Half of Health Advice on Social Media Is Wrong

The Internet Is a Sickness Machine

You scroll past a post claiming that sunscreen causes cancer. Then one that says a certain vitamin can cure the disease your aunt just died from. Then a video of someone who "reversed" their diabetes with celery juice. None of this feels random. It feels like the internet is trying to kill you.

It might be. At least, it is trying to make you confused enough to not know what helps and what hurts. A 2020 systematic review by Víctor Suárez-Lledó and Javier Álvarez-Gálvez, published in the Journal of Medical Internet Research, combed through 69 separate studies on health misinformation across social media platforms. The authors found something that should make you pause before you share that next health tip: the prevalence of misinformation varies wildly by topic, but for some of the most dangerous subjects, it is not a fringe problem. It is the dominant narrative.

On Twitter, the platform that politicians and journalists use to shape public discourse, misinformation about smoking products and drugs reached 87% in some studies (Suárez-Lledó & Álvarez-Gálvez, 2020). That means nearly nine out of every ten posts about vaping, marijuana, or opioids were wrong. Not slightly misleading. Wrong.

This is not an accident. It is a design feature.

How a Systematic Review Works (And Why This One Matters)

Before we dig into the numbers, you need to understand what this paper actually did. Suárez-Lledó and Álvarez-Gálvez searched four major databases (PubMed, MEDLINE, Scopus, and Web of Science) for studies published before March 2019 that specifically measured health misinformation on social media. They defined health misinformation as "a health-related claim that is based on anecdotal evidence, false, or misleading owing to the lack of existing scientific knowledge." That is a tight definition. It excludes opinion, satire, and genuine disagreement among experts. It catches only claims that are verifiably wrong.

They found 69 eligible studies. Then they categorized the misinformation by health topic and by platform. They also looked at how the studies were done: social network analysis (28%), evaluating content (26%), evaluating quality (24%), content/text analysis (16%), and sentiment analysis (6%). This is not a single experiment. It is a meta-analysis of the entire field's output. If there is a single number that captures how bad the problem is, this paper gives it.

The Big Numbers: Where Misinformation Lives

The authors grouped the misinformation into six categories. Here is what they found, in order of how much misinformation each topic attracted.

1. Smoking Products and Drugs (87% Misinformation)

This was the worst category. Studies focused on e-cigarettes, vaping, marijuana, and opioids. The authors found that posts with misinformation reached 87% in some studies (Suárez-Lledó & Álvarez-Gálvez, 2020). That number is staggering. Think about it: if you scroll through Twitter looking for information about vaping, you have an 87% chance of landing on something false. The platforms are not neutral conduits. They are amplifiers.

Why so high? Because the industries that profit from these products have learned to use social media the same way everyone else does. They create content. They buy ads. They pay influencers. And the algorithms that decide what you see reward engagement, not accuracy. A shocking claim about vaping being "safe" gets more clicks than a boring one about lung damage.

2. Vaccines (43% Misinformation)

This is the one everyone talks about, and for good reason. Vaccine misinformation was present in 43% of studies (Suárez-Lledó & Álvarez-Gálvez, 2020). The human papilloma virus (HPV) vaccine was the most affected. That is not random. The HPV vaccine prevents a virus that causes cervical cancer. Misinformation about it directly leads to fewer people getting vaccinated, which leads to more cancer. The authors found that the misinformation was not just about safety. It was about conspiracy theories, about government overreach, about the vaccine being "unnatural." The content was engineered to trigger an emotional response, not a rational one.

3. Diets and Eating Disorders (36% Misinformation)

This category surprised me. The authors found that 36% of posts about diets or pro-eating disorder arguments contained misinformation (Suárez-Lledó & Álvarez-Gálvez, 2020). This is not just about fad diets. It is about content that actively promotes disordered eating. The authors note that this category was "moderate" compared to the others, but moderate for a public health crisis is still a crisis. If a third of what you read about nutrition is wrong, you are not making informed choices. You are gambling.

