Communicating COVID-19 information on TikTok: a content analysis of TikTok videos from official accounts featured in the COVID-19 information hub (2024)

Data collection

The TikTok COVID-19 information hub featured videos from official accounts of eight public health and United Nation agencies (see TableI). For each account, we reviewed all of its videos posted since November 2019 and identified COVID-19 videos. A COVID-19 video was a video that included at least one of the following hashtags, namely ‘COVID-19’, ‘coronavirus’, ‘GoAwayCorona’ and ‘GoAwayCoronaChallenge.’ For videos without any hashtags, the video content and captions were reviewed. Those explicitly mentioning ‘COVID-19’ or ‘coronavirus’ in the audio and/or texts were counted as COVID-19 videos. In total, 331 COVID-19 videos were identified and downloaded on 5 May 2020.

Coding scheme

We developed a codebook based on existing content analysis studies of TikTok [16, 17], prior research on video-based health communication [18, 19], health behavior theories, such as the health belief model and appraisal theory, and expert consensus. Expert consensus involved open coding [20], discussions and revisions of the codebook within the research team and review by our interdisciplinary team of collaborators with substantial social media content analysis expertise. The codebook focused on user engagement, video format, video type and video content. First, we recorded the number of views, likes, comments and shares of each video to quantify its user engagement. Second, we collected information on video format (see TableII), including video length, subtitle, text, spoken language, caption and music. Except for video length, each variable was coded as present (1) or absent (0). If a video included captions, the number of hashtags in the captions was recorded.

Next, we identified the video types and examined video content by coding the major themes, primary characters, emotions and COVID-19 risk information presented in the videos. The categories of video type (see TableIII) and theme (see TableIV) were developed based on an open coding process, in which research team members independently generated a list of video types and themes, compared the degree of overlap between the lists and developed the final coding categories through discussions and expert review. We coded the type of a TikTok video into one of the following categories, including acting, animated infographics, documentary, news, oral speech, pictorial slideshow and TikTok dance.

In terms of video content, each video was categorized as one of the following themes, including anti-stigma/anti-rumor, disease knowledge, encouragement, personal precautions, recognition, societal crisis management and work report. If a video included human subjects, we examined the characteristics of the primary characters, which are individuals who played the leading role in the video. The primary characters’ general occupation was coded into five categories: the general public, health professionals, public figures, essential workers/volunteers and others. COVID-19-related emotions were also coded. We explored whether alarm/concern, hope and humor were present (1) or absent (0) in the videos. In addition, we recorded whether the videos conveyed susceptibility and severity of COVID-19, effectiveness of precautionary measures (i.e. response efficacy) and the impacts of COVID-19 on the economy and people’s normal life (see TableV).

Video coding

Two coders received four 2-h training sessions. The coders coded five videos for each training session and discrepancies were discussed and resolved. Training videos were included in the final data analysis. To establish intercoder reliability, about 30% of videos (100/331, 30.21%) were randomly selected and independently coded by each coder. The intercoder reliability was high, with Krippendorff’s Alpha values ranging from 0.91 to 1.00 (see TablesII–V for intercoder reliability scores for each variable). The remaining videos were then evenly and randomly divided and independently coded by the two coders.

Descriptive analyses

As of 5 May 2020, the 331 COVID-19 videos received a total of 907 930 000 views, 29 640 000 likes, 168 880 comments and 781 862 shares (see TableI).

The videos ranged from 9 to 115 s in length (M = 32.87, SD = 17.63, median =29). Five videos were longer than 60 s. About 11.48% of videos included subtitles, 78.55% contained textual information in the video, 33.84% had a spoken language, 95.47% used captions and 89.73% had background music (see TableII). The language used in subtitles (37/38 videos), texts (255/260) and oral messages (98/112) were predominantly English. The majority (96.17%, 315) of videos had at least one hashtag (M = 5.01, SD = 2.36, median =5, range: 0–11). Seven types of COVID-19 TikTok videos were identified (see TableIII). About one in three videos were acting videos, one in five were documentaries and oral speeches. Other video types included animated infographics, TikTok dance, news and pictorial slideshow.

