BRIEF COMMUNICATIONS Screening Psychiatric Patients for Axis II Disorders* DAVID NUSSBAUM, Ph.D. I AND RiCHARD ROGERS, Ph.D. 2 This study examined the usefulness of the SCID-PQ on a sample of 82 inpatients as a screening measure for Axis II disorders. The SCID-PQ proved effective (i.e., resulted in a very low false negatives and moderate false positives) in screening for seven of the II DSM-III-R personality disor- ders. Of the four remaining disorders, minor modifications were suggested which resulted in improved classification rates for dependent and schizoid personality disorders. T he emergence of multiaxial diagnoses in the DSM-III and the DSM-III-R is largely an acknowledgement of the importance of personality disorders in the effective treatment of psychiatric patients. As observed by Frances (I), recent investigations have underscored the way in which the com- bination of Axis I and Axis II disorders shapes both the patient's clinical presentation and treatment response. The rather formidable task facing clinicians is how to screen psychiatric patients effectively for each of the 11 Axis II disorders. The diagnosis of personality disorders reflects their rich and divergent origins in psychoanalytical and personological theory which, in turn, have strongly influenced the successive changes in diagnostic nosology (unpublished manuscript, 1988). Problems which naturally arise from such a theory- driven endeavour are I. overlapping and sometimes indistinct boundaries between specific personality disorders (1-3); 2. emphasis on categorical rather than dimensional diagnosis (3,4) which may lead to a bewildering array of symptom variations for a single diagnosis (5); and 3. conceptual diffi- culties in establishing the inter-relationships among the dis- orders that substantially contribute to problems in establishing reliable inclusion/exclusion criteria (6-9). It is not surprising that clinicians have had great difficulties *Manuscript received June 1991, revised December 1991. Jpsychologist-in-Charge and Coordinator of Research, METFORS/Clarke Institute of Psychiatry; Lecturer, Department of Psychiatry, University of Toronto, Toronto, Ontario. 2 Associate Professor, Department of Psychology, University of North Texas, Denton, Texas; formerly, Senior Psychologist and Coordinator of Research, METFORS/Clarke Institute of Psychiatry; Associate Professor of Psychiatry and Psychology, University of Toronto, Toronto, Ontario. Address reprint requests to: Dr. D. Nussbaum, METFORS/Clarke Institute of Psychiatry, 1001 Queen Street West, Toronto, Ontario M6J IH4 Can. J. Psychiatry Vol. 37, November 1992 658 achieving satisfactory reliability for Axis II disorders (range = 0.56 to 0.64) (10). Spitzer and Williams (11), in an. effort to improve the diagnostic reliability, operationalized the DSM-III-R criteria using a structured interview, the Structured Clinical Interview for DSM-III-R (SCID-R). Earlier efforts with standardized interviews have demonstrated the value of the interviews in achieving satisfactory to excellent interrater reliability (12,13). One major component of the SCID-R is devoted to the diagnoses of Axis II disorders (14). This diagnostic ap- proach appears promising for classifying the full range of personality disorders (15). In order to ease the substantial burden placed on clinicians by multiaxial diagnoses, Spitzer and Williams (16) also devised a self-administered question- naire, the SCID Personality Questionnaire (SCID-PQ), which parallels the SCID-II-R. If effective, the SCID-PQ would greatly facilitate the diagnosis of personality disorders by screening psychiatric patients for co-existing Axis II disor- ders. Until now however, there have been no reported studies evaluating the effectiveness of the SCID-PQ as a screening device (17). This study investigated the usefulness of the SCID-PQ as a screening measure for Axis II disorders with inpatients. Our goal was to study the personality disorders which are amena- ble to cost-effective screening using the SCID-PQ. In addi- tion, we investigated alternative cut-off scores for the SCID to improve classification rates for the four diagnoses for which the SCID-PQ was less effective. Method A sample of 82 inpatients was recruited from METFORS, a forensic assessment unit at the Clarke Institute of Psychia- try. METFORS is comprised of the outpatient Brief Assess- ment Unit (BAU), where patients referred by the courts are screened by interview and chart review (if available) for their competence to stand trial and returned to the correctional facility the same day; and a 23 bed inpatient unit for those patients whose fitness for trial is unclear. Evaluations of criminal responsibility, and occasionally, examinations of potential for violence are also conducted at METFORS. Inpatients typically stay on the unit for a period of two weeks. We therefore had a diverse clinical sample of patients referred by the courts; not a sample preselected on the basis of diag- nosis or treatment modality. The 82 subjects we studied were consecutively admitted patients who were able to provide informed consent and were willing to participate in the re- search project. The participants completed an informed con-