O-060: Social factors predict cognitive outcomes cross-sectionally, but not longitudinally, among older Irish adults

Abstract

Objectives: Evidence exists of an association between social factors and cognitive functioning in later life. We wanted to investigate whether (a) this association holds when health behaviours and mental health status are controlled for, and (b) whether the association is present in cross-sectional and/or longitudinal data. Methods: Data from 8504 older (aged 50+) participants in waves 1 and 2 of the TILDA (Irish Longitudinal Study on Ageing) dataset were analysed using hierarchical linear regressions. Outcomes were immediate and delayed word recall, and scores on the Mini Mental State Examination (MMSE), Age, gender, education level, health factors (blood pressure, waist-hip ratio, smoking and alcohol intake, exercise, triglycerides), and mental health (depressive and anxiety symptomatology) factors were included as covariates. Predictors of interest were social participation, social connectedness, and loneliness. Results: Controlling for covariates, social factors predicted immediate recall (Loneliness β = -0.064, p<0.01; Social participation β = 0.045, p<0.01), delayed recall (Loneliness β = -0.08, p<0.01; Social Participation β = 0.034, p<0.05) and MMSE outcomes (Loneliness β = -0.064, p<0.01, Social Participation β 0.056, p<0.01), cross-sectionally, but not longitudinally. We investigated potential reverse causality (cognition predicting change in social factors over time) but this possibility was ruled out. Conclusions: There is a cross-sectional but no longitudinal association between social participation, loneliness and cognitive outcomes among older Irish adult. Social factors may have a short-term impact on cognition only, or potentially a key confound was not considered in this analysis. Further research is warranted to investigate other potential confounds of the association.