"Statistical Methods for Recurrent Events Data in the Presence of a Terminal Event and Missing Covariate Information"
Shankar Viswanathan, DrPH, University of North Carolina at Chapel Hill
February 8, 2011 @ 3:30pm - 4:30pm
Location: 701 Blockley Hall
Biostatistics
"Statistical Methods for Recurrent Events Data in the Presence of a Terminal Event and Missing Covariate Information"
In many clinical and epidemiological studies, recurrent events of interest such as infections in immunocompromised patients or injuries in athletes occur. Often the interest is to examine the relationship between covariates and recurrent events. However, in many studies, some of the covariates collected involve missing information due to various reasons. Under such missingness, commonly practiced method is to analyze complete cases only in which the estimated parameters may be biased or inefficient. We present a method for estimating the parameters in the marginal rate model for analyzing recurrent event data in the presence of a terminal event. We adopt a weighted estimating equation approach with missing data assumed to be missing at random (MAR) for estimating the parameters. The parameters are estimated via weighted expectation-maximization (EM) algorithm. Simulation studies showed that our proposed estimators for the regression parameters are in general approximately unbiased and the variance estimates perform well. Our method is effective in reducing the bias and improving the efficiency of parameter estimates. We applied the proposed method to the India renal transplant cohort data for illustration.
