Statistics Seminars: Data combination in seamless Phase II/III clinical trial designs
21 February 2011 15:15 in CM221
In a seamless Phase II/III clinical trial, one of several treatments or doses is selected in Phase II for further study in Phase III. The final decision rule for declaring the selected treatment superior to control must protect the family-wise type I error rate for comparisons of all treatments against the control. When a combination rule is applied to P-values from data in the two phases, we find that overall power does not always increase beyond that obtained by simply ignoring the Phase II data! This raises the question of how one should combine information to maximise power while protecting family-wise type I error.
We present a formulation of the problem which is amenable to analysis by decision theory. Hence, we derive optimal data combination rules for particular objectives and find some surprising results. We conclude by identifying decision rules with robust efficiency across a variety of scenarios and, hence, quantify the effective information obtained from the Phase II data.
Host: Allan Seheult
Contact firstname.lastname@example.org for more information