This rejoinder points out that the psychometric properties of trait domains have, at least at present, limited clinical relevance to treatment choice for the most severe personality disorders. Since patients with PDs are often misunderstood and mistreated, it is important to retain classifications that usefully guide treatment methods. The author also points out that the ICD-11 system for classifying personality disorders remains provisional, and does not have enough research behind it to solve the complex problems of PD classification. Since borderline personality disorder as a category can guide the choice of effective therapy, it is fortunate that it has been retained in ICD-11.

Psychiatry faces the problem of multiple systems of classifying personality disorders, reflecting longstanding controversies as to whether dimensions or categories are the best way of understanding these conditions. The current options are a continuation of the use of categories, a purely dimensional system, or a hybrid model. Most PD research has focused on either borderline personality or antisocial personality. These categories describe severe psychopathology, and do not readily fit into a dimensional model. As it stands, research has focused on the diagnoses with those that are most associated with psychopathology and dysfunction. The choice among current alternatives is unclear. While the DSM-5 Section III alternative system attempts to draw on the best aspects of all models, it may be too complex for clinical use. Some are concerned that a more “scientific” but less user-friendly system could interfere with access to evidence-based treatments for BPD.