Quality,engagement,and predictive validity of acne-related short videos on Chinese platforms Bilibili and TikTok: A cross-sectional content analysis

Abstract

Objective

To evaluate the informational quality and user engagement of acne-related videos on Bilibili and TikTok, and examine associations with uploader characteristics, disease-related topics, presentation formats, and factors linked to high-quality content.

Methods

A cross-sectional analysis was conducted on 272 videos (122 from Bilibili, 150 from TikTok) retrieved in May 2025. Video characteristics, uploader types, disease-related topics, and presentation formats were recorded. Quality was assessed using the Journal of the American Medical Association benchmark criteria (JAMA), modified DISCERN instrument (mDISCERN), Global Quality Scale (GQS), and Video Information and Quality Index (VIQI). Engagement metrics (likes, comments, collections, shares) were analyzed. Correlation and predictive modeling were applied to examine associations between quality and engagement.

Results

Bilibili videos were longer (median 409 s vs 51 s; P < 0.001) and scored higher on VIQI (12.05 ± 2.94 vs 10.90 ± 1.97; P = 0.002). TikTok videos were more often uploaded by verified (68.10% vs 33.33%) and professional accounts (65.52% vs 25.64%), achieved higher JAMA (1.28 ± 0.45 vs 0.92 ± 0.98; P < 0.001) and mDISCERN scores (2.10 ± 0.52 vs 2.02 ± 0.93; P = 0.009), and demonstrated higher daily engagement. High-quality content was primarily produced by verified and professional uploaders, particularly in anatomy/physiology topics and doctor monologues. Official media, epidemiology, and television programs/documentaries achieved the greatest engagement. VIQI and GQS were strongly correlated (ρ = 0.755). VIQI (area under the curve [AUC] 0.922) and collections (AUC 0.901) were the strongest discriminators of high-quality content.

Conclusions

Acne-related videos on Bilibili and TikTok were generally of suboptimal quality. Bilibili favored coherence and accuracy, while TikTok favored transparency and engagement. Quality assessments outperformed engagement metrics in identifying high-quality content. These findings highlight the need to improve credentialing and promote engaging, evidence-based formats to enhance the reliability and impact of dermatologic information on short-video platforms.

Keywords

Acne vulgaris short-video platforms video quality user engagement predictive validity

Introduction

Acne vulgaris is a chronic inflammatory skin disease of the pilosebaceous unit, characterized by a range of lesions including comedones, papules, pustules, nodules, and cysts.¹ It often leads to post-inflammatory hyperpigmentation, permanent scarring, and psychological comorbidities such as anxiety and depression, contributing to a substantial cosmetic and psychosocial burden.^1,2 Epidemiological studies have identified acne as the eighth most prevalent disease globally, with a particularly high burden among adolescents in East Asia.³ Recent advances have expanded the understanding of acne pathophysiology and treatment strategies.⁴ However, a persistent gap remains between evidence-based recommendations and real-world treatment decisions.⁵

Currently, the global acne treatment market has expanded steadily, driven by increasing healthcare demand and the emergence of new therapeutic strategies. At the same time, short-video platforms have become influential channels for health information dissemination,⁶ particularly among younger users.⁷ In China, platforms such as Bilibili and TikTok attract millions of daily active users,⁸ where dermatologic topics—including acne—are widely discussed.

However, concerns remain regarding the accuracy and credibility of health-related content on these platforms. Videos created by non-medical individuals often receive high engagement despite containing unverified or misleading claims,^9–11 highlighting the fragmented nature of online health communication. Although similar quality concerns and patterns of dermatologic misinformation have been reported across multiple social media platforms,^11–13 evidence specifically focusing on acne-related information on Chinese short-video platforms remains limited. Moreover, few studies have systematically assessed content quality using validated instruments in conjunction with audience engagement metrics.¹³

Given the growing role of short-form videos in public health communication, a clearer understanding of how video characteristics influence both informational quality and user engagement is urgently needed. This study systematically evaluates acne-related videos on Chinese short-video platforms (Bilibili and TikTok) according to uploader characteristics, disease-related topics, and presentation formats. Content quality is assessed using four validated tools: the Journal of the American Medical Association benchmark criteria (JAMA), a modified DISCERN instrument (mDISCERN), the Global Quality Scale (GQS), and the Video Information and Quality Index (VIQI), alongside a set of engagement metrics with correlation and regression analyses applied. The study aims to characterize the landscape of dermatologic information on Chinese short-video platforms, quantitatively evaluate the alignment and divergence between content quality and popularity, and provide evidence-based guidance for improving the accuracy and educational value of online health information.

Methods

Figure 1 provides an overview of the study workflow and analytical framework.

Figure 1.

Overview of the study workflow and analytical framework. JAMA: Journal of the American Medical Association benchmark criteria; mDISCERN: modified DISCERN instrument; GQS: Global Quality Scale; VIQI: Video Information and Quality Index.

Ethical considerations

This study was reviewed and approved by the Ethics Committee of the First Affiliated Hospital of Xi’an Jiaotong University (Approval No. KYLLSL-2026-076-01).

All data analyzed in this study were obtained from publicly accessible short-video platforms, Bilibili and TikTok, in compliance with the respective platforms’ terms of service. No personal identifiers (e.g., user IDs, usernames, or other sensitive information) were collected, stored, or disclosed in the manuscript or supplementary materials. In accordance with institutional review board guidance and the Committee on Publication Ethics (COPE) guidelines, individual informed consent was not required.

Search strategy

From May 1 to 5, 2025, systematic searches were conducted using the Chinese keywords “痤疮” (acne vulgaris), “粉刺” (comedones), and “青春痘” (pubertal acne) on Bilibili and TikTok. To minimize algorithm-driven personalization and potential selection bias, searches were performed using newly registered accounts with no prior browsing history or user interaction.

For each platform, the top 150 videos returned by the default search algorithm were collected, resulting in a total sample of 300 videos.^11,14 A ranking-based sampling strategy was adopted to enhance ecological validity, as the sampling frame consisted of content prioritized by platform search and recommendation systems. Although this approach is inherently influenced by algorithmic ranking mechanisms, it more closely reflects the acne-related information that users are likely to encounter during routine browsing.¹⁵ Videos uploaded within the 7 days prior to data collection were excluded to allow engagement metrics to stabilize and to reduce temporal variability associated with early-stage algorithmic promotion,^16,17 thereby improving the reliability of analyses examining the relationship between informational quality and user engagement.

