Predictors of Dementia and Memory Decline: Lifecourse Socioeconomic Status and Diabetes

Abstract

This dissertation examines the relationship between both lifecourse socioeconomic status and diabetes and memory decline and dementia using data from the Health and Retirement Study. Chapter 1 examines associations between lifecourse measures of socioeconomic status (SES) and late life memory function and decline. Memory function was modeled longitudinally from 2002-2012 and decline was modeled for the same years via growth curves. SES measures included childhood SES, high school and college education, as well as late life household income. Of our SES measures, education was the strongest predictor of level of memory function, whereas rate of decline was driven by late life income, suggesting that late life interventions focused on SES or its downstream consequences – like health behaviors – may have the potential to slow cognitive decline. Chapter 2 assesses socioeconomic status (SES) at three points in the lifecourse in order to evaluate alternative lifecourse models with respect to the association between SES and risk of dementia. We used generalized estimating equations with a logit link to model dementia probability from 2002-2012. SES measures included childhood SES, high school education, and late life household income. We found that high school completion was the strongest predictor of decreased dementia risk, consistent with either of two lifecourse models: a sensitive period in early adulthood and/or cumulative risk. Chapter 3 explores how type 2 diabetes (T2D) and blood glucose levels relate to cognitive decline, which is important for understanding the mechanisms driving higher dementia rates among individuals with T2D. Memory decline was modeled via generalized estimating equations using linear growth curves with diabetes, HbA1c, and covariates as predictors. Diabetes and higher levels of HbA1c both predicted faster memory decline after controlling for an extensive battery of potential social confounders in a large, U.S. dataset. Our results add support for a causal pathway from diabetes to risk of dementia. Blood glucose concentration may partially explain the association between diabetes and both dementia and cognitive decline, although precise biological mechanisms are still unknown.