17(6).vp

Journal of Personality Disorders, 17(5), 550-561, 2003 2003 The Guilford PressAssessment of personality disorders in OCD TENNEY ET AL.
ASSESSMENT OF DSM-IV PERSONALITY
DISORDERS IN OBSESSIVE-COMPULSIVE
DISORDER: COMPARISON OF CLINICAL
DIAGNOSIS, SELF-REPORT QUESTIONNAIRE,
AND SEMI-STRUCTURED INTERVIEW
Nienke H. Tenney, PhD, Chris K.W. Schotte, PhD,Damiaan A.J.P. Denys, MD, MA,Harold J.G.M. van Megen, MD, PhD,and Herman G.M. Westenberg, PhD In patients with obsessive-compulsive disorder, personality disorders arenot many times assessed according to DSM-IV criteria. The purpose ofthe present study is to examine the prevalence of personality disordersdiagnosed according to the DSM-IV in a severely disordered OCD popula-tion (n = 65) with three different methods of assessing personality disor-ders (structured interview, questionnaire, and clinical diagnoses).
Furthermore, correspondence between these different methods was in-vestigated and their construct validity was examined by relating the threemethods to external variables. Each method resulted in a predominanceof Cluster C personality disorders, and obsessive-compulsive personalitydisorder had the highest prevalence. However, there was generally lowcorrespondence regarding which patient had which personality disorder.
Results concernign the relation of external variables were the mostpromising for the structured clinical interview.
INTRODUCTION
In clinical samples of patients with obsessive-compulsive disorder (OCD)
the prevalence of personality disorders was found to be at least 50%, with a
predominance of cluster C personality disorders (e.g., Steketee, 1990;
Horesh, Dolberg, Kirschenbaum Aviner, & Kotler, 1997; Bejerot, Ekselius,
& von Knorring, 1998). However, which of the specific Cluster C diagnoses
obtains the highest prevalence depends, among others, on which version of
the DSM is used for the assessment of personality disorders. This is illus-
trated by the research of Baer et al. (1990), who showed that the prevalence
of obsessive-compulsive personality disorder in a sample of OCD patients
From the Department of Psychiatry, University Medical Center, Utrecht (N.H.T., D.A.J.P.D,H.J.G.M.v.M., H.G.M.W.) and Department of Psychiatry, University Hospital Antwerp; Facultyof Psychology, Free University Brussels (C.K.W.S.).
Address correspondence to Nienke Tenney, Developmental Psychology and Psychopathology,Vrije Universiteit Amsterdam, Van der Boechorststraat 1, 1081 BT Amsterdam, The Nether-lands, E-mail: [email protected].
increased when DSM-III-R criteria were applied instead of DSM-III criteria.
Assessments of personality disorders in OCD are scarce according toDSM-IV criteria. In one study using the DSM-IV, a predominance of ClusterC personality disorders was found, with the majority of patients fulfillingcriteria for obsessive-compulsive personality disorder (Samuels et al.,2000).
In the absence of a gold standard, the most valid assessment method for personality disorders remains under debate (Zimmerman, 1994). Unstan-dardized clinical diagnoses, self-report inventories, and (semi-) structuredinterviews are the most commonly used methods. The reliability of clinicalAxis II diagnoses is poor, whereas standardized instruments tend to possesshigher levels of interrater and short-term test-retest reliability (Zimmerman,1994). The convergent validity between instruments assessing the DSMAxis II disorders is generally poor, while within-method correspondence isslightly higher than between-method correspondence (Clark, Livesley, &Morey, 1997; Perry, 1992; Schotte, 2000). Because Axis II diagnoses are notsignificantly comparable across methods beyond chance, results are not in-terchangeable (e.g., Kennedy et al., 1995; Mann et al., 1999). More insightwith regard to the validity of methods could be gained by investigating theirrelationship with variables like treatment outcome, biological markers, anddemographic and/or clinical characteristics, variables supposedly relatedto the presence of a personality disorder. In OCD, patients with comorbidpersonality disorders were found to have more depressive and anxioussymptoms and more impairment in functioning compared to OCD patientswithout comorbid personality disorders (e.g. Mavissakalian, Hamann, &Jones, 1990; Matsunaga et al., 1998). On the contrary, comorbid personal-i t y d i s o r d e r s obsessive-compulsive symptoms (e.g. Steketee, 1990; McKay, Neziroglu,Todaro, & Yaryura-Tobias, 1996; Cavedini, Erzegovesi, Ronchi, & Bellodi,1997).
