Methodologic studies for improved results in epidemiologic research on the dementia of aging

Growing evidence suggests that the burden of Alzheimer's disease (AD) will worsen during this century as longevity continues to increase and may effect as many as 14 million individuals in the U.S. by the year 2050. This fifty-year forecast assumes an often-cited exponential increase in the incidence of AD through age 90 and beyond. However, due to a paucity of data, there has been limited consideration to the possibility that the incidence of AD might decline in extreme old age, for example after age 90. This decline may reflect results from a “mixed” population that includes some individuals who are relatively invulnerable to AD.; If this decline is real, it could have substantial implications for the projected burden of disease over time. Reliable methods are therefore needed for the estimation of population dementia rates, particularly after age 90.; This thesis examines several methodological and analytical aspects of population studies of dementia and AD. Chapter one is a review of basic concepts and methodology for the epidemiology of dementia and AD. Chapter two presents a receiver operating characteristic (ROC) analysis. This report describes the accuracy and economy of available screening methods to detect cases in the population of dementia in old age. Building upon these findings, chapter three presents the results from an autopsy analysis. Here, we demonstrate the validity of the resulting diagnoses and their ability to differentiate various dementing disorders. Finally, chapter four is a reanalysis of preliminary findings (from prevalence data) suggesting that these methods might reveal an inflection or decline in the incidence of AD in extreme old age.