Quality, content, and reliability of lumbar spondylolisthesis–related videos on short-video platforms: A cross-sectional study of TikTok and Bilibili

Abstract

Short-video platforms have become prominent sources of health-related information for musculoskeletal conditions, yet the quality and reliability of content concerning lumbar spondylolisthesis remain insufficiently evaluated. This study aimed to systematically assess the characteristics, content composition, educational quality, and reliability of lumbar spondylolisthesis–related videos on TikTok and Bilibili. A cross-sectional analysis was conducted on videos retrieved from TikTok and Bilibili on December 25, 2025. Eligible videos were evaluated for general characteristics, uploader type, presentation format, thematic content, and audience engagement. Educational quality and information reliability were assessed using the Global Quality Score, modified DISCERN (mDISCERN), and Journal of the American Medical Association (JAMA) benchmark criteria. Correlations between video characteristics, engagement metrics, and quality scores were analyzed. A total of 199 videos were included (TikTok: n = 105; Bilibili: n = 94). Overall, videos demonstrated moderate educational quality and limited reliability. Treatment-related content predominated, whereas epidemiology, prevention, and rehabilitation were infrequently addressed; notably, etiology ranked as the third most common topic on both platforms. Videos from orthopedic practitioners had higher mDISCERN and JAMA scores but low overall quality; Global Quality Score differed by uploader type and was not highest for orthopedic practitioners. Scientific popularization videos exhibited superior educational quality, while engagement metrics did not reliably reflect informational value. Video duration – particularly on Bilibili – was positively correlated with quality and reliability scores, whereas engagement indicators showed weak or negligible associations. Lumbar spondylolisthesis–related short videos on major platforms provide moderate educational value but show substantial deficiencies in reliability and content balance. Greater professional involvement and improved platform-level quality oversight are warranted to enhance the dissemination of accurate and high-quality health information.

1. Introduction

Lumbar spondylolisthesis is a common degenerative spinal disorder characterized by the anterior displacement of 1 vertebral body relative to the adjacent segment, frequently affecting the aging population.^[1,2] Clinically, it often presents with chronic low back pain, radicular symptoms, neurogenic claudication, and functional limitation, which can substantially impair quality of life and daily activity.^[3,4] Epidemiological evidence indicates that the prevalence of lumbar spondylolisthesis increases with age, particularly among older adults and women, and that the overall disease burden is likely to increase with population aging.^[5,6] Although many patients initially experience mild or nonspecific symptoms, delayed diagnosis and inadequate disease understanding may lead to symptom progression, reduced treatment effectiveness, and increased healthcare utilization.^[7] Enhancing public awareness of lumbar spondylolisthesis – including its risk factors, clinical manifestations, diagnostic approaches, and management strategies – is therefore essential for promoting early recognition, appropriate treatment selection, and improved long-term outcomes.

In recent years, social media and short-video platforms have become increasingly prominent sources of health-related information for the general public.^[8,9] These platforms are characterized by extensive user bases, algorithm-driven content recommendation systems, and high levels of audience engagement, enabling rapid dissemination of medical information.^[10] In China, platforms such as TikTok^[11] and Bilibili^[12] have emerged as influential channels through which users seek knowledge about musculoskeletal conditions, treatment options, and rehabilitation experiences. However, the decentralized and largely unregulated nature of health information dissemination on these platforms raises concerns regarding content accuracy, educational quality, and reliability. Prior studies evaluating short videos related to various medical conditions – including spinal disorders,^[13] cancer,^[14] and chronic diseases^[15] – have consistently reported high audience engagement accompanied by substantial variability in informational quality and frequent deficiencies in transparency and evidence-based reporting. Such limitations may contribute to misinformation, inappropriate self-management, and delayed professional consultation.