The present study aims first to investigate the prevalence of personality disorders in an OCD population diagnosed according to DSM-IV definitions.
Three different methods are used to assess personality disorders;semi-structured interview, questionnaire, and clinical diagnosis based onan unstructured interview. Second, the convergence between these threemethods for the categorical DSM-IV Axis II diagnoses will be investigated.
Third, the construct validity of the three methods will be examined by relat-ing the results of each method to external variables. For each method the re-lationship with the level of obsessive-compulsive, depressive, and anxioussymptoms, and the level of global functioning is examined. Higher levels ofconstruct validity imply that for each method, patients with a comorbid per-sonality disorder are hypothesized to have more anxious and depressivesymptoms, and a lower level of functioning, compared to patients without ac o m o r b i d p e r s o n a l i t y d i s o r d e r . W e e x p e c t n o d i f f e r e n c e s i nobsessive-compulsive symptoms between patients with and patientswithout a comorbid personality disorder.
ASSESSMENT OF PERSONALITY DISORDERS IN OCD Sixty-five patients with a primary diagnosis of obsessive-compulsive disor-der according the DSM-IV (American Psychiatric Association, 1994) enteredthis study. These 65 subjects were part of a larger sample participating in adrug trial in which they received a standardized, 12-week treatment with ei-ther paroxetine (after 6 weeks a fixed dose of 60 mg/day) or venlafaxine (af-ter 6 weeks a fixed dose of 300 mg/day) (Denys, van der Wee, van Megen, &Westenberg, 2002).The MINI International Neuropsychiatric Interview(Sheehan et al., 1998) was used to establish the diagnosis. The study wascarried out at the outpatient Clinic for Anxiety Disorders at the UniversityMedical Center Utrecht. This unit offers specialized treatment for anxietydisorders.
The sample was predominantly female (60%) and had a mean age of 35.2 years (SD = 10.9). Regarding marital status, 48% were single, 51% were mar-ried or living together, and 2% were widowed. Fifty percent of the patientswere employed, 37% were unemployed, and 12.5% were students. The meanage of onset of obsessive-compulsive symptoms was 18.7 years (SD = 9.9),with an illness duration of 16.1 years (SD = 11.4). The mean Yale-Brown Ob-sessive-Compulsive Scale (Y-BOCS; Goodman et al., 1989) total score was25.9 (SD = 5.5); this mean score reflects a severe level of obsessive-compul-sive symptomatology (Goodman & Price, 1992). Most patients (82%) weretreated previously for OCD.
Exclusion criteria were: comorbid Axis I conditions (major depression, bi- polar disorder, schizophrenia or any other psychotic disorder, tic disorder,substance-related disorder during the past 6 months), a Hamilton Depres-sion Rating Scale (Hamilton, 1967) score higher than 16, treatment with an-tidepressants or neuroleptics for 2 weeks and cognitive-behavior therapy for3 months before the screening visit, intake of psychotropic drugs, with theexception of 30 mg of oxazepam or equivalent dose of any otherbenzodiazepine, during the trial.
INSTRUMENTS: ASSESSMENT OF PERSONALITY DISORDERS Clinical Diagnosis of Personality Disorders. Two trained and experienced psychiatric residents conducted an unstructured interview to assess thepresence of Axis II disorders.
