Lupa finds nearly 90% of online health content was false

By Priya Raman

Nearly 90% of health-related content that circulated on social media from 2017 through June 2025 was false, according to the fact-checking agency Lupa’s report “Raio X da desinformação em saúde” (“X-ray of health misinformation”), supported by the medical education group Afya.

Lupa’s research also found that the volume of health-misinformation posts grew tenfold in 2020 and 2021, the first two years of the Covid-19 pandemic.

Before Covid-19, vaccine misinformation represented 9.1% of Lupa’s fact checks. The share rose to 33.2% during the pandemic and reached 38.9% from 2022 through 2024. It then stayed around 20% from 2025 through June 2025 rather than returning to its pre-Covid level.

The report identified several ways health misinformation spread:

• Testimonials from fabricated medical or scientific authorities.
• Promises of cures and simple solutions to complex medical problems.
• Fear-based and alarmist language.
• Distortion of studies and data.

False medical or scientific authority was the most common tactic in Lupa’s findings.

In one case cited by Lupa, fraudsters used the image of Dr. Drauzio Varella to sell supplements and medicinal substances while promising miraculous results. The report also described an online course that used his voice to teach people how to produce false advertisements.

Varella has long denounced the practice. In an O Globo article, he said the fraud began with his photographs in advertisements for “remédios miraculosos” (“miracle cures”). The ads purported to cure back pain, joint pain, diabetes, high blood pressure, Alzheimer’s and other diseases.

Varella also said he repeatedly asked Meta to remove the false posts without success. According to Varella, he reported the case to the Public Prosecutor’s Office of São Paulo and pursued it in court.

In the week before publication, authorities launched an operation with support from the Public Prosecutor’s Office of Rio de Janeiro against a group accused of using AI to forge videos with the images and voices of famous people, including Varella, O Globo reported.

In October, the 6th Chamber of Private Law of the São Paulo Court of Justice unanimously increased the award for moral damages against those responsible for the course that used Varella’s image and voice, O Globo reported.

At the federal level, Brazil’s Office of the Attorney General of the Union notified YouTube in September about AI videos that simulated doctors to spread treatments without scientific evidence, O Globo reported. YouTube removed the channels and sent information about profiles linked to scams to the Federal Police and the Federal Council of Medicine, according to O Globo.

O Globo’s editorial argued that digital platforms must act preventively and stop serving as channels for health scams. It ended with a blunt line: “Não é difícil” (“It is not difficult”).

What does this change for marketers and platforms?

The practical change is prevention before distribution, rather than reaction after a complaint. Platform teams need to detect misuse of medical authority, cloned voices and videos in which AI simulates doctors. Marketers should treat cure claims, fabricated experts and synthetic endorsements as evidence-review and brand-safety triggers rather than waiting for a public complaint or takedown request.

Varella’s experience with Meta shows why the sequence matters: he had to pursue prosecutors and courts after removal requests failed. YouTube responded to the federal notice by removing channels and referring scam-linked profiles to law-enforcement and medical authorities, but that action came after government intervention, according to O Globo.

The next test is whether Meta, YouTube and other platforms can turn prevention into fewer violations, rather than leaving doctors and public prosecutors to pursue each case after publication.