Unblinding Concern: Projecting Last Event in > 2 Arm Study....

#1
In a two arm time to event study one can use various blinded projection methods using the aggregate event rate to estimate when the last event may occur.

Now, lets say one has a three arm study where two different doses of an experimental drug are compared to a control. In addition each testing pair against control needs to acquire a certain, protocol defined, number of events.

The blinded statistician can no longer project when the last event may occur based on the aggregate event rate.

So here is the question, if the unblinded statistician where to provide to the blinded group an estimate of when the last event would occur based on looking at the two aggregate event rates (i.e. Aggregate of Exp. Arm 1 and Control, Aggregate of Exp. Arm 2 and Control) would that cause unblinding concerns for the blinded group?

Any references to help investigate this or an answer would be greatly appreciated.