Facebook's policy on anti-COVID vaccine content didn't stop users from finding it, study suggests

But there were no significant changes in engagement with anti-vaccine page content (RR, 0.73), and the volume of false claims rose. Engagement with anti-vaccine groups was 33% higher than expected on the basis of prepolicy trends (RR, 1.33), but this was not significant relative to pro-vaccine group trends (RR, 1.22).

A robustness check on a second sample of anti-vaccine groups identified on July 28, 2021, showed that engagement counts were similar to those before the Facebook policy, which the study authors said raises the possibility that the platforms policy shifted—rather than drove down—engagement.

"This finding—that people were equally likely to engage with vaccine misinformation before and after Facebook’s extensive removal efforts—is incredibly concerning," senior author Lorien Abroms, PhD, said in a George Washington University news release. "It shows the difficulty that we face as a society in removing health misinformation from public spaces."

Largest rise in claims about vaccine's side effects

The proportions of several topics that seemed to violate Facebook's community standards for pages rose after the policy was announced.

The largest increase occurred in allegations that COVID-19 vaccination results in severe adverse reactions (odds ratio [OR], 1.41). Increases were also noted in reports of vaccine-related hospitalization and death (OR, 1.23), promotion of alternative medicine (OR, 1.32), allegations of negative effects of vaccines on immunity due to toxic ingredients (OR, 1.24), and posts focusing on children (OR, 1.34).

Other proliferating topics included school vaccine mandates (OR, 2.02), other vaccine mandates (OR, 1.21), legislation opposing vaccination (OR, 1.06), and anti-vaccine medical advice (OR, 1.14). Anti-vaccine groups also discussed misleading information—particularly those pertaining to vaccine safety, effectiveness, and mandates—more often than before the policy. "These increases were significantly smaller in magnitude than increases in provaccine groups, suggesting that they were not attributable to Facebook's policies," the researchers wrote.

The likelihood of a user engaging with posts containing low-credibility links was 2.7 times higher than expected based on prepolicy patterns, and the odds that groups had posts with such links was 1.2 times greater.

Compared with prepolicy trends, the probability that a post contained a link to a politically polarizing website climbed in both pages (OR, 1.35) and groups (OR, 1.51), and engagement with these links rose on anti-vaccine pages (OR, 2.37).

'Like,' 'angry' reactions boost exposure

"Prior work suggests that removal of antivaccine Facebook groups may have been associated with increased activity on other platforms, such as Twitter," they wrote. Also, "links to 'alternative' social media platforms—BitChute, Rumble, and Gab—as YouTube and Twitter [increased after they] began removing COVID-19 misinformation on 14 October 2020 and 1 March 2021, respectively."

Similarly, a simulation model built and calibrated to prepolicy data best reproduced data from after the prepolicy period when the researchers assumed that removals shifted demand for antivaccine content to remaining pages and groups rather than lowering it. "Together, these results suggest that content and account removals may not have dissuaded audiences from seeking out antivaccine misinformation," they wrote.

Rather, Facebook's system architecture, which determines how information flows through the platform, may have enabled anti-vaccine content producers and users to forge new paths to such content, such as through pages administered by anti-vaccine opinion leaders, discussion groups that coordinate to repost content, and interaction with posts to promote them in news feeds, the researchers said.

"Individuals that are highly motivated to find and share anti-vaccine content are just using the system the way it's designed to be used, which makes it hard to balance those behaviors against public health or other public safety concerns. You have to change the architecture if you want to create that balance," lead author David Broniatowski, PhD, said in the release.

The researchers added that Facebook's newsfeed algorithm, designed to promote content that has generated "meaningful social interaction" as measured by reactions such as "like" and "angry," could also have increased exposure to false claims.

"Just as the products of building architecture must conform to building codes to protect public health and safety, social media platform designers must consider the public health consequences of their architectural choices and perhaps even develop codes that are consistent with best scientific evidence for reducing online harms" the authors concluded.