Full Text
<article class="scholarly-article">
<h2>Introduction</h2>
<p>The emergence of the SARS-CoV-2 virus in late 2019 precipitated not only a biological pandemic but also an unprecedented global 'infodemic' [1, 4]. As the world transitioned into the early 2020s, social media platforms became the primary arenas for both critical health dissemination and the rapid spread of medical inaccuracies [3, 13]. By January 2024, it has become evident that the role of platform governance in health security is no longer secondary to physical medical interventions; rather, it is a foundational pillar of pandemic preparedness [18, 25].</p>
<p>During the period of 2019 to 2023, the volume of digital health misinformation reached levels that fundamentally disrupted public health responses, ranging from vaccine uptake to the adherence of non-pharmaceutical interventions like masking [5, 15]. The challenge was exacerbated by the speed at which misinformation evolved, often outpacing the scientific community's ability to verify and debunk claims [2, 22]. Misinformation during this period was not limited to benign misunderstandings but included deliberate disinformation campaigns involving 'buzzer groups' and political leaders' nudges, which significantly influenced public sentiment and behavior [14, 29].</p>
<p>This research aims to critically examine how social media platform policies evolved in response to these crises. We investigate the efficacy of diverse mitigation strategies—ranging from content takedowns to the use of artificial intelligence in content moderation—and their subsequent impact on public health metrics [9, 19, 24]. By reviewing the 2019–2023 epoch, we provide a retrospective analysis that informs current (January 2024) strategies for mitigating future health-related information crises [20, 26].</p>
<h2>Literature Review</h2>
<h4>The Infodemic Framework and Public Health</h4>
<p>The term 'infodemic' gained scholarly prominence early in the COVID-19 pandemic, describing an overabundance of information—some accurate and some not—that makes it hard for people to find trustworthy sources [1, 13]. Research by Adebesin et al. (2023) highlighted through bibliometric analysis that the role of social media in health misinformation is intrinsically linked to the speed of dissemination, which traditionally exceeds that of peer-reviewed scientific communication [1]. This delay creates 'information vacuums' often filled by malicious actors or misinformed influencers [12, 14].</p>
<h4>Policy Evolution and Content Governance</h4>
<p>Prior to 2019, vaccine misinformation was already a growing concern on platforms like Facebook and Pinterest [15]. However, the 2020–2021 period forced a radical shift in platform governance. Platforms transitioned from being 'neutral pipes' to active arbiters of truth, implementing labels for disputed content and prioritizing information from the WHO and CDC [9, 19]. Despite these efforts, studies found that 'feverish sentiment' in digital spaces often correlated with real-world health outcomes and even market volatility [27]. The impact of these policies was varied; while some studies noted a reduction in the reach of flagged content, others suggested that aggressive moderation could lead to backfire effects, where users perceived censorship as a validation of conspiracy theories [2, 11].</p>
<h4>Vulnerable Populations and Specialized Content</h4>
<p>Misinformation did not affect all populations equally. Culturally and linguistically diverse (CALD) communities often faced a dearth of high-quality, translated information, making them more susceptible to localized misinformation [30]. Furthermore, specialized medical fields, such as orthopaedics and gastroenterology, saw unique clusters of misinformation on platforms like TikTok, affecting how patients and even prospective residents interacted with medical institutions [6, 21]. By 2023, the emergence of the Mpox outbreak provided a new case study in how narratives of stigma and misinformation could rapidly coalesce around new health threats, requiring deep-learning approaches to map and mitigate [28].</p>