Inclusion criteria were as follows: (1) Chinese-language content; and (2) a primary focus on acne-related health information. Exclusion criteria included: (1) Duplicate videos across platforms; (2) content that was primarily commercial or promotional; and (3) videos unrelated to acne or dermatologic health.

Data collection

For each eligible video, a standardized set of variables was extracted. Video characteristics included length (in seconds), duration (days from upload to data collection), and user engagement metrics: number of likes, comments, collections, and shares. Daily engagement metrics were calculated by dividing total engagement counts by the number of days from video upload to data collection. Uploader characteristics comprised verification status, and professional background. Uploaders were categorized as professionals—comprising Western medicine doctor, other medical workers or medical students, hospital, official media, and traditional Chinese medicine (TCM) practitioner—or non-professionals, including self-media and for-profit company. Each video was classified by content category, based on disease-related topics: Anatomy/physiology, causes/prevention, pathology/mechanism, epidemiology, symptoms, diagnosis/examination, treatment, prognosis, and TCM. Presentation format was documented independently, encompassing: Doctor monologue, patient video blog (video log (VLOG)), interactive question-and-answer (Q&A) session, PowerPoint (PPT) or lecture explanation, animation, television (TV) program/documentary, and other formats. Detailed definitions are provided in the Supplemental Methods.

Video quality assessment

Video quality was evaluated using four validated instruments that have been widely applied and demonstrated acceptable validity and reliability in assessing audiovisual health information on video-based social media platforms.^14,18,19

The JAMA benchmark criteria, originally developed for text-based medical information, assess transparency and reliability across four domains—authorship, attribution, disclosure, and currency—with a total score ranging from 0 to 4.²⁰ The mDISCERN tool, originally designed for written health information, evaluates the reliability of digital health content using five equally weighted items (total score: 0–5).^21,22 Together, these instruments assess core aspects of informational quality that are applicable regardless of content length or presentation format. The GQS, also rated on a 5-point Likert scale, offers a subjective overall assessment of video quality from a user-oriented perspective, considering clarity, educational value, and ease of understanding.^22–24 The VIQI measures four aspects—information flow, accuracy, production quality, and use of visual aids—using a 5-point Likert scale, providing insights into both audiovisual and presentation-related quality.^19,22 Detailed definitions and scoring criteria are provided in Tables S1–S4.

All videos were independently evaluated by two trained raters. Any discrepancies were resolved through discussion with a third investigator until consensus was reached. Inter-rater reliability was assessed using an overall weighted Cohen's kappa coefficient (κ), with agreement interpreted as almost perfect (κ > 0.80), substantial (κ = 0.60–0.80), or moderate (κ = 0.40–0.60).²⁵

Statistical analysis

Data distribution was assessed using the Shapiro–Wilk test. Continuous variables were reported as mean ± standard deviation (SD) or median (interquartile range, IQR), as appropriate. Group comparisons were conducted using the Mann–Whitney U test for continuous variables, and the Chi-square or Fisher's exact test for categorical variables. Bonferroni correction was applied for multiple pairwise comparisons.

Spearman rank correlation was used to explore associations between video quality scores and engagement metrics, with correlation strength categorized as negligible (<0.20), weak (0.20–0.40), moderate (0.40–0.60), strong (0.60–0.80), or very strong (>0.80).

To identify predictors of high-quality videos, defined as a GQS score ≥ 4,²⁶ univariable logistic regression models were applied to individual video attributes, including quality assessment scores and engagement metrics, followed by a multivariable model incorporating multiple predictors. Discriminative performance was evaluated using receiver operating characteristic (ROC) curve analyses with the area under the curve (AUC) and 95% confidence intervals (CIs).

All statistical analyses were performed using SPSS version 18.0 (IBM Corp.), R version 4.3.2, and GraphPad Prism version 9.0. A two-tailed P value < 0.05 was considered statistically significant.

Results

Characteristics of acne-related videos

A total of 272 acne-related videos were included in the final analysis, with 122 videos from Bilibili and 150 from TikTok. Significant differences in video characteristics were observed between the two platforms.

The length of videos on Bilibili was substantially longer than those on TikTok (median [IQR]: 409 [204–637] seconds vs. 51 [38–93] seconds; P < 0.001). Similarly, Bilibili videos had a markedly longer duration since upload (461 [160–966] days vs. 26 [9–117] days; P < 0.001).

Regarding uploader characteristics, the proportion of verified accounts was significantly lower on Bilibili than on TikTok (33.33% vs. 68.10%), and professional accounts contributed a smaller share of content on Bilibili compared to TikTok (25.64% vs. 65.52%) (Figure S1, Table S5). In terms of uploader identity, videos on both platforms were primarily produced by Western medicine doctors and self-media (Figure 2A). Among the 78 Bilibili uploaders, self-media accounted for the largest share (42.31%), whereas among the 116 TikTok uploaders, Western medicine doctors were most common (56.03%), followed by self-media (22.41%) and for-profit companies (11.21%) (Figure 2B, Table S5).

Figure 2.

Distribution of acne-related videos on Bilibili and TikTok. (A and B) Distribution by uploader types: Overall proportions across both platforms (A) and platform-specific comparison (B). (C and D) Distribution by disease-related topics: Overall proportions across both platforms (C) and platform-specific comparison (D). (E and F) Distribution by presentation formats: Overall proportions across both platforms (E) and platform-specific comparison (F). VLOG: video log; PPT: PowerPoint; Q&A: question and answer; TV: television.

With respect to disease-related topics, since a single video could address multiple categories, the most frequently covered themes were treatment and symptoms, while epidemiology and prognosis received less attention (Figure 2C). Anatomy/physiology was more often featured on Bilibili (7.38%), whereas diagnosis and examination content was more prevalent on TikTok (17.33%). Content related to TCM appeared in about 10% of videos on both platforms (Figure 2D, Table S6).

As for presentation formats, doctor monologues were the most common overall (Figure 2E). Additionally, patient VLOGs were more frequently observed on TikTok (22.00%) (Figure 2F, Table S7).

Quality and engagement across subgroups

Inter-rater agreement between the two reviewers was excellent (Cohen's κ = 0.889).

