Bayesian Models for Multiple Outcomes in Domains with Application to the Seychelles Child Development Study

• Activity Number: 20
• Type: Topic Contributed
• Date/Time: Sunday, July 29, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Luo Xiao and Sally Thurston*+ and David Ruppert and Tanzy Love and Philip W. Davidson
• Companies: Cornell University and University of Rochester and Cornell University and University of Rochester and University of Rochester
• Keywords: Latent variable model ; Markov chain Monte Carlo ; Methylmercury

Abstract

The Seychelles Child Development Study (SCDS) examines the effects of prenatal methylmercury exposure on multiple neuodevelopmental outcomes. Previously a model was fit to 20 SCDS outcomes in four outcome classes or "domains": cognition, memory, motor, and behavior. Each outcome was preassigned to exactly one domain, however some outcomes appeared to be related to multiple domains. Using a Bayesian MCMC approach we examine exposure and covariate effects in the SCDS when outcomes may belong to multiple domains and when the data help determine the assignment of outcomes to domains. Each domain is defined by a sentinel outcome; all other outcomes may belong to multiple domains. Our model allows exposure and covariate effects to differ across domains and across outcomes within domains, and includes random subject-specific effects. Results show effective determination of sparse domain assignment with increased power to detect exposure and covariate effects relative to separate models for each endpoint. The partial domain assignments in the SCDS were considered reasonable and checks of model misspecification were improved relative to a model with each outcome in a single domain.