We identified seven categories of video themes (see TableIV). Videos focusing on personal precautions were the most prevalent, followed by encouragement, disease knowledge, anti-stigma/anti-rumor, social crisis management, recognition and work report videos. About 88.22% (292) of the videos had human characters, most of whom were the general public (138, 41.69%) or essential workers/volunteers (104, 31.4%). Nearly 10.88% (36) of the videos primarily focused on health professionals, and 4.23% (14) featured public figures, such as politicians, celebrities, social media influencers and business leaders.

We also examined the emotions and cognitive variables conveyed in the video content (see TableV). One in three videos expressed hope (32.63%). About the same number of videos expressed alarm/concern (12.99%) and humor (12.39%). While over one-third of videos conveyed the response efficacy of COVID-19 precautions (35.65%), 5.44% and 9.06% of videos mentioned the susceptibility and severity of COVID-19, respectively. About 8.16% of videos also mentioned the impact of COVID-19 on the economy and people’s normal life.

Substantial analyses

Correlation and ANOVA analyses explored how quantitative indicators of user engagement (numbers of views, likes, comments and shares) were related to different attributes of video format and content. Due to space limits, only significant results were reported.

First, we examined whether user engagement differed as a function of the accounts. A series of one-way ANOVA showed that the number of views, F(7, 323) = 5.29, P < 0.001; likes, F(7, 323) = 4.98, P < 0.001, comments, F(7, 323) = 8.02, P < 0.001 and shares, F(7, 323) = 6.24, P < 0.001; differed between accounts. Post hoc analysis using the Bonferroni correction indicated that videos posted by the World Health Organization (WHO) received significantly more views, likes, comments and shares than any other accounts. Thus, a dummy variable (0= accounts other than WHO, 1= WHO) was computed and controlled for in all subsequent data analyses.

The earlier a TikTok video was posted, the higher user engagement it may accumulate; thus, we correlated the number of days a video was posted (M = 38.20, SD = 19.74, range: 1–106) with the quantitative indicators of user engagement. Results showed that the number of posted days was positively correlated with the number of views, r(331) = 0.41, P < 0.001; likes, r(331) = 0.43, P < 0.001; comments, r(331) = 0.46, P < 0.001; and shares, r(331) = 0.15, P = 0.005. Thus, the number of posted days was controlled for in all subsequent data analyses.

A partial correlation analysis examined the associations between quantitative indicators of user engagement, video length and the number of hashtags. Controlling for accounts and the number of posted days, the number of hashtags was positively correlated with the number of likes, r(327) = 0.13, P = 0.018. Other user engagement measures were not significantly correlated.

A series of one-way ANOVA explored whether user engagement differed between videos with and without subtitles, texts, spoken language, captions and music. Results showed that controlling for accounts and the number of posted days, videos with subtitles (M = 4 830 000 SD = 10 250 000) received more shares than those without subtitles (M = 2 040 000 SD = 16 380 000), F(1, 327) = 6.98, P = 0.009. There were fewer comments for videos with spoken language (M = 448, SD = 890) than those without spoken language (M = 542, SD = 1152), F(1, 327) = 4.05, P = 0.045. Videos with spoken language (M = 1808, SD = 6298) also had fewer shares than those without spoken language (M = 2646, SD = 18 917), F(1, 327) = 4.32, P = 0.039.

A series of one-way ANOVA also explored whether quantitative indicators of user engagement differed between different categories of video type and video theme. We found that the number of shares was significantly different across the seven types of TikTok videos, F(6, 322) = 4.13, P = 0.001. Post hoc analysis using the Bonferroni correction indicated that TikTok dance videos (M = 15 687 SD = 59 999) were shared significantly more than news (M = 5467, SD = 10 374), pictorial slideshow (M = 3155, SD = 8194), documentary (M = 1486, SD = 4355), oral speech (M = 1132, SD = 4572), animated infographic (M = 1116, SD = 1896) and acting videos (M = 989, SD = 2917). No differences in video views, likes and comments were found across video types. There were also no significance differences in user engagement between different message theme categories.

