construct validity and prevalence rate … normal peers, and the bpd diagnosis was associated with...

Journal of Personality Disorders, 23(5), 494-513. 2009© 2009 The Gullford Press

CONSTRUCT VALIDITY AND PREVALENCE RATEOF BORDERLINE PERSONALITY DISORDERAMONG CHINESE ADOLESCENTS

See-Wai Leung. MPhil, and Freedom Leung. PhD

The construct of borderline personality disorder (BPD) among adoles-cents is a controversiai topic. No research has examined the BPD con-struct among Chinese adolescents because tlie Chinese Classiñcationof Mental Disorders-IIl rejects BPD as a valid diagnostic category. Thepresent study explored the construct validity and prevalence of BPDamong Chinese adolescents in Hong Kong. A total of 4.110 high schoolstudents completed the McLean Screening Instrument for borderlinepersonality disorder (MSI-BPD) and other measures assessing variousBPD traits twice over a one-year period. DSM-IV-TR BPD criteria setas measured by tbe MSI-BPD demonstrated good internal consistency,concurrent validity and test-retest reliability. Confirmatory factor anal-ysts of the MSI-BPD revealed four theoretically meaningful factors,namely affect dysregulation, impulsivity, interpersonal disturbances,and self/cognitive disturbances. Prevalence rate of BPD. according to astringent simulated diagnostic procedure used in this study, was esti-mated to be 2% among Chinese adolescents in Hong Kong. Findingssupport that BPD is a valid elinical construct among Chinese adoles-cents. More research on BPD among the Chinese population is war-ranted.

Borderline personality disorder (BPD) Is characterized by mood and im-pulse dysregulation, chaotic interpersonal relationships, and self-imagedisturbances. The prevalence rate of BPD has been found to be about1-2% in general population (Torgersen. Kringlen, & Cramer, 2001; Widiger& Weissman, 1991). According to DSM-IV-TR (American Psychiatric As-sociation, 2000), diagnosis of personality disorders can be given to ado-lescents. However, since some borderline features such as emotional in-stability and identity confusion are quite common among adolescents,differentiating genuine BPD symptoms from normative personality fea-tures during adolescence is no easy task. Some researchers were therefore

From The Chinese University of Hong Kong.This research was suppiorted by the Direct Research Grant of the Ciilnese Universi^ of HongKong.Address correspondence to Freedom Leung. Phi,D., Department ofPsyciioIogy, Chinese Uni-versity of Hong Kong, Shatin, Hong Kong: E-malÍ:fyÍdeung@psy,cuhÍc,edu,hk

494

BORDERLINE FEATURES IN ADOLESCENTS 495

skeptical about the applicability of the BPD diagnosis in adolescents (Kut-cher & Korenblum. 1992; Marton et ai., 1989). Others (Paris. 2003; Zanar-ini. Frankenburg. Khera. & Bleichmar. 2001). on the other hand, arguedthat adolescence is a critical period when symptoms of personality disor-ders first emerge and should be studied accordingly.

CONSTRUCT VALIDITY OF BPD AMONG ADOLESCENTSSeveral studies have examined the construct validity of BPD among ado-lescents. Internal consistencies of DSM-IIl-R BPD criteria as assessed bypersonality disorder examination (PDE) were found to be .76 In adolescentpsychiatric inpatients. with mean inter-item correlations of .28 (Becker etal.. 1999). Studies examining concurrent validity revealed that adolescentswith BPD reported significantly more problems at school or at work thantheir normal peers, and the BPD diagnosis was associated with more sui-cide attempts, anxious/depressed mood, and internalizing and externaliz-ing behavior problems (Bernstein et al.. 1993: Westen. Shedler. Durett,Glass, & Martens, 2003).

In a factor analytic study of the DSM-III-R BPD criteria set assessed byPDE among hospitalized adolescents. Becker, McGiashan, and Grilo(2006) found four subcomponents of BPD symptoms. They were (1) irrita-bility (affective instability, inappropriate anger, identity disturbances). (2)impulsivity (impulsive acts, identity disturbances). (3) poorly modulatedrelationships (unstable relationships, abandonment fears), and (4) self-negation (self-mutilating or suicidal behaviors, boredom, and emptiness).Even though this four-factor solution was based on DSM-III-R BPD criteriaset. which did not include the criterion of transient dissociative or psy-chotic symptoms, it resembles the four-factor model proposed by Lieb. Za-narini, Schmahl. Linehan, and Bohus (2004). According to Lieb et al., thecore pathology of BPD can be conceptually divided into four major do-mains: (1) affective disturbances (affective instability, inappropriate anger,chronic emptiness): (2) impulsivity (impulsive acts, self-mutila ting, or sui-cidal behaviors): (3) unstable relationships (unstable relationships, fear ofabandonment): and (4) disturbed cognition (identity disturbances, tran-sient dissociative/psychotic symptoms).

The lack of diagnostic stability over time among adolescents creates an-other challenge to the application of the BPD diagnosis to adolescents.Bernstein et al. (1993) reported that only 29% and 24% adolescents whoreceived moderate or severe BPD diagnosis in their study persisted to havethe diagnosis two years later. Other researchers (Meijer. Goedhart. & Tref-fers, 1998), however, indicated that while the BPD diagnosis may changeover time, different BPD symptoms differ in terms of stability. For example,self-mutilation, suicidal gesture, and transient dissociative/psychoticsymptoms were less stable, persisting in only 30% of adolescent inpatientswho endorsed them at baseline. Symptoms like mood lability and angerdyscontrol were more stable, persisting in 70% of the patients. Perhaps

496 • LEUNG AND LELÍNG

individuals vulnerable for BPD may display stable trait of mood dysregula-tion. but they show acting-out behavior, self-mutilation or suicidal ges-ture, or transient dissociative features only in time of extreme emotionaldistress (McGlashan et al., 2005).

PREVALENCE OF BPD IN ADOLESCENTSThe prevalence of BPD in adolescents is uncertain as few systematic epide-miological studies have been conducted. In a U.S. community sample ofadolescents (Bernstein et al.. 1993). participants (n = 733) completed abattery of diagnostic instruments including Personality Diagnostic Ques-tionnaire (Hyler et al., 1988) and Quality of Life Interview (Cohen, 1986).Diagnostic scales for personality disorders were retrospectively developedby selecting items from the battery of instruments. Among the 17 itemsselected to assess the DSM-III-R BPD criteria, none matched the criterionof identity disturbances. In this study, 7.8% of subjects was classified asmoderate BPD and 3.0% severe BPD. No gender difference was observed.Findings from this study should be interpreted with caution since the di-agnostic items were derived from scales not originally intended to measureBPD features.

