Cognitive-Behavioral Therapy for Adolescent DepressionProcesses of Cognitive Change

The substantial and often recurrent distress and impairment associated with major depressive disorder (MDD) in youth has prompted increased interest in the identification and dissemination of effective treatment models. Evidence supports the use of several antidepressant medications, specific psychotherapies, and, in the largest treatment study of depressed teenagers, the combination of fluoxetine and cognitive-behavioral therapy (CBT) as effective treatments.1-3 CBT is the most extensively tested psychosocial treatment for MDD in youth, with evidence from reviews and meta-analyses that supports its effectiveness in that population.3-5

The substantial and often recurrent distress and impairment associated with major depressive disorder (MDD) in youth has prompted increased interest in the identification and dissemination of effective treatment models. Evidence supports the use of several antidepressant medications, specific psychotherapies, and, in the largest treatment study of depressed teenagers, the combination of fluoxetine and cognitive-behavioral therapy (CBT) as effective treatments.1-3 CBT is the most extensively tested psychosocial treatment for MDD in youth, with evidence from reviews and meta-analyses that supports its effectiveness in that population.3-5

CBT is a time-limited, problem- focused intervention that seeks to reduce emotional distress through the modification of maladaptive beliefs, assumptions, attitudes, and behaviors.6 As outlined by Lewinsohn and Clarke,7 different CBT interventions emphasize different techniques to effect change, with some interventions primarily targeting cognitive factors and others primarily targeting behavioral factors. Reflecting this distinction, there have been 2 major theoretical approaches to CBT with adolescents.

The first approach, identified with the cognitive therapy of Beck and his colleagues,8 is based on a model in which cognitive processes are seen as the major maintaining variables associated with MDD. In Beck's model, the task of the therapist is to enhance the patient's ability to monitor mood; identify connections between mood and cognition; and identify, challenge, and modify automatic thoughts, assumptions, and core beliefs that sustain MDD. This model has been adapted by Brent and colleagues9 for treating adolescents.

The second approach is more behavioral and multifactorial. Associated with Lewinsohn and colleagues,10 this approach has been used with adults and adolescents and is based on the assumptions that behavior and thoughts sustain depressed mood, and that either behavioral or cognitive change can serve as the engine of therapeutic progress.

Despite different areas of emphasis, both of the major CBT approaches posit that cognitive factors such as cognitive distortions, negative automatic thoughts, dysfunctional attitudes, negative attributional style, and hopelessness play a major role in the onset and maintenance of depressive symptoms.6 However, the specific cognitive processes through which CBT works to counter MDD in young people are not well understood. Understanding treatment mechanisms is of both theoretical and practical importance. A clear understanding of therapeutic mechanisms would allow psychosocial treatment developers to focus on the intervention components that are most effective, thereby promoting amplification of the more active components and reduction or removal of the less active elements. In addition, a review by Kazdin and Nock11 proposes that studying mechanisms of therapeutic change can assist the adoption of effective treatments by practicing clinicians.

In this article we focus on cognitive factors that have been shown to influence the effectiveness of CBT for adolescent MDD. We identify factors that affect CBT treatment outcomes by examining 3 types of variables: predictors, moderators, and mediators. Predictors are variables present before treatment that influence treatment outcomes across all treatment conditions. For example, if, in a comparison of CBT and a different psychotherapy for MDD, adolescents with high levels of hopelessness had worse outcomes across both conditions, then hopelessness would be considered a predictor.

Moderators are pretreatment factors that interact with the treatment conditions to predict treatment outcome, and thereby represent factors that predict a differential response to CBT. For example, if adolescents with high levels of hopelessness had better outcomes with CBT than with an alternative psychotherapy, hopelessness would be a moderator. Mediators are factors that account for or explain the process of therapeutic change during CBT. For example, if it were shown that CBT treatment outcomes were attributable to a reduction in hopelessness during treatment, then hopelessness would be a mediator.

Mediational analyses are those that are most valuable in highlighting processes of change associated with effective treatment. The evaluation of mediators can help address the following 3 questions, which have been explored in the literature on adults with MDD and are only beginning to be studied in adolescents with MDD. First, does CBT lead to changes in cognition associated with MDD? Second, if CBT leads to cognitive change, does the cognitive change lead to a reduction in symptoms of MDD? Third, is change in depressive cognitions unique to CBT, or is it associated with other effective treatments for MDD? In this article, we explore the first 2 questions with reference to studies of CBT that emphasize cognition, studies of multifactorial CBT, and the largest treatment study of CBT for adolescent MDD to date.