Despite the growing clinical relevance of lumbar spondylolisthesis and the widespread use of short-video platforms for health information seeking, systematic evaluations of lumbar spondylolisthesis–related video content remain scarce. In particular, little is known about how information quality and reliability vary across platforms, uploader backgrounds, and video presentation formats, or whether audience engagement reflects the educational value of the content. Therefore, this study aimed to conduct a comprehensive assessment of lumbar spondylolisthesis–related videos on TikTok and Bilibili by analyzing video characteristics, content themes, information quality, and reliability using validated evaluation instruments. These instruments use standardized, item-based scoring to assess key credibility-related domains in social media health content, including overall educational quality and usefulness (Global Quality Score, GQS),^[16] reliability of consumer health information (modified DISCERN, mDISCERN),^[17,18] and transparency indicators of scientific credibility (JAMA benchmark criteria: authorship, attribution, disclosure, and currency).^[19] By identifying strengths and deficiencies in existing online content, this study seeks to provide evidence-based insights to guide healthcare professionals, platform administrators, and viewers in optimizing the dissemination of accurate, reliable, and high-quality educational information on lumbar spondylolisthesis.

2. Methods and materials

2.1. Study design and ethical considerations

This cross-sectional study was conducted to evaluate the characteristics, content composition, educational quality, and reliability of short videos related to lumbar spondylolisthesis on 2 major Chinese video-sharing platforms: TikTok (https://www.douyin.com) and Bilibili (https://www.bilibili.com). Data collection was performed on December 25, 2025. All analyzed data were obtained from publicly accessible online sources, and no personal, private, or identifiable user information was collected. Consequently, ethical approval and informed consent were not required.

2.2. Video search strategy and data collection

A systematic search was conducted on TikTok and Bilibili using a predefined Chinese keyword related to lumbar spondylolisthesis. The exact search term was “腰椎滑脱.” To reduce potential bias introduced by personalized recommendation algorithms, all searches were performed on a desktop computer using web browsers while not logged into any user account, and browser cache and browsing history were cleared before each search session. On each platform, videos were retrieved according to the default comprehensive relevance ranking. Videos were screened sequentially based on ranking order, and each candidate video was watched in full before eligibility decisions and data extraction. The detailed screening and selection process is illustrated in Figure 1. For each eligible video, the following data were extracted using a standardized data collection form: platform, video URL, upload date, video duration, and engagement metrics, including platform-reported counts of likes, collections, comments, and shares. In this study, “audience engagement” refers to these metrics recorded at the time of data extraction.

Figure 1.

The flow chart of this study. TCM = traditional Chinese Medicine.

2.3. Inclusion and exclusion criteria

Videos were included if they met all of the following criteria:

1. The primary content focused on lumbar spondylolisthesis;

2. The video provided information related to disease mechanisms, clinical manifestations, diagnosis, treatment, prevention, rehabilitation, or prognosis;

3. The video was presented in Chinese and contained complete audiovisual content.

Videos were excluded if they:

1. Were unrelated to lumbar spondylolisthesis;

2. Were duplicated across searches;

3. Had been uploaded for less than 1 week at the time of data collection;

4. Consisted primarily of advertisements without substantive medical or educational information.

2.4. Classification of videos

2.4.1. Classification by uploader type

Based on uploader identity, videos were categorized into 3 groups: orthopedic clinicians, traditional Chinese medicine (TCM) practitioners, and other uploaders. Classification was determined according to uploader profiles, self-introductions, and video content. “Other uploaders” referred to uploaders who were neither orthopedic clinicians nor TCM practitioners, including physical therapists, radiologists, pain physicians, and nonprofessional individual users. Because each of these subgroups accounted for a small number of videos, they were combined into a single “other” category to ensure sufficient sample size for group comparisons. Each video was independently classified by 2 reviewers, with discrepancies resolved through discussion.

2.4.2. Classification by video presentation format

According to the visual context and presentation style, videos were further classified into 3 presentation formats: scientific popularization scenarios, outpatient clinic scenarios, and other clinical scenarios. This classification reflected differences in recording environment and content delivery style. All videos were independently categorized by 2 reviewers, and consensus was reached through discussion when disagreements occurred.

2.5. Video content evaluation

Each video was assessed for content coverage across predefined domains, including epidemiology, etiology, clinical manifestations, diagnosis, treatment, prevention, rehabilitation, and prognosis. A single video could be assigned to multiple content categories where applicable. Content classification was performed independently by 2 reviewers.

