UReCA: The NCHC Journal of Undergraduate Research and Creative Activity 2020 Edition
The Proposal and Evaluation of the ICD-11 Classification System for Personality Disorders
Diagnostic criteria are present and necessary for all disorders, ranging from medical to psychiatric. The International Classification of Diseases, volume 11 (ICD-11) is the most recent edition of a diagnostic manual that is set to be implemented by 2022 (World Health Organization, 2018). One area of this manual that has been revised extensively is the classification of personality disorders (PD’s). A new dimensional model has been proposed, a vast difference from the existing categorical model. With such a dramatic change, researchers and clinicians have engaged in conversations regarding the efficacy and suitability of this model to sufficiently capture diagnosis of individual PD’s. Although there are still further questions that research needs to answer, existing studies address a myriad of concerns surrounding a model of this nature, and successfully provide evidence that a dimensional model is most appropriate to capture the diverse nature of PD’s.
Personality Disorders are reasonably prevalent, with 12% of the general population, 25% of primary care patients, and 50% of psychiatric care patients meeting criteria for at least one (Bach et al., 2017; Bach & First, 2018). Therefore, it is important for clinicians to have proper assessment tools and classification criteria to accurately diagnose patients. There is some consensus that the present categorical systems in the DSM-5 and ICD-10 are inadequate (Tyrer et al., 2014). With use of these systems, there is disproportional representation of borderline personality disorder (BPD) and antisocial personality disorder (ASPD) in published statistics, and many patients are diagnosed comorbid for at least two disorders or placed in the residual category. Clinicians also cite arbitrary thresholds and considerable overlap of symptoms as additional problems (Bach & First, 2018). The World Health Organization (WHO) assembled a committee to develop a new model for the classification of PD’s for the ICD-11 (Reed, 2018). Due to the release of this proposed model, many researchers are now conducting studies to determine its efficacy, compare it to previous models, and develop measures that may aid in diagnosis for clinicians. Literature over the past five years reveals a great deal of progress and supports that the ICD-11 dimensional model will be more descriptive and accurate for the diagnosis of PD’s.
The Proposed Model
The proposed model, hereafter referred to as the ICD-11 model, follows a hierarchical structure that is parallel to the diagnostic process (Bach et al., 2017). First, the clinician is to determine whether or not any PD is present in the patient by assessing general symptoms common to all PD’s. Second, the severity of the PD is assessed by measuring the extent of interpersonal dysfunction, occupational dysfunction, and risk of harm to self or others (Olajide et al., 2018; Tyrer et al., 2014). Third, trait domains are used to include additional relevant information to the clinician about the specific maladaptive behaviors of the patient (Bach & First, 2018). These trait domains line up with the five-factor model (FFM) that is generally accepted for personality organization, as well as traits defined in the DSM-5 model of PD’s (see Table 1). Additionally, bipolar pattern is offered as a qualifier in order to account for the symptomatology that may be present across multiple PD’s (Bach & First, 2018). Trait qualifiers and the bipolar pattern qualifier allow for clinicians to include more specific detail about patients, addressing the common concern that dimensional models do not include enough information about an individual’s particular issues. The proposed model will be presented in May 2019 and take effect by 2022 (World Health Organization, 2018).
Comparison to Previous Models
As discussions over the ICD-11 have continued, one area of interest is how this model compares to previous accepted models (e.g. the ICD-10 and DSM-5). Research has been conducted to compare diagnoses by the ICD-11 versus the ICD-10 and DSM-5 for three samples of patients (Tyrer et al., 2014). Patients were drawn from two existing studies involving psychiatric disorders as well as from one in-patient study. Measures were collected and scored to reflect classification by all three models. Prevalence of PD’s was 7% higher when utilizing ICD-11 criteria, which was an unexpected result. This may imply that the ICD-11 model is more sensitive than previous models. Importantly, this study provides evidence that patients currently diagnosed under previous models will likely not experience a difference in diagnosis upon changing to the ICD-11 model. This study effectively demonstrates these points, however the three samples did not use consistent measures (e.g. one sample used the PAS-Q to record severity of personality dysfunction, one sample used full ICD-10 and ICD-11 criteria, and the data from the remaining sample was recorded in the 80s, therefore current measures were not used). Despite these limitations, this study provides important comparisons in the application of the ICD-11 to previous accepted models.
