Clinical predictions of dangerousness: two-year follow-up of 408 pre-trial forensic cases.

The past decade has witnessed a number of serious attempts at empirically investigating the ability of mental health professionals, especially of psych iatrists, to discriminate between those individuals who will not act dangerously and those individuals who eventually will act dangerously .1-:1 These investigations have concentrated on the assessment of patients in hospitals for the criminally insane since it is in this context mental health professionals are called on routinely to offer their opinion regarding future post-release dangerousness. Based on data arising from these studies, less-thanfavorable conclusions have been drawn regarding the accuracy of dangerousness assessments and the associated clinical skills of the predicting psychiatrists. The problem of "over-prediction" has been widely documented in the literature so that, to use Monahan's recent summary of the findings, "psychiatrists and psychologists are accurate in no more than one out of three predictions of violent behavior over a several-year period among institutionalized populations that had both committed violence in the past (and thus had high base rates for it) and who were diagnosed as mentally ill."4 With the publication of these results has come. needless to say. a surge of criticism regarding the role of psychiatry in the courtroom, nowhere more thoroughly and persuasively argued than in the work of Ennis and Litwack.:; As convincing an argument, however, has been leveled at the interpretation of these results and the manner in which they are purported to apply to the task of clinically predicting dangerousness. Shah and, more recently, Monahan7 are very clear in their conceptions of the issue. Each maintains that the conclusions of investigations in this area that seek to discredit the clinical ability of the mental health professional to predict future dangerousness in individuals may not be altogether justified given (a) the' 'yes" versus "no" requirement of the prediction task imposed on the clinician to date in the area of dangerousness represents a far more stringent demand than does