Is dimensional scoring of borderline personality disorder important only for subthreshold levels of severity?

Studies comparing dimensional and categorical representations of personality disorders (PDs) have consistently found that PD dimensions are more reliable and valid. While comparisons of dimensional and categorical scoring approaches have consistently favored the dimension model, two reports from the Rhode Island Methods to Improve Diagnostic Assessment and Services (MIDAS) project have raised questions as to when dimensional scoring is important. In the first study, Asnaani, Chelminski, Young, and Zimmerman (2007) found that once the diagnostic threshold for borderline PD was reached the number of criteria met was not significantly associated with indices of psychosocial morbidity. In the second study, Zimmerman, Chelminski, Young, Dalrymple, and Martinez (2012) found that patients with 1 criterion of borderline PD had significantly more psychosocial morbidity than patients with 0 criteria. The findings of these two studies suggest that dimensional ratings of borderline PD may be more strongly associated with indicators of illness severity for patients who do not versus do meet the DSM-IV criteria for borderline PD. In this third report from the MIDAS project, we tested this hypothesis in a study of 3,069 psychiatric outpatients evaluated with semi-structured diagnostic interviews. In the patients without borderline PD the number of borderline features was significantly associated with each of 6 indicators of illness severity, whereas in the patients with borderline PD 3 of the 6 correlations were significant. The mean correlation between the number of borderline PD criteria and the indicators of illness severity was nearly three times higher in the patients without borderline PD than the patients with borderline PD (0.23 versus 0.08), and 4 of the 6 correlation coefficients were significantly higher in the patients without borderline PD. These findings suggest that dimensional scoring of borderline PD is more important for subthreshold levels of pathology and are less critical once a patient meets the diagnostic threshold. The implications of these findings for DSM-5 are discussed.