A comparison of the DSM-5 Section II and Section III personality disorder structures

Abstract

The DSM-5 Section III includes a hybrid model for the diagnosis of personality disorders, in which sets of dimensional personality trait facets are configured into personality disorder types. These PD types resemble the Section II categorical counterparts with dimensional traits descriptive of the Section II criteria to maintain continuity across the diagnostic systems. The current study sought to evaluate the continuity across the Section II and III models of personality disorders. This sample consisted of 397 undergraduate students, administered the Personality Inventory for the DSM-5 (Krueger et al., 2012) and the Structured Clinical Interview for the DSM-IV Axis II Disorders–Personality Questionnaire (First et al., 2013). We examined whether the DSM-5 Section III trait facets for the PDs would be associated with their respective Section II counterparts, as well as determining whether additional facets could augment the prediction of the Section II disorders. Results revealed that, generally, the DSM-5 Section II disorders were most strongly associated with their Section III traits. Results also showed evidence to support the addition of facets not included in the Section III diagnostic criteria in the prediction of most disorders. These results show general support for the Section III model of personality disorders, however, results also show that additional research is needed to replicate these findings.

Introduction

The Diagnostic and Statistical Manual of Mental Disorders, 5th edition (DSM-5) retained the 10 DSM-IV-TR (American Psychiatric Association, 2000) personality disorders for Section II (Diagnostic Criteria and Codes) of the manual, which is used for formal diagnosis. Since the DSM-III, however, the current diagnostic system for personality pathology has been met with substantial criticism (e.g. Clark, 2007, Clemence et al., 2009, Vinnars and Barber, 2008, Watson et al., 2008, Widiger and Mullins-Sweatt, 2010). Researchers have called for a model that is aligned with the empirical literature and include solutions to problems such as excessive diagnostic comorbidity, heterogeneity within diagnoses, temporal instability of the disorders, lack of variance in the model, and poor convergent and discriminant validity across personality disorders (see e.g., Clark, 2007, Skodol et al., 2011, Widiger and Trull, 2007, for reviews). As such, the time is ripe for substantial changes to this imperfect system.

The DSM-5 Personality and Personality Disorder Workgroup proposed an alternative system for the diagnosis of personality disorders. The APA Board of Trustees included both models in the DSM-5 “to preserve continuity with current clinical practice, while also introducing a new approach that aims to address numerous shortcomings of the current approach to personality disorders” (pg. 811; APA, 2013). However, the model was rejected as the primary model for diagnosing personality disorders, and was, instead, placed in Section III for “emerging models and measures.” The DSM-5 Section III model indicates several significant revisions to address the criticisms to previous versions of the DSM, which result in a system that is more consistent with the personality pathology literature (APA, 2013; Skodol, 2012). This model was a product of a significant amount of research that consistently identifies four to five broad trait domains existent in personality psychopathology (De Clercq et al., 2006, Harkness and McNulty, 1994, Krueger et al., 2011, Livesley et al., 1998, Tackett et al., 2008).

The two-prong model is a hybrid dimensional and categorical system, which employs functional impairment criteria and dimensional personality traits mapping onto one of six categorical personality disorders, with a seventh diagnosis (Personality Disorder-Trait Specified; PD-TS) to replace the former Personality Disorder NOS. The first criterion (Criterion A) considers impairment in the domains of self (identity or self-direction) and interpersonal (empathy or intimacy) functioning. The second criterion (Criterion B) dictates that an individual must also exhibit maladaptive personality traits based on a model of five dimensional personality domains and their accompanying set of three to seven facets. These trait domains include Antagonism, Psychoticism, Disinhibition, Negative Affectivity, and Detachment (American Psychiatric Association, 2011, Skodol et al., 2011), and are grounded in literature showing the empirical validity of dimensional models for maladaptive personality functioning (Harkness and McNulty, 1994, Krueger et al., 2011, Samuel and Widiger, 2008, Watson et al., 1994, Widiger and Simonsen, 2005; among others). This model has also shown strong associations with other models of personality such as the Personality Psychopathology Five (PSY-5) model (Anderson et al., 2013; see also Harkness et al., 2012), and the Five Factor Model (FFM) (Gore and Widiger, 2013; Thomas et al., 2013; Widiger et al., 2013). For instance, Negative Affectivity aligns well with the PSY-5 domain of Negative Emotionality/Neuroticism and the FFM domain of Neuroticism. This convergence shows evidence that the DSM-5 Section III personality model is reflective of other empirically driven personality models. Additionally, the association between the Section III model and the FFM is significant given that the FFM is generally conceived as a model of “normal” personality. The Section III model was, in part, designed in order to reflect the extreme range of these normal personality domains (Krueger et al., 2012) and, therefore, these empirical associations establish the important relationship between normal and pathological personality traits.

