Available online at www.sciencedirect.com

Comprehensive Psychiatry 53 (2012) 940–945www.elsevier.com/locate/comppsych

Which DSM-IV personality disorders are most strongly associated withindices of psychosocial morbidity in psychiatric outpatients?

Mark Zimmerman⁎, Iwona Chelminski, Diane Young, Kristy Dalrymple,Jennifer Martinez, Theresa A. Morgan

Department of Psychiatry and Human Behavior, Brown University School of Medicine, Rhode Island Hospital, Providence, RI

Abstract

The DSM-5 Work Group for Personality and Personality Disorders (PDs) recommended retaining 6 specific PD “types” (antisocial,avoidant, borderline, narcissistic, obsessive-compulsive, and schizotypal) and eliminating the other 4 PDs currently included in DSM-IV(paranoid, schizoid, histrionic, and dependent). One important clinical aspect of PDs is their association with indices of psychosocialmorbidity. Because the literature on the relationship between PDs and psychosocial morbidity in psychiatric patients is limited, we undertookthe current analysis of the Rhode Island Methods to Improve Diagnostic Assessment and Services project database to examine which PDswere most strongly associated with a variety of measures of psychosocial morbidity. We tested the hypothesis that the disordersrecommended for retention in DSM-5 would be associated with more severe morbidity than the disorders recommended for deletion. A totalof 2150 psychiatric outpatients were evaluated with semistructured diagnostic interviews for DSM-IV Axes I and II disorders and 7 measuresof psychosocial morbidity. We examined the correlation between each PD dimensional score and each measure of morbidity and thenconducted multiple regression analyses to determine which PDs were independently associated with the indices of morbidity. For the 6 PDsproposed for retention in DSM-5, 36 (85.7%) of the 42 correlations were significant, whereas for the 4 PDs proposed for deletion, 26 (92.9%)of the 28 correlations were significant. In the regression analyses for the 6 PDs proposed for retention in DSM-5, 19 (45.2%) of the 42 βcoefficients were significant, whereas for the 4 PDs proposed for deletion, 7 (25.0%) of the 28 β coefficients were significant. The results ofthe present study, along with the results of other studies, do not provide clear evidence for the preferential retention of some PDs over othersbased on their association with indices of psychosocial morbidity.© 2012 Elsevier Inc. All rights reserved.

1. Introduction

Through the years, there have been many critiques of theDSM-III, DSM-III-R, and DSM-IV approaches towardclassifying the personality disorders (PDs). These critiqueshave emphasized the problem of diagnostic overlap [1-5], thelack of a clear boundary between normality and abnormality[5-7], the failure to take into account findings from normalpersonality research [1,8], and the lack of diagnostic stabilityover time [2,5,7].

Because of these problems, the DSM-5 Work Group forPersonality and PDs recommended a reformulation of thePDs section [9]. The suggested changes followed the general

⁎ Corresponding author. Bayside Medical Center, Providence, RI02905.

E-mail address: mzimmerman@lifespan.org (M. Zimmerman).

0010-440X/$ – see front matter © 2012 Elsevier Inc. All rights reserved.doi:10.1016/j.comppsych.2012.02.008

guidelines for revising the DSM as proposed by Kendler et al[10]. Kendler et al distinguished between antecedent (eg,prior psychiatric history), concurrent (eg, patterns ofcomorbidity), and predictive (eg, course of illness) “criteriafor change” validators and suggested that to changediagnostic entities or criteria, sufficient clinically relevantevidence must be presented for past, present, and future(predictive) correlates of the disorder.

One of the Work Group's recommendations was toeliminate 4 of the DSM-IV PDs from DSM-5 (schizoid,paranoid, histrionic, and dependent) while retaining theremaining 6 DSM-IV PDs (antisocial, avoidant, borderline,narcissistic, obsessive-compulsive, and schizotypal). Thestated reason for reducing the number of PDs was to reducethe excessive diagnostic comorbidity among the PDs [9,11].The empirical support for the preferential retention of these 6disorders and eliminating schizoid, paranoid, histrionic, anddependent PDs is limited [12-15]. Part of the justification for

able 1emographic characteristics of 2150 psychiatric outpatients

haracteristic N %

exFemale 1310 60.9Male 840 39.1ducationbHigh school 178 8.3Graduated high school 1343 62.5≥Graduated college 629 29.2arital statusMarried 869 40.4Living with someone 127 5.9Widowed 36 1.7Separated 112 5.2