In a French study (Chabrol. Montovany, Chouicha. Callahan, & Mullet,2001), a sample of 1,363 high school students completed the self-reportScreening Test for Comorbid Personality Disorders (STCPD: Dowson,1992). Among them, 107 volunteered to be interviewed using the RevisedDiagnostic Interview for Borderlines (DIB-R: Zanarini. Gunderson, Frank-cnburg. & Chauncey, 1989). A regression analysis was conducted usingthe DIB-R score as criterion and STCPD score as predictor. The BPD cut-oiT point for STCPD was then derived by substituting the DIB-R clinicalcut-off point into the regression equation. According to the derived STCPDcut-off point, about 10% of boys and 18% of girls were classified as BPD.A major limitation of this study is that the BPD diagnostic cut-off forSTCPD was inferred from the smaller volunteered sample to the larger ran-dom sample, which may not be comparable.

Findings of these two available studies reported prevalence rates of BPDamong adolescents much higher than the 1-2% observed among generalpopulation (Torgersen et al.. 2001; Widiger & Weissman, 1991), Due tomethodological limitations of these studies, it is difficult to assess to whatextent these estimated figures have been confounded by false positives.More stringent epidemiological studies in the future are clearly needed.

BPD AMONG THE CHINESE POPULATIONThe BPD construct met strong resistance in the psychiatric profession inChina. Committee members of the Chinese Classification of Mental Disor-ders-III (CCMD-III: Chinese Psychiatry Association, 2001) argued that BPDis a vague construct and some of its diagnostic features (e,g,, fear of aban-


donment) may not be appropriate culturally when used in China. As aresult, the CCMD-III does not include the BPD diagnosis in its nomencla-ture, and consequently most clinicians and researchers in China are notfamiliar with this clinical construct (Zhong & Leung, 2007). Preliminaryempirical studies examining the construct validity of BFD. however, sug-gested that the clinical syndrome of BPD as observed among Chinese adultpsychiatric patients was comparable to that reporied in the West (Leung.Cheung. & Cheung, 2004). The DSM-IV-TR BPD criteria set, as measuredby the Chinese Personality Disorder Inventory, showed acceptable internalconsistency (a= .71) and good concurrent validity among Chinese adultpsychiatric patients (Leung et al.. 2004: Leung, Chan, & Cheung, 2007),These findings suggest that BPD patients do exist in China and systematicresearch to study the characteristics of this population is clearly needed.This is an imporiant clinical topic because the prevalence of BPD has beenestimated to be at 1-2% among general population in the West (Torgersenet al.. 2001; Widiger & Weissman, 1991). and if this prevalence figure isgeneralizable to China, a country with 1.3 billion people, it means 13 to26 million Chinese could be suffering from BPD without ever been properlydiagnosed or treated.

PURPOSES OF THIS STUDYPrevious studies in the West indicated that the BPD construct shows ac-ceptable internal consistency and concurrent validity among adolescents.Studies conducted with Chinese adult psychiatric patients also providepreliminary empirical suppori to the construct validity of BPD among theChinese. To our knowledge, no study has examined the BPD constructamong Chinese adolescents. The present study explored the construct va-lidity and prevalence of BPD in a large sample of Chinese adolescents inHong Kong over a one-year period. We also conducted confirmatory factoranalyses (CFA) to examine the factorial structure of the DSM-IV-TR BPDcriteria set as assessed by McLean Screening Instrument for BorderlinePersonality Disorder (MSI-BPD: Zanarini et al., 2003) among our adoles-cent sample.

Several models were available to guide our analyses (Lieb et al.. 2004:Sanislow et al., 2002). Lieb et al.'s four-factor model suggested that BPDsymptoms can be conceptually divided into four major domains: (1) af-fective disturbances (affective instability, inappropriate anger, chronicemptiness): (2) impulsivity (impulsive acts, se If-mutilating or suicidal be-haviors): (3) unstable relationships (unstable relationships, fear of aban-donment); and (4) disturbed cognition (identity disturbances, transientdissociative/psychotic symptoms).

In a CFA study of DSM-fV-TR BPD symptoms among adult psychiatricpaüents, Sanislow et al, (2002) reported that the BPD symptoms can beexplained both by a single-factor model as well as a three-factor model.Detailed analysis suggested that the three-factor model provided a better

498 LEUNG AND LEUNG

fit. These factors included (1) affective dysregulation (afTective instability,inappropriate anger, abandonment fears). (2) behavioral dysregulation(impulsive acts, self-mutilating or suicidal gestures) and (3) disturbed re-latedness (unstable relationships, identity disturbance, chronic empti-ness, transient dissociative/paranoid symptoms). Sanislow et al. suggestedthat these three factors describe the physiological traits, symptomatic be-haviors, and personality deficits of BPD patients. Sanislow et al., however,warned that abandonment fear, which loaded unexpectedly with the affec-tive dysregulation factor, could be a statistical artifact because of theitem's relatively low endorsement rate.

In the present study, we used CFA to compared five different models.They included: (1) Lieb et al.'s (2004} four-factor model: (2) a modified four-factor model in which we assumed chronic emptiness would cluster betterwith cognitive rather than affective symptoms as we speculated emptinessmay refiect dissociative symptoms (Sanislow et al.. 2002): (3) Sanislow etal.'s three-factor model; (4) a modified three-factor model in which we as-signed abandonment fear to disturbed relatedness rather than affectivedysregulation factor; and (5) Sanislow et al.'s (2002) one-factor model.Moreover, since the paranoid ideation item of the MSI-BPD (i.e.. I haveoften been distrustful of other people when emotionally distressed) clearlyassesses interpersonal distrust, we hypothesized that it would cluster withimstable interpersonal relationship and fear of abandonment to form thefactor of interpersonal disturbances.

METHODSUBJECTS ,- .

Participants came from five high schools in Hong Kong. At Timel. a totalof 5.224 adolescents, aged between 12 and 20 years (M 14.6, SD= 1.80}were tested. Among them. 68.7% {N= 3.318) were females and 31.3% [N =1.906) were males. At Time 2. a total of 5.461 adolescents (62.7% femalesand 37.3% males), aged between 12 and 20 years (M = 14.7. SD= 1.99).were tested. Among the Time 1 sample. 4. HO participants (63.6% femalesand 38.4% males) were successfully retested at Time 2. Attrition of samplewas mainly due to graduation of senior students.

