Abstract
Purpose:
Assess the new alternative Diagnostic and Statistical Manual of Mental Disorders, fifth edition (DSM-5) model for personality disorders (PDs) as it is seen by its creators and critics.
Method:
Follow the DSM revision process by monitoring the American Psychiatric Association website and the publication of pertinent journal articles.
Results:
The DSM-5 PD Work Group’s proposal was not included in the main diagnostic section of the new DSM, but it was published in the section devoted to emerging models. The alternative DSM-5 PD constructs are radically different from those found in DSM, fourth edition, text revision.
Discussion:
There are some positive conceptual changes in the new model, but reliability and validity are not generally improved. However, social workers may be able to benefit from the use of the personality trait domains/facets of the alternative model.
Introduction
Used widely by social workers, psychiatrists, psychologists, and other mental health professionals, personality disorder (PD) diagnoses came to the forefront in the third edition of the Diagnostic and Statistical Manual of Mental Disorders, third edition (DSM-III) with their relocation to the newly introduced Axis II. The three DSM versions that followed made many changes to the PD criteria, with some being major and some being minor. For example, changing some PD criterion sets from monothetic—that is, requiring that all criteria be met—to polythetic—that is, requiring that only some criteria be met—was a major change. More minor changes included rewording specific PD criteria. Even the most minor of these changes had a powerful impact on the prevalence of PD diagnosis in clinical settings. For example, the prevalence of schizotypal PD based on DSM-III was 11%, but that prevalence dropped to 1% with the DSM-III-R revision of increasing the minimum required criteria from four to five (Widiger, 2007). Bearing this sensitive nature in mind, any changes to the PD criteria should be cautiously and deliberately made. Contrary to most expectations, it appears that, with regard to PD, the American Psychiatric Association (APA) acted accordingly in the newly published DSM.
The multiaxial structure has been removed in DSM, fifth edition (DSM-5), while the DSM, fourth edition, text revision (DSM-IV-TR) PD categories have been retained with only minor text revisions. However, in an effort to introduce “a new approach that aims to address numerous shortcomings of the current approach to personality disorders,” an “alternative DSM-5 model for personality disorders” was included (APA, 2013, p. 761). This alternative approach uses a dimensional severity/trait model to describe PD, wherein four PD categories have been eliminated completely. The inclusion of two inconsistent diagnostic systems for the same categories in DSM-5 is indicative of a lack of satisfaction with the DSM-IV-TR PDs on the part of researchers and clinicians. It also reflects the wide range of opinions about the best way of improving PD diagnosis. Most importantly, though, it signifies the APA’s present reluctance to make radical changes to the PD diagnoses without sufficient time to build empirical support for them.
Arguments about the need for major changes to the PD diagnostic criteria in the new DSM are, themselves, nothing new, and have been described by many authors. For a full exposition of the problems with the DSM-IV-TR PDs, see O’Donohue, Fowler, and Lilienfeld (2007). Looking forward, we must ask this question: Is the alternative DSM-5 model for PD an improvement over the DSM-IV-TR system that has been retained in the main diagnostic section of DSM-5? This article will assess how PD is viewed diagnostically in the alternative DSM-5 model by explicating the constructs and highlighting their most recent and relevant critiques. After a brief overview of PD in DSM-IV-TR, each of the changes to PD diagnosis in the alternative DSM-5 model will be laid out in order of their appearance in the general PD criteria. After discussing these changes and their support, relevant criticisms will be considered.
PDs in DSM-5
The Context of the New Changes: PD in DSM-IV-TR
Traditionally, PDs have been conceptualized categorically: “The diagnostic approach used in [DSM-IV-TR] represents the categorical perspective that Personality Disorders are qualitatively distinct clinical syndromes” (APA, 2000, p. 689). With the publication of DSM-III, all PDs were assigned to Axis II in an attempt to differentiate them from most other mental disorders. There were 10 specific categories of PD in DSM-IV-TR: paranoid, schizoid, schizotypal, antisocial, borderline, narcissistic, histrionic, dependent, avoidant, and obsessive-compulsive. However, one could be diagnosed with “personality disorder: not otherwise specified” using the general PD criteria even if they were not diagnosable with any of the 10 specific categories (APA, 2000).