The World Health Organization (WHO) and American Psychiatric Association (APA) agree that the DSM and ICD systems should be harmonized, providing the basis for an important line of research (Bach et al., 2017). The objective of this study was to examine the trait qualifiers in the ICD-11 and how they relate to the traits in the DSM-5 model. The researchers used the Personality Inventory for DSM-5 (PID-5) and conducted Exploratory Structural Equation Modeling (ESEM) to determine an ideal structure for the ICD-11 model. This measure was developed for the alternative model for PD’s in the DSM-5, which is a dimensional model that is provided in addition to the standard categorical model. Statistical analysis nicely captured the hierarchical nature of personality pathology (see Figure 1), and the resulting five factor model captured each of the trait qualifiers described by the ICD-11 model. Because this study utilizes a measure developed specifically for the DSM-5, it connects the ICD-11 and DSM-5 models nicely. In addition, the ESEM factors traits out of overall maladaptive personality in a very logical, easy to understand manner. These researchers did choose to use only 16 out of 25 subscales from the PID-5, implying that the alternative model from the DSM-5 may provide more detailed information for clinicians. A further limitation is that this study looked only at the trait qualifiers rather than including severity, which is a cornerstone component to the ICD-11 model.
Application of the ICD-11 Model
Seeing the need for published results of the application of the ICD-11 model, researchers have applied the classification criteria to a series of cases (Bach & First, 2018). This study addressed several concerns: first, clinicians and researchers alike have debated how many trait qualifiers should be present, citing five as too few to be specific; second, there has been discussion over the necessity of a model that may be implemented by clinicians without specialized training, which the ICD-11 is aiming to be; third, the topic of comorbidity has been cited as an issue with previous models—a problem that the ICD-11 model aims to solve. For each of the five cases presented in this paper, the presenting problem is described followed by classification by ICD-11 along with reasoning for choices of severity and trait qualifiers. Discussions over the five cases show that the ICD-11 model was well-suited for a variety of presenting problems. In particular, the researchers noted that the trait qualifiers on their own may be less telling than when one views them in conjunction with one another (e.g. negative affectivity paired with dissociality results in externalizing behavior while negative affect paired with dependency results in internalizing behavior). This paper addresses the important concern of having too few trait qualifiers to collect sufficient information to make decisions on treatment. In addition, the researchers conclude that although behavior patterns are unlikely to change, the level of severity of a PD (and therefore the behaviors leading to respective trait qualifiers) is able to change. Therefore, clinicians may want to focus on reducing the severity of the PD and work toward developing adaptive coping strategies for the patient rather than trying to reduce specific traits. This study brings the hypothetical model to life in its application to real patients, which allows readers to evaluate its usefulness and simplicity.
Developing Measures for the ICD-11 Model
With the development of a new model, new measures must be validated to assess patients according to new criteria. In line with this necessity, a study was conducted to validate a measure for severity of PD, the Standardized Assessment of Severity of Personality Disorder (SASPD; Olajide et al., 2018). This measure was created based off the Standardized Assessment of Personality—Abbreviated Scale (SAPAS; Moran et al., 2003), which measures presence or absence of a PD, but not its severity. The SASPD modifies the items of the SAPAS to ask about the extent of the issue rather than its existence. The objective of this study was to determine thresholds for mild, moderate, or severe PD’s. The sample did not contain many patients with severe PD’s, therefore a threshold for severe PD was unable to be determined. For mild and moderate PD’s, however, the SASPD was accurate 80% of the time. One benefit of this measure is that it is self-report, further building on the objective of the ICD-11 model to be easier to use by clinicians without specialized training. In addition, the validation of a measure of severity is important to this model as there are few, if any, measures available to assess this dimension of PD. The validation is incomplete, however, as there is not a cut-off point established for severe PD given the lack of variety in the sample. With further research, this measure will allow clinicians to easily assess the severity of a patient’s PD, filling an important gap in the development of the ICD-11 model.
