Violent crime, suicide, and premature mortality in patients with schizophrenia and related disorders: a 38-year total population study in Sweden

Abstract

Background: People with schizophrenia and related disorders are at an increased risk of adverse outcomes, including conviction of a violent offence, suicide, and premature mortality. However, the rates of, and risk factors for, these outcomes need clarification as a basis for population-based and targeted interventions. We aimed to determine rates and risk factors for these outcomes, and investigate to what extent they are shared across outcomes and are specific to schizophrenia and related disorders.

Methods: We undertook a total population cohort study in Sweden of 24 297 patients with schizophrenia and related disorders between January, 1972 and December, 2009. Patients were matched by age and sex to people from the general population (n=485 940) and also to unaffected sibling controls (n=26 357). First, we investigated rates of conviction of a violent offence, suicide, and premature mortality, with follow-up until conviction of a violent offence, emigration, death, or end of follow-up (Dec 31, 2009), whichever occurred first. Second, we analysed associations between these adverse outcomes and sociodemographic, individual, familial, and distal risk factors, for men and women separately, with Cox proportional hazards models. Finally, we assessed time trends in adverse outcomes between 1972 and 2009, for which we compared patients with unaffected siblings, and analysed associations with changes in the number of nights spent in inpatient beds in psychiatric facilities nationwide.

Findings: Within 5 years of their initial diagnosis, 13·9% of men and 4·7% of women with schizophrenia and related disorders had a major adverse outcome (10·7% of men and 2·7% of women were convicted of a violent offence, and 3·3% of men and 2·0% of women died prematurely of any cause). During the study, the adjusted odds ratio of any adverse outcomes for patients compared with general population controls was 7·5 (95% CI 7·2-7·9) in men and 11·1 (10·2-12·1) in women. Three risk factors that were present before diagnosis were predictive of any adverse outcome: drug use disorders, criminality, and self-harm, which were also risk factors for these outcomes in unaffected siblings and in the general population. Over the period 1973-2009, the odds of these outcomes increased in patients with schizophrenia and related disorders compared with unaffected siblings.

Interpretation: Schizophrenia and related disorders are associated with substantially increased rates of violent crime, suicide, and premature mortality. Risk factors for these three outcomes included both those specific to individuals with schizophrenia and related disorders, and those shared with the general population. Therefore, a combination of population-based and targeted strategies might be necessary to reduce the substantial rates of adverse outcomes in patients with schizophrenia and related disorders.

Figure 1

Incidence of conviction of a violent offence, suicide, premature mortality, and any adverse outcome within 1, 2, and 5 years of first diagnosis in patients with schizophrenia and related disorders

Figure 2

Ratio of odds ratios of adverse outcomes, by year of diagnosis The graph presents 3-year rolling averages of the ratios of odds ratios of adverse outcomes (comparison of analyses of patients with schizophrenia and related disorders vs unaffected siblings). In statistical analyses, the ratio of odds ratios for conviction of a violent offence increased significantly by 1·1% (95% CI 0·1–2·2%) per year. This ratio of odds ratios also increased, but not significantly, for suicide (1·7% increase per year [95% CI −0·9 to 4·3]) and did not change significantly for premature mortality (0·2% increase per year [−1·3% to 1·6%]).