941M. Zimmerman et al. / Comprehensive Psychiatry 53 (2012) 940–945

the retention of these 6 is that 3 of them (borderline,antisocial, and schizotypal) have the greatest empiricalevidence of validity and clinical use. It was further arguedthat schizotypal and borderline PDs are more severelyimpairing than other “types,” although avoidant andobsessive-compulsive PDs were also noted to be impairing.The only other justification for retaining these 6 PDs wasprovided for obsessive-compulsive PD, which was noted tobe among the most common in community [16] and clinicalsamples [17], associated with increased mental healthtreatment use [18], and associated with the highest economicburden of all PDs [19].

We are aware of only 2 studies of psychiatric patients thathave compared the level of past, present, or future(predicted) psychosocial morbidity associated with the fullrange of PDs. Nakao et al [20] examined the associationbetween DSM-III-R PDs and functional impairment asmeasured by the Global Assessment of Functioning (GAF)scale in 149 psychiatric outpatients. All disorders excepthistrionic and obsessive-compulsive PDs were significantlyassociated with lower GAF scores. Soeteman et al [21]examined the association between the DSM-IV PDs andquality of life in a large sample of outpatients, day hospitalpatients, and inpatients receiving treatment for personalitypathology. Only depressive PD was significantly indepen-dently associated with lower quality of life. In anotheranalysis from this study, Soeteman et al [19] examined theeconomic burden associated with the PDs. In the regressionanalyses, they found that borderline and obsessive-compul-sive PDs were the only disorders significantly associatedwith overall higher direct and indirect costs, and these 2disorders along with paranoid PD were independentlyassociated with increased direct medical costs.

Because the literature on the relationship between PDsand psychosocial morbidity in psychiatric patients is limited,we undertook the current analysis of the Rhode IslandMethods to Improve Diagnostic Assessment and Servicesproject database to examine which PDs are most stronglyassociated with clinical indices of psychosocial morbidity.As markers of psychosocial morbidity, we examined thenumber of Axis I disorders, suicidal ideation at the time ofthe evaluation, history of suicide attempts, history ofpsychiatric hospitalizations, GAF ratings, time missed fromwork because of psychiatric disorders, and best level ofsocial functioning during the past 5 years. We tested thehypothesis that the disorders recommended for retention inDSM-5 would be associated with more severe morbiditythan the disorders recommended for exclusion.

Divorced 325 15.1Never married 681 31.7aceWhite 1952 90.8Black 95 4.4Hispanic 58 2.7Asian 21 1.0Other 24 1.1ge (y) mean, 38.5 SD, 12.9

2. Method

A total of 2150 psychiatric outpatients were evaluatedwith semistructured diagnostic interviews for DSM-IV AxesI and II disorders in the Rhode Island Hospital Department ofPsychiatry outpatient practice. This private practice group

predominantly treats individuals with medical insurance(including Medicare but not Medicaid) on a fee-for-servicebasis, and it is distinct from the hospital's outpatientresidency training clinic that predominantly serves lowerincome, uninsured, and medical-assistance patients. The datain Table 1 show the demographic characteristics of thesample. Most subjects were white, female, married or single,and had some college education. The mean age of the samplewas 38.5 years (SD, 12.9 years). The most frequent currentDSM-IV diagnoses were major depressive disorder (43.0%),social phobia (26.8%), generalized anxiety disorder (19.3%),and panic disorder (17.7%) (Table 2).

The patients were interviewed by a trained diagnosticrater who administered the Structured Interview for DSM-IVPersonality (SIDP-IV) [22] and a modified version of theStructured Clinical Interview for DSM-IV (SCID) [23]. TheRhode Island Hospital Institutional Review Committeeapproved the research protocol, and all patients providedinformed, written consent. Only a minority of patientsevaluated in the practice received the SCID and SIDP-IVbecause of the lack of available diagnostic raters or patients'preference for a less time-consuming standard clinicalevaluation. Patients who did and did not participate in thestudy were similar in sex, education, marital status, andscores on self-administered symptom questionnaires [24].

We integrated into the SCID interview the items from theSchedule for Affective Disorders and Schizophrenia (SADS)[25] on the amount of time missed from work because ofpsychiatric reasons during the past 5 years and best level ofsocial functioning during the past 5 years. The SCID/SADSinterview also included assessments of prior psychiatrichospitalizations, current suicidal ideation (rated on a 0-6scale on the SADS), and lifetime history of suicide attempts.

