511997 2013
ANP48410.1177/0004867413511997ANZJP ArticlesSellbom et al.
Research
Convergence between DSM-5 Section II and Section III diagnostic criteria for borderline personality disorder Martin Sellbom1, Randy A Sansone2,3, Douglas A Songer2,4 and Jaime L Anderson5
Australian & New Zealand Journal of Psychiatry 2014, Vol. 48(4) 325–332 DOI: 10.1177/0004867413511997 © The Royal Australian and New Zealand College of Psychiatrists 2013 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav anp.sagepub.com
Editor’s Choice
Abstract Objective: Borderline personality disorder (BPD) is a common mental health condition in psychiatric settings. The current study examined the overlap between the operationalization of BPD listed in Section II (Diagnostic Criteria and Codes) and the alternative, dimensional personality trait-based operationalization listed in Section III (Emerging Measures and Models) of the current Diagnostic and Statistical Manual of Mental Disorders (DSM-5). Moreover, the unique contributions of specific personality traits for indexing the traditional BPD operationalization were also evaluated, including conceptually relevant traits not originally proposed for Section III BPD. Method: Participants were 145 consecutive patients from a psychiatric unit in a hospital in the USA. These individuals completed a series of questionnaires that index both traditional (DSM-IV/DSM-5 Section II) and alternative (DSM-5 Section III personality traits) diagnostic criteria for BPD. Results: Structural equation modeling results revealed that latent constructs representing the Section II and Section III operationalizations of BPD, respectively, overlapped substantially (r = 0.86, p < 0.001). Hierarchical latent regression models indicated that at least five of the seven traits proposed to define Section III BPD uniquely accounted for variance (69%) in a latent Section II BPD variable. Finally, at least one other conceptually relevant trait (Perceptual Dysregulation) augmented the prediction of latent BPD scores. Conclusions: The proposed personality traits for Section III BPD are clearly aligned with traditional conceptualizations of this important personality disorder construct. At least five of the seven dimensional traits proposed to define Section III contributed uniquely to the characterization of Section II BPD, and these traits can be augmented by Perceptual Dysregulation. If replicated in other settings, these findings might warrant some modification to the operationalization of DSM-5 Section III BPD. Keywords Borderline personality disorder, DSM-5, personality traits
Introduction Borderline personality disorder (BPD) is a severe mental health condition that is estimated to affect 2–6% of the general population and from 10% to 25% of individuals undergoing mental health treatment (American Psychiatric Association (APA), 2013; Gunderson and Links, 2008; Lenzenweger et al., 2007; Torgersen et al., 2001). The diagnostic criteria for BPD have traditionally been criticized for various reasons, including, but not limited to, the use of an arbitrary polythetic criterion approach (i.e. five of nine symptoms required for the disorder, resulting in extreme heterogeneity among individuals diagnosed with this disorder) as well as inordinately high comorbidity with other mental
health disorders (e.g. Fyer et al., 1988; Tyrer, 2009). These types of concerns have led some authorities to question 1Research
School of Psychology, The Australian National University, Canberra, Australia 2Wright State University School of Medicine, Dayton, USA 3Kettering Medical Center, Kettering, USA 4Miami Valley Hospital, Dayton, USA 5The University of Alabama, Tuscaloosa, USA Corresponding author: Martin Sellbom, Research School of Psychology, Building 39, The Australian National University, Canberra, ACT 0200, Australia. Email: [email protected].
Australian & New Zealand Journal of Psychiatry, 48(4)
326 whether BPD should be considered a diagnosis at all (e.g. Tyrer, 2009, 2013; Tyrer et al., 2011). In preparation for the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5; APA, 2013), the Personality and Personality Disorders workgroup offered a proposal that would move the personality disorder system (including BPD) away from the theoretically and practically flawed categorical approach to one that emphasizes dimensional personality traits (and resulting impairment in functioning) as core criteria (Krueger et al., 2011; Skodol, 2012). The five trait domains in this latter model include Antagonism, Psychoticism, Disinhibition, Negative Affectivity, and Detachment (Skodol et al., 2011). Each trait domain is comprised of specific facets, allowing a higher fidelity description of more specific personality trait elements. In the pursuit of diagnostic clarity, but also to maintain continuity with the DSM-IV, these traits are then combined to define six specific personality disorders, which include Antisocial, Avoidant, Borderline, Narcissistic, Obsessive-Compulsive, and Schizotypal personality disorders. For all possible configurations that do not match one of the specific personality disorder types, Personality Disorder: Trait Specified is an additional diagnosis that makes use of the full personality trait model, and serves as a replacement to the current category of Personality Disorder Not Elsewhere Classified. For the diagnosis of BPD in particular, the personality traits associated with the dimensional model include Anxiousness, Depressivity, Emotional Lability, Separation Insecurity, Hostility (from the Negative Affectivity domain), Impulsivity, and Risk Taking (from the Disinhibition domain). (Note: Hostility was originally conceived as a facet of Antagonism, but most empirical evidence indicates that it loads more strongly with the Negative Affectivity domain (e.g. Krueger et al., 2012).) With the ultimate publication of the manual, the DSM-5 retained the personality disorder categories previously listed in DSM-IV-TR (APA, 2000) for Section II (Diagnostic Criteria and Codes) without any substantive modification. The initial proposal for revision of the DSM-5 personality disorders was maintained, yet relegated to Section III (Emerging Measures and Models) as an alternative model. The main impetus for this decision was presumably that the alternative model required further scientific inquiry and validation prior to being implemented in clinical practice (e.g. DSM-5.1 or beyond). With respect to BPD, limited research is available to inform whether the alternative trait-based approach will ultimately be useful in capturing the construct of BPD. In a broader evaluation of the dimensional personality disorder proposal, Hopwood and colleagues (2012) found in a large sample of undergraduate students that six of the seven proposed traits for Section III BPD correlated moderately to strongly with Section II BPD as indexed by the Personality Diagnostic Questionnaire-4 (PDQ-4; Hyler, 1994). Risk Australian & New Zealand Journal of Psychiatry, 48(4)
ANZJP Articles Taking was the lone trait facet that was not meaningfully associated with Section II BPD. In a mixed community and psychiatric sample, Miller and colleagues (2012) used traits from the five factor model of personality to emulate the DSM-5 proposal and showed that these traits could capture a substantial amount of variance in BPD. Although these findings are very promising with respect to evaluating the continuity of BPD and eventually moving from DSM-5 Section II to Section III, more research is needed in clinical/psychiatric settings that diagnostically epitomize the actual traits listed in DSM-5 to determine their unique contribution to the BPD construct, and whether additional conceptually relevant traits can augment this characterization. The current investigation sought to examine the convergence between the DSM-5 Section II and Section III operationalizations of BPD to determine: (1) the degree of continuity in diagnosis moving from the polythetic criterion approach to dimensional personality traits, and (2) whether additional conceptually and empirically relevant personality traits might further augment the operationalization of Section III BPD. More specifically, we examined traits we deemed conceptually relevant to the diagnostic construct of BPD, including submissiveness in interpersonal relationships, lack of restricted affectivity, interpersonal suspiciousness, and perceptual dysregulation, with the latter two possibly being best connected to Section II BPD criterion 9.