2.6. Assessment of video quality and reliability

Video quality and reliability were evaluated using 3 validated assessment tools: the Global Quality Score (GQS),^[16] the mDISCERN instrument,^[17,18] and the Journal of the American Medical Association (JAMA) benchmark criteria.^[19] The GQS assesses overall educational quality on a 5-point scale ranging from 1 (poor quality) to 5 (excellent quality). The mDISCERN score evaluates the reliability of health information based on 5 yes/no questions, with higher scores indicating greater reliability. The JAMA benchmark criteria assess authorship, attribution, disclosure, and currency, yielding a total score ranging from 0 to 4. Two reviewers independently assessed all videos after standardized training. Discrepancies were resolved through discussion until consensus was achieved, and the consensus scores were used for the final analyses. Inter-rater reliability of the initial independent ratings was assessed separately for GQS, total mDISCERN, and total JAMA using intraclass correlation coefficients (2-way random-effects model, absolute agreement), reporting single-measure ICCs with 95% confidence intervals.

2.7. Statistical analysis

Statistical analyses were conducted using R software (version 4.5.0; R Foundation for Statistical Computing). The Shapiro–Wilk test was applied to assess data normality. Continuous variables were summarized as medians with interquartile ranges (IQRs) due to non-normal distributions. The Kruskal–Wallis H test was applied for comparisons among multiple groups. When the Kruskal–Wallis test was significant, post hoc pairwise comparisons were performed using independent-samples Mann–Whitney U tests with Bonferroni correction for multiple comparisons. Spearman’s rank correlation coefficient was employed to evaluate associations between video characteristics, engagement metrics, and quality assessment scores. Correlation strength was interpreted as negligible (|r| < 0.2), weak (0.2–0.4), moderate (0.4–0.6), or strong (>0.6). A 2-sided P value < .05 was considered statistically significant.

3. Results

3.1. Video characteristics and baseline features

After screening according to the predefined inclusion and exclusion criteria (Fig. 1), a total of 199 videos focusing on lumbar spondylolisthesis were included in the final analysis, comprising 94 videos from Bilibili and 105 from TikTok. The general characteristics, audience engagement metrics, and information quality assessments are summarized in Table 1. Overall, the median video duration was 110.00 seconds (IQR, 64.00–249.50). Videos on Bilibili were significantly longer than those on TikTok (216.50 vs 74.00 seconds, P < .001). In terms of audience engagement, TikTok videos demonstrated significantly higher numbers of likes, comments, shares, and collections than those on Bilibili (all P ≤ .006). Regarding information quality, the median GQS score across all videos was 3.00 (IQR, 3.00–4.00), indicating moderate educational quality. No significant differences were observed between platforms in GQS or mDISCERN scores. However, TikTok videos achieved higher JAMA scores than Bilibili videos (median 1.00 vs 0.00, P < .001). Figure 2 illustrates the distribution of video uploaders (Fig. 2A) and presentation formats (Fig. 2B) across platforms, revealing clear differences between TikTok and Bilibili and providing the basis for subsequent subgroup analyses.

Table 1.

Video characteristics.

Variables	Total (n = 199)	Bilibili (n = 94)	TikTok (n = 105)	P
General information
Video length (s), M (Q1, Q3)	110.00 (64.00, 249.50)	216.50 (92.50, 497.50)	74.00 (51.00, 135.00)	<.001
Likes, M (Q1, Q3)	101.00 (20.50, 291.00)	19.50 (4.00, 105.25)	223.00 (87.00, 647.00)	<.001
Collections, M (Q1, Q3)	48.00 (13.00, 204.00)	25.00 (10.00, 158.50)	71.00 (14.00, 314.00)	.006
Comments, M (Q1, Q3)	10.00 (1.00, 32.00)	1.50 (0.00, 12.00)	21.00 (7.00, 58.00)	<.001
Shares, M (Q1, Q3)	22.00 (5.00, 93.00)	12.00 (2.25, 60.75)	40.00 (10.00, 209.00)	<.001
Video quality
GQS, M (Q1, Q3)	3.00 (3.00, 4.00)	3.00 (3.00, 4.00)	3.00 (3.00, 4.00)	.352
vmDISCERN, M (Q1, Q3)	2.00 (2.00, 2.00)	2.00 (2.00, 2.00)	2.00 (2.00, 2.00)	.826
JAMA, M (Q1, Q3)	0.00 (0.00, 1.00)	0.00 (0.00, 1.00)	1.00 (0.00, 1.00)	<.001

GQS = Global Quality Score, JAMA = Journal of the American Medical Association, M = Median, mDISCERN = modified DISCERN, Q₁ = 1st quartile, Q₃ = 3rd quartile.

