Methodology

Evidence tiers

• Strong — multiple human RCTs converging on the same effect, ideally including meta-analyses.
• Promising — limited human trials with positive but not yet replicated results, or strong mechanistic evidence with early human data.
• Traditional — long history of use and mechanistic plausibility; human RCT evidence is thin or mixed.

Every tier — including Traditional — must cite at least one source.

Cost-per-effective-dose

We translate the studied dose, the product's mg per serving, the price, and the servings per container into a monthly cost at the studied dose.

• Range handling. When a studied dose is reported as a range, we use the midpoint.
• Label cap. If a product is underdosed at the maximum label-recommended servings, we say so explicitly. We do not silently bump servings past the label.
• Missing studied dose. When no studied dose exists, the calculator is hidden for that ingredient and we say "no established studied dose."
• Stale prices. We show "price last verified [date]" when the snapshot is older than 90 days.

Editorial sourcing

Each ingredient page maintains a private claim → source map. Strong/Promising tiers require PubMed-linked human studies; Traditional accepts pharmacopoeia entries or peer-reviewed traditional-medicine references.