Method Participants The sample consisted of 167 consecutive, newly admitted patients in a psychiatric unit located in a midwestern US hospital, who were approached during the study duration. Potential participants were excluded if they exhibited obvious signs of intellectual (e.g. severe intellectual disability), medical (e.g. pain), psychiatric (e.g. acute psychosis), and/ or cognitive impairment (e.g. dementia) that would preclude the successful completion of a survey. This exclusionary process was informally undertaken by the subject recruiter, an experienced psychiatrist. This led to the exclusion of 14 patients (three with vision problems, four with an intellectual disability or illiteracy, four deemed too psychotic, and three who were judged to be malingering). Moreover, eight patients declined to participate (a response rate of 95% among eligible participants). This resulted in a working sample of 145 patients, who anonymously completed the assessment protocol. The final sample consisted of 55 men and 89 women (one participant declined to report gender), with a mean age of 38.06 (SD = 13.06; range = 18–71) years. Most participants self-identified as non-Hispanic white (73.6%); the remaining participants were black individuals (16.0%) or of other or mixed ethnicities (10.1%). In terms of
327
Sellbom et al. education, most participants (87.4%) had at least a high school education, with 38.5% reporting a 4-year college degree or higher. Approximately 41% of participants reported never having married, with 17.2% being married, 26.9% divorced, 9.9% separated, and 2.8% widowed. The majority of patients had government insurance (51%), with the remainder having either private insurance (15%) or being self-pay (34%). Moreover, 64.3% of participants reported at least one previous psychiatric hospitalization (M = 4.05, SD = 5.52). Although diagnoses were not formally assessed due to research ethics board restrictions, patients in this facility are typically diagnosed with schizophrenia, bipolar disorder, substance use disorders, and BPD. With regard to BPD specifically, 97 (66.9%) patients scored above standard cut-offs on at least one of the three DSM-5 Section II BPD measures (see ‘Measures’ below), 78 (53.8%) scored above cut-offs on at least two, and 60 (41.3%) patients scored above cut-offs on all three measures. As such, a large proportion of patients were deemed likely to meet at least partial criteria for BPD, which is consistent with the general facility population.
Measures Personality Inventory for DSM-5 (PID-5). The PID-5 (Krueger et al., 2012) is a 220-item self-report inventory developed to index the five DSM-5 Section III personality domains (Negative Affectivity, Detachment, Antagonism, Disinhibition, and Psychoticism) and their respective 25 facets. Item responses are based on a Likert scale ranging from 0 (very false or often false) to 3 (very true or often true). The factor structure and other evidence for external construct validity for this measure has been documented in numerous studies (e.g. Anderson et al., 2013; Hopwood et al., 2012; Thomas et al., 2013; Wright et al., 2012). An abbreviated 124-item version, that covers the PID-5 trait facets pertaining to BPD as well as other conceptually relevant facets, was administered in the current study. Personality Diagnostic Questionnaire-4 (PDQ-4). The BPD scale of the PDQ-4 (Hyler, 1994) is a nine-item, true/false, self-report measure that reflects the diagnostic criteria for BPD that are listed in the DSM-IV (APA, 1994). A score of 5 or higher is highly suggestive of the diagnosis of BPD. Previous versions of the PDQ have been found to be useful screening tools for BPD in both clinical samples (Dubro et al., 1988; Hyler et al., 1990) and non-clinical samples (Johnson and Bornstein, 1992), including the use of the freestanding BPD scale (Patrick et al., 1995). Structured Clinical Interview for DSM-IV Axis II Disorders – Personality Questionnaire (SCID-II-PQ). The BPD scale of the SCID-II-PQ (First et al., 1997) is a 15-item, true/false, selfreport measure that assesses BPD according to the DSM-IV (APA, 1994) diagnostic criteria. Although initially designed
as a screening device (Carey, 1994), a number of studies have shown that the scales have reasonable diagnostic validity (e.g. Neal et al., 1997). McLean Screening Inventory for Borderline Personality Disorder (MSI-BPD). The MSI-BPD (Zanarini et al., 2003) is a 10-item, yes/no, self-report questionnaire that explores borderline personality symptomatology. All endorsements are in the pathological direction and scores of seven or higher are suggestive of the disorder. This measure has undergone limited clinical study and is recommended by the authors as a screening measure for BPD; however, initial validity results show promising classification accuracy (Zanarini et al., 2003).
Procedure Shortly following admission, potential candidates were approached by one of the authors (DAS) and invited into the study. Participants were informed that the study was designed to (1) explore personality features and (2) compare these features across various tests administered as part of the present study. All materials were packaged into a research booklet and provided to participants. All measures were self-report in nature and completed privately on the inpatient unit during hospitalization. Because of anonymity and the self-report methodology, participants were not required to sign formal consent forms. However, the elements of informed consent were provided on the cover page of the research booklets, which indicated that completion of materials functioned as informed consent. Participants were advised to remove and retain the cover page for their own records. This project was approved by the Institutional Review Board of the sponsoring hospital.
Results Table 1 shows the descriptive statistics, reliabilities, and zero-order correlations among all study measures. The correlations indicate that the DSM-5 Section II measures are strongly associated, as expected. The DSM-5 Section III traits used to define the alternative version of BPD were also associated at a large magnitude, and these traits were also significantly and meaningfully associated with the PDQ-4, MSI-BPD, and SCID-II-PQ BPD scale scores. Structural equation modeling was used to address primary research questions. First, we estimated a measurement model to determine the association between latent constructs representing Section II and Section III BPD, respectively. Maximum Likelihood was used in Mplus 7.1 to estimate parameters and missing data. In this model, the PDQ-4, SCID-II-PQ, and MSI-BPD total scores served as indicators for the Section II BPD factor whereas the seven PID-5 traits scores were indicators for the Section III BPD factor. After applying two conceptually defensible modification indices, Australian & New Zealand Journal of Psychiatry, 48(4)
328
ANZJP Articles
Table 1. Descriptive statistics, reliabilities, and inter-correlations among study measures. M
SD
Range
1
2
3
4
5
6
7
8
9
10
1. PDQ-4
4.85
2.47
0–9
(0.74)
0.73
0.67
0.46
0.48
0.62
0.52
0.59
0.53
0.35
2. SCID-II-PQ
8.64
4.33
0–10
(0.86)
0.83
0.53
0.49
0.69
0.59
0.68
0.60
0.47
3. MSI-BPD
6.03
2.93
0–10
(0.78)
0.62
0.61
0.61
0.60
0.60
0.53
0.36
4. Anxiousness
1.87
0.80
0–3
(0.90)
0.71
0.68
0.63
0.50
0.43
0.11
5. Depressivity
1.48
0.82
0–3
(0.93)
0.61
0.55
0.46
0.40
0.16
6. Emotional Lability
1.73
0.84
0–3
(0.87)
0.71
0.68
0.58
0.31
7. Separation Insecurity
1.49
0.87
0–3
(0.85)
0.49
0.52
0.39
8. Hostility
1.35
0.76
0–3
(0.89)
0.59
0.44
9. Impulsivity
1.28
0.84
0–3
(0.87)
0.60
10. Risk Taking
1.33
0.63
0–3
(0.85)
Internal consistency reliabilities (coefficient alpha) are in parentheses. |r| > 0.27 is statistically significant at p < 0.001. PDQ-4: Personality Diagnostic Questionnaire-4; SCID-II-PQ: Structured Clinical Interview for DSM-IV Axis II Disorders – Personality Questionnaire; MSI-BPD: McLean Screening Inventory for Borderline Personality Disorder.