Figure 2.

Distribution of video uploaders and formats on Bilibili and TikTok. (A) Distribution of video uploaders on Bilibili and TikTok. (B) Distribution of video formats on Bilibili and TikTok.

3.2. Video quality stratified by uploader type

Based on uploader identity, videos were classified into orthopedic clinicians, TCM practitioners, and other uploaders, as defined during screening (Fig. 1). “Other uploaders” included non-orthopedic and non-TCM sources (e.g., physical therapists, radiologists, pain physicians, and nonprofessional individual users) and were analyzed as a single group due to the small number of videos in each subgroup. Comparisons of engagement metrics and information quality across uploader groups are presented in Table 2, with score distributions shown in Figure 3. Inter-rater reliability for the initial independent ratings was excellent, with single-measure ICCs (2-way random-effects model, absolute agreement) of 0.971 (95% CI: 0.965–0.976) for GQS, 0.808 (95% CI: 0.758–0.854) for total mDISCERN, and 0.745 (95% CI 0.675–0.8) for total JAMA. Significant differences were observed in video characteristics and quality among uploader types. Videos uploaded by orthopedic clinicians were shorter (median 86.00 seconds, IQR 56.50–142.50) than those from other uploaders and TCM practitioners (P < .001). Despite their shorter duration, orthopedic clinician-uploaded videos had higher engagement in terms of likes and comments than the other groups (both P < .001). In contrast, the Other Uploader group showed a higher number of shares (median 36.0 vs 23.5 for orthopedic clinicians; P = .022). In terms of information quality, quality scores varied across uploader groups. GQS and mDISCERN scores differed significantly across uploader groups, with videos from TCM practitioners showing lower scores (P < .01 for both). JAMA scores also differed significantly, with orthopedic clinician-uploaded videos having higher JAMA scores, suggesting better authorship disclosure and source transparency (P < .001). These patterns are further illustrated in Figure 3.

Table 2.

Comparison of characteristics among different uploaders.

Variables	Orthopedic practitioner (n = 118)	Other uploader(n = 34)	TCM practitioner(n = 47)	P
General information
Video length (s), M (Q1, Q3)	86.00 (56.50,142.50)	187.00 (92.25,626.00)	260.00 (87.00,571.50)	<.001
Likes, M (Q1, Q3)	158.00 (48.25,426.50)	75.00 (16.00,355.75)	20.00 (4.00,110.50)	<.001
Collections, M (Q1, Q3)	52.00 (10.00,170.50)	84.50 (19.00,348.25)	25.00 (14.50,164.50)	.259
Comments, M (Q1, Q3)	13.50 (4.00,42.00)	8.00 (2.00,33.50)	1.00 (0.00,12.00)	<.001
Shares, M (Q1, Q3)	23.50 (7.00,94.25)	36.00 (5.50,180.00)	8.00 (2.50,51.50)	.022
Video quality
GQS, M (Q1, Q3)	3.00 (3.00,4.00)	3.50 (3.00,4.00)	3.00 (2.00,3.00)	<.001
mDISCERN, M (Q1, Q3)	2.00 (2.00,2.00)	2.00 (2.00,2.00)	2.00 (1.00,2.00)	.002
JAMA, M (Q1, Q3)	1.00 (0.00,1.00)	0.00 (0.00,0.00)	0.00 (0.00,0.00)	<.001

GQS = Global Quality Score, JAMA = Journal of the American Medical Association, M = Median, mDISCERN = modified DISCERN, Q₁ = 1st quartile, Q₃ = 3rd quartile, TCM = traditional Chinese medicine.

Figure 3.

