تحلیل عاملی اکتشافی از معیارهای اختلال شخصیت مرزی در نوجوانان بستری در بیمارستان
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|38455||2006||7 صفحه PDF||سفارش دهید||4550 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Comprehensive Psychiatry, Volume 47, Issue 2, March–April 2006, Pages 99–105
Abstract Objective The authors examined the factor structure of borderline personality disorder (BPD) in hospitalized adolescents and also sought to add to the theoretical and clinical understanding of any homogeneous components by determining whether they may be related to specific forms of Axis I pathology. Method Subjects were 123 adolescent inpatients, who were reliably assessed with structured diagnostic interviews for Diagnostic and Statistical Manual of Mental Disorders, Revised Third Edition Axes I and II disorders. Exploratory factor analysis identified BPD components, and logistic regression analyses tested whether these components were predictive of specific Axis I disorders. Results Factor analysis revealed a 4-factor solution that accounted for 67.0% of the variance. Factor 1 (“suicidal threats or gestures” and “emptiness or boredom”) predicted depressive disorders and alcohol use disorders. Factor 2 (“affective instability,” “uncontrolled anger,” and “identity disturbance”) predicted anxiety disorders and oppositional defiant disorder. Factor 3 (“unstable relationships” and “abandonment fears”) predicted only anxiety disorders. Factor 4 (“impulsiveness” and “identity disturbance”) predicted conduct disorder and substance use disorders. Conclusions Exploratory factor analysis of BPD criteria in adolescent inpatients revealed 4 BPD factors that appear to differ from those reported for similar studies of adults. The factors represent components of self-negation, irritability, poorly modulated relationships, and impulsivity—each of which is associated with characteristic Axis I pathology. These findings shed light on the nature of BPD in adolescents and may also have implications for treatment.
. Introduction Much attention, over the past quarter century, has been focused on refining the “borderline” construct. Based in part on the work of Gunderson and Singer  and Spitzer et al , the Diagnostic and Statistical Manual of Mental Disorders, Third Edition (DSM-III)  subdivided this area of psychopathology into borderline and schizotypal personality disorders. Despite this improvement—and despite subsequent adjustments to the diagnostic criteria in the Diagnostic and Statistical Manual of Mental Disorders, Revised Third Edition (DSM-III-R)  and the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) —the borderline personality disorder (BPD) construct remains heterogeneous . This heterogeneity is partly inherent in the polythetic nature of the diagnosis . In addition, patients with BPD comprise a heterogeneous group, often manifesting a wide variety of comorbid Axes I and II disorders  and . Indeed, some of these comorbidity patterns have been used to characterize the nature of BPD within certain populations or to suggest BPD subtypes. For example, various investigators have considered the interface of BPD with mood, anxiety, somatization, and substance use disorders , ,  and . Another approach to examining this clinical heterogeneity has been through factor analytic techniques. Factor analysis can empirically identify meaningful components or latent elements within a diagnostic construct. Five such studies, using Diagnostic and Statistical Manual of Mental Disorders (DSM) criteria for BPD, have been reported , , ,  and . All examined adult populations: one used DSM-III criteria in college students , 2 used DSM-III-R criteria in adult psychiatric inpatients  and , and 2 studied DSM-IV criteria in diverse groups of adult patients  and . Although one set of findings  was consistent with a unidimensional construct, the other 4 studies suggested multiple dimensions. The analysis of Rosenberger and Miller  revealed 2 factors: the first including interpersonal and identity criteria and the second encompassing dysregulation of behavior and affect. The study of Clarkin et al  revealed 3 factors: interpersonal relationships and identity, affective dysregulation (including suicidality), and impulsivity. Using an adult inpatient group from the Yale Psychiatric Institute (YPI), Sanislow et al  also found 3 factors: disturbed relatedness, behavioral dysregulation, and affective dysregulation. Sanislow et al  subsequently validated this 3-factor model via confirmatory factor analysis, using DSM-IV