Key points are not available for this paper at this time.
The aim of this study was to assess interrater reliability and provide initial data bearing on the validity of a method of assessing personality disorders (PDs) that does not presume that patients can accurately self-report personality pathology. In a sample of 24 outpatients, two clinician-judges independently applied the Shedler-Westen Assessment Procedure-200 (SWAP-200; Westen & Shedler, 1999a, 2000), a 200-item Q-sort procedure for assessing personality pathology, to data from the Clinical Diagnostic Interview (Westen, 2002), a systematic clinical interview that mirrors and standardizes methods used by experienced clinicians to diagnose personality. In 16 of the 24 cases, the treating clinician also independently described the patient using the SWAP-200 Q-sort, based on longitudinal knowledge of the patient over the course of treatment, blind to the interview data. Interrater reliability was uniformly high, with median correlations between interviewers at r > .80. Interviewer-treating clinician correlations were also high, with median convergent validity coefficients at r > .80. Diagnostic overlap (discriminant validity) was moderate for dimensional DSM-IV diagnoses, reflecting extensive comorbidity among disorders, but minimal for empirically derived diagnoses identified in prior research. Treating clinicians' dimensional PD diagnoses using this method also strongly predicted interviewer-rated measures of adaptive functioning. The findings provide preliminary support for the reliability and validity of an alternative to structured interviews for diagnosing personality pathology, and suggest that the way to improve validity of personality diagnosis may not be to minimize clinical inference but to quantify it using psychometric instruments.
Westen et al. (Fri,) studied this question.