Credibility of health-related content on Chinese short-video platforms: A scoping review on accuracy, risks, and quality metrics

Abstract

Background

Short videos have emerged as a significant medium for disseminating health information. However, misleading content can lead to poor health decisions, undermining national efforts to enhance health knowledge and public health literacy.

Objective

This study aims to systematically evaluate the quality and reliability of health-related videos on Chinese short-video platforms and to offer insights for the regulation of digital health on a global scale.

Methods

A comprehensive search was conducted across the China Biology Medicine Database, PubMed, Wanfang, China National Knowledge Infrastructure, and VIP databases for articles published between January 2021 and December 2024. Twenty-five articles meeting the inclusion criteria were included and analyzed to evaluate the quality and credibility of health-related short videos on Chinese platforms, as well as the correlation between the type of creator and video quality.

Results

Health-related videos created by healthcare professionals or institutions demonstrated higher reliability and accuracy. The DISCERN was the most commonly used tool in the evaluation of these videos. Overall, video quality was generally substandard, with the prevalence of inaccurate information ranging from 10.1% to 100% across various health topics.

Conclusions

This review identified substantial deficiencies in the accuracy of health information disseminated through Chinese short-video platforms. The presence of low-quality content has negative impact on public health decision-making. These findings align with evaluation of health-related videos on international platforms, such as YouTube and Instagram. Therefore, it is imperative to adopt comprehensive strategies, including content moderation, creator verification, and responsible algorithm management to improve video quality and ensure the public's access to reliable digital health information.

Introduction

The Internet and social media have become significant sources of healthcare information, with novel online health education technologies gaining widespread acceptance. According to a survey, approximately 80% of the global population engages with health-related content on social media platforms. ¹ The availability of accurate online health information serves to enhance public health literacy and empowers individuals to exercise greater autonomy in health-related decision-making. For example, authoritative information disseminated via social media has been demonstrated to significantly promote hepatitis B virus screening and liver cancer prevention. ² Additionally, it has been shown to improve glycemic control among diabetic patients, thereby enhancing their self-management capabilities and overall quality of life. ³

Although social media has significantly enhanced the public's access to information, it has also facilitated the widespread dissemination of inaccurate and false health-related content. A study indicated that over 70% of individuals reported that information acquired from social media had influenced their health status. ⁴ In a survey conducted by Rahila Iftikhar et al., 42.6% of patients discontinued treatment due to misleading online content. ⁵ Waszak et al. examined health information across multiple platforms and discovered that approximately 40% of the content was identified as false or misleading. ⁶ Similarly, Sommariva et al. analyzed social media posts concerning the Zika virus and found that more than half contained misinformation. ⁷ Furthermore, a report evaluating videos related to gynecological cancers on TikTok revealed that the majority of such content was unreliable.^8,9 Previous studies have reported inaccuracies in health-related information across diverse topics on social media platforms, including nutrition (98%), ¹⁰ attention-deficit/hyperactivity disorder (52%), ¹¹ mask usage during the Coronavirus Disease 2019 pandemic (10.6%), ¹² pediatric urological diseases (77.8%), ¹³ prostate cancer (47%), ¹⁴ and breast cancer (63.08%). ¹⁵ Therefore, comprehensive understanding of the quality and credibility of different categories of health-related short videos is essential to improving the public's ability to discriminate online health information.

This review focused on the status of the quality and credibility of health science-related videos on Chinese short-video platforms. It provided an in-depth analysis of the potential threats to the public health security arising from the dissemination of distorted online health information and offered corresponding recommendations and strategies. The objective was to establish a robust theoretical foundation and practical guidance for the standardized dissemination of health science information on social media platforms, with the ultimate goal of enhancing public health literacy.

Methods

Literature search and inclusion criteria

A comprehensive search was conducted across literature databases most likely to contain studies on the quality and credibility of health science-related videos on social media, particularly in the fields of medicine, public health, and health education. The selected databases included China Biology Medicine Database (CBM), PubMed, Wanfang, China National Knowledge Infrastructure (CNKI) and VIP databases. The search strategy is based on four main selected keywords: “short videos,” “health,” “quality assessment,” and “reliability analysis.” Studies were selected based on inclusion criteria designed to identify articles relevant to the quality evaluation and credibility assessment of health information short videos.