- • *

MEASURESMSI-BPD. The Chinese version of the McLean Screening Instrument for

Borderline Personality Disorder (MSI-BPD; Zanarini et al.. 2003) was usedto measure the construct of BPD (Wang. Leung, & Zhong. 2008). In theMSI-BPD. each BPD diagnostic criterion was assessed by one item, withthe exception of transient psychotic feature which was assessed by twoseparate items. According to Zanarini et al.. MSI-BPD had adequate one-week test-retest reliability (r= .72). good internal consistency (a = .74).


and item-total correlation (ranged between .45 and ,63). In this study, par-ticipants rated their level of symptom severity on a four-point scale, i.e., 1(very disagree); 2 (disagree): 3 (agree}; 4 (very agree). Summation of theten item ratings gives a dimensional score. When categorical response wasrequired (i.e., determining item endorsement), a rating of 1 or 2 was con-verted into a "no" response and a rating of 3 or 4 into a "yes" response. Thediagnostic cutoff of the MSI-BPD is seven. Subjects who endorsed seven ormore items could be considered as a BPD case.

AJfective Instability. Affective instability was measured by "Reactivity toSituations" subscale of the Mood Survey (Underwood & Froming, 1980). Ithad adequate 7-week test-retest reliability (r = .83) and concurrent validitywith emotionality (r^.69). It consists of 7 items, e.g.. Sometimes mymoods swing back and forth very rapidly. Responses were made on a four-point scale ranging from 1 (very disagree) to 4 (very agree). Higher scoresindicate more labile affect. This scale had a Cronbach's alpha of .92 in thisstudy.

Impulsive Behavior. A ten-item impulsivity scale extracted and modifiedfrom the Revised Diagnostic Interview for Borderlines (DIB-R; Zanarini etal., 1989) was used to assess impulsive behavior. Participants rated howfrequent they displayed various behaviors (e.g., binge eating, verbal out-burst) over the past two years on a 4-point scale from 1 (never) to 4 (six ormore times). Higher scores indicate more impulsive behaviors. This scalehad a Cronbach's alpha of .76 in this study.

Instability of Self. Instability of sense of self was measured by iive itemsmodified from the Rosenberg's Stability of Self Scale (Alaska & Olweus,1986). Ratings were made on a 4-point scale from 1 (very disagree) to 4(very agree). Higher scores indicate a more unstable sense of self. Thisscale had a Cronbach's alpha of ,90 in this study.

SelJ-Deßation Proneness. The Self-Deflation Proneness scale (SDPS;Poon. 2003), a 10-item instrument, was used to assess the fragility of one'ssense of self. A sample item is "Even minor setbacks can shatter my self-confidence." The SDPS had excellent internal consistency (a = .94) andgood one-year test-retest reliability (r^.63). Items were rated from 1(strongly agree) to 5 (strongly disagree). Higher scores reflect a more fragilesense of self. This scale had a Cronbach's alpha of .95 in this study.

Disturbed Inter^yersonal Relationships. Disturbed interpersonal relation-ships were assessed by five items extracted and modified from the DIB-R(Zanarini et al., 1989). A sample item was "1 repeatedly worried about be-ing abandoned by someone close to me." Items were rated from 1 (verydisagree) to 4 (very agree). Higher scores reflect more problematic interper-sonal relationships. This scale had a Cronbach's alpha of .78 in this study.

Depression. The Chinese version of the Depression Subscale of Symp-toms Checklist-90 (SCL-90; Derogatis, Lipman. & Covi, 1973) was used toassess depressive symptoms. The original scale had 13 items. One item,"l-oss of sexual interest or pleasure," was deleted in this study as schoolauthorities considered it inappropriate for younger adolescents. Responses

500 • LEUNG AND LEUNG

were made on a five-point scale from 1 (never) to 5 (always). This scale hada Cronbach's alpha of .92 in this study.

With the exception of the MSI-BPD. Self-Deflation Proneness Scale, andSCL-90-Depression Scale for which Chinese versions were available, allother measures were first translated into Chinese by the first author andthen back translated independently by two bilingual researchers. Fidelitywas judged to be very high to place confidence in the translation as awhole. The whole questionnaire package was then reviewed and modifiedby the second author to make sure all items would be culturally relevantand understandable by young Chinese adolescents in Hong Kong.

Simulated Diagnostic Methad In this study, we were not able to conductdiagnostic interview with our subjects as we originally planned becauseschool authorities of the participating schools expressed serious concernsabout the possible labeling effect and insisted on strict principle of ano-nymity. To deal with this constraint, we developed a stringent simulateddiagnostic procedure to assess BPD diagnosis during the Time 2 testing.According to this procedure, we selected three highly relevant items to as-sess each of the BPD diagnostic criteria (see Appendix 1). Some of theseitems were extracted from measures we used to assess relevant BPD fea-tures. Other items were selected from the DIB-R (Zanarini et al., 1989).Pariicipants had to endorse all three relevant items to be considered asmeeting a specific BPD criterion. For instance, to meet the criterion foridentity disturbance, participants had to endorse the following threeitems: "I am confused about my own self," "My values change quickly."and "My vocational goal changes often." Following the DSM-IV-TR cut-offrule, participants had to meet at least five criteria to be considered as aBPD case. -.

PROCEDURE .

The study required written parental consent for student participation andfollowed standard data collection protocols approved by the Ethics in Hu-man Research Committee of the Chinese University of Hong Kong. Sub-jects completed the questionnaires in classrooms during a 45-minute pe-riod. Students absent from school on the day of testing were administeredquestionnaires later under the supervision of school personnel. The studywas framed as a "Study of Emotion and Mental Health among Adoles-cents." Strict confidentiality of the study was emphasized. To protect theprivacy of the subjects, students were not required to provide full nameson the questionnaires. Instead, they were asked to put down their sur-names and date of birih, and from that a unique ID for each subject (e.g..CHAN 101492) was created for data-matching purpose over repeated test-ing. Because of the cooperation of the school authorities and their strongencouragement for their students to participate in the study, overall stu-dent pariicipation rates were close to 99% for both Time 1 and Time 2testing.


RESULTSPreliminary analyses were first done separately for female and male sub-jects. Since result patterns were comparable, the two samples were com-bined for most analyses.

CONSTRUCT VALIDITY OF BPD AMONG CHINESE ADOLESCENTSInternal Consistency. Table 1 shows the item-total correlations and in-

teritem correlations for MSI-BPD. For Year 1 data, item-total correlationsranged from .48 to .66. Interitem correlations ranged from .25 to .60 (p<0.01), and mean interitem correlation was .39. These findings indicatedthat BPD features as measured by MSI-BPD clustered well together as aclinical syndrome among Chinese adolescents. The highest interitem cor-relation was between inappropriate anger and affective instability (r = .60).Similar results were replicated in the Time 2 data, Cronbach's alpha coeffi-cient was .86 and .87 for Time 1 and Time 2, respectively.

