# Explaining racial and ethnic Alzheimer’s disease disparities with joint estimation of life course social inequities and measurement bias in verbal memory assessment

> **NIH NIH F31** · UT SOUTHWESTERN MEDICAL CENTER · 2022 · $37,083

## Abstract

Project Summary/Abstract
 Marginalized populations have both greater rates of Alzheimer’s disease (AD) and greater risk of
misdiagnosis, which can delay access to treatment and caregiving. These differences reflect a life course
cascade of social inequity in the United States that disproportionately impacts historically marginalized racial
and ethnic groups and results in a host of physical and cognitive health disparities. Racial and ethnic
differences in neuropsychological test performance, which is commonly used to diagnose AD, are well-
documented and may contribute to racial/ethnic differences in the detection and diagnosis of AD.
 The effects of this structural racism begin in early life experiences and extend to educational and
occupational opportunities, social mobility, wealth accumulation, and stress. Differences in social classes of
origin (SCO) often persist across the lifespan and develop into differences in adult socioeconomic status
(SES). Associated with adult SES are additional contextual SES factors that can mediate health risks like food
scarcity and access to healthcare. Beyond SES and associated financial stressors, racial and ethnic minorities
also face greater social stress from experiences of discrimination, which have been linked to greater cognitive
risk. These differences in risk compound measurement biases in neuropsychological testing, making accurate
diagnosis of impairment more difficult and contributing to the observed racial/ethnic differences in scores.
 The proposed study will examine these complex life course factors to better understand why racial and
ethnic disparities arise in AD and how clinicians can reduce the risk of misdiagnosis. In order to study each of
these factors in the necessary complexity, the Health and Retirement Study and Harmonized Cognitive
Assessment Protocol datasets will be used to study the effects of SCO, adult SES, contextual SES factors,
stress, and test bias on verbal memory, which robustly declines early in the AD disease course. The first two
study aims utilize Bayesian explanatory item response theory to separate group-level racial/ethnic differences
attributable to measurement bias in the tests themselves from the effects of life course social inequities as a
cause of the group-level differences. The final aim will test the resulting models’ ability to improve diagnostic
accuracy for AD among minority populations. In line with the National Institute of Aging’s Health Disparities
Research Framework, these models leverage environmental, sociocultural, and behavioral factors to
understand the interactions and pathways that life course differences have in normal and abnormal aging,
dementia risk, and measurement bias. The fellowship will support training and mentoring in the complexities of
normal and abnormal cognitive aging, sophisticated statistical methods, and role of social equity in AD
assessment to support my growth as a researcher in AD disparities and the development of ne...

## Key facts

- **NIH application ID:** 10387926
- **Project number:** 1F31AG076284-01
- **Recipient organization:** UT SOUTHWESTERN MEDICAL CENTER
- **Principal Investigator:** William Goette
- **Activity code:** F31 (R01, R21, SBIR, etc.)
- **Funding institute:** NIH
- **Fiscal year:** 2022
- **Award amount:** $37,083
- **Award type:** 1
- **Project period:** 2022-09-01 → 2024-08-31

## Primary source

NIH RePORTER: https://reporter.nih.gov/project-details/10387926

## Citation

> US National Institutes of Health, RePORTER application 10387926, Explaining racial and ethnic Alzheimer’s disease disparities with joint estimation of life course social inequities and measurement bias in verbal memory assessment (1F31AG076284-01). Retrieved via AI Analytics 2026-06-02 from https://api.ai-analytics.org/grant/nih/10387926. Licensed CC0.

---

*[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*