By platform

Overall, both platforms exhibited suboptimal video quality, but significant inter-platform differences were observed (Table 1). In terms of the JAMA benchmark criteria, TikTok videos achieved a significantly higher total score than Bilibili videos (1.28 ± 0.45 vs. 0.92 ± 0.98; P < 0.001). All TikTok videos met the authorship disclosure criterion (item 1: 100.00% vs. 39.34%; P < 0.001), whereas Bilibili videos demonstrated significantly higher currency (item 4: 14.75% vs. 1.33%; P < 0.001) (Figure 3A, Table 1). The average mDISCERN score was slightly higher on TikTok (2.10 ± 0.52 vs. 2.02 ± 0.93, P = 0.009), though absolute scores were low on both platforms. Item-level comparisons showed that TikTok videos were significantly more likely to clearly state their aims (item 1: 98.00% vs. 67.21%; P < 0.001), but less likely to cite information sources (item 2: 4.00% vs. 18.85%; P < 0.001). Almost all videos on both platforms fulfilled item 3 (balanced and unbiased content), while performance on items 4 and 5 remained generally poor, with no significant difference between platforms (Figure 3B, Table 1). The GQS, a single-item measure on a 5-point Likert scale, showed no significant difference between the platforms (3.06 ± 1.19 for Bilibili vs. 2.93 ± 0.75 for TikTok; P = 0.624) (Table 1). In contrast, Bilibili videos achieved a significantly higher total VIQI score than TikTok (12.05 ± 2.94 vs. 10.90 ± 1.97; P = 0.002). Specifically, Bilibili outperformed TikTok in flow of information (item 3: 1.97 ± 1.02 vs. 1.20 ± 0.69; P < 0.001) and accuracy (item 4: 3.61 ± 1.09 vs. 3.07 ± 0.55; P < 0.001), indicating better logical structure and content reliability (Figure 3C, Table 1). In summary, TikTok videos demonstrated greater transparency in authorship, whereas Bilibili videos were characterized by better structural coherence and informational accuracy.

Figure 3.

Item-level video quality scores and engagement metrics on Bilibili and TikTok. (A–C) Radar plots showing item-level scores for the JAMA (A), the mDISCERN (B), and the VIQI (C). (D–G) Violin plots with embedded boxplots showing the distributions of log-transformed engagement metrics [log10(x + 1)] for likes (D), comments (E), collections (F), and shares (G). JAMA: Journal of the American Medical Association benchmark criteria; mDISCERN: modified DISCERN instrument; VIQI: Video Information and Quality Index.

Table 1.

Quality assessment of acne-related videos on Bilibili and TikTok.

Variables	Total (mean ± SD or n)	Bilibili n (%) (N = 122)	TikTok n (%) (N = 150)	P-value
JAMA
Sum (mean ± SD)	1.12 ± 0.75	0.92 ± 0.98	1.28 ± 0.45	<0.001^a
Item 1 (authorship)	198	48 (39.34)	150 (100.00)	<0.001^b
Item 2 (attribution)	19	13 (10.66)	6 (4.00)	0.032^b
Item 3 (disclosure)	66	31 (25.41)	35 (23.33)	0.691^b
Item 4 (currency)	20	18 (14.75)	2 (1.33)	<0.001^b
mDISCERN
Sum (mean ± SD)	2.11 ± 0.73	2.02 ± 0.93	2.17 ± 0.51	0.009^a
Item 1	229	82 (67.21)	147 (98.00)	<0.001^b
Item 2	29	23 (18.85)	6 (4.00)	<0.001^b
Item 3	271	121 (99.18)	150 (100.00)	0.449^d
Item 4	31	13 (10.66)	18 (12.00)	0.729^b
Item 5	14	8 (6.56)	6 (4.00)	0.342^b
GQS (mean ± SD)	2.99 ± 0.98	3.06 ± 1.19	2.93 ± 0.75	0.624 ^a
VIQI
Sum (mean ± SD)	11.42 ± 2.52	12.05 ± 2.94	10.90 ± 1.97	0.002^a
Item 1	3.66 ± 1.10	3.58 ± 1.00	3.72 ± 1.18	0.150^a
Item 2	2.89 ± 0.76	2.86 ± 0.94	2.91 ± 0.58	0.089^a
Item 3	1.54 ± 0.93	1.97 ± 1.02	1.20 ± 0.69	<0.001^a
Item 4	3.31 ± 0.88	3.61 ± 1.09	3.07 ± 0.55	<0.001^a

JAMA: Journal of the American Medical Association benchmark criteria; mDISCERN: modified DISCERN instrument; GQS: Global Quality Scale; VIQI: Video Information and Quality Index; SD: standard deviation. Values are expressed as mean ± SD or n (%), as appropriate.

Mann–Whitney U test.

Chi-squared test.

Continuity correction.

Fisher's exact test.

Regarding engagement metrics, although the absolute numbers of likes, collections, and shares were generally higher on TikTok compared to Bilibili, only the difference in shares reached statistical significance (P = 0.002) (Figure 3D-G, Table S8). To account for differences in video exposure time, engagement metrics were normalized by video duration. Under this framework, TikTok videos demonstrated markedly higher daily engagement, with a mean of 79 likes, 3 comments, 26 favorites, and 21 shares per day, whereas Bilibili videos received only 2 likes, 0.21 comments, 1 favorite, and 0.27 shares per day (Table S8). These findings suggest that TikTok videos garnered substantially more user interaction per day, reflecting broader audience reach and heightened responsiveness.

By uploader characteristics

Videos from verified uploaders exhibited significantly higher quality scores across all metrics (JAMA, mDISCERN, GQS, and VIQI; all P < 0.001) (Figure S2 A–D, Table S9). These videos also showed greater engagement in terms of likes, favorites, and shares (all P < 0.001), except for comments (Figure S2 E–H, Table S10). Similarly, content from professional accounts scored higher in video quality assessments (all P < 0.001) (Figure S3 A–D, Table S9). For engagement metrics, non-professional uploaders received more likes (P < 0.001), whereas professionals had higher numbers of collections and shares (both P < 0.001) (Figure S3 E–H, Table S10). These findings suggest that verified and professional sources tend to contribute higher-quality content, but non-professionals may sometimes achieve greater popularity in specific forms of user engagement.