The inclusion criteria were as follows: (1) Publication date between 1 January 2021 and 31 December 2024; (2) The study explicitly focused on health science-related videos originating from Chinese short-video platforms; (3) The study included systematic quality assessment or credibility evaluation of the short video content, covering at least one of the following dimensions: accuracy and scientific validity of information, completeness of information, source credibility, objectivity and evidence support of content, or alignment with professional health communication or clinical guidelines. The exclusion criteria for the literature were as follows: (1) Duplicate publications; (2) Studies solely addressing media characteristics, user behavior, production strategies, social support, current trends, or influencing factors; (3) Studies lacking a clear framework or methodology for quality or credibility assessment.

Two reviewers independently screened the literature according to the above criteria. Discrepancies were initially resolved through discussion and consensus. If agreement could not be reached, a third senior researcher made the final decision based on the established criteria.

Evaluating methodologies

Bibliometric analysis and content analysis methods were employed to conduct a retrospective analysis of the basic information and research content of the included literature. Microsoft Excel was used to summarize key research information for each included study, including publication date, short-video platforms analyzed, specific health topics covered, sample size of short videos, quality assessment tools applied, extent of misinformation identified, and main findings. Radar charts, bar charts, and percentage stacked bar charts were generated using Origin software to analyze the engagement with health information short videos, the proportion of distorted information and the sources and quality of health science-related videos.

Result

Literature screening and basic characteristics

This study conducted a search among 2608 potentially relevant articles from databases including CBM, PubMed, Wanfang, CNKI and VIP using keywords such as “short videos,” “health,” “quality assessment,” and “reliability analysis.” After screening, 2583 articles were excluded based on the exclusion criteria, resulting in 25 eligible studies included in the final analysis (Figure 1). The included studies analyzed 4321 health-related short video samples totally, with a median sample size of 90 videos per study. All studies focused on Chinese social media platforms, covering platforms such as TikTok, Kuaishou, Weibo, Bilibili, Xiaohongshu, WeChat, and Jinri Toutiao, with TikTok being the most common platform across all studies and emerging as a core medium for disseminating health-related short videos. The research topics covered various health domains, including knowledge related to conditions such as osteoarthritis, breast cancer, stroke, nutrition, hypertension, diabetes, and inflammatory bowel disease, as well as health management topics such as adolescent vision health (Table 1).

Figure 1.

PRISMA flow diagram.

Table 1.

Key information on the included articles.

