Our Standards

How we decide what counts as verified — and what each label means.

These three axes are independent. We never infer one from another: strong evidence is not the same as approval, and approval is not the same as a guideline recommendation.

1. Evidence stage

How strong the underlying human evidence is.

2. Treatment / regulatory status

What regulators have actually decided for the stated cancer and use.

3. Care / guideline status

What clinical guidelines say.

Plain-language glossary

Observational evidence
Shows association, not cause and effect. We require a non-causation note on observational stories.
Surrogate endpoints
A stand-in measure (like a lab marker) that may not translate into living longer or better. We flag it.
Preprints
Posted before peer review. We label them as not yet peer-reviewed.
Conference abstracts
Early, brief, often not peer-reviewed. We label them.
Clinical-trial access
The treatment is available through a trial. Eligibility and enrollment are never guaranteed.
Corrections
A published fix to the source. We record it and show it.
Retractions
A source withdrawn by the journal. A retracted load-bearing source blocks verification and suppresses the evidence label.
Last-reviewed date
When a human editor last checked the story against its sources.
Primary vs secondary sources
Primary = the actual study, registry, regulator, or guideline. Secondary = press releases and news, which are context, not evidence.
Funding / conflicts
Who paid for the research and any competing interests.
What This Means / What This Doesn't Mean
Plain-language statements of the finding and its limits, required on every verified story.

See the full Understanding Research glossary for more terms