model fit was generally acceptable, χ2 = 78.35, df = 32, p < 0.001, confirmatory fit index (CFI) = 0.953, Tucker–Lewis index (TLI) = 0.934, root mean square error of approximation (RMSEA) = 0.099, standardised root mean residual (SRMR) = 0.051, though mediocre from the perspective of the RMSEA statistic. However, the RMSEA is highly sensitive to small sample sizes and small models (MacCallum et al., 1996; see also Kenny et al., 2011). Figure 1 shows the final measurement model; as evident, the association between the two latent constructs was quite large (r = 0.86, p < 0.001). We next calculated the associations between individual DSM-5 Section III BPD traits and the latent Section II BPD factor. These correlations are shown in Table 2 and indicate large effect sizes, except for PID-5 Risk Taking, which is considered medium (r = 0.46). We subsequently examined the degree to which the seven PID-5 scores uniquely contributed to this prediction. Thus, we regressed the latent Section II BPD variable onto the seven traits. The overall model fit was mediocre to acceptable: χ2 = 33.26, df = 14, p = 0.002, CFI = 0.960, TLI = 0.940, RMSEA = 0.097, SRMR = 0.021. Although a substantial proportion of variance was accounted for in latent BPD scores, only three (PID-5 Emotional Lability, Hostility, and Risk Taking) of the seven PID-5 scores contributed uniquely to this prediction. As this was likely a statistical power issue, we pruned the model by fixing the parameter associated with the smallest magnitude (PID-5 Anxiousness) to zero, which did not significantly reduce model fit (Δχ2 = 1.01, df = 1, p > 0.05). At this point, four of the remaining six parameters Australian & New Zealand Journal of Psychiatry, 48(4)
were statistically significant predictors of the latent BPD variable. We pruned the model again by fixing the parameter currently associated with the smallest effect size (PID-5 Impulsivity) to zero, which did not result in a significant reduction of model fit relative to the original model (Δχ2 = 3.39, df = 2, p > 0.05). Now, all five remaining PID-5 scales were unique predictors of latent Section II BPD scores and, in combination, they accounted for 69.4% of variance in this latent variable. Figure 2 shows the final model, including standardized beta weights associated with each predictor. Finally, we examined whether additional PID-5 trait scores could augment the prediction of latent Section II BPD scores. We examined four traits in total: PID-5 Submissiveness, Suspiciousness, (Lack of) Restricted Affectivity, and Perceptual Dysregulation. As shown in Table 2, these four target PID-5 traits showed moderate (PID-5 Submissiveness, Suspiciousness, (Lack of) Restricted Affectivity) to large (PID-5 Perceptual Dysregulation) correlations with this latent variable. Because of concerns about statistical power for the regression model, we used the final pruned model displayed in Figure 2 as the baseline model and examined whether each trait added to the prediction of latent BPD scores individually (i.e. we estimated four separate models) above and beyond those already established in the previous step. We found that PID-5 Suspiciousness (z = 2.14, p = 0.027) and PID-5 Perceptual Dysregulation (z = 2.50, p = 0.014), but not (Lack of) Restricted Affectivity or Submissiveness, added incrementally to this prediction in separate models.
329
Sellbom et al.
Figure 1. All covariance parameters are statistically significant (p < 0.001). (Sec II: DSM-5 Section II BPD; Sec III: DSM-5 Section III BPD; imp: Impulsivity; risk: Risk Taking; sepins: Separation Insecurity; host: Hostility; emo: Emotional Lability; dep: Depressivity; anx: Anxiousness; msi: McLean Screening Inventory for Borderline Personality Disorder; scid2: Structured Clinical Interview for DSM-IV Axis II Disorders – Personality Questionnaire; pdq: Personality Diagnostic Questionnaire-4.)
Table 2. Correlations between PID-5 trait facet scores and Latent Section II BPD scores.
only, accounted for 71.3% of the variance in the latent BPD variable. This final model is shown in Figure 3, which includes standardized beta weights for all predictors.
PID-5 scale
Latent BPD score
Anxiousness
0.57
Discussion
Depressivity
0.54
Emotional Lability
0.70
Separation Insecurity
0.61
Hostility
0.69
Impulsivity
0.61
Risk Taking
0.46
Submissiveness
0.36
(Lack of) Restricted Affectivity
0.37
Suspiciousness
0.47
Perceptual Dysregulation
0.60
The current investigation provides further evidence that the DSM-5 Section III personality trait criteria can capture a substantial proportion of variance in the traditional conceptualization of the BPD construct. The latent representations of the Section II and Section III BPD constructs shared almost 74% of the variance. Moreover, a linear combination of the traits facets assigned to Section III BPD captured over 69% of latent Section II BPD variance. These findings add to the growing literature (e.g. Hopwood et al., 2012; Miller et al., 2012), indicating that these designated personality traits can indeed provide some continuity from the outdated and highly problematic categorical diagnosis of BPD to one that emphasizes dimensional personality traits. To our knowledge, this is the first study to examine the unique predictive contributions of the individual Section III traits in accounting for variance in the traditional conceptualization of BPD. The current findings indicate that at least five of the seven traits (Depressivity, Emotional Lability, Hostility, Separation Insecurity, and Risk Taking) contribute independently to this prediction. However, it is noteworthy that the two that did not, Anxiousness and Impulsivity, evidenced large zero-order correlations with
All correlations are statistically significant (p < 0.001). PID-5: Personality Inventory for DSM-5; BPD: borderline personality disorder.
When adding PID-5 Suspiciousness and Perceptual Dysregulation simultaneously, only the latter added significantly to the prediction of latent BPD scores (z = 2.01, p =0.045). The final model, adding Perceptual Dysregulation
Australian & New Zealand Journal of Psychiatry, 48(4)
330
ANZJP Articles
Figure 2. All regression parameters are statistically significant (p < 0.001). (risk: Risk Taking; sepins: Separation Insecurity; host: Hostility; emo: Emotional Lability; dep: Depressivity; imp: Impulsivity; anx: Anxiousness; bpd: borderline personality disorder; msi: McLean Screening Inventory for Borderline Personality Disorder; scid2: Structured Clinical Interview for DSM-IV Axis II Disorders – Personality Questionnaire; pdq: Personality Diagnostic Questionnaire-4.)
the latent BPD construct (see also Hopwood et al., 2012). It is possible that not all traits currently included to define Section III BPD are necessary to capture the traditional BPD conceptualization. Indeed, perhaps there is sufficient overlap between Impulsivity and Risk Taking that only one is necessary to capture these disinhibited aspects of the disorder. Moreover, Anxiousness tends to overlap largely with the other Negative Affectivity traits (see, for instance, Table 1) and might therefore not uniquely contribute to BPD. Nevertheless, we cannot ignore the potential for Type II error, and, thus, these findings should be considered tentative until replicated in other samples. Furthermore, we also found that two traits (Perceptual Dysregulation and Suspiciousness) contributed incrementally to the prediction of the latent BPD construct in separate models, and Perceptual Dysregulation continued to be a unique predictor when both of these traits were considered in the same model. This finding is not surprising given that the current DSM-5 Section III traits for BPD do not include a parallel trait for Section II BPD criterion 9, which involves transient, stress-induced dissociative and/or paranoid states. Moreover, the current results add to previous
Australian & New Zealand Journal of Psychiatry, 48(4)
findings in other clinical samples that have clearly shown that the broader personality domain of psychoticism tends to be strongly associated with BPD in addition to negative affectivity (e.g. Bagby et al., 2008; Sellbom et al., 2013; Wygant et al., 2006). Therefore, pending replication, future revisions of the personality trait constellation for BPD should consider these trait facets in addition to those already included. The current conclusions must be considered in light of some potential limitations upon which future studies can build. First, given the complex modeling, we were limited in statistical power in our regression models. As such, it is difficult to rule out Type II error for the failure of two of the seven traits in contributing uniquely to the prediction of Section II BPD, and future research studies need to follow up on this issue using larger samples. Moreover, given practical design issues, we were forced to rely upon selfreport measures for all constructs, which likely inflated the magnitude of associations across constructs. Future research needs to consider other measurement modalities (e.g. interviews, clinician-ratings, peer-ratings) when replicating these findings. Additionally, the study sample
331
Sellbom et al.