Authors (year)	Health information	Media platform	Data collection time	Sample size of videos	Assessment tools	Misinformation	Related results
Huang Ning 44 （2024）	Osteoarthritis	TikTok	23–30 July 2023	218	Goobie DISCERN	/	Health professionals provided the majority of the videos (66.51%). The total score of DISCERN was (49.53 ± 9.87), which was “fair” level, and the scores of health professionals were higher than those of other types of authors.
Ming Shuai 35 （2023）	Adolescent vision health	TikTok	12 March 2022	168	DISCERN PEMAT-A/V	10.1% of misinformation	Videos were mostly produced by doctors (45.8%). PEMAT-A/V found that TikTok videos had relatively high understandability (84.7%) and moderate actionability (74.9%). Based on the reliability DISCERN score, the overall scores were (19.8 ± 4.4).
Zhang Shiyi 15 （2023）	Breast cancer	TikTok	19 April 2022	156	DISCERN	36.92% of misinformation	Medical professionals were the main source of videos（85.26%）. The professional medical institutions received the highest score in the DISCERN assessment, with a score of (35.5 ± 0.00).
Wang Yaping 36 （2023）	Stroke nutrition	TikTok	18 October 2023	143	DISCERN PEMAT-A/V	11.89% of misinformation	The total score of DISCERN is (44.32 ± 8.05), PEMAT-A/V comprehension score is (79.07 ± 14.41), and the operability score is (71.47 ± 21.45). The overall quality of short stroke nutrition videos on the platform is fair.
Cai Ying 45 （2024）	Gastroesophageal reflux disease	TikTok, Bilibili	10–15 January 2024	TikTok: 83 Bilibil: 81	GQS DISCERN	/	Nongastrointestinal specialists contribute the majority of videos (56.71%). Videos scored 2 and 3 on DISCERN and GQS, respectively, indicating low quality and reliability.
Lai Yong kang 46 （2024）	Nonalcoholic fatty liver disease	TikTok	December 2018–September 2022	497	DISCERN	/	The videos were mostly posted by health care professionals (93%). The health professionals received the highest score in the DISCERN assessment, with a score of (20.65 ± 0.00)
Song Shijie 40 （2021）	Chronic obstructive pulmonary disease	TikTok	6–10 December 2020	199	DISCERN	/	The majority of the videos are uploaded by individual users (84.4%). The nonprofit organizations received the highest score in the DISCERN assessment, with a score of (66.8 ± 3.80)
Gong Xun 47 （2023）	Heart failure	TikTok	6 August 2023	141	DISCERN JAMA GQS	/	Medical professionals were the main source of videos (92.2%). Based on all 141 videos, the median scores were 2, and 2 for JAMA, and GQS. The total score of DISCERN is (26.50 ± 0.00).
Tan Wei 29 （2023）	Helicobacter pylori	TikTok	15 July 2022	116	GQS DISCERN	/	DISCERN classified 0.9% of the videos as “very poor,” 5.2% as “poor,” 68.7% as “fair,” 20% as “good” and only 5.2% as “excellent.”
He Zixuan 38 （2023）	Dietary guidance for inflammatory bowel disease	TikTok, Bilibili, Kwai	3–5 May 2022	125	GQS DISCERN Dietary guidance for IBD	Less than 30% of the correct dietary recommendations	Most of the video publishers are medical professionals（58%）. The overall quality of the videos was substandard quality. Videos from medical professionals were highly reliable.
Xue Xiaoqiang 37 （2022）	Genitourinary cancer	TikTok	13–20 September 2021	61	HONCode DISCERN	36.07% of misinformation	The most videos were contributed by healthcare practitioners (47.54%). According to both the HONCode scale and the DISCERN instrument, 96.72% of TikTok videos were poor to mediocre in quality.
Xun Gong 48 （2024）	Coronary heart disease	TikTok	29 August–2 September 2022	145	DISCERN JAMA GQS	/	The majority of videos originated from health professionals (88.3%). The overall quality of the videos was generally described as “very poor.” All videos scored on DISCERN score, JAMA and GQS assessment at 27, 2, and 2.
