The reach-rigor paradox in health communication: a comparative analysis of congenital heart disease content quality and engagement on TikTok and Bilibili Authors: Ming-Tang Ye, Yang Wang, Wen-Xin Lin, Al-wajih Hamzah, Yu-zhong Yang, Ji-rong Qi Published: 2026 ABSTRACT AbstractBackgroundCongenital heart disease (CHD) is the most common birth defect and a major contributor to childhood morbidity and mortality. Short-video platforms—particularly TikTok and Bilibili—have become prominent sources of public health information. However, the quality and reliability of CHD-related educational content on these platforms remain insufficiently characterized.MethodsWe conducted a cross-sectional analysis in October 2025 by retrieving the top 100 CHD-related videos from TikTok and Bilibili (n = 200). Two independent reviewers assessed each video using the Global Quality Scale (GQS) and a modified DISCERN instrument. Video characteristics, uploader type, and user engagement indicators were extracted. Analyses included descriptive statistics, nonparametric group comparisons, and Spearman correlation tests.ResultsOverall video quality was moderate to high, with a median GQS score of 4 (IQR 3-4), while reliability was moderate, with a median modified DISCERN score of 3 (IQR 3-4). Medical professionals were the most common uploaders (49%). A notable tension between reach and scientific rigor was observed: TikTok videos achieved substantially higher engagement (median likes 733 vs 60; p < 0.001) yet exhibited lower average GQS scores than Bilibili. Narrative/storytelling formats generated the highest interaction (median likes 3,248 on TikTok) but received the lowest quality and reliability ratings, whereas expert explanations and animation/diagram-based videos showed superior credibility. Engagement indicators displayed weak to negligible negative correlations with GQS and modified DISCERN scores. After 2023, uploader profiles shifted toward professionalization (medical professional uploaders increased from 26.3% to 62.9%), but this change was not associated with statistically significant improvements in quality metrics.ConclusionAlthough CHD-related short videos on major platforms demonstrate generally acceptable quality, content popularity remains poorly aligned with scientific reliability. These findings highlight the need to balance engagement with accuracy in digital health communication. Future efforts should strengthen science communication training for professional creators and implement platform-level quality assurance mechanisms to enhance the public health value of short-video content.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12872-026-06019-w.Keywords: Congenital heart disease, Short video, Health information, Quality, Reliability, TikTok, Bilibili, Content analysis Introduction Congenital heart disease (CHD) refers to structural abnormalities of the heart or great vessels due to aberrant embryonic development, representing the most prevalent congenital malformation, with a global birth incidence around 0.8% to 1.2% [1]. Clinical manifestations vary widely from asymptomatic presentations to life-threatening heart failure, impaired growth, and mortality, constituting a leading cause of pediatric morbidity and mortality worldwide [2]. Advances in prenatal screening, perioperative management, and surgical or interventional techniques have significantly improved survival outcomes, shifting CHD from a childhood disease to a chronic, lifelong condition [3]. Many survivors face residual lesions and long-term complications—such as arrhythmias, ventricular dysfunction, and the need for reintervention—emphasizing the importance of structured, lifelong follow-up [4]. In China, despite notable progress in diagnosis and treatment, persistent challenges in long-term care coordination, regional disparity, and health education highlight the need for strengthened public awareness and systematic follow-up systems [5]. Health education plays a critical role in chronic disease management. For CHD, effective education enhances caregivers’ understanding of disease pathophysiology and treatment options, reduces anxiety, and improves adherence to medical recommendations. Studies show that video-based educational interventions can improve knowledge retention and reduce anxiety in hospitalized patients and caregivers compared to traditional methods [6]. Moreover, multimedia and visual aids have been shown to enhance attention, comprehension, and recall of medical information, which is particularly valuable when conveying complex cardiac anatomy and procedures to non-specialist audiences [7]. Short-video