Regarding different types of uploaders, video quality varied notably across groups (Figure 4A-D, Table S9). Videos uploaded by other medical workers or medical students received the highest scores in JAMA (1.46 ± 0.78), GQS (3.69 ± 0.86), and VIQI (13.31 ± 1.70), followed by those from official media and Western medicine doctors. By contrast, videos uploaded by for-profit companies, self-media creators, and TCM practitioners tended to receive lower quality scores across most indicators. Significant differences were observed across uploader types in GQS (P < 0.001) and VIQI (P = 0.001), while those in JAMA and mDISCERN were not statistically significant (P = 0.280 and P = 0.063, respectively).

Figure 4.

Comparison of video quality and engagement metrics across uploader types. (A–D) Mean scores of four quality evaluation tools across uploader categories: Journal of the American Medical Association benchmark criteria (A), modified DISCERN instrument (B), Global Quality Scale (C), and Video Information and Quality Index (D). (E–H) Mean values of engagement metrics across uploader categories, including number of likes (E), comments (F), collections (G), and shares (H). A shared legend is shown in panels A and E for all subpanels in each row. Statistical significance: *P < 0.05, **P < 0.01, ***P < 0.001.

Regarding engagement metrics, videos uploaded by official media consistently received the highest number of likes, comments, collections, and shares, significantly outperforming other uploader types, particularly in terms of comments and shares (P < 0.01). Videos from self-media ranked second in engagement performance, while those from other sources showed relatively low levels across all indicators (Figure 4E–H, Table S10).

By disease-related topics

Among the different types of disease-related topics, videos focusing on anatomy/physiology received the highest quality scores, particularly in GQS (4.11 ± 0.78) and VIQI (14.78 ± 2.28), followed by those addressing epidemiology (mDISCERN: 3.00 ± 1.73; GQS: 4.00 ± 1.00) and pathology/mechanism (GQS: 3.51 ± 0.75; VIQI: 13.11 ± 2.27). In contrast, videos on prognosis exhibited the lowest scores across most indicators, especially GQS (2.50 ± 0.55) and VIQI (9.17 ± 2.79), indicating limited usefulness and presentation quality. Statistical analyses revealed significant differences in mDISCERN, GQS, and VIQI scores among topic categories (all P < 0.001), while the variation in JAMA was not statistically significant (P = 0.206) (Figure 5A–D, Table S11).

Figure 5.

Comparison of video quality and engagement metrics by disease-related topics. (A–D) Mean scores of four quality evaluation tools across topic categories: Journal of the American Medical Association benchmark criteria (A), modified DISCERN instrument (B), Global Quality Scale (C), and Video Information and Quality Index (D). (E–H) Mean values of engagement metrics across topic categories, including number of likes (E), comments (F), collections (G), and shares (H). A shared legend is shown in panels A and E for all subpanels in each row. Statistical significance: *P < 0.05, **P < 0.01, ***P < 0.001.

With respect to engagement, videos focusing on epidemiology, although not statistically significant, achieved the highest levels of engagement, especially in comments and shares. Other topic types generally showed lower engagement performance (Figure 5E–H, Table S12).

By presentation formats

Across different presentation formats, videos featuring doctor monologues demonstrated the highest overall quality, with strong scores in GQS (3.45 ± 0.86), and VIQI (12.46 ± 2.29). Both interactive Q&A and PPT or lecture explanations performed well in comparable mDISCERN scores (2.44 ± 0.88 and 2.47 ± 0.92, respectively) and favorable GQS ratings (3.11 ± 0.60 and 3.27 ± 0.59) though the latter showed a relatively low JAMA score (0.80 ± 0.78), suggesting potential limitations in credibility or authorship transparency. Videos presented in animation style achieved the highest VIQI score (12.67 ± 1.75), reflecting relatively strong production quality. In contrast, patient VLOGs showed the lowest scores across most quality indicators, particularly GQS (2.49 ± 0.64) and VIQI (9.97 ± 2.25), reflecting reduced educational value and presentation quality. Statistical analyses revealed significant differences in JAMA, GQS, and VIQI scores across different formats (all P < 0.001), while differences in mDISCERN scores were not statistically significant (P = 0.315) (Figure 6A–D, Table S13).

Figure 6.

Comparison of video quality and engagement metrics by presentation formats. (A–D) Mean scores of four quality evaluation tools across different presentation formats: Journal of the American Medical Association benchmark criteria (A), modified DISCERN instrument (B), Global Quality Scale (C), and Video Information and Quality Index (D). (E–H) Mean values of engagement metrics across presentation formats, including number of likes (E), comments (F), collections (G), and shares (H). A shared legend is shown in panels A and E for all subpanels in each row. Statistical significance: *P < 0.05, **P < 0.01, ***P < 0.001.

With respect to engagement, TV programs/documentaries showed the highest overall user interaction, particularly in shares (P < 0.01). Doctor monologues, interactive Q&A, and patient VLOGs also showed relatively higher engagement levels compared to other formats (Figure 6E–H, Table S14).

Correlation and predictive analysis

Spearman correlation analysis was conducted to explore relationships between video quality indicators and engagement metrics (Figure 7). Correlations among quality measures revealed a strong association between GQS and VIQI (ρ = 0.755), whereas JAMA showed only weak correlations with both VIQI and GQS (ρ = 0.309; ρ = 0.383, respectively). Very strong intercorrelations were observed among engagement metrics (ρ = 0.816–0.944, all P < 0.001), particularly between likes and collections (ρ = 0.944) and between likes and shares (ρ = 0.924). Among the quality indicators, VIQI showed the strongest correlations with engagement, particularly with collections (ρ = 0.717) and likes (ρ = 0.673). GQS also demonstrated strong correlations, especially with collections (ρ = 0.745). In contrast, JAMA was only weakly correlated with engagement metrics (ρ = 0.209–0.257), suggesting limited alignment between source transparency and user behavior.

Figure 7.

Correlations among video quality scores and engagement metrics. Spearman correlation heatmap showing the associations between engagement metrics (likes, comments, collections, and shares) and video quality assessment scores, including the JAMA, mDISCERN, GQS, and VIQI. Color intensity represents the strength of correlation, with numerical values indicating Spearman correlation coefficients. Asterisks denote statistical significance. JAMA: Journal of the American Medical Association benchmark criteria; mDISCERN: modified DISCERN instrument; GQS: Global Quality Scale; VIQI: Video Information and Quality Index. Statistical significance: *P < 0.05, **P < 0.01, ***P < 0.001.