Liu Hui 49 （2024）	Breast cancer	TikTok, Bilibili	28 July 2023	TikTok: 100 Bilibili: 100	GQS DISCERN	/	The average GQS and identification assessment scores for TikTok videos are 2 and 1, respectively. The median GQS and DISCERN of Bilibili Video is 3 and 2, respectively. TikTok and Bilibili videos are mainly posted by doctors, accounting for 40% and 56% respectively.
He Wenjie 50 （2024）	Cerebral palsy	TikTok, Kwai, Weibo, Bilibili, RED	21–31 August 2023	TikTok: 77 Kwai: 73 Weibo: 63 Bilibili: 64 RED: 67	DISCERN JAMA GQS	/	The average for JAMA, GQS, and DISCERN stood at 1.62, 2.05, and 1.26, respectively, while RED received the lowest average score.
Wu Minxia 39 （2024）	Hypertension and diabetes	WeCha, TikTok	15 August 2022–15 February 2023	WeChat: 60 TikTok: 60	DISCERN JAMA	100% of misinformation	Among all 120 videos, only 10 videos scored above 38 points in DISCERN, with 45 videos rated as “very poor.” The average JAMA rating for videos was 2.08 points.
Wang Li 51 （2023）	Thyroid cancer	TikTok	20 March 2022	56	DISCERN	/	87.5% of videos were uploaded by physicians. The average DISCERN score was (3.44 ± 0.72), while the average content score was (5.19 ± 0.95).
Wang Menghui 52 （2024）	Gastric cancer	TikTok, Bilibili, YouTube	1 April 2023	TikTok: 100 Bilibili: 100 YouTube: 100	DISCERN JAMA GQS	/	53% of the videos are from medical professionals. In overall quality evaluations by GQS, JAMA, and DISCERN, the videos average scored 2.6/5, 2.41/4, and 2.71/5, respectively.
Zhang Juanjuan 53 （2024）	Cervical cancer	TikTok, Kwai	/	TikTok: 82 Kwai: 81	GQS DISCERN	A patient claims that the HPV vaccine is invalid	74.4% of the videos on TikTok and 48.8% of the videos on Kawi were uploaded by medical professionals, respectively. Videos posted by specialists with the highest quality received GQS and DISCERN scores of 3 and 2, respectively.
Zheng Shuseng 54 （2023）	Liver cancer	TikTok, Bilibili	2 March 2023	TikTok: 100 Bilibil: 100	GQS DISCERN	/	In TikTok, the median GQS was 3 and DISCERN was 5. The median GQS and DISCERN of Bilibili Video is 2 and 4, respectively.
Chen Zeyang 31 （2022）	Anal fissures	TikTok, YouTube	30 May 2022	TikTok: 62 YouTube: 77	DISCERN PEMAT-A/V GQS	/	Based on DISCERN and GQS ratings, YouTube obtained scores of 40.16 and 33.86, while TikTok received scores of 2.92 and 2.93, respectively. As results of DISCERN, 71% of TikTok videos were evaluated as being of “very poor” or “poor” quality, whereas only 3.2% were assessed as possessing “good quality."
Chen Chang 32 （2024）	Osteosarcoma	TikTok, Bilibili	June to July 2023	TikTok: 44 Bilibili: 51	DISCERN JAMA	/	JAMA's average score was (3.8 ± 0.3), while DISCERN's average was (43.8 ± 13.4).
Bai Gaochen 30 （2022）	Pediatric urology	TikTok, Bilibili, Weibo	29 March–31 March 2022	TikTok: 90 Bilibili: 90 Weibo: 90	DISCERN JAMA	/	65.8% of the videos were provided by medical professionals. DISCERN's scores was 36.56 and JAMA's scores was 2.68. DISCERN assessed 60.6% of the videos as “very poor” or “poor,” and only 3.9% as “good” or “excellent.”
Li Aoying 33 （2023）	Osteoporosis	TikTok	28 March 2023	100	DISCERN	/	DISCERN scored 37.69, mostly in the “poor” (54%) and “appropriate” (43%) rating ranges, indicating an overall low quality.
Mao Tianyang 34 （2024）	Acute pancreatitis (AP)	TikTok	20 September 2023	111	DISCERN HONcode APCS	/	The video is mainly published by doctors (89.18%). According to DISCERN, HONcode, and APCS, the average scores achieved were 33.05 ± 7.87, 3.09 ± 0.93, and 1.86 ± 1.30, respectively. TikTok videos related to AP generally exhibited low quality.
Qiu Jun 55 （2024）	Heatstroke	TikTok	5 October 2022	90	DISCERN	/	According to the DISCERN evaluation, the short videos’ reliability scores were between 10.5 and 18.4 on a scale of 40, while their information quality scores ranged from 1.5 to 2.7 on a scale of 5. The videos published by nonprofit organizations are of the highest quality.