TD

C

S

E

M

R

A

Table 2Current DSM-IV diagnoses of 2150 psychiatric outpatients

DSM-IV diagnosis n %

Major depressive disorder 924 43.0Bipolar disorder 111 5.2Dysthymic disorder 179 8.3Generalized anxiety disorder 415 19.3Panic disorder 381 17.7Social phobia 576 26.8Specific phobia 225 10.5Obsessive-compulsive disorder 138 6.4Posttraumatic stress disorder 247 11.5Adjustment disorder 149 6.9Schizophrenia 8 0.4Eating disorder 143 6.6Alcohol abuse/dependence 207 9.6Drug abuse/dependence 103 4.8Somatoform disorder 167 7.8Attention deficit disorder 138 6.4Impulse control disorder 123 11.8Paranoid PD 69 3.2Schizoid PD 18 0.8Schizotypal PD 10 0.5Histrionic PD 17 0.8Borderline PD 204 9.5Antisocial PD 41 1.9Narcissistic PD 40 1.9Dependent PD 39 1.8Obsessive-compulsive PD 147 6.8Avoidant PD 285 13.2

Individuals could be given more than 1 diagnosis.

942 M. Zimmerman et al. / Comprehensive Psychiatry 53 (2012) 940–945

Based on the results of the SCID/SADS and SIDP-IVinterviews, the GAF was rated.

The SIDP-IV focuses on the individual's “usual self” overthe past 5 years. Each DSM-IV criterion is rated 0 (criterionnot present), 1 (subthreshold; some evidence of trait but notsufficiently pervasive or severe to be considered present), 2(criterion present; clearly evident for the last 5 years at least50% of the time), or 3 (criterion strongly present). Consistentwith the recommended scoring guidelines of the SIDP-IV,each criterion rated 2 or 3 was counted as present, whereasratings of 0 or 1 indicated that the criterion was absent. Thequestions on the SIDP-IV are grouped thematically intosimilar content areas, such as interpersonal relationships,interests and activities, social conformity, and emotions.Such an interview is less prone to halo effects in whichratings of individual criteria are influenced by how close theindividual is to meeting criteria for a particular disorder.

Diagnostic raters included PhD-level psychologists andresearch assistants with college degrees in the social orbiological sciences. Research assistants received 3 to 4months of training during which they observed at least 20interviews, and they were observed and supervised in theiradministration of more than 20 evaluations. Psychologistsonly observed 5 interviews, and they were observed andsupervised in their administration of 15 to 20 evaluations.During the course of the training the senior author met witheach rater to review the interpretation of every item on the

SCID. In addition during the training, every interview wasreviewed on an item-by-item basis by the senior rater whoobserved the evaluation and by the senior author whoreviewed the case with the interviewer. At the end of thetraining period, the raters were required to demonstrate exactor near exact agreement with a senior diagnostician on 5consecutive evaluations. Throughout the Methods to Im-prove Diagnostic Assessment and Services project, ongoingsupervision of the raters consisted of weekly diagnostic caseconferences involving all members of the team. In addition,every case was reviewed by the senior author.

Reliability of PD diagnoses was examined in 47 patients.A joint-interview design was used, in which 1 rater observedanother conducting the interview, and both raters indepen-dently made their ratings. The reliabilities of any PD (κ =0.90) and any cluster A (κ = 0.79), B (κ = 0.79), or C PD (κ =0.93) were good to excellent. Too few patients werediagnosed with individual PDs to calculate κ coefficients.However, intraclass correlation coefficients (ICCs) ofcriterion count dimensional scores were high (paranoid:ICC, 0.92; schizoid: ICC, 0.95; schizotypal: ICC, 0.82;antisocial: ICC, 0.95; borderline: ICC, 0.95; histrionic: ICC,0.91; narcissistic: ICC, 0.91; avoidant: ICC, 0.96; depen-dent: ICC, 0.97; and obsessive-compulsive: ICC, 0.90).