Of note, most clinical trials are sufficiently powered to test their main hypotheses, which pertain to outcome, and to identify predictors. By contrast, most trials are not sufficiently powered to test adequately for moderators or mediators. Therefore, most analyses for moderators or mediators must be considered exploratory.

CBT emphasizing cognition
The first major stream of CBT research in adolescent MDD has been the work of Brent and colleagues,9 who used Beck's theoretical approach to cognitive therapy. They adapted the Beck model for adolescents by including psychoeducation, problem solving, affect regulation, and social skills training. Nevertheless, the primary focus of treatment is on identification and modification of dysfunctional, depressogenic thinking.

One randomized trial has evaluated the effectiveness of the Brent model of individual CBT relative to 2 alternative psychotherapy conditions: Systemic Behavioral Family Therapy and Nondirective Supportive Therapy.9 In the sample of adolescents with MDD, CBT was found to be more effective in alleviating depression after 12 to 16 weeks of treatment than either of the 2 comparison psychotherapy conditions. Cognitive variables measured in this study, which were directly targeted by the CBT intervention, included cognitive distortions and hopelessness. Cognitive distortions were measured using the overall score on the Children's Negative Cognitive Errors Questionnaire,12 while hopelessness was measured using the Beck Hopelessness Scale.13

An investigation of predictors of treatment effect revealed that higher levels of both of these cognitive variables at intake predicted continued depression and failure to attain remission at termination.14 An additional predictor of negative treatment outcome was entering the study as a clinical referral, rather than in response to an advertisement. The effect of clinical referral status on treatment outcome was mediated by hopelessness, which reflects that adolescents who were referred to the study had higher levels of hopelessness than did adolescents who answered an advertisement. When controlling for adverse predictors including cognitive variables the efficacy of nondirective supportive therapy declined significantly, but not that of the other treatments. These findings support the overall robustness of CBT with patients who have dysfunctional cognitive processes.

More recent analyses of this trial provided additional information about the process of cognitive change during CBT and the other 2 psychosocial treatments.15 During treatment, CBT had a significantly greater impact in reducing cognitive distortions than did either of the other psychotherapies. Furthermore, when analyses included data from a 2-year follow-up assessment, CBT demonstrated significantly greater and more consistent improvement for cognitive distortions in children than did nondirective supportive therapy. By contrast, there was no difference among treatments in their impact on hopelessness.

Despite evidence of treatment- specific effects, cognitive distortion failed to meet the criteria for mediation because of the absence of group differences over time in the severity of depressive symptoms. As noted by the investigators, mediation in a treatment arm may have been difficult to show within the relatively brief time frame of acute treatment, particularly when considering the episodic nature of MDD among adolescents. The lack of significant mediators was congruent with the results of the literature on adults, which has generally been unsuccessful in identifying mediators of treatment effects for CBT interventions among patients with a confirmed diagnosis of MDD .16 It is of interest to note that Ackerson and colleagues17 found evidence that dysfunctional attitudes mediated response to a form of cognitive bibliotherapy based on the Beck cognitive model in a sample of 22 adolescents with depressive symptoms (not diagnoses). However, support for mediation was found only when symptoms and cognitions were both measured by self-report.

Multifactorial CBT
A second stream of research has been built on Lewinsohn's multifactorial model of MDD, which identifies multiple behavioral, cognitive,and biological factors that contribute to MDD.10 To simultaneously address these factors, Clarke and colleagues18 designed the Adolescent Coping With Depression Course (CWD-A), a course of CBT that is group-administered, psychoeducational, and skills-oriented in nature.

Two randomized trials have assessed the effectiveness of the CWD-A group CBT intervention compared with CWD-A plus a weekly parent psychoeducation group and a wait-list control.10,19 The first study involved 59 adolescents who met diagnostic criteria for MDD or intermittent depression, whereas the second study involved 96 adolescents who met diagnostic criteria for MDD or dysthymia. Both of these trials found that rates of recovery, as defined by no longer meeting criteria for initial diagnosis, were greater in the 2 CBT conditions than in the wait-list group. Across both studies, higher levels of rational (nondepressive) thinking at baseline, as measured by the Subjective Probability Questionnaire (R. Munoz and P. Lewinsohn, unpublished data, 1976), were a predictor of better treatment outcomes. No analyses of moderators or mediators were reported in these trials.

