ویژگی های شناختی از بیماران مبتلا به اختلال شخصیت مرزی: توسعه و اعتبار از یک پرسشنامه خودگزارشی

Abstract Based on cognitive concepts of personality disorders as well as on the bio-social model of borderline personality disorder (BPD), a 34-item instrument, the questionnaire of thoughts and feelings (QTF) was developed for the assessment of feelings, strategic cognitions, and assumptions characteristic for BPD. In different studies, item- and factor analyses were conducted with a dataset of N=646N=646 clinical and non-clinical participants. Cross-sectional data as well as longitudinal data are available including several other measures on personality and personality disorders: DSM-IV SCID II dimensional score, a personality inventory (PSSI), the German version of the personality disorder beliefs questionnaire (PBQ), as well as overall severity of symptoms (GSI) of the SCL-90-R. Internal consistency of the scale is excellent, Cronbach's α=0.91α=0.91. One-week test-retest reliability was also high rtt=0.81rtt=0.81. Good convergent and discriminant validity was obtained for correlation with SCID dimensional scores and discrimination of BDP patients and other clinical and non-clinical groups. A comparison of scores in a sample of 26 patients with BPD during crisis intervention and 8 months later indicated the sensitivity of the QTF to measure change over time. Overall, the questionnaire appears to be a methodologically sound measure, valuable for research purposes as well as in clinical practice.

Introduction In cognitive models of borderline personality disorder (BPD) the role of underlying maladaptive assumptions for the development and especially the maintenance of the disorder is emphasized (Arntz (1994) and Arntz (2004); Pretzer, 1990; Young, Klosko, & Weishaar, 2003). In the bio-social model developed by Linehan (1993), cognitions also play a major role for the conceptualisation and for treatment planning of this severe mental disorder. At the same time, this model stresses the importance of behavioral aspects. Recently, two valid and reliable self-report measures have been published based on cognitive concepts of BPD (Arntz, Dietzel, & Dreessen, 1999; Beck et al., 2001). In accordance with the cognitive model, focus of these instruments is to assess the dysfunctional basic assumptions and beliefs of patients with BPD. More behavioral aspects are not included in these instruments. A more global approach to assessment of borderline relevant symptoms and related psychopathology is described by Bohus et al. (2001). These authors constructed a questionnaire designed to measure the degree of borderline symptomatology during the last week. The authors followed current diagnostic criteria and expert ratings, as well as clinical observations and patients’ statements in the construction of the inventory. Although reliability and validity estimates are satisfactory, the instrument comprising 95 items appears to be somewhat long and redundant with interview measures assessing diagnostic criteria of BPD. The aim of the present studies was to develop an economic self-report questionnaire combining emotional aspects, assumptions, and strategic cognitions closely related to overt behaviors characteristic for BPD. Furthermore, the instrument should be applicable for treatment planning and evaluation, thus sensitive to measure change

