Health Information and Social Media: How Platforms Handle Contested Claims and What Consumers Should Know

In 2019, YouTube removed thousands of videos from health-related channels, including Natural News, which had accumulated over 1,700 videos on topics ranging from herbal remedies to vaccine skepticism. The mass removal was part of a broader platform crackdown on what YouTube termed “medically unsubstantiated” health claims. By 2025, the same platform had reversed course, allowing previously banned users to create new accounts and re-upload content that had contributed to their termination, provided the content did not violate the platform’s current (more permissive) rules.

This reversal illustrates the fundamental instability of social media content moderation as a mechanism for managing health information. Platforms have oscillated between aggressive enforcement and permissive free-expression policies, often in response to political pressure rather than scientific consensus. This article examines how major social media platforms handle health claims, what the shift away from fact-checking means for public health, and what consumers should know about evaluating health information online.

How Social Media Platforms Handle Health Content

Each major platform has developed its own approach to health-related content, and these approaches have changed significantly since 2020.

Meta (Facebook, Instagram, WhatsApp)

Meta operated a third-party fact-checking program from 2016 to January 2025, using independent fact-checkers certified through the International Fact-Checking Network (IFCN) to review and label potentially false claims. When a post was rated “False” by fact-checkers, it was not removed but was labeled with a warning, had its distribution reduced by the algorithm, and was paired with links to authoritative sources. Users who repeatedly shared content rated as false faced reduced visibility for their posts.

In January 2025, Meta CEO Mark Zuckerberg announced the end of the third-party fact-checking program in the United States, replacing it with a “Community Notes” system modeled on the approach used by X (formerly Twitter). Under this system, users themselves add contextual notes to posts they believe are misleading. Notes are only displayed if they receive agreement from users across the political spectrum, using a bridging algorithm designed to surface notes that people with different viewpoints both find helpful.

The shift was widely interpreted as a response to political pressure. Republican lawmakers had investigated whether the Biden administration had pressured tech companies to remove content, and the Supreme Court reviewed these allegations in Murthy v. Missouri (2024), ruling that the plaintiffs lacked standing without determining whether the content removals violated free speech. Meta’s announcement explicitly cited a desire to reduce “censorship” and restore “free expression.”

YouTube

YouTube’s health content policies have undergone the most dramatic swings of any major platform. During the COVID-19 pandemic, YouTube removed over 1 million videos for violating its medical misinformation policies, which prohibited content that contradicted World Health Organization (WHO) guidance on COVID-19 treatment, prevention, and vaccines. In September 2023, YouTube quietly ended its COVID-19 misinformation policy. In 2025, the platform went further, allowing previously banned users to create new accounts and re-upload content — including the kind of health-related content that had triggered their bans.

YouTube continues to prohibit content that “contradicts local health authority guidance about specific health conditions and substances,” but the scope of this prohibition has narrowed considerably. The platform still removes content promoting harmful substances as cures for serious diseases, but the category of content that receives only reduced distribution (rather than removal) has expanded.

X (formerly Twitter)

After Elon Musk acquired Twitter in October 2022, the platform dramatically reduced its content moderation operations, cutting trust and safety staff by approximately 50%. Community Notes became the primary mechanism for addressing misleading content. X does not have a specific health misinformation policy and relies on Community Notes to provide context on disputed health claims.

The Section 230 Question

Section 230 of the Communications Decency Act (1996) is the legal foundation that enables social media content moderation. The law provides two key protections: (1) platforms are not liable for content posted by their users (“No provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider”), and (2) platforms can moderate content in good faith without losing their liability protection.

Section 230 has been the subject of intense political debate. Some lawmakers have proposed removing liability protection from platforms that moderate content based on viewpoint, arguing that moderation constitutes censorship. Others have proposed removing protection from platforms that fail to moderate harmful content, arguing that the liability shield enables the spread of misinformation. The Supreme Court has addressed aspects of this debate in several recent cases, including Murthy v. Missouri (2024) and Gonzalez v. Google (2023), without issuing a definitive ruling on the scope of Section 230 in the context of algorithmic amplification.

