B0975
Title: Nonparametric mixture model: Application in contaminated trials
Authors: Zi Ye - Lehigh University (United States) [presenting]
Abstract: In personalized medicine, investigating the differential effect of treatments in groups defined by patient characteristics is of paramount importance. In a randomized clinical trial, participants are first classified using diagnostic tools, but such classifiers may not be perfectly accurate. The issue of diagnostic misclassification has recently become prominent and has produced severely biased estimations of treatment effects. The focus is on this problem in a pre-stratified randomized placebo-controlled repeated measures design. We develop a fully nonparametric method for estimating and testing the treatment effect for ordinal, discrete, or skewed outcomes. Consistent estimators and asymptotic distributions are provided for the misclassification error rates as well as the treatment effect. Simulation studies are conducted to compare the new method with traditional methods. The results show significant advantages of the proposed methods regarding bias reduction, coverage probability, and power.