Skip to main content
NCBI home page
Medicine logoLink to Medicine
. 2026 Sep 25;105(39):e50773. doi: 10.1097/MD.0000000000050773

Evaluating the quality and reliability of rotator cuff injury related videos on Bilibili and TikTok

A cross-sectional analysis

Junyu Bai a, Jinqiao Zhang b,*
PMCID: PMC13619200  PMID: 42798091

Abstract

Rotator cuff injury (RCI) is a common age-related musculoskeletal disorder, yet public awareness remains limited. With short-video platforms emerging as key health information sources, concerns about content quality have grown. This study evaluated the quality, reliability, and comprehensiveness of RCI-related videos on TikTok and Bilibili and examined the impact of uploader characteristics. On October 13, 2025, Chinese-language videos containing “rotator cuff injury” were retrieved from TikTok and Bilibili. Following strict inclusion criteria, 231 videos were analyzed. Video characteristics, engagement metrics (likes, comments, shares, collections), and uploader information were collected. Quality and content integrity were assessed using the Video Information and Quality Index (VIQI), Global Quality Scale (GQS), and modified DISCERN (mDISCERN). Statistical comparisons and Spearman correlation analyses were performed. Of 231 videos, 96 were from Bilibili and 135 from TikTok. TikTok videos were shorter (median 93 seconds vs 234.5 seconds, P < .001) but demonstrated higher engagement across all metrics (P < .001). Physicians contributed most content (TikTok: 79%, Bilibili: 64%). Diagnosis and treatment dominated, while epidemiology and prevention were rarely addressed. GQS and mDISCERN scores did not differ between platforms. TikTok scored higher on VIQI information flow (P < .001) and visual quality (P = .044); Bilibili excelled in information accuracy (P = .006). Individual user videos showed slightly higher GQS (P < .05) and accuracy (P = .033); physician videos received more comments (P = .002) and shares (P = .042). VIQI and GQS correlated positively with engagement metrics (VIQI: R = 0.21–0.40; GQS: r = 0.20–0.28; all P < .001), whereas mDISCERN showed weak associations. Video duration correlated positively with quality scores (GQS: r = 0.56; VIQI: R = 0.34; mDISCERN: r = 0.33; all P < .001) but negligibly with engagement. The analysis of RCI-related videos on TikTok and Bilibili reveals moderate content quality and a significant lack of preventive information. These findings underscore the need for strategies to enhance the scientific rigor and comprehensiveness of health information on short-video platforms.

Keywords: Bilibili, health communication, rotator cuff injury, short video, TikTok

1. Introduction

The rotator cuff is a complex of shoulder muscles, consisting of the supraspinatus, infraspinatus, subscapularis, and teres minor, and their associated tendons, which attach to the humeral head to form a dynamic stabilizing apparatus that maintains shoulder joint stability during various movements.[1,2] Beyond providing stability, the rotator cuff enables fine motor control of the shoulder through distinct muscular subdivisions.[3] Rotator cuff injury (RCI) represents one of the most prevalent shoulder disorders.[4,5] Its pathogenesis is multifactorial, involving oxidative stress and chronic inflammatory responses that may lead to tendon degeneration and tearing.[6] RCI not only restricts patients’ range of motion, but also adversely affects multiple aspects of daily life, diminishing overall quality of life.[7,8]

With the advancement of digital technology, social media platforms have emerged as important channels for public access to health knowledge.[9] TikTok leverages its high user engagement to effectively capture public attention toward health information,[10] while Bilibili shapes viewers’ perceptions and attitudes toward health topics through diverse creator perspectives – for instance, by fostering nuanced discussions that have contributed to vaccine uptake.[11] Orthopedic patients demonstrate substantial demand for educational materials and frequently turn to the internet for information.[12] Given the specialized knowledge required for RCI diagnosis, treatment, and rehabilitation, patients urgently need accessible and practical health guidance.[13] However, health communication via short-video platforms faces significant challenges: specialized content often becomes marginalized by algorithms due to its less engaging format, whereas entertaining content is preferentially recommended. This disconnect between high-quality science communication and user engagement may facilitate the spread of low-quality videos, potentially exposing viewers to health risks.[14-16] As short videos become ubiquitous in health science communication, identifying reliable information sources and optimizing dissemination strategies has emerged as a pressing concern.

Given that RCI is a common musculoskeletal condition, the quality of short-video science communication on this topic directly influences patients’ health decisions. Yet, systematic evaluations of content quality differences across uploader types and the impact of platform characteristics on viewer engagement remain lacking. Therefore, this cross-sectional study analyzed RCI-related short videos on TikTok and Bilibili, employing the modified DISCERN (mDISCERN), Global Quality Scale (GQS), and Video Information and Quality Index (VIQI) instruments to assess content quality and information reliability. Our findings aim to inform evidence-based strategies for optimizing health communication in this domain and to contribute to the standardization of high-quality health information dissemination on short-video platforms.

2. Materials and methods

2.1. Search strategy and data extraction

This cross-sectional study was conducted on October 13, 2025, using desktop computers, searching for videos with the keyword “rotator cuff injury” (肩袖损伤) on the Chinese versions of TikTok and Bilibili platforms. To minimize the bias introduced by personalized recommendation algorithms, new user accounts were created for each platform. Searches were performed using the default sorting method (no restriction on publication date), which presents results based on a comprehensive algorithm incorporating video completion rate (percentage of viewers watching more than 5 seconds), number of likes, comments, followers, and upload time, thereby prioritizing recent and popular content. A total of 300 videos were initially identified.

The following videos were excluded from the analysis: videos uploaded within 7 days prior to data collection (to ensure data stability); videos irrelevant to RCI, or those with superficial content, defined as videos that mentioned RCI only briefly (e.g., in passing within the first or last few seconds) without providing any substantive information on its epidemiology, etiology, clinical manifestations, diagnosis, treatment, or prevention; videos primarily for advertising or promotional purposes; and duplicate or re-uploaded videos that had already been analyzed. After applying these criteria, 231 videos were included for analysis (Fig. 1).

Figure 1.

Figure 1.

Open in a new tab

Flow chart of subject selection.

