US health officials relied on a single computer method to scan reports of COVID vaccine side effects, even though they knew it could miss real warning signs, according to an investigation published today in The BMJ. The method failed to flag myocarditis, pericarditis, Bell’s palsy and tinnitus after the Pfizer and Moderna shots.
The practical point: if something unexpected happens after any vaccine, tell your doctor and file your own report to VAERS, the federal Vaccine Adverse Event Reporting System. Do not assume a computer will catch it. This story does not reveal a hidden harm. Myocarditis was caught in 2021 by other checks and confirmed as a rare side effect. Whether to get a shot this fall still rests on the known risks, talked through with your doctor.
Key takeaways
- In England, a second Moderna dose added about 10 myocarditis cases per 1 million people in the next 28 days.
- VAERS reports of heart inflammation came in particular after mRNA vaccination.
- Myocarditis was still found in 2021 through VAERS and other reporting systems.
What happened
The investigation, by journalist David Willman, draws on government emails released under the Freedom of Information Act and to US Senate investigators, plus interviews.
The FDA method asks whether a problem is reported more often for one vaccine than expected from the whole database. If the Pfizer and Moderna vaccines both raised the risk of something like myocarditis by about the same amount, the “observed” and “expected” numbers came out similar, and no alert fired.
A scientist who raised concerns was asked to “cease and desist,” according to The BMJ.
Dr. Kumar’s take
VAERS is a passive system. Patients, nurses and doctors send in reports, and nobody first checks whether the vaccine caused the problem. That makes it fast but noisy. Software then hunts for patterns worth a closer look.
Every one of those pattern-hunting methods has blind spots. Monash University statistician Michael Lydeamore pointed out that PRR has the opposite problem: when reports flood in for a new vaccine, it can overestimate risk “by a large amount.” That is why you run more than one method and then test the signals in proper studies.
That second step did happen for myocarditis. By 4 November 2021, VAERS held 1,783 reports of heart inflammation in people aged 12 to 29 after COVID vaccination. A study of more than 38 million vaccinated people in England, published in Nature Medicine, then measured the risk. In the 28 days after a second Moderna dose, it found about 10 extra myocarditis cases per 1 million people. After a first Pfizer dose it was about 1 extra case per million. A positive COVID test carried about 40 extra cases per million. The added risk from the mRNA vaccines showed up only in people younger than 40. In people 40 and older, a first dose of the AstraZeneca (ChAdOx1) vaccine raised myocarditis risk by about 33% in the following 28 days.
So the failure here is not that a harm went unseen. It is that officials pointed to “no alerts” as reassurance while knowing the tool could not produce alerts for a shared problem. As University of Auckland vaccinologist Helen Petousis-Harris put it, the absence of a signal from one algorithm “does not prove that no risk exists.” Sidelining the scientist who offered a fix is what damages trust.
The reverse error matters too. Hundreds of events meeting an alert threshold does not mean hundreds of conditions were caused by the vaccine. A signal means “look closer,” not “proven.”
What it means for you
- Report it yourself. If you notice a new symptom after any vaccine, tell your doctor and file a VAERS report. Your report is the raw material every method depends on.
- Use what is measured, not what is alleged. The myocarditis risk is real, rare, concentrated in people under 40, and smaller than the risk after a positive COVID test in the England data. This report does not change those numbers.
- Leave the call with your doctor. Your age, health history and past reactions shape whether a fall shot makes sense for you.