A recent application of the CWD-A compared it with a life skills condition for adolescents with comorbid MDD and conduct disorder.20 Adolescents treated with CWD-A had higher recovery rates than those treated with the life skills condition (39% vs 19%). In the context of this controlled trial, the cognitive variables of hopelessness, negative automatic thoughts, and dysfunctional attitudes were assessed. Over the 1-year follow-up, lower levels of hopelessness and negative thoughts, but not dysfunctional attitudes, were found to predict faster recovery.21 In analyses of this trial, potential mediators of treatment outcome were explored in adolescents who received CBT intervention.22 One of the cognitive factors, automatic negative cognitions, which was measured with the Automatic Thoughts Questionnaire developed by Hollon and Kendall,23 mediated the effects of CBT on depression outcomes. None of the other variables that were tested (dysfunctional attitudes, relaxation, social skills, pleasant activities, or problem-solving) fulfilled the criteria for mediation of treatment effects. This finding suggests that reducing negative thinking may be the primary mechanism through which this form of CBT reduces depression among youths with conduct disorder and depression.

Another notable finding was that reducing negative thinking significantly reduced symptoms of depression, even though CBT did not appear to have an impact on underlying dysfunctional attitudes, as measured by the Dysfunctional Attitude Scale.24 Counter to the findings of the cognitive bibliotherapy trial, these findings suggest that it may not be necessary for the clinician to engage in depth-oriented cognitive techniques for this form of CBT to be effective.17

Taken together, results of the second stream of research provide evidence that CBT is associated with change in a specific cognitive factor (automatic negative thoughts) and that this change accounts for the effect of CBT on adolescent outcomes. Results of this stream of research are consistent with the results of 2 prevention studies for adult depression that found that measures of negative thoughts were the strongest mediators of reductions in depressive symptoms following treatment with CBT.

In the first study, Munoz and colleagues25 randomly assigned 150 predominantly minority, low-income primary care patients to an experimental CBT condition or 1 of 2 control conditions: information only or no treatment. At posttreatment and 6-month follow-up, reduction in depressive symptoms was significantly mediated by a reduction in negative thoughts, while at the 12-month follow-up, reduction was mediated by reduced negative thoughts and increased optimism.

In the second study, Allart-van Dam and colleagues26 randomized adults at risk for developing MDD to the Coping With Depression CBT course or an assessment and advice only condition. One month after acute treatment, negative automatic thoughts and self-esteem were found to mediate a reduction in symptoms of MDD. By contrast, measures of the frequency of interpersonal behavior and pleasant events were not found to mediate the results of CBT, which provided further indication that change in cognition may be the most important mechanism of change.

A large adolescent treatment study
The Treatment of Adolescents With Depression Study (TADS)[2] represents the largest treatment study of adolescents with MDD to date. Before TADS, a critical limitation of CBT research was its isolation from developments in pharmacotherapy research.1 TADS was therefore designed to compare CBT, fluoxetine, and their combination as treatments for moderate to severe MDD in adolescents.2 Response rates for CBT and medication in earlier studies were about 50% to 60%, indicating considerable room for improvement in treatment outcomes. A primary hypothesis of TADS was that the combination of CBT and fluoxetine would be more effective than either therapy alone.

In TADS, 439 adolescents with moderate to severe MDD were randomly assigned to receive CBT, fluoxetine, a CBT and fluoxetine combination, or clinical management with a pill placebo for 12 weeks. The CBT used in TADS combined the 2 streams of American CBT for adolescents: skills training from the Lewinsohn model was embedded in individual psychotherapy sessions that followed a structure from the Beck model. (The TADS CBT model has been described by Rohde and colleagues.27)

Acute treatment results supported combination treatment as the most effective intervention, as measured by rate of change of scores on the Children's Depression Rating Scale-Revised assessed by an independent evaluator uninformed of treatment assignment.28,29 Fluoxetine, but not CBT, was also superior to placebo at the end of 12 weeks of treatment. A secondary analysis indicated superiority of combination treatment on speed of response rated by treating clinicians.30

The TADS team explored possible predictors or moderators of acute treatment outcome, based on a literature review of previous studies of psychotherapy or pharmacotherapy for childhood or adolescent internalizing disorders (anxiety or depression).31 Two of these were cognitive variables. One of these, hopelessness, proved to be a predictor of acute outcome. Lower levels of hopelessness at baseline predicted more favorable response to treatment than did higher levels of hopelessness, regardless of treatment arm.

A second cognitive variable was a moderator of acute outcome: cognitive distortion. Among adolescents with higher levels of cognitive distortion, the addition of CBT to fluoxetine led to improved results. By contrast, the addition of CBT to fluoxetine did not improve treatment outcome in those with lower levels of distortion. As noted by Curry and colleagues,31 one potential interpretation of these results is that CBT addresses a vulnerability to depressogenic thinking, but that its benefits for moderately to severely depressed adolescents are only evident in the short term when combined with fluoxetine.

