AN EMPIRICAL-METHOD TO REFINE PERSONALITY-DISORDER CLASSIFICATION USING STEPWISE LOGISTIC-REGRESSION MODELING TO DEVELOP DIAGNOSTIC-CRITERIA AND THRESHOLDS

This study of DSM-III-R personality disorder (PD) classification provi des an empiricial approach to determine (1) the discriminative power o f each criterion and (2) the optimal number of criteria needed to diag nose the presence of each PD. A semistructured assessment of 110 outpa tients was performed for the 11 PDs and their 104 diagnostic criteria. Sensitivity, specificity, and predictive powers were calculated for e ach criterion item. Logistic regression was performed to determine (1) univariate weightings of the individual criteria as applied to a give n diagnosis, and (2) multivariate measures of the criteria that signif icantly improved the chi-square value in a stepwise fashion. The signi ficant items were then equally weighted to calculate the optimal numbe r needed to diagnose category membership. Of 104 PD criteria, 41 discr iminated at a significance level of .05 or less, and each PD could be optimally diagnosed with fewer criteria than currently required. We ca n empirically reduce the number of criteria combinations comprising in dividual categories, decrease heterogeneity, and narrow diagnostic bou ndaries. This increases the likelihood of identifying etiological fact ors, predictors of clinical course, specific treatments, familial aggr egation, and neurobiological correlates for the PD taxa. Copyright (C) 1994 by W.B. Saunders Company