Predictive validity of the suicide probability scale among adolescents in group home treatment

Objective: To determine the predictive validity of the Suicide Probability Scale (SPS). Method: Prospective design, predicting subsequent suicide attempts, suicide verbalizations, and minor self-destructive behaviors from SPS scores at admission to a group home in a sample of 855 adolescents. Results: SPS scores significantly predicted all three suicide-relevant measures. Only 48% of subsequent attempts could have been predicted from SPS criteria that would have put 21% of newly admitted adolescents in the at-risk group. Although the prediction is far from perfect, it can prevent some suicide attempts when combined with effective suicide precautions. Conclusions: This is the first known evidence of predictive validity of any measure of suicide risk in adolescents. Its predictive ability may have been reduced by interventions designed to reduce suicide risk. Despite the large number of false-positives and modest predictive validity, a large body of literature indicates that quantitative predictions perform better than clinical judgments in predicting psychological outcomes. However, the SPS measures only a subset of factors predictive of suicide, including suicidal ideation, hopelessness, and social isolation. Risk factors not assessed by the SPS include previous suicide attempts, conduct disorder, substance abuse, and reasons for living. These additional risk factors may be important to improve upon the predictive validity of the SPS.