Topic portal·16 articles
Portal:Clinical trials
The evidence base, and how to read it
Introduction
The incretin field is unusually well served by large randomised trials, several of them powered for hard outcomes rather than surrogates. That makes it a good place to learn how to read a trial — and a good place to see how easily a well-conducted trial is misreported.
This portal covers the named programmes, the individual outcome trials, and the methodological articles that explain what an estimand, an intention-to-treat population and a number needed to treat actually are.
Picture of the month
Selected article
Blind sampling — Blind sampling draws a test sample under a procedure neither the supplier nor the purchaser controls. No report reaching this wiki arose from one, which limits what all of them support.
33 revisions since 17 September 2024. Discussion · history
Articles in this portal
Things you can do
- Add confidence intervals wherever a point estimate is currently given alone.
- Check that no post-hoc subgroup analysis is presented in an article lead.
Items here are deliberately small enough to finish in one sitting. See Project:Verifiability before adding sources.