Date Available

6-28-2013

Year of Publication

2013

Document Type

Doctoral Dissertation

Degree Name

Doctor of Philosophy (PhD)

College

Public Health

Department/School/Program

Epidemiology and Biostatistics

Advisor

Dr. Richard J. Kryscio

Abstract

Dementia is increasingly recognized as a major and growing threat to public health worldwide, and there is a critical need for prevention and treatment strategies. However, it is necessary that appropriate methodologies are used in the identification of risk factors. The purpose of this dissertation research was to develop further the body of literature featuring Markov chains as an analytic tool for data derived from longitudinal studies of aging and dementia.

Data drawn from 649 participants in the University of Kentucky’s Alzheimer’s Disease Center’s (UK ADC) Biologically Resilient Adults in Neurological Studies (BRAiNS) cohort, which was established in 1989 and follows adults age 60 years and older who are cognitively normal at baseline to death, were used to conduct three studies. The first study, “Mild cognitive impairment: Statistical models of transition using longitudinal clinical data,” shows that mild cognitive impairment is a stable clinical entity when a rigorous definition is applied. The second study, “Self-reported head injury and risk of cognitive impairment and Alzheimer’s-type pathology in a longitudinal study of aging and dementia,” shows that when the competing risk of death is properly accounted for, self-reported head injury is a clear risk factor for late-life dementia and is associated with increased beta-amyloid deposition in the brain. The third study, “Incorporating prior-state dependence among random effects and beta coefficients improves multistate Markov chain model fit,” shows that the effect of risk factors, like age, may not be constant over time and may be altered based on the subject’s cognitive state and that model fit is significantly improved when this is taken into account.

Download

Share

COinS