OBJECTIVE

Little is known about the fundamental structure of core personality disorder psychopathology in the general population. The current study employed confirmatory factor analysis to evaluate competing models of patterns of personality disorder diagnoses in a nationally representative sample.

METHOD

DSM-IV and alternate models of the structure of personality disorder psychopathology were evaluated using data from the National Epidemiologic Survey on Alcohol and Related Conditions conducted between 2001 and 2002 (N = 43,093). Dimensional versus categorical representations of DSM-IV personality disorder structure were also tested. Face-to-face interviews were conducted in the United States. Participants were community-based respondents aged 18 years and older. Diagnoses and dimensional scores were made for antisocial, avoidant, dependent, histrionic, obsessive-compulsive, paranoid, and schizoid personality disorders.

RESULTS

Multiple goodness-of-fit indicators provided support for a DSM-IV-based hierarchical model of personality disorders. In this model, the individual personality disorders were viewed as belonging to 1 of 3 latent factors or clusters (A, B, or C). In all of the models, the individual personality disorders were allowed to be an indicator for only a single latent cluster, and errors were not allowed to correlate with each other. In turn, these 3 clusters were viewed as comprising a single higher-order "Axis II personality disorder factor." The DSM-IV model was largely invariant across gender, Axis I comorbidity, and treatment-seeking status. A dimensionally based form of assessment of the DSM-IV personality disorders produced excellent goodness-of-fit indicators and produced low Akaike information criterion values (which are indicative of better-fitting models).

CONCLUSIONS

The results from this confirmatory factor analysis in a large, nationally representative mental health survey supported the DSM-IV hierarchical organization of Axis II personality disorders. This model was significantly superior to viable alternative models of Axis II personality psychopathology. There was also evidence to suggest this model could obtain even stronger support if a dimensionally based form of diagnostic assessment was adopted in place of the dichotomous form of assessment (presence/absence) of personality disorders currently in use in the DSM-IV.