Basic information for each included video – such as title, uploader details, duration, content type, user engagement metrics, and publication date – was extracted using the platforms’ public APIs (application programming interface) to ensure data accuracy and consistency. All data were systematically recorded in Microsoft Excel for subsequent analysis.

2.2. Uploader characteristics

Based on the professional background of the video creators, the sources of the videos were classified into 2 categories: individual users (nonprofessional Physicians) and licensed physicians. Uploaders were considered “verified” if their account displayed an official verification badge and their profile explicitly stated their medical credentials or institutional affiliation. Detailed information was collected for each uploader, including their identity, number of followers, verification status, and type.

2.3. Ethical considerations

This study only included publicly available short videos on rotator cuff injuries from the Bilibili and TikTok platforms. As no personal or sensitive data were involved, ethical review was not required.

2.4. Assessment of video quality and reliability

The completeness of video content was evaluated across 6 predefined domains: epidemiology, etiology or mechanisms, clinical manifestations, diagnostic methods, treatments, and preventive measures. Each domain was scored using a binary system: 1 point if the video adequately covered the domain, and 0 points if the domain was not mentioned or only briefly mentioned without substantive information. All target videos were collected and downloaded by a designated researcher (BJY). To avoid potential bias related to scoring order, all videos to be evaluated were randomly reordered before assessment. Two reviewers with relevant medical backgrounds independently evaluated the videos and received standardized training before scoring to ensure consistent assessment criteria and minimize subjective bias. Any discrepancies in scoring were resolved by a third arbitrator (WS), and final consensus was achieved among all authors. The inter-rater agreement between the 2 reviewers was quantified using Cohen’s κ coefficient, with reference to the criteria proposed by Landis and Koch: κ > 0.8 indicating excellent agreement, 0.6 to 0.8 substantial agreement, 0.4 to 0.6 moderate agreement, and κ < 0.4 poor agreement.[17]

Video quality was assessed using the Video Information and Quality Index (VIQI), Global Quality Scale (GQS), and the modified DISCERN (mDISCERN) instrument.[18-20] The VIQI tool covers 4 dimensions: information flow (VIQI 1), accuracy of information (VIQI 2), quality (including scores for images, animation, interviews, captions, and summaries; VIQI 3), and precision (level of consistency between the video title and content; VIQI 4). Each criterion is rated on a 1 to 5 scale, with higher scores indicating better quality. The overall information quality of videos was evaluated using the Global Quality Scale (GQS), which consists of 5 items scored from 1 to 5, where higher scores represent better quality. The reliability of video content was assessed using the mDISCERN instrument, with higher scores indicating greater reliability.

2.5. Statistical analysis

Descriptive statistics were used to summarize video characteristics. Non-normally distributed variables were presented as median with interquartile range (IQR). The Mann–Whitney U test was employed to compare video quality scores (including GQS, mDISCERN, and VIQI) between different uploader groups and between different platform groups. Spearman correlation coefficient was used to analyze the correlations between engagement metrics (likes, collections, comments, shares), video duration, and quality scores. All statistical analyses were performed using R software (version 4.3.2; R Foundation for Statistical Computing), with a P-value < .05 considered statistically significant.

3. Results

3.1. Video characteristics and uploader types

Among the 231 videos analyzed, 96 (42%) were from Bilibili and 135 (58%) were from TikTok. The median duration of videos on Bilibili was significantly longer than that on TikTok (234.50 vs 93.00 seconds, P < .001). However, the median number of likes (158.00 vs 3411.00, P < .001), collections (243.00 vs 2443.00, P < .001), comments (1.00 vs 184.00, P < .001), and shares (86.50 vs 958.00, P < .001) were all significantly lower for videos on Bilibili compared to those on TikTok (Table 1).

Table 1.

Comparison of characteristics and user engagement of rotator cuff injury-related videos on short-video platforms.

Variables Total (n = 231) Bilibili (n = 96) TikTok (n = 135) Statistic P
Video length, M (Q1, Q3) 134.00 (74.50, 248.00) 234.50 (139.25, 495.50) 93.00 (58.50, 157.50) Z = −7.47 <.001
Like, M (Q1, Q3) 1052.00 (176.00, 4359.50) 158.00 (35.25, 491.75) 3411.00 (1071.00, 6851.50) Z = −9.91 <.001
Collections, M (Q1, Q3) 1099.00 (216.00, 3552.00) 243.00 (65.25, 1094.00) 2443.00 (699.50, 6018.00) Z = −7.51 <.001
Comment, M (Q1, Q3) 53.00 (3.00, 296.00) 1.00 (0.00, 9.50) 184.00 (60.00, 517.50) Z = −10.62 <.001
Share, M (Q1, Q3) 341.00 (81.00, 1886.50) 86.50 (16.75, 240.75) 958.00 (310.00, 3098.50) Z = −8.62 <.001
Open in a new tab

Bold P values indicate statistical significance.

Z: Mann–Whitney test.

M: median, Q1: first quartile, Q3: third quartile.

Among the 231 videos analyzed, 90 (39.0%) were uploaded by individual users, while 141 (61.0%) were uploaded by physicians. On Bilibili, 64% of the videos were created by physicians and 36% by individual users. On TikTok, the proportion of physician uploaders was even higher, accounting for 79%, while individual users contributed 21%. Overall, the majority of videos on both platforms were uploaded by physicians, with the proportion of physician-generated content higher than that of individual users (Fig. 2).

Figure 2.

Figure 2.

Open in a new tab

Percentage stacked bar chart illustrating the proportion of individual users and licensed physicians on Bilibili and TikTok platforms

Videos uploaded by individual users were significantly longer in duration (median: 216.50 seconds) compared to those uploaded by physicians (median: 112.00 seconds; P < .001). The physician group received a significantly higher number of comments (median: 81.00 vs 13.50; P = .002) and shares (median: 605.00 vs 212.00; P = .042) than the individual user group. However, no statistically significant differences were observed between the 2 groups in terms of the number of likes (P = .098) or collections (P = .675; Table 2).

Table 2.

Comparison of video metrics between individual and licensed physicians.