To further assess the predictive validity, ROC curve analyses was performed using a GQS threshold of ≥ 4 to define high-quality videos. Among quality metrics, VIQI demonstrated the strongest discriminative ability in univariable analyses (AUC = 0.922, 95% CI: 0.892–0.952), while in multivariable analysis, a logistic regression model incorporating all three quality scores further improved performance (AUC = 0.936, 95% CI: 0.908–0.963) (Figure 8A). For engagement metrics, collections had the strongest predictive value (AUC = 0.901, 95% CI: 0.863–0.940), whereas the combined engagement model yielded an AUC of 0.879 (95% CI: 0.836–0.921) (Figure 8B). These results indicate that structured quality assessments provide higher discriminative precision, whereas user interaction patterns can only moderately reflect content quality.

Figure 8.

Predictive performance of video quality and engagement metrics for identifying high-quality videos. Receiver operating characteristic curves evaluating the discriminative performance of video quality assessment tools (JAMA, mDISCERN, and VIQI) (A) and engagement metrics (likes, comments, collections, and shares) (B) for identifying high-quality videos (defined as GQS ≥ 4) based on univariable and multivariable logistic regression models. JAMA: Journal of the American Medical Association benchmark criteria; mDISCERN: modified DISCERN instrument; VIQI: Video Information and Quality Index; AUC: area under the curve; GQS: Global Quality Scale.

Discussion

Acne vulgaris is a highly prevalent dermatologic condition with substantial psychosocial impact, making its management a subject of widespread public concern. In recent years, short-video platforms have emerged as influential tools for health communication, especially among younger populations. In this context, this study systematically evaluated acne-related videos on two major Chinese platforms, Bilibili and TikTok, by integrating four validated quality assessment tools (JAMA, mDISCERN, GQS, and VIQI) with four engagement metrics (likes, comments, collections, and shares).

A total of 272 videos were analyzed. Overall, video quality on both platforms was suboptimal. However, the observed inter-platform differences reflected complementary strengths rather than consistent superiority. Bilibili videos were generally longer, less interactive, and demonstrated slightly better structural coherence and informational accuracy, as indicated by higher VIQI scores. In contrast, TikTok videos were shorter, attracted significantly greater daily engagement, and were more frequently uploaded by verified and professional accounts, with modestly higher transparency, as indicated by higher JAMA and mDISCERN scores. Beyond platform-level differences, uploader characteristics, disease-related topics, and presentation formats were also associated with variation in quality and engagement. Videos from other medical workers or medical students achieved the highest quality scores, while those uploaded by official media attracted the greatest engagement. Content related to anatomy/physiology received the highest quality ratings, whereas epidemiology-themed videos elicited the most interaction. Among presentation styles, doctor monologues demonstrated superior quality, while TV programs/documentaries generated the highest levels of engagement. Correlation analysis demonstrated a strong association between GQS and VIQI, while engagement metrics were highly intercorrelated. Among them, collections showed notable correlations with both GQS and VIQI. Predictive modeling confirmed that structured quality assessments outperformed engagement metrics in identifying high-quality content. Together, these findings provide an empirical basis for interpreting the observed patterns of quality and engagement in acne-related short-video content.

Platform-specific differences

Given the overall suboptimal quality and cross-platform differences, item-level findings offered detailed insights. TikTok videos universally fulfilled the JAMA authorship criterion (item 1), compared to less than half on Bilibili—aligning with previous research that underscores TikTok's stricter credentialing and emphasis on source transparency.²⁷ However, TikTok videos scored poorly on JAMA attribution and disclosure (items 2 and 3), underscoring that authorship transparency does not necessarily translate into reliable referencing or conflict-of-interest disclosure. This pattern is consistent with previous observations that social media platforms often prioritize identity verification over citation practices.^10,28 TikTok also outperformed Bilibili on mDISCERN item 1, suggesting more explicitly framed content. In contrast, Bilibili scored higher on VIQI domains such as information flow (item 3) and accuracy (item 4), reflecting stronger structural coherence and content reliability.

These patterns may be partly explained by differences in platform design and algorithmic priorities.²⁹ Bilibili's longer video duration and greater content persistence may facilitate more systematic, knowledge-oriented communication and long-term accessibility of educational materials. The platform is structured around a professional user-generated video model, with interest-based communities, durable content archives, and participatory features such as real-time “danmu” bullet comments, which are commonly associated with sustained attention and communal knowledge sharing.³⁰ In contrast, TikTok's short-form, algorithm-driven interface emphasizes immediacy and user responsiveness. Its “For You” feed ranks videos based on user engagement signals—such as likes, shares, watch time, and replays—combined with video content and user profile information, thereby promoting the rapid diffusion of highly engaging content.³¹ This structural logic may favor concise and visually stimulating formats, which can pose challenges for delivering in-depth and comprehensive medical information. Consistent with this, previous studies have noted TikTok's advantage in visibility and engagement^32,33 but relatively lower citation completeness and content comprehensiveness.^34,35

From a public health perspective, these observations highlight a fundamental trade-off. TikTok enables rapid dissemination and high audience engagement, supported by greater transparency of authorship, whereas Bilibili offers more coherent and accurate educational content but struggles to achieve comparable reach. These complementary characteristics suggest that platform-specific communication strategies may be warranted, such as concise, high-impact content for TikTok, and comprehensive, structured resources for Bilibili—to improve both accuracy and accessibility in dermatologic communication.

Influence of uploader characteristics

TikTok featured a higher proportion of verified and professionally affiliated contributors, reflecting its stricter credentialing system, which typically requires physicians to hold senior clinical titles or affiliations with professional societies. In contrast, Bilibili applies looser verification standards and promotes creative diversity, resulting in a broader range of contributors, particularly self-media and non-clinical accounts.²³ This structural contrast between a credential-based model and a community-driven framework may partly explain the observed differences in video quality and user engagement across the two platforms.

