Severity

All categories of personality disorders were abolished. In their place there was a general description of a personality disorder conceptualised along a dimension of severity. To qualify for a diagnosis of a personality disorder, an individual needed to have problems with functioning of aspects of the self and/or interpersonal dysfunction manifest in various patterns of emotional expression and maladaptive behaviour across a range of situations (see World Health Organization (WHO)^{Reference World Health Organization1} for a full definition). The diagnosis is then further specified as ‘mild’, ‘moderate’ or ‘severe’. Assessment of severity is based on the prominence of abnormal traits and their impact on the individual's social and occupational functioning, as well as the risk they pose to themselves and others (see WHO^{Reference World Health Organization1} for full definitions of mild, moderate and severe).

Personality difficulty

Classifying personality function on a single dimension from normality to severe implies that some individuals have personality dysfunction but do not qualify for a diagnosis of personality disorder. The term ‘personality difficulty’ was introduced; a subsyndromal condition related to ‘problems associated with interpersonal interactions’.^{Reference Mulder and Tyrer2} Table 1 outlines the differences between personality difficulty and personality disorder.

Table 1

Personality difficulty versus personality disorder

Reproduction of Table 9.1 Differences between personality difficulty and personality disorder.^{Reference Bach and Mulder3}

This new term is largely untested but the condition appears very common (up to half the population), and results in more healthcare seeking and increased distress.^{Reference Mulder and Tyrer2} There are concerns that the term may over-medicalise difficult behaviour or become reified as a diagnosis, but others have argued it will help in understanding the concept of a personality spectrum and reduce stigma when it is realised how common personality dysfunction is, and is not confined to a few deviant individuals.

Measuring severity

A number of measures of ICD-11 personality disorder severity have been developed. These include the Standardised Assessment of Severity of Personality Disorders (SASPD), Preliminary Scales for ICD-11 Personality Disorder and Personality Disorder Severity ICD-11 (PDS-ICD-11). The latter has been tested and seems to capture a single dimension of personality dysfunction and is convergent with other established measures of personality functioning.^{Reference Bach and Mulder3}

Trait domain qualifiers

While the shift to an emphasis on personality disorder severity was radical, it fitted with most researchers’ and clinicians’ views of the available evidence (a similar proposal was advocated by the DSM-5 classification committee⁴). However, there was less agreement about describing personality pathology. Here, those with a special interest in personality disorders believed that individual categories, such as narcissistic, borderline and antisocial, had clinical meaning and were difficult to translate into dimensional variation. DSM-5 had tried to keep some categories alongside a dimensional description of personality traits, leading to a complex classification that was eventually rejected. ICD-11 opted to abolish all categories and describe personality pathology across five dimensions.

The derivation of these dimensions attempted to be as empirical as possible but data was limited. Most studies of personality pathology used categories, or more commonly, chose one category and ignored the others. However, a number of researchers had attempted to study the underlying structure of categorical personality disorder symptoms using a variety of methodologies, most commonly factor analysis. Despite different samples, various means used to assess personality and different models of personality pathology, the results were surprisingly consistent. A systematic review^{Reference Mulder, Newton-Howes, Crawford and Tyrer5} reported three to five domains; all studies had a general personality pathology dimension, as predicted, which divided into internalising and externalising domains. Nearly all reported a schizoid/aloof domain and most a compulsivity domain. These four factors were not only reported reasonably consistently across most studies, but also they had good face validity and appeared linked to normal personality domains, such as the Big Five.

Attempted validation studies reported that the externalising domain encompassed a wide range of behaviours, including antisocial behaviour, psychopathy and impulsivity, and performed relatively poorly. After much debate, a factor incorporating non-psychopathic externalising behaviours – disinhibition – was introduced for further study. ICD-11 therefore has five broad descriptions of personality pathology called ‘trait domain qualifiers’. These are not categories but rather qualifiers used to describe the most prominent characteristics that contribute to personality disturbance. As many domains as necessary can be used to describe personality functioning. The more severe the personality disorder the more domains tend to be involved.

A brief description is given below.^{Reference Mulder and Tyrer2}

1. (a) Negative affectivity – a tendency to experience a broad range of negative emotions with a frequency and intensity out of proportion to the situation, which may include emotional lability, negative attitudes and low self-esteem.

2. (b) Dissociality – a disregard for the rights and feelings of others, encompassing both self-centredness and lack of empathy.

3. (c) Detachment – a tendency to maintain interpersonal and emotional detachment from others.

4. (d) Disinhibition – a tendency to act rashly based on immediate external and internal stimuli without consideration of potential negative consequences.

