Inference on treatment-covariate interaction based on a nonparametric measure of treatment effects and censored survival data

Shan Jiang; Bingshu Chen; Dongshengn Tu

doi:10.1002/sim.6907

Inference on treatment-covariate interaction based on a nonparametric measure of treatment effects and censored survival data

Stat Med. 2016 Jul 20;35(16):2715-25. doi: 10.1002/sim.6907. Epub 2016 Feb 16.

Authors

Shan Jiang¹, Bingshu Chen², Dongshengn Tu²

Affiliations

¹ Department of Mathematics and Statistics, Queen's University, Kingston, Ontario, Canada.
² Canadian Cancer Trials Group, Queen's University, Kingston, Ontario, Canada.

PMID: 26887976
DOI: 10.1002/sim.6907

Abstract

The investigation of the treatment-covariate interaction is of considerable interest in the design and analysis of clinical trials. With potentially censored data observed, non-parametric and semi-parametric estimates and associated confidence intervals are proposed in this paper to quantify the interactions between the treatment and a binary covariate. In addition, comparison of interactions between the treatment and two covariates are also considered. The proposed approaches are evaluated and compared by Monte Carlo simulations and applied to a real data set from a cancer clinical trial. Copyright © 2016 John Wiley & Sons, Ltd.

Keywords: biomarker; clinical trial; confidence interval; density ratio model; empirical likelihood; interaction; non-parametric inference.

MeSH terms

Humans
Monte Carlo Method*
Survival Analysis