ABSTRACT
This chapter discusses sample size calculations for non-inferiority trials, where the objective is to show one treatment is no worse than another treatment. With non-inferiority trials, we are only concerned with one of the margins and demonstrate that the difference between investigative and control treatment is sufficiently far away from this to declare non-inferiority.
A key assumption with the sample size calculation is made with respect to the true mean difference between the two treatment arms. Even if a small difference in favour of the investigative treatment can be assumed, it will reduce the estimate of sample size for study.
If we are sufficiently far away to also be able to declare the investigative treatment is statistically superior, i.e. greater than 0, this is a good result too. Such trials are described as “As good as or better” trials and will be revisited towards the end of this chapter.
