A Bayesian Hierarchical Model for Combining Multiple Longitudinal Aging Studies

• Activity Number: 139
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Nuoo-Ting Molitor*+ and Sylvia Richardson and John Molitor and Nicky Best
• Companies: Imperial College, London and Imperial College, London and Imperial College, London and Imperial College, London
• Keywords: ageing ; Bayesian analysis ; Hierarchical modelling ; grown-curve modelling ; cognitive reserve

Abstract

In studies of cognitive decline, one is often faced with the issue of cognitive reserve, a term used to describe an individual's apparent resilience to damage of the brain. Often an individual's level of education mask or delay the start of measured decay of brain function. We present a Bayesian model that combines data from several studies in a manner that allows for study and global-level tests for the impact of education level on results from cognitive abilities tests, especially the Mini Mental State Examination (MMSE). Further, our unified Bayesian model also allows us to deal with issues related to ceiling and retest effects often associated with tests of cognitive abilities. We illustrated our model by using data obtained from Integrative Analysis of Longitudinal Studies on Aging (IALSA).