How we assess evidence.
The seven-step evidence process AES uses to evaluate ingredients, formulations and claims — from question definition through approved public language.
Define the question
Every review begins with a written question capturing ingredient, ingredient form, quantity, population, intended outcome, duration, comparison and relevant safety considerations.
Search the evidence
Searches record the date, search terms, databases, inclusion and exclusion criteria, reviewer and next search update date.
Assess relevance
Studies are only credited to a claim when they match the exact ingredient, chemical form, quantity, product format, population, duration and outcome.
Assess study quality
Design, sample size, comparator, blinding, duration, outcome quality, attrition, funding, conflicts of interest, replication, statistical limitations and generalisability are recorded per study.
Separate evidence levels
These are internal evidence classifications, not clinical grades.
Document limitations
Every summary states what is known, what remains uncertain, whether the exact product was studied, whether the quantity matches, whether the population is comparable, key safety limitations and the date of review.
Approve public language
Claims are only published with a source record, evidence classification, regional approval, scientific review, regulatory review and a reassessment date.
Read the evidence framework →
A companion page defines each classification, the review cadence and the correction process.