Evaluating the reliability and quality of osteoporosis content on TikTok and BiliBili: A cross-sectional content analysis

Abstract

Objective

This study evaluated the information quality and user engagement of osteoporosis-related videos on Bilibili and TikTok, and examined their associations with uploader characteristics, content topics, and video quality factors.

Methods

On October 11, 2025, we conducted a systematic evaluation of the information quality and reliability of the top 100 Chinese-language short videos related to osteoporosis on TikTok and BiliBili platforms, ultimately including 171 valid videos for analysis. Using the Global Quality Scale (GQS) and a modified DISCERN instrument, we assessed multiple dimensions of video content. Furthermore, Spearman correlation analysis and the Kruskal-Wallis test were employed to examine how platform type, uploader category, and content characteristics influence video quality.

Results

A total of 171 videos were included in the analysis, comprising 82 from Bilibili and 89 from Tiktok. Bilibili videos exhibited significantly longer durations compared to Tiktok videos (median 307.5 seconds vs. 151.0 seconds; P < 0.001). Conversely, Tiktok videos demonstrated significantly higher user engagement metrics, including median number of likes (3407.0 vs. 65.0; P < 0.001), collections (1432.0 vs. 78.0; P < 0.001), and shares (677.0 vs. 48.5; P < 0.001). Regarding uploader characteristics, professional institutions contributed only 4.1% of the total sample. Nevertheless, videos uploaded by professional institutions achieved the highest median GQS score (4.50) and mDISCERN score (4.00), significantly surpassing those uploaded by professional individuals and non-professional individuals (P = 0.011 and P < 0.001, respectively). User engagement metrics strongly intercorrelated ( $r = 0.87$ –0.94, all $P < 0.001$ ) but correlated only weakly with quality scores ( $∣ r ∣ < 0.27$ ).

Conclusions

Bilibili videos feature longer durations and more detailed content, whereas TikTok videos demonstrate superior user engagement. Videos uploaded by professional institutions attained higher quality ratings compared to other uploader types, although this finding is based on a small subsample (n=7) and should be interpreted with caution. Medication-related content attracted the greatest public attention. Nevertheless, the weak correlation between user engagement and quality scores indicates that high popularity does not equate to high informational reliability. These findings underscore the need to strengthen professional credentialing mechanisms and optimize algorithmic recommendations to enhance both the scientific accuracy and communicative reach of osteoporosis health information on short video platforms.

Keywords

osteoporosis mDISCERN GQS TikTok BiliBili short-video platforms health information quality

1. Introduction

Osteoporosis is a common metabolic bone disease characterized by reduced bone mineral density and microstructural deterioration, which contribute to increased bone fragility and susceptibility to fracture.^1,2 Fragility fracture represents the most clinically significant consequence of this condition. As the global population ages, the prevalence of osteoporosis has continued to rise, establishing it as a major public health concern that substantially diminishes quality of life among middle-aged and older adults.^3,4 A Canadian study reported that more than two million individuals are affected by osteoporosis,⁵ underscoring its substantial global burden. However, because bone loss typically progresses asymptomatically in its early stages, many patients remain undiagnosed until they present with a fracture, at which point the optimal window for intervention has often been missed.⁶ Moreover, even after diagnosis, a considerable proportion of patients fail to receive standardized and systematic treatment.^1,6 Despite increasing public awareness of the disease, significant gaps persist in both early prevention and long-term treatment adherence.

In recent years, to enhance the accessibility of osteoporosis screening, researchers have begun exploring machine learning models based on routine clinical data, aiming to enable efficient risk assessment without reliance on dual-energy X-ray absorptiometry (DXA). A stacked ensemble model developed by Carvalho and Gavaia⁷ achieved 93% accuracy using routine clinical data from the National Health and Nutrition Examination Survey (NHANES). Subsequent studies by the same group further validated the model’s stability across diverse populations⁸ and successfully applied it to mortality prediction in orthopedic patients.⁹ These advances suggest that accessibility in osteoporosis management extends beyond the dissemination of health information to include the decentralization and simplification of screening tools.

Within this context, short-video platforms such as TikTok and Bilibili have emerged as important channels through which the public—particularly younger populations—access health-related knowledge, owing to their intuitive and visually engaging content formats as well as their substantial communicative reach.^10,11 As a leading short-video platform in China, TikTok has cultivated a vast user base and become deeply integrated into daily life. Bilibili, by contrast, specializes in medium-to long-format videos and hosts a large community of science communication content creators, rendering it especially popular among younger audiences.¹² Together, these platforms provide diverse avenues for the creation and dissemination of medical education content. Nevertheless, the quality of online health information has long been a subject of concern.¹³ The quality of existing video content varies considerably, and many videos contain professional inaccuracies that have the potential to mislead the public.¹⁴ Previous studies have evaluated the quality of videos related to various conditions, including hypertension, radiotherapy health information, and liver cancer, revealing significant heterogeneity across different platforms.^12,15,16

Given that the prevention and treatment of osteoporosis emphasize early intervention, a systematic evaluation and comparison of the quality of osteoporosis-related information across different platforms remains an underexplored area of research.¹⁷ While prior studies have focused on single platforms 17^,18, this study provides three incremental contributions: (1) direct comparison between TikTok and Bilibili; (2) systematic analysis of uploader type differences (professional vs. nonprofessional); and (3) the first evaluation of osteoporosis content on Bilibili.

Therefore, the present study aims to conduct a cross-sectional analysis to systematically evaluate and compare the content characteristics, information quality (assessed using the Global Quality Score [GQS] and the modified DISCERN [mDISCERN] instrument^18–20), and source composition of osteoporosis-related videos on TikTok and Bilibili. This study aims to investigate the following core question: how do platform affordances and algorithmic mechanisms shape the structural relationship between information quality and user engagement in short-form video health communication? By systematically evaluating the content characteristics and quality of osteoporosis-related videos on two major domestic platforms, this study seeks to provide key evidence for the development of more effective online health communication strategies.

2. Methods

2.1. Search and data collection

This study employed a cross-sectional content analysis design and was reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement.²¹ For the content coding and quality assessment components, the recommendations of the Standards for Reporting Qualitative Research (SRQR) were consulted. A keyword search using the term “osteoporosis” was conducted on Tiktok and Bilibili, with the top 100 ranked videos from each platform constituting the initial sample. To minimize the influence of algorithm-driven personalized recommendations and potential selection bias, the searches were performed using newly created accounts with no prior browsing history or user interaction, and access was conducted in guest mode where applicable.²² During the screening process, duplicate videos and those irrelevant to the research topic were excluded (detailed in Figure 1: 11 videos were removed from TikTok and 18 from Bilibili). The final analytical sample was restricted to the top 100 videos per platform, a threshold consistent with prior studies suggesting that videos ranked beyond the top 100 contribute marginally to overall content analysis.²³

Figure 1.