Figure 3. All regression parameters are statistically significant (p < 0.001). (risk: Risk Taking; sepins: Separation Insecurity; host: Hostility; emo: Emotional Lability; dep: Depressivity; perdys: Perceptual Dysregulation; imp: Impulsivity; anx: Anxiousness; bpd: borderline personality disorder; msi: McLean Screening Inventory for Borderline Personality Disorder; scid2: Structured Clinical Interview for DSM-IV Axis II Disorders – Personality Questionnaire; pdq: Personality Diagnostic Questionnaire-4.)
represents an inpatient population and, therefore, only one specific segment of the entire BPD population; whether some findings (e.g. Perceptual Dysregulation) will generalize to patient samples with lower rates of pathology is unknown. Finally, although we believe that these findings do have implications for the proposed personality trait constellation for Section III BPD, it is important to highlight that we did not measure the other major criterion for Section III BPD diagnosis, impairment in self- and inter-personal functioning, and, as such, our conclusions are specifically limited to the trait criteria for BPD. Once reliable and valid measurements of these impairment criteria are available, future research should incorporate such instruments when evaluating the Section III BPD diagnosis. Funding This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
Declaration of interest The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
References American Psychiatric Association (1994) Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition. Washington, DC: American Psychiatric Association. American Psychiatric Association (2000) Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision. Washington, DC: American Psychiatric Association. American Psychiatric Association (2013) Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition. Washington, DC: American Psychiatric Association. Anderson JL, Sellbom M, Bagby M, et al. (2013) Examining the MMPI2-RF PSY-5 Scales for the Assessment of DSM-5 Personality Trait Dimensions. Assessment 20: 286–294. Bagby RM, Sellbom M, Costa PT, et al. (2008) Predicting Diagnostic and Statistical Manual of Mental Disorders–IV personality disorders with the five-factor model of personality and the Personality Psychopathology Five. Personality and Mental Health 2: 55–69.
Australian & New Zealand Journal of Psychiatry, 48(4)
332 Carey KB (1994) Use of the Structured Clinical Interview for DSM-III–R Personality Questionnaire in the presence of severe Axis I disorders: A cautionary note. Journal of Nervous and Mental Disease 182: 669–671. Dubro AF, Wetzler S and Kahn MW (1988). A comparison of three selfreport questionnaires for the diagnosis of DSM-III personality disorders. Journal of Personality Disorders 2: 256–266. First MB, Gibbon M, Spitzer RL, et al. (1997) Structured Clinical Interview for DSM-IV Axis II Personality Disorders (SCID-II). Washington, DC: American Psychiatric Press, Inc. Fyer MR, Frances AJ, Sullivan T, et al. (1988) Comorbidity of borderline personality disorder. Archives of General Psychiatry 45: 348–352. Gunderson JG and Links P (2008) Borderline Personality Disorder: A Clinical Guide, 2nd edition. Washington, DC: American Psychiatric Press. Hopwood CJ, Thomas KM, Markon KE, et al. (2012) DSM-5 personality traits and DSM-IV personality disorders. Journal of Abnormal Psychology 121: 424–432. Hyler SE (1994) Personality Questionnaire, PDQ-4+. New York: New York State Psychiatric Institute. Hyler SE, Skodol AE, Kellman H, et al. (1990) Validity of the Personality Diagnostic Questionnaire–Revised: Comparison with two structured interviews. The American Journal of Psychiatry 147: 1043–1048. Johnson JG and Bornstein RF (1992) Utility of the Personality Diagnostic Questionnaire–Revised in a nonclinical population. Journal of Personality Disorders 6: 450–457. Kenny DA, Kaniskan B and McCoach DB (2011) The performance of RMSEA in models with small degrees of freedom [unpublished paper]. University of Connecticut, Storrs, CT. Krueger RF, Eaton NR, Clark LE, et al. (2011) Deriving an empirical structure of personality pathology for DSM-5. Journal of Personality Disorders 25: 170–191. Krueger RF, Derringer J, Markon KE, et al. (2012) Initial construction of a maladaptive personality trait model and inventory for DSM-5. Psychological Medicine 42: 1879–1890. Lenzenweger MF, Lane MC, Loranger AW, et al. (2007) DSM-IV personality disorders in the National Comorbidity Survey replication. Biological Psychiatry 62: 553–564. MacCallum RC, Browne MW and Sugawara HM (1996) Power analysis and determination of sample size for covariance structure modeling. Psychological Methods 1: 130–149. Miller JD, Morse JQ, Nolf K, et al. (2012) Can DSM-IV borderline personality disorder be diagnosed via dimensional personality traits?
Australian & New Zealand Journal of Psychiatry, 48(4)
ANZJP Articles Implications for the DSM-5 personality disorder proposal. Journal of Abnormal Psychology 121: 944–950. Neal LA, Fox CC, Carroll NN, et al. (1997) Development and validation of a computerized screening test for personality disorders in DSMIII-R. Acta Psychiatrica Scandinavica 95: 351–356. Patrick J, Links P, Van Reekum R, et al. (1995) Using the PDQ–R BPD scale as a brief screening measure in the differential diagnosis of personality disorder. Journal of Personality Disorders 9: 266–274. Sellbom M, Smid W, De Saeger H, et al. (2013) Mapping the Personality Psychopathology Five Domains onto DSM-IV Personality Disorders in Dutch Clinical and Forensic Samples: Implications for the DSM-5. Journal of Personality Assessment. Epub ahead of print 5 Sept 2013. DOI: 10.1080/00223891.2013.825625. Skodol AE (2012) Personality disorders in DSM-5. Annual Review of Clinical Psychology 8: 317–344. Skodol AE, Bender DS, Morey LC, et al. (2011) Personality disorder types proposed for DSM-5. Journal of Personality Disorders 25: 136–169. Thomas KM, Yalch MM, Krueger RF, et al. (2013) The convergent structure of DSM-5 personality trait facets and five-factor model trait domains. Assessment 20: 308–311. Torgersen S, Kringlen E and Cramer V (2001) The prevalence of personality disorders in a community sample. Archives of General Psychiatry 58: 590–596. Tyrer P (2009) Why borderline personality disorder is neither borderline nor a personality disorder. Personality and Mental Health 3: 86–95. Tyrer P (2013) The classification of personality disorders in ICD-11: Implications for forensic psychiatry. Criminal Behaviour and Mental Health 23: 1–5. Tyrer P, Crawford M, Mulder R, et al. (2011) The rationale for the reclassification of personality disorder in the 11th revision of the International Classification of Diseases (ICD-11). Personality and Mental Health 5: 246–259. Wright AG, Thomas KM, Hopwood CJ, et al. (2012) The hierarchical structure of DSM-5 pathological personality traits. Journal of Abnormal Psychology 121: 951–957. Wygant DB, Sellbom M, Graham JR, et al. (2006) Incremental validity of the MMPI-2 PSY-5 scales in assessing self-reported personality disorder criteria. Assessment 13: 178–186. Zanarini MC, Vujanovic A, Parachini EA, et al. (2003) A screening measure for BPD: The McLean Screening Instrument for Borderline Personality Disorder (MSI-BPD). Journal of Personality Disorders 17: 568–573.
