BMJ Probe: US Knew Vaccine Safety System Was Flawed

U.S. health officials knowingly relied on a flawed vaccine safety algorithm to monitor for harm from mRNA COVID-19 vaccines, and they shut down efforts to fix it, according to an investigation published by The BMJ. The investigation draws on government emails released under the Freedom of Information Act and in response to Senate investigators, along with interviews with public health officials and scientists.

Before the vaccine rollout began in December 2020, the Centers for Disease Control and Prevention told health care professionals and the public that its Vaccine Adverse Event Reporting System, known as VAERS, could quickly identify harmful reactions after vaccination. The agency said it would use two data mining techniques to scan the reports: proportional reporting ratios, known as PRRs, and an empirical Bayesian method supplied by the Food and Drug Administration, which operates VAERS jointly with the CDC. The two agencies were supposed to compare and discuss their findings.

But the CDC did not run PRR analyses until 2022, according to a letter from CDC Director Rochelle Walensky. Instead, the CDC and the FDA relied entirely on the FDA's Bayesian method.

The BMJ found that the FDA's algorithm failed to flag a possible link between mRNA COVID vaccination and myocarditis, inflammation of the heart muscle; pericarditis, inflammation of the fluid-filled sac around the heart; Bell's palsy; tinnitus; and other reported conditions, because of a flaw in how the method detects signals. More than 90 percent of the early VAERS reports involved the new Pfizer and Moderna mRNA vaccines. When both vaccines raised the rate of an event such as myocarditis by a similar amount, the observed and expected frequencies looked alike, so the system issued no automated alert.

Internal records show that FDA officials understood this limitation before and during the pandemic, and that CDC officials were told about it during the vaccine rollout.

The FDA doctor who warned them

In early 2021, FDA medical officer Dr. Ana Szarfman and statistician William DuMouchel, who had developed the Bayesian algorithm, warned top officials about the flaw and proposed an updated algorithm that caught the signals. Szarfman was told to "cease and desist." Officials kept pointing to the absence of system alerts as they reassured clinicians and the public that the vaccines were safe.

Szarfman told The BMJ she had not wanted to undermine public support for the COVID vaccines. "Very few people understand the statistics. That's the problem," she said.

DuMouchel said he could not explain officials' resistance to switching to the corrected method and that, in his view, they had been wrong. He added that he regretted that FDA officials rejected Szarfman's proposed fixes for VAERS: "If they had paid attention to Ana, they would have done better."

What the delayed checks found

When the CDC finally ran PRR analyses in 2022, Walensky said the results showed "no additional unexpected safety signals." But The BMJ examined the analyses and found hundreds of adverse events that met the agency's own alert criteria, including myocarditis, pericarditis, Bell's palsy, and tinnitus. The FDA's Bayesian method had never flagged them.

In October 2023, the FDA's pharmacovigilance chief acknowledged in an email to colleagues that the agency had known about the detection problem since the vaccines rolled out, more than two years earlier. The chief did not respond to requests for comment.

Why it matters

VAERS is still the country's main system for tracking vaccine safety, so the failure is not just history. The next time a new vaccine or booster rolls out quickly, the same infrastructure will be the one asked to confirm it is safe. The investigation shows the automated safety net was structurally blind to the signals it was meant to catch, and the scientists who saw the blindness were silenced.

The findings also arrive as public health policy from the pandemic era faces re-examination. The CDC no longer recommends routine COVID-19 vaccination for children, advising individual-based decision-making instead, after an FDA study found a higher risk of febrile seizures in young children following the Moderna shot. That reassessment is part of a broader moment for public health reporting, from new research on diabetes and everyday life to surveys linking worker fatigue to lost productivity.

The full investigation, by investigative journalist David Willman, is published in The BMJ: https://www.bmj.com/content/394/bmj-2026-100806. Further coverage of the findings is available at News-Medical.net.