Rawnoc---recall this from Siavoche's blog:
"As some of you know, many (including Mangrove Partners' Nathaniel August and several others) have tried "modeling" the HEAT study using enrollment data and events recorded at the interim analysis date and now for 380. According to some, if the control arm is truly 12 months, the separation in the curves could be fairly large due to unconfirmed guesses/estimates by many that the pooled median for the entire trial is in the 20-25 month range. I try not to speculate at all in this area (everything is purely just that, speculation based on fancy excel spreadsheets), as it is not an area of expertise of mine nor do I think such models carry much utility given all the variables that need to be accounted for. Sorry for the digression on this point."