The EU’s Different Approach: The Digital Services Act

The European Union’s Digital Services Act (DSA), which took full effect in February 2024, represents a fundamentally different approach to platform governance. The DSA requires very large online platforms (those with more than 45 million EU users) to conduct systemic risk assessments of how their services may amplify illegal content, misinformation, and threats to public health. Platforms must provide transparency reports detailing their content moderation decisions, give researchers access to data for studying systemic risks, and face substantial fines (up to 6% of global annual revenue) for non-compliance.

The DSA’s transparency database, which logs content moderation actions across major platforms, recorded roughly one billion moderation actions per month in 2024 — providing an unprecedented view into the scale and patterns of platform content decisions. This level of transparency has no equivalent in the United States, where platforms disclose their moderation practices voluntarily and selectively.

What This Means for Health Information Consumers

The shift toward less restrictive content moderation has significant implications for anyone who encounters health information on social media. Content that was previously labeled as false, reduced in distribution, or removed outright may now circulate freely. Community Notes, while useful as a supplementary information source, are reactive (they appear after content has already spread), incomplete (only a fraction of misleading posts receive notes), and dependent on the composition and motivation of the user base that creates them.

For consumers evaluating health claims encountered on social media, the following principles are well-established:

Check the source. Health claims should be traceable to peer-reviewed research, government health agencies (CDC, NIH, FDA, WHO), or recognized medical institutions. Claims attributed to a single doctor, practitioner, or anonymous source — without reference to published research — warrant skepticism regardless of how confident the presentation appears.

Look for consensus, not outliers. Individual studies can produce surprising results that are not replicated. Scientific consensus is built on the convergence of multiple studies using different methods in different populations. A single study claiming to overturn decades of research is almost always either preliminary, methodologically flawed, or misrepresented.

Beware of commercial motivation. Content creators who sell supplements, treatments, or health products have a financial incentive to present their products as solutions to health problems. This does not mean their claims are automatically false, but it does mean they should be evaluated with additional scrutiny — particularly when the creator combines alarming health claims with product promotions.

Understand the difference between correlation and causation. Many viral health claims are built on correlations (two things that happen to increase at the same time) presented as causal relationships. The fact that autism diagnoses and vaccine coverage have both increased over the same period does not mean vaccines cause autism — ice cream sales and drowning rates also correlate, because both increase in summer.

The Platform Accountability Gap

The fundamental challenge of health information on social media is structural: platforms are designed to maximize engagement, and emotionally charged, alarming, or contrarian health content generates more engagement than careful, nuanced, evidence-based health communication. A study published in Science in 2018 found that false information spread faster, farther, and more broadly than true information across all categories studied — and that health misinformation was particularly prone to rapid viral spread.

This creates what researchers call an “information asymmetry”: the creators of misleading health content have algorithmic tailwinds (engagement-optimized distribution), while evidence-based health communicators face algorithmic headwinds (nuanced, hedged, scientifically cautious content generates less engagement and therefore receives less distribution). No content moderation policy — whether strict or permissive — addresses this underlying dynamic, because it is a feature of the platform’s business model rather than a failure of its moderation system.

The Kids Online Safety Act (KOSA), passed by Congress in 2024, requires platforms to take reasonable steps to prevent minors from accessing harmful content, including content that promotes self-harm, eating disorders, and substance abuse. Whether KOSA’s mandate will extend to health misinformation directed at minors remains to be determined through regulatory implementation and judicial interpretation.

For the foreseeable future, the responsibility for evaluating health information will continue to rest primarily with consumers rather than with platforms or regulators. Building the media literacy and scientific reasoning skills needed to navigate this environment is not just a personal health issue — it is a public health priority.

