How VAERS Actually Works — What It Is, What It Isn't, and How to Read It Without Being Fooled

A parent searches a vaccine database at 2 a.m. and finds a scary number. Here's what that number really means — and the one question that separates a signal from a story.

A parent is up at 2 a.m. after a kid's shot, scrolling a government website. They type in the vaccine, and a number comes back: reports of fevers, seizures, even deaths. The stomach drops. Is this the thing nobody told them about? Or is it noise dressed up as a verdict?

That moment is where most VAERS confusion is born — and where both sides of the vaccine fight love to plant a flag. NU's job here isn't to tell you shots are safe or dangerous. It's to show you exactly what that database is, what it can and can't prove, and how to read it so nobody — pharma cheerleader or anti-vax influencer — can play you with it.

What VAERS Actually Is

VAERS stands for the Vaccine Adverse Event Reporting System. It's run jointly by the CDC and the FDA, and it has existed since 1990. Think of it as a national suggestion box for "something happened after a vaccine." It is an early-warning tripwire, not a courtroom and not a scoreboard.

The key word is passive surveillance. VAERS doesn't go looking for problems. It waits for people to file. And almost anyone can file — doctors, nurses, manufacturers (who are legally required to report certain events), and ordinary members of the public. You don't need proof. You don't even need a medical opinion that the vaccine caused anything. You just need to report that an event followed a vaccination.

That open door is a feature, not a bug. The whole point is to cast a wide net so a rare, unexpected pattern can surface fast. The system's job is to catch a whisper early, so scientists can go decide whether it's real.

The One Thing It Cannot Do

Here is the sentence that should be tattooed on every VAERS argument: a report shows that an event happened after a vaccine, not that the vaccine caused it.

Correlation in time is not causation. If millions of people get a shot, some of them will, by pure chance, have a heart attack, a stroke, a miscarriage, or die the next week — exactly as some would in any random week without a shot. Those events can legally and legitimately land in VAERS. Their presence in the database is not evidence the vaccine did it.

This cuts against bad-faith uses on both ends:

• The misleading "millions of injuries" claim. Pull a raw count of reports, call them "vaccine injuries," and you've quietly swapped "reported after" for "caused by." That's the most common manipulation, and it's wrong by definition of the system.
• The misleading "it's all junk, ignore it" dismissal. The flip side is treating VAERS as worthless because it's unverified. It isn't worthless. It's the tripwire that has caught real problems — its unverified nature is the cost of speed, not proof of uselessness.

Both moves abuse the same fact: VAERS is a starting point, never an ending point.

When VAERS Has Actually Worked

The honest case for VAERS is that the tripwire has tripped on real things. Its most cited success: in 1999, VAERS reports helped flag a link between an early rotavirus vaccine (RotaShield) and intussusception, a serious bowel problem in infants. Follow-up investigation confirmed the risk, and the vaccine was pulled. That's the system doing its exact job — whisper, investigate, act.

More recently, the rare clotting and heart-inflammation signals discussed around COVID vaccines were the kind of thing surveillance systems are built to surface and then study with harder tools. The pattern is always the same: VAERS (or a system like it) raises a hand; real epidemiology decides whether the hand was right.

How the Pros Actually Use It

VAERS isn't the only system, and it isn't meant to stand alone. The CDC and FDA pair it with active systems — networks like the Vaccine Safety Datalink and the FDA's BEST/Sentinel infrastructure — that comb real medical records and can compare vaccinated and unvaccinated groups. Those systems can do what VAERS can't: estimate whether an event happens more often after a vaccine than you'd expect by chance.

A "safety signal," in the professional sense, is a statistical flag that says go look harder — not a finding of harm. The integrity of the whole enterprise lives in one question: when a signal appears, does someone actually investigate it honestly, or does it get explained away? That's the legitimate fight worth having — and it doesn't require pretending a raw report count proves causation.

How to Read VAERS Without Getting Played

A field guide for that 2 a.m. moment:

1. Translate every number. "X reports" means "X times someone said this happened after a shot." It does not mean "X people were harmed by a shot." Mentally insert "reported, unverified" before any total.
2. Watch for the word "associated." It's doing heavy lifting. "Associated with" = "showed up near in time," not "caused by."
3. Ask: compared to what? A scary count means nothing without a baseline. How often does this event happen anyway, in a population that size, vaccine or not?
4. Look for the follow-up, not the filing. The real news is never the raw report — it's whether an active system investigated and what it found.
5. Distrust anyone who stops at the raw count — in either direction. The honest reader treats VAERS as a question, not an answer.

NU's Bottom Line

VAERS is a smoke detector, not a fire marshal. A beeping detector is worth taking seriously — and it's also not the same as your house being on fire. The system is genuinely valuable precisely because it's noisy and fast, and it's genuinely dangerous in the hands of anyone who quotes its raw numbers as a body count.

So when someone waves a VAERS figure at you — to scare you or to reassure you — ask the one question that separates a signal from a story: did anyone go check? That's the part people feel before they see a chart, and it's the part that keeps this honest. Records over spin.