A recent secondary analysis of the TADS trial compared the rate of cognitive change across the 12 weeks of treatment for each of the following conditions: combination, fluoxetine, CBT, and placebo.32 Cognitive change was measured using factor analytically derived scores representing cognitive distortions, cognitive avoidance, and solution-focused thinking, based on a previous analysis by Ginsburg and colleagues.33 Combination treatment led to a significantly greater rate of reduction in cognitive distortions and avoidance than did other conditions. Thus, when analyzing data with factor-derived scores, the pattern of cognitive change paralleled that for change in depression.

Analysis of treatment mediators in TADS is currently under way. In addition, more detailed analyses of specific cognitive variables including negative automatic thoughts, dysfunctional attitudes, and attributional style will track changes in these processes across the different treatments.

Cognitive variables that predict favorable treatment outcomes in adolescents with depression include higher levels of rational thoughts, less hopelessness, fewer negative thoughts, and fewer cognitive distortions at baseline.10,14,22 In an analysis of TADS data, cognitive distortions at baseline were also shown to moderate the effects of combined CBT and fluoxetine, such that adolescents with high levels of cognitive distortion were more likely to experience incremental benefits from CBT as part of combination treatment.31 In an analysis of the Brent study, hopelessness was found to mediate the predictive effects of referral source on outcome across 3 types of psychotherapy.14

The only variables that have been found to mediate the results of CBT specifically are negative automatic thoughts, and in a small subclinical sample, dysfunctional attitudes, which suggests that a reduction in these variables may be important mechanisms of change in CBT.17,23 Mediational analyses have not yet been conducted on the TADS data, which precludes conclusions about the mechanisms of change in pharmacological and combination treatments. Further research is needed to uncover the specific strategies used in therapy that reframe negative cognition, modify dysfunctional beliefs, and improve depressive symptoms.



