Fosco it changes the fit by a few percent survival for the most recently dosed patients. i haven't simulated but I modified you're spreadsheet for "survival model 3" to reflect the Weibull fit and the benefit was modified pretty significantly. If the trial ended today, I get your fit producing a 20% benefit, the Weibull fit produces a 13% benefit. I haven't looked at the result with dropouts.
I recomputed the sheet "Model 3 with attrition", nominally 4% attrition, using the Weibull fit and got 5% benefit if the trial ended today. It gets to 10% in September. The Weibull fit hits the 1y and 5y Taiwan survival numbers almost exactly but is high at 3 years, almost 66% as opposed to the empirical 63%.
News 
Market Data 
Discover 