Search strategy and video filtering program.

For each included video, the following data were extracted and recorded: title, uploader name, video content description, duration, and user engagement metrics including number of likes, favorites, and shares. All data were systematically entered and managed using Microsoft Excel. The data collection process was completed on October 11, 2025.

2.2. Classification of videos

Based on the background and attributes of the video uploaders, they were categorized into the following four groups:

(1) Professional individuals: This category includes individuals with formal medical or nursing qualifications (e.g., licensed physicians, physical therapists, etc.) or those possessing relevant academic backgrounds engaged in related research (e.g., medical researchers specializing in orthopedics).

(2) Nonprofessional individuals: This group comprises individuals without systematic medical or nursing education, such as patients sharing personal experiences with the condition, or fitness enthusiasts lacking clinical credentials.

(3) Professional institutions: These are organizations whose core mission is to provide medical or nursing services, including various types of hospitals, university-affiliated medical centers, and health administrative authorities.

(4) Nonprofessional institutions: This category consists of organizations whose primary operations are not related to healthcare, such as general media agencies, fitness brands, and community groups unassociated with clinical practice.

Furthermore, based on the core information conveyed in the videos, they were classified into the following six mutually exclusive content themes:

(1) Disease Prevention

(2) Disease Diagnosis

(3) Clinical Symptoms

(4) General Information

(5) Lifestyle Interventions

(6) Medication

Each video was assigned to only one category. For videos covering multiple topics, classification was based on the topic accounting for the largest proportion of the content or representing the most prominent core theme. Discrepancies in classification were resolved by a third researcher.

2.3. Evaluating methodologies

In this study, a modified version of the mDISCERN instrument (see Supplementary File 2) was utilized to evaluate the overall reliability of the video content. The mDISCERN tool is well-established in academic research as a means to help both consumers and healthcare providers assess the quality of health-related information. It has been widely applied in numerous studies evaluating online health information across various medical fields, including cancer, arthritis, and other chronic conditions.^14–16

The revised DISCERN instrument evaluates video quality based on five specific criteria. Clarity refers to whether the content is presented in a clear, concise, and understandable manner. Relevance describes the extent to which the content aligns with the topic of osteoporosis. Traceability reflects whether the information provides citations to credible sources. Comprehensiveness pertains to the depth and objectivity of the information presented. Finally, impartiality examines whether the content is free from bias. Each of these criteria was scored using a binary scale, where “yes” responses were assigned 1 point and “no” responses received 0 points. As a result, the total score for each video ranged from 0 to 5, offering a quantitative basis for assessing content reliability.^22–24

In addition, the Global Quality Scale (GQS; see Supplementary File 1) was applied to evaluate the overall quality of the videos. This instrument is a recognized measure for assessing the reliability and quality of health-related information disseminated through digital video platforms. The GQS uses a 5-point scoring system, in which a score of 1 indicates poor quality content and a score of 5 represents high quality.^25,26

2.4. Evaluation process

The video content evaluation was conducted by two graduate students (Rater A: S.Z.; Rater B: H.B.) affiliated with the orthopedics department of a tertiary teaching hospital. Prior to formal assessment, both raters received standardized training on the mDISCERN instrument and GQS scale. To ensure consistent interpretation of the scoring criteria, a calibration exercise involving 10 pilot videos was carried out.

All 171 videos were independently evaluated by the two raters. Each reviewer recorded scores for every video according to the mDISCERN criteria—clarity, relevance, traceability, comprehensiveness, and impartiality—as well as an overall GQS score. A discrepancy was defined as a difference of one point or more in either the total mDISCERN score or the GQS rating.

In cases where scoring discrepancies occurred, the two raters first conducted a joint reevaluation of the video content, comparing their assessments and identifying potential sources of disagreement, such as differing interpretations of specific criteria like “comprehensiveness.” If consensus could not be reached through discussion, a senior orthopedic expert (M.Z.) with over 30 years of clinical experience was consulted. This expert reviewed the video and the initial evaluations, facilitated further discussion, and made the final scoring decision based on current clinical guidelines and evidence-based practice.

2.5. Statistical analysis

Given the non-normal distribution characteristics of the data, this study will present statistical results using medians and interquartile ranges (IQR). For comparisons between two groups, the Mann-Whitney U test will be applied, while the Kruskal-Wallis H test will be employed when comparing three or more groups. For the Kruskal-Wallis test, where the result was statistically significant, Dunn’s test was further performed for post hoc pairwise comparisons, and the Bonferroni method was applied to adjust the P-values to control for type I error. In analyzing the GQS and mDISCERN scores, Fisher’s exact test will be utilized.

To evaluate inter-rater reliability, Cohen’s kappa coefficient will be calculated (Table 1). Due to the non-parametric nature of the data distribution, Spearman’s rank correlation analysis will be conducted to examine associations among various video-related variables and their relationships with assessment scores. A p-value of less than 0.05 will be considered statistically significant for all analyses.

Table 1.

Inter-rater reliability analysis.

		Cohen’s kappa
Tiktok	GQS	0.819
Tiktok	mDISCERN	0.805
Bilibili	GQS	0.852
Bilibili	mDISCERN	0.827

Abbreviations: GQS, Global Quality Scale; mDISCERN: modified DISCERN instrument.

2.6. Ethics statement

This study used publicly available data from TikTok and BiliBili. All data were collected in compliance with the platforms’ terms of service and applicable guidelines. No personally identifiable information was included in the analysis. As the study involved only publicly available content, institutional ethics review was waived.

2.7. Use of artificial intelligence

The authors declare that no artificial intelligence tools were used in the preparation of this manuscript.

3. Results

3.1. Video characteristics and uploader types

This study included a total of 171 videos, of which 82 (48.0%) were from BiliBili and 89 (52.0%) from TikTok. The median number of likes in the BiliBili group was significantly lower than that in the TikTok group (65.00 vs. 3407.00, p<0.001), while the median video duration was notably longer on BiliBili (307.50 seconds vs. 151.00 seconds, p<0.001), indicating distinct differences in content presentation between the two platforms. Furthermore, the BiliBili group also showed lower median values for collections (78.00 vs. 1432.00, p<0.001) and shares (48.50 vs. 677.00, p<0.001) compared to the TikTok group, with all these differences being statistically significant. The general characteristics of the videos are summarized in Table 2.

Table 2.

Characteristics of the videos on TikTok and BiliBili.

