tigative teams have published controlled trials in youth who meet diagnostic criteria for depression and are often clinically referred. We will begin our discussion with these more recent and well-developed treatment programs, and then briefly review historical work with subclinical samples of community youth. A summary of studies on clinic samples is shown in Table 2.1.

Wood, Harrington, and Moore (1996) compared the impact of a five-to eight-session CBT intervention with a comparable dose of relaxation training in the treatment of early-to middle-adolescent outpatients with depressive disorders, with 54% of the CBT group and 26% of the relaxation group remitting by the end of treatment. Similar results were obtained on self-report measures of depressive symptoms, self-esteem, and general psychosocial adjustment. Upon 6-month follow-up, the groups had converged, because of continued improvement in the relaxation group and symptomatic relapse in the CBT group. Younger age of diagnosis and higher level of functioning at intake were associated with better outcome (Jayson, Wood, Kroll, Fraser, Harrington, 1998). The addition of a median of six monthly booster CBT sessions after acute treatment resulted in a much lower relapse rate than acute treatment alone (20% vs. 50%; Kroll, Harrington, Jayson, Fraser, & Gowers, 1996).

Using a similar CBT treatment package to that of Wood et al. (1996), Vostanis, Feehan, Grattau, & Bickerton (1996) randomized depressed outpatients to either individual CBT or an attention-placebo condition termed nonfocused intervention (NFI). CBT and NFI were equivalent with regard to the proportion not meeting depressive criteria at the end of treatment (87% vs. 75%) and at 9-month follow-up (71% vs. 75%) (Vostanis et al., 1996a). However, 46% of patients across conditions reported experiencing a “depressive episode” at some point during the follow-up period (Vostanis et al., 1996). On average, patients in both the CBT and NFI conditions attended six sessions with a therapist, but the range of sessions was from two to nine, occurring over a 1-to 5-month period.

Brent and colleagues (1997) tested CBT, derived from Beck et al. (1979), against systemic behavior family therapy (SBFT) and a nondirective supportive therapy (NST), using a primarily clinically referred sample (2 3 vs. 1 3 from newspaper advertisements) of depressed adolescents. In comparison to the treatment used by Wood et al. (1996) and Vostanis et al. (1996), these treatments were much longer and more regular (12 to 16 weekly sessions).

At posttreatment assessment, significantly fewer of those subjects receiving CBT (17%) than NST (42%) continued to have diagnosable MDD. Remission, as defined by the absence of MDD and at least three consecutive BDI scores < 9, was more common in the CBT cell (60%) than in either SBFT (38%) or NST (39%). Reductions in suicidality and improvements in general psychosocial adjustment were not different across groups. Cognitive behavior therapy resulted in greater change in cognitive distortions than did either SBFT or NST, although changes in depressive symptoms were not mediated by changes in cognitive style (Kolko, Brent, Baugher, Bridge, & Birmaher, 2000). Across treatment cells, poorer response was predicted by greater cognitive distortion, more severe depression at intake, and referral to an advertisement rather than by clinical

referral. Removal from the study and dropout were predicted by double depression (depression comorbid with dysthymia) and greater hopelessness, respectively (Brent et al., 1998).