Concurrent Validity. Pearson correlations between MSI-BPD and (1) af-fective instability. (2) impulsive behaviors. (3) instability of self/self-defia-tion proneness, and (4) disturbed interpersonal relationships, the four ma-jor domains of BPD pathology, were .74. .48. .75/.63, and .78 (p < .001)at Time 1 and .74. .51. .66/.62, and .75 (p< .001) at Time 2, respectively.MSI-BPD score was also positively correlated with depression (r= .70 and.72 for Time 1 and Time 2, respectively, p< .001). Partial correlations be-tween MSI-BPD and affective instability, impulsive behaviors, instabilityof self/self-defiation proneness, and disturbed interpersonal relationship,after controlling for depression, were .53. .31, .54/.29, and .62 (p < ,001)at Time 1 and .51, 33, .41/.26, and .56.(p< ,001) at Time 2, respectively.These findings indicated that the relationships between MSI-BPD andother cross-validating measures remain, albeit at an attenuated degree.even when the effect of depressed mood was partialled out.

Factorial Validity. Table 2 presents the CFA results of the MSI-BPD inour samples. The CFAs were run by EQS. Since the data violated the as-sumption of multivariate normality, analyses were rescaled by SatorraBentler modification (Satorra, Chou, & Bentler, 1991). For all multifactormodels, factors were allowed to correlate with each other. For model com-parison, chi-square, which was susceptible to large sample size, was notsuitable for comparing nonnested model. Therefore, CAIC. which adjustedfor sample size and was useful for comparing nonnested model, was exam-ined. Similar to chi-square, lower CAIC indicated better fit.

Results revealed that all models provided acceptable fit indices, withNFL NNFI, and CF! above .90 and RMSEA below .08. Fit estimates of the 1-factor model indicated adequate fit for data from both time points, thoughmarginally acceptable for Time 2 (NFI = .917, CFi = .920, NNFI = .900,RMSEA = ,078 at Time 2). However, the largest, x^ CAIC. RMSEA, and low-est NFI, CFI. and NNFI. together, indicated that the single factor modelwas not as good in terms of model fit when compared to other multifactormodels.

8V

I

apppis01 X

in o

CO CO —

o CN (0^ CO CO

LO - I i n CO NCO CO CO CO CO

• * 03 - * IN CO i nCS CO ' T CS ^ CO

r- 00 o in 01 • *N CO 1* CS CO CO

tn es CO 01 N O)CO ^ ^ co CO

f f CO 00 O> h- CD —CO CO CO CO *}" CO ' t

C C^ CD CD "—* oCO CO CO CO CO '^ ^

ooo —

CDOOCSCOOOíOCOCÍ( D i D c o n ^ œ n m o t D

oc 9-

(Q c 3

i c

u

aSa


TABLE 2. Confirmatory Factor Analyses of DSM-IV-TR BPD Criteria Setas Assessed by MSI-BPD in Chinese Adolescents

ModelTime I data (W = 5.224)

Lieb 4-factorModified Lleb 4-factorSanislow 3-factorModified Sanislow 3-factorOne-factor

Time 2 data (N= 5,461)Lieb 4-factorModified Lieb 4-factorSanislow 3 factorModified Sanislow 3-factorOne factor

572.7299.5471.3341.0812.7

826.5600.9774.1678.1

U77.0

4f

2929323235

2 92 9323235

P

<.OOO!<.OOO1<.OOO1<.OOO1<.OOO1

<.OO01<.O001<.OOO1<.OOO1<.00Ol

CAIC

295.822.7

165.935.6

478.8

548.4322.9467.3371.2841.4

NFI

.953

.975

.961

.972

.933

.942

.958

.946

.952

.917

CFI

.955

.978

.964

.975

.936

.944

.960

.948

.955

.920

NNFI

.931

.965

.949

.964

.918

.913

.938

.927

.936

.900

RMSEA

.060

.043

.052

.043

.066

.072

.061

.066

.061

.078

Comparing the two three-factor models [dj= 32), the modiñed three-factor model yielded a lower chi-square and CAIC (x = 341.0. CAIC =35.6). Goodness of fit indices were also more satisfaetory (NFI - .972. CFI= .975. NNFI = .964, RMSEA = .043}. suggesting that it fits better thanSanislow et al.'s (2002) original three-factor model. In the modilied model,intereorrelations between factors ranged from .76 to .83. All faetor load-ings were above .50 (see Figure 1).

Comparing with Lieb et al.'s (2004) four-factor model, the modified four-factor model yielded a lower chi-square and CAIC (x = 299.5. CAIC = 22.7,df= 29). whereas goodness-of-ñt indices were also more satisfactory (NFI =.975. CFI = .978. NNFI = .965. RMSEA = .043). This suggested that themodified four-factor model fits better than Lieb et al.'s model. In the modi-fied model, intereorrelations between factors ranged from .76 to .97. Allfactor loadings were above .50 (see Figure 2),

.B3f.ei)

( .76(.74| .77(.77| .63(-61)

Affcclivi-iH'iiahiliiy

Angerdyscnnlml hehaviois

Scrf-muUlfltion/

sutcidc

.70(.65) ' .G9(.67}.70(.69),

Chronic Idemiiydisturbance

Dissociât weHymptoms

57(,57).51(,52).66(.65)

\ X --Un.iUhlc

fcMI

Pnrnnoitt

Values in parenthesas are tor the model lasted using the Time 2 daia.

FIGURE 1. Standardized Factor Loadings and Correlations of the Modified 3-factor Model.Values in parentheses are for the model tested using the Time 2 data.

504 LEUNG AND LEUNG

.86(.89)

.B1(.8Î) .781,751 .77(.76) .64(.62)

Affectivejnsinhility

Angerdescontrol

ImpuMvebehaviors

Serf-mutitalion/

suicide

.71[.6e) 71(,66) .71(.7O) -5e(.571 ,5O(.55) .e6(.65)

Chrimitemptiness

Identitydisturbunce

Dissivialivcrelationship

Values in paremtieses asB for itie model lesied using the Time 2 dala. . , •

FIGURE 2. Standardized Factor Loadings and CorrelaUons of the Moiitfled 4-factor ModeLValues in parentheses are for the model tested using the Time 2 data.

Comparing the modified three-factor model and the modified four-factormodel, the latter had lower CAIC and slightly higher NFI, CFI, and NNFI.This suggested that the modified four-factor model provides a slightly bet-ter fit. However, the same value of RMSEA revealed tliat the two modelswere very similar in terms of degree of fitness. Thus, it is open to interpre-tation as to which model is better conceptually. While the modified three-factor model is more parsimonious, the modiñed four-factor model providesa more detailed account of symptom composition of BPD. It is importantto note that even the modified four-factor model seems to provide a betterconceptual organization of the various BPD symptoms, the high correla-tion between self-identity and interpersonal disttirbances (r=.97) indi-cated that these two factors were closely related empirically in our adoles-cent sample. .