5. (e) Anankastia – a tendency towards perfectionism and orderliness and emotional and behavioural constraint.

Measuring trait domains

One problem with such a radical change in classification is that previous studies cannot be directly translated into the new system. This was compounded by the fact that ICD-11 was not accompanied by a measure to formally operationalise the classification. However, four of the five trait domains were shared with the DSM-5 Alternative Model for Personality Disorders (AMPD), the exception being anankastia in ICD-11 versus psychoticism in the AMPD. Bach et al^{Reference Bach, Sellbom, Kongerslev, Simonsen, Krueger and Mulder6} developed an algorithm for delineating the five ICD-11 trait domains (including anankastia) using the well-established Personality Inventory for DSM-5 (PID-5). They were then able to show that the five-factor structure was valid across US and Danish samples, as well as psychiatric patients. In addition, the domain demonstrated expected associations with personality disorder categories.

The Personality Inventory for ICD-11 (PiCD) was introduced in 2018, consisting of 60 items used to calculate the five ICD-11 domain scores. The PiCD has been evaluated in a number of studies supporting the validation of the five trait domains. Importantly, the ICD-11 domains showed meaningful and expected relationships with normal personality dimensions measured using the Big Five model. Negative affectivity correlated with neuroticism, detachment with low extraversion, dissociality with low agreeableness, anankastia with high conscientiousness and disinhibition with low conscientiousness.^{Reference Bach and Mulder3}

Also, taking advantage of the similarity between the ICD-11 and DSM-5 AMPD personality domain measures, a 34 item Personality Inventory for DSM-5 Brief Form Plus (PID5BF+), which aims to capture both ICD-11 and DSM-5 trait domains, was proposed. This was modified to comprise 36 items and appears robust across different population samples. A brief clinical interview is being developed that aims to include ICD-11 personality disorder severity as well as the five domains.^{Reference Bach and Mulder3}

Borderline pattern qualifier

The borderline pattern qualifier was introduced as a pragmatic solution to appease clinicians and researchers specialising in the field of personality disorders. They were concerned that losing this diagnosis would have substantial effects on research funding and treatment provision. While, as they pointed out, borderline personality disorder is the most studied personality disorder pathology in relation to treatment, the research essentially tells us the host of treatments are similarly effective and none has shown a specific efficacy for borderline personality disorder as opposed to general psychological distress and dysfunction. In addition, factor analytic studies have failed to support a distinct borderline personality disorder domain^{Reference Sharp7} and its features are more symptoms than personality traits. Nevertheless, allowing clinicians to specify a ‘borderline pattern qualifier’ (not a diagnosis) was felt necessary until there was sufficient data showing that its pathology could be accounted for by the ICD-11 severity and domain trait model.^{Reference Tyrer, Mulder, Kim and Crawford8}

Implications for treatment: clinical utility

On the face of it, the ICD-11 classification model seems more ‘true’ to existing evidence about personality pathology. As we have noted, the classification has been operationalised and appears to have robust construct validity and predictable convergence with other personality measures, and demonstrates reliability cross-culturally. However, while construct validity is an essential requirement, it is not sufficient. The most important consequence of a paradigm shift in diagnostic models is to aid the development and evaluation of treatments. A number of frameworks have been proposed that suggest careful assessment of severity and trait domains can lead to a coherent and holistic formulation that can be shared with the patient and a consensual approach to treatment adopted.^{Reference Mulder, Bach and Huprich9}

The general dimension of personality severity may be a good target for intervention and monitoring efficacy and a better way of measuring progress than specific personality features, which are more stable. Clinicians appear to value the level of personality function more than specific personality disorder categories when formulating treatments and discussing them with patients. Severity is also associated with treatment alliance and risk of drop-out as well as boundary confusion and increased negative countertransference.^{Reference Mulder, Bach and Huprich9}

Similarly, assessment of domain traits encourages a collaborative therapeutic approach: helping patients identify their own traits and how they are demonstrated in everyday life and conveying the idea that traits can be changed into something more adaptive.^{Reference Mulder, Bach and Huprich9} Acknowledging the adaptive significance of traits when the context is considered may be important. For example, a patient with detachment may be less emotionally responsive, which is problematic in some social situations but may be useful when cool-headed, self-absorbed behaviour is called for. The general aim of treatment is not to transform individual trait domains but to encourage adaptation and, to some degree, acceptance.

In summary, it might be expected for an author of a new classification to promote it, so any praise should be seen in that context. However, the evidence to date suggests that the new ICD-11 personality disorder construct is preferred by clinicians, is understood by patients and may be a spur to better and more sophisticated management of personality pathology.