The general criteria of PD in DSM-IV-TR required an enduring pattern of behavior or inner experience in two of the following realms: cognition, affect, interpersonal functioning, or impulse control. The pattern must have begun by early adulthood and be inflexible across time and circumstance. It must cause distress or functional impairment, must not be better accounted for by a general medical condition or the effects of a substance, and must defy cultural expectations (APA, 2000).
Outline of the Changes in the Alternative DSM-5 Model for PD
There are eight primary changes to the PD diagnostic constructs in the alternative DSM-5 model: (a) removal of four PD categories, (b) introduction of a dimensional conceptualization of PDs, (c) introduction of a functional impairment severity rating scale, (d) introduction of pathological personality trait descriptors, (e) elimination of the strict temporal stability criterion, (f) elimination of the Axis I exclusion, (g) elimination of the conduct disorder requirement for antisocial PD, and (h) introduction of the category PD-trait specified (PDTS).
Removal of Four PD Categories
The DSM-5 PD Work Group chose to eliminate four DSM-IV-TR PD categories, namely, paranoid, schizoid, histrionic, and dependent PDs (APA, 2012). The Work Group justifies the deletion of the first three of these PDs by first stating that “there are almost no empirical studies focused explicitly on [them]” and second by noting that the trait compositions of all four were too simple for their inclusion. For example, the single DSM-5 trait facet of suspiciousness covers all of the DSM-IV-TR paranoid PD criteria (APA, 2012, p. 9). Their conductive argument for retaining the remaining six PDs includes “prevalence in community and clinical populations, associated functional impairment, treatment and prognostic significance, and…neurobiological and genetic studies” (APA, 2012, p. 9).
Introduction of a Dimensional Conceptualization of PDs
The most serious criticism of the PD constructs of DSM-IV-TR is that they are entirely based on a false premise, namely, that PDs are discrete diagnostic entities that are categorical in nature. But this contradicts the findings of taxometric analysis of the PD constructs, which find no discrete PD categories (Bradley, Conklin, & Westen, 2007). For example, Conway, Hammen, and Brennan (2012) used latent class analysis, latent trait modeling, and factor mixture modeling to determine whether borderline PD (BPD) is best understood categorically, dimensionally as the expression of one or more traits, or as a mixture of both. They found that a dimensional latent trait model best represented the disorder, and that a single trait (dubbed “BPD-ness”) best explained the variation and severity in subjects (Conway, Hammen, & Brennan, 2012).
A dimensional conceptualization potentially serves to resolve the problem of comorbidity across, and heterogeneity within, the individual DSM-IV-TR PD categories by using transcategorical and subcategorical factors to describe PD. A dimensional understanding can provide more coverage than the existing categories by allowing for the description of atypical personality pathologies. At the same time, a potential deficit to such an understanding is the necessary establishment of, perhaps, arbitrary cutoffs that are essential to the binary decisions of clinical practice. Examples of these binary choices include the decisions to admit or not admit and treat or not treat. However, cutoffs for other continuous variables and their related disorders in medical practice, such as blood pressure and hypotension, have been established with high treatment utility, so this may not prove to be a serious problem for a dimensional conceptualization of PD (Dowson & Grounds, 1995; Widiger, 2007).
Introduction of a Functional Impairment Severity Rating Scale
The DSM-5 PD Work Group incorporated into the new general PD criteria a 5-point rating scale to specify the type and degree of functional impairment found in patients. The Criterion A requires “Moderate or greater impairment in personality functioning, manifested by difficulties in two or more of the following areas: (1) Identity, (2) Self-direction, (3) Empathy, or (4) Intimacy” (APA, 2013, p. 770). The criterion is polythetic in that significant impairment in two categories is required to meet the criterion. The 5-point scale ranges from 0 (healthy functioning) to 4 (extreme impairment; APA, 2013, pp. 775–778).