2.1. Data analysis

The Work Group for DSM-5 recommended the adoptionof dimensional ratings of PDs; therefore, we computedcorrelation coefficients between a dimensional representa-tion of each PD (using criterion counts) and 7 indices ofpsychosocial morbidity (no. of current Axis I disorders,lifetime psychiatric hospitalizations, lifetime suicide at-tempts, suicidal ideation at the time of the evaluation, GAFratings, amount of time unemployed during the past 5 yearsbecause of psychiatric reasons, and best level of socialfunctioning during the past 5 years). Because we werecomputing 70 correlation coefficients, the significance levelwas set at P b .01. After the univariate correlation analyses,we conducted a stepwise multiple regression analysis todetermine which of the PD dimensions were independentlyassociated with each measure of psychosocial morbidity.Only those PD dimensions that were significant in theunivariate analyses were included in the regression analysis.

3. Results

A little less than one-third of the patients were diagnosedwith 1 of the 10 DSM-IV PDs (29.3%, n = 629). The data inTable 2 show that avoidant, borderline, and obsessive-compulsive PDs were the most frequent PDs in the sample.

Paranoid, schizoid, schizotypal, borderline, antisocial,and dependent PD dimensional scores were significantlyassociated with all 7 indices of psychosocial morbidity, andavoidant PD was associated with 6 of the 7 indices ofseverity (Table 3). All 10 PDs were significantly associated

Table 3Correlation between DSM-IV PD dimensions and indicators of psychosocial morbidity in 2150 psychiatric outpatients

DSM-IV PD No. of currentAxis I disorders

No. lifetimepsychiatrichospitalizations

No. lifetimesuicide attempts

Currentsuicidalideation

GAF Time unemployed forthe past 5 yearsa

Best social functioningfor the past 5 years

Paranoid 0.31⁎ 0.10⁎ 0.10⁎ 0.20⁎ −0.26⁎ 0.14⁎ 0.22⁎

Schizoid 0.17⁎ 0.13⁎ 0.06⁎ 0.17⁎ −0.23⁎ 0.14⁎ 0.44⁎

Schizotypal 0.24⁎ 0.12⁎ 0.08⁎ 0.18⁎ −0.30⁎ 0.21⁎ 0.37⁎

Histrionic 0.18⁎ 0.10⁎ 0.05 0.16⁎ −0.17⁎ 0.11⁎ 0.03Borderline 0.44⁎ 0.21⁎ 0.18⁎ 0.31⁎ −0.36⁎ 0.27⁎ 0.23⁎

Antisocial 0.18⁎ 0.16⁎ 0.09⁎ 0.10⁎ −0.20⁎ 0.20⁎ 0.09⁎

Narcissistic 0.18⁎ 0.04 0.02 0.13⁎ −0.15⁎ 0.09⁎ 0.14⁎

Dependent 0.32⁎ 0.06⁎ 0.08⁎ 0.22⁎ −0.22⁎ 0.14⁎ 0.18⁎

Obsessive-compulsive 0.22⁎ 0.01 0.02 0.13⁎ −0.13⁎ 0.01 0.16⁎

Avoidant 0.46⁎ 0.04 0.08⁎ 0.24⁎ −0.28⁎ 0.18⁎ 0.36⁎

⁎ P b .01.a Individuals not expected to work (ie, retired, student, housewife, and physically ill) were excluded from this analysis (final n = 1927).

943M. Zimmerman et al. / Comprehensive Psychiatry 53 (2012) 940–945

with a higher number of Axis I disorders, lower GAF scores,and current suicidal ideation. All but obsessive-compulsivePD were significantly associated with time unemployed inthe past 5 years, and all but histrionic PD were associatedwith impaired social functioning in the past 5 years. For the 6PDs proposed for retention in DSM-5, 36 (85.7%) of the 42correlations were significant, whereas for the 4 PDsproposed for deletion, 26 (92.9%) of the 28 correlationswere significant.

Results from the stepwise regression analyses are shownin Table 4. Borderline PD was a significant predictor of all7 indices of psychosocial morbidity, and it was the only PDthat was independently associated with the lifetime numberof suicide attempts. Schizoid and avoidant PDs weresignificant independent predictors of 5 morbidity indices.Paranoid, histrionic, and narcissistic PDs were notindependently associated with any of the measures ofpsychosocial morbidity, and obsessive-compulsive PD wasindependently associated with only 1 measure of morbidity.For the 6 PDs proposed for retention in DSM-5, 19(45.2%) of the 42 β coefficients were significant, whereasfor the 4 PDs proposed for deletion, 7 (25.0%) of the 28β coefficients were significant.

