The prediction of suicide: why is it so difficult?

Methods for selecting those persons in a clinical population who will later commit suicide inevitably include large numbers of false positives--persons with similar characteristics who will not take their lives. This is true because of the absence of unique predictors and the statistical properties of infrequent events. At a clinical level, the focus is on risk detection rather than on specific behavior prediction. Since suicide is intimately related to certain psychiatric illnesses, effective treatment of those illnesses can prevent suicides. This must already be taking place. Prevention, however, generates no data. If suicide is difficult to predict, its prevention is even more difficult to detect.