3. Results Item analyses, reliability and validity estimates for the questionnaire of thoughts and feelings are described below. 3.1. Item analysis Item difficulty: At the individual item level of the QTF scale, item difficulty ranged between 0.24 and 0.79. According to Bortz & Döring (1995, p. 199) this represents an optimal range of difficulty. Item-test-correlation : The mean item-test-correlation for the different samples was rit=0.50rit=0.50. Indicating homogeneity but little redundancy of the items. Only for three items coefficients were < 0.30. Homogeneity : Calculation of mean item intercorrelation resulted in r=0.27r=0.27, being in the acceptable range as suggested by Briggs and Cheek (1986). Factor structure : The intercorrelations of the QTF item-scores of the total sample (N=646)(N=646) were factor-analyzed (principal component analysis). Formal criteria yielded a one-factor solution, explained variance was 50%. Thirty-three of the 34 items loaded higher than 0.40 with the factor. For 26 items loadings were higher than 0.60. Thus, the QTF can be considered as a homogenous, uni-dimensional measure. 3.2. Reliability Internal consistency : In the different samples, Cronbach's αα coefficients were consistently >0.89. Thus, the internal consistency of the QTF is excellent. Test-retest reliability : Thirty-six patients with substance abuse who were in long-term treatment in a drug rehabilitation center completed the QTF twice, with an interval of one week. Pearson correlation for the QTF score in this sample was r=0.81r=0.81. 3.3. Validity Construct validity: Correlations of the QTF with other measures of personality disorders are displayed in Table 2. Table 2. Correlations of the QTF with PBQ and PSSI scales for a sample of psychotherapy outpatients (n=88)(n=88) Personality style PBQ PSSI Borderline —a 0.82*** Avoidant 0.68*** 0.72*** Dependent 0.63*** 0.58*** Histrionic 0.61*** −0.12 Obsessive-compulsive 0.58*** 0.06 Paranoid 0.56*** 0.43*** Narcissistic 0.48*** 0.56*** Antisocial 0.43*** 0.16 Passive-aggressive /negativistic 0.34** 0.60*** Schizoid/schizotypal 0.29** 0.28**/0.08 Depressive —a 0.74*** Altruisticb —a 0.36** Optimisticb —a −0.19 Note : ***p<0.001***p<0.001,**p<0.01**p<0.01, *p<0.05*p<0.05. a Personality style not assessed by PBQ. b Personality style assessed by the PSSI, not included in DSM-IV. Table options In general, correlations of the QTF with the scales of Beck's PBQ( German version B-IKS, Fydrich, 2002) were high (see Table 2). PBQ scales assessing beliefs characteristic for avoidant, dependent or histrionic personality disorders were especially related to the QTF score. Additional information about construct validity is derived from the correlation coefficients of the QTF with the PSSI scales (Persönlichkeits-Stil- und Störungsinventar, Kuhl & Kazén, 1997, Table 2). The included borderline scale showed the highest correlation with the QTF score. Depressive, avoidant, and negativistic PSSI scales were also highly related to the QTF score. Criterion Validity : The SCID II was administered to all psychiatric inpatients with BPD (n=30)(n=30). The correlation between the QTF and the BPD dimensional score of the SCID II was significant r=0.43r=0.43. According to the expectation, patients with BPD scored higher on the QTF than patients with different psychiatric disorders other than BPD and the non-clinical control group. Mean QTF scores for the different samples are presented in Table 1. To estimate discriminant validity QTF and SCL-90 R GSI scores were compared in three different groups: inpatients with BPD, psychiatric inpatients with mainly affective and psychotic disorders, and a non-clinical control group (see Fig. 1). Comparison of mean QTF and SCL-90-R GSI Scores. Note: z-values based on ... Fig. 1. Comparison of mean QTF and SCL-90-R GSI Scores. Note: z -values based on non-clinical sample, ***=p<***=p< 0.001, BPD=borderline personality disorder, Clinical=psychiatric inpatients with mixed diagnoses. Figure options As expected, patients with BPD and inpatients with different psychiatric disorders showed significantly higher scores on the GSI and the QTF than the non-clinical control group. However, on the SCL-90-R GSI , both clinical groups did not differ significantly from each other. In contrast, the QTF score discriminated between the two clinical groups. Mean QTF score of the BPD patients was significantly higher (T=4.16,p<0.001)(T=4.16,p<0.001) than that of the group with different psychiatric disorders. In the international Internet study on BPD, participants (n=273)(n=273) were asked to rate whether they were affected by borderline symptoms or not. QTF scores of this sample were used to classify participants as “affected” or “non-affected”. Discriminant analysis revealed a 94% correct classification, with sensitivity of 96% and specificity of 91%. Sensitivity to change : In a study examining stability of aspects of BPD, the QTF was administered twice to BPD and depressed inpatients as well as a non-clinical control group. For the clinical groups, first assessment was during inpatient treatment, for all three groups second assessment took place approximately 8 months later, when patients were not in an acute crisis and not hospitalised. Only for BPD patients, significant differences in QTF scores were detected (Mt1=3.40,Mt2=3.08,T=2.93,p=0.007)(Mt1=3.40,Mt2=3.08,T=2.93,p=0.007), indicating a decrease in self-reported dysfunctional beliefs and behaviors. For patients with MDD QTF differences (Mt1=2.70,Mt2=2.48)(Mt1=2.70,Mt2=2.48) and the non-clinical control group (Mt1=1.80,Mt2=1.75)(Mt1=1.80,Mt2=1.75) were not significant.