Evidence methodology

Candipedia Research evaluates an exact claim, outcome, and population. A single paper can contribute to a conclusion but does not receive or determine a body-of-evidence level by itself.

Source hierarchy

We prefer current clinical guidelines and public-health or regulator sources for care and safety rules; systematic reviews and meta-analyses for synthesis; and controlled trials and well-designed observational studies for specific effects. Laboratory, animal, small pilot, expert-opinion, and anecdotal material can provide context but cannot be presented as established human benefit.

Sources are checked for identity, study design, population, applicability, date, funding or conflicts when reported, and correction, retraction, withdrawal, or supersession status. Important conclusions link to their underlying sources.

Evidence levels

• Strong: consistent, directly applicable evidence with low enough uncertainty to support a stable conclusion.
• Moderate: useful human evidence supports the conclusion, but important limitations or uncertainty remain.
• Early: preliminary, indirect, small, or not yet replicated evidence that is not ready to carry a firm practical conclusion.
• Unclear: evidence is insufficient, conflicting, or too poorly applicable for a dependable conclusion.

The benefit or effect verdict, strength of evidence, and safety status are separate judgments. Uncertainty and meaningful disagreement are shown beside the conclusion they affect.

Updates and reassessment

Health conclusions are reviewed by their stated review date and sooner when a relevant guideline, regulator notice, product-access change, major study, correction, retraction, source expiry, broken source link, or safety signal appears. A reassessment may record that the practical answer did not change. Material changes and corrections are dated in the public changes or corrections log.