Variables Total (n = 231) Individual users (nonprofessional physicians; n = 90) Licensed physicians (n = 141) Statistic P
Video length, M (Q1, Q3) 134.00
(74.50, 248.00)		 216.50
(90.75, 504.50)		 112.00
(66.00, 179.00)		 Z = −4.91 <.001
Like, M (Q1, Q3) 1052.00
(176.00, 4359.50)		 529.00
(148.25, 4154.00)		 1441.00
(309.00, 4400.00)		 Z = −1.66 .098
Collections, M (Q1, Q3) 1099.00
(216.00, 3552.00)		 1009.50
(224.50, 2834.00)		 1263.00
(205.00, 3841.00)		 Z = −0.42 .675
Comment, M (Q1, Q3) 53.00
(3.00, 296.00)		 13.50
(1.00, 137.00)		 81.00
(10.00, 341.00)		 Z = −3.16 .002
Share, M (Q1, Q3) 341.00
(81.00, 1886.50)		 212.00
(85.25, 666.75)		 605.00
(79.00, 2279.00)		 Z = −2.04 .042
Open in a new tab

Z: Mann–Whitney test.

M: median, Q1: first quartile, Q3: third quartile.

3.2. Content analysis

A total of 231 videos were included in this study, with 96 assigned to the Bilibili group and 135 to the TikTok group. Differences in the distribution of content categories were observed between the 2 platforms. In the epidemiology domain, both groups had a median score of 0.00, but the difference was statistically significant (P < .001). For the clinical manifestations domain, the median score was 0.00 in the Bilibili group and 1.00 in the TikTok group, with a statistically significant difference observed (P = .039). However, no significant differences were found between the 2 groups in terms of etiology, diagnosis, treatment, or prevention domains (all P > .05; Table 3).

Table 3.

Comparison of medical content completeness between Bilibili and TikTok platforms.

Variables Total (n = 231) Bilibili (n = 96) TikTok (n = 135) Statistic P
Epidemiology, M (Q1, Q3) 0.00 (0.00, 0.00) 0.00 (0.00, 0.00) 0.00 (0.00, 0.00) Z = −3.81 <.001
Etiology, M (Q1, Q3) 0.00 (0.00, 1.00) 0.00 (0.00, 1.00) 0.00 (0.00, 1.00) Z = −0.65 .515
Clinical manifestations, M (Q1, Q3) 0.00 (0.00, 1.00) 0.00 (0.00, 1.00) 1.00 (0.00, 1.00) Z = −2.07 .039
Diagnosis, M (Q1, Q3) 1.00 (0.00, 1.00) 1.00 (0.00, 1.00) 1.00 (0.00, 1.00) Z = −0.21 .836
Treatment, M (Q1, Q3) 1.00 (0.00, 1.00) 1.00 (0.00, 1.00) 1.00 (0.00, 1.00) Z = −0.07 .944
Prevention, M (Q1, Q3) 0.00 (0.00, 0.00) 0.00 (0.00, 0.00) 0.00 (0.00, 0.00) Z = −1.20 .228
Open in a new tab

Z: Mann–Whitney test.

M: median, Q1: first quartile, Q3: third quartile.

Within each platform, both Bilibili and TikTok primarily featured content related to “treatment” and “diagnosis,” with similar and relatively high proportions. The proportion of videos covering “etiology” was also comparable between the 2 platforms (17% for Bilibili and 16% for TikTok). Notably, TikTok had a higher proportion of content related to “clinical manifestations” (26%) compared to Bilibili (19%). In contrast, Bilibili included 6% of videos on “epidemiology” and 3% on “prevention,” while TikTok had no content addressing “epidemiology” and only 1% of videos focused on “prevention.” (Fig. 3).

Figure 3.

Figure 3.

Open in a new tab

Bar chart comparison of medical content completeness dimensions between TikTok and Bilibili platforms

Among the 231 videos, 39.0% (90 videos) were categorized as individual users, and 61.0% (141 videos) as licensed physicians. There were notable differences in the distribution of content scores between the 2 groups. In the epidemiology domain, both groups had a median score of 0.00, but the difference was statistically significant (P = .004). In the diagnosis domain, the physician group had a median score of 1.00, while the individual user group had a median score of 0.00, with a significant difference between groups (P = .022). No statistically significant differences were observed between the 2 groups in the domains of etiology, clinical manifestations, treatment, or prevention (all P > .05; Table 4).

Table 4.

Comparison of medical content completeness dimensions between individual users (nonprofessional physicians) and licensed physicians.

Variables Total (n = 231) Individual users (nonprofessional physicians; n = 90) Licensed physicians
(n = 141)		 Statistic P
Epidemiology, M (Q1, Q3) 0.00 (0.00, 0.00) 0.00 (0.00, 0.00) 0.00 (0.00, 0.00) Z = −2.88 .004
Etiology, M (Q1, Q3) 0.00 (0.00, 1.00) 0.00 (0.00, 1.00) 0.00 (0.00, 1.00) Z = −0.12 .902
Clinical manifestations, M (Q1, Q3) 0.00 (0.00, 1.00) 0.00 (0.00, 1.00) 1.00 (0.00, 1.00) Z = −1.84 .066
Diagnosis, M (Q1, Q3) 1.00 (0.00, 1.00) 0.00 (0.00, 1.00) 1.00 (0.00, 1.00) Z = −2.29 .022
Treatment, M (Q1, Q3) 1.00 (0.00, 1.00) 1.00 (0.00, 1.00) 1.00 (0.00, 1.00) Z = −1.53 .125
Prevention, M (Q1, Q3) 0.00 (0.00, 0.00) 0.00 (0.00, 0.00) 0.00 (0.00, 0.00) Z = −1.39 .165
Open in a new tab

Z: Mann–Whitney test.

M: median, Q1: first quartile, Q3: third quartile.

Within the physician group, content related to diagnosis accounted for the largest proportion (29%), followed by treatment (27%) and clinical manifestations (25%), both representing relatively high proportions. The proportions for etiology, prevention, and epidemiology were 16%, 1%, and 1%, respectively, with prevention and epidemiology being extremely low. In the individual user group, treatment-related content was most prevalent (33%), followed by diagnosis (22%), clinical manifestations (20%), and etiology (16%). The proportions for epidemiology and prevention were 5% and 3%, respectively, both higher than those observed in the physician group (Fig. 4).

Figure 4.

Figure 4.

