Editorial standards, transparently.

The Heme editorial team.

Every article on Heme has a byline. Every reviewer is named. Conflict-of-interest disclosures live on contributor pages. Bios below are placeholders during launch — the named team builds out across Q2–Q4 2026.

Medical reviewers.

Any page making a clinical claim — biomarker explanations, symptom hubs, test interpretations — is reviewed by a qualified clinician before publication, and re-reviewed at least annually. We disclose the reviewer's name, credentials and review date on the page itself.

Want to be considered as a reviewer? Email editorial@heme.co with your credentials and the area you'd like to cover.

The seven-criterion review framework.

Every at-home test, comprehensive panel, telehealth platform and supplement we cover runs through the same seven criteria. Each criterion is scored 1–10. The total weighted score becomes the Heme Score. We publish the score whatever it is — no editing for commercial outcome.

The full scoring rubric — what each 1–10 looks like, how the weights work, how ties are broken — lives on methodology.

Fact-checking.

Every medical claim in a Heme article is traceable to a source. We rely on peer-reviewed research, national clinical guidelines, and primary lab documentation — not other wellness blogs.

Sources we use.

• Peer-reviewed journals. NEJM, JAMA, The Lancet, Obstetrics & Gynecology, Thyroid, Journal of Clinical Endocrinology & Metabolism, and topic-specific titles.
• National clinical guidelines. ACOG, the Endocrine Society, the American Heart Association, the British Menopause Society, NICE, the WHO.
• Primary lab and assay documentation. Reference-range methodology directly from the lab or assay manufacturer.
• Patient-led associations where they synthesise the above (e.g. Iron Disorders Institute, North American Menopause Society).

What we don't cite.

• Other wellness blogs as a primary source for medical claims.
• Brand-owned research presented as independent.
• Single small studies framed as settled science.
• Influencer commentary as evidence — even from credentialed influencers.

When evidence is mixed or evolving — which it often is in women's health — we say so. We use compliance language ("may", "could", "worth discussing with your provider") rather than overclaiming.

Corrections policy.

Mistakes happen. When they do, we treat the correction as part of the editorial — not something to bury.

• Public corrections. Every correction is published at the top of the affected page with the date and a short description of what changed.
• Version history. Every page tracks its "Last updated" date. Major revisions are logged.
• Two-person rule on score changes. A published Heme Score cannot be revised by a single editor. Any post-publication change requires the Reviews Editor plus one other editor, with the reason logged.
• Reader-flagged corrections. Spotted something wrong? Email corrections@heme.co. Every message is read by a senior editor. If we got it wrong, we correct it publicly.

The editorial firewall.

The thing that makes everything else credible. Two principles, in plain English:

Conflicts of interest.

We don't pretend we have none. We make money — the full picture is on how we make money. What we owe readers is disclosure, not denial.

• Affiliate links. Disclosed on every page where they appear. Editorial coverage is independent of commission rate. When a low-paying product wins on the framework, it gets the pick.
• Heme Reviewed badge program. Brands pay an annual licence to display the badge after passing the framework. Badge is revocable and re-assessed yearly. The fact that a brand pays for the licence is disclosed on their review.
• Sponsored reviews. A brand can pay to be reviewed. They cannot pay for the score. Sponsored reviews are labelled "Sponsored review" at the top of the page and go through the same seven-criterion framework.
• Newsletter sponsorships. One sponsor per edition of The Depletion Report. The sponsor section is labelled. The editorial section is independent.
• Lead-gen partnerships. Where the Heme Quiz routes a reader to a partner clinic, we disclose that the partner pays per lead. We don't route on payment — we route on geographic and clinical relevance.
• Personal disclosures. Individual contributor conflicts (consulting work, board roles, equity) live on contributor pages.

AI use.

The honest version: yes, drafts may be assisted by AI tools. Research, outlines, first drafts and copy-edit suggestions sometimes start in an AI workspace.

Three guardrails, every time:

• Humans write and review every published page. No AI-generated text is published without a named editor working it over and signing it off.
• Clinical pages are reviewed by clinicians before publication. AI does not have the final word on a biomarker claim, a reference range, or a women's-health recommendation. A qualified human does.
• Sources are verified by humans. AI tools can hallucinate citations. We confirm every cited study, guideline and source against the primary record before it goes up.

If you're reading a Heme page, you can be confident a person took responsibility for it.

Reader feedback.

Our standards are sharper because readers tell us what's wrong, what's missing, and what the framework doesn't yet capture. If you've found a mistake, want us to cover something, or think the framework is wrong about a brand you've used — write to us.

• Editorial feedback: feedback@heme.co
• Corrections: corrections@heme.co
• Anything else: heme.co/contact

— Hannah Reilly, Reviews Editor