Semi-structured interview. The Dutch version of the Structured Clinical Interview for DSM-IV Axis II Personality Disorders (SCID-II; First, Spitzer,Gibbon, Williams, & Benjamin, 1994; Weertman, Arntz, & Kerkhofs, 1997)was used for the assessment of DSM-IV personality disorders. The SCID-II isa semi-structured interview of 140 items, organized by diagnosis, coveringthe ten personality disorders included in the DSM-IV Axis II and the two per-sonality disorders (passive-aggressive and depressive personality disorder)proposed for further study. Most studies on the reliability and validity of theSCID-II relate to the DSM-III-R version of the interview (e.g., Arntz et al.,1992; Dreessen & Arntz, 1998) and mention an adequate interrater reliabil-ity and a reasonable level of test-retest reliability. Maffei et al. (1997) re- ported adequate levels of interrater and internal consistency reliability forthe DSM-IV version of the SCID-II.
Two interviewers, who were blind to the clinical diagnosis and question- naire outcomes, conducted the SCID-II interviews. One of the two interview-ers, a qualified psychologist, received training in the administration andscoring of the SCID-II by the authors of the Dutch translation. In order to as-sess interrater reliability between the two interviewers, they conducted tenjoint interviews, in which they took turns being the first, then second, raters.
Good interrater reliability was shown; a median kappa of .89 for thecategorical diagnoses was found.
Questionnaire. The Assessment of DSM-IV Personality Disorder (ADP-IV; Schotte, de Doncker, Vankerckhoven, Vertommen & Cosyns, 1998) wasadministered as a questionnaire to assess DSM-IV personality disorders.
The ADP-IV consists of 94 items that represent the 80 criteria of the 10DSM-IV personality disorders and the 14 research criteria of the depres-sive and passive-aggressive personality disorders in a randomized order.
Each DSM-IV item is scored on a 7-point trait scale, ranging from 1 (totallydisagree) to 7 (totally agree). Furthermore, when a person acknowledgesthe presence of the trait by a score of 5 (rather agree) or higher on a traitquestion, he/she has to fill in a distress question as well. This distressquestion is: “Has this characteristic ever caused you or others distress orproblems?” The answer to this question is rated on a 3-point scale: 1 (to-tally not), 2 (somewhat), 3 (most certainly). In summary, the ADP-IV as-sesses for each personality disorder criterion the self-judged typicality ofthe criterion by means of a 7-point trait question; the distress,maladaptivity, and suffering of the subject or of the people around him orher as a consequence of the presence of the trait criterion is subsequentlyassessed with the 3-point distress question.
This structure allows dimensional and categorical scoring formats. The categorical diagnostic evaluation joins the DSM-IV personality disorder def-inition by combining the trait and distress scores in scoring algorithms. Inthe algorithm that we used for the present study, an item scores posi-tive/pathological and represents a DSM-IV criterion only when a trait scoreof 5 (rather agree), 6 (agree), or 7 (totally agree), and a distress score of 2(somewhat) or 3 (most certainly) are obtained simultaneously. Subse-quently, categorical personality disorder diagnoses are obtained accordingto the DSM-IV thresholds.
The Global Assessment of Functioning Scale (GAF) was used as an overallmeasure of functional impairment. The Y-BOCS (Goodman et al., 1989) wasadministered to rate the severity of obsessive-compulsive symptoms. De-pression was assessed with the Hamilton Depression Rating Scale (HDRS;Hamilton, 1967), and anxiety symptoms with the Hamilton Anxiety Scale(HAS; Hamilton, 1959).
ASSESSMENT OF PERSONALITY DISORDERS IN OCD Following the unstructured clinical interview and the confirmation of theprimary OCD DSM-IV diagnosis with the administration of the MINI, writteninformed consent was obtained for participation. During this unstructuredclinical interview by one of two psychiatric residents, Axis II disorders wereassessed, and the Y-BOCS, GAF, HAS, and HDRS, were administered. Thenext week, the SCID-II interview was administered by one of two interview-ers, who were unaware of the clinical Axis II diagnoses, and patients re-ceived the ADP-IV questionnaire to complete at home within 2 weeks.