A violin diagram comparing 3 assessment tools among different uploaders. The width of each violin reflects the kernel density (distribution) of scores. The embedded boxplot shows the median (horizontal line) and the interquartile range (box: Q₁–Q₃). Statistical significance is indicated as ns (not significant), *P < .05, **P < .01, ***P < .001, and ****P < .0001. Round dots indicate outliers beyond the whiskers (1.5 × IQR); groups without outliers therefore show no dots. GQS = Global Quality Score, IQR = interquartile range, JAMA = Journal of the American Medical Association, mDISCERN = modified DISCERN.

3.3. Video quality stratified by presentation format

According to presentation context, videos were categorized as scientific popularization, outpatient clinic recordings, or other clinical scenario footage. Comparative results are summarized in Table 3, with violin plots shown in Figure 4. Video duration differed significantly across formats (P < .001), with scientific popularization videos being the longest (median 194.00 seconds, IQR 91.00–400.00). Outpatient clinic and other clinical scenario videos were substantially shorter. No significant differences were observed among formats for likes, comments, or shares.

Table 3.

Comparison of characteristics among different formats.

Variables	Other clinical scenario (n = 13)	Outpatient Clinic scenario (n = 69)	Scientific Popularization scenario (n = 117)	P
General information
Video length (s), M (Q1, Q3)	73.00 (61.00,80.00)	70.00 (51.00,114.00)	194.00 (91.00,400.00)	<.001
Likes, M (Q1, Q3)	145.00 (70.00,225.00)	106.00 (36.00,210.00)	96.00 (12.00,416.00)	.847
Collections, M (Q1, Q3)	22.00 (8.00,64.00)	29.00 (6.00,93.00)	79.00 (16.00,314.00)	<.001
Comments, M (Q1, Q3)	17.00 (7.00,33.00)	11.00 (3.00,23.00)	9.00 (1.00,42.00)	.620
Shares, M (Q1, Q3)	8.00 (5.00,26.00)	15.00 (4.00,55.00)	31.00 (5.00,128.00)	.101
Video quality
GQS, M (Q1, Q3)	3.00 (2.00,3.00)	3.00 (2.00,3.00)	4.00 (3.00,4.00)	<.001
mDISCERN, M (Q1, Q3)	2.00 (1.00,2.00)	2.00 (2.00,2.00)	2.00 (2.00,2.00)	.002
JAMA, M (Q1, Q3)	1.00 (1.00,1.00)	1.00 (0.00,1.00)	0.00 (0.00,1.00)	<.001

JAMA = Journal of the American Medical Association, GQS = Global Quality Score, M = Median, mDISCERN = modified DISCERN, Q₁ = 1st quartile, Q₃ = 3rd quartile.

Figure 4.

A violin diagram comparing 3 assessment tools among different formats. The width of each violin reflects the kernel density (distribution) of scores. The embedded boxplot shows the median (horizontal line) and the interquartile range (box: Q₁–Q₃). Statistical significance is indicated as ns (not significant), *P < .05, **P < .01, ***P < .001, and ****P < .0001. Round dots indicate outliers beyond the whiskers (1.5 × IQR); groups without outliers therefore show no dots. GQS = Global Quality Score, IQR = interquartile range, JAMA = Journal of the American Medical Association, mDISCERN = modified DISCERN.

In contrast, information quality varied significantly by presentation format. Scientific popularization videos achieved the highest GQS scores (median 4.00, IQR 3.00–4.00; P < .001). MDISCERN and JAMA scores also differed significantly across formats (P ≤ .002), with videos recorded in clinical settings demonstrating relatively better performance in terms of source transparency. These differences are clearly depicted in Figure 4.

3.4. Video content

As shown in Figure 5, treatment-related content was the most frequently addressed topic across both platforms. Clinical manifestations were also commonly discussed, whereas diagnostic information was less frequently included. In contrast, epidemiology and preventive strategies were rarely covered, while etiology ranked as the third most common topic on both platforms. Rehabilitation-related content accounted for only a small proportion of videos. Overall, content distribution was similar between TikTok and Bilibili, with a predominant focus on treatment and symptoms and limited coverage of disease background and prevention.

Figure 5.

Comparative analysis of video content distribution for lumbar spondylolisthesis between TikTok and Bilibili.