ANP48410.1177/0004867413511997ANZJP ArticlesSellbom et al.
Research
Convergence between DSM-5 Section II and Section III diagnostic criteria for borderline personality disorder Martin Sellbom1, Randy A Sansone2,3, Douglas A Songer2,4 and Jaime L Anderson5
Australian & New Zealand Journal of Psychiatry 2014, Vol. 48(4) 325–332 DOI: 10.1177/0004867413511997 © The Royal Australian and New Zealand College of Psychiatrists 2013 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav anp.sagepub.com
Editor’s Choice
Abstract Objective: Borderline personality disorder (BPD) is a common mental health condition in psychiatric settings. The current study examined the overlap between the operationalization of BPD listed in Section II (Diagnostic Criteria and Codes) and the alternative, dimensional personality trait-based operationalization listed in Section III (Emerging Measures and Models) of the current Diagnostic and Statistical Manual of Mental Disorders (DSM-5). Moreover, the unique contributions of specific personality traits for indexing the traditional BPD operationalization were also evaluated, including conceptually relevant traits not originally proposed for Section III BPD. Method: Participants were 145 consecutive patients from a psychiatric unit in a hospital in the USA. These individuals completed a series of questionnaires that index both traditional (DSM-IV/DSM-5 Section II) and alternative (DSM-5 Section III personality traits) diagnostic criteria for BPD. Results: Structural equation modeling results revealed that latent constructs representing the Section II and Section III operationalizations of BPD, respectively, overlapped substantially (r = 0.86, p < 0.001). Hierarchical latent regression models indicated that at least five of the seven traits proposed to define Section III BPD uniquely accounted for variance (69%) in a latent Section II BPD variable. Finally, at least one other conceptually relevant trait (Perceptual Dysregulation) augmented the prediction of latent BPD scores. Conclusions: The proposed personality traits for Section III BPD are clearly aligned with traditional conceptualizations of this important personality disorder construct. At least five of the seven dimensional traits proposed to define Section III contributed uniquely to the characterization of Section II BPD, and these traits can be augmented by Perceptual Dysregulation. If replicated in other settings, these findings might warrant some modification to the operationalization of DSM-5 Section III BPD. Keywords Borderline personality disorder, DSM-5, personality traits
Introduction Borderline personality disorder (BPD) is a severe mental health condition that is estimated to affect 2–6% of the general population and from 10% to 25% of individuals undergoing mental health treatment (American Psychiatric Association (APA), 2013; Gunderson and Links, 2008; Lenzenweger et al., 2007; Torgersen et al., 2001). The diagnostic criteria for BPD have traditionally been criticized for various reasons, including, but not limited to, the use of an arbitrary polythetic criterion approach (i.e. five of nine symptoms required for the disorder, resulting in extreme heterogeneity among individuals diagnosed with this disorder) as well as inordinately high comorbidity with other mental
health disorders (e.g. Fyer et al., 1988; Tyrer, 2009). These types of concerns have led some authorities to question 1Research
School of Psychology, The Australian National University, Canberra, Australia 2Wright State University School of Medicine, Dayton, USA 3Kettering Medical Center, Kettering, USA 4Miami Valley Hospital, Dayton, USA 5The University of Alabama, Tuscaloosa, USA Corresponding author: Martin Sellbom, Research School of Psychology, Building 39, The Australian National University, Canberra, ACT 0200, Australia. Email: [email protected].
Australian & New Zealand Journal of Psychiatry, 48(4)
326 whether BPD should be considered a diagnosis at all (e.g. Tyrer, 2009, 2013; Tyrer et al., 2011). In preparation for the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5; APA, 2013), the Personality and Personality Disorders workgroup offered a proposal that would move the personality disorder system (including BPD) away from the theoretically and practically flawed categorical approach to one that emphasizes dimensional personality traits (and resulting impairment in functioning) as core criteria (Krueger et al., 2011; Skodol, 2012). The five trait domains in this latter model include Antagonism, Psychoticism, Disinhibition, Negative Affectivity, and Detachment (Skodol et al., 2011). Each trait domain is comprised of specific facets, allowing a higher fidelity description of more specific personality trait elements. In the pursuit of diagnostic clarity, but also to maintain continuity with the DSM-IV, these traits are then combined to define six specific personality disorders, which include Antisocial, Avoidant, Borderline, Narcissistic, Obsessive-Compulsive, and Schizotypal personality disorders. For all possible configurations that do not match one of the specific personality disorder types, Personality Disorder: Trait Specified is an additional diagnosis that makes use of the full personality trait model, and serves as a replacement to the current category of Personality Disorder Not Elsewhere Classified. For the diagnosis of BPD in particular, the personality traits associated with the dimensional model include Anxiousness, Depressivity, Emotional Lability, Separation Insecurity, Hostility (from the Negative Affectivity domain), Impulsivity, and Risk Taking (from the Disinhibition domain). (Note: Hostility was originally conceived as a facet of Antagonism, but most empirical evidence indicates that it loads more strongly with the Negative Affectivity domain (e.g. Krueger et al., 2012).) With the ultimate publication of the manual, the DSM-5 retained the personality disorder categories previously listed in DSM-IV-TR (APA, 2000) for Section II (Diagnostic Criteria and Codes) without any substantive modification. The initial proposal for revision of the DSM-5 personality disorders was maintained, yet relegated to Section III (Emerging Measures and Models) as an alternative model. The main impetus for this decision was presumably that the alternative model required further scientific inquiry and validation prior to being implemented in clinical practice (e.g. DSM-5.1 or beyond). With respect to BPD, limited research is available to inform whether the alternative trait-based approach will ultimately be useful in capturing the construct of BPD. In a broader evaluation of the dimensional personality disorder proposal, Hopwood and colleagues (2012) found in a large sample of undergraduate students that six of the seven proposed traits for Section III BPD correlated moderately to strongly with Section II BPD as indexed by the Personality Diagnostic Questionnaire-4 (PDQ-4; Hyler, 1994). Risk Australian & New Zealand Journal of Psychiatry, 48(4)
ANZJP Articles Taking was the lone trait facet that was not meaningfully associated with Section II BPD. In a mixed community and psychiatric sample, Miller and colleagues (2012) used traits from the five factor model of personality to emulate the DSM-5 proposal and showed that these traits could capture a substantial amount of variance in BPD. Although these findings are very promising with respect to evaluating the continuity of BPD and eventually moving from DSM-5 Section II to Section III, more research is needed in clinical/psychiatric settings that diagnostically epitomize the actual traits listed in DSM-5 to determine their unique contribution to the BPD construct, and whether additional conceptually relevant traits can augment this characterization. The current investigation sought to examine the convergence between the DSM-5 Section II and Section III operationalizations of BPD to determine: (1) the degree of continuity in diagnosis moving from the polythetic criterion approach to dimensional personality traits, and (2) whether additional conceptually and empirically relevant personality traits might further augment the operationalization of Section III BPD. More specifically, we examined traits we deemed conceptually relevant to the diagnostic construct of BPD, including submissiveness in interpersonal relationships, lack of restricted affectivity, interpersonal suspiciousness, and perceptual dysregulation, with the latter two possibly being best connected to Section II BPD criterion 9.