4. Diseases: Noncommunicable Diseases and Pandemics (40% Misinformation)

This category covered cancer, heart disease, diabetes, and infectious disease outbreaks. The authors found misinformation rates around 40% (Suárez-Lledó & Álvarez-Gálvez, 2020). Cancer was the most affected. That makes sense. Cancer is terrifying. People want a cure. They want hope. And the internet is happy to sell them hope in the form of miracle cures that do not work. The authors note that this misinformation often comes from sources that look legitimate. A website that looks like a medical journal. A doctor who has been stripped of their license but still has a Twitter account. The packaging is convincing. The content is not.

5. Medical Treatments (30% Misinformation)

This was the lowest category, but 30% is still bad. The authors found that misinformation about medical treatments, including prescription drugs and alternative therapies, was present in about a third of studies (Suárez-Lledó & Álvarez-Gálvez, 2020). This includes claims that antibiotics can cure viruses, that chemotherapy is more dangerous than cancer, that herbal supplements can replace insulin. The authors note that this category was the most heterogeneous. Some treatments were accurately represented. Others were not.

The Platform Problem: Twitter Is Ground Zero

The authors did not just count misinformation by topic. They also looked at which platforms were most affected. Twitter was the worst. The authors found that "the prevalence of health misinformation was the highest on Twitter" (Suárez-Lledó & Álvarez-Gálvez, 2020). This is not because Twitter is uniquely evil. It is because Twitter's format rewards brevity and emotional punch. A 280-character claim about a drug or a vaccine fits perfectly. A nuanced correction does not. The platform's algorithm also amplifies content that gets reactions, and nothing gets reactions like fear or anger.

Facebook and YouTube were also major vectors, but the authors note that the studies on these platforms were less common. That might be because Twitter is easier to scrape and analyze. Or it might be because the misinformation there is simply more visible to researchers.

What This Does Not Prove

This paper is a systematic review. It tells you what the research says about the prevalence of misinformation. It does not tell you how many people believe what they see. It does not tell you how many people act on it. It does not tell you what the long-term health effects are. Those are separate questions, and they are harder to answer.

The authors also note that their definition of misinformation excludes claims that are simply not supported by evidence. A claim that is "based on anecdotal evidence" counts as misinformation. But anecdotal evidence is not always wrong. Sometimes a person really does get better after trying something. That does not mean it works for everyone. The authors are careful to say that their definition is about the claim itself, not about the person making it.

There is also a question of timing. The studies in this review were published before March 2019. That means they predate the COVID-19 pandemic. The pandemic changed everything about health misinformation. It made it more urgent, more political, and more visible. The authors acknowledge this limitation. Their findings are a baseline. They are not the final word.

What This Actually Means

Here is what you can do with this information. Not what you should feel, but what you can do.

• ▸Check the source before you share. If a health claim comes from a person with no medical credentials, or from a website that looks like it was designed in 1998, assume it is wrong until proven otherwise. The authors found that misinformation is most common on platforms that reward speed over accuracy. You can slow down.

• ▸Be skeptical of anything that sounds too good to be true. The authors found that cancer misinformation was especially prevalent. That is because cancer is scary and people want easy answers. The easy answers are almost always wrong. Real medicine is slow, boring, and incremental.

• ▸Treat Twitter as a source of entertainment, not education. The authors found that Twitter had the highest prevalence of misinformation. If you are getting health advice from Twitter, you are getting it from the platform most likely to be wrong. Use it for jokes. Use it for news. Do not use it for medical decisions.

• ▸Watch out for smoking and drug content. The authors found that this category had the highest misinformation rate. If you are a parent, a teacher, or anyone who talks to young people about vaping or marijuana, know that the information they are seeing online is almost certainly wrong. You need to counter it with facts.

• ▸Do not assume that a high number of likes means the claim is true. The authors found that misinformation spreads because it is engaging. It triggers emotion. A post with a million likes might be a post that is perfectly designed to deceive you. Likes are not evidence. They are a measure of emotional resonance.

The internet is not a library. It is a machine that runs on attention. And attention does not care about the truth. It cares about what makes you stop scrolling. The authors of this paper have given you a map of where the traps are. Now you have to decide whether to walk into them.