3.5. Correlation analysis between video engagement metrics and quality scores

Figure 6 presents the correlation matrices between engagement metrics and quality scores on TikTok and Bilibili. On both platforms, GQS and mDISCERN scores were positively correlated, indicating consistency between educational quality and information reliability (TikTok: r  = 0.70; Bilibili: r = 0.59). On TikTok (Fig. 6A), video duration showed weak correlations with GQS (r = 0.14) and mDISCERN (r = 0.17). Engagement metrics were strongly intercorrelated (e.g., likes and comments, r = 0.84) but demonstrated weak or negligible associations with quality scores. JAMA scores showed no meaningful correlations with video length or engagement metrics.

Figure 6.

Correlation matrix of video engagement metrics and quality scores on TikTok and Bilibili. (A) Correlation matrix of video engagement metrics and quality scores on TikTok. (B) Correlation matrix of video engagement metrics and quality scores on Bilibili. GQS = Global Quality Score, JAMA = Journal of the American Medical Association, mDISCERN = modified DISCERN.

On Bilibili (Fig. 6B), video duration was strongly correlated with GQS (r = 0.75) and moderately correlated with mDISCERN (r = 0.40). Engagement metrics were highly intercorrelated (e.g., collections and shares, r = 0.97), yet remained weakly associated with quality scores. JAMA scores again showed minimal correlation with engagement or duration. Overall, video duration – particularly on Bilibili – was more closely associated with educational quality and reliability, whereas engagement metrics primarily reflected user interaction rather than informational quality.

4. Discussion

This cross-sectional study provides a systematic evaluation of lumbar spondylolisthesis–related videos on TikTok and Bilibili, focusing on video characteristics, content composition, information quality, reliability, and their relationships with audience engagement. By integrating platform-level comparisons, uploader identity, presentation formats, and correlation analyses, this study offers a comprehensive overview of how lumbar spondylolisthesis–related information is currently disseminated through short-video platforms. Overall, the findings indicate that although these platforms generate substantial audience engagement, the educational quality of the content remains moderate, and informational reliability is limited. Importantly, professional background and presentation context were more closely associated with content quality than popularity metrics, underscoring critical challenges in short-video–based health communication.

Consistent with prior studies examining musculoskeletal^[20] and chronic disease–related content on social media, TikTok videos demonstrated significantly higher engagement than those on Bilibili. This difference is likely attributable to platform-specific recommendation algorithms, content consumption patterns, and user demographics. However, higher engagement did not translate into better informational quality or reliability. Correlation analyses showed that likes, comments, shares, and collections were strongly intercorrelated but exhibited weak or negligible associations with GQS, mDISCERN, or JAMA scores. These findings reinforce the growing consensus that audience engagement primarily reflects viewer interest and platform mechanics rather than the educational value of health-related content, cautioning against equating popularity with credibility.^[21,22] Our overall pattern of moderate GQS but low mDISCERN/JAMA scores is also broadly aligned with recent short-video assessments in other musculoskeletal topics (e.g., knee osteoarthritis and ankle sprain) that reported limited transparency and reliability-related content despite substantial engagement.^[23,24] Differences across studies may reflect variations in topic complexity, sample composition, and the scoring instruments used.

Uploader type was associated with differences in information quality–related measures in our sample. In particular, orthopedic clinician-uploaded videos showed higher transparency (JAMA), whereas videos from TCM practitioners tended to have lower GQS and mDISCERN scores; GQS and mDISCERN distributions were broadly comparable between orthopedic clinicians and other uploaders. Notably, absolute mDISCERN and JAMA scores were low across all uploader groups, suggesting that reliability-related elements and transparency were limited overall. Therefore, these between-group differences should be interpreted as associations rather than evidence of uniformly high-quality content from any single uploader type. Similar associations between uploader background and quality–related scores have been reported in other musculoskeletal short-video analyses.^[23,24] Despite being shorter in duration, clinician-generated videos achieved higher engagement, suggesting that professional credibility and focused messaging may enhance both reach and quality. In contrast, videos from non-clinician uploaders – particularly TCM practitioners – showed lower quality-related scores, which may increase the likelihood of incomplete explanations and warrants caution for viewers when professional oversight is limited. These findings emphasize the importance of encouraging greater participation by orthopedic specialists in online health communication related to lumbar spondylolisthesis.

