(I didn't consider those figures because I wanted a conservative analysis but it is even better if they do indeed reflect the study population)
Btw the old sheet link is still valid but is available on google request, because I think it holds obsolete assumptions to be widespread, I prefer the new sheet
What to say then : This kind of survival curve must imply an even better OSI of Multikine group or a high dropout rate, but even the higher dropouts rates imply efficacy, drop outs would have to be well above 20% to imply so much delay in 298th occurrence