Diagnostic validity of ICD-10 personality dimensions: a multitrait-multimethod analysis of two self-report questionnaires and a structured interview

Abstract

Background/aims: There is growing evidence that the validity of diagnostic methods for personality disorders (PD) may be insufficient. Although the ICD-10 classification system is widely used, there is little data concerning its validity for diagnosing PD.

Methods: To examine convergent and discriminant validity of ICD-10 PD, corresponding dimensions were calculated using the Inventory of Clinical Personality Accentuations (self-rating) and the International Personality Disorder Examination ICD-10 module (interview and screening questionnaire). These were administered to 42 psychiatric patients. A multitrait-multimethod analysis of the 2 self-administered questionnaires and the structured interview was conducted.

Results: Correspondence between methods is significant at the level of 0.01 for the schizoid, borderline, anankastic, anxious and dependent PD dimensions. The investigation of discriminant validity within and between instruments reveals several strong relationships between different PD dimensions.

Conclusion: Convergent validity can be considered sufficient for 5 of the 9 examined PD dimensions. Non-convergence must partly be interpreted as caused by poor discriminant validity inherent in the ICD-10 classification. Conceptualization of PD as completely distinguishable entities may not reflect the clinical reality of a partial overlap between personality disorders.