Presentation format also played a significant role in shaping information quality. Videos designed for scientific popularization demonstrated the highest educational quality, likely due to their longer duration and more structured explanatory approach. In contrast, videos recorded in outpatient clinics or other clinical scenarios, although potentially authentic and engaging, showed lower quality and transparency scores. These results suggest that while clinical setting–based recordings may enhance realism and viewer trust, structured educational formats remain better suited for comprehensive patient education. Optimizing content by combining clinical authenticity with structured explanations may represent a balanced strategy for improving both engagement and educational value.

Analysis of content themes revealed a pronounced emphasis on treatment-related information, with relatively limited coverage of diagnosis, prevention, etiology, and epidemiology. This imbalance mirrors patterns observed in short-video studies across multiple medical disciplines and reflects a broader tendency toward intervention-focused narratives.^[25,26] For lumbar spondylolisthesis, insufficient emphasis on prevention and early recognition is particularly concerning, as conservative management, risk factor modification, and timely diagnosis play critical roles in disease progression and patient outcomes. The underrepresentation of preventive and background information represents a missed opportunity for public education and highlights an important area for content improvement. For lumbar spondylolisthesis, such omissions may have practical consequences. Overly simplified or non-individualized exercise recommendations may be inappropriate for certain patients (e.g., those with neurologic symptoms or suspected instability), and failure to discuss diagnostic work-up (e.g., indications for imaging, grading, or assessment of instability) may delay appropriate evaluation. In addition, limited discussion of warning signs (e.g., progressive neurologic deficits or bowel/bladder symptoms) may reduce timely care-seeking. Therefore, improving the completeness and safety framing of short-video content is particularly important for this condition.

Correlation analyses further clarified the relationship between video characteristics and informational value. In our analysis, video duration showed a positive association with educational quality and reliability scores, particularly on Bilibili. A similar positive relationship between video length and quality-related scores has also been observed in short-video health information studies on other medical topics, which may reflect that longer videos allow more complete explanations and context.^[27] In contrast, engagement metrics were poor indicators of content quality across both platforms. These findings suggest that modestly increasing video length – when paired with clear structure and evidence-based messaging – may improve educational outcomes without necessarily compromising viewer interest.

Taken together, this study highlights the dual nature of short-video platforms as both powerful dissemination tools and imperfect sources of health information. While TikTok and Bilibili offer unprecedented reach and user engagement for lumbar spondylolisthesis–related content, notable deficiencies persist in informational reliability, transparency, and preventive education. Addressing these challenges will require coordinated efforts from healthcare professionals, content creators, and platform administrators to promote evidence-based, transparent, and balanced health communication. Future research should explore longitudinal trends, evaluate intervention-based content optimization strategies, and assess the impact of professional training or platform-level quality controls on video quality.

Several limitations should be acknowledged. First, this study analyzed content from only 2 platforms, which may limit generalizability to other social media ecosystems. Second, the cross-sectional design captures a snapshot in time and cannot account for temporal changes in content or algorithms. Third, although standardized assessment tools were used, some degree of subjectivity in quality evaluation is unavoidable. Future studies incorporating automated assessment methods and larger, more diverse samples may help address these limitations.

5. Conclusion

This study systematically assessed the characteristics, content themes, information quality, and reliability of lumbar spondylolisthesis–related videos on TikTok and Bilibili. Overall, the videos exhibited moderate educational quality but limited reliability, with treatment-focused content predominating and preventive and background information being underrepresented. Videos uploaded by orthopedic clinicians and those presented in structured scientific popularization formats demonstrated higher quality, whereas audience engagement metrics were poorly aligned with informational value. Video duration showed a stronger association with educational quality than engagement indicators. These findings highlight the need for greater involvement of healthcare professionals, improved content structuring, and enhanced transparency to optimize the dissemination of reliable and evidence-based information on lumbar spondylolisthesis through short-video platforms.

Author contributions

Data curation: Keli Guo.

Formal analysis: Keli Guo.

Investigation: Keli Guo.

Methodology: Keli Guo.

Software: Keli Guo.

Supervision: Yanzheng Gao.

Writing – original draft: Keli Guo.

Writing – review & editing: Yanzheng Gao.