Not all assessments were available—one patient’s HDRS was missing, three patients’ HASs were missing, and no GAF scores were available for 12patients.
Cohen’s kappa (Cohen, 1960) was used to estimated agreement on categori-cal personality disorder diagnosis made using the three methods. The cate-gorical variables were (1) presence of one or more full-blown personalitydisorders, regardless of type; (2) presence of one or more cluster A, B, or Cpersonality disorders; (3) presence of specific personality disorders.
The reliability of the κ coefficient is influenced by the illness base rate: a few diagnostic disagreements have a more pronounced effect on reliabilitywhen the base rate is low (Zimmerman, 1994). Therefore, the following ill-ness base rate requirements were made. Kappa was calculated only if atleast five subjects were diagnosed with a personality disorder by everymethod. Generally, kappa values larger than 0.75 indicate excellent agree-ment, values below 0.40 indicate poor agreement, and values in betweenindicate fair to good agreement (Fleiss, 1986).
For to every method, the OCD patient group was divided into a group with and a group without comorbid personality disorders. Consequently, thesegroups were compared with regard to obsessive-compulsive symptoms,anxiety symptoms, depressive symptoms, and GAF score with studentT-tests.
RESULTS
The percentages of patients with personality disorder diagnoses according
the SCID-II, ADP-IV, and clinical assessment are presented in Table 1. Ac-
cording to the clinical diagnosis, 29.2% of the patients were diagnosed with
at least one Axis II diagnosis; most of those diagnoses occured in Cluster C,
with the dependent and obsessive-compulsive personality disorders obtain-
ing the highest prevalence. With the SCID-II and ADP-IV instruments, re-
spectively, 50.8% and 55.4% of the patients received at least one personality
disorder diagnosis. However, with the ADP-IV, compared to the SCID-II, al-
most twice the number of patients received two or more comorbid personal-
ity disorder diagnoses instead of just one, namely 26.1% versus 13.8%.
Comparable to the clinical diagnosis, with the ADP-IV and the SCID-II, ob-
sessive-compulsive personality disorder obtained the highest prevalence,
TABLE 1. Prevalence of Personality Disorder Diagnoses (%) in a Sample of 65 OCD
Patients According to Clinical Diagnosis, SCID–II, and ADP–IV
Clinical diagnosis
SCID–II
ADP–IV
PD present
Cluster A
Cluster B
Cluster C
but in contrast with it, dependent personality disorder did not; however theavoidant personality disorder was found to be the second most prevalentCluster C personality disorder diagnosis. The major difference between theADP-IV and the two other methods concerns the prevalence of cluster B per-sonality disorders and in particular borderline personality disorder. Thepercentages found with the ADP-IV are, respectively, between two and fourtimes higher than with the SCID-II and the clinical diagnosis.
Table 2 shows the correspondence between the clinical diagnosis and the SCID-II and ADP-IV diagnosis, respectively. Except for the correspondenceof dependent personality disorder between the ADP-IV and the clinical diag-nosis, the kappa values are all below 0.40. Hence, correspondence betweenthe clinical diagnosis and the two other methods can be considered poor. InTable 3, the correspondences between the SCID-II and the ADP-IV areshown. The range of kappa values for the presence of any personality disor-der, avoidant personality disorder, and presence of any Cluster C personal-ity disorder lies between 0.44 and 0.54, suggesting a fair to good agreement.
Regarding the presence of any Cluster B personality disorder, borderlinepersonality disorder, and obsessive-compulsive disorder, kappa valueswere below 0.40, indicating a level of poor agreement.