platforms such as TikTok and Bilibili have emerged as prominent channels for health information dissemination in China. Recent analyses indicate that these platforms influence health behaviors but exhibit considerable variability in content quality and reliability [8]. Systematic reviews of health information on TikTok note widespread inconsistencies in medical accuracy and highlight the frequent use of instruments like DISCERN for quality assessment, underscoring the need for disease-specific evaluations [9]. However, the rapid proliferation of short-form content and algorithm-driven distribution also amplifies exposure to misleading or incomplete health claims. The World Health Organization has identified “infodemics”—an overabundance of information including misinformation—as a barrier to effective public health communication, capable of delaying appropriate care and undermining trust in health systems [10]. Despite the widespread availability of CHD-related content on short-video platforms, comprehensive, cross-platform assessments focusing specifically on CHD remain scarce. To address this gap, this study collected and evaluated CHD-related short videos from TikTok and Bilibili using standardized quality appraisal tools to assess information reliability, characterize content formats and sources, and examine dissemination patterns. By mapping the current landscape of CHD-related health communication on major Chinese video platforms, this work aims to inform strategies for improving content quality, platform governance, and health literacy among patients and caregivers. Methods Search strategy and data processing To minimize potential biases introduced by personalized recommendation algorithms, searches were conducted using newly registered accounts in incognito mode on the web versions of both platforms, without applying any filtering criteria. Search results were displayed according to each platform’s default comprehensive ranking algorithm, which integrates video completion rate (the proportion of viewers who watched more than 5 s), like rate, comment rate, follow rate, and upload time, to recommend relevant videos. In this cross-sectional study, “先天性心脏病” (congenital heart disease), “先心病”, and “小儿心脏病” were used as keywords to search for the top 100 videos on the Chinese versions of TikTok and Bilibili between October 20 and October 25, 2025 (Fig. 1). The analysis was limited to the top 100 videos, as previous studies have confirmed that videos beyond this threshold do not significantly impact analysis results [15]. Duplicate videos, those lacking clear authorship attribution, and those meeting other predefined exclusion criteria were removed. For multi-part series, all parts were consolidated and analyzed as a single unit. This screening process continued until 100 eligible videos were identified from each platform (Fig. 1). ༎Search strategy and video screening procedure The following essential information was extracted from each video: video title, upload date, uploader identity (e.g., official medical institution, medical professional, health science popularization media, commercial organization, individual user), video type (e.g., expert explanation, Animation/diagram demonstration, live-action recording/demonstration, narrative/storytelling, hybrid type), Video content (e.g., treatment, postoperative rehabilitation, etiology, disease knowledge, diagnosis, advertisement or other information) (Supplementary Table 1) and the number of likes, saves, comments, duration, days since published and shares. All data were recorded and organized in Microsoft Excel for subsequent compilation and analysis. Assessment methods and procedure Two researchers independently reviewed the content of all videos and evaluated them using the Global Quality Scale (GQS) and a modified version of the DISCERN instrument (Supplementary Tables 2 and Supplementary Table 3). The GQS is a standardized tool widely used to assess the quality of health information in videos, with a scoring range of 1 to 5 (1: very poor quality, 5: excellent quality) [10–11]. The modified DISCERN instrument was employed to evaluate the reliability and evidence base of the videos. DISCERN is a well-validated tool originally consisting of 15 questions (each scored on a 5-point Likert scale) designed to assess the quality of health information on treatment options available to patients. Among the 15 items, 8 questions pertain to the reliability of publications, which were adopted in this study. The remaining 7 questions address the quality of information regarding treatment choices and were not applicable to our analysis. The