sample size: noninferiority and superiority in the same study

Hi all,

I'm struggling with the following: A study is planned to compare the effects of two surgical procedures (a new method vs established old method), and the aim is to evaluate both superiority (from a point of view of complications, ie the new surgical procedure is expected to have less complications), and non-inferiority (the new method should not be inferior to the old method regarding the outcome the surgery aims to cure).

Should the significance level be adjusted when determining the sample size, and if so how? I mean there are two different hypotheses tested, but the noninferiority/superiority on different things makes me unsure.

Any ideas or pointers for literature?

(Btw I tried to search the forums but I got the following error message:
This page isn’t working didn’t send any data.


Omega Contributor
The search feature is currently down, but will be address in the future with a site upgrade.

Interesting question, I would imagine a correction on alpha could be deemed appropriate (e.g., Bonferroni or just use a smaller value). Then run a sample size calculation for both tests and use the larger of the two results.