4. Discussion

Our results provide only modest evidence to suggest thatthe disorders recommended for retention in DSM-5 (antiso-cial, avoidant, borderline, narcissistic, obsessive-compulsive,and schizotypal) were more strongly associated with indicesof psychosocial morbidity than disorders recommended fordeletion (schizoid, paranoid, histrionic, and dependent).Instead, we found that all PDs were associated with greaterpsychosocial morbidity across several domains. This wasparticularly true for borderline PD, which was significantlyrelated to all 7 psychosocial morbidity measures in allanalyses. In the zero-order correlation analyses, 6 disorders

were significantly associated with all 7 measures ofmorbidity, 3 of which (paranoid, schizoid, and dependent)have been recommended for deletion in DSM-5. In contrast,narcissistic and obsessive-compulsive PD criteria yielded thefewest significant results in the zero-order analyses, both ofwhich were recommended for retention. In the regressionanalyses, borderline, schizoid, and avoidant PDs were themost frequently significantly associated with morbidity.Paranoid, histrionic, narcissistic, and obsessive-compulsivePDs were independently associated with none or only 1 of the7 measures of morbidity. These 4 disorders represent 2recommended for retention in DSM-5 and 2 recommendedfor deletion.

Looking at the data in another way, more of the zero-ordercorrelations were significant for the deleted than the retainedPDs (93% vs 86%), whereas in the multiple regressionanalyses, a higher percentage of the β coefficients weresignificant for the retained PDs (45% vs 25%). This can beinterpreted as providing modest support for the DSM-5proposal, although these findings do not show robustdifferences between the retained vs deleted groups of disorders.

We are aware of only 2 other studies of the associationbetween each of the DSM PDs and measures of psychosocialmorbidity in psychiatric patients. These studies reportalternately that all DSM-III-R PDs except histrionic andobsessive-compulsive PDs were significantly associatedwith lower GAF scores [19] or that the total number of PDdiagnoses—and, to a lesser extent, depressive PD—is theonly a significant correlate of quality of life [18,20].Although informative, these studies were limited either bysample size or by limited analyses of outcome variables.

Three other studies of nonpatient samples have examinedthe correlates of a broad range of individual PDs. Cramer et al[26] examined the relationship between quality of life andDSM-III-R PDs in a random sample of residents of Oslo,Norway. Avoidant, schizotypal, paranoid, and schizoid traitswere the most strongly associated with global quality of lifeand also related to the highest overall number of quality of life

Table4

Stepw

iseregression

forDSM-IV

PD

criteriacountsandindicators

ofpsychosocial

morbidity

in2150

psychiatricoutpatients

PD

No.

ofcurrent

AxisIdisorders

No.

oflifetim

epsychiatric

hospitalizations

No.

oflifetim

esuicideattempts

Current

suicidal

ideatio

nGAF

Tim

eun

employed

forthepast5

yearsa

Bestsocial

functio

ning

βSE

βSE

βSE

βSE

βSE

βSE

βSE

Paranoid

0.05

0.04

−0.03

0.03

0.04

0.08

0.03

0.03

−0.14

0.23

−0.10

0.05

−0.01

0.03

Schizoid

−0.05

0.05

0.15

0.04

0.10

0.11

0.14

0.04

−1.00

0.31

0.08

0.07

0.44

0.04

Schizotyp

al0.07

0.05

0.00

0.04

−0.00

0.11

0.03

0.04

−1.46

0.32

0.31

0.07

0.20

0.04

Histrionic

0.03

0.03

0.0.02

0.03

−0.12

0.08

0.06

0.03

−0.41

0.24

0.01

0.05

−−

Borderline

0.22

0.02

0.10

0.02

0.23

0.05

0.14

0.02

−1.16

0.13

0.20

0.03

0.04

0.02

Antisocial

0.07

0.04

0.12

0.03

0.12

0.08

−0.05

0.03

−0.76

0.25

0.27

0.05

−0.00

0.03

Narcissistic

−0.06

0.03

––

––

−0.01

0.03

0.44

0.20

−0.10

0.04

0.05

0.02

Dependent

0.08

0.03

−0.01

0.02

0.04

0.07

0.06

0.03

−0.43

0.20

0.05

0.04

0.00

0.02

Obsessive-com

pulsive

0.07

0.02

––

––

0.00

0.02

−0.01

0.16

––

0.00

0.02

Avo

idant

0.27

0.02

––

0.04

0.04

0.07

0.02

−0.60

0.13

0.08

0.03

0.12

0.01

Dashedlin

esindicate

that

thePD

dimension

was

notsignificantin

thecorrelationanalysisand,

therefore,notincluded

intheregression

analysis.β

Coefficientsareunstandardized.P≤

.01appearsin

boldface.

aIndividualsnotexpected

towork(ie,

retired,student,housew

ife,

andphysically

ill)wereexcluded

from

thisanalysis(final

n=1927).