criteria and a separate adult sample. Inasmuch as such factors may reflect core dimensions of borderline psychopathology, this type of analysis has important theoretical and clinical implications . For instance, delineating homogeneous BPD components may elucidate the boundaries between BPD and comorbid conditions, clarify etiologic pathways, and provide more specific targets for treatment . Although BPD has been studied far less in adolescents than in adults, the past decade has brought several empirical investigations of the BPD construct within this age group ,  and . Our own reports from the YPI Adolescent Follow-up Study suggest that personality disorders in this population, including BPD, can be reliably diagnosed, occur frequently, and have concurrent validity, but have only modest predictive validity and stability over time ,  and . These general findings for hospitalized adolescents are consistent with those of other studies involving community samples of adolescents  and  and are also consistent with overall findings from the adult literature . Some findings, however, suggest that the BPD construct may represent a more diffuse range of psychopathology in adolescents than in adults. Specifically, compared with an analogous group of adult inpatients, we found that BPD in adolescents had a broader pattern of criterion overlap with other personality disorders , a broader pattern of Axis II diagnostic comorbidity , and greater variability in the diagnostic efficiency of its criteria . Such findings highlight the need to explore BPD heterogeneity in adolescents and suggest the potential utility of factor analytic methods. To our knowledge, only 1 factor analysis of BPD in adolescents has previously been reported , although this study did not use DSM criteria. These investigators performed an exploratory factor analysis of borderline symptoms in a nonclinical adolescent sample and found 3 factors: one encompassing aspects of affective disturbance and psychoticlike experiences, another involving impulsive action, and a third comprised of various aggressive manifestations. The aim of the present study was to explore the factor structure of the BPD criteria in hospitalized adolescents who had been reliably assessed with a semistructured interview for DSM-III-R personality disorders. We also sought to add to our theoretical and clinical understanding of any homogeneous components by determining whether they may be related to specific forms of Axis I pathology. This aspect of the study was prompted by research in adults suggesting that BPD has broad Axis I comorbidity (eg, see Refs. ,  and ), as well as by studies from our own group ,  and  and others  and  indicating that BPD in adolescents may be associated with various Axis I disorders, including depression, conduct disorder, and substance use disorders.
نتیجه گیری انگلیسی
. Results Table 1 provides the frequencies of the Axis I disorders and disorder groups under consideration. BPD was diagnosed in 65 subjects—45% of the boys and 63% of the girls. These proportions were not statistically different. Table 1. Frequencies of Axis I disorders and disorder groups in 123 adolescent inpatients N % Major depression 80 65 Dysthymia 37 30 Anxiety disorders 29 24 Conduct disorders 68 55 Oppositional defiant disorder 23 19 Attention-deficit hyperactivity disorder 36 29 Alcohol use disorders 58 47 Drug use disorders 49 40 Table options Coefficient α for the BPD criterion set was 0.67, suggesting adequate internal consistency. The coherence of the criterion set is further supported by the strength of the intercorrelations among the individual BPD criteria, shown in Table 2. Although all criteria were significantly correlated with at least one other criterion, none was correlated with all other criteria, suggesting that no individual item clearly represents a core feature of this disorder. Table 2. Intercorrelations between DSM-III-R BPD criteria in 123 adolescent inpatients BPD Criterion Correlation (r) 1 2 3 4 5 6 7 8 1. Unstable relationships – 2. Impulsiveness 0.21⁎ – 3. Affective instability 0.25⁎⁎ 0.06 – 4. Uncontrolled anger 0.24⁎⁎ 0.21⁎ 0.36⁎⁎⁎ – 5. Suicidal threats or gestures 0.15 0.20⁎ 0.18⁎ 0.35⁎⁎⁎ – 6. Identity disturbance 0.21⁎ 0.23⁎⁎ 0.25⁎⁎ 0.29⁎⁎⁎ 0.11 – 7. Emptiness or boredom 0.29⁎⁎⁎ 0.11 0.17 0.26⁎⁎ 0.36⁎⁎⁎ 0.28⁎⁎ – 8. Abandonment fears 0.24⁎⁎ 0.12 0.08 0.18 0.13 0.15 0.06 – All tests are 2-tailed. ⁎ P < .05. ⁎⁎ P < .01. ⁎⁎⁎ P < .001. Table options Results of the principal components factor analysis are shown in Table 3. A 4-factor