Quantitative analysis revealed a significant association between uploader credentials and video quality. Across all four evaluation tools (JAMA, mDISCERN, GQS, and VIQI), videos uploaded by verified or professional contributors consistently outperformed those from uncertified or non-professional sources, aligning with recent findings in the field of digital health communication.^32,34,36

Engagement metrics, however, revealed a more nuanced pattern. Non-professional creators, particularly self-media, tended to receive more immediate interaction in the form of likes and comments.^34,37 In contrast, professional and verified accounts performed better in terms of collections and shares, suggesting that their content may be perceived as more credible, valuable, or worth revisiting and redistributing,¹⁰ although direct evidence linking professional credentials to specific engagement behaviors remains limited.³⁴ This reflects a familiar tension between popularity and informational reliability: Emotionally appealing or entertaining content may garner more attention, even when its scientific rigor is limited.^10,37,38 Notably, institutional accounts, such as official media, stood out as a distinctive subgroup. Although representing a small share of the dataset, their videos consistently achieved the highest engagement across all metrics, indicating that institutional authority remains a strong signal of trustworthiness and audience engagement.³⁹

In summary, platform governance plays a pivotal role in shaping both the quality and dissemination of medical information. TikTok could promote professionally verified content through algorithmic adjustments, while Bilibili might improve overall quality by strengthening credentialing requirements for health-related uploads.

Impact of disease-related topics and presentation formats

As for disease-related topics, videos covering anatomy and physiology received the highest quality scores, likely due to their structured content and suitability for expert-led explanations. In contrast, videos on prognosis tended to score lowest across most dimensions. Although overall engagement was similar across topics, content focusing on epidemiology showed relatively higher interaction, suggesting audience interest in broader contextual understanding of acne.

Presentation styles significantly influenced video performance.^32,40 Doctor monologues, PPT or lecture explanations, and animations consistently demonstrated higher quality, while TV programs/documentaries—though less frequent—achieved the strongest engagement levels, reflecting the communicative strength of institutionally produced, narrative-driven formats.⁴¹

Taken together, these findings suggest that professionally led content and formats are associated with higher informational quality, while diverse narrative and documentary styles are linked to greater audience engagement. Rather than representing competing approaches, these complementary strengths highlight the value of integrating professional expertise with engaging presentation formats to balance credibility and communication impact.

TCM-related content accounted for approximately 10% of acne-related videos across both platforms, whereas TCM uploaders made up only about 3% of all contributors. In TCM theory, acne is typically attributed to internal imbalances, such as “overheating of the lung or stomach” or “blood stasis,” and is treated through herbal medicine, acupuncture, or dietary regulation.⁴² Despite its cultural relevance and long-standing presence in dermatologic care, TCM videos consistently received lower quality scores, particularly in the JAMA and mDISCERN domains. This may reflect a predominance of uncertified contributors and limited adherence to standardized, evidence-based communication frameworks. Similar shortcomings have been reported in TCM-related content in other medical fields, particularly regarding source transparency and educational rigor.^43,44 To improve the reliability of TCM content on digital platforms, efforts should be made to encourage participation from certified practitioners and to promote collaboration with professional health communicators.

Association between content quality and user engagement

Correlation and predictive analyses revealed a nuanced relationship between content quality and user engagement. Among quality measures, VIQI and GQS showed strong internal correlation, indicating that presentation quality and perceived usefulness often co-occur. Both also demonstrated strong associations with engagement metrics, particularly collections and likes, suggesting that users may be more responsive to visually coherent and educationally valuable content. In contrast, JAMA and mDISCERN, which emphasize source attribution and reliability, exhibited only weak correlations with both other quality tools and engagement indicators, highlighting a disconnect between source transparency and audience attention. Multivariable logistic regression confirmed that structured quality metrics more accurately identified high-quality videos compared to engagement indicators. However, the strong univariable predictive value of collections suggests that certain forms of user behavior may still reflect perceived trustworthiness and informational value. These findings underscore a persistent challenge in short-video health communication: popular content is not always reliable, and reliable content is not always popular.⁴⁵ To bridge this gap, platforms may need to move beyond engagement-based algorithms and adopt quality-aware strategies, such as expert labeling, credential verification, or algorithmic weighting based on validated tools to ensure that dermatologic information is not only broadly disseminated, but also accurate and trustworthy.

The divergence between popularity and informational rigor warrants particular attention in adolescent acne communication, as adolescents are both heavily affected by acne^46,47 and constitute a major user group of short-video platforms.⁴⁸ In our study, self-media videos, typically demonstrating moderate rather than high informational quality, were observed to achieve disproportionately high engagement, suggesting that in adolescent-oriented digital environments, medically rigorous information should not be disadvantaged in terms of accessibility or dissemination competitiveness.

This study has several strengths. First, it provides a structured, cross-platform evaluation of acne-related videos on two widely used Chinese short-video platforms, offering timely insight into dermatologic health communication within a fast-evolving media environment. Second, by incorporating uploader characteristics, disease-related topics, and presentation formats into the analytical framework, the combined use of four validated quality tools (JAMA, mDISCERN, GQS, VIQI) and four user engagement metrics enabled a comprehensive analysis linking informational quality with audience response. All videos were independently rated by trained dermatology physicians, with high inter-rater agreement, and analytic robustness was ensured through correlation analysis, ROC modeling, and multivariable prediction. Third, algorithmic and sampling bias were minimized through predefined keyword queries, newly registered accounts, and standardized inclusion criteria. Collectively, these strengths provide a reproducible methodological framework for future investigations and contribute to the development of platform-aware strategies to improve the quality and impact of digital dermatologic content.

Several limitations should also be acknowledged. First, the cross-sectional design limits causal inference and reflects only a snapshot of content within a rapidly evolving media landscape. Second, by focusing on Chinese-language content and top-ranked videos, the study may not fully represent lower-ranked or non-Chinese content, limiting generalizability across linguistic and regional contexts. Third, although four validated instruments were employed, some (e.g., JAMA, mDISCERN) were originally developed for static, text-based health information and primarily focus on informational transparency and reliability. As a result, certain platform-specific audiovisual and narrative features of short-form videos may not be fully captured. In addition, some categories, including TCM uploaders and epidemiology-related videos, were represented by relatively small sample sizes, which may limit the robustness of subgroup comparisons. Accordingly, related findings should be interpreted with caution. Despite excellent inter-rater reliability, subjective scoring and potential misclassification of uploader identity cannot be fully excluded. Additionally, key engagement indicators such as view counts, dislikes, and watch duration were not accessible due to platform restrictions, which may have constrained the comprehensiveness of user interaction analysis. Lastly, algorithmic amplification may skew engagement metrics, complicating the interpretation of user preferences.⁴⁹ These limitations underscore the need for future longitudinal, multi-platform studies incorporating user behavior data and algorithmic transparency to better understand how dermatologic content is encountered, interpreted, and trusted.

