It is puzzling but, in my understanding, un-blinding the data...

It is puzzling but, in my understanding, un-blinding the data compromises the integrity of a study. The issue came up recently with ACW and their interim review for the Xanadu phase2 trial, the CEO Dr Ketelbey had this to say to Alan Kholer:

Any sneak previews at this point?
We’d love to know, wouldn’t we? We did an interim analysis of the first 50 patients on the trial back in May and that was done on both efficacy and safety data. It was important that we maintained the blinding of the trial, maintained the integrity of the trial. We had to be sure that no patients, investigators, research staff or us for that matter, the company, became aware of who was on what treatment. The interim analysis was done by an independent body called the Data Safety Monitoring Board who had an independent charter and their brief was to unblind the data, analyse it and comment back to us with what their findings were. Their conclusion, their recommendation, was to continue the trial without modification. That’s probably the best that we could hope for because what it indicated was that all the assumptions that we had made around design of the trial, statistical powering of the trial, outcome assumptions, that generally they appeared to be correct and that the trial in its current form was progressing as expected.
We were very pleased with that outcome. It doesn’t indicate obviously whether the drug works or not, it does give a reasonable indication that safety is not a problem but we will know whether our drug works and whether the trial has proved that our drug works we will know that at the end, only at the end.

https://hotcopper.com.au/threads/an...4423508/page-23?post_id=35784494#.W8b5rTFRdaQ