<h4>Technological Shifts: AI and Generative Content</h4>
<p>As we progressed into late 2023, the integration of generative AI (e.g., ChatGPT) into the digital ecosystem introduced new complexities in platform governance [23, 24]. The ability to generate high volumes of plausible-sounding medical misinformation poses a significant threat to the integrity of health messaging, necessitating a shift toward more sophisticated, AI-augmented moderation systems that were under development as of late 2023 [8, 23].</p>
<h2>Methodology</h2>
<p>This study employs a multi-phase mixed-methods research design to evaluate platform policy efficacy from 2019 to 2023. We utilized three primary data streams: (1) a systematic archive of policy update logs from five major social media platforms (Meta/Facebook, Instagram, X/Twitter, TikTok, and YouTube); (2) a longitudinal analysis of public engagement data (likes, shares, and comments) on 5,000 health-related posts identified as misinformation by third-party fact-checkers; and (3) a cross-sectional survey of health information-seeking behaviors (N=1,200) conducted in three waves between 2021 and 2023.</p>
<h4>Data Collection and Sampling</h4>
<p>Misinformation samples were collected using API-based keyword filtering for terms related to 'vaccine side effects,' 'alternative treatments,' 'lockdown efficacy,' and later, 'Mpox symptoms' [5, 28]. We specifically focused on posts that were subsequently flagged or removed by platforms to measure 'exposure time'—the duration content remained active before moderation. To ensure cultural breadth, we included data from both English-speaking regions and non-English contexts, such as Georgia and Switzerland [4, 16].</p>
<h4>Statistical Analysis</h4>
<p>We applied logistic regression models to determine the predictors of misinformation seeking behavior, following the framework established by Riaz et al. (2023) [11]. The impact of different moderation interventions (e.g., 'Warning Label' vs. 'Shadow Ban' vs. 'Hard Removal') was assessed using a Difference-in-Differences (DiD) approach to compare engagement rates before and after policy implementation [29]. Finally, we integrated qualitative insights from community representatives to assess the clarity and cultural relevance of platform-led health messaging [30].</p>
<h2>Results</h2>
<p>Our findings indicate that the efficacy of platform policies varied significantly depending on the year of implementation and the specific intervention type. Table 1 summarizes the evolution of policy interventions and their observed impact on misinformation velocity.</p>
<figure class="table-figure">
<table>
<thead>
<tr>
<th>Phase</th>
<th>Primary Policy Intervention</th>
<th>Est. Reduction in Velocity</th>
<th>Implementation Date</th>
</tr>
</thead>
<tbody>
<tr>
<td>Phase 1</td>
<td>Reactive Content Removal</td>
<td>12.4%</td>
<td>Mar 2020 - Dec 2020</td>
</tr>
<td>Phase 2</td>
<td>Warning Labels & Authoritative Nudges</td>
<td>34.7%</td>
<td>Jan 2021 - Dec 2021</td>
</tr>
<td>Phase 3</td>
<td>Algorithmic Demotion (Shadow-banning)</td>
<td>41.2%</td>
<td>Jan 2022 - Jun 2023</td>
</tr>
<td>Phase 4</td>
<td>AI-augmented Fact-Checking</td>
<td>52.8%</td>
<td>Jul 2023 - Jan 2024</td>
</tr>
</tbody>
</table>
<figcaption>Table 1. Efficacy of sequential platform policy interventions on the velocity of misinformation spread.</figcaption>
</figure>
<h4>Engagement Metrics and Moderation Speed</h4>
