| Abstract: | The publication of the DSM-III by the American Psychiatric Association has fueled interest among psychologists in developing alternative classifications. A popular approach is to use multivariate methods known as cluster analysis to identify clinical syndromes. Empirical clustering techniques offer potential for making diagnoses more objective, reliable, and predictive of outcome from differential treatment. However, problems exist at the interface between this statistical technology and psychiatric classification. To date, cluster-analytic studies have had limited impact on psychiatric nomenclature, and clinicians have not found the cluster-derived syndromes to be relevant to everyday practice. The present authors discuss reasons that underlie the lack of impact and describe directions for upgrading both the scientific quality and clinical acceptance of cluster-analysis research. It is concluded that more emphasis must be placed on fostering the clinical use of empirically derived classifications. (69 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved) |