Functional impairment of identity is measured according to the stability of sense of self, adherence to role-appropriate boundaries, and level, regulation, and tolerance of self-esteem. Impairment of self-direction considers coherence of goals, standards of behavior, and awareness of subjective experience. Degree of empathic impairment is measured by the level of understanding others’ motivations and experiences, level of understanding and toleration of others’ perspectives, and awareness of one’s own effects on others. Finally, level of intimacy functioning is determined by existence and maintenance of multiple satisfying relationships, desire for close relationships, capacity for cooperation, and mutuality of regard reflected in interpersonal behavior (APA, 2013, p. 762).
According to the Work Group, Criterion A is beneficial to the PD constructs because research “indicates that generalized severity is the most important single predictor of concurrent and prospective dysfunction in assessing personality psychopathology…” (APA, 2012, p. 5). The self-other differentiation enables the development and implementation of superior treatment interventions and informs a more accurate prognosis (APA, 2012). Furthermore, self and interpersonal functioning is not only consistent with the theoretical aspects of PD, such as cognitive-behavioral, interpersonal, psychodynamic, attachment, developmental, social cognitive, and evolutionary theories, but is also widely considered “a key aspect of personality pathology in need of clinical attention” (APA, 2012, p. 7). Finally, the test–retest reliability of Criterion A used by untrained, experienced clinicians was minimally adequate (intraclass correlation coefficient [ICC] = .416) in field trials (APA, 2012). This reliability may increase as DSM-5 training becomes more prevalent.
Many of the DSM-IV-TR criteria can be described using Criterion A. For example, BPD chronic emptiness and identity disturbance are functional impairments related to the self, while the characteristic “stable instability” of relationships can be considered a functional impairment relating to both empathy and intimacy, the two subcategories of interpersonal functioning. The problem of comorbidity among PDs in DSM-IV-TR is addressed by Criterion A in that multiple syndromes can share the same types of functional impairment without full-fledged co-occurrence. In addition, Criterion A helps describe genuine categorical subtypes found within DSM-IV-TR PDs, such as overt and covert narcissistic PD (see Bradley et al., 2007), as well as other heterogeneity, with relevant accuracy and parsimony.
Introduction of Pathological Personality Trait Descriptors
The next major change to PDs in the alternative DSM-5 model is the addition of a set of pathological personality trait descriptors. The new Criterion B requires “One or more pathological personality trait domains OR specific trait facets within domains…” (APA, 2013, p. 770). The Work Group has devised a set of five trait domains, which are further defined by 25 trait facets (see APA, 2013, pp. 779–781). For each patient, the clinician will address all trait domains and facets, giving a dimensional rating according to the level of expression of each trait. Criterion B is versatile in that it can be used to describe maladaptive personality traits of all patients, whether or not they are diagnosable with a PD.
The DSM-5 PD Work Group claims that the Criterion B trait model is “an extension of the Five Factor Model” (FFM) of normal personality traits, but also includes the more extreme and maladaptive personality traits that are necessary to accurately describe dispositions associated with PD (APA, 2012, p. 7). That the trait model is somehow related to the FFM grants it credibility, as there is a significant amount of empirical literature that relates the DSM-IV-TR PD categories to the FFM (APA, 2012; Miller, Morse, Nolf, Stepp, & Pilkonis, 2012). Field trials have demonstrated relatively good correlations (r = .62 to .75) between the DSM-5 traits and the retained DSM-IV-TR PD categories, while the test–retest reliability of the five trait domains has been positive (ICC = .765 to .857; APA, 2012, p. 8).
A major benefit of including Criterion B is that the excessive comorbidity among PDs is greatly reduced or, perhaps, even resolved. This is because PD traits can and do overlap across PD categories as with, for example, impulsivity in BPD and antisocial PD. But this does not threaten the construct validity of the diagnostic constructs because any comorbidity that exists is an accurate measurement of real phenomena. Criterion B also serves to describe reliable variation within PDs, helping, in conjunction with Criterion A, to differentiate between different PD categories, as well as between subtypes within PDs (APA, 2012).
