Aging
Title: Association of perceived neighborhood social characteristics with cognitive function and decline in the Health and Retirement Study Xu Zhang* Kendra Sims Xu Zhang M. Maria Glymour
Background: Perceiving lack of neighborhood cohesion or limited social ties may reduce later-life cognitive function and accelerate decline. We hypothesized that adverse neighborhood context harms cognitive health for men more than women. Methods: We analyzed 14119 Health and Retirement Study participants (aged 51-101, 74% Non-Hispanic White, 58% women) who responded to scales on neighborhood social cohesion and social ties in 2006 or 2008. Data on our outcomes of interest, biennial scores on immediate and delayed word recall batteries, were collected from 2006 through 2018. With gender-stratified linear regressions and linear mixed-effect models adjusted for social determinants of health, we evaluated the association of each neighborhood characteristic with standardized recall scores at baseline and 12 years. Results: The baseline mean neighborhood social cohesion score was 0.59 (SD=0.23); and mean social ties score was 0.53 (SD=0.32). Neighborhood social cohesion was positively associated with baseline memory performance for both genders: e.g., scoring higher on social cohesion was associated with a 0.13-SD (95% CI: 0.10; 0.16) increase in immediate recall. Women reporting more social ties had higher baseline immediate recall scores (β: 0.04; 95% CI: 0.01, 0.06), but this association was reversed among men (β: -0.03; 95% CI: -0.05, 0.00). Over 12 years, neighborhood characteristics were not associated with cognitive trajectories: e.g., the per-visit change in delayed recall for those reporting more social ties was 0.01 (95% CI: -0.03; 0.04) for women and -0.02 (95% CI: -0.07; 0.02) for men. Conclusions: Unexpected associations between perceived social ties and cognitive function indicate future need to evaluate differential reporting on psychosocial scales, selection bias, and gender-specific strategies to retain older adults when evaluating neighborhood determinants of cognitive health.