The Rise and Fall of Health Content Moderation: A Timeline

The trajectory of how social media platforms have handled health content over the past decade reveals a pattern of reactive, politically driven policy changes rather than a coherent long-term strategy.

2016–2018: Early interventions. Following the 2016 U.S. presidential election, platforms faced intense public and congressional pressure over their role in enabling misinformation. Facebook launched its third-party fact-checking partnership in December 2016. YouTube introduced “information panels” on videos about topics prone to misinformation, linking to authoritative sources like Wikipedia and the Encyclopedia Britannica. Pinterest took perhaps the most aggressive approach, blocking all search results for certain health-related queries (including “vaccine” and “anti-vax”) and replacing them with content from recognized health organizations.

2019: The anti-vax crackdown. A measles outbreak linked to vaccine-hesitant communities in Washington State and New York prompted platforms to take more aggressive action against anti-vaccination content. YouTube demonetized anti-vaccine channels (removing their ability to earn advertising revenue), Facebook reduced the distribution of anti-vaccine groups in recommendations, and Amazon removed anti-vaccination documentaries from its Prime Video service. This was the period in which Natural News and other alternative health channels were removed from YouTube.

2020–2021: COVID-19 and maximum enforcement. The COVID-19 pandemic triggered the most aggressive content moderation intervention in platform history. YouTube removed over one million videos for violating its COVID-19 misinformation policy. Facebook applied warning labels to over 190 million pieces of COVID-related content and removed 20 million pieces of content that violated its policies. Twitter introduced a strike system for COVID misinformation, permanently suspending accounts after five violations. These actions were coordinated with guidance from the WHO and national health authorities, creating what some critics characterized as a “censorship-industrial complex” and what public health advocates described as a necessary response to a deadly pandemic.

2022–2023: The pullback begins. Following Elon Musk’s acquisition of Twitter in October 2022, the platform dramatically reduced its content moderation operations. Twitter (rebranded as X) ended its COVID-19 misinformation policy in November 2022 and reinstated previously banned accounts, including those that had been suspended for spreading health misinformation. YouTube quietly ended its COVID-19 misinformation policy in September 2023, though it maintained broader health misinformation guidelines. Republican investigations into alleged government pressure on platforms to moderate content created a political environment in which aggressive health content moderation became increasingly untenable.

2024–2025: The permissive era. The Supreme Court’s ruling in Murthy v. Missouri (2024), while finding that the plaintiffs lacked standing, validated concerns about government involvement in platform moderation decisions. Meta’s January 2025 announcement ending its fact-checking program and YouTube’s decision to allow previously banned users to return completed the shift toward a more permissive approach. By mid-2026, no major U.S. platform maintains the kind of active, expert-reviewed health misinformation enforcement that was standard during 2020–2021.

What the Research Says About Health Misinformation Online

The academic literature on health misinformation and its effects is substantial and growing. Several key findings are relevant to the current policy debate:

Misinformation spreads faster than accurate information. The landmark 2018 study by Vosoughi, Roy, and Aral published in Science analyzed 126,000 stories shared by 3 million users on Twitter and found that false stories reached 1,500 people roughly six times faster than true stories. False information was 70% more likely to be retweeted than accurate information. The researchers found that this spread advantage was driven by human behavior (false stories were more novel and emotionally arousing) rather than by bots.

Health misinformation affects behavior. A 2020 systematic review by Suarez-Lledo and Alvarez-Galvez in Social Science & Medicine analyzed 69 studies on health misinformation in social media and found that exposure to health misinformation was associated with reduced vaccine uptake, increased use of unproven treatments, delayed medical care-seeking, and decreased trust in healthcare providers. The effects were strongest among individuals with lower health literacy and those who relied primarily on social media for health information.

Correction is less effective than prevention. Research consistently finds that fact-checks and corrections, while valuable, are less effective at changing beliefs and behavior than preventing exposure to misinformation in the first place. A 2021 study by Pennycook et al. in Psychological Science found that even brief prior exposure to misinformation reduced the corrective effect of subsequent fact-checks — a phenomenon known as the “continued influence effect.” This finding has significant implications for platforms that have shifted from proactive content removal to reactive community notes, because the community note typically appears only after the content has already been seen and shared.

