Selection Factors Influencing Participation in National Alzheimer’s Coordinating Center (NACC) Studies Compared to the Nationally Representative Health and Retirement Study (HRS)

Background
Racial/ethnic disparities in ADRD are well documented such that prevalence of ADRD and known risk factors (e.g., cardiovascular disease, educational attainment) disproportionately impact Black and Latinx populations compared to their White counterparts. To generalize findings on the mechanisms and disease course and mitigate racial/ethnic disparities in ADRD, it is critical that diverse populations are well represented in studies of ADRD. We examined selection factors across racial/ethnic groups included in National Alzheimer’s Coordinating Center data (NACC) compared to the general U.S. population using data from the Health and Retirement Study (HRS).

Method
Analyses included baseline NACC data from 33,998 participants (Mage = 74.1yrs±8.0; Medu = 15.5±6.9; 56.7% women) and the 2010 wave of HRS weighted up to the 2010 US population aged 60+ years (weighted n = 48,726,070; Mage = 71.3yrs; Medu = 12.9yrs; 54.3% women). We assessed covariate balance between NACC and the 2010 US population aged 60+ by calculating standardized mean differences across harmonized sociodemographic factors, self-reported health characteristics, depressive symptoms, and subjective memory concerns stratified by race/ethnicity. Standardized mean differences greater than +/- 0.25 were considered strong selection factors into NACC.

Result
Across all racial/ethnic groups, NACC participants were more likely to be older, have higher educational attainment, and report subjective memory concerns and hearing difficulties compared to the 2010 US population aged 60+ years (Figure). Standardized mean differences for age, educational attainment, and memory concerns were larger among non-Latinx Whites compared to differences among non-Latinx Black and Latinx participants. There was greater representation of non-Latinx Black women in NACC, and non-Latinx Black and Latinx NACC participants were less likely to endorse depressive symptoms. Additionally, results suggested non-Latinx White participants were less likely to report diabetes and hypertension compared to their 2010 aged 60+ US population counterparts.

Conclusion
Results suggest that NACC participants are not representative of the U.S. population in key demographic and health factors and self-reported memory concerns. These selection factors differed across racial/ethnic groups and are suggestive of selection bias. Future studies should leverage statistical tools that aimed at generalizing findings to general populations. Studies should also consider mitigating these sources of bias through inclusive recruitment efforts across Alzheimer’s Disease Centers to improve generalizability.