1. Emslie GJ, Rush AJ, Weinberg WA, et al. A double blind, randomized, placebo-controlled trial of fluoxetine in children and adolescents with depression. Arch Gen Psychiatry. 1997;54:1031-1037.
2. Treatment for Adolescents With Depression Study Team. Treatment for Adolescents With Depression Study (TADS): rationale, design, and methods. J Am Acad Child Adolesc Psychiatry. 2003;42:531-542.
3. Weisz JR, McCarty CA, Valeri SM. Effects of psychotherapy for depression in children and adolescents: a meta-analysis. Psychol Bull. 2006;132:132-149.
4. Curry JF. Specific psychotherapies for childhood and adolescent depression. Biol Psychiatry. 2001;49: 1091-1100.5. Reinecke M, Ryan N, DuBois D. Cognitive-behavioral therapy of depression and depressive symptoms during adolescence: a review and meta-analysis. J Am Acad Child Adolesc Psychiatry. 1998;37:26-34.
6. Freeman A, Reinecke M. Cognitive therapy. In: Gurman AS, Messer SB, eds. Essential Psychotherapies: Theory and Practice. New York: Guilford Press; 1995: 182-225.
7. Lewinsohn PM, Clarke GN. Psychosocial treatments for adolescent depression. Clin Psychol Review. 1999; 19:329-342.
8. Beck AT, Rush J, Shaw BF, Emery G. Cognitive Therapy of Depression. New York: Guilford Press; 1979.
9. Brent D, Holder D, Kolko D, et al. A clinical psychotherapy trial for adolescent depression comparing cognitive, family, and supportive treatment. Arch Gen Psychiatry. 1997;54:877-885.
10. Lewinsohn PM, Clarke GN, Hops H, Andrews J. Cognitive-behavioral treatment for depressed adolescents. Behav Ther. 1990;21:385-401.11. Kazdin A, Nock M. Delineating mechanisms of change in child and adolescent therapy: methodological issues and research recommendations. J Child Psychol Psychiatry. 2003;44:1116-1129.
12. Leitenberg H, Yost L, Carroll-Wilson M. Negative cognitive errors in children. J Consult Clin Psychol. 1986;54:528-536.13. Beck A, Weissman A, Lester D, Trexler L. The measurement of pessimism: the Hopelessness Scale. J Consult Clin Psychol. 1974;42:861-865.
14. Brent D, Kolko D, Birmaher B, et al. Predictors of treatment efficacy in a clinical trial of three psycho-social treatments for adolescent depression. J Am Acad Child Adolesc Psychiatry. 1998;37:906-914.
15. Kolko D, Brent D, Baugher M, et al. Cognitive and family therapies for adolescent depression: treatment specificity, mediation, and moderation. J Consult Clin Psychol. 2000;68:603-614.
16. Hollon S, DeRubeis R, Evans M. Causal mediation of change in treatment for depression: discriminating between nonspecificity and noncausality. Psychol Bull. 1987;102:139-149.
17. Ackerson JD, Scogin F, McKendra-Smith N, Lyman RD. Cognitive bibliotherapy for mild and moderate adolescent depressive symptomatology. J Consult Clin Psychol. 1998;66:685-690.
18. Clarke GN, Lewinsohn PM, Hops H. Adolescent Coping With Depression Course. Eugene, OR: Castalia Press; 1990.
19. Clarke GN, Rohde P, Lewinsohn PM, et al. Cognitive- behavioral treatment of adolescent depression: efficacy of acute group treatment and booster sessions. J Am Acad Child Adolesc Psychiatry. 1999;38: 272-279.
20. Rohde P, Clarke G, Mace D, et al. An efficacy/effectiveness study of cognitive-behavioral treatment for adolescents with comorbid major depression and conduct disorder. J Am Acad Child Adolesc Psychiatry. 2004;43:660-668.
21. Rohde P, Seeley P, Kaufman N, et al. Predicting time to recovery among depressed adolescents treated in two psychosocial group interventions. J Consult Clin Psychol. 2006;74:80-88.
22. Kaufman N, Rohde P, Seeley J, et al. Potential mediators of cognitive-behavioral therapy for adolescents with comorbid depression and conduct disorder. J Consult Clin Psychol. 2005;73:38-46.
23. Hollon SD, Kendall PC. Cognitive self-statements in depression: development of an automatic thoughts questionnaire. Cognit Ther Res. 1980;4:383-395.
24.Weissman M, Beck A. Development and validation of the Dysfunctional Attitude Scale. Presented at: the 12th Annual Meeting of the Association for the Advancement of Behavior Therapy; November 1978; Chicago.
25. Munoz R, Ying Y, Bernal G, et al. Prevention of depression with primary care patients: a randomized controlled trial. Am J Community Psychol. 1995;23: 199-222.
26. Allart-van Dam E, Hosman C, Hoogduin C, Schaap C. The Coping With Depression Course: short-term outcomes and mediating effects of a randomized controlled trial in the treatment of subclinical depression. Behav Ther. 2003;34:381-396.
27. Rohde P, Feeny NC, Robins M. Characteristics and components of the TADS CBT approach. Cogn Behav Pract. 2005;12:186-197.
28. March J, Silva S, Petrycki S, et al; Treatment for Adolescents With Depression Study (TADS) Team. Fluoxetine, cognitive-behavioral therapy, and their combination for adolescents with depression: Treatment for Adolescents With Depression Study (TADS) randomized controlled trial. JAMA. 2004;292:807-820.
29. Pozanski E, Freeman L, Mokros H. Children’s Depression Rating Scale–Revised. Psychopharmacol Bull. 1985;21:979-989.
30. Kratochil C, Emslie G, Silva S, et al. Acute time to response in the Treatment for Adolescents With Depression Study (TADS). J Am Acad Child Adolesc Psychiatry. 2006;45:1412-1418.
31. Curry JF, Rohde P, Silva S, et al. Predictors and moderators of acute outcome in the Treatment for Adolescents With Depression Study (TADS). J Am Acad Child Adolesc Psychiatry. 2006;45:1427-1439.
32. Curry JF, Becker SB, Sanchez CS, et al. Processes of cognitive change in adolescents receiving treatment for depression. Poster presented at: the 53rd Annual Meeting of American Academy of Child and Adolescent Psychiatry; 2006; San Diego.
33. Ginsburg GS, Silva SG, Tonev S, et al. Cognitive measures of adolescent depression: unique or unitary constructs? Manuscript under review.

Evidence-Based References
Kaufman N,Rohde P,Seeley J,et al.Potential mediators of cognitive-behavioral therapy for adolescents with comorbid depression and conduct disorder. J Consult Clin Psychol.2005;73:38-46.
March J, Silva S, Petrycki S, et al; Treatment for Adolescents With Depression Study (TADS) Team.Fluoxetine, cognitive-behavioral therapy,and their combination for adolescents with depression: Treatment for adolescents With Depression Study (TADS) randomized control trial JAMA. 2004; 292:807-820