Open in a new tab

Grouped bar chart comparison of medical content dimensions between individual users (nonprofessional physicians) and licensed physicians

Overall, physician-uploaded videos focused more on professional content such as diagnosis, treatment, and clinical manifestations, while content related to prevention and epidemiology was rarely included. In contrast, individual users contributed a higher proportion of videos on epidemiology and prevention, though treatment remained their primary focus.

3.3. Comparison of GQS, mDISCERN, and VIQI scores between Bilibili and TikTok

To investigate differences in GQS, mDISCERN, and VIQI scores between TikTok and Bilibili, data distributions were visualized using violin plots combined with box plots. For TikTok, GQS scores were primarily concentrated in the range of 2 to 4, with a median of approximately 3. Similarly, GQS scores on Bilibili were also centered within the 2 to 4 range, with an additional minor density peak observed between 5 and 6; the median was likewise around 3. Statistical analysis indicated that there was no significant difference in GQS scores between the 2 platforms (Fig. 5).

Figure 5.

Figure 5.

Open in a new tab

Violin plot of GQS distributions between TikTok and Bilibili platforms. GQS = Global Quality Scale, ns = not significant (statistically non-significant).

For TikTok (in red), mDISCERN scores were mainly concentrated in the range of 2 to 3, with a median of approximately 2.5. Likewise, the distribution of mDISCERN scores for Bilibili (in blue) was also centered around the 2 to 3 range, with some data points extending to both lower and higher values; notably, the median was also about 2.5. Statistical analysis indicated that there was no significant difference in mDISCERN scores between the 2 platforms (Fig. 6).

Figure 6.

Figure 6.

Open in a new tab

Violin plot of mDISCERN distributions between TikTok and Bilibili platforms. mDISCERN = modified DISCERN, “ns” = not significant (statistically non-significant).

For TikTok (represented in red), the median total VIQI score was approximately 12.5, primarily concentrated in the range of 10 to 15, with some data points extending to both 0 and 20. For Bilibili (represented in blue), the median total VIQI score was about 12, also centered in the 10 to 15 range. Statistical analysis indicated a significant difference in overall VIQI scores between the 2 platforms (Fig. 7).

Figure 7.

Figure 7.

Open in a new tab

Violin plot of VIQI distributions between TikTok and Bilibili platform. The asterisk (*) indicates a statistically significant difference between the 2 groups (P < .05). VIQI = Video Information and Quality Index.

3.4. VIQI sub-score comparison

Regarding the VIQI scores, the TikTok group had a significantly higher median total VIQI score (13.0, IQR 12.0–14.0) compared to the Bilibili group (12.0, IQR 10.0–14.0; P = .021). For the information flow sub-score (VIQI 1), the TikTok group had a significantly higher median (4.0, IQR 4.0–4.0) compared to the Bilibili group (3.0, IQR 2.0–3.0; P < .001). Regarding information accuracy (VIQI 2), the Bilibili group demonstrated a higher median score (4.0, IQR 3.0–4.0) than the TikTok group (3.0, IQR 3.0–4.0; P = .006). The visual quality scores (VIQI 3) differed significantly between the TikTok and Bilibili groups (P = .044), with median scores of 2.0 (IQR 1.0–2.0) in both groups. No statistically significant difference was observed in content precision score (VIQI 4) between the 2 groups (P = .067), with Bilibili having a median of 4.0 (IQR 3.0–5.0) and TikTok 4.0 (IQR 3.0–4.0; Table 5).

Table 5.

VIQI dimensions: comparisons between Bilibili and TikTok platforms.

Variables Total (n = 231) Bilibili (n = 96) TikTok (n = 135) Statistic P
VIQI sum, M (Q1, Q3) 12.00 (11.00, 14.00) 12.00 (10.00, 14.00) 13.00 (12.00, 14.00) Z = −2.31 .021
VIQI 1, M (Q1, Q3) 4.00 (3.00, 4.00) 3.00 (2.00, 3.00) 4.00 (4.00, 4.00) Z = −9.52 <.001
VIQI 2, M (Q1, Q3) 3.00 (3.00, 4.00) 4.00 (3.00, 4.00) 3.00 (3.00, 4.00) Z = −2.75 .006
VIQI 3, M (Q1, Q3) 2.00 (1.00, 2.00) 2.00 (1.00, 2.00) 2.00 (1.00, 2.00) Z = −2.01 .044
VIQI 4, M (Q1, Q3) 4.00 (3.00, 4.00) 4.00 (3.00, 5.00) 4.00 (3.00, 4.00) Z = −1.83 .067
Open in a new tab

Z: Mann–Whitney test; M: median, Q1: first quartile, Q3: third quartile.

VIQI 1: Information flow; VIQI 2: Accuracy of information; VIQI 3: Quality of production; VIQI 4: Precision.

VIQI = Video Information and Quality Index.

3.5. Comparison of GQS, mDISCERN, and VIQI scores by uploader group

This violin plot compares the distribution of Global Quality Scores (GQS) – a key metric for evaluating video smoothness, informational completeness, and practical value to the audience – between the physician group and the individual user group. The kernel density distribution shows that the individual user group has a significantly higher density in the high scoring range (1–2 points) than the physician group, and the median GQS for the individual user group is closer to the higher end of the vertical axis. A statistically significant difference in GQS scores was found between the 2 groups (P < .05; Fig. 8).

Figure 8.

Figure 8.

Open in a new tab

Violin plot of GQS distributions between individual users (nonprofessional physicians) and licensed physician uploaders. The asterisk (*) indicates a statistically significant difference between the 2 groups (P < .05). GQS = Global Quality Scale.

This violin plot compares the distribution of mDISCERN scores – a core metric assessing the quality of information discernment in videos – between physician users and individual users. The kernel density distributions of the 2 groups show substantial overlap, with the median (indicated by the horizontal line within each violin) and interquartile ranges (the core density regions) not showing any marked shift between groups. There was no statistically significant difference in mDISCERN scores between the 2 groups (Fig. 9).

Figure 9.

Figure 9.

Open in a new tab

Violin plot of mDISCERN distributions between individual users (nonprofessional physicians) and licensed physician uploaders. mDISCERN = modified DISCERN, ns = not significant (statistically non-significant).