Another problem that Criterion B resolves is that of the temporal stability of PDs. In DSM-IV-TR, the general criteria for PD stated that PD is stable across time (APA, 2000, p. 689). The problem was that within the categorical conceptualization of PD of DSM-IV-TR, there was no explanation for this stability. Personality trait research, however, tells us that traits become stable to a significant degree in some individuals by the age of 3, obtain an adult-like structure by school-age, and increase in stability from then until at least 50 years (APA, 2012). By using traits to describe PDs, DSM-5 improves upon the explanatory scope and power of DSM-IV-TR, because it is the stability of the traits themselves that are responsible for the stability of the PDs (APA, 2012).
Elimination of the Strict Temporal Stability Criterion, the Axis I Exclusion, and the Conduct Disorder Requirement for Antisocial PD
Criterion D of the general criteria for a PD in DSM-IV-TR reads, “The pattern is stable and of long duration…” (APA, 2000, p. 689). The corresponding Criterion C in DSM-5 adds the word “relatively,” and reads, “The impairments in personality functioning and the individual’s personality trait expression are relatively [emphasis added] stable across time…” (APA, 2013, p. 761). This change makes for a more accurate portrayal of PD, and is based on empirical evidence showing PD symptoms to fluctuate in their expression across time, as well as the clinical course of PD patients, which tends toward remission (APA, 2012).
Some PDs in DSM-IV-TR, such as paranoid, schizoid, schizotypal, and antisocial PDs, have an “Axis I Exclusion.” Criterion D of antisocial PD, for example, states that “the occurrence of antisocial behavior is not exclusively during the course of schizophrenia or a manic episode” (APA, 2000, p. 706). The DSM-5 PD Work Group found these exclusions to be inconsistent and failing to take into account the possibility of heterotypic continuity, and, therefore, deleted them from the criteria of all PD categories (APA, 2013, pp. 764–770). Consequently, within the alternative DSM-5 model, an individual can be diagnosed with antisocial PD, even when the symptoms are only present during manic episodes.
Finally, the Work Group eliminated the DSM-IV-TR requirement of conduct disorder for the diagnosis of antisocial PD (APA, 2000, 2013, pp. 764–765). To justify this change, the Work Group notes that antisocial PD is no different from any other PD in terms of pathological antecedents or precursors in childhood or adolescence (APA, 2012). By deleting this from the alternative criteria, the PD categories are more consistent and streamlined.
Introduction of PDTS
All PD categories in the alternative DSM-5 model follow the structure of the general criteria and only differ in the levels and types of functional impairment (Criterion A) and the presence and expression of maladaptive personality trait domains and facets (Criterion B). For all PDs that do not fall within the six specific categories, there is the category of PDTS. The criteria for this diagnosis are the same as those for the general PD diagnosis, but the clinician must include the patient’s specific levels and types of functional impairment, as well as the patient’s specific maladaptive personality trait domains or trait facets (APA, 2013, pp. 770). PDTS will provide more coverage for atypical or unique PD.
Criticisms of the DSM-5 PD Constructs
There have been far too many criticisms of the alternative DSM-5 PD constructs to include them all in this article. However, most criticisms have been addressed in the Work Group’s Rationale for the Proposed Changes to the Personality Disorders Classification in DSM-5, published on the DSM-5 website in May 2012 (APA, 2012). Therefore, the following section will focus on relevant critiques that have been written or printed after May 2012. Other than Livesley and Verheul, none of the authors of the following criticisms were affiliated with the DSM-5 PD Work Group.
A “Hybrid” Model
According to the Work Group, DSM-5 uses a “hybrid” categorical/dimensional model for the PD constructs (APA, 2012, p. 4). This amounts to an implicit assertion that PDs make up discrete classes of persons but are ultimately constituted by incremental trait expression and functional impairment. However, according to Conway et al. (2012), this assertion is incorrect, at least when attributed to BPD, which is one of the DSM-IV-TR PD categories with the most existing research. To reach this conclusion, they used latent trait, latent class, and factor mixture modeling on data obtained from 700 twenty-year-olds using the Structured Clinical Interview for DSM-IV Axis II PDs (Conway et al., 2012).