Conclusion

This study demonstrated that the overall quality of acne-related videos on both Bilibili and TikTok was suboptimal, with each platform exhibiting distinct strengths. Bilibili videos showed slightly greater structural coherence and informational accuracy, whereas TikTok videos were more frequently uploaded by verified professionals, displayed higher transparency, and achieved substantially higher levels of daily engagement. Videos from professional contributors received higher quality scores, while content from official media and self-media accounts garnered more user interaction. Both topic selection and presentation format significantly influenced video quality and engagement metrics. Although user engagement showed only moderate correlation with content quality, predictive modeling confirmed the value of structured quality assessments and collections among engagement metrics in identifying high-quality content. These findings underscore the need for platform-specific strategies that account for differences in content structure, dissemination mechanisms, and user engagement patterns, in order to enhance the visibility and reliability of dermatologic information.

Supplemental Material

sj-docx-1-dhj-10.1177_20552076261430231 - Supplemental material for Quality, engagement, and predictive validity of acne-related short videos on Chinese platforms Bilibili and TikTok: A cross-sectional content analysis

Supplemental material, sj-docx-1-dhj-10.1177_20552076261430231 for Quality, engagement, and predictive validity of acne-related short videos on Chinese platforms Bilibili and TikTok: A cross-sectional content analysis by Yuhan Xie, Qinxiao Li, Wenmin Deng, Yuxin Yan, Longmei Duan, Yuting Chen, Yusheng Wan, Kainian Han, Heni Ma and Yan Zheng in DIGITAL HEALTH

Supplemental Material

sj-docx-2-dhj-10.1177_20552076261430231 - Supplemental material for Quality, engagement, and predictive validity of acne-related short videos on Chinese platforms Bilibili and TikTok: A cross-sectional content analysis

Supplemental material, sj-docx-2-dhj-10.1177_20552076261430231 for Quality, engagement, and predictive validity of acne-related short videos on Chinese platforms Bilibili and TikTok: A cross-sectional content analysis by Yuhan Xie, Qinxiao Li, Wenmin Deng, Yuxin Yan, Longmei Duan, Yuting Chen, Yusheng Wan, Kainian Han, Heni Ma and Yan Zheng in DIGITAL HEALTH

Footnotes

Acknowledgements

The authors thank the video uploaders for their contributions to public health.

ORCID iD

Yuhan Xie

Ethics approval

This study was reviewed and approved by the Ethics Committee of the First Affiliated Hospital of Xi’an Jiaotong University (Approval No. KYLLSL-2026-076-01).

Consent to participate

All data were obtained from publicly accessible short-video platforms (Bilibili and TikTok) in compliance with the respective platforms’ terms of service. No personal identifiers or private user information were collected or reported; therefore, individual informed consent was waived in accordance with Institutional Review Board Guidance and COPE guidelines.

Authors’ contributions

Yuhan Xie and Qinxiao Li conceived the study, contributed to study design and data interpretation, and drafted the manuscript. Wenmin Deng and Yuxin Yan collected and organized the data. Longmei Duan and Yuting Chen performed statistical analyses. Yusheng Wan and Kainian Han contributed to data validation and manuscript revision. Heni Ma assisted with manuscript preparation and formatting. Yan Zheng supervised the study, provided critical revisions, and served as the corresponding author. All authors read and approved the final manuscript.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Availability of data and materials

The datasets generated and analyzed during the current study are available from the corresponding author upon reasonable request.

Supplemental material

Supplemental material for this article is available online.

References

Eichenfield

Sprague

Eichenfield

. Management of acne vulgaris: a review. JAMA 2021; 326: 2055–2067.

Shields

Nock

, et al. Evaluation of stigma toward individuals with acne. JAMA Dermatol 2024; 160: 93–98.

Saurat

J-H

Halioua

Baissac

, et al. Epidemiology of acne and rosacea: a worldwide global study. J Am Acad Dermatol 2024; 90: 1016–1018.

Abdulbari

Ali

Abu-Raghif

, et al. Impact of azithromycin on specific biochemical markers and sebum composition in acne vulgaris patients. Arch Dermatol Res 2025; 317: 798.

Ertekin

Salici

Manav Bas

, et al. Influence of social media and internet on treatment decisions in adult female acne patients: a cross-sectional survey study. Dermatol Pract Concept 2024; 14: e2024156.

Zhu

Zhang

, et al. How health communication via TikTok makes a difference: a content analysis of TikTok accounts run by Chinese provincial health committees. Int J Environ Res Public Health 2019; 17: 192.

Roberts

David

. Technology affordances, social media engagement, and social media addiction: an investigation of TikTok, Instagram Reels, and YouTube Shorts. Cyberpsychol Behav Soc Netw 2025; 28: 318–325.

Gao

. What drives me to use TikTok: a latent profile analysis of users’ motives. Front Psychol 2022; 13: 992824.

Chou

W-YS

Gaysynsky

Cappella

. Where we go from here: health misinformation on social media. Am J Public Health 2020; 110: S273–S275.

10.

Kington

Arnesen

Chou

W-YS

, et al. Identifying credible sources of health information in social media: principles and attributes. NAM Perspect 2021: 10.31478/202107a.

11.

Mueller

Hongler

VNS

Jungo

, et al. Fiction, falsehoods, and few facts: cross-sectional study on the content-related quality of atopic eczema–related videos on YouTube. J Med Internet Res 2020; 22: e15599.

12.

Iyengar

Saldanha

Chong

. #Acne: a thematic qualitative analysis of acne content on TikTok. Australas J Dermatol 2025; 66: 127–134.

13.

Zheng

Ning

Levoska

, et al. Acne and social media: a cross-sectional study of content quality on TikTok. Pediatr Dermatol 2021; 38: 336–338.

14.

HO-Y

Bailey

Huynh

, et al.

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

BMJ Glob Health 2020; 5: e002604.

15.

Morahan-Martin

. How internet users find, evaluate, and use online health information: a cross-cultural review. Cyberpsychol Behav 2004; 7: 497–510.

16.

Liang

Yang

, et al. Quality and reliability of prostate cancer videos on TikTok and Bilibili: cross-sectional content analysis study. Digit Health 2025; 11: 20552076251376263.