This violin plot illustrates the distribution of total Video Information and Quality Index (VIQI) scores for videos uploaded by physician users and individual users. The kernel density contours of the 2 groups exhibit considerable overlap, with no distinct separation in the medians or interquartile ranges. Statistical analysis indicated no significant difference in total VIQI scores between the 2 groups (Fig. 10).

Figure 10.

Figure 10.

Open in a new tab

Violin plot of VIQI sum distributions between individual users (nonprofessional physicians) and licensed physician uploaders. ns = not significant (statistically non-significant), VIQI = Video Information and Quality Index.

Among the 231 participants, 39.0% (90) were classified as individual users, and 61.0% (141) as licensed physicians. The results of group comparisons for each scoring index are as follows: Information flow score (VIQI 1): Individual users had a median of 3.00 (IQR 3.00–4.00), while licensed physicians had a median of 4.00 (IQR 3.00–4.00), with no statistically significant difference (P = .067). Information accuracy score (VIQI 2): Individual users had a median of 4.00 (IQR 3.00–4.00), which was higher than that of licensed physicians (median 3.00, IQR 3.00–4.00), and this difference was statistically significant (P = .033). Visual quality score (VIQI 3): Both groups had a median of 2.00 (IQR 1.00–2.00), with no statistically significant difference (P = .690). Content precision score (VIQI 4): Individual users had a median of 4.00 (IQR 4.00–4.75), while licensed physicians had a median of 4.00 (IQR 3.00–4.00), with a statistically significant difference (P = .036; Table 6).

Table 6.

Descriptive statistics and group comparisons of VIQI between individual users (nonprofessional physicians) and licensed physicians.

Variables Total (n = 231) Individual users (nonprofessional physicians; n = 90) Licensed physicians (n = 141) Statistic P
VIQI sum, M (Q1, Q3) 12.00 (11.00, 14.00) 13.00 (11.00, 14.75) 12.00 (11.00, 14.00) Z = −1.09 .277
VIQI 1, M (Q1, Q3) 4.00 (3.00, 4.00) 3.00 (3.00, 4.00) 4.00 (3.00, 4.00) Z = −1.83 .067
VIQI 2, M (Q1, Q3) 3.00 (3.00, 4.00) 4.00 (3.00, 4.00) 3.00 (3.00, 4.00) Z = −2.13 .033
VIQI 3, M (Q1, Q3) 2.00 (1.00, 2.00) 2.00 (1.00, 2.00) 2.00 (1.00, 2.00) Z = −0.40 .690
VIQI 4, M (Q1, Q3) 4.00 (3.00, 4.00) 4.00 (4.00, 4.75) 4.00 (3.00, 4.00) Z = −2.09 .036
Open in a new tab

Z: Mann–Whitney test; M: median, Q1: first quartile, Q3: third quartile.

VIQI 1: Information flow; VIQI 2: Accuracy of information; VIQI 3: Quality of production (images, animation, etc); VIQI 4: Precision (title-content consistency).

VIQI = Video Information and Quality Index.

3.6. Correlation heatmap analysis

Spearman correlation analysis was conducted to examine the relationships among various video metrics, GQS, mDISCERN, and VIQI scores in RCI-related videos. The results are as follows: Total VIQI scores were moderately positively correlated with all engagement metrics (r = 0.21–0.40, all P < .001), with the strongest correlation observed for collections (r = 0.40). GQS exhibited weak to moderate positive correlations with engagement metrics (r = 0.20–0.28, all P < .001), most strongly with collections (r = 0.28), and showed no significant correlation with shares. mDISCERN demonstrated generally weak correlations with engagement indicators, displaying only modest positive associations with likes (r = 0.14, P < .05) and comments (r = 0.16, P < .05).

Video duration was significantly positively correlated with quality scores (GQS: r = 0.56; VIQI: r = 0.34; mDISCERN: r = 0.33; all P < .001) but showed negligible correlations with engagement metrics, except for a weak positive association with collections (r = 0.13, P < .05). Moderate positive correlations were observed among all quality assessment tools (r = 0.30–0.47, all P < .001), with the strongest correlation between GQS and total VIQI scores (r = 0.47; Fig. 11).

Figure 11.

Figure 11.

Open in a new tab

Correlation heatmap of content quality indicators, user interaction behaviors, and video length in social media health information. Statistical significance is denoted as follows: *P < .05, **P < .01, ***P < .001, ****P < .001. GQS = Global Quality Scale, mDISCERN = modified DISCERN, VIQI = Video Information and Quality Index.

4. Discussion

This study revealed that TikTok videos achieved higher engagement despite shorter duration, while Bilibili excelled in information accuracy. Physicians contributed most content, yet preventive information was scarce. Video quality showed weak-to-moderate correlations with engagement. Notably, video duration correlated positively with quality but negligibly with engagement, highlighting a decoupling between popularity and informational value on short-video platforms.

4.1. Platform attributes and user engagement

This study found significant differences between Bilibili and TikTok in terms of user engagement features and content presentation. Videos on TikTok were notably shorter in length but achieved significantly higher numbers of likes, comments, collections, and shares compared to Bilibili. This finding is consistent with previous studies, which reported that TikTok videos are generally shorter than those on Bilibili while displaying higher user engagement, a difference closely related to the platforms’ inherent characteristics.[21,22] Furthermore, shorter videos are more likely to stimulate continuous viewing, leading to increased user interactions and broader dissemination.[23] In contrast, Bilibili is characterized by longer videos and relatively lower interaction frequencies; its orientation toward “knowledge-based content” causes users to prioritize content depth over instant engagement.[24]

Among the specific interaction indices for the 2 platforms, “likes” and “collections” were the main user engagement forms, with TikTok users showing a significantly higher tendency to share compared to Bilibili. This may be associated with different sharing function designs and audience behaviors.[25-27] This divergence not only highlights platform-specific patterns in user engagement but also underscores critical design and operational considerations for social media platforms aiming to optimize user experience and participation.[28,29]

4.2. Uploader type and content structure

On both platforms, the proportion of videos uploaded by physicians exceeded that of individual users; this phenomenon was more pronounced on TikTok, reflecting a strong initiative among medical professionals for health communication via new media. This aligns with Kong et al,[30] who reported that Chinese health professionals have actively promoted diabetes-related knowledge on TikTok. Moreover, TikTok’s identity verification mechanism ensures that more physicians can participate in health information dissemination as verified professionals, thereby enhancing viewers’ trust in the information.[31]

The current results demonstrate that videos posted by physician users had higher numbers of comments and shares and shorter durations, suggesting that physicians tend to deliver concise summaries and highlight key messages. Notably, videos from individual users received higher GQS scores than those from physicians, possibly because content from individuals is more relatable and presented in a manner that better fits audience preferences. These differences may also reflect distinct content production goals and audience psychological needs.[32] Mueller et al[33] found that patients are more inclined to adopt advice from other patients rather than physicians, a conclusion not entirely consistent with some studies,[34] indicating the need for further investigation into the influence of cultural context and platform attributes on health communication.

