Back To Index Previous Article Next Article Full Text

Statistica Sinica 24 (2014), 1411-1432

ANALYSIS OF DEPENDENTLY CENSORED DATA
BASED ON QUANTILE REGRESSION
Shuang Ji1, Limin Peng2, Ruosha Li3, and Michael J. Lynn2
1University of Texas Health Science Center at Houston,
2Emory University and 3University of Pittsburgh

Abstract: Dependent censoring occurs in many biomedical studies and poses considerable methodological challenges for survival analysis. We develop a new approach for analyzing dependently censored data by adopting quantile regression models. We formulate covariate effects on the quantiles of the marginal distribution of the event time of interest. Such a modeling strategy can accommodate a more dynamic relationship between covariates and survival time compared to traditional regression models in survival analysis, which usually assume constant covariate effects. We propose estimation and inference procedures, along with an efficient and stable algorithm. We establish the uniform consistency and weak convergence of the resulting estimators. Extensive simulation studies demonstrate good finite-sample performance of the proposed inferential procedures. We illustrate the practical utility of our method via an application to a multicenter clinical trial that compared warfarin and aspirin in treating symptomatic intracranial arterial stenosis.

Key words and phrases: Copula model, dependent censoring, empirical process, martingale, regression quantile.

Back To Index Previous Article Next Article Full Text