17.

Yao

, et al. Short video platforms as sources of health information about HPV vaccine: a content and quality analysis. Digit Health 2025; 11: 20552076251379340.

18.

Ferhatoglu

Kartal

Ekici

, et al. Evaluation of the reliability, utility, and quality of the information in sleeve gastrectomy videos shared on open access video sharing platform YouTube. Obes Surg 2019; 29: 1477–1484.

19.

Nagpal

SJS

Karimianpour

Mukhija

, et al. YouTube videos as a source of medical information during the Ebola hemorrhagic fever epidemic. Springerplus 2015; 4: 457.

20.

Silberg

Lundberg

Musacchio

. Assessing, controlling, and assuring the quality of medical information on the Internet: caveant lector et viewor—Let the reader and viewer beware. JAMA 1997; 277: 1244–1245.

21.

Charnock

Shepperd

Needham

, et al. DISCERN: an instrument for judging the quality of written consumer health information on treatment choices. J Epidemiol Community Health 1999; 53: 105–111.

22.

Albayrak

Büyükçavuş

. Does YouTube offer high-quality information? Evaluation of patient experience videos after orthognathic surgery. Angle Orthod 2023; 93: 409–416.

23.

Bernard

Langille

Hughes

, et al. A systematic review of patient inflammatory bowel disease information resources on the World Wide Web. Am J Gastroenterol 2007; 102: 2070–2077.

24.

Yeung

AWK

. Evaluation of content quality of online health information by Global Quality Score: a case study of researchers misnaming it and citing secondary sources. Publications (Basel) 2025; 13: 23.

25.

Tang

Zhang

, et al. Kappa coefficient: a popular measure of rater agreement. Shanghai Arch Psychiatry 2015; 27: 62–67.

26.

Niu

Hao

Yang

, et al. Quality of pancreatic neuroendocrine tumor videos available on TikTok and Bilibili: content analysis. JMIR Form Res 2024; 8: e60033.

27.

Liu

Chen

Lin

, et al. YouTube/Bilibili/TikTok videos as sources of medical information on laryngeal carcinoma: cross-sectional content analysis study. BMC Public Health 2024; 24: 1594.

28.

Geels

Graßl

Schraffenberger

, et al. Virtual lab coats: the effects of verified source information on social media post credibility. PLOS One 2024; 19: e0302323.

29.

Klinger

Svensson

. The end of media logics? On algorithms and agency. New Media Soc 2018; 20: 4653–4670.

30.

Bilibili Inc. Bilibili environmental, social and governance report 2024. Shanghai: Bilibili Inc., 2024, https://ir.bilibili.com/media/gxtmixwi/bilibili-2024environmental-social-and-governance-report.pdf (accessed 14 September 2025).

31.

TikTok. How TikTok recommends content. TikTok Help Center, https://support.tiktok.com/en/using-tiktok/exploring-videos/how-tiktok-recommends-content (accessed 14 September 2025).

32.

Guan

J-L

Xia

S-H

Zhao

, et al. Videos in short-video sharing platforms as sources of information on colorectal polyps: cross-sectional content analysis study. J Med Internet Res 2024; 26: e51655.

33.

Wang

Yao

Wang

, et al. Bilibili, TikTok, and YouTube as sources of information on gastric cancer: assessment and analysis of the content and quality. BMC Public Health 2024; 24: 57.

34.

Ren

, et al. The quality and reliability of short videos about premature ovarian failure on Bilibili and TikTok: cross-sectional study. Digit Health 2025; 11: 20552076251351077.

35.

Cui

Cao

, et al. Quality assessment of TikTok as a source of information about mitral valve regurgitation in China: cross-sectional study. J Med Internet Res 2024; 26: e55403.

36.

Peng

, et al. Comparative analysis of NAFLD-related health videos on TikTok: a cross-language study in the USA and China. BMC Public Health 2024; 24: 3375.

37.

Loeb

Sengupta

Butaney

, et al. Dissemination of misinformative and biased information about prostate cancer on YouTube. Eur Urol 2019; 75: 564–567.

38.

Vraga

Bode

. Defining misinformation and understanding its bounded nature: using expertise and evidence for describing misinformation. Polit Commun 2020; 37: 136–144.

39.

Zhang

Mei

Song

, et al.

Why do citizens engage with the TikTok accounts of public hospitals in China?

SAGE Open 2021; 11: 21582440211061568.

40.

Zheng

Jin

, et al. Quality, reliability, and dissemination of lung cancer information on short-video platforms in China: a cross-platform content analysis of TikTok, Kwai, and Rednote. Front Public Health 2025; 13: 1683561.

41.

Xiao

Sun

Lei

, et al. How do features in health short-form videos impact viewer engagement: an empirical study. Ind Manag Data Syst 2025: 1–32.

42.

Cao

Yang

Wang

, et al. Complementary therapies for acne vulgaris. Cochrane Database Syst Rev 2015: CD009436.

43.

Woon

Chia

Kwan

, et al. Evaluation of the quality of YouTube videos on traditional Chinese medicine and inflammatory arthritis. Eur J Integr Med 2021; 47: 101380.

44.

Liu

Jin

Hsu

, et al. Comparing smartphone apps for traditional Chinese medicine and modern medicine in China: systematic search and content analysis. JMIR MHealth UHealth 2021; 9: e27406.

45.

Chou

W-YS

Klein

WMP

. Addressing health-related misinformation on social media. JAMA 2018; 320: 2417–2418.

46.

Layton

Ravenscroft

. Adolescent acne vulgaris: current and emerging treatments. Lancet Child Adolesc Health 2023; 7: 136–144.

47.

Conte

Iorio

Esposito

, et al. Scrolling through adolescence: a systematic review of the impact of TikTok on adolescent mental health. Eur Child Adolesc Psychiatry 2025; 34: 1511–1527.

48.

Pew Research Center. Teens, social media and technology 2024. Washington, DC: Pew Research Center, 2024, https://www.pewresearch.org/internet/2024/12/12/teens-social-media-and-technology-2024/ (accessed 17 September 2025).

49.

Yuan

Shi

, et al. Resistance or compliance? The impact of algorithmic awareness on people’s attitudes toward online information browsing. Front Psychol 2025; 16: 1563592.

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.

0.00 MB

0.51 MB

0.60 MB