A comparison between the Personality Diagnostic Questionnaire-4+ (PDQ-4+) and Longitudinal, Expert, All Data (LEAD) standard diagnoses was made in a sample of 100 patients with a high prevalence of both axis I and axis II disorders. The internal consistency was considered acceptable (alpha >.70) for 3 of 12 PDQ-4+ scales. Diagnostic agreement between the assessment methods was poor for specific personality disorders (PDs). The PDQ-4+ gave many false-positive diagnoses. Changing the diagnostic thresholds only marginally increased the agreement between methods or the diagnostic efficiency of the PDQ-4+. Multiple regression analyses showed that the discrepancies between the methods were strongly associated with the current level of perceived symptoms (SCL-90-R).