944 M. Zimmerman et al. / Comprehensive Psychiatry 53 (2012) 940–945

indices such as poor self-realization and neighborhood quality.Grant et al [16] examined the association between DSM-IVPDs and impairment in the National Epidemiologic Survey onAlcohol and Related Conditions. The first phase of theNational Epidemiologic Survey on Alcohol and RelatedConditions study assessed 7 PDs (avoidant dependent,obsessive-compulsive, paranoid, schizoid, histrionic, andantisocial). The authors reported that after controlling forage, other PDs, substance use, anxiety and mood disorders, allPDs except histrionic and obsessive-compulsive PDs weresignificantly associated with the 3 mental disability scores ofthe Short-Form 12-ItemHealth Survey.Obsessive-compulsivePD was significantly associated with 1 of the 3 disabilityscores, and histrionic PD was not associated with any of themeasures of disability. Ullrich et al [27] examined theassociation between DSM-IV PDs and life success in acommunity sample of 48-year-old men. In their regressionanalyses, lower levels of status and wealth were significantlyassociated with schizoid, dependent, and antisocial traits. Ofinterest, higher levels of status and wealth were positivelyassociated with obsessive-compulsive and narcissistic traits.Successful intimate relationships were negatively associatedwith avoidant, schizoid, and borderline traits.

Taken together, the results of these 5 studies, along withthe results of the present study, do not provide clear evidencefor the preferential retention in DSM-5 of some PDs overothers based on their association with indices of psychoso-cial morbidity. The DSM-5 PD Work Group stated thatdecisions to delete or retain disorders were based primarilyon reduction of comorbidity and clinically relevant correlatesof disorders. Our data address only concurrent (no. of currentAxis I disorders, suicidal ideation at the time of theevaluation, and GAF ratings) and antecedent (lifetimepsychiatric hospitalizations, lifetime suicide attempts,amount of time unemployed during the past 5 years becauseof psychiatric reasons, and best level of social functioningduring the past 5 years) variables. Results from this studydifferentiate borderline, schizoid, and avoidant PDs as themost significantly associated with morbidity and paranoid,histrionic, narcissistic, and obsessive-compulsive PDs as theleast. Importantly, we only tested 7 possible correlates of 10much broader categories proposed by Kendler et al [10].Future research could better explicate the full range ofantecedent, concurrent, and predictive validators for all PDs.

This present study had several limitations. First, thesample was predominantly female and white, and it wasdrawn from a single clinical practice. The results shouldbe replicated in a clinical sample with more diversedemographic characteristics. Second, the same interviewersassessed all variables, including both Axes I and IIinterviews. Although it is possible that interviewers werebiased because the ratings were not blind, it is unlikely tohave affected the current study because the data collectionwas not specifically designed to address the questionsaddressed in the current manuscript. Third, because comor-bidity is representative of the current PD diagnostic system,

945M. Zimmerman et al. / Comprehensive Psychiatry 53 (2012) 940–945

we elected to retain comorbidity across Axis II in the currentsample. Further research could assess the relation betweenPD diagnosis and psychosocial morbidity in samples with asingle or hierarchically defined “primary” PD. Finally,criteria counts were used to represent dimensional scoringof PDs, which reflects an adaptation (as opposed to a directtranslation) of the proposal from the DSM-5 committee.However, this limitation is not fully addressable until theproposed changes for DSM-5 are finalized. The present studyalso has several notable strengths, including a large samplesize representative of the full range of psychopathology, thesimultaneous collection of multiple types of data, and the useof highly trained diagnostic interviewers to reliably admin-ister a semistructured diagnostic interview.

References

[1] Clark LA. Assessment and diagnosis of personality disorder: perennialissues and an emerging reconceptualization. Annu Rev Psychol2007;58:227-57.

[2] Trull TJ, Durrett CA. Categorical and dimensional models ofpersonality disorder. Annu Rev Clin Psychol 2005;1:355-80.

[3] Blashfield RK, Breen MJ. Face validity of the DSM-III-R personalitydisorders. Am J Psychiatry 1989;146:1575-9.

