Skip to main content
Cancer Explained
Donate

Where our information comes from — and how much of it you can check.

This site does not originate medical claims. It restates information published by federal health agencies — overwhelmingly the National Cancer Institute — in plainer language, and points you at the original so you can read it yourself. That is the whole basis on which we are asking you to trust anything here.

A promise like that is only worth something if you can test it. So rather than describe our standards, this page publishes the measurements — including the one we are not happy with. The numbers are generated from the pages themselves, not written by hand, and the build fails if they fall out of date.

Can you check what a page tells you?

Of 2,250 guides on cancer and its care:

1,887

cite a specific document — a direct link to the page the information came from. You can open it and compare. That is 84% of our guides.

363

name the publisher but not the document — a link to a section or front page rather than the source itself. You can see who says it, but not easily where. This is the gap, and it is 16% of our guides.

We would rather tell you this than not. A citation that reads “cancer.gov” is the name of a library, not a reference. The information on those pages still came from the National Cancer Institute and we have not invented anything — but you cannot verify that as easily as you should be able to, and until you can, that is our problem and not yours.

We are working through that list by hand, one page at a time: reading the article, finding the document that covers it, and checking the document really does say what the page says before recording it. Of the 363 still to resolve, 346 have not been looked at yet and 17 have been checked and have no suitable document — there simply is no federal page on, for instance, adjusting to life with a colostomy, or cancer screening for transgender people. We keep those two numbers apart because one is a backlog and the other is a ceiling, and reporting them together would flatter us.

The tempting shortcut here would be to guess a plausible NCI address for each page and watch the number go green. We are not doing that. A citation nobody checked is a worse lie than a vague one, because it looks like proof. The count above can only move upward, and the build fails if it does not.

Who publishes what we draw on

88% of our clinical guides trace to a federal government source. The remainder are chiefly international scientific bodies — for instance the IARC monographs, which are the actual authority on whether a substance causes cancer.

  • cancer.gov1,866 pages
  • cancer.org56 pages
  • monographs.iarc.who.int22 pages
  • fda.gov19 pages
  • acr.org16 pages
  • prescancerpanel.cancer.gov13 pages
  • medicare.gov12 pages
  • radiologyinfo.org12 pages

The pages that are not clinical guides

990 pages are biography and reporting — the story of a public figure’s illness, or how cancer has been portrayed on screen. These draw on journalism and reference works rather than clinical sources, because no federal health agency publishes the details of a musician’s diagnosis. Where such a page explains something medical, that explanation links to a guide that is federally sourced. We keep the two kinds of page apart deliberately, and count them separately here.

What we do not claim

  • That a doctor wrote this. Pages are drafted by an AI-assisted editorial system from the sources above. Where a page carries a “medically reviewed” label, a named reviewer checked it on the date shown. Most pages do not carry that label, and we will not pretend otherwise. See how we use AI.
  • That this is advice for you. Everything here describes what is generally true. Your situation has details no general page knows.
  • That we never get it wrong. We do. If you find an error, please tell us — corrections are logged publicly at /corrections.

Every figure on this page is computed from the site’s own content each time it is built, and a check refuses the build if this page disagrees with reality or if the number of verifiable citations goes down. Related reading: how we use AI, our editorial process, source methodology, and the medical disclaimer.