full set of 15 questions can be adapted to various topics in patient education. The original DISCERN instrument also includes a 16th question regarding the overall rating of the information material, which was not assessed in this section as it is essentially equivalent to the Global Quality Scale (GQS). This assessment covers the following eight dimensions: Is the purpose of the video clear? Does it achieve its purpose? Is the information relevant? Is the source of information clear? Is the information up-to-date? Is the information balanced and fair? Does it provide references or sources for additional information? Does it address areas of uncertainty or describe treatment choices [12, 13]? . These eight items cover key dimensions of information quality: clarity of purpose, relevance, source transparency, timeliness, balance, references, and acknowledgment of uncertainty. The two evaluators scored each video based on these eight dimensions using a 5-point Likert scale: 1 point (not at all), 2 points (partially, but with serious shortcomings), 3 points (partially, with some information), 4 points (mostly), and 5 points (completely). Each dimension receives a score ranging from 1 to 5. The total score for the modified DISCERN instrument is calculated by averaging the scores of all eight items and is also reported on a scale from 1 to 5 to facilitate comparison with the GQS score. To further aid interpretation, we classified the total scores into five reliability levels: 1 (unreliable, score range 5–11), 2 (less reliable, 12–18), 3 (fairly reliable, 19–25), 4 (relatively reliable, 26–32), and 5 (reliable, 33–40) ( Supplementary Table 4). This adaptation and scoring approach align with methodologies used in prior similar studies [14, 15]. Statistical analysis Categorical variables were expressed as frequencies and percentages and compared using the Chi-square test or Fisher’s exact test, as appropriate. Continuous variables were first tested for normality using the Kolmogorov-Smirnov test. Variables not conforming to a normal distribution were described using the median and interquartile range (IQR), whereas those conforming to a normal distribution were described using the mean and standard deviation (SD). Depending on the distribution of the data, group comparisons for continuous variables were conducted using Student’s t-test or the Mann-Whitney U test. The inter-rater agreement between the two independent evaluators was assessed using Cohen’s kappa coefficient (κ), with the specific methodology implemented in accordance with previous studies [16]. In this study, the Cohen’s κ coefficient was 0.85 for the GQS assessments and 0.82 for the modified DISCERN assessments, indicating excellent inter-rater agreement. Associations between different variables were examined using Spearman’s rank correlation analysis. For comparisons among multiple groups, the Bonferroni correction method was applied to adjust for multiple testing. All statistical tests were two-sided, and a P-value less than 0.05 was considered statistically significant. All statistical analyses were performed using SPSS version 26.0 (IBM Corp., Chicago, IL, USA) and R version 4.2.3 statistical software (www.r-project.org). Results Video characteristics A total of 342 relevant short videos were initially collected from the two platforms, Tiktok and Bilibili. After applying the exclusion criteria, 18 non-Chinese or non-English videos, 28 videos with advertising or conflicts of interest, 20 duplicate videos, 19 videos containing only imagery without narration or subtitles, and 57 videos unrelated to the core topic were excluded. Consequently, a final sample of 200 short videos was included for subsequent analysis and evaluation. Table 1 presents a comparative analysis of video characteristics and quality scores between the Bilibili and TikTok platforms. Regarding engagement metrics, videos on TikTok received significantly higher median counts of likes (733 vs. 60, p < 0.001), shares (49 vs. 26, p = 0.001), and comments (100 vs. 4, p < 0.001) compared to those on Bilibili, while no statistically significant difference was observed in the number of saves (p = 0.341). In terms of video attributes, TikTok videos were notably shorter in duration (median 80 s) than Bilibili videos (median 265 s) (p < 0.001). Additionally, the time since publication was significantly shorter for TikTok videos (median 81 days) compared to Bilibili videos (median 1,037 days) (p < 0.001). Table 1Characteristics of the videos on TikTok and BiliBiliCharacteristicBilibili, Median (Q1, Q3)Tiktok, Median (Q1, Q3)p-valueLikes (counts)60 (16, 334)733 (282, 2,273)< 0.001Saves (counts)91 (19, 311)90 (41, 343)0.341Shares (counts)26 (6, 128)49 (20, 230)0.001Comments (counts)4 (1, 25)100 (43, 386)< 0.001Duration (seconds)265 (150, 488)80 (55, 123)< 0.001Days since published1,037 (524, 1,404)81 (50, 135)< 0.001 Characteristics of the videos on TikTok and BiliBili Based on data presented in Supplementary Table 6, Supplementary Tables 6 and Fig. 2, a cross-platform analysis of video sources, content types, and engagement metrics for health-related content on TikTok and Bilibili, distinct platform-specific patterns emerge. On TikTok, videos were predominantly uploaded by medical professionals (67%), followed by individuals (15%), commercial organizations (9%), official medical institutions (6%), and health science popularization media (3%). In contrast, Bilibili exhibited a more balanced source distribution: commercial organizations (28%), health science popularization media (23%), medical professionals (31%), individuals (12%), and official medical institutions (6%). Regarding video type, TikTok content was dominated by expert explanations (40%) and narrative/storytelling (25%), while Bilibili featured more animation/diagram demonstrations (28%) and hybrid formats (25%). In terms of content focus, disease knowledge was the most common category on both platforms (TikTok: 35%, Bilibili: 38%), followed by treatment-related content (TikTok: 33%, Bilibili: 34%). Engagement metrics revealed notably higher median likes, saves, and comments on TikTok across most categories, whereas Bilibili videos generally had longer durations and older publication dates. These findings suggest that TikTok favors concise, professionally sourced explanatory content with higher immediate interaction, while Bilibili accommodates longer-form, diversely sourced content with sustained visibility over time. Percentage of videos on CHD from different sources, different types and with different contents in TikTok and Bilibili. A Sources of Tiktok videos. B Sources of Bilibili videos. C Content types of Tiktok videos. D Content types of Bilibili videos. E Types of Tiktok videos. F Types of Bilibili videos The results of the comparative analysis are presented in Table 2. Statistically significant differences were observed in several metrics between videos published in or before 2023 and those published after 2023. Videos published after 2023 exhibited significantly higher median numbers of likes (410 vs. 127, p = 0.006) and comments (52 vs. 10, p < 0.001). In contrast, videos from the earlier period (2023 and prior) demonstrated significantly higher median numbers of saves (173 vs. 62, p = 0.026) and longer median durations (195 s vs. 104 s, p < 0.001), while no statistically significant difference was found in shares (p = 0.199). Additionally, following 2023, there was a substantial increase in videos uploaded by medical professional (20 vs. 78) and in videos featuring expert explanations (16 vs. 51). Notably, no significant differences were observed between the two periods in the median Global Quality Score (p = 0.489) or the median DISCERN score (p = 0.394). Table 2Comparison of video characteristics between periods (before and after 2023)CharacteristicFor the years 2023 and priorN = 76For the years following 2023N = 124p-valueLikes, Median (Q1, Q3)127 (28, 802)410 (134, 1,408)0.006Saves, Median (Q1, Q3)173 (29, 489)62 (33, 175)0.026Shares, Median (Q1, Q3)73 (14, 170)32 (11, 174)0.199Comments, Median (Q1, Q3)10 (2, 104)52 (16, 248)< 0.001Duration, Median (Q1, Q3)195 (118, 400)104 (65, 223)< 0.001Days since published, Median (Q1, Q3)1,280 (981, 1,605)91 (54, 270)< 0.001Global quality score, Median (Q1, Q3)4.00 (3.00, 4.00)4.00 (3.00, 5.00)0.489DISCERN score, Median (Q1, Q3)3.00 (3.00, 4.00)3.00 (3.00, 4.00)0.394Video source, n (%)< 0.001Commercial organization21 (27.6%)16 (12.9%)Health science popularization media15 (19.7%)11 (8.9%)Individuals15 (19.7%)12 (9.7%)Medical professional20 (26.3%)78 (62.9%)Official medical institution5 (6.6%)7 (5.6%)Video type, n (%)0.003Animation/diagram demonstration21 (27.6%)15 (12.1%)Expert explanation16 (21.1%)51 (41.1%)Hybrid type20 (26.3%)21 (16.9%)Live-action recording/demonstration3 (3.9%)12 (9.7%)Narrative/storytelling16 (21.1%)25 (20.2%) Comparison of video characteristics between periods (before and after 2023) Video quality and reliability assessments Table 3 presents the detailed DISCERN and GQS results for these videos. For TikTok videos, the median GQS score was 4 (interquartile range [IQR]: 3–4), and the median DISCERN score was 3 (IQR: 3–4), indicating overall moderate-to-good quality and moderate reliability of the