Non-inferiority margins: how they are chosen and how they are abused — a second dataset posts 31–38

Risk of bias: structured appraisal of internal validity. Key things to assess: randomisation method (was it truly random or could someone predict the next assignment), concealment (could randomisation be subverted), blinding (who was blinded and why or why not), completeness of outcome reporting.

Everything in post #30 holds. The case it does not cover is the one I have.

I have no financial interest in anything named in this thread and I want to say so before I comment on Non-inferiority margins, because it is the sort of subject where it matters.

Practical answer on Non-inferiority margins, since the theoretical one is upthread: do the simplest check first, write down the result, and only then decide whether the complicated explanation is needed. It usually is not.

Non-inferiority margins are chosen, and the choice is an argument rather than a fact. A wide margin can make a worse treatment look acceptable.

That is what the documentation says. What happens in practice is usually close.

I read post #32 twice before replying, because I had assumed the opposite.

The failure mode on Non-inferiority margins is boring rather than dramatic. It is almost always the step everyone assumes was done correctly because it is too simple to get wrong.

Subgroup analyses are hypothesis-generating unless pre-specified and adequately powered, and almost none are the second. The interaction test matters more than the subgroup point estimate.

I checked the source rather than the summary, and they differ.

I would call the community position on Non-inferiority margins likely rather than established, and I would be comfortable defending that hedge.