Bayesian Inference of Incomplete Longitudinal Data: A Simple Method to Assess Sensitivity to Nonignorable Dropout

• Activity Number: 24
• Type: Contributed
• Date/Time: Sunday, August 2, 2009 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Hui Xie*+
• Companies: University of Illinois
• Keywords: Bayesian Statistics ; Hierachical Model ; Linear Mixed Model ; Nonignorable ; Sensitivity Analysis ; MAR

Abstract

Bayesian approach has been increasingly used for analyzing longitudinal data. When dropout occurs in the study, analysis often relies on the assumption of ignorable dropout. It is important to assess the impact of departure from the ignorability assumption on the key Bayesian inferences. In this paper, we extend a Bayesian index of local sensitivity to non-ignorability method proposed by Zhang and Heitjan to longitudinal data with dropouts. We derive formulas for Bayesian ISNI when the complete longitudinal data follow a linear mixed-effect model. The calculation of index only requires the posterior draws from the ignorable model or summary statistics of these draws that is available from standard analysis. One can use the method to evaluate which Bayesian parameter estimates or functions of these estimates in a linear mixed-effect model are susceptible to nonignorable dropout.