How OnlineDiets evaluates nutrition claims

Nutrition evidence is rarely captured honestly by a headline. We separate the type of evidence, the outcome measured, the length of follow-up, the population studied, and the practical size of the effect.

Our evidence ladder

Systematic reviews and meta-analyses: useful for summarizing a body of research, but only as strong as the included studies.
Randomized controlled trials: useful for comparing interventions, though adherence and duration often limit interpretation.
Prospective cohort studies: helpful for long-term patterns but cannot prove causation on their own.
Mechanistic and short feeding studies: useful for understanding biology, not always for predicting long-term outcomes.
Expert consensus and clinical guidance: valuable when transparent and evidence-linked.
Testimonials and influencer claims: useful for generating questions, not establishing typical effects.

Questions we ask of every claim

Compared with what?
In which population?
For how long?
Was the outcome weight, laboratory markers, symptoms, disease events, or self-report?
How many people completed the study?
Was the difference statistically significant and practically meaningful?
Were harms, dropouts, and conflicts of interest reported?

Evidence does not remove individual variation

A diet can perform well on average and still be a poor fit for a particular person. Medication, culture, food access, digestive tolerance, cooking ability, and adherence all affect real-world outcomes.

Our core reference sources

Government health agencies and public-health guidance
Peer-reviewed research and systematic reviews
University-based nutrition resources with transparent sourcing
Professional clinical guidelines where appropriate

We distinguish direct evidence from editorial interpretation and avoid claiming certainty where the evidence is mixed, short-term, or population-specific.