In the Journals

Current risk instruments fail to predict suicide

Several suicide risk scales lack the specificity to predict suicidal behavior in a psychiatric context, according to results of a prospective multicenter cohort study published in Journal of Clinical Psychiatry.

“The American Psychiatric Association guideline for the assessment and treatment of patients with suicidal behavior states that suicide assessment scales may be used as aids but not as substitutes for a thorough clinical evaluation,” Åsa U. Lindh, MD, PhD, of the Centre for Psychiatry Research at Karolinska Institute in Stockholm, and colleagues wrote. “In Great Britain, the National Institute for Health and Care Excellence guidelines on long-term management of self-harm classify the use of suicide risk assessment tools and scales to predict future suicide or repetition of self-harm as a ‘do not do’ recommendation but state that they might be of help in structuring the risk assessment.”

According to the researchers, few head-to-head designs have compared existing suicide risk scales. In the present study, they compared the predictive accuracy of the Suicide Intent Scale, the Suicide Assessment Scale, the Karolinska Interpersonal Violence Scale and the Columbia-Suicide Severity Rating Scale for suicides and suicide attempts within 3 and 12 months of a self-harm episode. They included data from 804 patients aged 18 to 95 years with a recent episode of self-harm who were assessed in psychiatric emergency settings in Sweden from 2012 to 2016.

The researchers considered a sensitivity of at least 80% combined with a specificity of at least 50% as minimally acceptable. Among the study population, at least one suicide attempt was recorded for 216 participants, as well as 19 suicides. The Suicide Assessment Scale and the Columbia-Suicide Severity Rating Scale classified the 114 suicide attempts that occurred during the first 3 months better than chance, with the latter instrument yielding a sensitivity of 79.8% and a specificity of 51.5% (P < .001). At 1-year follow-up, both scales performed better than chance, but no cutoff on either instrument exhibited the desired minimum specificity or sensitivity.

Lindh and colleagues noted that the Suicide Intent Scale was the only instrument able to correctly predict suicides, but this capability was primarily in the short timeframe of 3 months.

“Although acceptable accuracy statistics could be obtained for the data-driven cutoffs for the [Suicide Intent Scale] in predicting suicide and the [Columbia-Suicide Severity Rating Scale] in predicting suicide attempts, none of the instruments evaluated in this study was specific enough for prediction of suicidal behavior in a psychiatric context,” the researchers wrote. “Thus, our findings support the existing guidelines concerning the use of suicide risk assessment instruments. These instruments may still be of use in the education of physicians under training and inexperienced staff members and as a way of structuring the clinical assessment.” – by Joe Gramigna

Disclosures: The authors report no relevant financial disclosures.

Several suicide risk scales lack the specificity to predict suicidal behavior in a psychiatric context, according to results of a prospective multicenter cohort study published in Journal of Clinical Psychiatry.

“The American Psychiatric Association guideline for the assessment and treatment of patients with suicidal behavior states that suicide assessment scales may be used as aids but not as substitutes for a thorough clinical evaluation,” Åsa U. Lindh, MD, PhD, of the Centre for Psychiatry Research at Karolinska Institute in Stockholm, and colleagues wrote. “In Great Britain, the National Institute for Health and Care Excellence guidelines on long-term management of self-harm classify the use of suicide risk assessment tools and scales to predict future suicide or repetition of self-harm as a ‘do not do’ recommendation but state that they might be of help in structuring the risk assessment.”

According to the researchers, few head-to-head designs have compared existing suicide risk scales. In the present study, they compared the predictive accuracy of the Suicide Intent Scale, the Suicide Assessment Scale, the Karolinska Interpersonal Violence Scale and the Columbia-Suicide Severity Rating Scale for suicides and suicide attempts within 3 and 12 months of a self-harm episode. They included data from 804 patients aged 18 to 95 years with a recent episode of self-harm who were assessed in psychiatric emergency settings in Sweden from 2012 to 2016.

The researchers considered a sensitivity of at least 80% combined with a specificity of at least 50% as minimally acceptable. Among the study population, at least one suicide attempt was recorded for 216 participants, as well as 19 suicides. The Suicide Assessment Scale and the Columbia-Suicide Severity Rating Scale classified the 114 suicide attempts that occurred during the first 3 months better than chance, with the latter instrument yielding a sensitivity of 79.8% and a specificity of 51.5% (P < .001). At 1-year follow-up, both scales performed better than chance, but no cutoff on either instrument exhibited the desired minimum specificity or sensitivity.

Lindh and colleagues noted that the Suicide Intent Scale was the only instrument able to correctly predict suicides, but this capability was primarily in the short timeframe of 3 months.

“Although acceptable accuracy statistics could be obtained for the data-driven cutoffs for the [Suicide Intent Scale] in predicting suicide and the [Columbia-Suicide Severity Rating Scale] in predicting suicide attempts, none of the instruments evaluated in this study was specific enough for prediction of suicidal behavior in a psychiatric context,” the researchers wrote. “Thus, our findings support the existing guidelines concerning the use of suicide risk assessment instruments. These instruments may still be of use in the education of physicians under training and inexperienced staff members and as a way of structuring the clinical assessment.” – by Joe Gramigna

Disclosures: The authors report no relevant financial disclosures.