In Table 4, the mean scores of the Y-BOCS, HAS, HDRS, and GAF for pa- tients with and without comorbid personality disorders, according the threemethods are shown. When patients were categorized as personality disorderabsent or present according the clinical diagnosis, and according theADP-IV, no differences were observed in the HAS, HDRS, and GAF scores. Incontrast, the categorization with the SCID-II resulted in differences on thesethree measures in the predicted direction (i.e., higher depressive and anxi-ety symptoms and a lower GAF score for patients with a personality disorderdiagnosis compared to patients with no personality disorder). With regard tothe scores of the Y-BOCS, only categorization with the ADP-IV resulted in a ASSESSMENT OF PERSONALITY DISORDERS IN OCD TABLE 2. Correspondence between Clinical Diagnosis and ADP–IV and SCID–II
Expressed in Kappa Coefficients and in Percentage Agreement (Po)
SCID–II
ADP–IV
PD present
Cluster C
Note. Kappa values >.22 are significant at a p < .05 level.
difference between patients with and without a personality disorder diagno-sis (patients with a personality disorder diagnosis according the ADP-IVscored higher on the Y-BOCS compared to patients without a comorbidpersonality disorder).
DISCUSSION
The first goal of the present study was to examine the prevalence of person-ality disorders in an OCD population according to DSM-IV definitions. Wefound that each method for assessing personality disorders (the clinical di-agnosis, the semi-structured interview, and the self-report questionnaire)resulted in a predominance of Cluster C DSM-IV personality disorders in aseverely disordered OCD population. Of the Cluster C personality disorders,it was the obsessive-compulsive personality disorder that had the highestprevalence according every method. With the ADP-IV and the SCID-II, thesecond most prevalent cluster C personality disorder was avoidant person-ality disorder, in contrast to dependent personality disorder found with theclinical diagnosis. These findings are in line with most of the other studiesthat used previous formats of the DSM. In these studies, the prevalence ofpersonality disorders in OCD shows a predominance of obsessive-compul-sive personality disorder, and Cluster C personality disorders in general, aswell (e.g. Bogetto, Barzega, Bellino, Maina, & Ravizza, 1997; Cavedini et al.,1997; Dreessen, Hoekstra, & Arntz, 1997; Horesh, Dolberg et al., 1997;Bejerot et al., 1998; Samuels et al., 2000). Except for the last study, whereDSM-IV criteria were applied, DSM-III-R criteria were used. The presentstudy reveals that when applying DSM-IV criteria, obsessive-compulsivepersonality disorder is also the most prevalent personality disorder in clini-cal samples of OCD patients. This finding poses the question of whether therelationship between OCD and obsessive-compulsive personality disorderis true or a consequence of conceptual confusion between the criteria of bothdisorders.
We found that a clinical diagnosis resulted in a lower frequency of person- ality disorders compared to the structured interview and the self-reportquestionnaire. This is in accordance with Zimmerman and Mattia (1999),who have shown that a clinical diagnosis resulted in a lower frequency ofborderline personality disorder than a structured interview.
Our second goal was to examine correspondence between the three meth- ods. These results call into question the conclusion mentioned above con- TABLE 3. Correspondence between SCID–II and ADP–IV Diagnoses Expressed in
Kappa Coefficients and in Percentage Agreement (Po)
PD present
Cluster B
Cluster C
Note. All reported kappa values are significant at a p < .05 level.
cerning the predominance of obsessive-compulsive personality disorder.
Although the predominance of obsessive-compulsive personality disorderwas found with every method, methods did not correspond concerningwhich patients had an obsessive-compulsive personality disorder. Overall,correspondence between the clinical diagnosis on the one hand and theADP-IV and the SCID-II diagnoses on the other was low. The low correspon-dence of the clinical diagnosis with both standardized methods, and thesomewhat higher correspondence between the standardized methods, is notsurprising and is consistent with earlier reports (Perry, 1992; Schotte,2000). Both standardized methods employ inquiry into each criterion ofeach personality disorder diagnosis, in contrast to the clinical diagnosis,where assessments are made through more general descriptions given bythe patient. Correspondence between the ADP-IV and the SCID-II diagnosesreached a fair to good agreement for some diagnostic categories. Correspon-dence was never, however, excellent (i.e., a kappa value above 0.75). This isin line with the research of Schotte, de Doncker, Dmitruk, de Valck and vanMulders (2002), who found only moderate correspondence between theSCID-II and ADP-IV in a mixed psychiatric population.
