CopulaCenR: Copula based Regression Models for Bivariate Censored Data in R

Authors

Tao Sun, Renmin University of China, University of Pittsburgh
Ying Ding, University of Pittsburgh

Date of this Version

6-2020

Document Type

Article

Citation

The R Journal (June 2020) 12(1); Editor: Michael J. Kane

Comments

Copyright 2020, The R Foundation. Open access material. License: CC BY 4.0 International

Abstract

Bivariate time-to-event data frequently arise in research areas such as clinical trials and epidemiological studies, where the occurrence of two events are correlated. In many cases, the exact event times are unknown due to censoring. The copula model is a popular approach for modeling correlated bivariate censored data, in which the two marginal distributions and the between margin dependence are modeled separately. This article presents the R package CopulaCenR, which is designed for modeling and testing bivariate data under right or (general) interval censoring in a regression setting. It provides a variety of Archimedean copula functions including a flexible two-parameter copula and different types of regression models (parametric and semiparametric) for marginal distributions. In particular, it implements a semiparametric transformation model for the margins with proportional hazards and proportional odds models being its special cases. The numerical optimization is based on a novel two-step algorithm. For the regression parameters, three likelihood-based tests (Wald, generalized score and likelihood ratio tests) are also provided. We use two real data examples to illustrate the key functions in CopulaCenR.