Although the CBCL has been used extensively in youth with a wide variety of psychopathology, including JBD, to our knowledge, this is the first study to specifically examine its utilization in youth with a depressive episode associated with bipolar I disorder. A majority of prior studies focused on the diagnostic utility of the CBCL in bipolar youth with a manic or mixed mood state or who were euthymic. Despite some controversies about the sensitivity and specificity of its use in this population in general, previous studies found the CBCL to be useful in identifying individuals with bipolar disorder. Since Biederman et al. (1995)
proposed that youth with JBD have a specific CBCL-JBD profile, characterized by an elevation in Anxious/Depressed, Aggression, and Attention Problems subscales, the CBCL has been viewed as a behavioral assessment tool that is trait (bipolar disorder, per se) rather than state (manic, mixed, depressive, euthymic) specific.
Results of our study show that the CBCL profile in medication-free adolescents with bipolar depression, although showing elevated scores of at least one SD above the mean in the majority of the subscales, reached the cutoff score of 70, indicative of clinical significance, only in the internalizing subscale and total scores. The CBCL-JBD profile in our sample also failed to reach the cutoff score of 210 reported to be indicative of clinical significance.
Prior studies have explored the validity of the CBCL-JBD in youth. For example, Diler et al. (2009)
studied the CBCL-JBD in children, younger than 12 years old, who were diagnosed with bipolar disorders I, II and Not Otherwise Specified; current mood state was not specified. The authors did not find the CBCL-JBD useful in discerning BD from other psychopathology; while a negative screen might rule out BD, a positive screen was non-specific. Significant differences in CBCL subscale scores between our study and that of Diler et al. were found in the domains of Social Problems, Thought Problems, Aggressive Behaviors as well as CBCL-JBD profile scores. In another study, Dienes et al. (2002)
assessed offspring of bipolar parents, ages 6 to 18 years. Of these, 16 subjects were diagnosed with bipolar disorder themselves (BD group; again, current mood state was not specified). The authors concluded that the CBCL identified subjects with clinical disorders and reflected greater pervasive disturbance in the children with bipolar disorder. The CBCL profile of our sample differed from that of Dienes et al.'s in some domains, including Withdrawn, Somatic Complaints, Social Problems and Aggressive Behavior. Most importantly, in both comparison studies, the CBCL-JBD profile scores reached the level of clinical significance at 215.9 (SD=21.6) and 219.3 (SD=21.2), respectively, whereas ours did not ().
Comparison of CBCL subscales and total scores of our sample with other studies.
A literature search was performed for CBCL scores of children and adolescents with depression. Biederman et al., (1996a)
assessed the CBCL scales of children with major depression (mean age 13.8, SD 3.0). The authors concluded that the CBCL Anxious/Depressed and other internalizing scales adequately discriminated between unipolar depressed and non-depressed children regardless of ADHD comorbidity. With the exception of Rule Breaking and Aggressive Behaviors, syndrome sub-scale scores from our sample population were similar to those of Biederman and colleagues’ depressed subjects without comorbid ADHD. Hepperlin and colleagues’ (1990)
study of youth ages 11 to 18 years with unspecified depression concluded that the CBCL measures poorly predicted Children's Depression Inventory (CDI) scores. Our data corresponded with internalizing and externalizing broad-band scales and total scores in this study of depressed youth. Overall, our data in bipolar depressed youth correlated with that of studies involving depressed adolescents.
We conclude that the CBCL-JBD profile fails to identify youth in the depressive phase of bipolar disorder. In light of these findings, we propose diagnostic sensitivity of the CBCL-JBD profile is contingent upon the bipolar child presenting in a manic or mixed state; in other words, our results introduce the idea that this profile is state-specific rather than trait-specific. An assessment tool which supports diagnosis based on discerning and stable traits of true JBD would be ideal. This conclusion is perhaps consistent, however, with the purported dimensional paradigm of the CBCL (Biederman et al., 1996a
). Biederman et al. described the categorical paradigm, which is based upon discrete medical syndromes that require fulfillment of specific criteria for diagnosis. In contrast, the authors elaborate upon the dimensional approach of the CBCL whereby pathology is viewed on a continuum and focus is placed upon capturing deviation from a norm rather than a specific diagnosis.
One significant limitation of our comparison of our subjects to those in Diler's and Dienes’ studies is notable demographic difference; differences in sex, age and ADHD comorbidity limit the interpretability of our comparisons with other studies. Further investigation with a larger sample size as well as a control sample is warranted. Nonetheless, our study does contribute to the limited extant literature focusing on depression associated with bipolar disorder. Diagnostic tools that discern unipolar and bipolar depression are needed. Only after accurate and efficient diagnosis can appropriate treatment be implemented to minimize the significant morbidity and mortality that is inherent to untreated bipolar illness.