Psychometric Properties and Longitudinal Predictors of the Columbia Suicide Severity Rating Scale (C-SSRS) in a...

The Columbia Suicide Severity Rating Scale (C-SSRS) is widely used for assessing suicidal thoughts and behaviors, yet evidence on the validity of its scoring outputs in community-based samples from low- and middle-income countries remains scarce.

Participants included 2076 young adults (mean age 23.45, 47% female) from the Brazilian High-Risk Cohort Study at 15-year follow-up. We evaluated C-SSRS Ideation Severity (0-5) and Intensity (0-25) scores using internal consistency, confirmatory factor analysis, item response theory, Guttman scalability, convergent/discriminant validity, known-groups comparisons, ROC analysis, and longitudinal associations with childhood risk factors via adjusted count regression.

37% endorsed lifetime suicidal ideation; 16% reported attempts. Internal consistency was low for severity (KR-20 = 0.57) and intensity (ω = 0.67). A two-factor model improved fit over a unidimensional structure (CFI = 0.997 vs. 0.954). Guttman scalability was acceptable (CR = 0.940) but with violations at higher severity levels. C-SSRS severity discriminated attempt history (AUC = 0.93). Prior self-harm was the strongest longitudinal predictor (IRR = 2.06), with a dose-response across risk categories.

C-SSRS scores show low internal consistency and multidimensional structure, yet strong clinical discrimination and robust 15-year longitudinal associations. Caution is warranted when using these scores as continuous indices.