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 and psychosis-like experiences (PLEs). Few studies to... [ view full abstract ]

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 and psychosis-like experiences (PLEs). Few studies to date however have modelled the complexity of social isolation and its association with discrete PLEs in the general population.

Method & Results: Using data from the Adult Psychiatric Morbidity Survey 2007 (N = 7,403) the current study first attempted to identify distinct groups of individuals in the population characterised by the same profile of social isolation. Key indicators of loneliness, perceived social support, perceived quality of relationships, perceived social acceptance and social engagement were modelled using latent class analysis and three classes were identified. Class 1 (1.4%) was characterised by the highest endorsement probabilities across all social isolation indicators except loneliness. Class 2 (22%) was characterised by the highest endorsement probabilities for loneliness but also by near zero probabilities for endorsement of the perceived social support and perceived social acceptance indicators. Class 3 represented a baseline non-isolated class (77%). The current study then analysed associations between the latent classes and four discrete PLEs as measured by the psychosis screening questionnaire using multivariate binary logistic regression analysis. Compared to the baseline class, class 2 was significantly more likely to endorse all four PLEs (thought interference, OR=3.6; paranoia, OR=3.1; strange experiences, OR=2.3; hallucinations, OR=1.5) while class 1 was significantly more likely to endorse thought interference (OR=6.4) and strange experiences (OR=2.6) only. All analyses were adjusted for a range of known psychosis risk correlates.

Conclusions: Social isolation in the general population seems to vary between two distinct groups and is distinguishable by levels of loneliness, negative perceptions of social support and negative perceptions of social acceptance. Moreover, both groups seem to vary in relation to PLE risk.