Social isolation and psychosis-like experiences: a UK general population analysis

Background: Social isolation is a complex construct characterised by both objective and perceived components and has been commonly identified as a risk factor for psychosis-like experiences (PLEs). Few studies, however, have modelled the association between social isolation and PLEs in the general population. Method: Data from a UK general population survey (N = 7403) were analysed using latent class analysis to identify distinct groups of individuals characterised by the same profile of social isolation. Six objective and perceived indicators of social isolation (e.g. living alone; feeling socially isolated and lonely) were modelled. Associations between classes and PLEs were analysed using multivariate binary logistic regression analysis. Results: Three classes were identified: an isolated and lonely class (Class 1), a no communication or close relationships class (Class 2) and a baseline class (Class 3). Compared to the baseline class, Class 1 was significantly more likely to endorse thought interference (OR = 2.0) and paranoia (OR = 3.3), while Class 2 was significantly more likely to endorse paranoia (OR = 8.6) and hallucinations (OR = 1.9). Conclusions: Social isolation in the general population seems to vary between two distinct groups. PLEs, in turn, seem to vary depending on the nature of this isolation.