Implications of Response Patterns in Questions of Early Life Adverse Events as a Risk Factor for Lower Cognitive Function in the Multi‐Ethnic Study of Atherosclerosis

Abstract

AbstractBackgroundEarly‐life disadvantage may increase the risk for later cognitive decline. Recently, other aspects of early‐life adversity (ELA) have received attention but their associations with cognitive decline in later life has not been widely studied. Due to the sensitivity of such questions and the potential for systematic bias, we examined (1) predictors of response patterns to ELA questions, and (2) among those who completed ELA questions, we examined whether responses were associated with cognitive decline in the Multi‐Ethnic Study of Atherosclerosis (MESA).MethodELA was self‐reported via a 7‐item telephone assessment of socioemotional and physical threat or deprivation at MESA annual follow‐up call in 2019. Descriptive statistics were grouped by completion status: not completed, partially completed, and fully completed. Continuous variables were compared across groups with t‐tests; categorical variables were compared with chi‐square tests. Cognitive function was measured using the Cognitive Abilities Screening Instrument (CASI) at MESA Exams 5 (2010‐2012) and 6 (2016‐2018). Multinomial logistic regression was used to determine predictors of completion group membership. Cognitive decline from Exam 5 to 6 by completion group was assessed using multivariable linear regression.ResultOf the 3,990 participants interviewed, the mean age was 59.2y (SD = 8.8), 55% were female, and 40% were White (n = 1,604), 13% Chinese (n = 508), 26% Black (n = 1,036), and 21% Hispanic/Latino (n = 842). The odds of partial response was higher for those born outside of the U.S. (2.17, CI 95%: 1.53‐3.09), and among participants who spoke Spanish (2.82, CI 95%: 1.74‐4.58) or Chinese (2.04, CI 95%: 1.08‐3.88) at baseline. Among participants with completed ELA (n = 3,173), one component of childhood adversity: how often was your household well‐organized/managed, was associated with better Exam 6 CASI score (β = 0.80, SE = 0.33; p = 0.015) adjusting for age, race, education, gender, language, years in the U.S., socio‐economic status (SES), and parental birthplace. There was no significant association between cognitive change from Exam 5 to Exam 6 and ELA completion.ConclusionFactors associated with acculturation including country of participant’s birth and language preference predicted completion status on questions of ELA. Future studies should be mindful of the potentially differential response patterns and the implications for interpretation when analyzing ELA data.