Do you guys know what "locked" means?
In trials you "lock" the data after it has all been collected and verified before unblinding. The concept is that it is not acceptable to look for and fix errors after unblinding because there might be a bias in what errors you find, or how you decide it's an error.
What does this mean? If the numbers were not stat sig because of errors, IT'S TOO LATE TO FIX THE ERRORS FOR SUBMISSION TO THE FDA.
There is a remote chance they could pick another indepenedent CRO to analysis the raw data (blinded), verify, lock, unblind and redo the analysis. Perhapse the FDA would accept this if there was clear proof that the first group screwed up bad.
It seams highly unlikely that the (possibly errant) data is stat sig, or the company would just have released that news and dribbled the rest out later.