PREVALENCE OF BPD AMONG CHINESE ADOLESCENTS

When we used the MSI-BPD cutoff as specified by Zanarini et al. (2003).7.4% females and 4.2% males at Time 1 and 7.7% females and 5.0% malesat Time 2 met the BPD diagnosis. The prevalence rates were significantlyhigher for females in both years (x = 20.83 and 14.51. respectively, dj= 1.p < .001). When the stringent simulated diagnostic procedure was appliedduring the Time 2 testing, only 2.2% females and 1.8% males met the BPDdiagnosis, and no significant gender difference was observed (x'^^O.Sl.dj= l .p>.05).


STABILITY OF BPD FEATURES AS ASSESSED BY MSI-BPD

Among the 4,110 subjects who participated in both Time 1 and Time 2testing over a 12-month period, the total MSI-BPD score at Time 1 wassignificantly correlated with that at Time 2 (r= .60 for females and .51for males, respectively. p< .001). These test-retest correlations indicatethat BPD features among adolescents, when measured as a unitary di-mension, show reasonable degree of stability over a 12-month period.When a categorical approach was used, however, a different pictureemerged. Table 3 shows the percentages of subjects who repeatedly en-dorsed the BPD features as assessed by MSI-BPD over the one-year pe-riod. Out of the participants who scored above the diagnostic cutoff ofMSI-BPD at Time 1, only 36.8% females and 36.2% males scored abovethe diagnostic cutoff again at Time 2. When the endorsement rates ofeach BPD feature was examined individually, different pattern of resultsemerged for adolescent girls and boys. Comparing to girls, boys seem tohave a more unstable symptoms endorsement pattern over time, particu-larly in affective instability (65.2% vs. 48.4%), anger dyscontrol (50.5%vs. 40.4%). impulsive behaviors (43.4% vs. 25.9%). unstable rela-tionships (53.2% vs. 43.4%), and identity disturbances (52.2% vs. 41%).Among the various BPD features, paranoid ideation had the most stableendorsement pattern, with 63.5% girls and 62.1% boys endorsed thatitem repeatedly at both time points. Affective instability was also rela-tively stable, with 65.2% girls and 48.4% boys endorsed the item repeat-edly over time. The BPD features that were less stably endorsed includedimpulsive behaviors, fear of abandonment, and self-mutilating/suicidalbehaviors.

TABLE 3. Symptoms Endoraement Stability Patterns for Girls and BoysGirls Boys

Above MSI-BPD cutoffIndividual Symptoms

Aifective instabilityAnger dyscontrolImpulsive behaviorsUnstable relationshipsAbandonment fearIdentity disturbanceChronic emptinessSelf- m utilatlon/sulcideParanoid ideationDissociative symptom

NS = Not

Scored positiveat both

TI and T2 (%)36.8

65.250.543.453.231.752.246.344.463.549.7

Scored positiveat both

Tl and T2 (%]36.2

48.440.425.943.435.341.045.434.462.146.8

x".007

28.47.82

15.67.26

.7087.43

.0572.05

.214

.585

PNS

<.OO1<.O5<.OO1<.O5

NS<.O5

NSNSNSNS

506 LEUNG AND LEUNG

DISCUSSIONCONSTRUCT VALIDITY OF BPD AMONG CHINESE ADOLESCENTS

The present study examined the construct validity and prevalence of BPDamong Chinese adolescents in Hong Kong. Findings indicate that theDSM-IV-TR BPD criteria set as assessed by the MSI-BPD measures a reli-able and coherent personality construct among Chinese adolescents asdemonstrated by the high internal consistency and item-total correlations.The BPD construct as measured by MSI-BPD also has good concurrentvalidity and factorial validity. These findings lend further empirical sup-port to the generaiizability of the BPD construct among the Chinese popu-lation (Leung et a l . 2004: Leung et al-, 2007).

Results of item analysis indicate that among the ten MSI-BPD features,aflective instability had the highest item-total correlation and correlatedhighly with most of the other BPD symptoms, particularly anger dyscon-trol and impulsive behaviors. Similar results were also observed in adultpsychiatric patients (Grilo et al., 2001; Johanscn. Karterud. Pedersen,Gude. & Falkum. 2004). highlighting the importance of emotion dysregu-lation in the BPD construct (Leung & Zhong, 2006; Linehan, 1993).

According to DSM-IV-TR. BPD is characterized by frequent mood andimpulse dysregulation. unstable self-image, and chaotic interpersonal re-lationships (American Psychiatric Association. 2000). Findings of thisstudy lend support to the concurrent validity of the BPD construct as mea-sured by the MSI-BPD in our Chinese adolescent sample. As expected,significant moderate to high positive correlations were observed betweenMSI-BPD scores and independent measures of affective instability, impul-sive behaviors, unstable and fragile self-image, and disturbed interper-sonal relationships. Since BPD features and depressive symptoms arehighly correlated (Marton et al.. 1989; Zimmerman & MatUa. 1999). it ispossible that the correlation between BPD features and other personalitymeasures may be attributed to their common variance with depression.However, after controlling for depression in our analyses, all partial corre-lations between MSI-BPD and its concurrent correlates remained signifi-cant even though at an attenuated level. These findings are consistentwith previous studies which indicated that the BPD pathology was associ-ated with high mood lability (Zeigler-Hill & Abraham. 2006). fragile self-concept (Pinto. Grapentine. Francis. & Picarello. 1996). frequent interper-sonal conflicts, and low relationship satisfaction (Daley. Bürge & Hammen.2000).

CFA results in this study indicate that the DSM-IV-TR BPD criteria setas assessed by MSI-BPD can be represented by either a four-factor or athree-factor model among Chinese adolescents. The four-factor model in-cludes affective dysregulation. impulsivity, interpersonal disturbances.and self/cognitive disturbances. The three-factor model combines the in-terpersonal and self/cognitive disturbances components into a single fac-tor. While the three-factor model is more parsimonious, the four-factor


model provides a more detailed conceptualization of symptom compositionof the BPD pathology. Overall, findings of the CFA results are consistentwith the current conceptualization of the BPD pathology (Lieb et al.. 2004;Leung & Zhong, 2006: Linehan. 1993).

Many researchers believed that emotional vulnerability, which mani-fested as mood lability and frequent anger outburst (similar to the affectivedysreguiation factor in our CFA results), may represent an important pre-disposing personality trait in the development of BPD (Lieb et al., 2004:Leung & Zhong. 2006; Linehan, 1993). Frequent mood dysreguiation mayimpair interpersonal relationships (Linehan. 1993). The interpersonal dis-turbances factor found in this study, which included unstable relation-ships, interpersonai distrust, and fear of abandonment, may reflect thisambivalent relationship pattern (MUlon & Davis, 1996). Frequent mooddysreguiation may also impede the development of stable self-cognitionand disrupt other cognitive functioning (Zimmerman & Mattia. 1999). Ourself/cognitive disturbances factor which included unstable self-cognition,chronic feeling of emptiness, and dissociative symptoms may reflect prob-lems in the cognitive domain. Impulsive acts and self-mutilation may rep-resent maladaptive behavioral reactions in time of intense painful affects(Linehan. 1993). Our impulsivity factor which included impulsive acting-out behavior and suicidal acts may represent the impulsive componentwithin the borderline syndrome (Brodsky. Malone. Ellis. Dulit, & Mann,1997). These four factors correspond to the four major domains of BPDfunctioning as stated in Lieb et al.'s model.