In the analysis of the three modeling types, it was found that, while there could be as many as four classes of BPD, it is more likely that these classes represent severity levels, and that a dimensional trait model is the best fit for BPD. Concerning the use of a hybrid categorical/dimensional model in DSM-5, Conway et al. (2012) conclude that “the current analysis reveals relatively poor fit for a hybrid latent structure of BPD…” (p. 800). They add that their research suggests that different traits should be weighted differently in diagnosis, rather than uniformly as they are in DSM-5. Significant limitations to this study include the fact that all subjects in the sample were 20 years of age and that subjects with maternal depression were overselected. The former of these is potentially problematic because many symptoms of BPD, such as impulsivity and suicidality, vary according to age, while the overselection may have influenced the symptom presentation of BPD in the subjects (Conway et al., 2012).
Deleting Half of the DSM-IV-TR PD Categories
The PD Work Group removed 4 of the 10 PD categories found in DSM-IV-TR. According to the Work Group, “critiques of the DSM-5 proposal have almost universally been against the deletion of any of the DSM-IV PD types,” based on their treatment utility and heuristic value (APA, 2012, p. 12). Mullins-Sweatt, Bernstein, and Widiger (2012, pp. 691–692) make the most recent argument against deleting specific categories, pointing out that “there is perhaps nothing more significant to the field of personality disorder research than the inclusion of diagnostic categories within the professional society’s official diagnostic manual.” They argue that there is not “a broad consensus of expert clinical opinion” to eliminate any PD category; nor is there a sufficient lack of validity or utility in any PD to remove it from the DSM. Consensus and validity/utility were requirements explicitly declared by the PD Work Group for the deletion of any PD category.
To demonstrate this, Mullins-Sweatt et al. (2012) asked members of the Association for Research on Personality Disorders and the International Society for the Study of Personality Disorders to participate in an online survey concerning the utility and validity of the DSM-IV-TR PD categories (n = 520). There was a 28% response rate, but only 130 surveys were completed sufficiently to be included in analysis. The most important findings were that a substantial majority of respondents believed all PDs—especially antisocial and BPDs—to have validity, a majority considered all PDs except histrionic PD (HPD) to be “important” or “very important” in treatment decisions, a majority felt that all PDs except HPD should “probably” or “definitely” not be deleted, and no majority supported the deletion of any PD, including HPD.
To explain the dichotomy between “expert clinical opinion” and the Work Group’s decision to eliminate many of the PDs in DSM-5, the researchers point to a literature review on the retention of PD categories conducted by the Work Group members themselves. The review “does not really appear to have been systematic or comprehensive in its coverage of its validity and utility research” (Mullins-Sweatt, Bernstein, & Widiger, 2012, p. 698). Specifically, the Work Group members did not include a description of their inclusion/exclusion criteria and failed to cover a significant number of studies that are pertinent to the validity and utility of PDs. Furthermore, summaries found within the literature review were sometimes completely wrong (Mullins-Sweatt et al., 2012). As Mullins-Sweatt et al. (2012) take note, the Work Group asserts that there has been relatively little research done on the deleted categories to establish their utility and validity. In response, however, Mullins-Sweatt et al. make two points. First, absence of evidence is not evidence of absence. Second, the lack of pertinent research could just as easily be accounted for by a failure on the part of the PD field as by a lack of validity of the PD categories. Arguably, a consensus of opinion among researchers is not sufficient reason to retain a PD category, but it should inspire some caution in the process of deleting that category.
Co-Occurrence of PDs and Schizophrenia Spectrum Disorders (SSD) in DSM-5
It was one of the implicit goals of the DSM-5 PD Work Group to resolve the problem of excessive co-occurrence among the PDs. While the new alternative DSM-5 Criteria A and B have been successful in this regard concerning DSM-IV-TR Axis II disorders, the problem of high comorbidity between PDs and DSM-IV-TR Axis I disorders may have been exacerbated. Schroeder et al. (2012) conducted a study of day-clinic patients between 18 and 65 with SSD (n = 45) in order to find the prevalence of PDs among this population.
