Systematic review of risk factors for violence in psychosis: 10-year update

Abstract

Background

Understanding risk factors for violence in people with psychosis can inform risk management and violence prevention. However, much of the evidence comes from cross-sectional studies, and previous reviews require updating.

Aims

To synthesise evidence from longitudinal studies on risk factors for violence in people with schizophrenia-spectrum disorders, bipolar disorder or other affective psychoses.

Method

We searched five bibliographic databases up to June 2022. We identified longitudinal studies reporting risk factors for violence in individuals diagnosed with schizophrenia or other psychoses using DSM or ICD criteria. If ≥3 independent samples reported a risk factor, we conducted random-effects meta-analyses to provide a pooled estimate. We also meta-analysed risk factors by major domains.

Results

We identified 47 longitudinal studies on risk factors for violence in psychosis, representing 41 independent samples – 21 from the original and 20 from the updated review – and 203 297 individuals. A total of 30 risk factors were present in ≥3 independent samples. Criminal history factors were associated with the greatest risk of violent outcomes (pooled odds ratio 3.50, 95% CI = 2.37, 5.16), followed by substance misuse factors (odds ratio 2.36, 95% CI = 1.99, 2.80). Many treatment-related factors were protective (odds ratio 0.54, 95% CI = 0.34, 0.85). Effect estimates were attenuated in inpatient settings. We also identified novel risk factors, including cannabis use, in a secondary analysis (odds ratio 3.34, 95% CI = 2.32, 4.82).

Conclusions

Using longitudinal evidence, we have validated comorbid substance misuse and criminal history as major risk factors for violence in psychosis. Novel factors such as cannabis use need further replication. Several identified factors are possible intervention targets if associations are found to be causal.