There are five common prejudices that people have about comparing things in comparison to each other. These prejudices include: Lack of transparency, context, inferiority bias, and lack of detail. Luckily, these issues are easily addressed. Whether you are a fan of comparisons or a skeptic, there are ways to avoid these common errors.
In an Amazing Comparison, the test drug should be significantly better than the placebo. The lower bound of the C-T confidence interval should be lower than the NI margin, which is specified in advance. Alternatively, the M margin can be set to the entire known effect of the active control over the placebo.
The study employed modified intention-to-treat (mITT) and noninferiority criteria. The authors used the noninferiority margin of 0.5 units for the two groups. These methods are generally more conservative in estimating noninferiority, since they protect against the possible violation of the constancy assumption.
The CONSORT Statement recommends that the margin of safety (CI) of a comparison study be presented with the margin of noninferiority. The margin should reflect the effect of the comparator over the placebo. This margin must account for several factors and assumptions, and it is often calculated using a point estimate or a fixed margin.