Quantification of race/ethnicity representation in Alzheimer's disease neuroimaging research in the USA: a systematic review

Abstract

Background: Racial and ethnic minoritized groups are disproportionately at risk for Alzheimer's Disease (AD), but are not sufficiently recruited in AD neuroimaging research in the United States. This is important as sample composition impacts generalizability of findings, biomarker cutoffs, and treatment effects. No studies have quantified the breadth of race/ethnicity representation in the AD literature.

Methods: This review identified median race/ethnicity composition of AD neuroimaging US-based research samples available as free full-text articles on PubMed. Two types of published studies were analyzed: studies that directly report race/ethnicity data (i.e., direct studies), and studies that do not report race/ethnicity but used data from a cohort study/database that does report this information (i.e., indirect studies).

Results: Direct studies (n = 719) have median representation of 88.9% white or 87.4% Non-Hispanic white, 7.3% Black/African American, and 3.4% Hispanic/Latino ethnicity, with 0% Asian American, Native Hawaiian/Pacific Islander, and American Indian/Alaska Native, Multiracial, and Other Race participants. Cohort studies/databases (n = 44) from which indirect studies (n = 1745) derived are more diverse, with median representation of 84.2% white, 83.7% Non-Hispanic white, 11.6% Black/African American, 4.7% Hispanic/Latino, and 1.75% Asian American participants. Notably, 94% of indirect studies derive from just 10 cohort studies/databases. Comparisons of two time periods using a median split for publication year, 1994-2017 and 2018-2022, indicate that sample diversity has improved recently, particularly for Black/African American participants (3.39% from 1994-2017 and 8.29% from 2018-2022).

Conclusions: There is still underrepresentation of all minoritized groups relative to Census data, especially for Hispanic/Latino and Asian American individuals. The AD neuroimaging literature will benefit from increased representative recruitment of ethnic/racial minorities. More transparent reporting of race/ethnicity data is needed.

Fig. 1. Article selection flowchart.

Of 34,336 articles retrieved, 2464 articles were included in this review, divided into direct and indirect studies.

Fig. 2. Histogram of the racial/ethnic composition of Alzheimer’s Disease neuroimaging studies directly reporting race/ethnicity data.

Mean and median % representation of (a) white; (b) Non-Hispanic white; (c) Black/African American; and (d) Hispanic/Latino participants for studies directly reporting race/ethnicity data (n = 719). Count represents the number of studies reporting a given %.

Fig. 3. Study race/ethnicity composition by publication year for Alzheimer’s disease neuroimaging studies directly reporting race/ethnicity data.

Time by % representation of (a) white; (b) Non-Hispanic white; (c) Black/African American; and (d) Hispanic/Latino participants for studies directly reporting race/ethnicity data (n = 719). Each study represents a dot in this scatterplot. % Race refers to % of participants in each study. Total N (dot size) refers to the sample size of each study.

Fig. 4. Histogram of the racial/ethnic composition of Alzheimer’s Disease neuroimaging studies indirectly reporting race/ethnicity data.

Mean and median % representation of (a) white; (b) Non-Hispanic white; (c) Black/African American; and (d) Hispanic/Latino participants for cohort studies/databases (N = 44) from which studies indirectly report race/ethnicity data (n = 1745). Count represents the number of cohort studies reporting a given %.