Analyzing a Biomarker As A Proof Of Concept

Biomarker Analysis with the Mixed Modelling Methodology

Two doses of a therapeutic vaccine were tested simultaneously in a randomized phase II trial. While the trial’s primary endpoint was not sensitive enough to show efficacy of either dose or any difference between the two, using the advanced statistical methodology of mixed modelling on repeated measures, it was found that the high dose had a highly significant effect on a relevant biomarker, while the low dose had no such effect. This proof of concept allowed the sponsor to raise additional funding to support further trials.