Evidence Labels, Explained
At a Glance
- Every remedy and traditional-practice article on this site carries one of five evidence labels.
- A label describes how much is known, not how much we like the remedy.
- “Traditional practice, little clinical study” is not an insult. It is the honest answer for a great deal of what people ask us about.
- Labels are assigned by the writer and confirmed by a medical reviewer, and revisited annually.
- None of the labels means “this treats a disease”. No label on this site does.
Most health content online works in two modes. Either a remedy is presented as remarkable, or it is dismissed. Both are easier to write than the truth, which is usually that some research exists, it is smaller and messier than anyone would like, and it supports something narrower than the headline suggests.
The evidence label is our attempt to put that middle answer on the page, in a form you can read in two seconds.
The five labels
Well-studied, benefit supported
Multiple systematic reviews, or a consistent body of controlled trials, support a specific and usually modest benefit. The effect has been measured more than once, by more than one group, and it held up.
Read the specific claim carefully. “Supported” attaches to a particular effect, not to the ingredient in general. Something can be well-supported for one narrow purpose and completely unstudied for the purpose you had in mind.
Some evidence, limited studies
A small number of studies suggest a benefit, but the evidence base has real limitations — few participants, short duration, no control group, results that have not been replicated, or studies funded by a party with an interest in the outcome.
This is the most common honest label for popular remedies. It means “there might be something here, and we do not yet know.”
Traditional practice, little clinical study
Long cultural use, minimal or no controlled clinical research.
This label carries no judgement. A great deal of traditional practice has simply never been through a trial, often because there is no commercial reason to fund one. Absence of evidence is genuinely not evidence of absence — but it is also not evidence of benefit, and we will not present it as though it were.
Most of what people write to us about lands here.
Studied and not supported
Researchers examined the claim, and the evidence does not support it. This is a stronger statement than “unstudied” — somebody looked, and did not find what was expected.
We publish these deliberately. Knowing that something has been tested and did not work is useful, and it is information the rest of this content category rarely provides.
Potentially unsafe — see cautions
There is evidence of harm, meaningful interaction risk, contamination risk, or a pattern of adverse reports. The article will say what the specific concern is.
Heavy-metal contamination in some unregulated preparations, interactions with anticoagulants and glucose-lowering medication, and hepatotoxicity from certain concentrated extracts are all real, documented problems rather than theoretical ones.
What no label means
No label on this site means “this treats a disease.” Under India’s Drugs and Magic Remedies (Objectionable Advertisements) Act, 1954, no remedy may be advertised as curing the conditions in the Act’s Schedule, and we publish no such claims. A green label means the evidence supports a specific modest effect — better sleep onset, reduced nausea, improved skin barrier moisture. It never means a condition is treated. See our DMR compliance page.
The source hierarchy we use
Labels come out of a ranking. Higher tiers outrank lower ones when they disagree.
| Tier | Source type | Weight |
|---|---|---|
| 1 | Cochrane and other systematic reviews; meta-analyses in indexed journals; ICMR, AIIMS and NIN publications; WHO guidance; FSSAI regulations | Highest |
| 2 | Randomised controlled trials in indexed journals; government health portals; peer-reviewed observational studies | Substantial |
| 3 | Classical texts (Charaka Samhita, Sushruta Samhita, Ashtanga Hridaya); attributed practitioner interviews | Context only — never clinical evidence |
Tier 3 is the important one. Classical texts tell us what a tradition holds and how a practice has been used for centuries. That is genuine, valuable knowledge, and we cite it. It is not clinical evidence of efficacy, and treating it as though it were is the central epistemic failure of this whole content category.
So in our sources block, classical citations are visually distinguished from research citations. You can always see which kind of claim you are looking at.
We do not cite as evidence: other content-mill health sites, supplement brand marketing, social media claims, anecdote, Wikipedia (useful as a gateway, not as a source), or AI-generated summaries.
Who assigns a label
The writer proposes a label while researching. A medical reviewer confirms or changes it before publication. Where they disagree, the more cautious label wins.
Reviewers can decline to approve an article. Declined articles do not publish.
When labels change
Evidence moves. Labels are revisited when a health-substantive article comes up for its annual re-review, and immediately whenever a significant new systematic review is published on a topic we cover.
A changed label is a correction, and it is logged at our corrections page like any other. If a remedy moves from “some evidence” to “studied and not supported”, we would rather say so plainly than quietly delete the article.
How to read a label critically
- Check what the claim is attached to. “Supported” for reducing nausea in pregnancy says nothing about any other use.
- Check the quantity studied. Trials often use concentrations far above culinary amounts. Cooking with a spice is not the same intervention as a standardised extract.
- Check the population. A result in one group does not automatically transfer to another.
- Read the cautions. A supported benefit and a real interaction risk can coexist in the same ingredient.
Why we publish the method
A label is only worth something if you can see how it was arrived at and disagree with it. Publishing the hierarchy, the process and the review step means you can check our reasoning rather than take our word — which is the same standard we apply to everything else we write about.