As already mentioned in the Introduction, there is no consensus about which of the methods to assess personality disorders is more valid. Wewanted to address the question of validity by examining the relationship ofthese three methods with variables found to be related to the presence of apersonality disorder diagnosis in OCD, namely higher levels of depressiveand anxious symptoms and more functional impairment in OCD patientswith a comorbid personality disorder as compared to OCD patients withouta comorbid personality disorder (e.g. Mavissakalian et al., 1990; AuBuchon& Malatesta, 1994; Matsunaga et al., 1998). We found that patients with aSCID-II diagnosis of a personality disorder indeed had more depressivesymptoms (even while in the present study patients with a HDRS score of 16and higher were excluded), more anxiety symptoms, and a lower GAF scorethan patients without a comorbid personality disorder. Patients with andwithout a comorbid personality disorder diagnosis, according the clinical di-agnosis and the ADP-IV, showed no difference on any of the three variables.
With the ADP-IV, a difference in obsessive-compulsive symptoms was foundbetween patients with and without a personality disorder. However, otherstudies have shown that the severity of obsessive-compulsive symptoms is p
ADP–IV
p
SCID–II
p
Present/Absent Subgrouping as Independent Variable
Clinical diagnosis
TABLE 4. Results of the T–tests with Y–BOCS, HDRS, HAS and GAF Scores as Dependent Variables and the Diagnostic Axis II Disorde
Y–BOCS
not related to the presence or absence of personality disorders (Steketee,1990; McKay et al., 1996; Ricciardi et al., 1992). Hence, this finding sug-gests that answers received via a questionnaire might be more susceptible tointerference from OCD symptoms compared to those elicited by an inter-view. This may imply a greater value for a semi-structured interview com-pared to the questionnaire method for the assessment of personalitydisorders in OCD patients. Therefore, when these results are taken into con-sideration, assessment of personality disorders with SCID-II seems themost valid. However, we are aware that there are other variables that couldbe used as external criteria to validate the different methods as well, whichshould be examined first. Furthermore, the SCID-II and the HAS, HDRS,and GAF share the same method of assessing, namely an interview. Incontrast, the ADP-IV is a self-report questionnaire. This shared methodcould be another explanation for the findings with regard to the SCID-II incontrast to the ADP-IV.
To summarize, although all three methods used for assessing DSM-IV personality disorders in a OCD population obtained the highest prevalencefor Cluster C personality disorders and more particularly for obsessive-com-pulsive personality disorder, the convergence between the three methodswas rather low. Our attempt to examine which method was most valid by us-ing external measures is most promising with regard to the SCID-II.
REFERENCES
American Psychiatric Association (1994). Di- agnostic and Statistical Manual of Men- tal Disorders (5th ed.). Washington DC: Bogetto, F., Barzega, G., Bellino, S., Maina, Arntz, A., van Beijsterveldt, B., Hoekstra, R., Hofman, A., Eussen, M., & Sallaerts, S.
i t y d i m e n s i o n : A s t u d y r e p o r t .
European Journal of Psychiatry, 11, Cavedini, P., Erzegovesi, S., Ronchi, P., & AuBuchon, P. G. & Malatesta, V. J. (1994).
Neuropsychopharmacology, 7, 45-49.
Clark, L. A., Livesley, W. J., & Morey, L.
prehensive behavior therapy. Journal of Clinical Psychiatry, 55, 448-453.