PREVALENCE OF BPD FEATURESAMONG CHINESE ADOLESCENTS

Using only the MSÏ-BPD as a screening tool, 7.4% girls and 4,2% boyswere classified as suffering from BPD at Time 1. and 7.7% girls and 5.0%boys at Time 2. These prevalence figures should be viewed with caution asself-report instruments, when compared to structured clinical interviews.are prone to overdiagnose personality disorders (Hunt & Andrews, 1992;Hyler. Skodol. Kellman, Oldham, & Rosnick. 1990; Modestin, Emi, & Ob-erson, 1998; Zimmerman & Coreyell, 1990). Since the school authoritiesrefused to let us identify subjects for clinical interview, we developed astringent simulated diagnostic procedure for Time 2 testing with the inten-tion of achieving a more reliable estimate of the prevalence of BPD amongour adolescent sample. According to this simulated diagnostic procedure,subjects had to endorse three relevant behavior features to be consideredas meeting a specific BPD diagnostic criterion. We believe this stringentrequirement would have enhanced the possibility that subjects truly dis-play the concemed behavior characteristics. Based on this procedure.2.2% girls and 1.8% boys in our study met the BPD diagnosis. These prev-alence figures are highly comparable to those reported among general pop-ulation in the West (Torgersen et al.. 2001; Widiger & Weissman. 1991).

508 LEUNG AND LEUNG

Without collaborative information from structured clinical interviews,these prevalence figures should be regarded as preliminary estimatesawait eonñrmation from future studies. It is important, however, to realizethat even interview-based diagnoses display poor reliability and question-able validity as evidenced by the different findings yielded from differentstructured interviews (Hyler et al-, 1990). This is particularly relevantwhen subjects are young adolescents whose responses may be easily influ-enced by social desirability, hence under-reporting in face-to-face inter-\àew (Moum, 1998). Anonymous self-report methodology, when appliedwith well-planned cross-validating questions, may help eliciting more hon-est responses as well as eliminating interviewer bias.

STABILITY OF BPD FEATURES

Our findings indicate that the BPD pathology, when assessed as a dimen-sion with the MSI-BPD. displayed moderate one-year test-retest stability.This finding is consistent with what was reported in a previous study(Crawford. Cohen. & Brook. 2001). When the categorical approach wasused, however, only 36.8% girls and 36.2% boys in this study scored abovethe diagnostic cutoiT consistently at both time points. The finding that BPDas a diagnosis is not stable over time may appear counterintuitive as per-sonality disorders by DSM definition should be enduring. Some research-ers pointed out that BPD is a diagnosis with heterotypic continuity whichconsists of both state and trait sjonptoms (Mattanah. Becker, Levy. Edell.& McGlashan, 1995). While trait symptoms may persist over time, statesymptoms do not. Our findings reveal that affective instability and inter-personal distrust were the more stable features among our adolescentsample. Fear of abandonment, self-mutilating/suicidal behavior, and im-pulsive behaviors were less stable. Similar findings were also observed inadult patients (Meijer. Goedhart, & Trefiers. 1998). It is possible that affec-tive instability represents the predisposing vulnerability trait of BPD(Leung & Zhong. 2006: Linehan, 1993). This stable Lrait of affective insta-bility, even at subthreshold level, may impede the development of a stablesense of self and secure interpersonal bonding, making them more viiiner-able to full-blown BPD in the future (Morey & Zanarini. 2000). Based onthis reasoning, a person vulnerable for BPD may not display all the re-quired symptoms of BPD as some acute symptoms, like suicidal behaviorsand impulsive acts, may occur only in time of severe emotional distress(Linehan. 1993). Future research should examine the possible develop-mental sequences of different BPD symptoms to develop a better under-standing of the pathogenesis of this complex disorder,

STRENGTHS AND LIMITATIONS

There are several strengths in this study. First, the large sample size en-sures sufficient variability witliin the sample and enhances the reliability


of the findings. Second, our sample was followed longitudinally over a one-year period. The replication of results over time lends further support tothe reliability of our findings. Findings of this study, however, should beconsidered in light of certain limitations. First, our data were based solelyon adolescents' self-report. Because of strong objections from school ati-thorities. we were unable to conduct structured clinical interviews withour subjects as originally planned. Instead, we needed to develop a strin-gent simulated diagnostic procedure to assess prevalence of BPD amongour subjects. These findings await external cross-validation, either by in-terview data or data from other informants such as peers or parents infuture studies. Considering the strong concern with labeling or stigmatiz-ing effects of mental illness in the Chinese culture, other means of collect-ing cross-validating data should be explored in future studies. Second,our sample of school adolescents may not be fully representative of all theadolescents in Hong Kong. It is possihle that school drop-outs may containmore BPD individuals. Future studies examining BPD features amongschool drop-outs may be fhiitful.

Despite the limitations, findings from this study demonstrate that BPDis a reliable and coherent personality construct with good concurrent va-lidity, factorial validity, and acceptable test-retest reliability among Chi-nese adolescents in Hong Kong. A stringent simulated diagnostic proce-dure revealed that 2.2% female and 1.8% male adolescents met the BPDdiagnosis. The CCMD-III (Chinese Psychiatric Association. 2001) does notinclude BPD diagnosis in its nomenclature, claiming that the BPD diagno-sis is a vague construct that lacks precise boundaries and some of its diag-nostic features (e.g., fear of abandonment) may not be appropriate cultur-ally when used in China. Our findings, however. lend empirical support tothe BPD construct and suggest that cultural factors do not seem to affectthe expression of this syndrome among Chinese adolescents. Further re-search on this topic with the Chinese population is clearly warranted.

APPENDIX 1. ITEMS FOR CROSS-VALIDATING DSM-IV-TR BPDDIAGNOSTIC CRITERIA

1. Fear of abandonment— I always worried about being abandoned by others.— To avoid being rejected by others. I allowed other people to force me to do

things that I didn't want to do.— I am very dependent on others in relatlonsiiips. even feeling that I cannot

live without him/her.2. Unstable relationships

— I either love or hate other people in an extreme way.— My relationships with other people are very unstahle.— I had repeated breakups with oüier people.