Algorithmic amplification is the core problem. Multiple studies have found that social media recommendation algorithms disproportionately amplify sensational, emotionally charged, and contrarian content — precisely the characteristics that define health misinformation. A 2022 study published in Nature found that Facebook’s algorithm significantly shaped the political news users saw in their feeds, and similar dynamics have been documented for health content. The problem is not just that misinformation exists on platforms, but that the platforms’ business models actively promote it to the most receptive audiences.

The Natural Health Content Ecosystem

The removal of channels like Natural News from YouTube in 2019 did not eliminate the content — it displaced it to other platforms and distribution channels. Understanding how health content migrates across platforms when moderation intensifies is important for evaluating the effectiveness of platform-specific policies.

When YouTube removed Natural News, the channel’s content migrated to Rumble, Bitchute, and other platforms with less restrictive content policies. The channel also expanded its presence on Telegram, a messaging platform with minimal content moderation, and through its own website and email list. This pattern — deplatformed content migrating to less-moderated spaces rather than disappearing — has been documented repeatedly across different content categories.

Research by the Center for Countering Digital Hate (CCDH) found that the “disinformation dozen” — twelve individuals responsible for an estimated 65% of anti-vaccine content on social media in 2021 — maintained cross-platform presences that proved resilient to individual platform bans. When removed from one platform, these creators redirected their audiences to alternatives, often experiencing only temporary disruption in their reach.

This cross-platform resilience raises a fundamental question about the effectiveness of content moderation as a strategy for combating health misinformation. If removal from major platforms simply pushes content to less-moderated alternatives — where it may reach more radicalized audiences without any fact-checking or contextual information — the net effect on public health could be neutral or even negative. This “splinternet” dynamic is one reason some researchers advocate for media literacy education and demand-side interventions (teaching people to evaluate health claims) rather than supply-side interventions (removing content from platforms).

The Bluesky and Fediverse Experiment

The fragmentation of social media has introduced new models for content governance that may offer alternatives to the centralized moderation versus no-moderation binary. Bluesky, part of the AT Protocol decentralized social networking project, has pioneered what it calls “composable moderation” — a system in which users and communities can subscribe to different moderation services (called “labelers”) that tag content according to different criteria. A health-focused labeler could flag content that contradicts medical consensus, while a free-speech-focused labeler might flag only illegal content. Users choose which labelers to subscribe to, giving them control over their own information environment.

The Fediverse (the network of interconnected servers running Mastodon, Pleroma, and other ActivityPub-based social networking software) takes a similar approach at the server level: individual server administrators set content policies for their communities, and servers can choose whether to federate (exchange content) with other servers based on their moderation standards. A server focused on health professionals might defederate from servers that host unmoderated health misinformation, creating a segmented information environment where different communities operate under different rules.

These decentralized approaches are experimental and reach only a fraction of the social media audience. But they represent a conceptual alternative to the binary choice between platform-wide censorship and platform-wide permissiveness that has dominated the content moderation debate.

Media Literacy: The Long-Term Solution

Regardless of how platforms manage health content, the most sustainable approach to protecting consumers from health misinformation is building the critical evaluation skills that enable individuals to assess health claims independently. Media literacy — the ability to access, analyze, evaluate, and create media — has been identified by the WHO, the U.S. Surgeon General, and multiple academic reviews as a key intervention for reducing the impact of health misinformation.

Effective health media literacy includes several components:

Source evaluation. Learning to assess the credibility of information sources by examining their credentials, funding, track record, and potential conflicts of interest. A claim made by a board-certified oncologist publishing in a peer-reviewed journal carries different evidentiary weight than the same claim made by a supplement seller in a YouTube video.