Variables	Total (n = 171)	Bilibili (n = 82)	TikTok (n = 89)	P
Likes, M (Q1,Q3)	906.00(74.00,3623.50)	65.00(21.50,293.00)	3407.00(1571.00,8023.00)	<0.001
Collections, M (Q1,Q3)	361.00(78.00,1572.00)	78.00(36.00,216.00)	1432.00(524.00,3885.00)	<0.001
Shared, M (Q1,Q3)	205.00(46.50,757.50)	48.50(22.00,146.75)	677.00(247.00,1800.00)	<0.001
Duration(s), M (Q1,Q3)	216.00(112.50,315.50)	307.50(206.25,527.50)	151.00(90.00,222.00)	<0.001

Abbreviations: M, median; Q, quartile.

Note. P-values were derived from the Mann-Whitney U test for comparisons between the two platforms.

Among all 171 videos, 160 (93.6%) were produced by professional individuals, 7 (4.1%) by professional institutions, and 4 (2.3%) by non-professional individuals. No content was published by non-professional institutions. This distribution highlights the dominant role of professional individuals in creating and disseminating osteoporosis-related videos. In terms of platform-specific distribution, 98% of the videos on TikTok were uploaded by professional individuals, with other types of uploaders accounting for only 2%, indicating a highly concentrated source of content. In contrast, Bilibili exhibited greater diversity in uploader types: although professional individuals remained the primary contributors (approximately 55%), non-professional individuals accounted for 18%, professional institutions for 17%, and non-professional institutions for 10%. The proportion of non-professional uploaders was notably higher on BiliBili than on TikTok, as illustrated in Figure 2.

Figure 2.

General characteristics of osteoporosis-related videos from TikTok and BiliBili. (a) Pie chart displaying the distribution of uploader types across all platforms. (b) Stacked bar chart illustrating the proportional distribution of uploader types on each.

Regarding the different uploader categories, the overall median number of likes was 906.0. Specifically, non-professional individuals had a median of 20.0 likes, professional institutions had 517.0, and professional individuals had 1114.0. Although certain variations were observed in the medians for collections, shares, and video duration across uploader types, no statistically significant differences were detected in between-group comparisons. The overall median video duration was 216.0 seconds, with non-professional individuals at 266.0 seconds, professional institutions at 213.0 seconds, and professional individuals at 216.5 seconds. Detailed data are presented in Table 3. In summary, differences in likes among uploader types were observed, reflecting a pattern consistent with the video characteristics described earlier.

Table 3.

Characteristics of uploader types of videos on TikTok and BiliBili.

Variables	Total (n = 171)	Professional individuals (n =160)	Professional institutions (n =7)	Nonprofessional individuals (n =4)	P
Likes, M (Q1,Q3)	906.00(74.00,3623.50)	1114.00(82.50,4015.75)	517.00(206.00,899.50)	20.00(14.75,16767.25)	0.319
Collections, M (Q1,Q3)	361.00(78.00,1572.00)	405.50(80.00,1811.25)	182.00(116.00,583.00)	60.50(29.00,5818.25)	0.472
Shared, M (Q1,Q3)	205.00(46.50,757.50)	207.50(48.75,752.75)	215.00(102.00,507.00)	44.00(34.50,7538.25)	0.783
Duration(S), M (Q1,Q3)	216.00(112.50,315.50)	216.50(111.00,314.00)	213.00(145.50,317.50)	266.00(165.50,732.50)	0.567

Abbreviations: M, median; Q, quartile.

Note. P-values were derived from the Kruskal-Wallis H test for comparisons across the three uploader types.

Significant differences in video quality assessed by the GQS and mDISCERN instruments were observed across uploader types (Table 4). Professional institutions achieved the highest median scores for both GQS (4.50) and mDISCERN (4.00), followed by professional individuals (GQS: 4.00; mDISCERN: 3.00), whereas non-professional individuals scored the lowest (GQS: 3.00; mDISCERN: 2.50). These differences were statistically significant (GQS: P = 0.011; mDISCERN: P < 0.001). Post hoc pairwise comparisons using Dunn’s test with Bonferroni correction revealed that, for GQS scores, both professional institutions and professional individuals scored significantly higher than non-professional individuals (adjusted P = 0.028 and 0.004, respectively), with no significant difference detected between professional institutions and professional individuals (adjusted P = 1.000). For mDISCERN scores, no pairwise comparisons reached statistical significance after Bonferroni correction (all adjusted P > 0.05). Collectively, these findings indicate that video quality is closely associated with uploader type, underscoring the importance of professional background in ensuring information reliability.

Table 4.

Quality scores of osteoporosis-related videos by uploader type.

Uploader type	n (%)	GQS score, M (Q1,Q3)	mDISCERN score, M (Q1,Q3)
Professional Individuals	160 (93.6)	4.00 (3.00, 5.00)	3.00 (2.00, 4.00)
Professional Institutions	7 (4.1)	4.50 (4.00, 4.50)	4.00 (4.00, 4.00)
Nonprofessional Individuals	4 (2.3)	3.00 (2.25, 4.75)	2.50 (2.00, 3.00)
p		0.011	<0.001

Abbreviations: GQS, Global Quality Scale; M, median; Q, quartile; mDISCERN: modified DISCERN instrument.

Note. P-values were derived from the Kruskal-Wallis test for comparing the distributions of scores across the three uploader types.

3.2. Video content analysis

Figure 3 presents a content analysis of 171 videos categorized as follows: disease prevention (n = 16, 9.4%), disease diagnosis (n = 29, 17.0%), clinical symptoms (n = 15, 8.8%), general information (n = 33, 19.3%), lifestyle interventions (n = 37, 21.6%), and medication (n = 41, 24.0%). In terms of platform distribution, TikTok showed higher proportions of medication-related and general information videos, whereas BiliBili featured substantial content not only on medication but also on disease diagnosis and clinical symptoms. Additionally, both platforms included diverse themes such as lifestyle interventions and disease prevention, indicating a relatively broad coverage of topics across platforms.

Figure 3.

Content analysis of osteoporosis-related videos on TikTok and BiliBili. (a) Donut chart presenting the distribution of content categories across both platforms; (b) Radial bar chart showing content type proportions on TikTok; (c) Radial bar chart displaying content type proportions on BiliBili.

Regarding engagement metrics (Table 5), the overall median values were 906.0 (74.0–3623.5) for likes, 361.0 (78.0–1572.0) for collections, 205.0 (46.5–757.5) for shares, and 216.0 seconds (112.5–315.5) for video duration. Significant differences across content categories were observed for likes (P = 0.002) and collections (P = 0.005), with medication-related videos achieving the highest median likes and collections (2632.0 and 1524.0, respectively), whereas disease diagnosis–related videos had the lowest (137.0 and 100.0, respectively). Post hoc pairwise comparisons revealed that likes for disease diagnosis–related videos were significantly lower than those for general information and medication-related videos (adjusted P = 0.008 and 0.002, respectively), and collections for disease diagnosis–related videos were significantly lower than those for medication-related videos (adjusted P = 0.004). Although medication-related videos had the highest median number of shares (542.0), the differences across categories were not statistically significant (P = 0.066). No significant differences were found in video duration across content categories (P = 0.217); lifestyle intervention videos had the longest median duration (240.0 seconds), whereas disease prevention videos had the shortest (126.0 seconds).

