At the national level, there has been significant interest in addressing U.S. health disparities in Alzheimer's disease and related dementia. Although there is growing evidence to suggest race/ethnic disparities in dementia exists in the U.S., much of the evidence of disparities is indirect or comes from geographically- localized studies that may not be generalizable to the experience of Americans as a whole. Furthermore, there is little evidence to indicate whether disparities are widening or shrinking over time. Thus, the aims of this proposal are to (1) describe race/ethnic disparities in the incidence of experiencing Alzheimer's disease and related dementia at the national level in the U.S. and examine the extent to which race/ethnic disparities in incidence have changed over time, and (2) describe race/ethnic disparities in over- or under-diagnosis of prevalent Alzheimer's disease and related dementia in the medical setting at the national level in the U.S. and examine the extent to which race/ethnic disparities in under- or over-diagnosis have changed over time. To accomplish these aims, we will use data from the nationally representative U.S. Health and Retirement Study (HRS), the Aging, Demographics, and Memory Study (ADAMS), and linked Medicare data.
For Aim 1, we will build on previous work creating algorithms to predict dementia status from ADAMS using only HRS data (ADAMS participants are sampled from HRS) in order to create an improved algorithm with optimal validity across and within race/ethnic groups. We will then apply this algorithm to assign dementia status to each eligible HRS participant at each HRS interview. From this we will derive age-standardized estimates of dementia incidence in HRS by race/ethnic group (self-reported non-Hispanic white, non-Hispanic black, Hispanic) over two time periods, 1998-2004 and 2006-2012.
For Aim 2, we will use Medicare claims data to determine whether each eligible HRS participant had a diagnosis of dementia prior to or within 1 year after each HRS interview. We will then estimate agreement between the Medicare-based and algorithmic dementia diagnosis assigned at each HRS interview from 1998 to 2012 within race/ethnic groups to quantify disparities and temporal trends in over- or under-diagnosis of Alzheimer's disease and related dementia. This contribution is expected to be significant as it will provide (i) national and historical benchmarks against which to compare current and future data on local race/ethnic disparities in both the incidence and over- or under- diagnosis of Alzheimer's disease and related dementia; (ii) evidence summarizing temporal changes in race/ethnic disparities in Alzheimer's disease and related dementia; (iii) improved algorithms for diagnosis of dementia in HRS that maximize validity in minority populations; and (iv) evidence for the feasibility of linking of multiple data sources to provide valid estimates of race/ethnic disparities in Alzheimer's disease and related dementia in the U.S.
This study represents a critical first step towards addressing race/ethnic health disparities in Alzheimer's disease and related dementia in the U.S, as until we can quantify the extent of the problem and understand how it has changed with time, we cannot hope to adequately address it. Ultimately, the proposed work will inform future efforts to develop a national monitoring system for race/ethnic health disparities in Alzheimer's disease and related dementias.