Method Participants The sample consisted of 167 consecutive, newly admitted patients in a psychiatric unit located in a midwestern US hospital, who were approached during the study duration. Potential participants were excluded if they exhibited obvious signs of intellectual (e.g. severe intellectual disability), medical (e.g. pain), psychiatric (e.g. acute psychosis), and/ or cognitive impairment (e.g. dementia) that would preclude the successful completion of a survey. This exclusionary process was informally undertaken by the subject recruiter, an experienced psychiatrist. This led to the exclusion of 14 patients (three with vision problems, four with an intellectual disability or illiteracy, four deemed too psychotic, and three who were judged to be malingering). Moreover, eight patients declined to participate (a response rate of 95% among eligible participants). This resulted in a working sample of 145 patients, who anonymously completed the assessment protocol. The final sample consisted of 55 men and 89 women (one participant declined to report gender), with a mean age of 38.06 (SD = 13.06; range = 18–71) years. Most participants self-identified as non-Hispanic white (73.6%); the remaining participants were black individuals (16.0%) or of other or mixed ethnicities (10.1%). In terms of
327
Sellbom et al. education, most participants (87.4%) had at least a high school education, with 38.5% reporting a 4-year college degree or higher. Approximately 41% of participants reported never having married, with 17.2% being married, 26.9% divorced, 9.9% separated, and 2.8% widowed. The majority of patients had government insurance (51%), with the remainder having either private insurance (15%) or being self-pay (34%). Moreover, 64.3% of participants reported at least one previous psychiatric hospitalization (M = 4.05, SD = 5.52). Although diagnoses were not formally assessed due to research ethics board restrictions, patients in this facility are typically diagnosed with schizophrenia, bipolar disorder, substance use disorders, and BPD. With regard to BPD specifically, 97 (66.9%) patients scored above standard cut-offs on at least one of the three DSM-5 Section II BPD measures (see ‘Measures’ below), 78 (53.8%) scored above cut-offs on at least two, and 60 (41.3%) patients scored above cut-offs on all three measures. As such, a large proportion of patients were deemed likely to meet at least partial criteria for BPD, which is consistent with the general facility population.
Measures Personality Inventory for DSM-5 (PID-5). The PID-5 (Krueger et al., 2012) is a 220-item self-report inventory developed to index the five DSM-5 Section III personality domains (Negative Affectivity, Detachment, Antagonism, Disinhibition, and Psychoticism) and their respective 25 facets. Item responses are based on a Likert scale ranging from 0 (very false or often false) to 3 (very true or often true). The factor structure and other evidence for external construct validity for this measure has been documented in numerous studies (e.g. Anderson et al., 2013; Hopwood et al., 2012; Thomas et al., 2013; Wright et al., 2012). An abbreviated 124-item version, that covers the PID-5 trait facets pertaining to BPD as well as other conceptually relevant facets, was administered in the current study. Personality Diagnostic Questionnaire-4 (PDQ-4). The BPD scale of the PDQ-4 (Hyler, 1994) is a nine-item, true/false, self-report measure that reflects the diagnostic criteria for BPD that are listed in the DSM-IV (APA, 1994). A score of 5 or higher is highly suggestive of the diagnosis of BPD. Previous versions of the PDQ have been found to be useful screening tools for BPD in both clinical samples (Dubro et al., 1988; Hyler et al., 1990) and non-clinical samples (Johnson and Bornstein, 1992), including the use of the freestanding BPD scale (Patrick et al., 1995). Structured Clinical Interview for DSM-IV Axis II Disorders – Personality Questionnaire (SCID-II-PQ). The BPD scale of the SCID-II-PQ (First et al., 1997) is a 15-item, true/false, selfreport measure that assesses BPD according to the DSM-IV (APA, 1994) diagnostic criteria. Although initially designed
as a screening device (Carey, 1994), a number of studies have shown that the scales have reasonable diagnostic validity (e.g. Neal et al., 1997). McLean Screening Inventory for Borderline Personality Disorder (MSI-BPD). The MSI-BPD (Zanarini et al., 2003) is a 10-item, yes/no, self-report questionnaire that explores borderline personality symptomatology. All endorsements are in the pathological direction and scores of seven or higher are suggestive of the disorder. This measure has undergone limited clinical study and is recommended by the authors as a screening measure for BPD; however, initial validity results show promising classification accuracy (Zanarini et al., 2003).
Procedure Shortly following admission, potential candidates were approached by one of the authors (DAS) and invited into the study. Participants were informed that the study was designed to (1) explore personality features and (2) compare these features across various tests administered as part of the present study. All materials were packaged into a research booklet and provided to participants. All measures were self-report in nature and completed privately on the inpatient unit during hospitalization. Because of anonymity and the self-report methodology, participants were not required to sign formal consent forms. However, the elements of informed consent were provided on the cover page of the research booklets, which indicated that completion of materials functioned as informed consent. Participants were advised to remove and retain the cover page for their own records. This project was approved by the Institutional Review Board of the sponsoring hospital.
Results Table 1 shows the descriptive statistics, reliabilities, and zero-order correlations among all study measures. The correlations indicate that the DSM-5 Section II measures are strongly associated, as expected. The DSM-5 Section III traits used to define the alternative version of BPD were also associated at a large magnitude, and these traits were also significantly and meaningfully associated with the PDQ-4, MSI-BPD, and SCID-II-PQ BPD scale scores. Structural equation modeling was used to address primary research questions. First, we estimated a measurement model to determine the association between latent constructs representing Section II and Section III BPD, respectively. Maximum Likelihood was used in Mplus 7.1 to estimate parameters and missing data. In this model, the PDQ-4, SCID-II-PQ, and MSI-BPD total scores served as indicators for the Section II BPD factor whereas the seven PID-5 traits scores were indicators for the Section III BPD factor. After applying two conceptually defensible modification indices, Australian & New Zealand Journal of Psychiatry, 48(4)
328
ANZJP Articles
Table 1. Descriptive statistics, reliabilities, and inter-correlations among study measures. M
SD
Range
1
2
3
4
5
6
7
8
9
10
1. PDQ-4
4.85
2.47
0–9
(0.74)
0.73
0.67
0.46
0.48
0.62
0.52
0.59
0.53
0.35
2. SCID-II-PQ
8.64
4.33
0–10
(0.86)
0.83
0.53
0.49
0.69
0.59
0.68
0.60
0.47
3. MSI-BPD
6.03
2.93
0–10
(0.78)
0.62
0.61
0.61
0.60
0.60
0.53
0.36
4. Anxiousness
1.87
0.80
0–3
(0.90)
0.71
0.68
0.63
0.50
0.43
0.11
5. Depressivity
1.48
0.82
0–3
(0.93)
0.61
0.55
0.46
0.40
0.16
6. Emotional Lability
1.73
0.84
0–3
(0.87)
0.71
0.68
0.58
0.31
7. Separation Insecurity
1.49
0.87
0–3
(0.85)
0.49
0.52
0.39
8. Hostility
1.35
0.76
0–3
(0.89)
0.59
0.44
9. Impulsivity
1.28
0.84
0–3
(0.87)
0.60
10. Risk Taking
1.33
0.63
0–3
(0.85)
Internal consistency reliabilities (coefficient alpha) are in parentheses. |r| > 0.27 is statistically significant at p < 0.001. PDQ-4: Personality Diagnostic Questionnaire-4; SCID-II-PQ: Structured Clinical Interview for DSM-IV Axis II Disorders – Personality Questionnaire; MSI-BPD: McLean Screening Inventory for Borderline Personality Disorder.
