Multiplicity

#1
Hello everyone!

I will plan a clinical trial, double-blinded, phase 3, that long 24 weeks, and the primary endpoint will be assessed based on results at Week 12. And I want to add one interim analysis when all patients finished week 12, and conduct a final analysis when patients finished the study (Week 24). So, Should I use methods of control the type 1 error, and if yes, what tests or methods i should use?

Thank in advance!
 

fed2

Active Member
#3
So, Should I use methods of control the type 1 error
yes

I'm not sure though what method. The usual interim analysis would be by 'group sequential' method. That requires, usually addition of new subjects at each analysis point.

I would highly recommend ponying up for statistician on this thing, especially if there is alot on the line here. You could easily invalidate the whole trial by screwing it up.