Health misinformation sites are rare online - but a small group of older adults visits them repeatedly

The conventional picture of health misinformation online - omnipresent, algorithmically amplified, spreading virally through social networks - may need adjusting. At least for health content specifically, a study tracking actual web-browsing behavior of more than 1,000 U.S. adults found that low-credibility health sites are both scarce and largely unseen.

Researchers from the University of Utah's Department of Communication, working with colleagues from the School of Medicine and Huntsman Cancer Institute, installed browser tracking tools on the devices of 1,055 participants and recorded their online activity for four weeks. The resulting dataset covered approximately 9 million page views, including 500,000 YouTube videos. The team coded websites for health content and categorized the health sites by credibility, identifying those that trafficked in low-credibility medical claims.

Of the 1,055 domains tagged as health-related, just 78 - 6.8% - contained low-credibility content. Only 13% of study participants visited even one such site during the four-week window, and those visits comprised just 3% of all health-related browsing. The findings were published in Nature Aging.

But exposure is sharply concentrated

The aggregate numbers are reassuring. The distribution is not. Traffic to low-credibility health sites was highly concentrated in a small subset of participants: the top 10% of visitors accounted for more than three-quarters of all such traffic. And that concentrated group skewed older.

"It's sort of good news, though. Overall, the levels are pretty low," said lead author Ben Lyons, an associate professor in the Department of Communication. "Not all older adults are like this, but the outliers are concentrated among older adults."

Older adults make more medical decisions and carry greater health burdens than younger cohorts, making accurate health information more consequential for them. They are also, this study suggests, the population most likely to encounter dubious health claims online - a mismatch that has practical implications for public health communication.

Politics matters less for health than for news

The study identified a correlation between right-leaning political orientation and exposure to low-credibility health content - consistent with patterns seen in political misinformation research. But Lyons and his colleagues found that political affiliation drove health misinformation exposure less strongly than it drives political misinformation exposure.

"The age effect is way bigger for politics," Lyons said. "People see politics as way more entertaining than they would health-related content. So there's less of a motivation to want to share these things. You don't get a feeling of team identity from sharing health misinformation like you would for information that puts down your political opponents."

This distinction matters for understanding the misinformation landscape. Political content has partisan identity functions that can motivate sharing and seeking. Health misinformation, absent that social fuel, appears to spread differently - and less widely.

How people reach dubious health sites

One of the study's more counterintuitive findings concerned the pathways through which people arrived at low-credibility health sites. Lyons's team examined the "referral" sources - the prior websites that sent traffic to dubious health destinations. Social media platforms and partisan news sites, often cited as major vectors for misinformation spread, did not appear prominently in this data.

"Are people going through Google search, or are they being referred through Facebook? We're not really seeing that in this data," Lyons said. "What we found, at least in the referral data, is that it's a more insular type of thing. They're visiting these because they visit other low-credibility sites, they're clicking through, and they're spending more time on these sites. They're going to them directly."

This pattern - direct navigation to familiar low-credibility sources rather than algorithmic referral - suggests that the most consistent health misinformation consumers may already have established habits and preferences that drive them back to specific sites. Interventions focused on disrupting algorithmic amplification may be less relevant for this population than approaches addressing the underlying habits and information preferences.

People who already held false health beliefs or expressed more conspiratorial views were more likely to encounter low-credibility health content, reinforcing a pattern of selective exposure rather than passive exposure. The problem is not random contamination of otherwise well-informed people; it is a concentrated subset with existing tendencies toward non-mainstream health information.

A limitation of the study is its four-week window, which may not capture the full range of health-related online behavior or the patterns that emerge during a specific health event or public health campaign. The sample, while tracked through actual browsing behavior rather than self-report, may not represent all demographic groups equally.