Baer, L., Jenike, M. A., Ricciardi, J. N., Hol- ity. Journal of Personality Disorders, land, A. D., Seymour, R. J., Minichiello, W. E., & Buttolph, M. L. (1990). Stan- Cohen, J. (1960). A coefficient of agreement for nominal scale. Educational and Psy- o r d e r s i n o b s e s s i v e - c o m p u l s i v e chological Measurement, 20, 37-46.
disorder. Archives of General Psychia- Denys, D., van der Wee, N., van Megen, H. & Bejerot, S., Ekselius, L., & von Knorring, L.
disorder. International Journal of N e u r o p s y c h o p h a r m a c o l o g y , ASSESSMENT OF PERSONALITY DISORDERS IN OCD Dreessen, L., & Arntz, A. (1998). Short-inter- val test-retest interrater reliability of Disorders (SCID-II), version 2.0. Jour- (SCID-II) in outpatients. Journal of Per- sonality Disorders, 12, 138-148.
Dreessen, L., Hoekstra, R., & Arntz, A. (1997).
Takei, Y., & Yamagami, S. (1998). Per- sessive-compulsive disorder in Japan.
der. Journal of Anxiety Disorders, 11, Acta Psychiatrica Scandinavica, 98, First, M. B., Spitzer, R. L., Gibbon, M., Wil- Mavissakalian, M., Hamann, M. S., & Jones, liams, J. B. W., & Benjamin, L. S.
(1994). Structured Clinical Interview for DSM-IV Axis II Personality Disorders disorder. Comprehensive Psychiatry, McKay, D., Neziroglu, F., Todaro, J., & Fleiss, J. L. (1986). The design and analysis of clinical experiments. New York: Wiley.
Goodman, W. K. & Price, L. H. (1992). Assess- sive compulsive disorder. Psychiatric Perry, J. C. (1992). Problems and consider- sonality disorders. American Journal of Goodman, W. K., Price, L. H., Rasmussen, S.
Ricciardi, J. N., Baer, L., Jenike, M. A., Hill, C. L., Heninger, C. R., & Charney, Fischer, S. C., Scholtz, D., & Buttolph, use and reliability. Archives of General sive-compulsive disorder. American Journal of Psychiatry, 149, 829-831.
Hamilton, A. (1959). Diagnosis and rating of Samuels, J., Nestadt, G., Bienvenu, O. J., anxiety. British Journal of Psychiatry, 3, Costa, P. T., Jr., Riddle, M. A., Liang, K.
Y., Hoehn-Saric, R., Grados, M. A., & Hamilton, M. (1967). Development of a rating scale for primary depressive illness.
British Journal of Social and Clinical disorder. British Journal of Psychiatry, Aviner, N., & Kotler, M. (1997). Person- Schotte, C. K. W. (2000). New instruments for diagnosing personality disorders. Cur- rent Opinion in Psychiatry, 13, 605-609.
Schotte, C.K., de Doncker, D., Dmitruk, D. de Valck, E., & van Mulders, I. (2002).
Mendlowitz, S., Ralevski, E., & Clewes, bruikbaarheid. Tijdschrift Klinische methods. Journal of Nervous and Men- Maffei, C., Fossati, A., Agostoni, I., Barraco, C., Novella, L., & Petrachi, M. (1997).
characteristics: the ADP-IV. Psycholog- Weertman, A., Arntz, A., & Kerkhofs, M.
ical Medicine, 28, 1179-1188.
(1997). Nederlandstalige versie van de Sheehan, D. V., Lecrubier, Y., Sheehan, K. H., SCID-II (DSM-IV versie). Maastricht, Hergueta, T., Baker, R., & Dunbar, G.
Clinical & Experimental Psychology.
Zimmerman, M. (1994). Diagnosing personal- ity disorders. A review of issues and re- search methods. Archives of General and ICD-10. Journal of Clinical Psychia- Zimmerman, M., & Mattia, J. (1999). Differ- Steketee, G. (1990). Personality traits and derline personality disorder. American Journal of Psychiatry, 156, 1570-1574.

Source: http://www.ukmuenster.de/fileadmin/ukminternet/daten/kliniken/zahnprothetik/Psychosomatik/tenney_et_al._2003.pdf