Separate and joint associations of psychosis and criminal justice involvement with hospital-treated self-harm and injury: a population-based cohort study

Abstract Background People with psychotic disorders are overrepresented in criminal justice systems, and psychosis and justice involvement are each associated with self-harm and other injury. The excess risk where psychosis and justice involvement co-occur has not been quantified, and whether prevention should target the illness, the justice exposure, or both is unclear. This study estimated their separate and joint associations with hospital-treated self-harm and injury. Methods In a population-based cohort study, six statewide administrative databases from New South Wales, Australia, were linked. We identified 463,641 people in four mutually exclusive groups defined by psychosis and criminal justice involvement (justice involvement only, psychosis only, neither, and both). The primary outcome was a first hospital-treated episode of self-harm within 2 years of follow-up. To account for common causes of the exposures and of self-harm, analyses were restricted to the 103,400 people whose measured characteristics made membership of every group plausible (the common-support population). The groups were weighted to a common distribution of demographic, prior injury, health care use, and comorbidity covariates, and 2-year risks were standardised by g-computation. Results In the common-support population, 2-year risk of self-harm was 4.59% with both psychosis and justice involvement, 3.23% with justice involvement alone, 1.96% with psychosis alone, and 0.69% with neither. Psychosis added 1.37% points (95% CI 0.68 to 2.32) with justice involvement and 1.27 (0.93 to 1.60) without. Hospital-treated injury affected approximately 25% of both justice-involved groups, and justice involvement added 11.34% points (9.55 to 13.29) with psychosis and 11.48 (10.55 to 12.27) without. Conclusions Psychosis was associated with an increment in 2-year self-harm risk of approximately 1.3% points that was similar inside and outside the justice system, and the two exposures combined additively. Justice involvement was associated with an excess of more than 11% points in any hospital-treated injury regardless of psychosis, including assault and unintentional injury as well as self-harm. Injury prevention should target everyone leaving custody, and people with psychosis additionally require continuity of psychiatric care and self-harm prevention across custodial and community settings.