Notes: Scores for assessment tools are presented as reported in the original studies.

Due to variations in reporting formats across studies (e.g. some did not report measures of dispersion such as standard deviation or interquartile range), the score metrics could not be fully standardized.

Application of evaluation tools for health information quality and credibility

All 25 included studies employed standardized tools to assess the quality and credibility of health-related short videos. The DISCERN instrument was consistently used across all studies, with most combining two or more tools for cross-validation. Other commonly utilized tools included Goobie, the Global Quality Score (GQS), PEMAT-A/V, the Journal of the American Medical Association (JAMA) benchmarks, HONCode, and disease-specific clinical guidelines (Table 1). The main characteristics of each tool are summarized below:

1. DISCERN, developed by the University of Oxford in 1997, was designed to provide users with a method for evaluating the quality of online information when making decisions about treatment options for health issues. ¹⁶ The tool has been widely used to assess the quality of health-related videos across various social media platforms, such as TikTok, Twitter, and Facebook, and has been demonstrated to be effective and applicable for evaluating content on Chinese video-sharing platforms. ¹⁷ The DISCERN questionnaire consists of three core dimensions: information reliability (eight items), quality of treatment options (seven items), and overall quality (one item). Each item is scored on a scale from 1 (poorest) to 5 (best), with a total score ranging from 16 to 80. Higher scores indicate better video quality, categorized as follows: very poor (16–26), poor (27–38), moderate (39–50), good (51–62), and excellent (63–80). ¹⁸

2. Goobie was developed in 2019 to evaluate medical content on YouTube, and its assessment framework has been widely applied to the content analysis of online medical videos.^19,20 Goobie focuses on scoring the completeness of six core content categories related to diseases, including the definition of the disease, signs/symptoms, risk factors, diagnosis, management, and outcomes. Each category is rated on a scale of 0 (no content), 0.5 (minimal content), 1 (limited content), 1.5 (moderate content), and 2 (comprehensive content). ¹⁹

3. PEMAT-A/V emphasizes the understandability and actionability of information, with 13 items related to understandability and four items related to actionability. Each item is scored as “agree” (1 point), “disagree” (0 points), or “not applicable.” The final score is calculated as the percentage of all applicable items, with higher percentages indicating greater understandability or actionability. A score exceeding 70% suggests that the information is easily understandable. ²¹

4. JAMA employs a 0–4 rating scale to assess the credibility of video sources, with a median score of ≥2 indicating high quality. ²² The GQS scale, ranging from 1 (poor quality) to 5 (excellent quality), is widely recognized for evaluating web-based video content. Videos scoring 4 or 5 are considered high quality;a score of 3 indicates moderate quality, while scores of 1 or 2 reflect low quality. ²³

5. HONcode is a tool designed to harmonize and standardize the quality of online health information, incorporating eight procedural principles: authority, complementarity, privacy, attribution, justifiability, transparency, financial disclosure, and advertising policy. ²⁴ Each principle is evaluated numerically, with each question scored as 1 point. Scores of 0–2 indicate low quality, 3–5 indicate moderate quality, and 6–8 indicate high quality. ²⁵

Among the 25 included studies, the results from DISCERN and other evaluation tools consistently indicated that most health information-related short videos on Chinese platforms were of substandard quality (see Table 1). These videos exhibited significant shortcomings in terms of content completeness, practicality, and standardization, and a lack of credibility. Health science communication in these videos often fails to comprehensively cover all relevant information.

Status of the quality and credibility of online healthcare information on Chinese short-video platforms

The user base of short videos in China and the demand for online healthcare has shown synchronized growth. By December 2024, the number of short video users in China had reached 1.04 billion, accounting for 95.5% of the total internet user population. ²⁶ December 2020 to December 2024 witnessed continuous growth in both the number of short video users and online healthcare users (see Figure 2). Health-related short videos have demonstrated exceptionally high user engagement. Over the past year, medical and health-related videos on TikTok garnered 13.1 billion likes, 3 billion saves, and 1.93 billion shares, with the most-viewed single video exceeding 580 million views. ²⁷ On Kuaishou, nearly 230 million users, on average, viewed health and wellness-related videos daily in 2023, while over 80 million users engaged with such content for more than one minute per month. ²⁸ These statistics underscore the significant role that short videos play in the popularization of health information.

Figure 2.

The count of short video and online medical users in China (data from China internet information Center ²⁶ ).