The results showed that SSD subjects had a 20% likelihood of being diagnosable with a DSM-IV-TR PD category. While this finding pertains to the DSM-IV-TR categories, the high level of interrater agreement between the DSM-IV-TR and alternative DSM-5 categories makes this relevant to the new DSM-5 PD constructs. In addition, four correlations between study participants and PD traits were found that were both significant and medium to high in effect strength: schizotypy, paranoia, depressivity, and avoidance (Schroeder et al., 2012). These levels of co-occurrence constitute a baseline by which we can judge the DSM-5 PD constructs. Schroeder et al. (2012) argue that we can expect co-occurrence to increase with the use of DSM-5. This is because SSD patients typically meet PDTS Criterion A as a direct consequence of their SSD symptoms, often meet Criterion C, and necessarily meet Criteria D and E based on their overlap with the respective SSD criteria. The only other requirement is the presence of any of the 25 trait facets in Criterion B. This problem is a threat to the construct validity of the DSM-5 PD criteria because it indicates that they have less than optimal specificity. This study’s primary limitation is a small sample size.
Criticisms by Former DSM-5 PD Work Group Members
Two of the 11 members of the DSM-5 PD Work Group resigned in protest in 2012 and have since written scathing criticisms of the alternative DSM-5 PD constructs. Verheul (2012) calls attention first to the lack of empirical support for the new general criteria, especially Criterion A, the levels of personality functioning, the support for which is limited to a single, well-received study. Likewise, the specific dimensional ratings for the categories retained from DSM-IV-TR are based on the opinions of various Work Group members, rather than on specific research.
Verheul’s (2012) next criticism is that the DSM-5 PD constructs are far too complex to be considered user-friendly among clinicians. First, Criterion B combines both pure dimensions (e.g., suspiciousness and hostility) and multidimensional types (e.g., negative affectivity and detachment) within the general criteria and the various retained PD categories. Second, the alternative DSM-5 model for PDs has increased complexity when compared with DSM-IV-TR, as can be seen in terms of sheer number of criteria and words: DSM-IV-TR BPD involves 9 criteria and 180 words, while the alternative DSM-5 BPD involves 29 subcriteria and 375 words.
Third, the alternative DSM-5 criteria require significantly more inference and interpretation than those in DSM-IV-TR. For example, one criterion of BPD in DSM-IV-TR reads, “markedly and persistently unstable self-image or sense of self,” while the corresponding criterion in DSM-5 reads, “markedly impoverished, poorly developed or unstable self-image, often associated with excessive self-criticism” (APA, 2000, p. 710, 2013, p. 766). Finally, the new alternative DSM model draws on sundry theoretical frameworks, including trait psychology (e.g., unstable emotional experiences and frequent mood changes) and psychoanalytic theory (e.g., ongoing awareness of a unique self and maintains role-appropriate boundaries), making the criteria difficult to use for anyone without extensive knowledge in the pertinent areas. Not only does the complexity threaten acceptability among clinicians, but it also predicts poor interrater reliability, which has arguably been confirmed in field trials, depending on one’s statistical threshold of acceptability (Verheul, 2012). “Interrater reliability is conditional to validity and utility,” writes Verheul, “therefore, insufficient reliability is unacceptable and should be followed by a rejection of the proposal in its current form” (2012, p. 370).
John Livesley (2012), also a former member of the DSM-5 PD Work Group, insists that, while a radical change is needed to address the poor structural and discriminant validity of DSM-IV-TR, the final DSM-5 proposal, which is almost identical to the published alternative DSM-5 PD model, was too flawed to accept. The first problem Livesley has is with the use of categories in DSM-5 to describe PDs despite voluminous evidence that such a framework is incorrect. Especially perplexing is the Work Group’s own admission that this is the case when they say that most DSM-IV-TR PD problems “can well be understood as examples of the fact that personality features and psychopathological tendencies do not tend to delineate categories of persons in nature [emphasis added]…” (2012, p. 85). Livesley points out that the Work Group had a choice between tradition and empirical evidence and chose tradition.
The problem is compounded by the attempt to combine both categories and dimensions. Livesley (2012) objects to the Work Group’s claim that the model is a “hybrid” of these two frameworks. In fact, by their very definitions, the two cannot be united in a single system. The categorical framework assumes that the features of the construct are discontinuous, while the dimensional framework makes the opposite assumption that features are continuous.
