Modeling Arbitrarily Interval-Censored Data with Time-Dependent Covariates

• Activity Number: 419
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Wei Fang*+
• Keywords: Survival Analysis ; Arbitrarily Interval-censored Data ; Time-dependent Covariates ; Generalized Estimating Equations

Abstract

In regression modeling of survival data, arbitrarily interval-censored data provide more accurate information regarding the time when the event of interest has occurred than right-censored data. Time-dependent covariates provide dynamic information for the relationship between the event of interest and covariates, while time-independent covariates do not. However, the Cox proportional hazards (PH) model, which usually analyzes right-censored data, cannot directly account for arbitrarily interval-censored data, and although time-dependent covariates often arise in practice, most of the inference approaches developed for arbitrarily interval-censored data only apply to time-independent covariates. This paper presents a new framework for modeling arbitrarily interval-censored data with time-dependent covariates. Parameter estimates are obtained via Generalized Estimating Equations (GEE) with the independent working correlation. This paper also presents comparison analyses of an example data set and programs.