When we started pulling apart the FDA's supplement adverse-event data — more than 54,000 reports — we expected the most common complaint to be something dramatic. A liver problem, maybe, or a racing heart. It wasn't. The single most-reported reaction, by a wide margin, was choking. People struggling to swallow large pills, mostly older adults and kids. It's the kind of thing that never makes a headline, and it tells you something important about how to read this data: it rarely says what you assume it says.
So before you read a number on one of our product pages and draw a conclusion, here's how to read it the way it actually deserves — carefully, and without scaring yourself over something the data can't support.
A report is a notification, not a finding
Every figure on this site traces back to a report someone filed with the FDA saying a health problem happened while they were using a supplement. Those reports come from consumers, from doctors and pharmacists, and from the manufacturers themselves.
What trips most people up is this: the FDA does not investigate and confirm most of them. A report records that an experience was reported — not that anyone established the product was responsible. It's a notification, full stop.
That distinction isn't pedantic; it's the whole game. Take kratom, which has one of the highest death counts in our data. Read the underlying reports and a pattern emerges fast: most of them involve several substances at once. The data genuinely cannot tell you kratom alone was responsible for those deaths — only that kratom was named in reports where someone died. Read every reaction on this site that way: "this was reported alongside the product," never "this product causes this."
Why we refuse to show you a "risk rate"
You will not find a figure like "4% of users had a serious reaction" anywhere on this site, and that's deliberate. A number like that would be worse than useless — it would be actively misleading.
The reason is that report counts measure reporting, not risk. They climb when a product gets sued, when it's in the news, or simply when a lot of people take it. A blockbuster multivitamin collects more reports than a niche one for the obvious reason that more people use it — not because it's more dangerous. Divide serious outcomes by total reports and you get a percentage that looks like your personal odds but isn't, because the people in that denominator aren't "everyone who used it" — they're "the small slice who reported." So we show counts, plainly, and leave the false precision out.
What the timing usually tells you
One of the most useful things on a product page is the least obvious: when the reports were filed.
Hydroxycut is the clean example. Its reports don't trickle in steadily over the years — they pile up around 2009 to 2011, the window of its FDA recall over liver-injury concerns. That cluster isn't a measure of your risk if you bought it last week; it's the fingerprint of a specific historical event. We see the same shape again and again — a recall, a wave of litigation, or a batch of reports filed together, all compressed into one or two years. When a product's reports concentrate like that, we flag it on the page. Treat a spike as a reason to go read what happened that year, not as a verdict on today.
What these numbers can and can't do
They can show you that reports exist, roughly how many, what kinds of effects people described, and how serious those reports were. They're a real, public record of human experience, and that's worth something.
They can't tell you your personal odds of anything, whether the product actually caused a reported effect, or how common that effect is among everyone who takes it. The skill is holding both of those at once — taking the data seriously without overreading it.
Reading a product page, section by section
Each page is laid out to support exactly that careful reading:
The total reports number is just how many reports name that specific product. The reactions are the effects people described, translated out of medical jargon and shown as counts — a picture of what was reported, not a ranking of danger. Outcomes tell you how serious things got (a doctor's visit, a hospitalization) as counts, never rates. Who reported shows the ages and sexes on record. And reports over time is the timing signal from above — often the most revealing part of the page.
One more deliberate choice worth knowing: we list death only under outcomes, never as a "reaction." When we first generated these pages, a product like 5-Hour Energy led with "death" at the top of its reaction list, which is both miscategorized and needlessly alarming. Move death to where it belongs — outcomes — and the genuine signal surfaces underneath: for 5-Hour Energy, that's cardiac events, heart attack and chest pain, which actually match the concerns regulators raised about it. That's the data being useful instead of just frightening.
What to do with any of this
Think of a product page as the start of a conversation, not the end of one.
If something here worries you about a supplement you take, bring it to a doctor or pharmacist — someone who can weigh it against your real health, your other medications, and the broader evidence. This site is information, not medical advice, and a single alarming-looking count is exactly the thing you now know not to panic over. And if you've had a problem with a supplement, report it to the FDA through the Safety Reporting Portal or MedWatch. Every page on this site exists because people did exactly that.
Read counts as reports, not rates. Read reactions as "reported," not "caused." Read a spike as an event, not a trend. Do that, and this becomes what it should be — one honest input into a decision you make with a professional, and never a substitute for one.