Mini-mental state examination trajectories after age 50 by religious affiliation and practice in Ireland.

Religious attendance is sometimes associated with better health outcomes, although the link between religion and cognitive ageing is inconclusive. We aimed to assess differences in cognitive performance trajectories by religious affiliation and religious attendance. We further sought to test possible mechanisms for an association.Data from the Irish Longitudinal Study on Ageing (TILDA), a nationally representative study of the over 50 s population in Ireland, was used. We identified latent class trajectories of Mini Mental State Examination (MMSE) performance over five waves using Latent Growth Class Analysis (LGCA) on data from 7325 individuals. Multinomial logistic regression was used to estimate the likelihood of membership to each trajectory class by religious affiliation or non-affiliation, and by religious attendance and importance. Finally, we tested possible behavioural, psychological and social mechanisms. LGCA identified three trajectory classes, a 'high start' class, a 'medium start' class and a 'low start' class. There were no differences in class membership by religious affiliation or non-affiliation. Women who attended religious services were less likely to be in the low declining MMSE class. This effect was mediated by depressive symptoms, social network and smoking. Women who said religion was very important were more likely to be in the medium performing class, and this was not mediated. The cognitive trajectories of the over 50 s in Ireland vary. Variation was not influenced by religious affiliation. Religious attendance and importance had mixed effects on women's cognition trajectories. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s10433-020-00597-0.