Service utilization and suicide among people with schizophrenia spectrum disorders

Objective — To compare individuals with and without schizophrenia spectrum disorders (SSD) (schizophrenia, schizoaffective disorder, or psychotic disorder not otherwise specified) who die by suicide.

Method — This is a retrospective case control study which compared all individuals who died by suicide in Ontario, Canada with (cases) and without (controls) SSD between January 1, 2008 and December 31, 2012. Cases (individuals with SSD) were compared to controls on demographics, clinical characteristics, and health service utilization proximal to suicide. A secondary analysis compared the characteristics of those with SSD and those with severe mental illness (defined as those without SSD who have had a psychiatric hospitalization within the five-years before suicide (excluding the 30 days prior to death)).

Results — Among 5650 suicides, 663 (11.7%) were by individuals with SSD. Compared to other suicides, SSD suicides were significantly more likely to be between the ages of 25–34. SSD suicide victims were significantly more likely to reside in the lowest income neighbourhoods and to reside in urban areas. SSD victims were also significantly more likely to have comorbid mood and personality disorders and all types of health service utilization, including outpatient mental health service contact in the 30 days prior to death, even when compared only with those who had a history of mental health hospitalization.

Conclusions — Individuals with schizophrenia spectrum disorder account for over 1 in 10 suicide deaths, tend to be younger, poorer, urban, more clinically complex, and have higher rates of mental health service contact prior to death. The demographic and service utilization differences persist even when the SSD group is compared with a population with severe mental illness that is not SSD. Suicide prevention strategies for people with schizophrenia spectrum disorder should emphasize the importance of clinical suicide risk assessment during clinical encounters, particularly early in the course of illness.