How it works
Review checks that the claim is reproducible, non-duplicative and separated from exposed evidence. New data or prospective observations supply the challenge.
How to read it carefully
Compare the candidate specification with the exploration that produced it. Check planned sample, stopping and promotion rules.