Furthermore, per the DSM-5 Section III model, an individual may be assigned to a particular personality disorder (PD) category which can include Antisocial PD, Borderline PD, Narcissistic PD, Avoidant PD, Schizotypal PD, Obsessive Compulsive PD, and PD: Trait Specified. These personality disorder categories were included in the model, in part, to maintain continuity between the DSM-IV-TR personality disorder model and the DSM-5 Section III model. Each personality disorder has its own unique configuration of personality traits based on the DSM-5 Section III facet structure. For example, Borderline PD is operationalized via the facets Impulsiveness, Risk Taking, Emotional Lability, Anxiousness, Separation Insecurity, Hostility, and Depressivity. On the other hand, if an individual shows impairment in self and interpersonal functioning, but does not meet any specific configuration of traits for any diagnosis, that individual would be assigned a diagnosis of PD: Trait Specified.

Preliminary research has found initial support for the Criterion B trait-based model of personality disorders. Wright et al. (2012) found empirical support for the hierarchical structure of the personality trait model for DSM-5 using a very large (N=2461) sample of undergraduate students. Additionally, Miller et al. (2012) found support specifically for the validity of the DSM-5 in capturing Borderline PD personality traits in community participants, half of whom were receiving psychiatric treatment. Furthermore, Berghuis et al. (2012) found support for the DSM-5 Section III׳s trait model assessment of personality dysfunction. Finally, Hopwood et al. (2012) sought to evaluate the degree to which the DSM-5 Section III model maintained continuity with the DSM-IV-TR model in a large undergraduate sample (N=808). They examined how the DSM-5 Section III traits were associated with the DSM-IV-TR personality disorder. They found significant associations between the DSM-IV-TR and DSM-5 personality disorders, thus exhibiting continuity between these models. They also found that, generally, the facets for each personality disorder in the DSM-5 adequately described the disorder to which they were assigned and were the most influential predictors thereof. Hopwood et al. did not, however, examine additional configurations of trait facets or the predictive and discriminate validity of the constructs. They also did not examine the semi-partial correlations associated with these trait facets and whether the Section III facets contributed uniquely to the prediction of DSM-IV PDs. These issues are important given that this model has the potential to replace the current problematic system featured in Section II. Additionally, regardless of any limitations of the Hopwood et al. study, a greater amount of research is necessary before this model is fit to be a replacement of the current model, including in a variety of settings and geographic areas, particularly given that diversity was a focus of the field trials when this model was designed.

Also important to note is that the DSM has never included two separate ways to derive the same diagnosis previously to the fifth edition. Given that the DSM-5 now has two methods to diagnose personality disorders, with Section III posed as an “emerging model” that will potentially be included in later versions of the manual, it becomes especially important to evaluate the continuity between these models. Indeed, Section III is presumably the path to DSM-5 Section II diagnoses. Although Section III will not be used to formally diagnose individuals with personality pathology at this point, it would nonetheless be problematic if the DSM-5 includes contradictory methods to derive at one disorder.

The current study sought to examine the association between the DSM-5 Section II and Section III personality disorders as measured by Criterion B of the model. Specifically, this study aimed to examine the associations these two sections of the DSM-5 with regard to the continuity between the models. We sought to examine the specific facet profiles of the DSM-5 Section III personality disorder model. More specifically, we aimed to determine whether the traits included for each disorder were related to their DSM-5 Section II counterparts. Similarly, we aimed to examine whether additional facets might warrant consideration in understanding each DSM-5 Section II disorder. In addition, although Section III includes only six PD׳s, we examined all 10 of the Section II PD׳s in order to determine the associations between the Section III PD model and the Section II personality disorders in their entirety by utilizing a previously proposed “cross-walk” for the assessment of the four remaining PD׳s using the Section III domain and facet system.¹