Background There is a growing literature drawing on theories of cognitive reserve suggesting that factors relating to social networks, activity and support can predict dementia in older age. Much of this work has focused on social deficits (such as loneliness and isolation), but there is increasing evidence that engaging in social activities can be protective. The aim of this study was to compare the potential protective associations of between three different types of social activity (socialising, community group activities and cultural engagement) and dementia incidence.

Methods We used nationally-representative data from 9,550 adults aged 50+ from the English Longitudinal Study of Ageing followed-up over a 12 year period. We determined dementia occurrence using an algorithm that combined self- or informant-reported physician diagnosis of dementia with informant-reported score on the Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE). We ran three types of time-to-event analyses: Cox proportional hazards models (modelling time to dementia), competing risk regressions models (modelling time to dementia vs the competing risk of death), and modified Fine and Gray Subdistribution hazards models (modelling time to dementia or death with a high probability of dementia vs the competing risk of death with a low probability of dementia). Analyses controlled for all identified demographic, health-related, and social covariates.

Results Social and cultural engagement were both independently associated with a lower risk of developing dementia in older age in fully-adjusted models. Using time-to-event analyses, socialising once a week was associated with a 29% lower risk OR=0.71 95%CI 0.52–0.97 and cultural engagement every few months or more was associated with a 42% lower risk OR=0.58 95%CI 0.41–0.80. Using competing risk models, socialising OR=0.80 95%CI 0.59–1.08 and cultural engagement OR=0.65 95%CI 0.47–0.90. Using modified Fine and Gray Subdistribution hazards models, socialising OR=0.66 95%CI 0.53–0.82 and cultural engagement OR=0.42 95%CI 0.32–0.56. Community group activities were only associated with dementia in minimally-adjusted models. Results were robust to sensitivity analyses considering moderators, reverse causality, over-adjustment, and baseline cognitive function.

Conclusion The results presented here suggest that social and cultural engagement are independent risk-reducing factor for the development of dementia in older age. Even for those who lack contact with friends and family or who socialise infrequently, engagement with cultural venues, even on a less frequent basis, could be protective against the incidence of dementia. These findings align with broader findings relating to cognitive reserve and support the development of multimodal community-based interventions to promote healthy cognitive ageing amongst older adults.