A Class of Weighted Estimating Equations for Semiparametric Transformation Models with Missing Covariates

• Activity Number: 128
• Type: Contributed
• Date/Time: Monday, August 5, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Yang Ning*+ and Grace Y Yi and Nancy Reid
• Companies: University of Waterloo and University of Waterloo and University of Toronto
• Keywords: Augmented weighted estimating equations ; Missing covariates ; Transformation models ; Weighted estimating equations

Abstract

In survival analysis, covariate measurements often contain missing observations. Ignoring missingness in covariates can result in biased results. To conduct valid inferences, properly adjusting for missingness effects is usually necessary. We propose a weighted estimating equation approach to handle missing covariates under semiparametric transformation models for right censored data. The estimating equations have the same form as the weighted profile score functions. Besides the selection probability, we also allow additional time-specific and subject-specific weights in the estimating equations, which extends the inverse probability weighted approach. We consider both parametric and nonparametric models for the selection indicator. To further improve estimation efficiency, the weighted estimating equations are augmented by another sets of unbiased estimating equations. All our proposed estimators are shown to be consistent and asymptotically normal. We conduct simulation studies and apply our method to a real data set. The effect of model misspecification of the selection probability is considered.