Table 5.

Characteristics of content of videos on TikTok and BiliBili.

Variables	Total (n = 171)	Disease prevention (n = 16)	Disease diagnosis (n = 29)	Clinical symptoms (n = 15)	General information (n = 33)	Lifestyle interventions (n = 37)	Medication (n = 41)	P
Likes, M (Q1,Q3)	906.00 (74.00,3623.50)	873.00 (270.75,4586.00)	137.00 (21.00,464.00)	963.00 (235.00,3227.00)	1487.00 (391.00,3118.00)	481.00 (44.00,2870.00)	2632.00 (85.00,7573.00)	0.002
Collections, M (Q1,Q3)	361.00 (78.00,1572.00)	527.00 (245.50,2456.00)	100.00 (34.00,401.00)	275.00 (70.00,472.00)	444.00 (198.00,1253.00)	182.00 (61.00,1161.00)	1524.00 (112.00,5560.00)	0.005
Shared, M (Q1,Q3)	205.00 (46.50,757.50)	244.50 (60.25,1077.50)	84.00 (27.00,341.00)	156.00 (56.00,286.50)	243.00 (141.00,748.00)	169.00 (39.00,620.00)	542.00 (45.00,1800.00)	0.066
Duration(S), M (Q1,Q3)	216.00 (112.50,315.50)	126.00 (84.25,257.25)	284.00 (155.00,317.00)	153.00 (100.50,258.00)	213.00 (112.00,306.00)	240.00 (136.00,435.00)	199.00 (91.00,301.00)	0.217

Abbreviations: M, median; Q, quartile.

Note. P-values were derived from the Kruskal-Wallis H test for comparisons across the seven content categories.

3.3. Assessment of video information quality and reliability

The GQS and mDISCERN scores were categorized into three tiers: low (scores 1-2), moderate (score 3), and high (scores 4-5). When comparing the platforms based on the proportion of high-quality videos (scores ≥ 4), BiliBili demonstrated a notably higher percentage of videos meeting this threshold in mDISCERN scores (30.5% vs. 20.2%), though this difference did not reach statistical significance (p=0.121). In contrast, TikTok showed a marginally higher proportion of high-quality videos according to GQS scores (59.6% vs. 53.7%, p=0.436). The detailed comparison of high-quality video proportions is presented in Table 6.

Table 6.

Comparison of high-quality video proportions between BiliBili and TikTok.

Variable	Bilibili (n = 82)	TikTok (n = 89)	p
GQS score ≥4
No. of Videos	44.0	53.0
Percentage, %	53.7	59.6	0.436
mDISCERN score ≥4
No. of Videos	25.0	18.0
Percentage, %	30.5	20.2	0.121

Abbreviation: GQS, Global Quality Scale; mDISCERN: modified DISCERN instrument.

Note. P-values were derived from Z-test for the difference in proportions between two independent groups.

Significant differences were observed between the two platforms in GQS scores assigned by both reviewers (both p < 0.001). According to Reviewer A’s GQS assessment, Bilibili videos were predominantly concentrated in scores of 3 (37.8%), 4 (39.0%), and 5 (23.2%). In contrast, TikTok videos exhibited a polarized distribution, with higher proportions at the lower end (score 1: 4.5%; score 2: 5.6%) and the upper end (score 5: 39.3%). A similar pattern was observed in Reviewer B’s GQS scores: Bilibili videos were primarily clustered around scores of 3 (37.8%) and 4 (50.0%), whereas TikTok videos showed notably higher proportions at score 2 (12.4%) and score 5 (36.0%).

Regarding mDISCERN scores, Reviewer A’s assessment revealed significant inter-platform differences (p < 0.001). Bilibili videos were mainly distributed across scores of 3 (42.7%) and 4 (32.9%), while TikTok videos demonstrated substantially higher proportions at the lower end, particularly at score 1 (16.9%) and score 2 (33.7%). Reviewer B’s mDISCERN scores also showed significant differences between platforms (p < 0.001), with Bilibili videos concentrated at scores of 3 (46.3%) and 4 (37.8%), whereas TikTok videos were predominantly clustered at score 2 (43.8%). The detailed distribution is presented in Table 7.

Table 7.

Ratings of TikTok and Bilibili by different reviewers.

Reviewers’ score	Total (n = 171)	Bilibili (n = 82)	TikTok (n = 89)	P
A (GQS score), n(%)				<0.001
1	4 (2.3%)	0 (0.0%)	4 (4.5%)
2	5 (2.9%)	0 (0.0%)	5 (5.6%)
3	44 (25.7%)	31 (37.8%)	13 (14.6%)
4	64 (37.4%)	32 (39.0%)	32 (36.0%)
5	54 (31.6%)	19 (23.2%)	35 (39.3%)
B (GQS score), n(%)				<0.001
1	0 (0.0%)	0 (0.0%)	0 (0.0%)
2	11 (6.4%)	0 (0.0%)	11 (12.4%)
3	41 (24.0%)	31 (37.8%)	10 (11.2%)
4	77 (45.0%)	41 (50.0%)	36 (40.4%)
5	42 (24.6%)	10 (12.2%)	32 (36.0%)
A (mDISCERN score), n(%)				<0.001
1	19 (11.1%)	4 (4.9%)	15 (16.9%)
2	42 (24.6%)	12 (14.6%)	30 (33.7%)
3	56 (32.7%)	35 (42.7%)	21 (23.6%)
4	43 (25.1%)	27 (32.9%)	16 (18.0%)
5	11 (6.4%)	4 (4.9%)	7 (7.9%)
B (mDISCERN score), n(%)				<0.001
1	10 (5.8%)	4 (4.9%)	6 (6.7%)
2	47 (27.5%)	8 (9.8%)	39 (43.8%)
3	62 (36.3%)	38 (46.3%)	24 (27.0%)
4	44 (25.7%)	31 (37.8%)	13 (14.6%)
5	8 (4.7%)	1 (1.2%)	7 (7.9%)

Abbreviation: GQS, Global Quality Scale; mDISCERN: modified DISCERN instrument.

Note. P-values for A_GQS and B_mDISCERN were derived from Fisher’s exact test due to expected cell counts <5; P-values for the remaining scores were derived from Pearson’s chi-square test.

