Potential to have statistically significant treatment
effect estimate of -0.06 or better
•O’Brien-Fleming
efficacy boundary at third analysis:
–Terminate
if bias adjusted estimate -0.055 or better
–What is the chance of obtaining -0.06 or better at the
fourth analysis if study continues?
•If
true effect is -0.07, probability of 4.1% of BAM < -0.06
•If
true effect is -0.06, probability of 3.6% of BAM < -0.06