Despite advancement, significant deficiencies persist in the accuracy of health information available on Chinese short-video platforms, which continue to fall short of fulfilling the public's essential need for scientifically sound and reliable health knowledge. An analysis of 25 studies reveals that, concerning quality and reliability evaluation, 18 studies identified the quality of the health-related videos on the platforms as poor, five studies rated the quality as moderate or average, and only two studies—focusing on short videos about chronic obstructive pulmonary disease and thyroid cancer of TikTok—reported overall satisfactory quality. Figure 3 provides a visual presentation of DISCERN analysis results for short videos on topics such as Helicobacter pylori, ²⁹ pediatric urology, ³⁰ anal fissure, ³¹ osteosarcoma, ³² osteoporosis, ³³ and acute pancreatitis. ³⁴ For subjects like pediatric urology, anal fissure, osteoporosis, and acute pancreatitis, the majority of videos were rated as “very poor” or “poor” in quality, with only a small proportion achieving ratings of “moderate,” “good,” or “excellent.” Conversely, videos concerning Helicobacter pylori and osteosarcoma exhibited a higher proportion of ratings of “moderate” or above.

Figure 3.

Quality assessment of online healthcare information on Chinese short-video platforms.

Distorted health information on Chinese short video platforms

Distorted health information on Chinese short video platforms has been inadequately addressed in existing research, with a predominant focus on content comprehensiveness rather than the accuracy of health information and the underlying causes of errors. Among the 25 articles reviewed, only six systematically quantified the proportion of distorted health information. The findings indicated substantial variability in error rates across different health topics, ranging from 10.1% to 100%. Vision health had the lowest error rate at 10.1%, ³⁵ followed by stroke nutrition at 11.89%, ³⁶ genitourinary cancer at 36.07%, ³⁷ and breast cancer at 36.92%. ¹⁵ In contrast, dietary guidance for inflammatory bowel disease (IBD) exceeded 70%, ³⁸ and treatment recommendations for hypertension and diabetes reached a 100% error rate ³⁹ (see Figure 4).

Figure 4.

The proportion of inaccurate health information on China's short video platforms.

Misinformation predominantly appeared in three distinct forms: (1) contradiction with established authoritative guidelines such as dietary advice for IBD recommending avoidance of dietary fiber and dairy products ³⁸ ), (2) dissemination of outdated information, which involves referencing obsolete data or discontinued methodologies ³⁷ ), and (3) exaggerated claims, exemplified by assertions of “complete cure” or “miracle treatment.”

Association between source and quality of health information short videos

Previous research have shown that the quality of health education videos is influenced by professional background of the author. ⁴⁰ Among the studies reviewed only videos related to heatstroke predominantly identified news agencies as the primary source (43.4%). In contrast, the other 24 categories of health-related video were primarily produced by healthcare professionals, such as doctors, medical practitioners, comprising between 40.7% and 98.38% (see Figure 5). Videos originating from medical institutions, science communicators, health professionals, and nonprofits organizations were significantly more accurate and reliable compared to those produced by for-profit organizations, patients, or general users. Consequently, the public is more likely to get access to trustworthy health information in short videos disseminated by healthcare professionals.

Figure 5.

Source of the online health information on Chinese short-video platforms.

Discussion

Features of health information quality on Chinese social media short videos and international comparisons

Most studies have employed assessment tools such as DISCERN, Goobie, and GQS, alongside metrics including video duration and user interaction data (e.g. likes, favorites, and shares), to evaluate content quality. These efforts provide a foundational reference for enhancing public health literacy and standardizing the dissemination of medical information. Drawing on findings from video quality research conducted on platforms such as YouTube and Facebook,^41–45 several consistent issues have been identified. Numerous social media platforms worldwide exhibited problems with suboptimal health information quality. Commonly used assessment tools heavily rely on expertise of raters, while cross-sectional studies only capture the quality characteristics of a small video sample at a single time point. Furthermore, compared to international platforms such as YouTube, Chinese short-video platforms impose duration limits, which may restrict the ability to convey comprehensive information within the allotted time.

In conclusion, the limitations associated with subjective assessment tools and the inherent constraints of cross-sectional studies represent significant impediments to the advancement of health-related short video evaluations. Therefore, the development of objective and automated evaluation tools, coupled with the implementation of long-term longitudinal studies, is essential for addressing current challenges.