Analysis based on uploader type (Table 8) revealed that among the 171 videos analyzed, the majority were uploaded by professional individuals (n = 160, 93.6%), followed by professional institutions (n = 7, 4.1%) and non-professional individuals (n = 4, 2.3%). Reviewer A’s GQS scores differed significantly across uploader categories (p = 0.032), with professional institutions achieving the highest median score (5.0), followed by professional individuals (4.0), and non-professional individuals receiving the lowest median score (3.0). Post hoc pairwise comparisons revealed that GQS scores for both professional institutions and professional individuals were significantly higher than those for non-professional individuals (adjusted P = 0.028 and 0.004, respectively). Although Reviewer B’s GQS scores did not reach statistical significance (p = 0.550), a similar trend was observed, with professional institutions demonstrating a higher median score (4.0) compared to non-professional individuals (3.0).

Table 8.

Ratings by different commenters for different types of uploaders on TikTok and BiliBili.

Reviewers’ score	Total (n = 171)	Professional individuals (n =160)	Professional institutions (n =7)	Nonprofessional individuals (n =4)	P
A(GQS score), M (Q1,Q3)	4.00 (3.00, 5.00)	4.00 (3.00, 5.00)	5.00 (4.00, 5.00)	3.00 (3.00, 4.50)	0.032
B(GQS score), M (Q1,Q3)	4.00 (3.00, 4.00)	4.00 (3.00, 5.00)	4.00 (4.00, 4.00)	3.00 (3.00, 4.50)	0.550
A(mDISCERN score), M (Q1,Q3)	3.00 (2.00, 4.00)	3.00 (2.00, 4.00)	4.00 (4.00, 4.00)	2.50 (2.00, 3.00)	0.004
B(mDISCERN score), M (Q1,Q3)	3.00 (2.00, 4.00)	3.00 (2.00, 4.00)	4.00 (4.00, 4.00)	2.50 (2.00, 3.00)	0.004

Abbreviations: GQS, Global Quality Scale; Q, quartile.; mDISCERN: modified DISCERN instrument.

Note. P-values were derived from the Kruskal-Wallis H test for comparisons across the three uploader types.

For mDISCERN scores, both reviewers identified statistically significant differences across uploader types (both p = 0.004). Professional institutions consistently attained the highest median scores (both 4.0), followed by professional individuals (both 3.0), while non-professional individuals received the lowest median scores (both 2.5). These findings suggest that videos produced by professional institutions demonstrate superior information quality, whereas those created by non-professional individuals tend to exhibit lower quality.

The complete score distributions are detailed in Table 7. Furthermore, the box plots in Figures 4 and 5 provide additional visual evidence of the central tendencies and dispersion patterns for both reviewers’ scores.

Figure 4.

GQS distributions for osteoporosis-related videos. (a) Reviewer A’s GQS scores; (b) Reviewer B’s GQS scores; Abbreviations: GQS, Global Quality Scale.

Figure 5.

mDISCERN distributions for osteoporosis-related videos. (a) Reviewer A’s mDISCERN scores; (b) Reviewer B’s mDISCERN scores.

3.4. Spearman correlation analysis

A Spearman correlation analysis was conducted to examine the associations between video characteristics and quality assessment scores (GQS and mDISCERN) among osteoporosis-related videos. The results revealed statistically significant positive correlations among user engagement metrics. Specifically, the number of likes was strongly correlated with the number of collections (r = 0.92, p < 0.0001) and shares (r = 0.87, p < 0.001). A strong positive correlation was also observed between collections and shares (r = 0.94, p < 0.0001). In contrast, video duration demonstrated weak associations with user engagement indicators, with a modest negative correlation with likes (r = -0.41) and non-significant correlations with collections (r = -0.27) and shares (r = -0.24).

Regarding video quality assessment, both GQS and mDISCERN scores assigned by the two independent reviewers exhibited weak correlations with user engagement metrics, with all absolute correlation coefficients falling below 0.27. These findings suggest that video popularity, as reflected by user interactions, is not substantially associated with information quality. Similarly, video duration showed weak correlations with quality scores, with coefficients ranging from -0.026 to 0.30.

In terms of inter-rater consistency, moderate correlations were observed between the two scales completed by the same reviewer (Reviewer A: r = 0.56, p < 0.0001; Reviewer B: r = 0.52, p < 0.0001). Notably, strong inter-rater reliability was demonstrated for both assessment tools, with a high positive correlation for GQS scores between the two reviewers (r = 0.93, p < 0.0001) and a high positive correlation for mDISCERN scores (r = 0.96, p < 0.0001). These results indicate excellent consistency between the two reviewers in their quality assessments. Complete correlation coefficients are presented in Figure 6.

Figure 6.

Spearman correlation heatmap.

4. Discussion

Osteoporosis, a common metabolic bone disease, renders the quality of health information dissemination critical to public disease awareness and self-management. With short-video platforms emerging as a new frontier for health communication, this study systematically evaluated the content quality and user engagement of osteoporosis-related videos on two major Chinese platforms, Bilibili and Tiktok.

A total of 171 videos were analyzed. The findings revealed a marked disconnect between video popularity and information quality. Tiktok videos demonstrated significantly higher engagement metrics, indicating greater communicative reach, whereas Bilibili videos featured substantially longer durations, suggesting more detailed content. Inter-platform differences in quality assessment were complex, with uploader type emerging as a significant determinant of video quality. Although professional institutions accounted for a small proportion of the sample, their videos achieved the highest GQS and mDISCERN scores, underscoring the importance of professional background in ensuring information reliability. Unlike prior single-platform studies,^17,18 our cross-platform comparison (TikTok vs. Bilibili) and analysis of uploader ecosystem differences reveal how platform affordances shape the quality-popularity disconnect.

With respect to content topics, medication-related videos garnered the highest numbers of likes and collections, reflecting substantial public interest in treatment-focused information. Correlation analyses revealed strong positive intercorrelations among user engagement metrics, collectively reflecting video popularity; however, these metrics exhibited only weak correlations with quality scores, providing robust evidence that popularity does not equate to informational accuracy or reliability. Furthermore, the two reviewers demonstrated high consistency in their quality ratings, ensuring the robustness of the assessment.

4.1. Platform-specific differences and uploader characteristics

A total of 171 videos were included in this analysis, comprising 82 from Bilibili and 89 from Tiktok. Tiktok videos demonstrated significantly higher user engagement metrics—including likes, collections, and shares—compared to those on Bilibili, whereas Bilibili videos exhibited significantly longer durations, reflecting marked differences in content presentation styles between the two platforms.

Regarding uploader composition, Tiktok showed a highly concentrated source profile, with professional individual creators accounting for 98% of videos. In contrast, Bilibili demonstrated a more heterogeneous uploader base, characterized by a significantly higher proportion of non-professional contributors. This discrepancy likely reflects divergent creator ecosystems: Tiktok employs a more stringent credentialing system, whereas Bilibili adopts relatively lenient verification standards. Such structural differences may partially account for the observed inter-platform divergence in both video quality and user engagement. However, because TikTok is composed almost entirely of professional individuals (98%) while Bilibili has a more heterogeneous uploader mix, the observed between-platform quality differences are partially confounded by uploader composition and cannot be attributed solely to platform features.

