How to Judge Supplement Evidence — a Simple Tier System
Quick Answer: What Counts as Good Evidence?
Human trials, at the dose in the product, ideally randomised and controlled, and ideally more than one of them. Everything below that is a hint about a mechanism rather than a finding about people. The most common trick in supplement marketing is not lying — it is citing a real study from a lower tier and describing it in language reserved for a higher one.
- Top: meta-analyses and repeated randomised human trials.
- Middle: a single small human trial — useful, not settled.
- Bottom: cell and animal work, and traditional use.
The tiers, from strongest to weakest
Not all citations are equal, and a reference list is not evidence of anything by itself. Here is the hierarchy the research world works in, translated into the language of a supplement label.
Tier 1 — systematic reviews and meta-analyses
Someone has gathered every trial on a question and pooled the results. This is the strongest form of evidence a supplement ingredient can have, and very few of them have it in a form that matches a specific product's dose. When it exists, it is the thing to read first.
Tier 2 — randomised controlled trials in humans
People, randomly allocated, ideally blinded, with a placebo group. The details matter enormously: forty participants for six weeks is a different animal from four hundred for a year. A single positive RCT is a real finding and not a settled question.
Tier 3 — observational human studies
Populations tracked over time. Excellent for generating hypotheses, weak for causation, and the source of a large share of confidently-worded nutrition claims that later failed in trials.
Tier 4 — animal studies
Mice are not small people. Rodent studies are how promising mechanisms get identified and they routinely fail to replicate in humans. A supplement claim resting on mouse data is a hypothesis wearing a lab coat.
Tier 5 — in vitro, and traditional use
Cells in a dish exposed to concentrations that could never be reached in a bloodstream. Traditional use tells you something about safety and cultural history, and essentially nothing about efficacy for a modern claim.
The dose question that outranks the tier
Evidence tier is only half the assessment. The other half is whether the product contains the dose that produced the effect. An ingredient with two excellent human trials at 4 g per day, included at 100 mg, is not supported by those trials — the label is borrowing their credibility while delivering one fortieth of the intervention. Under-dosing is the single most common gap between a supplement's citation list and its capsule, and it is invisible unless you check both numbers.
The strain and form problem
For probiotics, findings are strain-specific: evidence for one strain does not transfer to another strain of the same species. For botanicals, extract ratio and standardisation matter for the same reason. And preparation can change everything — the best-known human trial on Akkermansia muciniphila, for instance, used a pasteurised preparation, which is not automatically interchangeable with a live one in a capsule.
Four phrases that should send you to the source
- “Clinically proven” — proven is not a word research uses. Ask: which trial, on what, at what dose, in whom?
- “Studies show” with no citation — the studies, if they exist, are usually tier 4 or 5.
- “Scientifically formulated” — a description of the formulation process, not a finding.
- “Doctor recommended” — a person in a white coat in a photograph is a stock convention, not a clinical endorsement.
How to run the check in two minutes
Take the one ingredient the product is really selling. Search its name plus “randomised trial” on PubMed. Look at what the trials used as a dose and compare it to the panel on the bottle. Then check whether any trial exists on the finished product rather than on the parts. That last question is where most supplements — including, for what it is worth, the microbiome formulas we cover — come up empty, and knowing that in advance is worth more than any review.
What a fair conclusion sounds like
Not “this works” and not “this is a scam”. Something more like: one ingredient has strong evidence at doses above this one, a second has a promising single human trial in a different preparation, the third is present at a fraction of the studied amount, and the combination has never been tested. That sentence is less satisfying than either extreme, and it is the honest shape of almost every supplement on the shelf.
Common questions
Is a study funded by the manufacturer worthless?
Not automatically, but industry funding is associated with more favourable results across many fields. Read it, note who paid for it, and weight it accordingly — particularly if it is the only trial that exists.
What does 'proprietary blend' do to evidence assessment?
It ends it. If you cannot see the dose of each ingredient, you cannot compare anything to a trial. A blend total is not a dose.
Does the absence of a trial mean a product does nothing?
No — it means nobody has tested it, which is a different statement. The honest position is uncertainty, not a verdict in either direction.
General information, not medical advice. These statements have not been evaluated by the Food and Drug Administration, and no dietary supplement is intended to diagnose, treat, cure or prevent any disease. Speak to a qualified healthcare professional before changing what you take.
See the official SlimTide offer
Affiliate link. Salimtide Evidence Room earns a commission on purchases made through it, at no extra cost to you. We are not the manufacturer.