videos. For Bilibili videos, the median GQS score was 4 (IQR: 4–4.5), and the median DISCERN score was 3 (IQR: 3–4), suggesting generally acceptable quality but moderate reliability. As shown in Fig. 3, a statistically significant difference was observed in GQS scores between TikTok and Bilibili videos (P < 0.05), while no statistically significant difference was found in DISCERN scores. However, both GQS and DISCERN scores were slightly higher for Bilibili compared to TikTok (Fig. 4). Table 3The GQS and DISCERN score based on different video sources, contents and formats about CHD on TikTok and BiliBiliVariableBilibiliTiktokGQS, Median (IQR)DISCERN score, Median (IQR)GQS, Median (IQR)DISCERN score, Median (IQR)Overall score4.0 (4.0, 4.5)3.0 (3.0, 4.0)4.0 (3.0, 4.0)3.0 (3.0, 4.0)Video typeAnimation/diagram demonstration5.0 (4.0, 5.0)4.0 (4.0, 4.0)4.0 (3.0, 4.5)3.0 (3.0, 4.0)Expert explanation4.0 (4.0, 4.0)3.0 (3.0, 4.0)4.5 (4.0, 5.0)4.0 (3.0, 4.0)Hybrid type4.0 (4.0, 4.0)3.0 (3.0, 4.0)3.0 (2.0, 3.0)3.0 (2.0, 3.0)Live-action recording/demonstration2.0 (1.5, 3.0)2.5 (1.5, 3.0)4.0 (3.0, 4.0)3.0 (3.0, 4.0)Narrative/storytelling3.5 (3.0, 4.0)3.0 (3.0, 3.0)2.0 (2.0, 3.0)2.0 (2.0, 3.0)Video contentAdvertisement or other information4.0 (4.0, 5.0)4.0 (3.0, 4.0)4.0 (3.0, 5.00)4.0 (3.0, 4.0)Diagnosis5.0 (3.5, 5.0)4.0 (3.5, 4.0)4.0 (3.0, 4.0)3.0 (3.0, 4.0)Disease knowledge4.0 (3.0, 4.0)3.0 (3.0, 4.0)3.0 (3.0, 4.0)3.0 (3.0, 4.0)Etiology4.0 (3.0, 5.0)3.0 (3.0, 4.0)3.0 (3.0, 4.0)3.0 (3.0, 3.5)Postoperative Rehabilitation4.0 (3.0, 4.0)3.0 (3.0, 3.0)3.5 (3.0, 5.0)3.0 (2.0, 4.0)Treatment4.0 (4.0, 5.0)3.0 (3.0, 4.0)4.0 (3.0, 5.0)3.0 (3.0, 4.0)Video sourcesCommercial organization4.0 (4.0, 4.0)3.0 (3.0, 3.0)4.0 (3.0, 4.0)3.0 (3.0, 4.0)Health science popularization media4.0 (3.0, 4.0)3.0 (3.0, 4.0)4.0 (4.0, 4.0)4.0 (3.0, 5.0)Individuals3.0 (2.5, 4.0)3.0 (3.0, 3.5)2.0 (1.0, 2.0)2.0 (1.0, 2.0)Medical professional5.0 (4.0, 5.0)4.0 (3.0, 4.0)4.0 (3.0, 5.0)3.0 (3.0, 4.0)Official medical institution4.5 (4.0, 5.0)4.0 (4.0, 4.0)4.0 (3.0, 5.0)3.5 (2.0, 4.0) The GQS and DISCERN score based on different video sources, contents and formats about CHD on TikTok and BiliBili Global quality scores, DISCERN scores, and quality/reliability distributions of short videos related to CHD on TikTok and Bilibili. A Comparison of global quality scores between TikTok and Bilibili videos. B Proportions of different levels of video quality. C Comparison of DISCERN scores between TikTok and Bilibili videos. D Proportions of different levels of video reliability. ns: not significant at P < 0.05; ** significant at P < 0.05 Comparison of Global Quality and DISCERN Scores Between TikTok and Bilibili Videos When comparing the eight individual items of the DISCERN instrument, TikTok scored higher than Bilibili only on items 5 and 8, whereas Bilibili outperformed TikTok on the remaining five items (Fig. 5). Comparison of DISCERN Scores Across Eight Items Between TikTok and Bilibili Videos We also compared GQS and DISCERN scores across videos from different sources, video types, and content categories. Videos featuring expert explanations and those utilizing animations/graphics demonstrated higher GQS and DISCERN scores compared to other formats (Fig. 6). Furthermore, videos covering diagnostic-related content had higher GQS scores than other content categories, while no statistically significant difference was observed in DISCERN scores (Fig. 7). Videos produced by professionals or professional institutions exhibited higher GQS and DISCERN scores than those created by other contributors (Fig. 8). Global Quality and DISCERN Scores of CHD Videos from Different Video types; (A) Global quality scores; (B) DISCERN Scores. *P<0.05, **P<0.01, ****P<0.001 Global Quality and DISCERN Scores of CHD Videos from Different contents, (A) Global quality scores; (B) DISCERN Scores. *P<0.05, **P<0.01, ****P<0.001 Global Quality and DISCERN Scores of CHD from Different Video sources, (A) Global quality scores; (B) DISCERN Scores. *P<0.05, **P<0.01, ****P<0.001 Correlation analysis Spearman correlation analysis was performed to assess the relationships between video interaction metrics (including likes, comments, saves, and shares) and quality assessment scores (GQS and DISCERN). As shown in Fig. 9, the results revealed significant positive correlations among the interaction metrics. For instance, strong positive correlations were observed between likes and comments (r = 0.90), between saves and shares (r = 0.89), between likes and shares (r = 0.76), and between likes and saves (r = 0.74). In contrast, correlations between interaction metrics and the quality