Account

Fact-Checking Policy

Effective Date: April 20, 2026

Why Fact-Checking Matters Here

RRM Academy is a medical education platform. Readers use our content to understand their health, evaluate treatment options, and make decisions with their clinicians. A hallucinated statistic or a fabricated citation is not just a quality problem -- it is an existential threat to a medical education site. Our fact-checking process is built accordingly.

What We Verify

Before publication, every article is checked for:

  • Citation existence. Every DOI, PMID, and URL is confirmed to resolve to the claimed source. We do not insert citations from model knowledge alone.
  • Citation accuracy. Every cited statistic is confirmed against the actual source. The claim in our article must match what the source says.
  • Statistic integrity. Numbers, percentages, sample sizes, and effect estimates are traced to their primary source.
  • Clinical accuracy. Treatment descriptions, diagnostic criteria, and disease definitions are reviewed by a credentialed clinician.
  • RRM alignment. Content is reviewed to ensure it reflects restorative reproductive medicine principles and does not misrepresent RRM practice.

How We Verify

Citation verification uses a multi-API cascade. Each cited URL passes through:

  1. HTTP resolution. The URL must return a live response.
  2. Source content check. The target page is fetched and the cited claim is located within it. Live sources beat memory.
  3. Bibliographic API cross-reference. For journal articles, metadata (title, authors, year, DOI) is cross-checked against CrossRef, PubMed, and Semantic Scholar.
  4. Rendering fallback. Bot-protected pages are verified through a headless browser when the direct fetch fails.

Any citation that fails this cascade blocks publication. Our CI pipeline enforces this automatically: scripts/verify-citations.mjs runs on every deploy and halts the build if any citation fails.

Research Library Integrity

The RRM Research Library holds more than 3,000 peer-reviewed articles. Every record is enriched from authoritative bibliographic APIs (CrossRef, PubMed, OpenAlex) and includes verified metadata, DOIs, and, where available, ORCID identifiers for authors and ROR identifiers for their institutions. Retracted articles are flagged through a weekly automated retraction scan and labeled on their library pages.

Clinical Review

Clinical content is reviewed by a credentialed RRM clinician before publication. Pillar guides receive review from our editorial lead, Dr. Naomi Whittaker, or from a designated reviewing clinician. Patient-facing content is additionally reviewed for voice and for alignment with the Editorial Policy.

What Happens When We Get It Wrong

Errors are corrected under our Corrections Policy. If you believe a claim on this site is not accurately sourced or does not match what the cited source says, please contact us and we will investigate.

What We Do Not Do

We do not generate citations from model knowledge. We do not cite sources we have not read or resolved. We do not infer statistics from memory. When a claim we want to make is not supported by a verifiable source, we either omit it or say explicitly that the evidence is limited.

Related Policies