510 LEUNG AND LEUNG

3. Identity disturbance ' '"— I am confused about my own identity.— My values change quickly.— My vocational goal changes often.

4. Impulsivity— 1 have deliberately damaged properties (e.g. smashed dishes or brolce fur-

niture).— I have gone on spending sprees where i spent a lot of money.— I have eaten uncontrollably to an extent that I felt sick.

5. Recurrent self-injurious or suicidal behavior . '— 1 have deliberately cut myself. . '— 1 bave deliberately bit myself.— 1 have attempted suicide.

6. AffecUve instability— My moods always vary.— Sometimes my moods swing back and forth very rapidly.— Compared to my friends, I'm more up and down in my mood states.

7. Chironic emptiness— 1 feel numb about many things around me.— Sometimes I don't bave any feeling at all. • , * ' '— Sometimes 1 feel quite empty inside me.

7. Inappropriate anger— I bave tbrealened to pbysically harm someone {e.g. tell someone that I

would punch him).— 1 iiave lost my temper and really shouted, yelled or screamed at someone.— I bave physically assaulted or verbally abused someone.

8. Transient dissociative features— In extreme emotional distress, sometimes I feel unreal.— Wben 1 feel extremely painful, sometimes I experience a feeling of detach-

ment from my body.— When I'm very disturbed emotionally, sometimes i feel tbings around me

are unreal.

REFERENCES

Alsaka. F., AOlweus. D. (1986). Assessment Becker, D. F., McGiashan. T. H.. & Grilo,of global negative self-evaluations and C. M, (2006). Exploratory factor analy-percelved stability of self in Norwegian sis of borderline personality disorderpréadolescents and adolescents. Jour- criteria ln hospitalized adolescents.nal of Early Adolescents. 6. 269-278. Comprehensive Psychiatry. 47. 99-105.

American Psychiatric Association. (20001. Bernstein, D. P., Cohen, P., Velez. C. N.,Diagnostic and statistical manual of Schwab Stone, M.. Siever. L. J.. &mental disorders (4th ed. Text-Revl- Shinsato. L. (1993). Prevalence andsion). Washington. DC: American Psy- stability of the DSM-IIl-R personalitychlatric Association Press. disorders in a community-based sur-

Becker. D. F., Grilo, C. M.. Morey. L. C vey of adolescents. American JournalWalker. M. L.. Edeli. W. S., & McGlas- ofPsychiatnj. 150. 1237-1243.han. T. H. (1999). Applicability of per- Brodsky. B. S., Malone. K. M.. Ellis, S. P.,sonality disorder criteria to hospital- DuUt. R. A.. & Mann. J. J. (1997).ized adolescents: Evaluation of internal Characteristics of borderline personal-consistency and criterion overlap. ity disorder associated with suicidalJournal of American Academy ofChikl behavior. American Journal ofPsychia-and Adolescent Psychiatry. 38. 200- try, 154. 1715-1719.205. Chabrol. H.. Montovany, A.. Choulcha, K.,


Callahan. S.. & Mullet. E. {2001). Fre-quency of borderline personality disor-der In sample of French high schoolstudents. Canadian Journal of Psychi-atry. 46. 847-849.

Ciilnese Psychiatric Association (2001). Chi-nese classification of mental disorders.(3rd ed.). Jinan: Shandong SciencePress.

Cohen. P. (1986}. TTie qualify of life interview.New York: New York State PsychiatricInstitute.

Crawtbrd. T. N.. Cohen. P.. & Brook. J. S.(2001). Dramatic-erratic personalitydisorder symptoms: 11. Developmentalpathways from early adolescence toadulthood. Journal of Personality Dis-orders. ¡5. 336-350.

Daley. S. E.. Bürge. D., Ä Hammen. C.(2000). Borderline personality disordersymptoms as predictors of 4-year ro-mantic relationship dysfunction inyoung women: Addressing issues ofspecificity. Journal of Abnormal Psy-chology. J09. 451-460.

Derogatis. L. R.. Lipman. R. S., & Covi. L.(1973). SCL-90. an outpaüent psychi-atric rating scale—preliminary report.Psychopharmacology Bulletin. 9. 13-28.

Dowson, J. H. (1992). Assessment of DSM-llI-R personality disorders by self-reportquestionnaire: The role of informantsand a s(?reening test for co-morbid per-sonality disorders (STCPD). BritishJour-nal of Psychiatry. 161. 344-352.

Grilo. C. M.. McGlashan. T. H.. Morey. L. C.Gunderson, J. G.. Skodol. A. E.. Tracie.S. M., et al. (2001), Internal consis-tency, lntercriterion overlap and diag-nostic efficiency of criteria sets forDSM-IV schizotypal, borderline, avoid-ant and obsessive-compulsive person-ality disorders. Acta Psychiatrica Scan-dinavica. 104. 264-272.

Hunt. C. & Andrews. G. (1992). Measuringpersonality disorder: The use of self-repori questionnaires. Journal of Per-sonaUty Disorders. 6. 125-133.

Hyler. S. E.. Skodol. A. E.. Kellman. H. D..Oldham. J. M., & Rosnlck. L. (1990).Validity of the personality diagnosticquestionnaire-revised: Comparison with

'- two simctured inierviews. AmericanJournal of Psychiatry. 147. 1043-1048.

Hyler. S. E.. Reider. R. O.. Williams. J.B.W..Spitzer. R. L. Hendler, J., & Lyons, M.( 1988). The personality diagnosUc

questionnaire: Development and pre-liminary results. Journal of PersonalityDisorders. 2. 229-237.

Johansen, M., Karterud. S., Pedersen. G.,Gude. T., & Falkum. E. (2004). An in-vestigation of the prototype validity ofthe borderline DSM-IV construct. ActaPsychiatrica Scandinavica. 109. 289-298.

Kutcher. S. P.. & Korenbhmi. M. (1992). Bor-derline personality disorder in adoles-cents: A critical overview, novel specu-lations and suggested future directions.In D. Siever. M. Rosenbluth (Eds.).HarKlbook of borderline personalitydisorder (pp. 535-552). Madison. CT:International Universities Press.

Leung. F., Chan, K.. & Cheung. F. (2007).Construct validity and prevalence ofborderline personality disorder amongChinese psychiatric patients in China.Paper presented at the World Congressof Behavioral and Cognitive Therapies.Barcelona. Spain.

Leung. F.. Cheung, F., & Cheung. S. F.(2004). CPAl-2 profiles of Chinese fe-male psychiatric patients displaying asigniflcant level of borderline personal-ity disorder features. Paper presentedat the 28th International Congress ofPsychology. Beijing. China.

