Citation

Cochrane-Brink KA, Lofchy JS, Sakinofsky I. Gen. Hosp. Psychiatry 2000; 22(6): 445-451.

Affiliation

Department of Psychiatry, University of Toronto, Toronto, Ontario, Canada.

Copyright

(Copyright © 2000, Elsevier Publishing)

DOI

unavailable

PMID

11072061

Abstract

This study explores the usefulness of clinical rating scales in the assessment of suicidal risk in an urban psychiatric teaching hospital. Admission for clinically evaluated suicide risk was the outcome variable because actual suicide occurs rarely. Six clinical scales identified high-risk patients: the Modified SAD PERSONS scale, revised Beck Depression Inventory, Beck Anxiety Inventory, Beck Hopelessness Scale, Beck Scale for Suicidal Ideation (BSS), and the High-Risk Construct Scale (NEW). It was hypothesized that patients who scored highly on the clinical scales were more likely to be admitted. Five of the scales had previously established psychometric properties, while one was new and untested. For our patient population, the established scales had 100% sensitivity and negative predictive value, but lower specificity and positive predictive value (range = 38-90% & 28-71%). We performed a correlation matrix and regression analysis to determine which scale(s) best predicted admission based upon suicidal concerns. The previously untested NEW scale was the best predictor followed by the BSS. Clinical rating scales cannot predict suicide in the individual and strict cut-off scores should not be used to dictate admission to hospital. However, the information provided can be a valuable adjunct to suicide risk assessment in psychiatric and non-psychiatric emergency settings

Language: en