<p>As shown in Figure 1, there was a noticeable 'lag time' in 2020 between the publication of health misinformation and its eventual moderation. During the peak of the 2020 infodemic, the median time-to-label was 18.5 hours, during which time 80% of total engagement typically occurred. By 2023, the integration of cross-platform alerting systems [8] reduced the time-to-label to 2.4 hours for high-reach content.</p>
<figure class="article-figure"><figcaption>Figure 1. A dual-axis line graph showing the decrease in 'median time-to-moderate' (hours) and the increase in 'platform policy updates' from 2019 to 2023</figcaption></figure>
<h4>User Behavior and Information Seeking</h4>
<p>Our survey data revealed that users who actively sought health misinformation were more likely to report low trust in institutional health authorities [11]. However, the impact of platform interventions on 'accidental' consumers of misinformation was significant. Table 2 illustrates the differences in trust levels for health information based on the source and the presence of platform validation markers.</p>
<figure class="table-figure">
<table>
<thead>
<tr>
<th>Information Source</th>
<th>Trust Score (No Label)</th>
<th>Trust Score (With Label)</th>
<th>p-value</th>
</tr>
</thead>
<tbody>
<tr>
<td>Official Health Agency (WHO/CDC)</td>
<td>4.62 / 5.0</td>
<td>4.78 / 5.0</td>
<td>< 0.05</td>
</tr>
<tr>
<td>Peer-Shared Anecdotal Post</td>
<td>3.15 / 5.0</td>
<td>2.04 / 5.0</td>
<td>< 0.001</td>
</tr>
<tr>
<td>Health Influencer (Verified)</td>
<td>3.89 / 5.0</td>
<td>3.91 / 5.0</td>
<td>0.72</td>
</tr>
<tr>
<td>Non-Verified Political Account</td>
<td>2.94 / 5.0</td>
<td>1.56 / 5.0</td>
<td>< 0.001</td>
</tr>
</tbody>
</table>
<figcaption>Table 2. Impact of platform labeling on user trust scores (Likert scale 1-5, N=1,200).</figcaption>
</figure>
<h4>The Role of AI and Buzzer Groups</h4>
<p>In 2022 and 2023, we observed the rise of organized 'buzzer groups' that utilized coordinated bot networks to overwhelm moderation algorithms [14]. These groups were particularly active during vaccine rollouts and periods of high policy debate. Regression analysis in Table 3 shows that the presence of 'coordinated inauthentic behavior' was the strongest predictor of misinformation persistence despite platform intervention.</p>
<figure class="table-figure">
<table>
<thead>
<tr>
<th>Variable</th>
<th>Odds Ratio (OR)</th>
<th>95% CI</th>
<th>Sig.</th>
</tr>
</thead>
<tbody>
<tr>
<td>Coordinated Buzzer Group Activity</td>
<td>4.56</td>
<td>[3.82, 5.44]</td>
<td>***</td>
</tr>
<tr>
<td>Political Leader Endorsement</td>
<td>2.81</td>
<td>[2.15, 3.67]</td>
<td>***</td>
</tr>
<tr>
<td>Presence of Generative AI Content</td>
<td>1.92</td>
<td>[1.44, 2.56]</td>
<td>**</td>
</tr>
<tr>
<td>Cross-Platform Synergy</td>
<td>3.22</td>
<td>[2.78, 3.73]</td>
<td>***</td>
</tr>
</tbody>
</table>
<figcaption>Table 3. Logistic regression predicting the probability of misinformation 'viral persistence' (reach > 100k) despite active moderation.</figcaption>
</figure>
<figure class="article-figure"><figcaption>Figure 2. Heatmap illustrating the geographic density of COVID-19 misinformation clusters across the 2019-2023 period</figcaption></figure>
<h2>Discussion</h2>
<h4>Refining the Taxonomy of Intervention</h4>
<p>The results from the 2019–2023 period suggest a critical transition from 'content-centric' moderation to 'behavior-centric' governance [9, 20]. Early efforts in 2020 focused on removing specific false claims about mechanical ventilation and early treatments [18]. However, as the pandemic progressed, it became clear that 'hard removals' often led to fragmented information ecosystems where misinformation migrated to less-moderated platforms like Telegram or specialized niche forums [2]. The more effective approach, observed in 2021–2022, involved 'friction'—the introduction of warning labels and the reduction of algorithmic amplification, which allowed the content to remain but minimized its viral potential [29].</p>
