A Bayesian Shrinkage Model for Incomplete Longitudinal Binary Data with Application to the Breast Cancer Prevention Trial

• Activity Number: 163
• Type: Topic Contributed
• Date/Time: Monday, August 2, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Chenguang Wang*+ and Michael J. Daniels and Daniel O. Scharfstein and Stephanie R. Land
• Companies: FDA and University of Florida and Johns Hopkins Bloomberg School of Public Health and University of Pittsburgh
• Keywords: Informative drop-out ; Intermittent missingness ; Prior elicitation

Abstract

A Bayesian Shrinkage Model for Incomplete Longitudinal Binary Data with Application to the Breast Cancer Prevention Trial

We consider inference in randomized longitudinal studies with missing data that is generated by skipped clinic visits and loss to follow-up. We assume a non-future dependence model for the drop-out mechanism and partial ignorability for the intermittent missingness. We posit an exponential tilt model that links non-identifiable distributions and distributions identified under partial ignorability. This model is indexed by non-identified parameters, which are assumed to have an informative prior distribution elicited from subject-matter experts. To avoid the curse of dimensionality, we model the distribution of the observed data using a Bayesian shrinkage model. Our methodology is motivated by and applied to data from the Breast Cancer Prevention Trial.