How we review content
Trust is the entire point of this project, so accuracy is enforced, not just intended. Alongside human editorial review, every page must pass an automated content-review step before it can publish. This page explains what that check does.
The automated content review
Our build pipeline runs a strict self-review script over every page. A page that fails any of the following checks cannot be published until it is fixed:
- No placeholder or filler text — drafts containing dummy content are blocked.
- No overstated medical claims — absolute or sensational language, such as promising a wonder cure, total absence of risk, or a treatment that works for everyone, is rejected.
- Statistics must be supported — any percentage or research claim must sit next to a citation.
- Citations must be authoritative — outbound source links are restricted to an allowlist of trusted medical and scientific domains; a link to any other domain fails review.
- No implausible references — for example, future-dated studies are flagged.
- The medical disclaimer must appear on every page.
Human editorial review
Automated checks catch mechanical problems; people catch nuance. Our editorial team reviews content for clinical accuracy against current guidance from authoritative bodies and records a last-reviewed date on each page. We welcome corrections — see our contact page.
Why this matters.
Automated checks reduce the risk that an error or an “AI hallucination” reaches a
reader. They are a safety net, not a replacement for professional medical judgment.
Sources
- MedlinePlus (U.S. National Library of Medicine). Evaluating Health Information. https://medlineplus.gov/evaluatinghealthinformation.html