In the Brent et al. trial (1998), comorbid anxiety was associated with more robust response to CBT, and maternal depressive symptoms were associated with a poorer response to CBT (Fig. 2.1). At posttreatment, CBT was superior to NST and SBFT, even in the presence of multiple adverse predictors. This suggests that CBT, in addition to being efficacious under controlled conditions, may also be efficacious for use in real-world service settings with clinically complex cases (Fig. 2.2). Lifetime suicidality was associated with a poorer response to supportive treatment, where CBT appeared robust (Barbe, Bridge, Birmaher, Kolko, & Brent, 2004b). A history of sexual abuse, however, was associated with a poorer response to CBT (Barbe, Bridge, Birmaher, Kolko, & Brent, 2004a). Subjects who entered the study via an advertisement fared much better than clinically referred subjects despite having nearly identifiable demographic, clinical, and family characteristics, and were 12 times less likely to have a recurrence of their depression on follow-up (Birmaher et al., 2000; Brent et al., 1998). The differential outcome of those who entered via an advertisement was in part mediated by hopelessness (Brent et al., 1998). At 2-year follow-up, differences between treatment groups on the presence of current MDD were not significant, although the descriptive data again favor CBT (6%) over SBFT (23%) and NST (26%). Recurrence of depression over the 2-year follow-up period was predicted by greater severity of depression symptoms at intake, higher levels of parent–child conflict, and a lifetime history of sexual abuse (Birmaher et al., 2000).

Table 2.2 shows a summary of the studies per

formed with community samples. The following set of studies used community samples of youth with diagnostic depression recruited either from schools or via advertisements.

Lewinsohn, Clarke, Hops, and Andrews (1990) randomized 59 participants to either Coping with Depression–Adolescent version (CWD-A) alone, CWD-A plus the parent group (consisting of 7 weekly sessions), or a wait-list control. The CWD-A consisted of 14 two-hour psychoeducationally oriented group sessions delivered over 7 weeks. The focus of CWD-A is on increasing social skills, pleasant events and activities, and problem-solving and conflict resolution skills, while reducing anxious and depressive cognitions. At the conclusion of treatment, 43% of the adolescent-only group and 48% of the adolescent-plus-parent group no longer met diagnostic criteria, compared to only 5% of the wait-list control subjects. Significantly decreased scores on self-report depression measures were also obtained for both treatment groups relative to the wait-list control group. Treatment gains persisted at 1-, 6-, 12-, and 24-month follow-up. The adolescent-plus-parent group did not result in better outcomes than the adolescent-only condition. Poor outcome was associated with greater depressive symptomatology, comorbid anxiety symptoms, and greater cognitive distortions at intake (Clarke, Hops, Lewinsohn, Andrew, & Williams, 1992).

Clarke, Lewinsohn, Rohde, Hops, and Seeley (1999) replicated these results, with 65%, 69%, and 48% of subjects in the adolescent-only, adolescent-plus-parent, and wait-list control conditions, respectively, no longer meeting diagnostic criteria for MDD or dysthymia posttreatment. In this trial, there was a re-randomization to a one-to two-session booster condition that did not reduce the rate of depression recurrence for those who had remitted by the end of treatment. However, for patients who had not yet recovered from depression at the end of the acute treatment phase, booster sessions did accelerate their rate of recovery. Across both studies assessing the effects of CWD-A with diagnosed samples, a positive treatment response was predicted by a less severe initial depression, higher initial engagement in and enjoyment of pleasant activities, and fewer irrational thoughts (Clarke et al., 1992; Rohde, Clarke, Lewinsohn, Seeley, & Kaufman, 2001).

Clarke et al. (2002) examined the efficacy of group CBT for depressed adolescents whose parents were depressed. The sample was obtained through sampling the offspring of depressed parents. Subjects were assigned to either a 16-session group CBT program plus treatment as usual (N = 41) or treatment as usual alone (N = 47). Assessments were conducted at intake, after treatment, and at 12-and 24-month follow-up. According to intent-to-treat analyses, there was no advantage of group CBT over treatment as usual on rate of depressive disorders, depressive symptoms, nondepressive symptoms, or functional outcomes at the end of treatment or over the follow-up period. These findings are consistent with those of Brent et al. (1998), who also found that CBT, delivered individually, was no more effective than either family or supportive therapy in the face of significant maternal depressive symptoms.

