Internal emails show FDA leaders were warned in early 2021 that the algorithm scanning COVID vaccine injury reports could not detect myocarditis and other harms from mRNA shots. Officials kept using it anyway and pointed to its silence as proof of safety.
The Independent Medical Alliance (IMA), a national coalition of independent physicians, researchers and clinicians, today called on Congress and the Justice Department to investigate how the CDC and FDA handled COVID vaccine safety surveillance.
The call follows an investigation published in The BMJ (formerly British Medical Journal) by reporter David Willman. Willman reconstructed the first year of the mRNA rollout using internal government emails released under the Freedom of Information Act and to U.S. Senate investigators.
The CDC promoted its Vaccine Adverse Event Reporting System (VAERS) as “the nation’s frontline system” and “the early warning system for detecting possible vaccine safety problems.” Behind the scenes, federal officials had been told that the method used to analyze those reports could not reliably detect harm from the two dominant mRNA products. They did not fix the method. They told the FDA scientist who offered a fix to stand down. Then they cited the absence of alerts as public reassurance. One reason given in writing for leaving the method alone: a change could have “unintended consequences” for “vaccine confidence.”
“Frontline doctors like me were seeing these injuries in growing numbers. When we raised the alarm, we were dismissed because the ‘official data’ said otherwise,” said Dr. Joseph Varon, President and Chief Medical Officer of the Independent Medical Alliance. “Now we know that data came from a tool the FDA had been warned was blind. This is a failure of our public health system that may well rise to criminal negligence, and patients paid the price.”
What The BMJ found
Two methods were supposed to run side by side: the CDC’s proportional reporting ratios (PRR) and the FDA’s empirical Bayesian data mining, with the agencies sharing results. In a little-noticed September 2022 letter, CDC Director Rochelle Walensky said her agency did not run PRR analyses until 2022, more than a year into the vaccination campaign. Both agencies, she wrote, “chose to rely” on the FDA’s Bayesian method, which she called “a more robust technique.”
In the weekly FDA reports The BMJ reviewed, that Bayesian method never signaled for myocarditis, pericarditis or tinnitus. The problem was a statistical effect called masking. In the first year, more than 90 percent of VAERS reports involved the two mRNA shots, so the algorithm was in effect comparing Pfizer against Moderna. If both raised the risk of the same injury, the elevated rate looked normal, and no alert fired.
Timeline
- February 2021: FDA medical officer Dr. Ana Szarfman emailed Dr. Peter Marks, director of the FDA’s Center for Biologics Evaluation and Research and the government’s top COVID vaccine regulator. She urged the agency to fix its monitoring and offered an updated algorithm developed by statistician William DuMouchel, who had designed the original Bayesian method the FDA was using.
- Late February 2021: The CDC and FDA received notice from Israel’s Ministry of Health of a large number of myocarditis reports in young people after vaccination.
- Two weeks after Szarfman’s email: Marks’s office held a video conference with Szarfman. Her slides showed the PRR method lighting up with signals and the FDA’s Bayesian method showing almost none: “You are not getting useful information with such low counts.” DuMouchel’s updated method, first published in 2012, caught signals without burying investigators in false alarms. Marks and his staff declined to change the FDA’s approach.
- Spring 2021: Israeli officials told a CDC-convened safety committee that two of 77 vaccinated patients with myocarditis or pericarditis had died. The following week, the Pentagon’s Defense Health Agency told the same committee about a cluster of myocarditis cases among male service members.
- April 27, 2021: Walensky told reporters, “We have not seen a signal [for myocarditis], and we’ve actually looked intentionally for the signal in the over 200 million doses we’ve given.”
- May 7, 2021: Craig Zinderman, an FDA medical policy official, asked Szarfman to “please hold off on creating and sending data mining reports and analyses using COVID-19 vaccine AE [adverse event] data.” Adjusting the method mid-campaign, he wrote, “could lead to confusion and have unintended consequences (eg, regarding vaccine confidence).”
- June 2021: A CDC presentation showed VAERS myocarditis and pericarditis reports in males ages 12 to 24 running 29 to 347 times higher than expected. The day before, Dr. Tom Shimabukuro, deputy director of the CDC’s Immunization Safety Office, received the FDA’s latest data-mining report, which still showed no alert. He emailed a colleague: “I’m perplexed that myocarditis isn’t alerting for either of the mRNA vaccines.”
- September 15, 2021: After Szarfman kept pressing, Marks wrote to his counterpart at the FDA’s drug center that she had been asked to “cease and desist, because the strategy that she is using could create erroneous conflicts that feed into anti-vaccination rhetoric.”
- 2022: When the CDC finally ran its PRR analyses, myocarditis, pericarditis, Bell’s palsy and tinnitus met the agency’s own signal criteria in all 13 consecutive weekly reports from May 6 to July 29. Hundreds of other event types also crossed the threshold. Walensky nonetheless wrote that the results were “generally consistent” with the Bayesian findings and revealed “no additional unexpected safety signals.” The worksheets The BMJ examined do not support that description.
- October 2023: Dr. Narayan Nair, head of the FDA’s vaccine pharmacovigilance division, wrote to CDC and FDA colleagues: “We were aware of this [data mining] limitation before and during the pandemic.”
“Officials protected the safety narrative instead of protecting patients,” continued Varon. “They served the industry, not the individual. Groups like the American Medical Association and the American Academy of Pediatrics joined the chorus to drown out frontline doctors. It is long past time for Congress and the Justice Department to find out how this happened and why. It must never happen again.”
The BMJ investigation is available at bmj.com. IMA’s analyses are available at IMAhealth.org. To arrange an interview with one of our doctors, contact [email protected].