Leung. F.. & Zbong. J. (2006). Understand-ing the pathogenesis of borderline per-sonality disorder using the intergener-ational brain-and-experience model.Journal of Chinese Clinical Psychology.J 4. 258-262.

Lieb. K.. Zanarini, M. C. Schmahl, C.Linehan. M. M., & Bohus. M. (2004).Borderline personality disorder. TheLancet. 364. 453-461.

Linehan. M. M. (1993). Cognitive behavioraltreatment of borderline personality dis-order. New York: Guilford Press.

Marton. P.. Korenblum. M., Kutcher. S.,Stein. B.. Kennedy, B.. & Pakes. J.(1989). Personality dysfuncUon In de-pressed adolescents. Canadian Jour-nal of Psychiatry, 34. 810-813.

Mattanah. J.J.F.. Becker. D. F.. Levy, K. N..Edelt. W. S.. & McGlashan, T. H.(1995). Diagnostic stability in adoles-cents followed up 2 years after hospi-talization. American Journal of Psychi-atry. 152. 889-894.

McGlashan. T, H., Grilo, C. M.. Sanislow.C. A.. Ralevski. E., Morey, L. C. Gund-

512 LEUNG AND LEUNG

erson, J, G.. et al. (2005). Two-yearprevalence and stability of individualDSM-IV criteria for schlzoiypal. bor-derline, avoidant, and obsessive-compulsive personality disorders: to-ward a hybrid model of axis II disor-ders. American Joumai of Psychiatry.162. 883-889.

Meijer. M.. Goedhart. A. W.. & Treffers,P.D.A. (1998). The persistence of bor-derline personality disorder in adoles-cence. Journal of Personality Disor-ders. 12. 13-22.

Millón. T.. & Davis, R D. (1996). Disorders ofpersanalUy: DSM-IV and beyond (2ndEd.) (pp. 645-690). John Wiley &Sons.

Modestin, J.. Eml. T.. & Oberson. B. (1998).A comparison of self-report and inter-view diagnoses of DSM-III-R personal-ity disorders. European Journal of Per-sonality. 12. 445-455.

Morey. L. C, & Zanarini. M. C. (2000). Bor-derline personality; Traits and disor-der. Journal of Abnormal Psychology.109. 733-737.

Moum. T. (1998), Model of admlnistraUonand interviewer elTects in self-reportedsymptoms of anxiety and depression.Social Indicator Research 45, 279-318.

Pads, J. (2003). Personality disorders overUme: FYccursors. course and outcome.Washington, DC, London: AmericanPsychiatric Publishing.

Pinto. A.. Grapentine. W. L., Francis. G.. &Picarello. C. M. (1996). Personality dis-order in adolescents: Affective andcognitive features. Journal of AmericanAcademy of Child and Adolescent Psy-chiatry. 35. 1338-1343.

Poon. W. F. (2003). The role of fragile senseof self in the development of disorderedeating behaviors among adolescentgirls in Hong Kong. Unpublished Mas-ter's Thesis. Tlie Chinese University ofHong Kong. Hong Kong.

Sanislow. C. A. Grilo. C. M.. Morey. L, C.Bender. D. S.. Skodol. A. E., Gunder-son, J. G.. et al. (2002). Conflnnatoryfactor analysis of DSM-IV criteria forborderline personality disorder: Find-Ings from the collaborative longitudinalpersonality disorders study. AmericanJournal of Psychiatry. ¡59. 284-290.

Satorra. A., Chou, C. C. & Bentler, P. M.(1991). Scaled test statistics and ro-bust standard errors for nonnormal

data in covariance structure analysis:A Monte Carlo study. British Joumai ofMathematical and Statistical Psychol-ogy. 44. 347-357.

Stembach. S. E., Judd, P. H.. Sabo, A. N.,McGlashan. T.. & Gunderson. J. G.(1992), Cognitive and perceptual dis-tortions in borderline personality' dis-orders and schizotypal personalitydisorder In a vignette sample. Compre-hensive Psychiatry. 33. 186-189.

Torgersen. S.. Krfnglen. E.. & Cramer. V.(2001). The prevalence of personalitydisorder in a community sample. Ar-chives of General Psycíiiaín/, 58. 590-596.

Underwood. B.. & Froming. W. J. (1980). Themood survey: A personality measure ofhappy and sad moods. Joumai of Per-sonality Assessment. 44. 404-414.

Wang, Y. Y.. Uung. F.. & Zhong, J. (2008).The adaptation of the McLean screen-ing instrument for borderline person-ality disorder for Chinese college stu-dents. Chinese Joiimal of ClinicalPsychology. 16,258-260.

Westen. D.. Shedler. J.. Durett, C, Glass. S..& Martens. A. (2003). Personality diag-nosis in adolescence: DSM-fV axis IIdiagnoses and an empirically derivedalternative. American Journal of Psy-chiatry, j 60. 952-966.

Widtger, T. A.. & Weissman. M. M. (1991),Epidemiology of borderline personalitydisorder. Hospital and CommunityPsychiatry. 42. 1015-1021.

Zanarini. M. C. Frankenburg. F. R., Kliera.G. S.. & Bklchmar. J. (2001). Treat-ment histories of borderline in pa-tients. Comprehensive Psychiatry. 42.144-150.

Slanarlni, M. C. Gunderson. J. G.. Franken-burg, F. R. & Chauncey. D, L. (1989).The revised diagnostic interview forborderlines: Discriminating BPD fromother axis II disorders. Journal of Per-sonality Disorders. 3. 10-18.

Zanarini. M, C, Vujanovic. A. A.. Parachini.E. A. Boulanger, J. L., Frankenburg.F. R.. & Hennen. J. (2003). A screeningmeasure for BPD: The McLean screen-ing instrument for borderline person-ality disorder (MSI-BPD). Joumai ofPersonality Disorders. 17. 568-573.

Zeigler-Hill, V.. & Abraham. J. (2006). Bor-derline personality features: Instabilityof self-esteem and affect. British Jour-


nal of Social and Clinical Psychology diagnostic comorbidity and borderline25. 668-687. personality disorder. ComprehensiDe

Zinunerman. M.. & Coreyell. W. H. (1990). Psychiatry. 49, 245-252.Diagnosing personality disorders In 21iong. J,. & I-eung. F. (2007). Should border-the community; A comparison of self- line personality disorder be included inreport and interview measures. Ar- the fourth edition of the Chinese classl-chiues of General Psychiatry. 47. 527- flcation of mental disorders? Chínese531. Medical Journal 120.77-82.

Zimmerman. M.. & Mattia. J. I. (1999). Axis 1

Copyright of Journal of Personality Disorders is the property of Guilford Publications Inc. and its content may

not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written

permission. However, users may print, download, or email articles for individual use.

construct validity and prevalence rate … normal peers, and the bpd diagnosis was associated with...

Documents