Quantile Regression Models for Interval Censored Data

• Activity Number: 604
• Type: Contributed
• Date/Time: Thursday, August 7, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Fang-Shu Ou*+ and Donglin Zeng and Jianwen Cai
• Companies: University of North Carolina at Chapel Hill and University of North Carolina at Chapel Hill and University of North Carolina at Chapel Hill
• Keywords: Coarsening at random ; Concave-convex procedure ; Interval censored data ; M-estimation ; Quantile regression ; Subsampling

Abstract

Interval censored Data data arise frequently in demography, epidemiology, and econometrics where the exact failure time cannot be determined but is known only to have occurred before or after random observation times. We propose a quantile regression model to analyze interval censored data because it relaxes the requirements on the error term and the coefficients are interpretable as direct regression effects on the failure time. Our model assumes that the conditional quantile of failure time is a linear function of covariates. We assume the conditional independence between the failure time and observation time. An M-estimator is developed for parameter estimation and the asymptotic distribution for the estimator is derived. The estimator is computed using the convex-concave procedure, and its confidence intervals are constructed using a subsampling method. The small sample performance of the proposed method is demonstrated via simulation studies. Finally, a real life data were analyzed to illustrate the proposed method.