[4] Shedler J, Westen D. Refining personality disorder diagnosis:integrating science and practice. Am J Psychiatry 2004;161:1350-65.

[5] Westen D, Shedler J. Revising and assessing axis II, part I: developinga clinically and empirically valid assessment method. Am J Psychiatry1999;156:258-72.

[6] Tyrer P. What's wrong with DSM-III personality disorders? J PersonalDisord 1988;2:281-91.

[7] Paris J. Outcome and epidemiological research on personality disorders:implications for classification. J Pers Disord 2005;19:557-62 [discussion594-56].

[8] Grove WM, Tellegen A. Problems in the classification of personalitydisorders. J Personal Disord 1991;5:31-41.

[9] American Psychiatric Association. Personality and personalitydisorders. American Psychiatric Association. Accessed on July26, 2010 from http://www.dsm5.org/ProposedRevisions/PagesPersonalityandPersonalityDisorders.aspx.

[10] Kendler KS, Kupfer DJ, Narrow WE, Phillips KA, Fawcett J.Guidelines for making changes to DSM-V, 2009. accessed on 10/14/10 from http://www.dsm5.org.

[11] Skodol AE, Bender DS. The future of personality disorders in DSM-V?Am J Psychiatry 2009;166:388-91.

[12] Bornstein R. Reconceptualizing personality pathology in DSM-5:limitations in evidence for eliminating dependent personalitydisorder and other DSM-IV syndromes. J Personal Disord2011;25:235-47.

[13] Clarkin JF, Huprich SK. Do DSM-5 personality disorder proposalsmeet criteria for clinical utility? J Pers Disord 2011;25:192-205.

[14] Miller JD, Widiger TA, Campbell WK. Narcissistic personalitydisorder and the DSM-V. J Abnorm Psychol 2010;119:640-9.

[15] Zimmerman M. Is there adequate empirical justification for radicallyrevising the personality disorders section for DSM-5? Pers DisordTheory Res Pract (in press).

[16] Grant BF, Hasin DS, Stinson FS, Dawson DA, Chou SP, RuanWJ, et al. Prevalence, correlates, and disability of personalitydisorders in the United States: results from the national epidemiologicsurvey on alcohol and related conditions. J Clin Psychiatry2004;65:948-58.

[17] Stuart S, Pfohl B, Battaglia M, Bellodi L, Grove W, Cadoret R. Thecooccurrence of DSM-III-R personality disorders. J Pers Disord1998;12:302-15.

[18] Bender D, Dolan R, Skodol A, Sanislow C, Dyck I, McGlashan T,et al. Treatment utilization by patients with personality disorders.Am J Psychiatry 2001;158:295-302.

[19] Soeteman DI, Hakkaart-van Roijen L, Verheul R, Busschbach JJ. Theeconomic burden of personality disorders in mental health care. J ClinPsychiatry 2008;69:259-65.

[20] Nakao K, Gunderson J, Phillips K, Tanaka N, Yorifuji K, Takaishi J,et al. Functional impairment in personality disorders. J PersonalDisord 1992;6:24-33.

[21] Soeteman DI, Verheul R, Busschbach JJ. The burden of disease inpersonality disorders: diagnosis-specific quality of life. J Pers Disord2008;22:259-68.

[22] Pfohl B, Blum N, Zimmerman M. Structured interview for DSM-IVpersonality. Washington, DC: American Psychiatric Press, Inc.;1997.

[23] First MB, Spitzer RL, Williams JBW, Gibbon M. Structured ClinicalInterview for DSM-IV (SCID). Washington, DC: American Psychi-atric Association; 1997.

[24] Zimmerman M, Mattia JI. Psychiatric diagnosis in clinicalpractice: is comorbidity being missed? Compr Psychiatry1999;40:182-91.

[25] Endicott J, Spitzer RL. A diagnostic interview: the Schedule forAffective Disorders and Schizophrenia. Arch Gen Psychiatry1978;35:837-44.

[26] Cramer V, Torgersen S, Kringlen E. Socio-demographic conditions,subjective somatic health, Axis I disorders and personality disorders inthe common population: the relationship to quality of life. J PersDisord 2007;21:552-67.

[27] Ullrich S, Farrington DP, Coid JW. Dimensions of DSM-IVpersonality disorders and life-success. J Pers Disord 2007;21:657-63.