model fit was generally acceptable, χ2 = 78.35, df = 32, p < 0.001, confirmatory fit index (CFI) = 0.953, Tucker–Lewis index (TLI) = 0.934, root mean square error of approximation (RMSEA) = 0.099, standardised root mean residual (SRMR) = 0.051, though mediocre from the perspective of the RMSEA statistic. However, the RMSEA is highly sensitive to small sample sizes and small models (MacCallum et al., 1996; see also Kenny et al., 2011). Figure 1 shows the final measurement model; as evident, the association between the two latent constructs was quite large (r = 0.86, p < 0.001). We next calculated the associations between individual DSM-5 Section III BPD traits and the latent Section II BPD factor. These correlations are shown in Table 2 and indicate large effect sizes, except for PID-5 Risk Taking, which is considered medium (r = 0.46). We subsequently examined the degree to which the seven PID-5 scores uniquely contributed to this prediction. Thus, we regressed the latent Section II BPD variable onto the seven traits. The overall model fit was mediocre to acceptable: χ2 = 33.26, df = 14, p = 0.002, CFI = 0.960, TLI = 0.940, RMSEA = 0.097, SRMR = 0.021. Although a substantial proportion of variance was accounted for in latent BPD scores, only three (PID-5 Emotional Lability, Hostility, and Risk Taking) of the seven PID-5 scores contributed uniquely to this prediction. As this was likely a statistical power issue, we pruned the model by fixing the parameter associated with the smallest magnitude (PID-5 Anxiousness) to zero, which did not significantly reduce model fit (Δχ2 = 1.01, df = 1, p > 0.05). At this point, four of the remaining six parameters Australian & New Zealand Journal of Psychiatry, 48(4)
were statistically significant predictors of the latent BPD variable. We pruned the model again by fixing the parameter currently associated with the smallest effect size (PID-5 Impulsivity) to zero, which did not result in a significant reduction of model fit relative to the original model (Δχ2 = 3.39, df = 2, p > 0.05). Now, all five remaining PID-5 scales were unique predictors of latent Section II BPD scores and, in combination, they accounted for 69.4% of variance in this latent variable. Figure 2 shows the final model, including standardized beta weights associated with each predictor. Finally, we examined whether additional PID-5 trait scores could augment the prediction of latent Section II BPD scores. We examined four traits in total: PID-5 Submissiveness, Suspiciousness, (Lack of) Restricted Affectivity, and Perceptual Dysregulation. As shown in Table 2, these four target PID-5 traits showed moderate (PID-5 Submissiveness, Suspiciousness, (Lack of) Restricted Affectivity) to large (PID-5 Perceptual Dysregulation) correlations with this latent variable. Because of concerns about statistical power for the regression model, we used the final pruned model displayed in Figure 2 as the baseline model and examined whether each trait added to the prediction of latent BPD scores individually (i.e. we estimated four separate models) above and beyond those already established in the previous step. We found that PID-5 Suspiciousness (z = 2.14, p = 0.027) and PID-5 Perceptual Dysregulation (z = 2.50, p = 0.014), but not (Lack of) Restricted Affectivity or Submissiveness, added incrementally to this prediction in separate models.
329
Sellbom et al.
Figure 1. All covariance parameters are statistically significant (p < 0.001). (Sec II: DSM-5 Section II BPD; Sec III: DSM-5 Section III BPD; imp: Impulsivity; risk: Risk Taking; sepins: Separation Insecurity; host: Hostility; emo: Emotional Lability; dep: Depressivity; anx: Anxiousness; msi: McLean Screening Inventory for Borderline Personality Disorder; scid2: Structured Clinical Interview for DSM-IV Axis II Disorders – Personality Questionnaire; pdq: Personality Diagnostic Questionnaire-4.)
Table 2. Correlations between PID-5 trait facet scores and Latent Section II BPD scores.
only, accounted for 71.3% of the variance in the latent BPD variable. This final model is shown in Figure 3, which includes standardized beta weights for all predictors.
PID-5 scale
Latent BPD score
Anxiousness
0.57
Discussion
Depressivity
0.54
Emotional Lability
0.70
Separation Insecurity
0.61
Hostility
0.69
Impulsivity
0.61
Risk Taking
0.46
Submissiveness
0.36
(Lack of) Restricted Affectivity
0.37
Suspiciousness
0.47
Perceptual Dysregulation
0.60
The current investigation provides further evidence that the DSM-5 Section III personality trait criteria can capture a substantial proportion of variance in the traditional conceptualization of the BPD construct. The latent representations of the Section II and Section III BPD constructs shared almost 74% of the variance. Moreover, a linear combination of the traits facets assigned to Section III BPD captured over 69% of latent Section II BPD variance. These findings add to the growing literature (e.g. Hopwood et al., 2012; Miller et al., 2012), indicating that these designated personality traits can indeed provide some continuity from the outdated and highly problematic categorical diagnosis of BPD to one that emphasizes dimensional personality traits. To our knowledge, this is the first study to examine the unique predictive contributions of the individual Section III traits in accounting for variance in the traditional conceptualization of BPD. The current findings indicate that at least five of the seven traits (Depressivity, Emotional Lability, Hostility, Separation Insecurity, and Risk Taking) contribute independently to this prediction. However, it is noteworthy that the two that did not, Anxiousness and Impulsivity, evidenced large zero-order correlations with
All correlations are statistically significant (p < 0.001). PID-5: Personality Inventory for DSM-5; BPD: borderline personality disorder.
When adding PID-5 Suspiciousness and Perceptual Dysregulation simultaneously, only the latter added significantly to the prediction of latent BPD scores (z = 2.01, p =0.045). The final model, adding Perceptual Dysregulation
Australian & New Zealand Journal of Psychiatry, 48(4)
330
ANZJP Articles
Figure 2. All regression parameters are statistically significant (p < 0.001). (risk: Risk Taking; sepins: Separation Insecurity; host: Hostility; emo: Emotional Lability; dep: Depressivity; imp: Impulsivity; anx: Anxiousness; bpd: borderline personality disorder; msi: McLean Screening Inventory for Borderline Personality Disorder; scid2: Structured Clinical Interview for DSM-IV Axis II Disorders – Personality Questionnaire; pdq: Personality Diagnostic Questionnaire-4.)
the latent BPD construct (see also Hopwood et al., 2012). It is possible that not all traits currently included to define Section III BPD are necessary to capture the traditional BPD conceptualization. Indeed, perhaps there is sufficient overlap between Impulsivity and Risk Taking that only one is necessary to capture these disinhibited aspects of the disorder. Moreover, Anxiousness tends to overlap largely with the other Negative Affectivity traits (see, for instance, Table 1) and might therefore not uniquely contribute to BPD. Nevertheless, we cannot ignore the potential for Type II error, and, thus, these findings should be considered tentative until replicated in other samples. Furthermore, we also found that two traits (Perceptual Dysregulation and Suspiciousness) contributed incrementally to the prediction of the latent BPD construct in separate models, and Perceptual Dysregulation continued to be a unique predictor when both of these traits were considered in the same model. This finding is not surprising given that the current DSM-5 Section III traits for BPD do not include a parallel trait for Section II BPD criterion 9, which involves transient, stress-induced dissociative and/or paranoid states. Moreover, the current results add to previous
Australian & New Zealand Journal of Psychiatry, 48(4)
findings in other clinical samples that have clearly shown that the broader personality domain of psychoticism tends to be strongly associated with BPD in addition to negative affectivity (e.g. Bagby et al., 2008; Sellbom et al., 2013; Wygant et al., 2006). Therefore, pending replication, future revisions of the personality trait constellation for BPD should consider these trait facets in addition to those already included. The current conclusions must be considered in light of some potential limitations upon which future studies can build. First, given the complex modeling, we were limited in statistical power in our regression models. As such, it is difficult to rule out Type II error for the failure of two of the seven traits in contributing uniquely to the prediction of Section II BPD, and future research studies need to follow up on this issue using larger samples. Moreover, given practical design issues, we were forced to rely upon selfreport measures for all constructs, which likely inflated the magnitude of associations across constructs. Future research needs to consider other measurement modalities (e.g. interviews, clinician-ratings, peer-ratings) when replicating these findings. Additionally, the study sample
331
Sellbom et al.