Quantitative analyses revealed a significant association between uploader credentials and video quality. Professional institutions ranked highest in quality scores, followed by professional individuals, with non-professional individuals scoring lowest, indicating that content producers with professional backgrounds are better positioned to ensure informational accuracy.^16,27 Notably, despite receiving the lowest quality ratings, videos uploaded by non-professional individuals had the longest durations, suggesting that creators lacking formal training may attempt to compensate for low information density by extending video length.²⁷

Engagement metrics presented a more nuanced picture. Although Tiktok significantly outperformed Bilibili in user engagement, this advantage was not accompanied by correspondingly higher video quality—a finding consistent with the well-documented “quality-popularity paradox” in health communication, wherein a video’s popularity does not necessarily reflect its informational reliability.^28,29 Despite producing high-quality content, professional institutions exhibited relatively low engagement levels, indicating that credible information may not receive proportionate visibility within algorithm-driven dissemination environments.^29,30

From a public health perspective, these observations underscore a fundamental tension: videos with high dissemination potential often lack adequate quality, whereas high-quality content frequently fails to achieve comparable reach.³⁰ This finding suggests the need for multifaceted interventions, including platform-level policy guidance, engagement from professional societies, and enhanced public health literacy, to strike a balance between broad information dissemination and the assurance of content accuracy.³⁰

4.2. Influence of content topics

Regarding content topics, videos addressing pharmacological management demonstrated the highest levels of user engagement across all metrics, whereas those focused on disease diagnosis exhibited the lowest.³¹ This discrepancy likely reflects the information preferences of the target audience: for a chronic condition such as osteoporosis, patients and the general public tend to prioritize practical information concerning treatment options that directly influence daily disease management, rather than content related to diagnostic processes, which may be perceived as more technically oriented and less immediately applicable.³² Of note, although videos centered on lifestyle interventions did not achieve particularly high engagement, they had the longest average duration. This finding suggests that content creators may employ extended formats to comprehensively address non-pharmacological strategies—including exercise and nutrition—consistent with the practical, instructional nature of such topics.

From a platform perspective, Tiktok featured a higher proportion of videos on pharmacological management and general information, whereas Bilibili hosted substantial content related to disease diagnosis and clinical manifestations. This variation may be attributable to differences in platform-specific viewing behaviors: users on Tiktok may favor concise, actionable health information, while those on Bilibili may demonstrate greater receptivity to more knowledge-oriented content, such as discussions of disease mechanisms and diagnostic criteria.

Collectively, these observations indicate that user engagement is closely associated with the perceived practicality of the content topic. The high interaction rates observed for medication-related videos underscore a strong public demand for accessible, actionable medical information.³³ Nevertheless, prior research has consistently emphasized that high engagement does not necessarily equate to high quality.¹⁷ Although pharmacological content attracts the most attention, its informational accuracy and reliability warrant careful scrutiny using validated assessment instruments. In future health communication practice, topic-specific strategies should be developed: for high-interest therapeutic content, enhanced professional review mechanisms are essential to ensure accuracy; for lower-engagement yet equally important topics such as diagnosis and prevention, more compelling presentation formats should be explored to improve public awareness and comprehension.

4.3. Correlation between video quality and user engagement

Correlation analysis revealed a complex relationship between user engagement and content quality in osteoporosis-related short videos. User engagement metrics demonstrated strong internal consistency, indicating that likes, collections, and shares collectively serve as indicators of video popularity¹⁷

Notably, video quality scores exhibited only weak correlations with user engagement metrics. This finding highlights a substantial disconnect between popularity and informational reliability within the domain of osteoporosis health communication—videos with high engagement do not necessarily contain high-quality medical content, nor do high-quality videos necessarily achieve corresponding levels of popularity. This observation aligns with the well-documented “quality-popularity paradox” in health communication, wherein users may be more readily attracted to emotionally resonant content or personal narratives than to educational material rigorously grounded in evidence-based medicine.¹⁷ Video duration demonstrated weak associations with both engagement metrics and quality scores, suggesting that duration alone is not a reliable indicator of content quality.^30,32

Regarding rater reliability, the two reviewers exhibited strong consistency in their assessments using the same instruments, demonstrating that the quality evaluation employed in this study possesses high reproducibility and reliability.¹⁷ In contrast, moderate correlations were observed when the same reviewer applied different assessment tools, indicating that although these instruments fall within the domain of quality evaluation, each emphasizes distinct dimensions of content quality.³⁴

These findings underscore a central challenge in short-form video health communication: popular content is not consistently reliable, nor is reliable content consistently popular.^32,35 To address this gap, platforms may consider incorporating quality-sensitive strategies into their algorithmic recommendation systems, thereby ensuring that osteoporosis-related educational content achieves both broad dissemination and informational accuracy.³⁰

4.4. Practical implications and future directions

The findings of this study offer actionable insights for multiple stakeholders. Platform developers should optimize algorithmic recommendation mechanisms to preferentially surface high-quality content and strengthen credentialing systems to mitigate the dissemination challenges posed by the quality-popularity paradox.^32,36 Medical professionals are encouraged to adopt platform-specific content strategies that balance informational accuracy with communicative reach—for instance, concise and practical formats on Tiktok versus more in-depth, structured presentations on Bilibili. Policymakers may consider establishing quality assessment standards for health-related short videos and promoting public health literacy initiatives.³² Patients and the general public should prioritize content from verified accounts and seek confirmation of treatment-related information through clinical consultation.

Enhancing the accessibility of osteoporosis management also requires the development of convenient screening tools. Recent research has introduced machine learning models based on routine clinical data that can efficiently assess osteoporosis risk without reliance on DXA.^7–9 Future efforts could explore the integration of such screening tools with short video platforms—for example, by delivering personalized educational content based on individual risk stratification—thereby establishing a seamless continuum from risk assessment to knowledge acquisition.

4.5. Strengths and limitations

This study has several methodological strengths. First, it conducted a structured cross-platform evaluation of osteoporosis-related videos on Bilibili and Tiktok, providing timely insights into the quality of bone health information within the digital health communication landscape. Second, by integrating uploader characteristics, content topics, two validated quality assessment instruments, and four user engagement metrics, the study achieved a comprehensive analysis that combined information quality with audience response; all videos were independently rated by trained reviewers, demonstrating excellent inter-rater reliability. Third, the use of predefined keyword searches, newly created platform accounts, and standardized inclusion criteria minimized potential biases arising from algorithmic recommendations and sampling procedures. These strengths offer a replicable methodological framework for future research.