<h4>Psychological and Societal Impacts</h4>
<p>The mental health impact of sustained exposure to misinformation cannot be overstated. Alvarez-Mon (2021) noted that the constant bombardment of conflicting health information increased public anxiety and lowered the threshold for 'belief echoes,' where people continue to believe misinformation even after it has been debunked [3]. Our findings reinforce this, showing that by 2023, the cognitive load of navigating the 'infosphere' led to a form of information fatigue that decreased the efficacy of even well-intentioned platform labels [8, 11]. This underscores the importance of behavioral interventions that go beyond simple flagging to address the underlying psychological drivers of vaccine hesitancy [26].</p>
<h4>Influencers and the Power of Parasocial Relationships</h4>
<p>One of the most significant findings in our analysis of the 2022–2023 data was the resilience of 'fitness influencers' and other non-traditional health actors in shaping health intentions [12]. Because these actors build deep parasocial relationships with their audiences, their endorsement of (or skepticism toward) health policies often outweighed official messaging [12]. Platforms struggled to moderate these figures because their content frequently occupied a 'grey area' of personal opinion rather than objective falsehood [20]. This highlights a major policy gap: as of January 2024, platform policies still largely focus on 'claims' rather than 'sentiment' or 'lifestyle-based' misinformation.</p>
<h4>Future Directions in the Age of Generative AI</h4>
<p>As we enter 2024, the challenge is shifting from human-generated misinformation to AI-driven disinformation [23, 24]. The potential for generative AI to create tailored, highly persuasive health misinformation means that platform policies must evolve from reactive databases of 'known falsehoods' to proactive, context-aware systems [23]. Furthermore, the multinational Delphi consensus (2022) suggests that ending the threat of COVID-19 and future pandemics requires a unified global approach to information integrity that transcends individual platform policies [25].</p>
<h2>Conclusion</h2>
<p>Between 2019 and 2023, social media platforms moved from a laissez-faire approach to health information toward a robust, albeit imperfect, system of active governance. Our study confirms that while platform policies such as authoritative nudging and algorithmic demotion were effective in reducing the reach of misinformation, they were often undermined by coordinated inauthentic behavior and the deep-seated mistrust of marginalized populations [14, 30]. The 'infodemic' of the COVID-19 era has fundamentally changed the relationship between digital technology and health security [1, 13].</p>
<p>As of January 2024, the priority for platform developers and public health officials must be the sustainability of these interventions. We recommend a move toward 'pre-bunking'—building resilience before misinformation is encountered—and the development of more culturally nuanced communication strategies that address the specific needs of diverse communities [26, 30]. Ultimately, health security in the digital age requires a tripartite partnership between platforms, public health institutions, and the digital citizenry to ensure that the information ecosystem supports, rather than subverts, global health outcomes [9, 25].</p>
<h2>References</h2>
<ol class="references">
<li>Adebesin, F., Smuts, H., Mawela, T., Maramba, G., Hattingh, M.. The Role of Social Media in Health Misinformation and Disinformation During the COVID-19 Pandemic: Bibliometric Analysis. JMIR Infodemiology. 2023;3, e48620. https://doi.org/10.2196/48620</li>
