Erik Falkum1, Geir Pedersen, Sigmund Karterud. 1. The Research Department, Clinic of Mental Health, Aker University Hospital, 0320 Oslo, Norway. erik.falkum@medisin.uio.no
Abstract
OBJECTIVE: This article examines reliability and validity aspects of the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) paranoid personality disorder (PPD) diagnosis. METHOD: Patients with personality disorders (n = 930) from the Norwegian network of psychotherapeutic day hospitals, of which 114 had PPD, were included in the study. Frequency distribution, chi(2), correlations, reliability statistics, exploratory, and confirmatory factor analyses were performed. RESULTS: The distribution of PPD criteria revealed no distinct boundary between patients with and without PPD. Diagnostic category membership was obtained in 37 of 64 theoretically possible ways. The PPD criteria formed a separate factor in a principal component analysis, whereas a confirmatory factor analysis indicated that the DSM-IV PPD construct consists of 2 separate dimensions as follows: suspiciousness and hostility. The reliability of the unitary PPD scale was only 0.70, probably partly due to the apparent 2-dimensionality of the construct. Persistent unwarranted doubts about the loyalty of friends had the highest diagnostic efficiency, whereas unwarranted accusations of infidelity of partner had particularly poor indicator properties. CONCLUSIONS: The reliability and validity of the unitary PPD construct may be questioned. The 2-dimensional PPD model should be further explored.
OBJECTIVE: This article examines reliability and validity aspects of the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) paranoid personality disorder (PPD) diagnosis. METHOD:Patients with personality disorders (n = 930) from the Norwegian network of psychotherapeutic day hospitals, of which 114 had PPD, were included in the study. Frequency distribution, chi(2), correlations, reliability statistics, exploratory, and confirmatory factor analyses were performed. RESULTS: The distribution of PPD criteria revealed no distinct boundary between patients with and without PPD. Diagnostic category membership was obtained in 37 of 64 theoretically possible ways. The PPD criteria formed a separate factor in a principal component analysis, whereas a confirmatory factor analysis indicated that the DSM-IV PPD construct consists of 2 separate dimensions as follows: suspiciousness and hostility. The reliability of the unitary PPD scale was only 0.70, probably partly due to the apparent 2-dimensionality of the construct. Persistent unwarranted doubts about the loyalty of friends had the highest diagnostic efficiency, whereas unwarranted accusations of infidelity of partner had particularly poor indicator properties. CONCLUSIONS: The reliability and validity of the unitary PPD construct may be questioned. The 2-dimensional PPD model should be further explored.