Calibration Problems with Panel Data Applied to Alzheimer's Disease Biomarkers

• Activity Number: 547
• Type: Contributed
• Date/Time: Wednesday, August 7, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Huanli Wang*+ and Danielle Harvey and Laurel Beckett
• Companies: University of California, Davis and UC Davis and UC Davis
• Keywords: Alzheimer's Disease ; Calibration ; Panel data ; Longitudinal Analysis

Abstract

Alzheimer's disease (AD) develops and progresses over many years, from earliest undetected brain damage to dementia and death. Clifford Jack (2010) has proposed a conceptual model for the onset sequence of pathological and clinical features of AD: abnormal processing of beta-amyloid (A-beta) peptide; formation of A-beta plaques in the brain; and after a lag period, neuronal dysfunction and neurodegeneration become the dominant pathological processes. However, data are not yet available measuring all biomarkers of interest on each individual through the entire disease process, which means only multivariate panel data are provided. Panel data covering shorter time windows on individuals can capture much of the disease trajectory. Our study goal is to estimate the disease trajectory using panel data and then use a calibration approach to estimate the onset timing of disease progression and the between person variation of stages of disease progression. We will also show the asymptotic theoretical properties, and simulations for realistic sample sizes. Finally, we will apply this approach in Alzheimer's Disease Neuroimaging Initiative (ADNI) data