The Use of Clustering Techniques for the Classification of Psychiatric Patients

Abstract

There has been a reawakening of interest in the classification of psychiatric patients. Clinicians and researchers alike have realized that it is impossible to evaluate methods of treatment, determine aetiology, or measure the course of illness of psychiatric disorders without adequate methods for diagnosis. On the other hand, it has been shown that conventional clinical methods for making psychiatric diagnoses are of distressingly low reliability, except for the broadest categories, and have only marginal relationships to such criteria of validity as common aetiology, common response to treatment, and common prognosis (Baldessarini, 1970; Beck et al., 1962; Jenkins, 1966). Klein (1967) has shown that the usual clinical diagnoses, because of low validity, can actually obscure important relationships between types of psychopathology and such crucial variables as response to treatment.