Objective: To investigate the prototype validity of the borderline personality disorder (BPD) DSM-IV construct.
Method: Patients (n = 930) from the Norwegian Network of Psychotherapeutic Day Hospitals. Exploratory/confirmatory factor analyses, correlation and reliability statistics, chi-square, and frequency distributions.
Results: Number of BPD criteria showed no distinct threshold between No-BPD and BPD patients among whom 136 different combinations of criteria occurred. Both factor analyses supported that one component/latent variable accounted for the variance of the BPD criteria, showing a high convergent and discriminant validity. The criterion of unstable relationships displayed the highest diagnostic efficiency, and that of chronic feelings of emptiness the lowest.
Conclusion: The prototype theoretical model for BPD fitted the data well and appeared to be satisfactory described by the current criteria. The emptiness criterion needs a more appropriate definition and the hierarchy of the criteria in DSM-IV should be revised.