Compared to videos without human characters (M = 1536, SD = 4721), videos with human characters (M = 8544, SD = 44 217) were more likely to be shared, F(1, 327) = 7.62, P = 0.006. The number of views, likes, comments and shares did not differ as a function of primary characters’ occupations. In terms of emotions, videos expressing alarm/concern (M = 1002, SD = 1501) received more comments than those without alarming emotions (M = 437, SD = 973), F(1, 327) = 4.24, P = 0.040. In contrast, hopeful videos (M = 1476, SD = 4657) had fewer views than those did not express hope (M = 3356, SD = 7954), F(1, 327) = 4.83, P = 0.029.

Compared to those without mentioning the susceptibility of COVID-19 (M = 82 963 SD = 209 290), those mentioning susceptibility (M = 203 837 SD = 471 419) received more likes, F(1, 327) = 3.89, P = 0.049. Videos mentioning the severity of COVID-19 (M = 970, SD = 1447) had more comments than those without severity information (M = 464, SD = 1017), F(1, 327) = 4.95, P = 0.027. Moreover, videos mentioning response efficacy of COVID-19 precautions (M = 4 920 000 SD = 9 830 000) received more views than those without response efficacy information (M = 1 540 000 SD = 4 590 000), F(1, 327) = 5.75, P = 0.017. There were more likes for videos with (M = 156 363 SD = 327 017) than without response efficacy information (M = 52 515 SD = 143 443), F(1, 327) = 4.30, P = 0.039. Videos mentioning response efficacy (M = 780, SD = 1363) also had more comments than those without response efficacy messages (M = 350, SD = 827), F(1, 327) = 10.90, P = 0.001.

As an alternative approach, regression analyses examined how video format variables (see TableII) and how emotions and cognitive variables (see TableV) all together predicted quantitative indicators of user engagement. Results were consistent with the ANOVA results.

Video format and user engagement

Almost every TikTok video had a hashtag, and a higher number of hashtags was related to more video likes. Social media research has shown that hashtag use assists users to seek, follow and share information [21, 22]. Health campaign-specific hashtags also help increase the credibility and dissemination of campaign messages [23, 24]. Similar to our results, a study of health advocacy tweets found that a higher number of hashtags predicted more retweets [25]. Thus, hashtags on TikTok seem to share similar functions of those on other social media platforms and should be utilized strategically to attract viewers and promote user engagement. Due to the exploratory nature of this study, we did not examine the content of the hashtags. Future research should examine whether user engagement differs as a function of different types of hashtags, such as between #GoAwayCorona and #coronaviurs.

About 1 in 10 videos included subtitles. Post hoc analyses showed that subtitles were predominately used in news and oral speech videos. In addition, videos with subtitles received more shares than those without subtitles. This may be because subtitle enables viewers to have a better understanding of video content [26], which may promote content sharing. In contrast, TikTok videos with spoken language attracted fewer comments and shares than those without spoken language. A recent study on government-generated health communication via TikTok showed that users preferred easily understood messages and entertaining videos than scientific jargon and overly formal presentations [16]. Delivering information via spoken language may be perceived as more formal than non-verbal performances like dances and thus, invite less user engagement. Future studies should further examine how the interactions of different video format features affect user engagement.

Video type and user engagement

Seven different types of COVID-19 videos emerged from the data, including acting, animated infographic, documentary, news, oral speech, pictorial slideshow and TikTok dance. The first five types were also identified in general public health communication on TikTok, with documentary and animated infographics being most frequently viewed [16]. We found that while the number of views, likes and comments did not differ as a function of video types, TikTok dance videos had more shares than any other types of videos. Audience-centered interactions are crucial for public health communication via social media [27]. Compared to other video types, TikTok dance videos are often featured with rhythmic and earworm music and easy-to-mimic lip-sync and body movement, all of which may encourage audience-centered interactions and increase shares. We encourage future studies to explicitly focus on public health communication via TikTok dances.

Video content and user engagement

We also identified seven major themes of COVID-19 videos, including anti-stigma/anti-rumor, disease knowledge, encouragement, personal precautions, recognition, societal crisis management and work report. A content analysis of government social media during the COVID-19 crisis revealed similar message content, including the latest updates about COVID-19, appreciation to front-line emergency services, guidance for stakeholders and information about the government’s handling of the crisis [28]. Moreover, compared to appreciation posts, latest pandemic updates and government crisis response lead to increased user engagement [28], as different message themes may meet community members’ demands in different ways [29, 30]. However, our study found no differences in video views, likes, comments and shares across the seven message themes. Studies with larger samples should continue to explore different message themes of COVID-19 across various social media platforms.

