01 / Method
A practical evidence ladder
Different study designs answer different questions; a mechanism, an association, a controlled human outcome, and a consumer report should not be given the same weight.
Controlled human studies can test whether an intervention caused a measured outcome under defined conditions. Well-designed randomized controlled trials generally provide the clearest product-level evidence for health-benefit claims.
Observational evidence can identify associations in real populations, but association does not by itself establish causation. Animal, mechanistic, and in-vitro research can explain biological plausibility and guide future studies, but does not establish a human product benefit. Consumer experience can identify questions and preferences, but testimonials do not substantiate objective health effects.
Finished product versus ingredient evidence
A study on the finished formula is the closest match to a claim about that formula. An ingredient study may be relevant background, but its result cannot automatically be transferred to a different dose, extract, combination, delivery format, or product.
When only ingredient or mechanistic research is available, the page should label it that way and avoid presenting it as a finished-product outcome.
03 / Relevance
Match the study to the claim
The tested product, formulation, dose, route, duration, population, and measured outcome should fit the statement being made.
Evidence can be scientifically sound yet still be a poor match for a specific claim. A result in a different population, a much larger dose, a different extract, or a different outcome may not substantiate the advertised use.
We therefore ask two separate questions: “Is the study reliable?” and “Does it actually support this exact statement?”
04 / Interpretation
Read the totality, not one result
A responsible summary accounts for study quality, consistency, null findings, limitations, and the complete body of relevant evidence.
One favorable result should not be isolated from better-designed, conflicting, or non-supportive research. Sample size, controls, blinding, attrition, outcome selection, and statistical analysis all affect how much confidence a result deserves.
Uncertainty is information. When evidence is preliminary, mixed, indirect, or absent at the finished-product level, the wording should say so plainly.
05 / Publication standard
How sources will be labelled
Every reference should identify what kind of source it is and what part of the page it supports.
Vitality Lab references distinguish government guidance, controlled human research, observational research, preclinical or in-vitro work, and editorial background. Source lists are visible to readers and repeated as structured citations for retrieval systems.
A source link is not a blanket endorsement of every statement on a page. Its placement and evidence label define its intended scope.
Source register
Sources and further reading
-
01
U.S. advertising guidance
Health Products Compliance Guidance (opens in a new tab)
Federal Trade Commission. Guidance on competent and reliable scientific evidence, study design, claim fit, and the totality of evidence.
-
02
U.S. regulatory reference
Structure/Function Claims (opens in a new tab)
U.S. Food and Drug Administration. Overview of structure/function claim categories and supplement-label requirements.
-
03
U.S. consumer guidance
FDA 101: Dietary Supplements (opens in a new tab)
U.S. Food and Drug Administration. General supplement oversight, label, interaction, and responsible-use background.