Although the present study did not directly assess the impact of audiovisual formats on older audiences, the observed differences between platforms highlight the relevance of this issue. Bilibili videos featured longer durations and a more diverse composition of uploaders, whereas Tiktok videos were characterized by shorter formats and higher user engagement. Given that osteoporosis predominantly affects middle-aged and older adults—a population that tends to rely on clear structural organization and moderate information density during information processing—existing literature suggests that older viewers often exhibit lower rates of information retention and comprehension when exposed to fast-paced, information-dense short videos.^37,38 Future research should further investigate the relationship between audiovisual characteristics and information comprehension across different age groups to optimize health communication strategies tailored to older populations.

Several limitations of this study should be acknowledged. First, the cross-sectional design precludes causal inferences, and data collection was completed within a single day, limiting the ability to capture the dynamic evolution of platform content. Second, the study focused exclusively on Chinese-language content and top-ranked videos, which constrains the generalizability of the findings across linguistic and cultural contexts. Third, the GQS and mDISCERN instruments were originally developed for static, text-based health information and may not fully capture dynamic features unique to short video platforms, such as audiovisual presentation and narrative techniques.^35,39 Additionally, the small sample sizes for professional institutions and non-professional individuals may affect the robustness of subgroup comparisons. Although inter-rater reliability was high, potential subjective bias in the scoring process and possible misclassification of uploader identities cannot be entirely excluded. Finally, due to platform-imposed data access limitations, deeper user behavior metrics—such as view counts and completion rates—could not be incorporated into the analysis. These limitations underscore the need for future longitudinal, multi-platform research to develop a more comprehensive understanding of the dissemination patterns and public reception of osteoporosis-related health information.

5. Conclusion

This study revealed a marked divergence between content quality and communicative reach in osteoporosis-related short videos, demonstrating that high popularity does not equate to high quality. The higher quality observed for videos uploaded by professional institutions (although based on a small subsample of n=7) suggests a potential role of professional background in ensuring information reliability. The high level of public attention directed toward medication-related content reflects the strong demand for treatment-focused information among patients. These findings challenge the conventional reliance on engagement metrics as proxies for content value and highlight the need for differentiated health communication strategies. Future efforts should focus on optimizing algorithmic recommendations, strengthening professional credentialing mechanisms, and enhancing public health literacy to ensure both broad information dissemination and scientific accuracy, thereby fostering the sustainable development of the osteoporosis health communication ecosystem.

Supplemental material

Supplemental material - Evaluating the reliability and quality of osteoporosis content on TikTok and BiliBili: A cross-sectional content analysis

Supplemental material for Evaluating the reliability and quality of osteoporosis content on TikTok and BiliBili: A cross-sectional content analysis by Siyuan Zhang, Xi Yang, Haowei Bai, Haohao Wang, Yao Chen, Pengyu Liu, Can Zhou, Hao Li, Min Zhang in DIGITAL HEALTH.

Supplemental material

Supplemental material - Evaluating the reliability and quality of osteoporosis content on TikTok and BiliBili: A cross-sectional content analysis

Footnotes

ORCID iDs

Siyuan Zhang

Haowei Bai

Author contributions

Siyuan Zhang: Investigation, Validation, Writing – Original Draft, Writing – Review & Editing. Xi Yang: Data Curation, Formal Analysis, Visualization, Writing – Review & Editing. Haowei Bai: Methodology, Software, Validation, Writing – Review & Editing. Haohao Wang and Pengyu Liu: Data Curation, Resources, Writing – Original Draft. Yao Chen and Can Zhou: Formal Analysis, Visualization, Writing – Review & Editing. Hao Li: Conceptualization, Funding Acquisition, Project Administration, Supervision, Writing – Review & Editing. Min Zhang: Conceptualization, Funding Acquisition, Project Administration, Supervision, Writing – Review & Editing.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Science and Technology Activity Project for Overseas Returnees in Shanxi Province, grant number: 20250050; Natural Science Foundation of Shanxi Province, grant number: 202203021221276; Natural Science Foundation of Shanxi Province, grant number: 202403021211110; Natural Science Foundation of Shanxi Province, grant number: 202403021211109.

Declaration of conflicting interests

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

Data Availability Statement

The authors confirm that the data supporting the findings of this study are available within the article. Further inquiries can be directed to the corresponding author.*

Provenance and peer review

Not commissioned, externally peer reviewed.

Guarantor

Hao Li and Min Zhang

Supplemental material

Supplemental material for this article is available online.

References

Liu

Huang

Tang

, et al. Prevalence of osteoporosis and associated factors among Chinese adults: a systematic review and modelling study. J Glob Health 2025; 15: 04009. https://doi.org/10.7189/jogh.15.04009

Gregson

Armstrong

Avgerinou

, et al. The 2024 UK clinical guideline for the prevention and treatment of osteoporosis. Arch Osteoporos 2025; 20(1): 119. https://doi.org/10.1007/s11657-025-01588-3

Zhou

Xiong

. A bibliometric analysis of the current state and future directions of osteoporosis pharmacological treatment. Front Med (Lausanne) 2025; 12: 1622425. https://doi.org/10.3389/fmed.2025.1622425

Qaseem

Mount

Campos

, et al. Quality Indicators for Osteoporosis in Adults: A Review of Performance Measures by the American College of Physicians. Ann Intern Med 2023; 176(10): 1386–1391. https://doi.org/10.7326/M23-1291

Morin

Feldman

Funnell

, et al. Clinical practice guideline for management of osteoporosis and fracture prevention in Canada: 2023 update. Cmaj 2023; 195(39): E1333–e1348. https://doi.org/10.1503/cmaj.221647

Walker

Shane

. Postmenopausal Osteoporosis. N Engl J Med 2023; 389(21): 1979–1991. https://doi.org/10.1056/NEJMcp2307353

Carvalho

Gavaia

. Enhancing osteoporosis risk prediction using machine learning: A holistic approach integrating biomarkers and clinical data. Comput Biol Med 2025; 192(Pt B): 110289. https://doi.org/10.1016/j.compbiomed.2025.110289

Carvalho

Gavaia

. Letter to the Editor: Robustness of osteoporosis risk prediction models with enhanced statistical analyses. Comput Biol Med 2025; 196(Pt A): 110711. https://doi.org/10.1016/j.compbiomed.2025.110711

Carvalho

Gavaia

Brito Camacho

. OrthoMortPred: Predicting one-year mortality following orthopedic hospitalization. Int J Med Inform 2024; 192: 105657. https://doi.org/10.1016/j.ijmedinf.2024.105657

10.