Public health risks associated with misleading health information

The high share ability and low comprehension threshold of short videos render them a pivotal medium for the dissemination of health information. However, the pervasive circulation of misleading content presents several public health challenges. Firstly, misinformation directly distorts public health decision-making, potentially resulting in inappropriate lifestyle choices or the abandonment of standard medical treatments. Secondly, exaggerated health rumors can heighten public anxiety and distort perception of risk. Thirdly, the persistent dissemination of low-quality information undermines the credibility of authoritative medical institutions, diminished public acceptance of scientifically validated health information, and hinders the effective implementation of public health policies.

Government should establishing authoritative safeguards and regulatory frameworks

Firstly, it is imperative to establish a dynamically updated authoritative health information database. This initiative should be spearheaded by the National Health Commission, in collaboration with tertiary hospitals and research institutions. The database should cover critical domains such as disease prevention, nutrition, and mental health, providing standardized content for various platforms. It should also undergo quarterly reviews to ensure the information remains current. Secondly, comprehensive regulations governing health information on short video platforms must be issued, clarifying certification standards for content creators, requiring, for instance, that medical practitioners must provide licensure, institutions must submit administrative permits. Additionally, a negative list for content moderation should be established, prohibiting exaggerated claims about health products and the dissemination of false medical advice. A tiered penalty system should be instituted to enforce those regulations, with penalties ranging from content removal and account restrictions to administrative sanctions.

Platforms should integrate technological empowerment with operational optimization

Firstly, the development of a collaborative moderation system involving AI and expert-input should be established. By using the authoritative database, models for text and image recognition can be trained to prescreen health-related short videos, identifying and flagging potentially misleading terms like “complete cure” or “miracle treatment.” This process should be followed by expert review, thereby reducing the moderation cycle to within 24 h. Secondly, the mechanisms for distributing health information content should be optimized. Content originating from verified authoritative accounts should be prioritized with recommendation algorithms, clearly labeled as “authoritative source” in search results, and equipped with mechanisms to trace misinformation and deliver targeted clarifications. Platforms should also incentivize healthcare professionals, including clinicians and nutritionists, to engage in the creation of high-quality content through project-based incentives and traffic support. Thirdly, dedicated sections such as “Health Science Highlights” should be established. These sections would curate high-quality content categorized by disease type or target population (e.g. children, elderly, pregnant women) and integrate online consultation services.

Researchers should strengthen foundational studies and practical contributions

Efforts should focus on typological analysis of health misinformation to inform moderation strategies. This includes developing objective quality assessment systems specifically designed for short video formats, such as weighted models for accuracy, evidence sufficiency, and comprehensibility. Additionally, longitudinal research should be conducted in collaboration with platforms to monitor the impact of quality health content on public behaviors. This approach will provide evidence-based guidance for content creation.

Conclusions

This review identified a suboptimal overall quality of health information disseminated through short-video platforms in China, with misleading content frequently leading to irrational health decisions among the general public. Conversely, videos produced by professional healthcare personnel demonstrated significantly higher quality and credibility. These findings are consistent with studies conducted on platforms such as YouTube and TikTok in other countries.

Currently, most existing research focuses on evaluating the comprehensiveness of health information, while insufficient attention is given to in-depth analyses of the prevalence, typology, and underlying causes of health misinformation. To enhance and standardize the quality of health-related content on short-video platforms, this review proposes three key directions for future research: First, systematic research should be conducted to classify types of health misinformation in short videos, including cognitive misconceptions, deliberate disinformation, and outdated information. Second, the development of artificial intelligence-based real-time monitoring and early warning models for health information is essential to improve the precision and reliability of content dissemination. Third, longitudinal studies spanning one to two years are recommended to thoroughly identify the causal relationships between exposure to health-related short videos and changes in public health awareness and health behaviors.

Moreover, improving the quality of health related videos through a multifaceted approaches including as the optimization of content review process, rigorous verification of creator qualifications, and the refinement of algorithmic recommendation systems constitutes a critical advancement in guaranteeing public access to standardized digital health information.