assessment scores (GQS and DISCERN) were weak to negligible, indicating minimal association between user engagement indicators and content quality as measured by these standardized tools. Correlation matrix of Spearman’s correlation analysis between video metrics and quality scores Discussion Key findings This study systematically evaluated the information quality, reliability, and dissemination dynamics of congenital heart disease (CHD)–related educational short videos on TikTok and Bilibili. Overall, the sampled videos demonstrated moderate-to-high quality and moderate reliability, with median Global Quality Score (GQS) and modified DISCERN (mDISCERN) scores of 4 and 3, respectively —two commonly used instruments for appraising online health information [17, 18]. A notable tension between reach and scientific rigor was observed: TikTok videos attracted substantially higher engagement (likes, shares, comments) yet exhibited lower average quality and reliability than Bilibili, where videos tended to score better on scientific rigor despite comparatively lower interaction metrics [19, 20]. This divergence underscores that popularity indicators on short-video platforms are not reliable proxies for informational credibility [21]. Video format and source emerged as key determinants of both dissemination and quality. Narrative/storytelling videos consistently generated the greatest interaction but scored lowest in quality and reliability [22]. In contrast, expert-led explanations and animation/diagram-based presentations achieved the strongest credibility profiles [19, 20]. Notably, animation-based videos also supported deeper engagement through higher rates of sharing and saving, suggesting that well-designed visual explanations can improve accessibility without sacrificing accuracy [19]. Finally, temporal analyses indicated an apparent professionalization trend after 2023, characterized by increased participation by medical professionals and a higher prevalence of expert-explanation formats [19, 20]. However, this shift did not translate into statistically significant improvements in overall GQS or mDISCERN scores, suggesting that professional identity alone does not ensure high-quality science communication [17, 18]. In parallel, engagement patterns evolved toward more immediate interactions (more likes and comments) but reduced saving behavior, consistent with lightweight, real-time consumption patterns on short-form video platforms [20, 21]. Quality assessment of chd-related educational short videos Platform-level patterns: engagement advantage vs. quality advantage The observed platform contrast may reflect differences in algorithmic distribution and user expectations. For instance, Douyin’s recommendation mechanisms and entertainment-oriented viewing contexts could preferentially amplify emotionally resonant or attention-optimized content, whereas Bilibili users might be more inclined toward structured, knowledge-seeking viewing, which could favor higher-quality educational materials [19, 20]. However, these interpretations are speculative, as our observational study did not directly measure the underlying causes of platform differences. Importantly, our findings reinforce that dissemination success and informational rigor can diverge, and both dimensions should be evaluated independently [21]. Format effects: the cost of virality and the value of structure Video format showed a robust association with both engagement and quality. Narrative/storytelling content, while highly shareable and interactive, frequently lacked components central to scientific communication—such as balanced framing, explicit evidence cues, clear uncertainty boundaries, and actionable guidance—thereby receiving lower GQS and mDISCERN scores [18, 22]. In these videos, persuasive strength appeared to arise primarily from emotional resonance rather than informational completeness [22]. By contrast, expert explanations and animation/diagram-based presentations achieved the highest quality and credibility ratings [19, 20]. Animations were particularly notable: beyond strong reliability scores, they were associated with higher levels of deeper engagement (e.g., saving and sharing), implying that visualization can promote understanding and retention of complex CHD concepts [19]. This supports the practical value of medically grounded visual aids in translating specialized knowledge to lay audiences [7]. Live-action demonstrations performed relatively poorly across metrics, which may be attributable to higher production barriers, limited didactic structure, and reduced suitability for short-video viewing dynamics [19, 20]. Source effects: authority–reach trade-offs Content source further revealed a clear trade-off between authority and reach. Videos published by official institutions generally achieved superior quality and reliability, yet were less likely to attain viral dissemination [19, 20]. Conversely, highly popular personal narratives often exhibited weaker scientific structure and fewer credibility signals [22]. These results suggest that improving public health communication on short-video platforms requires not only increasing professional participation, but also optimizing the packaging, clarity, and platform-fit of authoritative content [20, 21]. Temporal trends: professionalization without measurable quality gains Although uploader composition shifted toward medical professionals after 2023, overall quality metrics did not significantly improve (GQS p = 0.489; DISCERN p = 0.394), indicating that professional identity alone does not guarantee high-quality science communication [19, 20]. Persisting deficits likely include structural completeness, lay-friendly clarity, and explicit sourcing—all essential to reliability appraisal frameworks [17, 18]. Simultaneously, we observed an engagement shift characterized by increased likes and comments but decreased saves (median 173 vs. 62, p = 0.026), suggesting a move from deliberate information retention to instantaneous endorsement and discussion [21]. Practical implications As digital media increasingly mediates health knowledge acquisition, patients and the public have shifted from passive recipients to active seekers of information [23]. Evidence from randomized and quasi-experimental research suggests that video-based health education can improve knowledge and short-term self-management outcomes compared with traditional print materials, particularly when content is structured and tailored to audience needs [24, 25]. For pediatric conditions such as CHD, audiovisual education may be especially valuable for caregivers navigating complex anatomical, procedural, and longitudinal management concepts [26, 27]. However, the uneven quality of CHD short videos presents nontrivial risks: partial or unbalanced messages can reinforce misconceptions; inaccurate claims may delay appropriate care-seeking; and overly simplified narratives can obscure individualized decision-making and clinical heterogeneity [22]. Accordingly, we recommend developing a scientifically rigorous yet accessible CHD public education ecosystem on short-video platforms. First, establishing cross-platform certification and collaboration mechanisms involving platforms, tertiary hospitals, and academic societies may increase the visibility of credible materials [28]. Second, tiered, audience-tailored resource libraries should be curated and optimized for platform-specific viewing behaviors [23]. Third, hybrid governance models combining algorithmic distribution with professional review may help balance dissemination efficiency with quality assurance, particularly for high-reach content [28]. Beyond professional participation, efforts should prioritize enhancing clinicians’ science communication competencies, including narrative structuring, jargon reduction, evidence signaling, and visualization design [23]. Future directions and limitations Several limitations warrant consideration. First, restricting sampling to the top-exposure videos (100 per platform) improves ecological relevance but may underrepresent the long-tail content ecology [19, 20]. Second, focusing on Chinese-language platforms limits generalizability to other linguistic contexts; cross-linguistic comparisons with global platforms (e.g., YouTube) are needed [21]. Third, although GQS and mDISCERN are widely used, they entail subjective judgment and potential inter-study variability [17, 18]. Fourth, platform algorithms and rapid content turnover may affect temporal stability, supporting the need for longitudinal monitoring [20]. In addition, a substantial difference in video age was observed between platforms, with Bilibili videos having a significantly longer median publication lag (1,037 days) compared to TikTok videos (81 days), which may have influenced cumulative engagement metrics. This temporal discrepancy should therefore be taken into account when interpreting cross-platform comparisons. Future research should (i) develop platform-adapted CHD content standards and audit protocols [28], (ii) assess whether higher-quality content predicts measurable changes in caregiver knowledge, care-seeking behavior, or adherence [24, 25], and (iii) explore scalable assistance tools (e.g., automated claim detection