Livesley (2012) then calls attention to the fact that trait definitions are not uniform across all PD categories in the alternative DSM-5 PD model. Impulsivity, for example, is found in both antisocial and BPD categories; yet the phrase, “a sense of urgency and self-harming behavior under emotional distress,” is only found within the BPD category, and not the antisocial PD category (p. 86). This inconsistent use of traits breaches the basic tenet of trait psychology that trait definitions are universal across all people, with the only difference between individuals being their level of trait expression (Livesley, 2012). It is true that the word “impulsivity” is traditionally used equivocally in clinical settings, but Livesely sees this as another example of the Work Group choosing tradition over empirical research.
Finally, Livesley (2012) questions the basis of Criterion B. The Work Group chose its own model, rather than choosing one from the extant literature. This can be compared to the International Classification of Diseases (ICD)–11th Revision, which will base differences between PDs on the FFM (Livesley, 2012). In fact, within the section of the Work Group’s most recent rationale for the Criterion B trait model, of the seven citations made to lend it direct support, four of these were primarily authored within the last 3 years by Robert Kreuger, a member of the DSM-5 PD Work Group (APA, 2012, p. 7). It appears that the Work Group may have opted for self-promotion over empirical evidence in formulating Criterion B.
Discussion and Applications to Social Work
The question that this article aims to address (are the DSM-5 alternative PD diagnostic constructs an improvement over the DSM-IV-TR constructs?) is difficult to answer: In some ways they are better and in other ways they are not. It would seem that the poor construct validity of the DSM-IV-TR PDs evidenced by excessive comorbidity—they were too broadly constructed—has been replaced with a different type of poor construct validity—they are now too narrow. In a sense, coverage has been improved with the addition of the PDTS category, but coverage has been diminished by the deletion of four other categories that, according to the PD research community, were both highly useful and valid (Mullins-Sweatt et al., 2012). The diagnostic reliability of the new trait model (Criterion B) is a significant advancement, but the reliability of the functional impairment (Criterion A) rating system could be far worse than that for any of the DSM-IV-TR PD constructs.
The lack of empirical evidence for the validity and associated treatments of the PDs is as much of a problem for the alternative DSM-5 model as it was for DSM-IV-TR. The adjustment of Criterion C from “stable” to “relatively stable across time” is a small step toward correcting the pejorative connotations of the PDs. Widespread disagreement about specific PD criteria in the alternative DSM-5 model is already underway. Finally, while the use of dimensions to measure PDs in the alternative DSM-5 model is generally viewed positively, the continued use of categories alongside the dimensions is confounding.
The changes in the alternative DSM-5 PD model are no doubt the correct categories of changes. The idea of using levels of functional impairment to measure PD has been well received, as has been the dimensional use of personality traits, but the integral model selected by the Work Group appears to be seriously flawed. The state of PD diagnostics in the alternative model seems to be only minimally and superficially improved as compared with DSM-IV-TR. While this article should not be perceived as an endorsement of the alternative model, the decision to annex the alternative model to the main diagnostic section of DSM-5 may ultimately be a painful but necessary step that will bring us closer to a more valid DSM, sixth edition revision.
The multiaxial system of previous editions of the DSM has been eliminated in DSM-5. This involved the removal of Axis IV, which was devoted to psychosocial and environmental problems. The DSM-5 Task Force opted not to replace Axis IV with a new and original set of codes but rather refers clinicians to the relevant ICD–Tenth Revision–Clinical Modification and ICD-10 Z codes. Regardless of the acceptability of the ICD codes, the lack of attention devoted to psychosocial and environmental problems by the DSM-5 work groups indicates that the social work perspective was not given due consideration in the new DSM. Indeed, it appears from the list of Task Force members that no one with a master of social work degree was included in any of the various DSM-5 work groups and study groups.
However, it is still possible that social workers may be able to benefit from the alternative DSM-5 PD model. All clients can be assessed using the personality trait domains and facets, whether or not they have a PD or any mental disorder at all. Problems that maladaptive personality traits can pose for clients range from barriers to employment to alienation of family members or even social workers. Therefore, the DSM-5 system of personality trait domains (see APA, 2013, pp. 779–781), which have shown good reliability, may prove to be a very useful assessment tool for social workers and have the potential to improve treatment plans, even if their validity is doubtful on grounds other than reliability.
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.