Evaluation of DSM-III decision rules for case detection using joint conditional probability structures

Recent revisions of the psychodiagnostic nomenclature have relied on the specification of explicit criteria and decision rules for psychiatric diagnosis. Although both features have contributed to the increased reliability of psychiatric diagnosis, most empirical attention has been devoted to establishing the reliability of the criteria and the diagnoses, and relatively scant attention has been paid to the development of procedures for the optimization of the decision rules. The present report combines data from four independent studies of borderline personality disorder to illustrate a strategy for the optimization of psychodiagnostic decision rules which rely on polythetic criteria. The strategy considers the relationships among the criteria and the joint conditional probability structure of criteria combinations in optimizing the decision rule for the diagnosis. The advantages to the clinician, researcher, and theoretician of this, or any other, attempt to optimize the decision rules are discussed.