Revising the diagnosis of personality disorder: Can it be single,clinical and factorial?

The debate discussed by Peter Tyrer (2017) around classification of personality disorder is important as there is a lot at stake for patients and the field of study if we move to a classification system that does not help people get the treatment they need or impedes research progress.

The introduction of a systematic approach to the diagnosis of personality disorder in 1980 with the Diagnostic and Statistical Manual of Mental Disorders (3rd ed.; DSM-III) has spurred a huge growth in research. Yet, the vast bulk of studies are on borderline, and antisocial personality disorder categories, because they are clinically meaningful, describe well-recognised cohorts, but importantly have different treatment implications. We have, however, some ‘dead’ diagnoses, not used clinically or researched, such as histrionic personality disorder. The field needs a refresh.

What then about the draft International Classification of Disease (11th rev.; ICD-11) personality disorder classification? The ICD-11 proposal for a single classification of ‘personality disorder’ has merit. Most clinicians agree that the 10 categories in Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV) are obsolete because individual patients are complex combinations of categories and there is too much overlap between them. The current system creates high numbers of artificial comorbid personality disorders in patients because of this fundamental flaw in the categories. The ICD-11 has removed this problem of Personality Disorder (PD) comorbidity.

Significantly, there is increasing evidence for an underlying ‘general’ factor beneath all personality categories. Some research suggests that this general factor probably overlaps most with the criteria of borderline personality disorder (BPD; Sharp et al., 2015). Having a single overarching ‘general’ disorder also fits the conceptualisation of ‘borderline personality organisation’ found in the Psychodynamic Diagnostic Manual (2nd ed.; PDM-2), emphasising the core deficits in identity fragility and interpersonal dysfunction. What is clear is that personality disorder is a discrete disorder. Calls to reconceptualise the disorder as a complex traumatic stress disorder are inappropriate, because not all clients have trauma, and trauma therapies are not recommended for front-line treatment of personality disorder (Lewis and Grenyer, 2009).

The ICD-11 proposal is to also define personality disorder dimensionally by severity (mild, moderate, severe). Most clinicians and researchers recognise that severity information has useful implications for treatment planning, as well as describing how the patient is actually functioning in the community. In addition, severe patients may even induce metacognitive deficits in the treating practitioner’s capacity to think clearly (Bourke and Grenyer, 2017).

However, for a classification to be useful, it importantly must do three things: first identify a group of patients accurately (is there presence of personality disorder?), second be able to describe individual differences to personalise treatment (what kind of personality disorder traits?) and third provide severity information to assist care planning (is the patient mild, moderate or severe?). ICD-11 will do all three. Traditional DSM ‘BPD’ meets only the first and partially the second of these. The majority of people with BPD find the diagnosis meaningful to their phenomenological experience. There are now more than 33 randomised controlled trials of BPD treatment efficacy and reasonable prospects of clinical improvement. However, BPD itself is heterogeneous, as there are 256 combinations of the 9 DSM criteria meeting the disorder specifier, and there are at least three underlying factors along emotion (affect dysregulation), interpersonal (rejection sensitivity) and cognitive (mentalization failure) dimensions (Lewis et al., 2012). We thus need better ways to understand the individual patient that allow such features to be described.

Because the heterogeneity of personality disorder is well recognised, ICD-11 will allow the diagnosis of personality disorder to be personalised by allowing the rating of five ‘domain traits’. We now know these are highly similar to the ‘alternative’ DSM-5 traits, with the ICD/DSM concordance, respectively, being Dissociality/Antagonism, Detachment/Detachment, Disinhibition/Disinhibition, Negative Affectivity/Negative Affectivity, but with two mismatches – ICD-11 has the Anankastia (obsessive–compulsive) and DSM-5 has Psychoticism traits. Unfortunately, ICD-11 misses the opportunity to describe psychotic and schizotypal traits in personality disorder, despite them being recognised since Adolf Stern in 1938 described patients being on the ‘border line’ between neurosis and psychosis, and these being added as the ninth criterion of BPD in 1994. DSM-5 also misses the opportunity to fully describe obsessive–compulsive personalities despite their high prevalence, but rather this group gets described although aspects of perfectionism and perseveration are in the low disinhibition and negative affectivity domains.

Significantly, because ‘borderline’ is a personality disorder diagnosis recognised by health insurers across the world, who reimburse evidence-based BPD treatments, the ICD-11 committee have now agreed to allow a ‘borderline’ qualifier to be assigned once the general criteria for personality disorder diagnosis is met. This is good news for patients. The challenge for clinicians and researchers will be to reconceptualise how other prominent categories lost in translation, such as narcissistic personality disorder, can be understood within the new ICD-11 world of a single personality disorder, clinical severity rating and five factorial trait domains.

See Editorial by Tyrer. 51: 1077–1078.