Objective: Accurate prediction of suicide remains elusive due to lack of predictive measures. Given the Columbia-Suicide Severity Rating Scale's (C-SSRS) emerging "gold-standard" status for risk assessment, studies are needed to assess its psychometric properties, particularly predictive validity. The current study adds to the limited literature by assessing the C-SSRS's internal consistency, factor structure, concurrent validity, and predictive validity.
Methods: In this longitudinal study of 1,055 adults with DSM-IV diagnoses consecutively admitted to a specialized psychiatric hospital between July 1, 2012, and June 30, 2014, patients completed standardized assessments, including the C-SSRS, at admission and 2, 12, and 24 weeks postdischarge.
Results: The C-SSRS evidenced excellent internal consistency (ordinal α = .95). Principal components analysis (PCA) revealed a 2-factor solution, accounting for 65.3% of the variance across items. The severity of ideation and behavioral items loaded onto the first factor, and the intensity of ideation items loaded onto the second factor. The total score, factors, and the most severe ideation single item were moderately correlated with other measures of suicidality (0.27 ≤ r ≤ 0.58; P < .0001). The summary score from the ideation/behavior factor was found to be modestly correlated with any suicide-related behavior within the 6 months following hospitalization. Receiver operator characteristics indicated that the C-SSRS performed adequately in correctly classifying any suicide-related behavior within 6 months of discharge from the hospital (AUC = 0.757, P < .001) with the total score and summary score from the ideation/behavior factor providing the best balance between sensitivity (0.694) and specificity (0.652-0.674).
Conclusions: This study is the first to assess the factor structure of the C-SSRS in a large, high-risk sample. The measure has solid psychometric properties and merits use as a suicide risk assessment measure.
© Copyright 2016 Physicians Postgraduate Press, Inc.