Cognitive assessment in diverse populations: implications for Alzheimer’s disease clinical trials

Abstract

AbstractBackgroundBy 2060, racially/ethnically diverse individuals will constitute almost half of all older adults in the US and experience largest increases in dementia burden. Appropriate representation of diverse participants in clinical trials is critical to ensure generalizability of outcomes. Screening/inclusion criteria for trials often involve a brief cognitive assessment to indicate overall disease severity. We examined the Mini‐Mental State Examination (MMSE), the most widely used assessment, as a recruitment tool in diverse older adults.MethodData from 10,377 older adults (8,585 controls, 1,752 MCI; 232 Asian, 1,641 Black, 623 Hispanic, 7,841 White) were drawn from the National Alzheimer’s Coordinating Center (NACC) uniform data set. Only MCI participants for whom AD was presumed to be the primary etiology per NACC protocol were included. We performed logistic regressions to examine the likelihood of performing at/above a conventional MMSE cutoff of 24 in MCI participants. Multivariable regressions were performed to examine associations between 1) demographic variables and MMSE performance in controls, and 2) MMSE and CDR Sum of Boxes (CDR‐SB) in all participants.ResultCompared to White participants, Asian (OR = 0.46 [0.24‐0.99]), Black (OR = 0.53 [0.35‐0.80]), and Hispanic (OR = 0.38 [0.24‐0.61]) older adults with MCI were significantly less likely to perform at/above MMSE cutoff. Regressions in controls revealed a significant race/ethnicity by education interaction with greater magnitude of the association between education and MMSE in Black and Hispanic participants controlling for age, sex, and primary language. Associations between MMSE and CDR‐SB were significantly weaker in Black and Hispanic participants (Figure 1) and participants without high school education after controlling for age, sex, and primary language.ConclusionWe found that the MMSE has a greater education bias and weaker functional validity among Black and Hispanic older adults. Findings suggest that using an MMSE cutoff may be an inappropriate criterion for assessing disease stage among underrepresented participants. We discuss the implications of findings for the current state of diversity in clinical trials and for clinical diagnostic practice with diverse older adults. Focus points will include the need for culturally fair, sensitive, and psychometrically robust brief cognitive measures and the critical importance of incorporating social determinants of health into clinical and research practices.