Sources and how we grade
Citations last checked: October 29, 2026
What this page is
Every supplement in your results carries an evidence label: Strong, Moderate or Limited. This page says where that evidence comes from, how a label is decided, and what we do when a study is withdrawn. It is written in plain words; the papers themselves are listed under each supplement in the app.
How a label is decided
A label is given for one supplement and one use at a time, for example magnesium for falling asleep, not magnesium in general. The best studies on that exact question are read twice, by a grader and by a second reader whose job is to doubt, against the same written standard:
- Strong: a meta-analysis (a study that pools the results of many trials) of randomized controlled trials, or at least three independent trials, measured the thing you picked, in people like you, with a consistent and clear result, and no later pooling contradicts it.
- Moderate: a benefit was shown, but with one weakness: small or short trials, mixed results, a stand-in measurement instead of the thing you feel, a benefit only in a subgroup (for example only in people who were deficient), or a medical guideline without strong trials behind it.
- Limited: single or small trials, or early and mixed pooled results.
- No label: no human trials, or pooled results that found no benefit. These uses are never shown as a reason.
The label says how sure the evidence is, not how big the benefit is. A supplement can have strong evidence for a small effect; where we know the size of the effect we say so separately. A few rules keep the standard honest: trials are counted as trials, not papers, so two papers about one trial, or two poolings of the same trials, do not count twice; a paper that has been retracted or placed under an expression of concern supports nothing; when a pooled analysis is updated, the label follows the newer result; preprints, trial registrations without a published paper, conference abstracts and reference books never decide a label; and a guideline or regulator opinion can support Moderate at most, never Strong.
Safety works the other way round. A benefit needs strong proof before we claim it; a warning is given on weaker proof. Warnings about medicines, pregnancy and upper limits come from drug labels, regulator statements, NIH fact sheets and published reports, and they never add to a supplement’s score.
Where the research comes from
PubMed. Every paper behind a label is indexed in PubMed, the database of the U.S. National Library of Medicine (NLM) at the National Institutes of Health. We read the papers’ abstracts and conclusions ourselves; PubMed supplies the records, the publication type (trial, meta-analysis, review) and any notice attached to a paper, such as a correction or a retraction. We fetch these records through NLM’s public E-utilities. PubMed and the National Library of Medicine do not endorse or recommend Find Your Supplement or any product; NLM’s disclaimer and copyright notice applies to the data we use from it.
ClinicalTrials.gov. Trial registration numbers (the identifiers that let us tell one trial from another when it is published more than once) and trial records come from ClinicalTrials.gov, a resource provided by the U.S. National Library of Medicine. Data from ClinicalTrials.gov processed on October 29, 2026. We claim no ownership of any part of that database.
NIH Office of Dietary Supplements and NCCIH. Dosage guidance, safe upper limits and some interaction notes follow the fact sheets of the NIH Office of Dietary Supplements (ODS) and the National Center for Complementary and Integrative Health (NCCIH), which we link to under each supplement. These are public U.S. government publications. Linking to them, or quoting them, is not an endorsement by NIH, ODS or NCCIH of Find Your Supplement or of any product we show, including any product you can buy through our links.
FDA drug labels (DailyMed). Medication warnings quote what the medicine’s own FDA-submitted label says, as published on DailyMed by the National Library of Medicine, and say so in each warning.
Retraction checks. Each quarter, and whenever a label changes, we check every paper we cite against PubMed’s own notices, the Retraction Watch Database (Crossref) and OpenAlex. Retraction Watch data: Retraction Watch Database, Crossref.
When a study is withdrawn
A retracted paper, or one under an expression of concern (a journal’s public notice of doubt), supports nothing. The moment our check finds one, the paper is dropped from every label it supported, the label is decided again without it, and the row is held for a person to review before anything ships. A paper with a published correction stays, with a note to re-read the corrected figure. A pooled analysis that has been updated is replaced by the newer version, and the label follows the newer conclusion.
Reviewed each quarter
Every quarter we search for newly published trials and meta-analyses on each supplement and use we cover, re-run the retraction check, and re-check doses and upper limits against the current NIH fact sheets. Any label that changes is reviewed by a person before it reaches you. Any number we state about the research is one that a dated run produced. We are taking study counts out of our descriptions; a count you may still see in an older app store listing comes out with its next update.
Your data never goes to these sources
All of this reading and checking happens ahead of time, on our side. The app and the website only download the results. Your quiz answers, lab values, DNA data and medicines are never sent to PubMed, ClinicalTrials.gov, NIH, the FDA, Crossref, OpenAlex or any other research source. Our privacy policy says what we do keep.
Questions
Ask us at support@findyoursupplement.com. To see which supplements we cover, read Supplements we review; to see how a result is scored, read How it works.