Rohde et al. (2001) examined the impact of comorbidity on treatment outcome in an aggregated sample of depressed adolescents treated in the Lewinsohn et al. (1990) and Clarke et al. (1999) samples. Of 151 subjects enrolled in these trials, 40% had one or more comorbid diagnoses. Those with comorbid anxiety disorders showed a greater decrease in depression scores at posttreatment, similar to Brent et al. (1998). Lifetime substance abuse was associated with a slower time to recovery. Subjects with attention-deficit hyperactivity disorder (ADHD) and disruptive

behavior disorders were more likely to experience a depressive recurrence. In general, the authors concluded that the effects of CBT were robust to the impact of comorbidity.

There are several community-based studies of symptomatic children and adolescents in which CBT is compared with other active treatments (e.g., relaxation treatment, self-control, modeling) as well as a wait-list control (Butler, Meizitis, & Friedman, 1980; Kahn, Kehle, Jenson, & Clark, 1990; Liddle & Spence, 1990; Reynolds & Coats, 1986; Stark, Reynolds, & Kaslow, 1987; Weisz, Thurber, Sweeney, Proffitt, & LeGagnoux, 1997). In general, all active interventions are superior to wait-list control conditions and the impact is sustained on follow-up, but with few exceptions, CBT showed no differential impact compared to the other active interventions. These studies are important because they are some of the only CBT studies of depressive symptoms in children, are usually delivered in a group format, and, therefore, may guide future efforts to prevent early-onset depression.

Clarke and colleagues (1995) tested the efficacy of an adaptation of CWD-A, termed “coping with stress” (CWS), in preventing diagnosable depression in adolescents with subsyndromal depressive symptoms, but without a diagnos-able depressive disorder. One hundred fifty high school students who scored high on the Center for Epidemiological Studies–Depression scale (CES-D) but failed to meet diagnostic criteria for a current depressive disorder were randomly assigned to the 15-session CWS or “usual care.” At 12 months postintervention, fewer cases of depressive disorder (15%) had developed in the experimental group than in the control group (26%). Clarke et al. (2001) conducted a randomized clinical trial comparing the efficacy of group CBT plus treatment as usual versus treatment as usual alone for the adolescent offspring of depressed parents deemed to be at risk for the development of a depressive disorder (e.g., presence of subsyndromal symptoms of depression and/or a history of past mood disorder). Subjects could not meet criteria for a current mood disorder. Ninety-four adolescent subjects were randomized to the experimental treatment (N = 45) and to usual care (N = 49). At 12-month follow-up, the cumulative rate of incident major depression was much lower in the experimental condition than that with treatment as usual (9.3% vs. 28.8%). The odds of new-onset depression was nearly six times higher in the control than in the experimental group, after adjusting for gender, age, past history of depression, and intake self-reported symptoms of depression (adjusted odds ratio [OR] = 5.6, 95% confidence interval [CI] = 1.6-20.4). Significant treatment by time interactions favoring the experimental treatment were found for self-reported depression and global functioning. On average, those exposed to the experimental group experienced 33 fewer depressed days than those in the treatment-as-usual condition.

One study examined the effect of CBT “bibliotherapy,” or therapy involving pertinent reading materials, on adolescents with mild to moderate depression who were symptomatic volunteers (Ackerson, Scogin, McKendree-Smith, & Lyman, 1998). Twenty-two subjects were assigned to either reading the Feeling Good book by Burns (1980) or to a 4-week delayed-treatment condition, after which they received the bibliotherapy condition. Significant treatment by time interactions, favoring the experimental condition, were found for interview-, parent-and self-rated depression, and dysfunctional attitudes. Gains for the experimental condition were maintained at 1-month follow-up, and there was even evidence of continued improvement. Approximately 60% (59 to 64%) of subjects experienced a clinically significant improvement in their symptomatology, with evidence of sustained and even continued improvement on follow-up. Partial support was found for mediation of changes in depressive symptomatology by changes in dysfunctional thinking.

