Lifestyle activities and risk of incident dementia from linked insurance claims in the Baltimore Experience Corps Study.

OBJECTIVES: Lifestyle activity engagement in later life may protect against dementia, but studies have often been limited to short follow-up periods and predominantly White samples. We examined activity variety and frequency as predictors of incident dementia, leveraging 15 years of linked longitudinal insurance claims data in the Baltimore Experience Corps Study (BECS). METHODS: Participants were 482 individuals from BECS (2006-2013) with linked insurance claims (mean age=68.2, 84% women, 95% Black/African American). Participants self-reported baseline engagement in 28 activities. We calculated the number of unique activities (activity variety) and average frequency of engagement across reported activities (activity frequency). Incident dementia (2008-2022) was ascertained from CMS and private claims using the Bynum-EM algorithm. We used Cox proportional hazards models, adjusting for age and intervention group, and tested multiple follow-up periods (5-year, 15-year). RESULTS: The average baseline variety was 18.8 activities (SD = 3.3) and frequency was 13.1 days/month (SD = 3.2). For the 15-year follow-up only, higher activity variety was associated with lower risk of incident dementia (HR=.94, 95% CI: 0.89-0.99), equivalent to about a 6-month delay in diagnosis per additional activity. This association was attenuated after adjusting for additional demographics (e.g., sex, education), prevalent health conditions, and depression (HR=.96, 95% CI: 0.90-1.01). Activity frequency was not significant in any model. DISCUSSION: Higher activity variety, but not frequency, modestly predicted lower dementia risk over 15 years in a sample of primarily Black older adults. Findings suggest that engagement in a range of activities, rather than more frequent engagement in limited activities, may elicit greater cognitive benefits.