Biotech Values

[iwfal]

Everyone says how wonderful alexion trial was. They had an 18 percent reduction of events with a p value of .08 in a 3 or 4 thousand patient trial

In the primary pre-pivotal trial in CABG (PRIMO-CABG):

a) It was about a 24% reduction in MI/Death

b) The 30 day p value was 0.01


Now for the poser - if you had two trials, one of 1500 per arm with a p value of 0.01 and the other with two treatment arms of 300 (two arms allows some cherry picking) and p value of 0.028, which would you say was more likely to be repeatable (p value under 0.05 when trial is duplicated). Answer is, I believe, a. It is the p value that matters and Alexion had the better p value and yet failed utterly to repeat. This is the point of the warning - it would, IMO, be wise to understand why the Alexion trial failed. Change in SoC? Change in endpoint definition? And then know that MCU isn't going to have the same failings.