4.3. Content distribution and quality assessment

Most videos on both platforms focused on “diagnosis” and “treatment”; TikTok videos more broadly covered “clinical manifestations,” whereas Bilibili had more content on “epidemiology” and “prevention.” Further analysis by uploader type revealed that the overall proportions of preventive and epidemiological content were generally low across both groups, suggesting insufficient attention to preventive health education among the public and content creators. This finding is consistent with Hu Yaqin et al,[35] who reported that over half of the thyroid-related videos on TikTok emphasized diagnosis and symptoms rather than prevention. Given that the incidence of RCI increases with age,[36] the lack of preventive content may hinder public risk awareness and early intervention, potentially increasing treatment challenges.

In terms of quality, GQS and mDISCERN scores did not differ significantly between platforms and were generally mediocre; however, VIQI scores were slightly higher for TikTok, mainly due to superior information flow and visual quality. Zheng et al[37] argued that high-quality health videos should combine scientific accuracy, accessibility, and clarity without misleading information, yet both platforms exhibited shortcomings in informational rigor. Additionally, Sun et al[38] suggested that credible videos do not necessarily gain more attention, with TikTok’s algorithm favoring entertaining content, thus explaining why overall quality scores remain moderate.

4.4. Correlation among variables

The present study found that total VIQI scores were moderately positively correlated with user engagement metrics, whereas GQS showed weak-to-moderate correlations, and mDISCERN exhibited very weak associations. These findings confirm that high engagement does not necessarily equate to high information quality,[39] suggesting that user behavior and algorithmic recommendations are more closely tied to sensory stimulation or entertainment value than to informational rigor. Li et al[40] reported positive correlations among likes, comments, shares, and collections, while Chen et al[41] explicitly noted that high-quality videos do not always attract greater user engagement, partially explaining the decoupling between popularity and information quality.

Notably, video duration was positively correlated with content quality, indicating that longer videos tend to provide more complete and accurate information.[42,43] Contrary to previous findings,[44] the present study observed no significant association between video duration and user engagement behaviors such as likes, comments, or shares. This suggests that while producing high-quality health science content requires sufficient time to ensure informational completeness,[45] merely lengthening videos does not guarantee viewer completion or interaction. Therefore, health science creators face the challenge of balancing content comprehensiveness against user attention. A practical approach involves breaking complex topics into short, serialized videos, with each episode focusing on a single knowledge point.[42] This strategy preserves scientific rigor while adapting to users’ fragmented viewing habits.[46]

4.5. Implications for patients and public health

Our findings have several important implications for patients and public health practice. First, the weak correlation between user engagement and content quality suggests that patients should exercise caution when relying on popularity metrics such as likes or shares, as high viewership does not guarantee accurate information.[47]

Second, the notable lack of preventive and epidemiological content across both platforms is concerning. Given that rotator cuff injuries are common, particularly among older adults, the scarcity of preventive information may limit public awareness of risk factors and early symptoms, potentially delaying diagnosis and treatment.[48,49] Thus, public health campaigns should prioritize creating and disseminating high-quality preventive content on these platforms.

Third, the complementary strengths of the 2 platforms offer opportunities for tailored health communication strategies. TikTok’s high engagement could be leveraged for brief, impactful awareness messages, while Bilibili’s strength in information accuracy makes it suitable for more detailed educational content.[50,51] Collaboration between healthcare institutions and platform providers could help promote evidence-based content and improve the overall quality of health information available to the public.[52,53]

4.6. Study limitations and future perspectives

Several limitations should be noted. First, this study focused solely on RCI videos, limiting generalizability across health information domains and platforms. Second, video scoring remains partly subjective, although standardized protocols and multiple raters were employed. Third, the analyses revealed associations rather than causal relationships; future research should incorporate experimental or longitudinal approaches for more robust inference. Future studies should leverage emerging techniques such as AI-based content analysis to enhance objectivity while expanding to broader disease areas, additional platforms, and countries.[54,55] Furthermore, efforts to assess and reveal the mechanisms behind users’ health literacy promotion remain warranted.

5. Conclusions

This study found that the overall quality of rotator cuff injury–related short videos on TikTok and Bilibili was moderate, with most content focusing on diagnosis and treatment, while preventive information was limited. TikTok videos had higher user engagement and better visual quality, but information accuracy was higher on Bilibili. Physician-uploaded videos generated more interaction, whereas individual users’ videos had slightly higher quality and accuracy. Importantly, user engagement was only weakly related to content quality, highlighting the need to improve both scientific rigor and coverage of prevention in health communication on short-video platforms.

Acknowledgments

The authors would like to express their gratitude to the video uploaders for their contributions to public health.

Author contributions

Data curation: Junyu Bai.

Investigation: Jinqiao Zhang.

Methodology: Jinqiao Zhang.

Software: Junyu Bai.

Supervision: Junyu Bai.

Visualization: Junyu Bai.

Writing – original draft: Junyu Bai.

Writing – review & editing: Jinqiao Zhang.

Abbreviations:

Bilibili
Chinese video sharing platform
CI
confidence interval
GQS
Global Quality Scale
IQR
interquartile range
mDISCERN
modified DISCERN
RCI
Rotator cuff injury
SD
standard deviation
TikTok (Douyin)
Chinese short video platform
VIQI
Video Information and Quality Index.

The data used in this study were sourced from publicly available video content published on platforms such as Bilibili and TikTok. These videos are publicly accessible, and no personal privacy information was involved during the data collection process. All analyzed content was publicly available, and the study did not involve the collection or processing of users’ private information. In accordance with relevant ethical review guidelines, ethical approval for this study was not required.

