|Title||In Hochberg’s tradition: Selective Inference for Clinical Trials|
|Year of Publication||2014|
I shall review my joint work with Yosi Hochberg on the emergence of the False Discovery Rate (FDR) and his motivating example from clinical trials. I shall explain why the FDR has not found its way into the statistical analysis of clinical trials, and demonstrate that problems of selective inference manifest in earlier stages of medical research in spite of that avoidance. I shall then present the weighted false discovery rate as and a new hierarchical way of controlling it. It allows researchers analyzing Phase I and II clinical trials to address selective inference when facing primary and secondary endpoints, as well as when confronting heterogeneous treatment effect.