Thats not what it was. It said 85% chance of identifying a statistical difference when the trial is complete (interim analysis of the first 29 participants).