prevention-related content constituted only 7% of videos on Bilibili. The lack of emphasis on preventive measures for pulmonary embolism may lead to insufficient public vigilance when facing high-risk factors. The inadequate coverage of epidemiology and preventive measures could result in incomplete public knowledge regarding the prevention and management of pulmonary embolism, thereby affecting the implementation of preventive behaviors.^[35,36] Therefore, future short video content creators and platforms should strengthen their focus on areas such as prevention, diagnosis, and epidemiology of pulmonary embolism to ensure that audiences receive balanced and comprehensive health information.

4.3. Differences in quality and reliability between platforms

The overall quality of pulmonary embolism-related videos was acceptable, yet significant differences were observed between platforms. This study found that both the GQS and mDISCERN scores on Bilibili were significantly higher than those on TikTok, indicating superior content quality and information reliability in pulmonary embolism-related videos on Bilibili. This finding is consistent with previous research on diseases such as prostate cancer and knee osteoarthritis.^[37,38] As a platform primarily focused on knowledge sharing, Bilibili’s content creators tend to emphasize knowledge depth and logical presentation, and its review mechanisms are potentially more rigorous. Furthermore, a positive correlation was observed between video length and both quality and reliability on Bilibili, suggesting that longer videos facilitate more comprehensive explanations, citation of authoritative sources, and clearer organization of medical content. In contrast, TikTok, dominated by short-form videos, emphasizes fast-paced and entertaining content, which may constrain the delivery of depth and rigorous information.^[39-41] Therefore, it is recommended that the TikTok platform encourage creators to produce medium-length videos or adopt series formats to enhance content completeness and scientific rigor. Additionally, the platform should strengthen its review and tiered recommendation mechanisms for medical content, prioritizing the visibility of high-quality content published by users with verified professional certifications in relevant specialties. Such measures are crucial for improving overall video quality and fostering audience trust.

4.4. Differences in quality and reliability among different uploaders

Content from pulmonologists was notably associated with higher mDISCERN scores compared to that from other healthcare professionals and science communicators. This highlights the value of subject matter experts in disseminating accurate medical knowledge and their superior professionalism in conveying information reliability. However, the differences in GQS scores among the 3 groups did not reach statistical significance, indicating that the overall content quality still requires improvement. This finding aligns with previous research indicating that videos uploaded by professional physicians hold advantages in terms of medical accuracy and information completeness.^[42,43] Notably, although videos from pulmonologists demonstrated higher reliability, they exhibited lower audience engagement compared to videos from other healthcare professionals and science communicators. Furthermore, shortcomings were observed in areas such as the citation of sources of information, which affected their scientific rigor. Therefore, it is recommended that platforms establish standardized modules for publishing medical content, guiding uploaders to standardize content structure and encourage the citation of relevant authoritative guidelines, thereby fostering the creation of high-quality medical science communication and ensuring information accuracy and reliability. Furthermore, viewers should strengthen their ability to discern content sources and consider the professional background of uploaders when selecting videos to watch, thereby better obtaining reliable information.

4.5. Correlation analysis

Correlation analysis revealed that the interaction metrics exhibited strong positive intercorrelations (r-values ranging from 0.76 to 0.92), indicating a synergistic effect among user behaviors. Videos with higher popularity typically exhibited active engagement across multiple dimensions, a finding consistent with that reported by Li et al.^[44] However, no significant correlations were found between interaction metrics and either the GQS or mDISCERN scores. This “quality-popularity paradox” consistently observed in studies evaluating health information on social media, reaffirms the phenomenon that “popularity does not equate to quality.” It demonstrates that high interaction rates do not necessarily reflect superior content quality, which aligns with previous studies on medical short videos related to gallstone disease and hypertension.^[45,46] This indicates that users may find it challenging to judge the scientificity of content based solely on interaction data, implying that platform recommendation algorithms should deemphasize the weight of engagement metrics and strengthen the evaluation of content quality. Furthermore, this study identified no significant correlation between video duration and interaction metrics, indicating that user engagement is not driven by a preference for either longer or shorter videos but depends more on content appeal and dissemination strategies. Conversely, video duration showed positive correlations with both the GQS score (r = 0.63, P < .05) and the mDISCERN score (r = 0.51, P < .05). This result suggests that longer videos facilitate the delivery of more systematic, comprehensive, higher-quality, and more reliable information, providing crucial guidance for content creators in designing video length. This finding is consistent with studies on topics such as lymphedema and colorectal polyps,^[47,48] supporting the view that “medium-to-long videos are more conducive to the dissemination of medical knowledge.” This insight underscores the importance of balancing content depth and quality during video production. It is recommended that creators ensure informational richness while simultaneously maintaining the video’s appeal and audience engagement.

4.6. Limitations

Several limitations associated with the present study should be noted. First, there is a potential for selection bias, as the analysis focused exclusively on pulmonary embolism-related videos from 2 Chinese platforms (TikTok and Bilibili) and did not include content from international short-video platforms. Second, the sample size was limited; for each platform, only the top 100 videos displayed under the platform’s default comprehensive search ranking were screened. Although we standardized the search procedure by using newly created accounts, clearing browser cache, and avoiding prior viewing history, platform search and ranking algorithms are dynamic and may vary over time, even when standardized search procedures are used, which may affect the reproducibility of the sample and the external validity of the findings. Third, view/play counts, an important indicator of exposure scale, were not included because they were not consistently available across both platforms. Therefore, engagement metrics may not be directly comparable across platforms, as they can be affected by platform-level traffic, account follower base, recommendation algorithms, metric definitions, and time since upload. Additionally, the video evaluation process inevitably introduced a degree of subjectivity, which could influence the objectivity of the results. Finally, given the dynamic nature of social media, temporal bias is a relevant factor; the analyzed videos and engagement metrics represent a specific point in time, and engagement metrics may have changed after data extraction. Therefore, future research should expand the sample size and consider more diverse platforms, repeated sampling at different time points, view/play counts when consistently available, and additional evaluation criteria to enhance the comprehensiveness and accuracy of the study.

5. Conclusion

This study conducted the first systematic evaluation of interaction metrics, video quality, and reliability concerning pulmonary embolism-related content on the TikTok and Bilibili platforms. Our findings indicate that Bilibili generally provides higher-quality and more reliable content, while content uploaded by pulmonologists demonstrates superior reliability. However, the platform exhibits imbalanced content coverage, with important topics such as epidemiology and prevention being underrepresented. These findings suggest that the dissemination effectiveness of short videos is shaped by multiple factors, including platform characteristics, content format, and the professional background of uploaders. Short-video platforms hold significant potential for medical science communication. Future research should continue to investigate these variables to enhance the overarching quality level of the content and advance the dissemination of scientific, systematic, and trustworthy health knowledge.

Author contributions

Conceptualization: Wang Lv, Menglu Zhu.

Data curation: Wang Lv, Lijiang Wang, Miaolian Wu.

Formal analysis: Wang Lv, Lijiang Wang, Miaolian Wu.

Methodology: Wang Lv, Menglu Zhu.

Writing – original draft: Wang Lv.

Writing – review & editing: Menglu Zhu.