Along the same vein of research, another study was conducted a study to validate three measures: the Personality Inventory for ICD-11 (PiCD), the Standardized Assessment of Severity of Personality Disorder (SASPD; originally developed by Olajide et al., 2018), and the Borderline Pattern Scale (BPS; Oltmanns and Widiger, 2019). Both the PiCD and BPS had good internal consistency and convergent validity to existing measures. Validating the SASPD proved to be more challenging as current measures may not exhibit good discriminant validity, and the current measures do not only assess for severity. This study provided good initial validation for three new measures that would correspond to each of the three steps of diagnosis of a PD with the ICD-11 model. This paper also evaluated the ICD-11 model from the perspective of the accepted five-factor model (FFM) for personality (see Table 2). These researchers addressed two major points: evaluation of the new ICD-11 model as a whole and validation of three new corresponding diagnostic measures. While connected, these two points may have been more successful broken into separate papers. This paper reads as a literature review followed by a study, resulting in a scope that is far-reaching while attempting to maintain specificity. Despite this, the validation of diagnostic measures is incredibly necessary to allow for clinicians to make accurate diagnoses using the ICD-11 model. In addition, the description of the model was quite detailed and the comparison to existing models was useful for readers to understand the proposed changes.
Scientific discourse surrounding the development of the ICD-11 classification model for PD’s is incredibly extensive. Studies generally fall into one of three categories: application of the model, correspondence to existing models, and validation of measures that correspond to the new diagnostic criteria. Each of these categories lends important information to the community by addressing different concerns over classification systems for PD’s in general. For example, comparing and contrasting the ICD-11 model to previous established models demonstrates the improvements made in response to concerns raised by clinicians. Moreover, published studies continue to provoke further conversation about how to develop improved models that will allow for clinicians to treat patients to the best of their ability. There will always be controversy over how diagnostic criteria will be defined, but the ICD-11 model has received positive feedback and demonstrated its usefulness in a variety of the studies described here.
One concern over the ICD-11 model is that five trait qualifiers does not allow for enough specificity (Bach & First, 2018). ESEM was used to test models up to five trait qualifiers, however models containing more than five factors were not constructed to examine whether subsequent models would be appropriate (Bach et al., 2017). Specifically, some clinicians cite the psychoticism domain of the DSM-5 model as an area that is lacking in the ICD-11 model. The ICD lists psychotic-like symptoms under schizophrenia, rather than including a schizotypal PD similar to that of the DSM-5 (Oltmanns & Widiger, 2019). Empirical evidence supporting the choice to include psychotic symptoms elsewhere should be provided to address this concern.
Additionally, the development of validated measures to aid in diagnosis is still in progress. Several measures have been presented (i.e. the PiCD, BPS, and SASPD), but further evidence is needed to provide sufficient validation. The SASPD in particular has proven difficult to validate against current measures (see Oltmanns & Widiger, 2019). This measure in particular is crucial to the ICD-11 model, as a measure of severity is one of the key components that makes this model different than its predecessors. The introduction of validated, simple measures could do significant work to address any concerns stemming from the implementation of the ICD-11 model given the lack of available measures.
Future directions in research should aim to address the previously mentioned limitations of these studies. Additional published research demonstrating the accuracy of a five-factor trait qualifier model may provide further reassurance that five domains is sufficient to capture individual patient information. Presentation and validation of new and existing measures for the ICD-11 model will also help smooth the implementation process. A measure for severity—be it the SASPD or another measure—is crucial to this model. Future research may consider looking to measures of severity for other disorders to draw into the realm of PD’s. Finally, more published studies demonstrating the application of the ICD-11 model on actual patients (such as Bach & First, 2017) will further prove the model’s suitability for assessing PD’s.