Understanding study design. Recognizing the difference between anecdotes, case reports, observational studies, and randomized controlled trials — and understanding that each type of evidence has different strengths and limitations. Many viral health claims are based on animal studies, in vitro (test tube) experiments, or small observational studies that are presented as if they were definitive proof.

Recognizing logical fallacies. Identifying common reasoning errors in health content, including appeals to nature (the assumption that “natural” substances are inherently safe and effective), post hoc reasoning (assuming that because B followed A, A caused B), appeals to authority (citing credentials rather than evidence), and the ecological fallacy (drawing individual-level conclusions from population-level data).

Emotional awareness. Recognizing when health content is designed to provoke fear, outrage, or urgency rather than to inform. Emotional activation reduces critical thinking capacity and increases susceptibility to misinformation. Content that opens with alarming claims, uses inflammatory language, or creates a sense of emergency before presenting evidence should be evaluated with particular caution.

Several states have begun integrating media literacy into their K–12 education standards. Illinois became the first state to require media literacy education in 2022. California, Texas, New Jersey, and Colorado have since adopted similar requirements. The National Association for Media Literacy Education (NAMLE) has developed curriculum frameworks specifically addressing health misinformation, and the WHO’s “Let’s Flatten the Infodemic Curve” initiative provides open-access resources for health media literacy education.

What the Current Moment Means

The social media landscape in 2026 is characterized by a paradox: platforms are simultaneously more central to how people encounter health information and less willing to take responsibility for the quality of that information. The pandemic-era experiment in aggressive health content moderation has ended, replaced by a lighter-touch approach that relies on user-generated community notes and individual platform judgment rather than expert-reviewed fact-checking.

For public health, this creates both risks and opportunities. The risk is that health misinformation will circulate more freely, potentially influencing health behaviors and undermining trust in medical institutions. The opportunity is that the failure of top-down content moderation may accelerate investment in bottom-up approaches — media literacy education, health communication training for clinicians, and the development of trusted information intermediaries (health librarians, patient navigators, and community health workers) who can help individuals evaluate health claims in context.

The most important thing for consumers to understand is that social media platforms are not health information systems. They are advertising-supported engagement platforms that happen to carry health content alongside entertainment, news, opinions, and commerce. Treating them as authoritative sources of health guidance — whether the content is moderated or unmoderated — is a category error. For health decisions that matter, the appropriate sources remain peer-reviewed research, established medical institutions, and healthcare providers who know your individual situation.

The Role of Influencers in Health Information

The shift away from institutional content moderation has coincided with the rise of health influencers as primary sources of health information for millions of people. A 2024 survey by the American Medical Association found that 45% of adults aged 18–34 reported getting health information from social media influencers, compared to 38% from their primary care physicians. This inversion of traditional information hierarchies has significant implications for public health.

Health influencers occupy a spectrum from highly credentialed professionals (physicians, pharmacists, registered dietitians) who use social media to communicate evidence-based health information, to wellness entrepreneurs with no formal health training who present personal experience and commercial products as medical guidance. The platforms make no systematic distinction between these categories; a board-certified cardiologist and a self-described “wellness coach” selling supplements appear in the same feed, competing for the same attention, subject to the same (or lack of) content moderation standards.

The FTC has taken some enforcement action against health influencers who fail to disclose paid partnerships or make unsubstantiated health claims. In 2023, the FTC updated its endorsement guidelines to require clearer disclosure of material connections between influencers and the brands they promote. However, enforcement is reactive and resource-constrained: the FTC brings a handful of high-profile cases each year while thousands of influencers make unsubstantiated health claims daily without consequence.

The health influencer economy creates a structural incentive problem. Influencers who present nuanced, evidence-based information (“the evidence on this supplement is mixed; here’s what we know and don’t know”) generate less engagement than those who make dramatic claims (“this one supplement changed my life and Big Pharma doesn’t want you to know about it”). The algorithm rewards certainty, emotion, and novelty — qualities that are antithetical to careful scientific communication. The result is a marketplace where the most popular health content is often the least reliable.

