There is a need of identifying predictors of suicide. With effective treatment interventions for those at highest risk, suicide can be prevented.

To investigate variables possibly associated with suicide, by comparing patients in psychiatry deceased by suicide to matched controls. We compared the psychiatric ward consumption two years prior to suicide, previous medical and psychological treatments, somatic co-morbidity and the professions of health care staff in previous visits.

The 153 suicide cases, 101 (66%) men and 52 (34%) women, aged between 13 and 96 years, were identified from the National Cause of death registry. All cases had previous contact in psychiatry in Örebro County and died between January 1st 2007 through December 31st 2013. Each case had one gender-, age- and diagnosis matched patient control from psychiatry. Data on care usage were collected from medical records. Correlations were calculated between ward variables and the outcome variable suicide yes/no. Variables significantly correlated to suicide were used in an adjusted logistic regression model to analyse possible associations with suicide.

Among the cases deceased by suicide, we found significantly fewer psychiatric outpatient visits (P < .01), a smaller amount receiving systematic psychological treatment (P < .01), less pharmacological interventions (P < .01) and a smaller amount of treatment provided by a multi-professional team (P < .01). Less occurrence of psychological treatment (OR(adj): 0.459; 95% CI 0.224–0.874, P < .05) was significantly associated to suicide.

Results indicate that not only repeated visits in psychiatric care, but the content of visits (i.e. psychological interventions) seems to make an important difference between life and death.

The authors have not supplied their declaration of competing interest.