Although CBT interventions have received the most attention from researchers of youth de

pression, there are a number of clinical trials that focus on changing the interpersonal aspects of depression. Interpersonal psychotherapy for depressed adolescents, family therapy, and social skills training have all been investigated in at least one controlled trial (Mufson et al., 1994; Diamond et al., 2002; Fine, Forth, & Gilbert, 1991). All three treatments target behavioral and environmental processes, although the social domain targeted by the interventions differs, as do the techniques used to affect change.

Interpersonal psychotherapy for adolescents (IPT-A), an adaption of IPT, is a time-limited, focused psychotherapy that addresses common adolescent developmental issues that are closely related to depression: separation from parents, authority and autonomy issues in the parent–teen relationship, development of dyadic interpersonal relationships, peer pressure, loss, and issues related to single-parent families. Interpersonal therapy has been adapted and tested for depressed adolescents (Mufson et al., 1994; Mufson, Weissman, Moreau, & Garfinkel, 1999). In a controlled, 12-week, clinical trial of IPT-A, 48 adolescents (12–18 years old) with MDD were randomly assigned to either weekly IPT-A or biweekly to monthly 30-minute sessions of clini-cal monitoring. The sample was largely Hispanic and female. Thirty-two of the 48 patients completed the protocol (21 IPT-A-assigned patients and 11 patients in the control group). At termination, a much lower proportion of those in IPT-A still met criteria for major depression (12.5% vs. 41.6%). As measured by the Clinical Global Improvement scale, 95.5% of those treated with IPT-A were rated as significantly better than at intake, compared to 61.5% of those treated with clinical management. Moreover, only 4.5% of those treated with IPT-A were rated as significantly worse, compared to 23.1% of those assigned to clinical monitoring. Pa-tients who received IPT-A reported a significant decrease in depressive symptoms and greater improvement in overall social functioning, functioning with friends, and problem-solving skills. In the intent-to-treat sample, 18 (75%) of 24 patients who received IPT-A compared with 11 patients (46%) in the control condition met recovery criterion (HAM-D ≤6) at Week 12. These preliminary findings support the feasibil ity, acceptability, and efficacy of 12 weeks of IPT-A in acutely depressed adolescents in reducing depressive symptoms and improving social functioning and interpersonal problem-solving skills.

Mufson and colleagues assessed the effectiveness of IPT-A in school-based mental health clinics in New York City: they randomized 63 depressed adolescents referred for a mental health intake visit to IPT-A or treatment as usual, both performed by the clinic staff. The adolescents had either MDD, dysthymia, depressive disorder not otherwise specified (NOS), or adjustment disorder with a depressed mood. At termination, the adolescents in the IPT-A treatment experienced significantly higher symptomatic relief than the treatment-as-usual group (Mufson et al., 2004)

Rossello and Bernal (1999) compared the efficacy of a 12-week, individually administered CBT program to a similar dose of IPT and wait-list control in adolescents with diagnosed MDD and/or dysthymia referred by school personnel. Both active interventions were adapted to be culturally appropriate for use with Puerto Rican youth. Attendance problems occurred in both treatments, as only 68% of IPT cases and 52% of CBT cases completed over seven treatment sessions. To assess outcome, a clinical cutoff was selected on the Children's Depression Inventory (CDI) that was approximately 3 points lower than the mean intake CDI score for the sample. As measured by these criteria for clinically significant change (Jacobson and Truax, 1991), 59% of adolescents in the CBT condition and 82% of IPT cases achieved clinically significant improvement in depression symptoms by posttreatment; data were not provided for the wait-list control condition. Using an unspecified normative cutoff point for the CDI, the authors indicated that 56% of the IPT cases, 48% of the CBT cases, and 61% of wait-list cases were “severely depressed” at intake. At posttreatment, these percentages were 11%, 24%, and 34%, respectively, and 17% and 18% at 3-month follow-up (for IPT and CBT, respectively). Although this investigation has several methodological difficulties (e.g., substantial attrition and no intent-to-treat analyses conducted), these results provide some of the first information on the efficacy