Purpose of this methodology
The methodology exists to make supplement claims auditable. Readers should be able to understand why one claim is described as strong, another as preliminary, and another as unsupported. The process is designed for evidence-led editorial work, not for replacing clinical guidelines or conducting a formal systematic review unless a page explicitly follows such a protocol.
Source hierarchy
Regulatory and authoritative safety sources are prioritized for labeling, interactions, intake limits and public-health guidance. Systematic reviews and meta-analyses of relevant human studies carry more weight for efficacy than isolated trials. Individual randomized trials remain useful when they address a specific formulation or population. Manufacturer material is used for factual product information, not as independent proof of efficacy.
Evidence labels
Strong evidence requires a substantial and reasonably consistent human evidence base. Moderate evidence can support a cautious conclusion but may still have important limitations. Preliminary evidence reflects an early or small signal that needs replication. Insufficient evidence means the available research does not justify a confident benefit claim. A safety concern is tracked separately because an ingredient can have efficacy evidence and still be inappropriate in certain contexts.
Claim classification
An established fact, a study finding, an editorial interpretation and a manufacturer claim are not the same thing. Pages should identify which type of statement is being made. A study finding should preserve the study population and outcome; an editorial interpretation should not be written as though it were the authors' conclusion; and manufacturer statements should be attributed where relevant.
How dose and formulation are handled
Evidence is mapped to the form and amount actually studied whenever possible. A branded extract does not automatically validate a generic extract. A mineral salt may provide a different elemental amount from another salt. A study dose is treated as research context rather than a universal recommendation. If a product hides individual amounts in a blend, the uncertainty is retained.
Safety review
Safety is reviewed alongside benefit. Relevant issues include upper intake limits, medication interactions, contraindications, adverse effects, medical conditions, duplicate exposure across products and documented contamination concerns. FDA and NCCIH sources are used where appropriate because these facts can change and are more authoritative than secondary summaries.
Product reviews are desk research by default
Unless a page explicitly states otherwise, product reviews do not imply personal use, laboratory testing or clinical testing by this site. Desk research can assess labels, evidence alignment and documented quality claims. It cannot establish taste, tolerability or an individual's physiological response.
Commercial relationships and editorial separation
If affiliate or commercial relationships are introduced, they should be disclosed in a way that allows readers to understand the relationship. Commercial eligibility should not silently determine evidence labels or create a stronger efficacy conclusion. Ranking pages should use disclosed criteria rather than opaque commercial ordering.
Corrections and updates
Material factual errors should be corrected when identified. Time-sensitive information such as product formulations, prices, regulatory notices, safety warnings and new systematic reviews should be rechecked before publication and periodically thereafter. An older page should not preserve a strong conclusion if newer, higher-quality evidence materially changes it.
Applying the evidence to a real decision
A useful decision process has three layers. First, define the intended outcome precisely enough that it can be matched to research. Second, inspect the actual product or ingredient information—form, amount, serving size, duration and relevant safety details. Third, ask whether the population studied resembles the intended user and whether the outcome measured is meaningful for the goal. This prevents a common error in supplement content: moving from a plausible mechanism or a positive result in one narrow setting to a universal consumer promise. When evidence is uncertain, the uncertainty itself is decision-relevant. It may justify choosing a better-studied alternative, seeking professional guidance, or deciding that the expected benefit is too small or uncertain to justify the cost and complexity of another supplement.
How conflicts and uncertainty are written
When credible studies disagree, the content should not hide the disagreement by selecting only the result that supports a cleaner headline. The page should describe the direction of the evidence, relevant differences in population or formulation, and why the conclusion remains uncertain. Where a meta-analysis is available, its limitations still matter; pooling weak or heterogeneous trials does not automatically create strong evidence. Editorial language should become more cautious as heterogeneity, risk of bias or indirectness increase.
What AI assistance may and may not do
AI tools may assist with organization, extraction, comparison, drafting or quality checks, but they should not be treated as a source of scientific authority. Source citations must resolve to authentic materials, and key claims should be checked against the underlying publication or authoritative guidance. AI-generated wording should not invent study details, product facts, professional credentials or consensus positions. Human editorial responsibility remains necessary for deciding whether the final claim accurately reflects the evidence and the site's published methodology.
Why page-level citations matter
A methodology page can explain the hierarchy, but each substantive page still needs citations close enough to the claim that a reader can verify the basis. Source lists should not function as decorative bibliographies. The cited paper or official guidance should actually support the statement being made, and conflicting or newer evidence should not be omitted merely because an older source produces a cleaner narrative.