Figure 3. All regression parameters are statistically significant (p < 0.001). (risk: Risk Taking; sepins: Separation Insecurity; host: Hostility; emo: Emotional Lability; dep: Depressivity; perdys: Perceptual Dysregulation; imp: Impulsivity; anx: Anxiousness; bpd: borderline personality disorder; msi: McLean Screening Inventory for Borderline Personality Disorder; scid2: Structured Clinical Interview for DSM-IV Axis II Disorders – Personality Questionnaire; pdq: Personality Diagnostic Questionnaire-4.)
represents an inpatient population and, therefore, only one specific segment of the entire BPD population; whether some findings (e.g. Perceptual Dysregulation) will generalize to patient samples with lower rates of pathology is unknown. Finally, although we believe that these findings do have implications for the proposed personality trait constellation for Section III BPD, it is important to highlight that we did not measure the other major criterion for Section III BPD diagnosis, impairment in self- and inter-personal functioning, and, as such, our conclusions are specifically limited to the trait criteria for BPD. Once reliable and valid measurements of these impairment criteria are available, future research should incorporate such instruments when evaluating the Section III BPD diagnosis. Funding This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
Declaration of interest The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
References American Psychiatric Association (1994) Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition. Washington, DC: American Psychiatric Association. American Psychiatric Association (2000) Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision. Washington, DC: American Psychiatric Association. American Psychiatric Association (2013) Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition. Washington, DC: American Psychiatric Association. Anderson JL, Sellbom M, Bagby M, et al. (2013) Examining the MMPI2-RF PSY-5 Scales for the Assessment of DSM-5 Personality Trait Dimensions. Assessment 20: 286–294. Bagby RM, Sellbom M, Costa PT, et al. (2008) Predicting Diagnostic and Statistical Manual of Mental Disorders–IV personality disorders with the five-factor model of personality and the Personality Psychopathology Five. Personality and Mental Health 2: 55–69.
Australian & New Zealand Journal of Psychiatry, 48(4)
332 Carey KB (1994) Use of the Structured Clinical Interview for DSM-III–R Personality Questionnaire in the presence of severe Axis I disorders: A cautionary note. Journal of Nervous and Mental Disease 182: 669–671. Dubro AF, Wetzler S and Kahn MW (1988). A comparison of three selfreport questionnaires for the diagnosis of DSM-III personality disorders. Journal of Personality Disorders 2: 256–266. First MB, Gibbon M, Spitzer RL, et al. (1997) Structured Clinical Interview for DSM-IV Axis II Personality Disorders (SCID-II). Washington, DC: American Psychiatric Press, Inc. Fyer MR, Frances AJ, Sullivan T, et al. (1988) Comorbidity of borderline personality disorder. Archives of General Psychiatry 45: 348–352. Gunderson JG and Links P (2008) Borderline Personality Disorder: A Clinical Guide, 2nd edition. Washington, DC: American Psychiatric Press. Hopwood CJ, Thomas KM, Markon KE, et al. (2012) DSM-5 personality traits and DSM-IV personality disorders. Journal of Abnormal Psychology 121: 424–432. Hyler SE (1994) Personality Questionnaire, PDQ-4+. New York: New York State Psychiatric Institute. Hyler SE, Skodol AE, Kellman H, et al. (1990) Validity of the Personality Diagnostic Questionnaire–Revised: Comparison with two structured interviews. The American Journal of Psychiatry 147: 1043–1048. Johnson JG and Bornstein RF (1992) Utility of the Personality Diagnostic Questionnaire–Revised in a nonclinical population. Journal of Personality Disorders 6: 450–457. Kenny DA, Kaniskan B and McCoach DB (2011) The performance of RMSEA in models with small degrees of freedom [unpublished paper]. University of Connecticut, Storrs, CT. Krueger RF, Eaton NR, Clark LE, et al. (2011) Deriving an empirical structure of personality pathology for DSM-5. Journal of Personality Disorders 25: 170–191. Krueger RF, Derringer J, Markon KE, et al. (2012) Initial construction of a maladaptive personality trait model and inventory for DSM-5. Psychological Medicine 42: 1879–1890. Lenzenweger MF, Lane MC, Loranger AW, et al. (2007) DSM-IV personality disorders in the National Comorbidity Survey replication. Biological Psychiatry 62: 553–564. MacCallum RC, Browne MW and Sugawara HM (1996) Power analysis and determination of sample size for covariance structure modeling. Psychological Methods 1: 130–149. Miller JD, Morse JQ, Nolf K, et al. (2012) Can DSM-IV borderline personality disorder be diagnosed via dimensional personality traits?
Australian & New Zealand Journal of Psychiatry, 48(4)
ANZJP Articles Implications for the DSM-5 personality disorder proposal. Journal of Abnormal Psychology 121: 944–950. Neal LA, Fox CC, Carroll NN, et al. (1997) Development and validation of a computerized screening test for personality disorders in DSMIII-R. Acta Psychiatrica Scandinavica 95: 351–356. Patrick J, Links P, Van Reekum R, et al. (1995) Using the PDQ–R BPD scale as a brief screening measure in the differential diagnosis of personality disorder. Journal of Personality Disorders 9: 266–274. Sellbom M, Smid W, De Saeger H, et al. (2013) Mapping the Personality Psychopathology Five Domains onto DSM-IV Personality Disorders in Dutch Clinical and Forensic Samples: Implications for the DSM-5. Journal of Personality Assessment. Epub ahead of print 5 Sept 2013. DOI: 10.1080/00223891.2013.825625. Skodol AE (2012) Personality disorders in DSM-5. Annual Review of Clinical Psychology 8: 317–344. Skodol AE, Bender DS, Morey LC, et al. (2011) Personality disorder types proposed for DSM-5. Journal of Personality Disorders 25: 136–169. Thomas KM, Yalch MM, Krueger RF, et al. (2013) The convergent structure of DSM-5 personality trait facets and five-factor model trait domains. Assessment 20: 308–311. Torgersen S, Kringlen E and Cramer V (2001) The prevalence of personality disorders in a community sample. Archives of General Psychiatry 58: 590–596. Tyrer P (2009) Why borderline personality disorder is neither borderline nor a personality disorder. Personality and Mental Health 3: 86–95. Tyrer P (2013) The classification of personality disorders in ICD-11: Implications for forensic psychiatry. Criminal Behaviour and Mental Health 23: 1–5. Tyrer P, Crawford M, Mulder R, et al. (2011) The rationale for the reclassification of personality disorder in the 11th revision of the International Classification of Diseases (ICD-11). Personality and Mental Health 5: 246–259. Wright AG, Thomas KM, Hopwood CJ, et al. (2012) The hierarchical structure of DSM-5 pathological personality traits. Journal of Abnormal Psychology 121: 951–957. Wygant DB, Sellbom M, Graham JR, et al. (2006) Incremental validity of the MMPI-2 PSY-5 scales in assessing self-reported personality disorder criteria. Assessment 13: 178–186. Zanarini MC, Vujanovic A, Parachini EA, et al. (2003) A screening measure for BPD: The McLean Screening Instrument for Borderline Personality Disorder (MSI-BPD). Journal of Personality Disorders 17: 568–573.