and evidence-cue checks) to support human review [23]. Despite these limitations, this study provides baseline evidence on the CHD short-video information landscape and highlights a persistent gap between reach and rigor, offering actionable directions for improving the quality and public value of platform-based health education. Conclusion This study systematically assessed 200 congenital heart disease (CHD)–related science popularization videos on TikTok and Bilibili and identified a central tension in contemporary short-video health communication: content reach tends to increase as scientific reliability declines, though the strength of this inverse association is modest. Although medical professionals have become the predominant content creators, professional identity alone did not translate into a consistent improvement in overall informational quality. Instead, both content format and source were decisive in shaping dissemination and credibility. Narrative-style videos generated the highest levels of user interaction but frequently lacked scientific rigor, whereas expert-led explanations and animated/graphic presentations achieved stronger quality and reliability yet faced constraints in achieving comparable reach. Similarly, experience-based personal narratives drew substantial attention but often raised concerns regarding evidentiary support and balance, while videos released by reputable organizations were typically the most credible but struggled to attain equivalent dissemination impact. Collectively, these findings suggest that short-video health science communication has entered a stage in which optimizing both quality and dissemination effectiveness is essential. Future efforts should prioritize two complementary strategies. First, targeted training in media literacy and science communication for professional creators is needed to strengthen their ability to translate specialized knowledge into clear, audience-appropriate, and engaging content. Second, platforms should refine governance mechanisms by embedding professional review into algorithmic distribution, including scientific verification and risk-sensitive moderation for high-reach CHD content. By building an ecosystem that aligns professionalism, accessibility, and scalable reach, short-video platforms may better realize their public-health potential and provide more reliable informational support for CHD prevention, diagnosis, treatment, and long-term management. Supplementary Information Supplementary Material 1. Acknowledgements The authors would like to thank the creators of the videos on TikTok and Bilibili for making their content publicly available. Abbreviations Congenital Heart Disease Global Quality Scale Interquartile Range Standard Deviation Authors’ contributions MT.Y. conceived and designed the study. AW.H. collected the top 200 videos. WX.L. collected the characteristics of the videos and authors. MT.Y., YZ.Y., and Y.W. were responsible for reviewing, classifying, and scoring the videos. YZ.Y, JR.Q. analyzed the data. MT.Y. wrote the original draft. Y.W. reviewed and edited the manuscript. AW.H, JR.Q. critically revised the manuscript for intellectual content and offered the funding. All the authors contributed to manuscript writing and editing and approved the final draft for submission. Funding None. Data availability Raw data were generated at Children’s Hospital of Nanjing Medical University. Derived data supporting the findings of this study are available from the corresponding authors on request. Declarations Ethics approval and consent to participate No clinical data, human specimens, or laboratory animals were used in this study. All information was obtained from publicly available videos on TikTok and Bilibili. The study did not involve any interaction with video creators or users, and no personally identifiable information was collected or analyzed. Therefore, formal ethical approval was not required for this study. Consent for publication All authors have read and approved the final manuscript and consent to its publication. 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. Ming-Tang Ye , Yang Wang and Wen-Xin Lin contributed equally to this work and should be regarded as co-first authors. Contributor Information Yu-zhong Yang, Email: 373728125@qq.com. Ji-rong Qi, Email: 570683100@qq.com. References Associated Data Supplementary Materials Supplementary Material 1. Data Availability Statement Raw data were generated at Children’s Hospital of Nanjing Medical University. Derived data supporting the findings of this study are available from the corresponding authors on request.