Wonder if Dewophile, Clark or io have comments on any of those responses.
None right now. I need to look at previous similar trials and duplicate some of the math - but note the very precision of the answers implies a lot more thought went into this trial design than many in biotechland. The only other comment is to agree with fudoo that the power calculation for a trial presupposes randomness - no single factor bias compared to previous trials that are the basis for the powering calculation. It is why the FDA doesn't like single center trials. But how much of a risk this is is completely uncalibrated by me since I haven't seen a trial fail from this.
PS Sometime in the next month I will be looking at the bigger, more important trials. Just not yet.
PPS Part of the reason for looking at all of these trials is calibrate management. So far so great.
Clark
Market Data 
Markets 
AI