TikTok videos expressing alarm/concern received more comments than those without expressing the emotions. Videos mentioning susceptibility and severity of COVID-19 received more likes and comments, respectively. The results suggest that risk information may increase user engagement. Indeed, research on health information engagement demonstrated that risk information content increased likes, comments and shares of health communication posts [31]. Moreover, during infectious disease outbreaks, social media use predicts individuals’ prevention behaviors through risk perceptions and negative emotions [32]. Our results suggest that risk information and negative emotions in social media messages may be the underlying mechanisms accounting for the relationships between social media use and individuals’ risk perceptions and negative feelings. Finally, response efficacy of COVID-19 precautions attracted greater views, likes and comments. During the COVID-19 pandemic, community members are experiencing heightened uncertainty and fear. Videos mentioning the effectiveness of precaution measures help people reduce uncertainty and boost their confidence [2] and thus, have higher user engagement. In short, risk information and response efficacy are crucial for user engagement.

Practical implications

Our findings provide practical implications for public health agencies to use TikTok to communicate COVID-19 and health information in general. First, including more hashtags invited more likes for videos. Campaign-specific hashtags have also facilitated the dissemination of health communication messages. Thus, public health agencies can create various hashtags to summarize the video content and promote user engagement. Second, due to its audience-center features, TikTok dance videos are more engaging and easier to ‘duet’ than other types of videos. When possible, public health professionals are encouraged to communicate health information via TikTok dances. The viral handwashing dance suggests that TikTok dances are a feasible tool to demonstrate and ultimately promote health behaviors, such as washing hands, wearing face covers and social distancing. Third, TikTok videos mentioning risk information of COVID-19 and response efficacy of precautionary measures have higher user engagement. Thus, health message designers should consider incorporating those elements in their message design.

Limitations and future research

There are several limitations. First, we collected COVID-19 videos from official accounts featured in the information hub. While this approach helps ensure the credibility of the videos, the majority of TikTok videos are created by ordinary community members. The content, format and popularity of user-generated COVID-19 videos remain unstudied. Future research should examine COVID-19 information and misinformation created by average users. Moreover, this study is cross-sectional and explanatory in nature. While numbers of views, likes, comments and shares are common quantitative indicators of user engagement in social media studies [16, 17], the comment and share functions of a few TikTok videos were disabled by the accounts and thus, may not reflect the actual user engagement. Moreover, those measures do not indicate the message effect of TikTok videos on viewers’ health literacy, attitudes or behavioral change.

User comments were also not analyzed and thus, we do not know the number of comments that indicates approval or disapproval of the messages in the TikTok videos. Future studies are needed to examine users’ demographics and their cognitive, emotional and behavioral responses to TikTok health communication videos. In addition, epidemiology and our understanding of COVID-19 have also been changing over time. This study captures videos in the early stage of the pandemic and thus, reflects video content, format and popularity from that period only. Additional research is needed to examine how COVID-19 videos and user engagement change as community members become more knowledgeable about the COVID-19 pandemic progresses.

1. World Health Organization. Who Coronavirus (COVID-19) Dashboard, 2020. https://covid19.who.int. Accessed: 12 February 2021.

2. Finset A, Bosworth H, Butow P et al. Effective health communication – a key factor in fighting the COVID-19 pandemic. Patient Educ Couns 2020; 103:873–6. [PMC free article] [PubMed] [Google Scholar]

3. Neiger BL, Thackeray R, Van Wagenen SA et al. Use of social media in health promotion: purposes, key performance indicators, and evaluation metrics. Health Promot Pract 2012; 13:159–64. [PubMed] [Google Scholar]

4. Ramanadhan S, Mendez SR, Rao M et al. Social media use by community-based organizations conducting health promotion: a content analysis. BMC Public Health 2013; 13:1129. [PMC free article] [PubMed] [Google Scholar]

5. Conrad EJ, Becker M, Powell B et al. Improving health promotion through the integration of technology, crowdsourcing, and social media. Health Promot Pract 2020; 21:228–37. [PubMed] [Google Scholar]

10. Castagnoli R, Votto M, Licari A et al. Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection in children and adolescents: a systematic review. JAMA Pediatr 2020; 174:882. [PubMed] [Google Scholar]

