Title: “Estimation in two-stage adaptive threshold enrichment clinical trials”
Speaker: Dr Peter Kimani, University of Warwick
Abstract: It is now widely acknowledged that a patient’s characteristics may predict how they respond to treatments. Consequently several clinical trials assess the treatment effect in the full population and in biomarker-defined subpopulations. If a biomarker is continuous, a two-stage adaptive threshold enrichment trial incorporates determining the best threshold for defining a subpopulation. Stage 1 data are used to perform an interim analysis to determine if the new treatment is promising in the full population or in a subpopulation defined by one of the candidate biomarker threshold values. In stage 2, recruitment is restricted to the promising subpopulation. In this talk I will describe how to obtain point estimators that adjust for subpopulation selection associated with including stage 1 data in the confirmatory analysis.