An Estimating Equation for Semiparametric Frailty Models with Error-Prone Covariates

• Activity Number: 524
• Type: Contributed
• Date/Time: Wednesday, August 3, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Se Hee Kim*+ and Yi Li and Donna Spiegelman
• Companies: Harvard School of Public Health and Dana-Farber Cancer Institute/Harvard School of Public Health and Harvard School of Public Health
• Keywords: copula ; estimating equation ; measurement error ; proportional hazards model

Abstract

We consider semiparametric frailty models for the correction of measurement errors in survival data analysis. An estimating equation method based on the conditional expectation of the unobservable frailty is proposed for regression coefficients in the proportional hazard model with error-prone covariates. We allow general structures for the error-prone covariates, not restricted to a linear additive measurement error model or a Gaussian measurement error. The conditional distribution of the frailty given observed covariates is estimated nonparametrically through a copula as a function of the marginal distributions of the true exposure and the surrogate. The proposed approach via the copula can enhance the precision of estimation by utilizing all the available information on the surrogate from both main and validation studies. Asymptotic properties are established well, and finite sample properties are evaluated through extensive simulation studies. The proposed method is applied to data from the Nurses' Health Study and Health Professionals Follow-up Study.