I'm attempting to compute a Fannington-Manning power calculation in SAS.
I have a proportion of incidence of a disease based on a novel treatment at 15.7%, with the previous standard of care at 18.7%. I want to compute the power based on a non-inferiority margin of 9%; we reject null if upper side of 95% CI on Proportion treatment - Proportion control is less than 9%. The new 'treatment' has 311 patients while the standard of care would have 867.
The null hypothesis:
Null : Proportion treatment ≥ Proportion standard of care + margin Alternative: Proportion treatment < Proportion standard of care + margin
From my understanding, here is what the code should be:
proc power;
twosamplefreq test = fm
refproportion = 0.187
nullproportiondiff = 0.09
proportiondiff = -.03
alpha = .025
sides = L
groupns = (867 311)
power = . ;
run;
Based on this code I have a power of 99.6%. Is this correct?