The authors have no funding and conflicts of interest to declare.

The datasets generated during and/or analyzed during the current study are publicly available.

How to cite this article: Bai J, Zhang J. Evaluating the quality and reliability of rotator cuff injury related videos on Bilibili and TikTok: A cross-sectional analysis. Medicine 2026;105:39(e50773).

JB and JZ contributed to this article equally.

References

  • [1].Imhoff AB. [Rotator cuff]. Oper Orthop Traumatol. 2012;24:457. [DOI] [PubMed] [Google Scholar]
  • [2].Day A, Taylor NF, Green RA. The stabilizing role of the rotator cuff at the shoulder--responses to external perturbations. Clin Biomech (Bristol). 2012;27:551–6. [DOI] [PubMed] [Google Scholar]
  • [3].Cavanaugh E, Santillan A A, Hoshikawa K, Giambini H. Subregions of the rotator cuff muscles present distinct anatomy, biomechanics, and function. Sports (Basel). 2024;12:349. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [4].Tashjian RZ. Epidemiology, natural history, and indications for treatment of rotator cuff tears. Clin Sports Med. 2012;31:589–604. [DOI] [PubMed] [Google Scholar]
  • [5].Weber S, Chahal J. Case studies AAOS clinical practice guideline: management of rotator cuff injuries. J Am Acad Orthop Surg. 2021;29:e104–8. [DOI] [PubMed] [Google Scholar]
  • [6].Yuan T, Qian H, Yu X, et al. Proteomic analysis reveals rotator cuff injury caused by oxidative stress. Ther Adv Chronic Dis. 2021;12:2040622320987057. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [7].Manoso-Hernando D, Bailón-Cerezo J, Angulo-Díaz-Parreño S, Reina-Varona A, Elizagaray-García I, Gil-Martínez A. Shoulder mobility and strength impairments in patients with rotator cuff related shoulder pain: a systematic review and meta analysis. PeerJ. 2024;12:e17604. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [8].Baumer TG, Dischler J, Mende V, et al. Effects of asymptomatic rotator cuff pathology on in vivo shoulder motion and clinical outcomes. J Shoulder Elbow Surg. 2017;26:1064–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [9].Ghalavand H, Panahi S, Sedghi S. Opportunities and challenges of social media for health knowledge management: a narrative review. J Educ Health Promot. 2020;9:144. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [10].Wang S, Jia S, Su Y, Cheng R. Videos on YouTube, Bilibili, TikTok as sources of medical information on Hashimoto’s thyroiditis. Front Public Health. 2025;13:1611087. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [11].Gao H, Yin H, Peng L, Wang H. Effectiveness of social video platforms in promoting COVID-19 vaccination among youth: a content-specific analysis of COVID-19 vaccination topic videos on Bilibili. Risk Manag Healthc Policy. 2022;15:1621–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [12].Burrus MT, Werner BC, Starman JS, et al. Patient perceptions and current trends in internet use by orthopedic outpatients. HSS J. 2017;13:271–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [13].Miskiewicz M, Capotosto S, Ling K, Hance F, Wang E. Readability analysis of patient education material on rotator cuff injuries from the top 25 ranking orthopaedic institutions. J Am Acad Orthop Surg Glob Res Rev. 2024;8:e24.00085. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [14].Fernandez-Llatas C, Traver V, Borras-Morell JE, Martinez-Millana A, Karlsen R. Are health videos from hospitals, health organizations, and active users available to health consumers? An analysis of diabetes health video ranking in YouTube. Comput Math Methods Med. 2017;2017:8194940. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [15].Wang C, Wang H, Yang Z, Ding S. Divergent routes of health infotainment in changing public health attitudes: a GPT-2 analysis of users’ responses to health infotainment. J Health Commun. 2025;30:40–50. [DOI] [PubMed] [Google Scholar]
  • [16].Pothugunta K, Liu X, Susarla A, Padman R. Assessing inclusion and representativeness on digital platforms for health education: evidence from YouTube. J Biomed Inform. 2024;157:104669. [DOI] [PubMed] [Google Scholar]
  • [17].Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977;33:159–74. [PubMed] [Google Scholar]
  • [18].Singh AG, Singh S, Singh PP. YouTube for information on rheumatoid arthritis--a wakeup call? J Rheumatol. 2012;39:899–903. [DOI] [PubMed] [Google Scholar]
  • [19].Bernard A, Langille M, Hughes S, Rose C, Leddin D, Veldhuyzen van Zanten S. A systematic review of patient inflammatory bowel disease information resources on the World Wide Web. Am J Gastroenterol. 2007;102:2070–7. [DOI] [PubMed] [Google Scholar]
  • [20].Charnock D, Shepperd S, Needham G, Gann R. DISCERN: an instrument for judging the quality of written consumer health information on treatment choices. J Epidemiol Community Health. 1999;53:105–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [21].Hang YY, Hsu HQ, Lu M, Bai L, Qian J. The quality and reliability of short videos about uremia on Bilibili and TikTok in China: a cross sectional study. Medicine (Baltim). 2026;105:e47543. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [22].Qiu X, Zhou Y, Yuan T, et al. A cross-sectional analysis of the quality and characteristics of sleep apnea hypopnea syndrome videos on YouTube, Bilibili, and TikTok. Sci Rep. 2026;16:4070. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [23].Zheng C. Research on the flow experience and social influences of users of short online videos. A case study of DouYin. Sci Rep. 2023;13:3312. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [24].Han J, Li X, Li M, Wu X. Assessing the scientific quality of radiotherapy popularization videos on BiliBili and TikTok: a cross-sectional study. Front Public Health. 2026;14:1668189. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [25].Habibi SA, Salim L. Static vs. dynamic methods of delivery for science communication: a critical analysis of user engagement with science on social media. PLoS One. 2021;16:e0248507. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [26].Lau N, Srinakarin K, Aalfs H, Zhao X, Palermo TM. TikTok and teen mental health: an analysis of user-generated content and engagement. J Pediatr Psychol. 2025;50:63–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [27].Mosleh M, Pennycook G, Rand DG. Self-reported willingness to share political news articles in online surveys correlates with actual sharing on Twitter. PLoS One. 2020;15:e0228882. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [28].Wang L, Chen Y, Zhao D, Xu T, Hua F. Quality and dissemination of uterine fibroid health information on TikTok and Bilibili: cross-sectional study. JMIR Form Res. 2025;9:e75120. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [29].Peng G, Wang C, Zhang HW, Xu T, Di JZ. Evaluating the reliability and quality of bariatric surgery educational content on TikTok and Bilibili: a cross-sectional content analysis. Obes Surg. 2025;35:5111–9. [DOI] [PubMed] [Google Scholar]
