The time has come for dimensional personality disorder diagnosis

The committee revising the ICD-11 Mental or Behavioural Disorders section ‘Personality Disorders and Related Traits’ has proposed replacing categorical personality disorders with a severity gradient ranging from personality difficulties to severe personality disorder and five trait domains: negative affectivity, dissocial, disinhibition, anankastic and detachment ¹. While acknowledging that there are multiple potential pathways for moving toward a more evidence-based and clinically useful scheme for classifying personality dysfunction, we applaud and support the proposed transition from a categorical model of personality disorder types, which has proven to be empirically problematic and of limited clinical utility, to a dimensional model of personality disorder that has considerable connection to scientific evidence and potential for clinical application.

There is no evidence supporting the hypotheses that personality disorders are categorical ^2,3 or that there are 10 (or any other number of) discrete types of personality disorder ⁴. Well-established problems with categorical personality disorder diagnosis such as low reliability, diagnostic comorbidity and within-disorder heterogeneity complicate research and treatment ⁵. There are no validated interventions for most of the categorical personality disorders, and although several psychotherapies from different theoretical perspectives have been developed for borderline personality disorder that have evidence of moderate efficacy, none have proven to be relatively more effective than any of the others ⁶. Evidence for treatment mechanisms is sparse, and there is no evidence that existing approaches have specific efficacy for borderline personality disorder as opposed to general efficacy for a variety of psychiatric difficulties ⁷.

In contrast, there is a vast body of empirical literature supporting dimensional models of personality disorder that are closely aligned with the proposed model ^8–11, in addition to the emerging body of work on the specific dimensions proposed for ICD-11 ^12–19. The ICD-11 proposal has two elements. The severity dimension has ties to the psychodynamic tradition ^20,21, which has historically been at the forefront of personality disorder classification, and aligns with a number of empirical efforts to quantify general personality dysfunction (e.g. ^22–27). Research demonstrates that much of the predictive and prognostic value in personality disorder data can be derived from such a dimension ²⁸.

The personality trait model proposed for ICD-11 resembles other dimensional models of personality such as the Five-Factor Model or the DSM-5 Alternative Model for Personality Disorders ^16,29,30. Although there are some important differences between the ICD-11 proposal and these other models that will be adjudicated by future research, the more important point at this stage is that evidence consistently supports the validity of dimensional trait models for describing individual differences in personality. In contrast to the categorical model of personality disorder types, there is a large literature on the genetic underpinnings, cross-cultural validity, course, correlates and measurement of broad personality traits ^11,31. Dimensional models also address issues such as comorbidity and heterogeneity in a direct and empirically tractable manner ⁸; recapture but empirically reorganize the information provided by personality disorder types ³²; and have considerable potential for guiding and tracking treatment ^33,34. We would highlight that research has repeatedly shown that the borderline personality disorder construct in particular can be accounted for by empirically derived dimensions of personality traits and functioning ^35–40.

Nevertheless, some people in the field continue to argue in favour of personality diagnosis by categorical types. We are concerned about the implications of retaining a categorical system that has been so thoroughly shown to be empirically and clinically problematic. It is very difficult to justify allocating resources toward continued research on an approach that has proven to be fundamentally flawed, as opposed to a dimensional model that points to exciting new avenues for research on aetiology, mechanisms and treatment (e.g. ⁴¹). We are likewise concerned about the implications that retaining a demonstrably problematic model has for patients’ lives. It would be very unsettling to be told that one’s problems are due to a specific medical condition, only to learn later that the supposed condition had been abandoned by the medical community. It is probably already confusing for patients, who might discover via an internet search on their personality diagnosis that much of the field does not believe such a disorder actually exists. It would be far preferable to be straightforward with our patients about what we know and do not know regarding personality and its related problems than to label them with legacy diagnoses that will not stand the test of time.

Reasonable concerns have been expressed about challenges associated with the transition from a categorical to a dimensional model of personality disorder. Such concerns need to be balanced against several field surveys that show that a majority of clinicians and researchers support the transition to a more dimensional, evidence-based framework ^42–44. We acknowledge that the transition to a dimensional model needs to be thoughtful with regard to issues such as third-party reimbursement. Moreover, we recognize that legal, community mental health and other systems will need to be educated regarding how to translate from the old system to the new. However, we do not believe that these practical issues provide a compelling rationale for retaining a system that does not effectively capture individual differences in patients’ personality difficulties. In contrast, moving forward with an evidence-based framework for diagnosing personality disorders has significant potential to stimulate research that can lead to new treatments and aetiological models that will ultimately reduce the burden of personality disorders on patients, families and society. The changes proposed for ICD-11 also provide a generative model for conceptualizing the meta-structure of psychopathology. Indeed, there are clear phenotypic and genetic links between the dimensions proposed for ICD-11 and a number of mental health conditions beyond personality disorders ^45,46.

Past scientists believed that the sun revolved around the earth, the brain was organized according to the principles of phrenology, and spirits were responsible for psychiatric problems. It is a testament to science that these views gave way to a more accurate model of nature. The new perspectives that replaced them contributed to major advancements in astronomy, neuroscience and mental health. Likewise, the evidence is clear that personality disorders do not exist as 10 discrete types. The categorical model has become a hindrance to research and practice. As an example, see the unfortunate outcome of the DSM-5 revision process, in which a model that is not supported by evidence or the majority of the field was retained as the official diagnostic scheme despite the viable alternative proposed by the Personality and Personality Disorders Work Group, published in Section III of DSM-5. It is time for the field to transition to a model that fits research data and clinical reality. The ICD-11 proposal connects psychiatric classification of personality disorder manifestations with scientific evidence. The proposed changes would enhance diagnostic efficiency and patient care while spurring research that can further improve the assessment and treatment of psychopathology. As clinicians and researchers who have dedicated our careers to understanding and helping people with personality pathology, we urge the ICD-11 PD work group to remain committed to an evidence-based revision of personality disorder diagnosis.