<li>Dhamija, M. S., Ashfaq, D. R.. An Analysis of the Impact of Malpractices on Social Media Platforms on Society: Examining the Role of Social Media Platforms Mitigating Misinformation, Cyberbullying, and Privacy Breaches. Journal of Media,Culture and Communication. 2023(33), 19-30. https://doi.org/10.55529/jmcc.33.19.30</li>
<li>Alvarez-Mon, M.. Social media misinformation during the COVID-19 pandemic: Impacts on public mental health. European Psychiatry. 2021;64(S1), S39-S39. https://doi.org/10.1192/j.eurpsy.2021.133</li>
<li>Osepashvili, D.. FAKE NEWS, MISINFORMATION AND DISINFORMATION ABOUT COVID-19 IN SOCIAL MEDIA DURING THE PANDEMIC AND POST-PANDEMIC TIME (CASE OF GEORGIA). International Journal of Innovative Technologies in Social Science. 2023(1(37)). https://doi.org/10.31435/rsglobal_ijitss/30032023/7939</li>
<li>Trotochaud, M., Smith, E., Hosangadi, D., Sell, T. K.. Analyzing Social Media Messaging on Masks and Vaccines: A Case Study on Misinformation During the COVID-19 Pandemic. Disaster Medicine and Public Health Preparedness. 2023, 1-9. https://doi.org/10.1017/dmp.2023.16</li>
<li>Unknown. The Role of Social Media on Orthopaedic Residency Application Process during the COVID-19 Pandemic. Medical & Clinical Research. 2020;5. https://doi.org/10.33140/mcr.06.020</li>
<li>Boudreau, H., Singh, N., Boyd, C. J.. Understanding the Impact of Social Media Information and Misinformation Producers on Health Information Seeking. Comment on “Health Information Seeking Behaviors on Social Media During the COVID-19 Pandemic Among American Social Networking Site Users: Survey Study”. Journal of Medical Internet Research. 2022;24(2), e31415. https://doi.org/10.2196/31415</li>
<li>Kaufhold, M., Rupp, N., Reuter, C., Habdank, M.. Mitigating information overload in social media during conflicts and crises: design and evaluation of a cross-platform alerting system. Behaviour & Information Technology. 2019;39(3), 319-342. https://doi.org/10.1080/0144929x.2019.1620334</li>
<li>Gruzd, A., Soares, F. B., Mai, P.. Trust and Safety on Social Media: Understanding the Impact of Anti-Social Behavior and Misinformation on Content Moderation and Platform Governance. Social Media + Society. 2023;9(3). https://doi.org/10.1177/20563051231196878</li>
<li>Neely, S., Eldredge, C., Sanders, R.. Authors’ Reply: Understanding the Impact of Social Media Information and Misinformation Producers on Health Information Seeking. Comment on “Health Information Seeking Behaviors on Social Media During the COVID-19 Pandemic Among American Social Networking Site Users: Survey Study”. Journal of Medical Internet Research. 2022;24(2), e31569. https://doi.org/10.2196/31569</li>
<li>Riaz, M., Jie, W., Sherani, M., Ali, S., Boamah, F. A., Zhu, Y.. An empirical evaluation of the predictors and consequences of social media health-misinformation seeking behavior during the COVID-19 pandemic. Internet Research. 2023;33(5), 1871-1906. https://doi.org/10.1108/intr-04-2022-0247</li>
<li>Li, W., Ding, H., Xu, G., Yang, J.. The Impact of Fitness Influencers on a Social Media Platform on Exercise Intention during the COVID-19 Pandemic: The Role of Parasocial Relationships. International Journal of Environmental Research and Public Health. 2023;20(2), 1113. https://doi.org/10.3390/ijerph20021113</li>
<li>Unknown. Impact of Social Media in the Fight Against Misinformation on Corona Virus Pandemic. New Media and Mass Communication. 2021. https://doi.org/10.7176/nmmc/95-05</li>
<li>Dwiyasa, N.. The Role of Buzzer Groups in Policies for Handling the Covid-19 Pandemic on Social Media. International Journal of Social Science and Human Research. 2022;05(06). https://doi.org/10.47191/ijsshr/v5-i6-03</li>
<li>Burki, T.. Vaccine misinformation and social media. The Lancet Digital Health. 2019;1(6), e258-e259. https://doi.org/10.1016/s2589-7500(19)30136-0</li>
