Background
Schizophrenia is associated with premature mortality, partly through increased suicide rates.
Aims
To examine (1) if persecutory ideas, auditory hallucinations, and probable cases of psychosis are associated with suicidal thoughts or attempts cross-sectionally and prospectively, and (2) if such links are mediated by specific affective factors (depression, impulsivity, mood instability).
Method
We analysed the 2000, 2007, and 2014 British Adult Psychiatric Morbidity Surveys (APMS) separately. Measures of psychosis provided independent variables for multi-stage logistic regressions, with suicidal thoughts and attempts as dependent variables. We also conducted analyses to assess mediation by affective variables, and longitudinal analyses on a subset of the 2000 dataset.
Results
In every dataset, persecutory ideas, auditory hallucinations and probable psychosis were associated cross-sectionally with lifetime suicidal attempts and thoughts, even after controlling for confounders, with a single exception (persecutory ideation and suicide attempts were unconnected in APMS 2014). Cross-sectional associations between auditory hallucinations and suicidal phenomena were moderated by persecutory ideation. In the 2000 follow-up, initial persecutory ideas were associated with later suicidal thoughts (O.R. 1.77,
p
< 0.05); there were no other longitudinal associations. In the 2007 and 2014 datasets, mood instability mediated the effects of psychotic phenomena on suicidality more strongly than impulsivity; depression was also an important mediator. There were appreciable direct effects of positive symptoms on suicidal thoughts and behaviour.
Conclusions
Improving psychotic symptoms and ameliorating co-morbid distress may in itself be effective in reducing suicidal risk in schizophrenia. Given their potential mediating role, mood instability and depression may also be targets for intervention.