11. Tenforde MW, Kim SS, Lindsell CJ et al.; IVY Network Investigators. Symptom duration and risk factors for delayed return to usual health among outpatients with COVID-19 in a multistate health care systems network — United States. MMWR 2020; 69:993–8. [PMC free article] [PubMed] [Google Scholar]

12. Adams SH, Park MJ, Schaub JP et al. Medical vulnerability of young adults to severe COVID-19 illness-data from the national health interview survey. J Adolesc Health 2020; 67:362–8. [PMC free article] [PubMed] [Google Scholar]

16. Zhu C, Xu X, Zhang W et al. How health communication via TikTok makes a difference: a content analysis of TikTok accounts run by Chinese provincial health committees. Int J Environ Res Public Health 2019; 17:192. [PMC free article] [PubMed] [Google Scholar]

17. Serrano JCM, Papakyriakopoulos O, Hegelich S. Dancing to the partisan beat: a first analysis of political communication on TikTok. In: Proceedings of the 12th ACM Conference on Web Science. 2020, 257–66.

18. Gabarron E, Fernandez-Luque L, Armayones M et al. Identifying measures used for assessing quality of YouTube videos with patient health information: a review of current literature. Interact J Med Res 2013; 2:e6. [PMC free article] [PubMed] [Google Scholar]

19. Massey ZB, Li Y, Holli J et al. Modifications to electronic nicotine delivery systems: content analysis of youtube videos. J Med Internet Res 2020; 22:e17104. [PMC free article] [PubMed] [Google Scholar]

20. Krippendorff K. Content Analysis: An Introduction to Its Methodology. London: Sage, 2004. [Google Scholar]

21. Small TA. What the hashtag? Inf Commun Soc 2011; 14:872–95. [Google Scholar]

22. Chang H-C. A new perspective on Twitter hashtag use: diffusion of innovation theory. Proc Am Soc Inf Sci Technol 2010; 47:1–4. [Google Scholar]

23. Chung JE. A smoking cessation campaign on Twitter: understanding the use of Twitter and identifying major players in a health campaign. J Health Commun 2016; 21:517–26. [PubMed] [Google Scholar]

24. George N, Britto DR, Krishnan V et al. Assessment of hashtag (#) campaigns aimed at health awareness in social media. J Educ Health Promot 2018; 7:114. [PMC free article] [PubMed] [Google Scholar]

25. Saxton GD, Niyirora J, Guo C et al. The strategic use of hashtags in social media advocacy. Adv Soc Work 2015; 16:154–169. [Google Scholar]

26. Perego E, Del Missier F, Porta M et al. The cognitive effectiveness of subtitle processing. Media Psychol 2010; 13:243–272. [Google Scholar]

27. Heldman AB, Schindelar J, Weaver JB. Social media engagement and public health communication: implications for public health organizations being truly “social”. Public Health Rev 2013; 35:13. [Google Scholar]

28. Chen Q, Min C, Zhang W et al. Unpacking the black box: how to promote citizen engagement through government social media during the COVID-19 crisis. Comput Human Behav 2020; 110:106380. [PMC free article] [PubMed] [Google Scholar]

29. Wang Y, Wang T, Ye X et al. Using social media for emergency response and urban sustainability: a case study of the 2012 Beijing rainstorm. Sustainability 2015; 8:25. [Google Scholar]

30. Keib K, Himelboim I, Han J-Y. Important tweets matter: predicting retweets in the #blacklivesmatter talk on Twitter. Comput Human Behav 2018; 85:106–115. [Google Scholar]

31. Rahim AI, Ibrahim MI, Salim FNA et al. Health information engagement factors in Malaysia: a content analysis of Facebook use by the ministry of health in 2016 and 2017. Int J Environ Res Public Health 2019; 16:591. [PMC free article] [PubMed] [Google Scholar]

32. Oh S-H, Lee SY, Han C. The effects of social media use on preventive behaviors during infectious disease outbreaks: the mediating role of self-relevant emotions and public risk perception. Health Commun 2020; 1–10. 10.1080/10410236.2020.1724639. Accessed: 12 February 2021. [PubMed] [CrossRef] [Google Scholar]