>If IDIX's first P3 in naives is a failure, the second attempt could well be against a new SoC which has SVR rates of 80-90%, if VRTX gets lucky, so I think IDIX needs to get it right the first time.<
The official VRTX line for VX-950 is an NDA submission in 2008 and approval in 2009, so NVS and IDIX do not have their backs against the wall just yet. I agree, however, that they will need a modicum of luck in addition to considerable skill to make this work.
Whether the best design is a two-arm trial (with more alpha for each individual analysis) or a three-arm trial (which better addresses the commercial exigencies) is apt to be a tricky decision. I would place my bet on their opting for three arms and making the trial BIG (that’s what Big Pharma is for, after all), with the proviso that the FDA assigns a generous delta for the non-inferiority analysis.
Market Data 
Markets 
AI
