No abstract

Background Religious participation has been shown to be associated with a number of health outcomes in later life. Research into religion and cognitive decline has been inconclusive, although there is some evidence for a protective effect. There is a lack of evidence on mechanisms around this relationship, and what the implications are for those who are not religious. We aimed to assess whether religious affiliation or religious practice was associated with cognitive trajectories, and to test possible mechanisms for an association. Methods Data came from the Irish Longitudinal Study on Ageing (TILDA), a nationally representative study of the over 50s population in Ireland. A total of 7,331 had available data on all measures of interest. We used Latent Growth Class Analysis (LGCA) to identify different latent trajectory classes for cognition using the Mini Mental State Examination (MMSE) measure. Five waves of data collection were used to identify latent class trajectories. We then used multinomial logistic regression to assess the likelihood of membership to each trajectory class by religious affiliation or non-affiliation, and by level of religious attendance. We tested three possible mediation pathways to explain observed relationships; depressive symptoms, social network and smoking. Results Three MMSE trajectory classes were identified using LGCA. These included a 'high steady' class, a 'medium declining' class and a 'low declining' class. There were no differences in class membership by religious affiliation or nonaffiliation. Women who attended religious services were less likely to be in the low declining MMSE class (relative risk ratio=0.73, 95% confidence interval=0.55; 0.96). This effect was fully mediated by depressive symptoms, social network and smoking. No effects were found for men. Discussion Cognitive trajectories after age 50 are not uniform. Being religious or nonreligious in the over 50s in Ireland is not associated with the type of cognitive trajectory experienced. However, frequent religious attendance in women who are religious appears to have a small protective effect. This effect appears to be driven through better mental health, more social participation and lower rates of smoking.

negative (stigma, work load, negative impact on reputation) and the positive impact (detailed review of procedures, implementation of targeted approaches) of the outlier process. Participants felt that sharing experiences of outlying hospitals helps others to improve. They also suggested a 'buddy system' between better and worse performing hospitals. Many highlighted the importance of 'networks' to share experiences, either good or bad, as a vehicle for improving practice. Discussion The outlier process was generally accepted as a possible mechanism to improve practice. However, participants indicated that effective dissemination is key to ensuring that identifying poor outcomes in some hospitals (e.g. high-risk approach) can stimulate country-wide quality improvement (population approach).

No abstract