Liu

Yang

Tan

, et al. Web-Based Video Platforms as Sources of Information on Body Image Dissatisfaction in Adolescents: Content and Quality Analysis of a Cross-Sectional Study. JMIR Form Res 2025; 9: e71652. https://doi.org/10.2196/71652

11.

Song

Suo

, et al. Exploration for a knowledge translation model in public dissemination via social media: Insights from an innovative Cochrane evidence dissemination competition in China. Digit Health 2025; 11: 20552076251357396. https://doi.org/10.1177/20552076251357396

12.

Zheng

Tong

Wan

, et al. Quality and Reliability of Liver Cancer-Related Short Chinese Videos on TikTok and Bilibili: Cross-Sectional Content Analysis Study. J Med Internet Res 2023; 25: e47210. https://doi.org/10.2196/47210

13.

Morahan-Martin

. How internet users find, evaluate, and use online health information: a cross-cultural review. Cyberpsychol Behav 2004; 7(5): 497–510. https://doi.org/10.1089/cpb.2004.7.497

14.

Sittig

Wright

Coiera

, et al. Current challenges in health information technology-related patient safety. Health Informatics J 2020; 26(1): 181–189. https://doi.org/10.1177/1460458218814893

15.

Zhu

Cui

Zhou

, et al. The quality and reliability of short videos about pulmonary hypertension on Bilibili and TikTok: a cross-sectional study. BMC Cardiovasc Disord 2026; 26: 236.

16.

Guo

Ding

Zhang

, et al. Quality Assessment of Radiotherapy Health Information on Short-Form Video Platforms of TikTok and Bilibili: Cross-Sectional Study. JMIR Cancer 2025; 11: e73455. https://doi.org/10.2196/73455

17.

Zhang

, et al. Quality of information in osteoporosis videos on TikTok: a cross-sectional study. Arch Osteoporos 2025; 20(1): 115. https://doi.org/10.1007/s11657-025-01597-2

18.

Zhang

Yin

Liu

, et al. Videos in short-video sharing platforms as sources of information on osteoarthritis: cross-sectional content analysis study. Front Digit Health 2025; 7: 1622503. https://doi.org/10.3389/fdgth.2025.1622503

19.

Dai

Gong

, et al. The Quality and Reliability of Online Videos as an Information Source of Public Health Education for Stroke Prevention in Mainland China: Electronic Media-Based Cross-Sectional Study. JMIR Infodemiology 2025; 5: e64891. https://doi.org/10.2196/64891

20.

Yang

Jiang

Ning

, 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(6): e70341. https://doi.org/10.1111/1756-185X.70341

21.

Skrivankova

Richmond

Woolf

BAR

, et al. Strengthening the reporting of observational studies in epidemiology using mendelian randomisation (STROBE-MR): explanation and elaboration. Bmj 2021; 375: n2233. https://doi.org/10.1136/bmj.n2233

22.

Xie

Deng

, et al. Quality, engagement, and predictive validity of acne-related short videos on Chinese platforms Bilibili and TikTok: A cross-sectional content analysis. Digit Health 2026; 12: 20552076261430231. https://doi.org/10.1177/20552076261430231

23.

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(1): 1594. https://doi.org/10.1186/s12889-024-19077-6

24.

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(2): 105–111. https://doi.org/10.1136/jech.53.2.105

25.

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(9): 2070–2077. https://doi.org/10.1111/j.1572-0241.2007.01325.x

26.

Albayrak

Büyükçavuş

. Does YouTube offer high-quality ınformation? Evaluation of patient experience videos after orthognathic surgery. Angle Orthod 2023; 93(4): 409–416. https://doi.org/10.2319/042522-315.1

27.

Che

, et al. The quality and reliability of short videos about hypertension on TikTok: a cross-sectional study. Sci Rep 2025; 15(1): 25042. https://doi.org/10.1038/s41598-025-08680-1

28.

Yan

Jiang

Zha

, et al. Quality and reliability of hyperlipidemia-related short videos on mainstream social media platforms: a cross-sectional study. Sci Rep 2026.

29.

Melhem

Almomani

Mousa

, et al. Behind the counter, behind the discourse: The paradox of pharmacist influence in Arabic women's health online. Explor Res Clin Soc Pharm 2026; 21: 100697. https://doi.org/10.1016/j.rcsop.2025.100697

30.

Thomas

Mathew

. Learning to see clearly again: How critical media engagement can save public health. Front Public Health 2026; 14: 1731417. https://doi.org/10.3389/fpubh.2026.1731417

31.

Thrasher

Brewer

Niederdeppe

, et al. Advancing Tobacco Product Warning Labels Research Methods and Theory: A Summary of a Grantee Meeting Held by the US National Cancer Institute. Nicotine Tob Res 2019; 21(7): 855–862. https://doi.org/10.1093/ntr/nty017

32.

Dopelt

. The Impact of Social Media on Public Health and Education: Mechanisms, Vulnerabilities, and Policy Implications. European Journal of Investigation in Health, Psychology and Education 2026; 16(3): 43. https://doi.org/10.3390/ejihpe16030043

33.

Labban

Mulugeta-Gordon

Kravitz

, et al. TikTok's Role in HPV Education: A Content Analysis. Gynecology and Obstetrics Clinical Medicine 2026; 6(1): e000239. https://doi.org/10.1136/gocm-2025-000239

34.

Zhou

Huang

Wei

, et al. Benchmark evaluation of video large language models in quality assessment of science popularization videos for dry eye. Sci Rep 2026; 16(1): 8756. https://doi.org/10.1038/s41598-026-39444-0

35.

Sun

Guo

Zhao

, et al. Evaluating the information quality and reliability of ovarian cancer educational content on TikTok and Bilibili: a cross-sectional study. Sci Rep 2026; 16(1): 7296. https://doi.org/10.1038/s41598-026-37077-x

36.

Kuo

Chen

. Predicting User Engagement in Health Misinformation Correction on Social Media Platforms in Taiwan: Content Analysis and Text Mining Study. J Med Internet Res 2025; 27: e65631. https://doi.org/10.2196/65631

37.

Zhang

Gao

. Research on the influencing factors of using health science popularization short videos on the self-management behavior of older adult patients with chronic diseases. Frontiers in Public Health 2026; 14: 2026.

38.

Sun

Zhang

, et al. Preferences for Health Science Popularization Short Videos Among Older Adults in China: A Discrete Choice Experiment. J Appl Gerontol 2025; 7334648251398602. https://doi.org/10.1177/07334648251398602

39.

Wang

Jiang

Wang

. The quality and reliability of Kawasaki disease-related short videos on TikTok and Bilibili: a cross-platform content analysis using GQS, mDISCERN, and JAMA benchmarks. Ital J Pediatr 2026; 52: 69.

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.

0.27 MB

0.00 MB

0.10 MB