A Life Course Approach to Understanding Racial and Ethnic Disparities in Alzheimer's Disease and Related Dementias and Health Care

As the share of U.S. older population and number of people living with Alzheimer's Disease and Related Dementias (ADRD) continue to grow rapidly, marked gaps in prevalence and incidence of ADRD and ADRD-attributable health care persist across demographic groups. This study aims to deepen our understanding of differences in the risk of ADRD and related health care utilization among U.S. older adults, using a life course approach. We will utilize appropriate machine learning (ML) approaches to examine how life course factors, especially early-life circumstances, may accumulate over the life course in ways that differ across demographic groups to shape differential ADRD risks; how risk factors in midlife and later life may explain gaps in ADRD-attributable health care use and outcomes for persons with ADRD. Identifying ADRD risk in the preclinical stage is crucial, our holistic life course approach holds promise in enhancing prevention at the population level and addressing gaps across demographic groups.
Our overarching goal is to address ADRD-related health and health care gaps, guided by novel evidence starting from early stages of life, and ideally delay the onset or slow the progression of ADRD. To achieve our overall goal, we will adapt ML to multiple rich data sources linking longitudinal survey, national neighborhood data, medical claims, and life history in 1995-2020 Health and Retirement Study (HRS). ML has demonstrated large potential for early disease detection and cost containment, and may circumvent key statistical challenges.
We will pursue four specific aims: 1) develop and validate ML and other models for ADRD prediction, examining multifactorial influences of life course factors; 2) understand individual and collective contributions of early-life circumstances to ADRD and its gaps across demographic groups; 3) examine the effect of incident ADRD on health care use and its dynamics pre- and post- ADRD diagnosis, and gaps across demographic groups; 4) investigate the extent to which midlife and later-life factors may mediate the effects of ADRD on health care use and its gaps across demographic groups.
This study will add significant value to narrowing gaps in ADRD and its health care, by using ML algorithms to explore the role of a uniquely rich set of life course factors on gaps in ADRD across demographic groups; by augmenting a nationally representative longitudinal survey with administrative data to systematically examine ADRD and gaps in health care. Taken together, these findings will inform 1) development of risk prediction models for ADRD to offer a cost-effective approach for population-level screening in the preclinical stage, identification of risk factors and groups at elevated risk of ADRD for targeted preventive interventions; 2) products that can aid individuals and clinicians in making informative assessments; and 3) policies addressing ADRD-attributable health and health care gaps starting from early stages of life, leveraging midlife and later-life mediators, and ideally delaying the onset or progression of ADRD.

Grant Number: 5R01AG077529-04
NIH Institute/Center: NIH
Principal Investigator: Xi Chen