Noyemi evaluates health claims about supplements and natural remedies by reviewing published scientific research. This page explains how we find, assess, and rate the evidence behind every claim on this site.
All studies referenced on Noyemi come from PubMed and PubMed Central, the U.S. National Library of Medicine's databases of biomedical literature. These are the same databases used by physicians, researchers, and academic institutions worldwide. We prioritize peer-reviewed, published research and clearly label any preprints (from medRxiv or bioRxiv) when included.
Not all studies carry the same weight. We assess research using a hierarchy of evidence, giving more weight to study types that are more rigorous and less prone to bias:
Highest weight: Meta-analyses and systematic reviews (pooled data from multiple trials)
High weight: Randomized controlled trials (RCTs), especially double-blind, placebo-controlled
Moderate weight: Cohort studies and observational studies with large sample sizes
Lower weight: Case studies, pilot studies, and animal/in-vitro research
Noted but flagged: Preprints (not yet peer-reviewed)
We also consider sample size, study duration, dosage used, whether results have been replicated, and potential conflicts of interest (such as industry-funded studies).
Every ingredient-condition claim on Noyemi receives one of five evidence ratings. These ratings reflect the overall body of research, not any single study:
Strong Evidence
Multiple high-quality RCTs or meta-analyses with consistent, statistically significant results in human subjects. Effective dosages are well-established. The scientific community broadly recognizes the effect.
Moderate Evidence
Several human RCTs showing positive results, but with limitations such as small sample sizes, short durations, or some inconsistency between trials. Promising but not definitive. More research would strengthen the case.
Preliminary Evidence
Limited human studies (pilot trials, small RCTs) or strong animal/in-vitro evidence suggesting a plausible mechanism. Interesting but far from proven. Results need replication in larger, well-designed human trials.
Insufficient Evidence
Very few studies, only animal or in-vitro data, or studies with serious methodological flaws. The claim may be plausible but cannot be supported or refuted with current research. We include these claims for completeness but clearly flag the limitations.
Conflicting Evidence
Multiple studies exist but reach contradictory conclusions. Some trials show benefit while others show no effect or potential harm. We present both sides and note possible explanations for the discrepancy (differences in dosage, form, population, study design, etc.).
Noyemi uses AI tools to assist with research discovery, data extraction, and initial content drafting. However, every evidence rating, claim summary, and interaction warning is reviewed by a human editor before publication. AI helps us process research at scale — it does not make editorial decisions.
All food nutrition data on Noyemi comes from the USDA FoodData Central database, a public resource maintained by the U.S. Department of Agriculture. This is the same data used by nutrition researchers, dietitians, and food labeling compliance. We use the Standard Reference and Foundation datasets for whole foods.
Our interaction warnings draw from multiple sources: published pharmacological research on supplement-supplement and supplement-drug interactions, the FDA Adverse Event Reporting System (FAERS), NIH Office of Dietary Supplements fact sheets, and pharmacology reference databases. Interactions are tagged by severity (info, caution, warning, danger) and include the mechanism and practical recommendations.
Science evolves, and so does Noyemi. When new research changes the evidence picture for a claim, we update the rating and note what changed. If you find an error or believe we've misrepresented a study, please contact us — we take corrections seriously.