  • [30].Kong W, Song S, Zhao YC, Zhu Q, Sha L. TikTok as a health information source: assessment of the quality of information in diabetes-related videos. J Med Internet Res. 2021;23:e30409. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [31].Bravery BD, Stojkov J, Brown J. Supporting effective doctor-patient communication: doctors’ name badges. Med J Aust. 2020;213:500–1.e1. [DOI] [PubMed] [Google Scholar]
  • [32].Zhu W, He B, Wang X, Du Y, Young K, Jiang S. Information quality of videos related to esophageal cancer on tiktok, kwai, and bilibili: a cross-sectional study. BMC Public Health. 2025;25:2245. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [33].Mueller SM, Hongler VNS, Jungo P, et al. Fiction, falsehoods, and few facts: cross-sectional study on the content-related quality of atopic eczema-related videos on YouTube. J Med Internet Res. 2020;22:e15599. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [34].Wang H, Zhang H, Cao J, Zhang F, Xiong W. Quality and content evaluation of thyroid eye disease treatment information on TikTok and Bilibili. Sci Rep. 2025;15:25134. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [35].Hu Y, Cheng Z, Zhu H. Reliability and quality of hypothyroidism videos on TikTok and Bilibili: a cross-sectional content analysis in China. Medicine (Baltim). 2026;105:e47523. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [36].Teunis T, Lubberts B, Reilly BT, Ring D. A systematic review and pooled analysis of the prevalence of rotator cuff disease with increasing age. J Shoulder Elbow Surg. 2014;23:1913–21. [DOI] [PubMed] [Google Scholar]
  • [37].Zheng S, Tong X, Wan D, Hu C, Hu Q, Ke Q. Quality and reliability of liver cancer-related short Chinese videos on TikTok and Bilibili: cross-sectional content analysis study. J Med Internet Res. 2023;25:e47210. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [38].Sun F, Zheng S, Wu J. Quality of information in gallstone disease videos on TikTok: cross-sectional study. J Med Internet Res. 2023;25:e39162. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [39].Tutan D, Kaya M. Evaluation of YouTube videos as a source of information on hepatosteatosis. Cureus. 2023;15:e46843. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [40].Li A, Xing Q, Zhang Y, et al. Evaluation of the information quality related to osteoporosis on TikTok. BMC Public Health. 2024;24:2880. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [41].Chen Z, Pan S, Zuo S. TikTok and YouTube as sources of information on anal fissure: a comparative analysis. Front Public Health. 2022;10:1000338. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [42].Gong X, Dong B, Li L, Shen D, Rong Z. TikTok video as a health education source of information on heart failure in China: a content analysis. Front Public Health. 2023;11:1315393. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [43].Subramanian T, Araghi K, Akosman I, et al. Quality of spine surgery information on social media: a DISCERN analysis of TikTok videos. Neurospine. 2023;20:1443–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [44].Chen Q, Min C, Zhang W, Ma X, Evans R. Factors driving citizen engagement with government TikTok accounts during the COVID-19 pandemic: model development and analysis. J Med Internet Res. 2021;23:e21463. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [45].Kim YJ, Lin M, Davis RJ, Ayo-Ajibola O, Kwon D, Hur K. Evaluation of the quality of thyroidectomy-related posts on a video-based social media platform. Laryngoscope Investig Otolaryngol. 2023;8:1685–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [46].Al-Leimon O, Pan W, Jaber AR, et al. Reels to remembrance: attention partially mediates the relationship between short-form video addiction and memory function among youth. Healthcare (Basel). 2025;13:252. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [47].Efe IE, Tesch C, Subedi P. YouTube as a source of patient information on awake craniotomy: analysis of content quality and user engagement. World Neurosurg X. 2024;21:100249. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [48].Lin JC, Weintraub N, Aragaki DR. Nonsurgical treatment for rotator cuff injury in the elderly. J Am Med Dir Assoc. 2008;9:626–32. [DOI] [PubMed] [Google Scholar]
  • [49].Micallef J, Pandya J, Low AK. Management of rotator cuff tears in the elderly population. Maturitas. 2019;123:9–14. [DOI] [PubMed] [Google Scholar]
  • [50].Quinn B, Nichols L, Frazee J, Payton M, Linger RMA. Dissemination of information on selective serotonin reuptake inhibitors on TikTok: analytical mixed methods study of creator types, content tone, and user engagement. JMIR Ment Health. 2025;12:e77383. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [51].Yang Y, Jiang S, Ning X, et al. Assessing the video content quality of TikTok and Bilibili as health information sources for systemic lupus erythematosus: a cross-sectional study. Int J Rheum Dis. 2025;28:e70341. [DOI] [PubMed] [Google Scholar]
  • [52].Puljak L. Using social media for knowledge translation, promotion of evidence-based medicine and high-quality information on health. J Evid Based Med. 2016;9:4–7. [DOI] [PubMed] [Google Scholar]
  • [53].Sufrate-Sorzano T, Corton-Carrasco O, Garrote-Cámara ME, et al. Social networks as a tool for evidence-based health education: umbrella review. Nurs Rep. 2024;14:2266–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [54].Wang JW, Xun JJ, Zhao FF. The theoretical impact of AI-based quality evaluation of short-video health information on public cognition and treatment adherence: a case study of denosumab combined with PD-1/PD-L1 therapy for lung cancer bone metastasis. Front Public Health. 2025;13:1724546. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [55].Pienkowska A, Ravaut M, Mammadova M, et al. Understanding COVID-19 impacts on the health workforce: AI-assisted open-source media content analysis. JMIR Form Res. 2024;8:e53574. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Medicine are provided here courtesy of Wolters Kluwer Health

RESOURCES