CBT with Children and Adolescents: More May Be Better

Background and Purpose: Existing meta-analyses have established CBT as an EBP for childhood and adolescent mental health disorders. The clinical relevance of this literature is enhanced by the identification of moderators of treatment effectiveness. This study examined the structural characteristics of CBT related to treatment dose, specifically, session length in minutes, number of sessions, and use of booster sessions. The purpose of this study was to identify conditions under which treatment dose was related to stronger intervention effects.

Methods: A random effects meta-analysis was conducted with RCTs of cognitive behavioral therapy published from January 1, 2000 – December 31, 2009 (k=72). The primary inclusion criteria was the use of CBT with children or adolescents who were diagnosed with any mental health disorder. Data coded from articles included study characteristics such as sample characteristics (age, gender, diagnostic characteristics), CBT modality (group, family, individual, mixed), CBT dose (length of sessions, number of sessions, booster sessions), study quality (intent to treat analysis, blind), outcome measures, and pre-test/post-test means and standard deviations. Articles were coded by trained graduate research assistants who demonstrated high levels of reliability (interrater agreement > 90%). Cohen's d was calculated for pre-test/post-test change for each outcome reported which were then averaged across outcomes to yield a single effect size per study.

Results: Consistent with earlier research, CBT showed positive effects on study outcomes (k=63, d=.76, p<.05). Effect sizes were significant for group interventions (k=21, d=.85, p<.05), whereas effect sizes were not statistically significant for other modalities (i.e., individual and family modality). Studies that employed more than 12 sessions had statistically significant effect sizes (k=26, d=1.01, p<.05) and studies that employed longer sessions (longer than 60 minutes) also had statistically significant effect sizes (k=24, d=.90, p<.05). Similarly, studies that employed booster sessions had large effect sizes (k=17, d=1.11, p<.05). Studies that employed less than 12 sessions (k=23, d=.82, ns), whose sessions lasted for less than 60 minutes (k=25, d=.96, ns) and studies that did not employ booster sessions (d=50, d=.88, ns) did not achieve statistical significance. Follow-up analysis examined interactions between youth demographic and diagnostic characteristics.

Implications: Results of this study suggest that higher dose interventions may increase the effects of CBT on the mental health of children and adolescents. Toward this end, the findings related to booster sessions may stand out. Results of this study suggest strongly for the inclusion of follow-up treatment sessions devoted to the maintenance of treatment gains. Thus, findings of this study argue for a higher treatment dose when implementing CBT with children and adolescents in treatment for mental health disorders.