Reliability and validity of the Columbia Suicide Screen for suicide risk assessment in senior middle school students

Abstract

Objective To examine the reliability and validity of the Columbia Suicide Screen (CSS) for suicide risk assessment in senior middle school students. Methods 1 218 grade 1 to grade 3 students from two senior middle schools in Shanghai completed the CSS and Beck Depression Inventory (BDI). Cronbach's alpha coefficient was examined. Test–retest reliability was assessed in a subsample of 165 ones two weeks later. Content validity was evaluated using specialist analysis and sensibility analysis. The BDI was used to explore the correlative validity. 375 students were also interviewed by psychiatrists to assess whether they had suicidal tendency according to Mini International Neuropsychiatry Inventory (MINI). MINI was regarded as the gold standard to investigate the criterion validity of CSS. Results Students with suicide (either suicide attempts in the past or suicide within the previous year), and substance use were 225(18%, 225/1 218), 571(47%, 571/1 218)and 243(20%, 243/1 218). The Cronbach's alpha was estimated to be 0.83. Test–retest reliabilities for of suicide ideation, of substance use, of suicide problem and of suicide and depression were 0.90, 1.00, 0.88 and 0.78 (ICC). Alpha if item deleted was in the range of 0.82 to 0.83 by means of the sensibility analysis of CSS. Screening of depression and of suicide ideation were correlated with BDI. Correlative validity was 0.43 and 0.70(Kappa), respectively. The CSS algorithm that gave the best sensitivity (0.88) but the worst specificity (0.49) was of suicide problem. However of suicide and depression, of suicide and BDI≥5 yielded the balanced sensitivity/specificity trade–off: 72%/62%, 85%/60%. Conclusion The CSS has adequate internal consistency and content validity. But it demonstrates common sensitivity and specificity. Combination with screening of or diagnostic scale would be needed to improve its capability in identifying high–risk cases of suicide in senior middle school students. Key words: Questionnaires; Students; Suicide; Columbia Suicide Screen; Reliability; Validity