solution accounted for 67.0% of the overall variance and was felt to be conceptually appropriate. Factor 1 consists of “suicidal threats or gestures” and “emptiness or boredom.” This factor reflects self-negating or depressive aspects of the borderline presentation. Factor 2 embodies aspects of affective dysregulation or irritability and consists of “affective instability” and “uncontrolled anger,” along with “identity disturbance.” Factor 3 reflects the borderline patient's interpersonal dysregulation and consists of “unstable relationships” and “abandonment fears.” Lastly, “impulsiveness” loads most heavily on factor 4, although “identity disturbance” loads on this factor as well. Table 3. Varimax solution with 4 factors for DSM-III-R BPD criteria in 123 adolescent inpatients a BPD criterion Factor loadingb Factor 1c Factor 2d Factor 3e Factor 4f Unstable relationships 0.163 0.338 0.531 0.211 Impulsiveness 0.158 −0.070 0.126 0.879 Affective instability 0.117 0.834 0.083 −0.117 Uncontrolled anger 0.447 0.521 0.175 0.142 Suicidal threats or gestures 0.873 −0.011 0.103 0.049 Identity disturbance 0.017 0.575 0.027 0.555 Emptiness or boredom 0.695 0.235 0.007 0.112 Abandonment fears 0.024 −0.007 0.924 0.018 a Total percent of variance, 67.0%. b Bolded values represent the criteria that correspond to each factor. c Eigenvalue, 2.47; percent of variance, 30.9%. d Eigenvalue, 1.03; percent of variance, 12.9%. e Eigenvalue, 0.96; percent of variance, 12.0%. f Eigenvalue, 0.90; percent of variance, 11.2%. Table options We also considered a 3-factor solution, which accounted for 55.7% of the variance. We found this solution less satisfying theoretically, because it contained a component in which the criteria did not fit together as well. In the 3-factor solution, factors 1 and 2 were essentially the same as in the 4-factor solution, although factor 3 consisted of “unstable relationships,” “impulsiveness,” and “abandonment fears.” Table 4 shows the results of the logistic regression analyses, using the 4 BPD factors as independent variables and the Axis I disorders as dependent variables. Results for attention-deficit hyperactivity disorder are not reported, as the regression model for this disorder was not statistically significant (χ24 = 6.1, P = .19). For the remaining Axis I disorders, the table notes provide the χ2 results and also a goodness-of-fit index (Nagelkerke R2) to assist in evaluating the overall ability of the factor model to predict these disorders. Major depression and dysthymia are significantly associated with factor 1 only. Anxiety disorders are predicted by factors 2 and 3. Conduct disorder is predicted only by factor 4. Oppositional defiant disorder is associated with factor 2 and is inversely associated with factor 4. Alcohol use disorders are predicted by factors 1 and 4, and drug use disorders are predicted by factor 4 alone. Table 4. Logistic regression analyses with BPD factors as independent variables and Axis I disorders as the dependent variables, in 123 adolescent inpatientsa Factor 1 Factor 2 Factor 3 Factor 4 Odds ratio 95% CI Odds ratio 95% CI Odds ratio 95% CI Odds ratio 95% CI Major depressionb 1.76⁎⁎ 1.20-2.60 1.42 0.97-2.10 1.37 0.89-2.11 0.97 0.66-1.44 Dysthymiac 1.83⁎ 1.15-2.92 1.29 0.85-1.97 1.30 0.89-1.91 1.01 0.68-1.51 Anxiety disordersd 0.69 0.44-1.09 2.33⁎⁎ 1.35-4.03 1.54⁎ 1.01-2.34 0.89 0.56-1.41 Conduct disordere 1.14 0.76-1.69 0.90 0.60-1.34 1.22 0.81-1.82 2.85⁎⁎⁎ 1.83-4.44 Oppositional defiant disorderf 1.03 0.63-1.66 1.86⁎ 1.07-3.24 0.91 0.54-1.53 0.60⁎ 0.37-0.95 Alcohol use disordersg 1.77⁎ 1.13-2.75 0.95 0.62-1.45 1.14 0.75-1.74 3.73⁎⁎⁎ 2.24-6.20 Drug use disordersh 1.26 0.81-1.94 1.21 0.79-1.85 1.31 0.87-1.98 3.83⁎⁎⁎ 2.22-6.61 CI indicates confidence interval. a Results are presented as odds ratios “Odds ratio” represents the effect of the independent variable on the odds ratio of the probability of occurrence of the dichotomous dependent variable; it is calculated as eB, where B is the slope coefficient on the independent variable.. b For the overall model, χ24 = 13.9, P < .01; Nagelkerke R2 = 0.147. c For the overall model, χ24 = 10.4, P < .05; Nagelkerke R2 = 0.115. d For the overall model, χ24 = 17.1, P < .01; Nagelkerke R2 = 0.196. e For the overall model, χ24 = 28.3, P < .001; Nagelkerke R2 = 0.275. f For the overall model, χ24 = 10.7, P < .05; Nagelkerke R2 = 0.134. g For the overall model, χ24 = 42.2, P < .001; Nagelkerke R2 = 0.388. h For the overall model, χ24 = 37.5, P < .001; Nagelkerke R2 = 0.356. ⁎ P < .05. ⁎⁎ P < .01. ⁎⁎⁎ P < .001.