Premium
Examining risk of dementia in lifestyle engagement subgroups of community‐dwelling older adults: A latent class approach
Author(s) -
Moored Kyle D,
Parisi Jeanine M,
Snitz Beth E,
DeKosky Steven T,
Williamson Jeff D,
Fitzpatrick Annette L,
Carlson Michelle C
Publication year - 2020
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1002/alz.043167
Subject(s) - dementia , latent class model , hazard ratio , psychology , confounding , proportional hazards model , social class , gerontology , cognition , clinical psychology , demography , medicine , developmental psychology , disease , psychiatry , confidence interval , statistics , mathematics , sociology , political science , law
Background Engagement in lifestyle activities may be neuroprotective, but it remains unclear what aspects of engagement are most beneficial. For self‐reported activity inventories, examining response patterns may better characterize both quantitative (e.g., number) and qualitative (e.g., characteristic/motivational) differences in engagement. Method Participants were 3069 individuals from the Gingko Evaluation of Memory (GEM) Study who were evaluated for dementia and completed neuropsychiatric testing over 8 years. We used a subset of 18 physical, intellectual, or social activities from the Lifestyle Activity Questionnaire. We performed latent class analysis to characterize subgroups with distinct activity response patterns and examined whether they have differential risk of incident dementia and domain‐specific cognitive impairments over time. Cox proportional hazards models were used to examine subgroup differences in risk of cognitive impairments. Result A three‐class model was chosen based on fit statistics and interpretability. Class 1 (Highly intellectually/socially active) participated in more social leisure activities (e.g., movies) than Classes 2 (Socially/less intellectually active) and 3 (Less intellectually/socially active). Class 2 participated in more institutional social activities (e.g., church) than Class 3. In an unadjusted Cox proportional hazard model, Class 1 had a reduced risk of dementia compared to Classes 2 ( HR =1.32, 95% CI : [1.1,1.6], p =.013) and 3 ( HR =1.53, 95% CI : [1.2,1.9], p <.001). After adjusting for potential demographic and health confounders, Class 1 had a statistically significant reduced risk of dementia compared to Class 3 ( HR =1.40, 95% CI : [1.1,1.8], p =.003), but the reduction in risk compared to Class 2 was only trending ( HR =1.26, 95% CI : [1.0,1.6], p =.051). Conclusion These results suggest that subgroups characterized by less engagement in leisure social and intellectual activities have increased risk for dementia. Our findings also highlight subgroup differences in activity types (e.g., participation in social institutions in Class 2) that may be leveraged by further behavioral interventions to reduce risk of dementia with age.