Examining the Properties of the Suicidal Ideation Attributes Scale within a Text-Based Intervention Service Triage.

Abstract

Objective: Text-based crisis services often gather information on users' suicide risk prior to intervention. Suicide.ca's crisis chat service uses a triage questionnaire combining a screening question on suicidal intent with the Suicidal Ideation Attributes Scale (SIDAS) to help identify users at highest risk. This study aimed to: (1) evaluate the SIDAS's psychometric properties in a real-world setting; and (2) assess its predictive validity in identifying users likely to be classified as high risk by counselors; and (3) examine whether combining the SIDAS with the suicidal intent question improves detection of high-risk users.

Method: At triage, users completed the French version of the SIDAS along with questions on age, gender, and suicidal intent. Data from 1,480 users who subsequently received crisis intervention were analyzed. Predictive validity was evaluated using counselors' suicide risk assessments.

Results: Confirmatory factor analysis replicated the unifactorial structure of the SIDAS. All items showed satisfactory factor loadings except item 2. Internal consistency was good (ω = .80). ROC analysis showed low sensitivity and specificity in differentiating high- vs low-risk users. Adding the SIDAS (cutoff ≥ 32) to the suicidal intent item in a hierarchical regression yielded a statistically significant but modest increase in explained variance.

Conclusions: The SIDAS showed similar psychometric performance in a real-world context. While a revised cutoff improved prediction slightly, its added value in identifying users deemed high risk by counselors appears limited. The clinical utility of the SIDAS and implications of using a revised cutoff for triage in text-based intervention services will be discussed.