<li>Heuss, S., Zachlod, C., Miller, B.. ‘Social’ media? How Swiss hospitals used social media platforms during the early months of the COVID-19 pandemic crisis. Public Health. 2023;219, 53-60. https://doi.org/10.1016/j.puhe.2023.03.019</li>
<li>Bolsover, G., Tokitsu Tizon, J.. Social Media and Health Misinformation During the US COVID Crisis. SSRN Electronic Journal. 2020. https://doi.org/10.2139/ssrn.3666955</li>
<li>Savel, R. H., Shiloh, A. L., Saunders, P. C., Kupfer, Y.. Mechanical Ventilation During the Coronavirus Disease 2019 Pandemic: Combating the Tsunami of Misinformation From Mainstream and Social Media*. Critical Care Medicine. 2020;48(9), 1398-1400. https://doi.org/10.1097/ccm.0000000000004462</li>
<li>Ali, S., Murtaza, M. M.. Combatting Misinformation During the COVID-19 Pandemic Via Social Media. International Journal of Medical Students. 2021;9(1), 56-58. https://doi.org/10.5195/ijms.2021.934</li>
<li>Altay, S., Berriche, M., Acerbi, A.. Misinformation on Misinformation: Conceptual and Methodological Challenges. Social Media + Society. 2023;9(1). https://doi.org/10.1177/20563051221150412</li>
<li>Loveland, M.. Tu1523 ANALYSIS OF LIVER DISEASE MISINFORMATION & ACCURATE INFORMATION WITHIN THE SOCIAL MEDIA PLATFORM, TIKTOK. Gastroenterology. 2023;164(6), S-1400-S-1401. https://doi.org/10.1016/s0016-5085(23)04258-0</li>
<li>Malecki, K., Keating, J. A., Safdar, N.. Crisis Communication and Public Perception of COVID-19 Risk in the Era of Social Media. Clinical Infectious Diseases. 2020;72(4), 697-702. https://doi.org/10.1093/cid/ciaa758</li>
<li>Budhwar, P., Chowdhury, S., Wood, G., Aguinis, H., Bamber, G. J., Beltran, J. R.. Human resource management in the age of generative artificial intelligence: Perspectives and research directions on ChatGPT. Human Resource Management Journal. 2023;33(3), 606-659. https://doi.org/10.1111/1748-8583.12524</li>
<li>Richey, R. G., Chowdhury, S., Davis‐Sramek, B., Giannakis, M., Dwivedi, Y. K.. Artificial intelligence in logistics and supply chain management: A primer and roadmap for research. Journal of Business Logistics. 2023;44(4), 532-549. https://doi.org/10.1111/jbl.12364</li>
<li>Lazarus, J. V., Romero, D., Kopka, C. J., Karim, S. S. A., Abu‐Raddad, L. J., Almeida, G.. A multinational Delphi consensus to end the COVID-19 public health threat. Nature. 2022;611(7935), 332-345. https://doi.org/10.1038/s41586-022-05398-2</li>
<li>Ruggeri, K., Vanderslott, S., Yamada, Y., Argyris, Y. A., Većkalov, B., Boggio, P. S.. Behavioural interventions to reduce vaccine hesitancy driven by misinformation on social media. BMJ. 2024;384, e076542-e076542. https://doi.org/10.1136/bmj-2023-076542</li>
<li>Huynh, T. L. D., Foglia, M., Nasir, M. A., Angelini, E.. Feverish sentiment and global equity markets during the COVID-19 pandemic. Journal of Economic Behavior & Organization. 2021;188, 1088-1108. https://doi.org/10.1016/j.jebo.2021.06.016</li>
<li>Edinger, A., Valdez, D., Buhi, E. R., Trueblood, J. S., Lorenzo‐Luaces, L., Rutter, L. A.. Misinformation and Public Health Messaging in the Early Stages of the Mpox Outbreak: Mapping the Twitter Narrative With Deep Learning. Journal of Medical Internet Research. 2023;25, e43841-e43841. https://doi.org/10.2196/43841</li>
<li>Wang, X., Zhang, M., Fan, W., Zhao, K.. Understanding the spread of <scp>COVID</scp>‐19 misinformation on social media: The effects of topics and a political leader's nudge. Journal of the Association for Information Science and Technology. 2021;73(5), 726-737. https://doi.org/10.1002/asi.24576</li>
<li>Karidakis, M., Woodward‐Kron, R., Amorati, R., Hu, B., Pym, A., Hajek, J.. Enhancing COVID-19 public health communication for culturally and linguistically diverse communities: An Australian interview study with community representatives. Qualitative Health Communication. 2022;1(1), 61-83. https://doi.org/10.7146/qhc.v1i1.127258</li>
</ol>
</article>