Abstract:OBJECTIVE: To investigate whether the effects of leisure activities on cognition differ among socioeconomic groups and the effects of changes in the frequency of habitual leisure activities on cognition.METHODS: We included 5869 older adults from the Chinese Longitudinal Healthy Longevity Survey (2008-2014). Five typical leisure activities were used to calculate the Leisure Activity Index (LAI). The Mini-Mental State Examination was used to assess cognition. Latent class analysis was used to construct the overall socioeconomic status (SES). Cox proportional hazard regression was applied to explore the associations and further stratified the analysis by SES.RESULTS: Participants in the highest quartile of LAI had a 50 % lower risk of developing cognitive impairment compared with the lowest quartile of LAI (HR: 0.50, 95 % CI: 0.40-0.62) (Ptrend < 0.01). Compared with participants who participated in leisure activities with increased frequency during the follow-up period, the HR (95 % CIs) of participants with no change in frequency was 3.10 (2.39-4.01), and that of participants with decreased frequency was 2.34 (1.81-3.04). A significant interaction between LAI and SES were found (Pinteraction = 0.02). The association between LAI and cognitive function was more pronounced in participants with high SES (HRQ4 vs Q1: 0.31, 95 % CI: 0.16-0.59) than in participants with low SES (HRQ4 vs Q1: 0.61, 95 % CI: 0.45-0.83).CONCLUSIONS: Leisure activities may have significant benefits in preventing cognitive impairment. However, promoting leisure activities alone may not substantially reduce socioeconomic inequalities in cognitive health. Measures addressing the social determinants of cognitive health still need to be further explored.

Engagement in leisure activities and cognitive function by socioeconomic groups in China: A prospective cohort study