Suicidal ideation is insensitive to suicide risk after emergency department discharge: Performance characteristics of the Columbia-Suicide Severity Rating Scale Screener

Objectives: We describe the Columbia-Suicide Severity Rating Scale (C-SSRS)–Clinical Practice Screener’s ability to predict suicide and emergency department (ED) visits for self-harm in the year following an ED encounter. Methods: Screening data from adult patients’ first ED encounter during a 27-month study period were analyzed. Patients were excluded if they died during the encounter or left without being identified. The outcomes were suicide as reported by the state health department and a recurrent ED visit for suicide attempt or self-harm reported by the state hospital association. Multivariable regression examined the screener’s correlation with these outcomes. Results: Among 92,643 patients analyzed, eleven (0.01%) patients died by suicide within a month after ED visit. The screener’s sensitivity and specificity for suicide by 30 days were 0.18 (95% confidence interval [CI] = 0.00 to 0.41) and 0.99 (95% CI = 0.99 to 0.99). Sensitivity and specificity were better for predicting self-harm by 30 days: 0.53 (95% CI = 0.42 to 0.64) and 0.97 (95% CI = 0.97 to 0.97), respectively. Multivariable regression demonstrated that screening risk remained associated with both suicide and self-harm outcomes in the presence of covariates. Suicide risk was not mitigated by hospitalization or psychiatric intervention in the ED. Conclusions: The C-SSRS screener is insensitive to suicide risk after ED discharge. Most patients who died by suicide screened negative and did not receive psychiatric services in the ED. Moreover, most patients with suicidal ideation died by causes other than suicide. The screener was more sensitive for predicting nonfatal self-harm and may inform a comprehensive risk assessment. These results compel us to reimagine the provision of emergency psychiatric services.