Psychosis as a risk factor for suicidal thoughts and behaviors: a meta-analysis of longitudinal studies

• Published in: Psychological medicine 2018-04, Vol.48 (5), p.765-776
• Main Authors: Huang, X., Fox, K. R., Ribeiro, J. D., Franklin, J. C.
• Format: Article
• Language: English
• Subjects: Artificial intelligence; Behavior; Death; Death & dying; Hallucinations; Humans; Longitudinal Studies; Meta-analysis; Negative symptoms; Original Articles; Psychiatry; Psychosis; Psychotic Disorders - complications; Psychotic Disorders - epidemiology; Psychotic Disorders - physiopathology; Psychotic symptoms; Schizophrenia; Statistical analysis; Suicidal Ideation; Suicide; Suicide - statistics & numerical data; Suicide, Attempted - statistics & numerical data; Suicides & suicide attempts; Systematic review
• ISSN: 0033-2917; 1469-8978
• DOI: 10.1017/S0033291717002136

Research has long noted higher prevalence rates of suicidal thoughts and behaviors among individuals with psychotic symptoms. Major theories have proposed several explanations to account for this association. Given the differences in the literature regarding the operationalization of psychosis and sample characteristics, a quantitative review is needed to determine to what extent and how psychosis confers risk for suicidality. We searched PsycInfo, PubMed, and GoogleScholar for studies published before 1 January 2016. To be included in the analysis, studies must have used at least one psychosis-related factor to longitudinally predict suicide ideation, attempt, or death. The initial search yielded 2541 studies. Fifty studies were retained for analysis, yielding 128 statistical tests. Suicide death was the most commonly studied outcome (43.0%), followed by attempt (39.1%) and ideation (18.0%). The median follow-up length was 7.5 years. Overall, psychosis significantly conferred risk across three outcomes, with weighted mean ORs of 1.70 (1.39-2.08) for ideation, 1.36 (1.25-1.48) for attempt, and 1.40 (1.14-1.72) for death. Detailed analyses indicated that positive symptoms consistently conferred risk across outcomes; negative symptoms were not significantly associated with ideation, and were protective against death. Some small moderator effects were detected for sample characteristics. Psychosis is a significant risk factor for suicide ideation, attempt, and death. The finding that positive symptoms increased suicide risk and negative symptoms seemed to decrease risk sheds light on the potential mechanisms for the association between psychosis and suicidality. We note several limitations of the literature and offer suggestions for future directions.