Government Health Communication in the Platform Era

Public health agencies face an asymmetric challenge in the social media environment. The CDC, NIH, FDA, and WHO all maintain active social media presences, but their content consistently underperforms compared to sensational health misinformation. A 2021 study by the Reuters Institute found that during the COVID-19 pandemic, content from health authorities received a fraction of the engagement generated by misleading health content from unofficial sources.

Several factors contribute to this asymmetry. Public health agencies communicate with institutional caution — hedging claims, acknowledging uncertainty, using technical language — that is accurate but engagement-poor. They operate on government timelines, often responding to misinformation days or weeks after it has gone viral. They are constrained by political dynamics; the CDC’s credibility was damaged during the pandemic by perceived political interference in its guidance, and the agency has struggled to rebuild public trust.

Some public health communicators have adapted successfully to the platform environment. The CDC’s social media team has experimented with more engaging formats (short videos, infographics, humor), and individual clinicians with large social media followings (such as emergency physician Dr. Eric Feigl-Ding and dermatologist Dr. Dustin Portela) have demonstrated that evidence-based health communication can achieve significant reach when delivered in platform-native formats. However, these individual successes have not solved the structural problem: institutional health communication is systematically disadvantaged by engagement-optimization algorithms.

The Business Model Problem

The fundamental challenge underlying all social media health content issues is the platforms’ advertising-based business model. When revenue depends on maximizing user engagement (time on platform, clicks, shares, comments), and when emotionally charged, alarming, and contrarian content generates more engagement than careful, nuanced content, the economic incentives of the platform are structurally misaligned with the public health interest in accurate health information.

This misalignment cannot be fixed through content moderation alone, because content moderation addresses the symptoms (individual pieces of misinformation) rather than the cause (an engagement-optimization system that systematically advantages misinformation). Proposals for addressing the root cause include: requiring algorithmic transparency (so researchers and regulators can study how recommendation systems affect health information exposure), creating “circuit breakers” that slow the viral spread of health-related content until it can be verified, redesigning recommendation algorithms to prioritize information quality rather than engagement intensity, and developing alternative business models (subscription-based platforms, public-interest social media) that do not depend on engagement maximization.

The EU’s Digital Services Act represents the most ambitious regulatory attempt to address these structural issues, requiring very large platforms to assess and mitigate systemic risks including the amplification of health misinformation. Whether the DSA model will produce meaningful changes in platform behavior — and whether anything similar will be adopted in the United States — remains to be seen. The U.S. political environment, in which content moderation has become a partisan issue, makes comprehensive platform regulation unlikely in the near term.

In the meantime, individual health literacy remains the most reliable defense against health misinformation. Understanding how platforms work, why certain content goes viral, and how to evaluate health claims using basic scientific reasoning provides a level of protection that no content moderation policy — however well-designed — can match. Teaching these skills at scale, through K–12 education, public health campaigns, and healthcare provider training, is both the most effective and the most durable response to the challenge of health misinformation in the social media era.

References

1. KFF. Social Media Content Moderation Policies and False Claims. November 2025. kff.org
2. Congressional Research Service. Social Media: Content Dissemination and Moderation Practices. R46662. congress.gov
3. Berkeley Technology Law Journal. Meta’s Fact-Checking Rollback. May 2025. btlj.org
4. Vosoughi S, Roy D, Aral S. The spread of true and false news online. Science. 2018;359(6380):1146–1151.
5. European Commission. Digital Services Act. ec.europa.eu
6. Stanford Ethics in Society. Case Study: Community Moderation. 2025. ethicsinsociety.stanford.edu
7. Supreme Court of the United States. Murthy v. Missouri, 603 U.S. ___ (2024).
8. U.S. Surgeon General. Advisory on Building a Healthy Information Environment. 2021.

Last updated: September 26, 2026