Biotech Values

[DewDiligence]

AMLN—Was "Not Non-inferior" also "statistically inferior"? Was it also clinically inferior?

The answer to your first question merely requires statistical grunt work (which I have not done) and a few assumptions; the answer to your second question, however, is subjective.

Say, for the sake of discussion, that the non-inferiority margin in the Bydureon trial was 0.5%. I.e. Bydureon’s failing to achieve a finding of non-inferiority means that the “bad” boundary of the 95% confidence interval on the difference between the two drugs was a point where Bydureon was more than 0.5% worse than Victoza.

Is it clinically meaningful that Bydureon could be worse than Victoza by more than 0.5% with a probability of 2.5% or greater? I would say that it could be. Regards, Dew

p.s. The above shows why north40000’s post in #msg-60560038 is off-base; north40000 failed to realize that it’s the boundary of the 95% CI—not the difference in the means—that is the issue in this discussion.