Mediators of Cognitive Behavioral Therapy for Anxiety-Disordered Children and Adolescents: Cognition, Perceived Control, and Coping

Abstract

The purpose is to investigate whether a change in putative mediators (negative and positive thoughts, coping strategies, and perceived control over anxious situations) precedes a change in anxiety symptoms in anxiety-disordered children and adolescents receiving cognitive behavioral therapy (CBT). Participants were 145 Dutch children (8–18 years old, M = 12.5 years, 57% girls) with a primary anxiety disorder. Assessments were completed pretreatment, in-treatment, posttreatment, and at 3-month follow-up. Sequential temporal dependencies between putative mediators and parent- and child-reported anxiety symptoms were investigated in AMOS using longitudinal Latent Difference Score Modeling. During treatment an increase of positive thoughts preceded a decrease in child-reported anxiety symptoms. An increase in three coping strategies (direct problem solving, positive cognitive restructuring, and seeking distraction) preceded a decrease in parent-reported anxiety symptoms. A reciprocal effect was found for perceived control: A decrease in parent-reported anxiety symptoms both preceded and followed an increase in perceived control. Using a longitudinal design, a temporal relationship between several putative mediators and CBT-outcome for anxious children was explored. The results suggest that a change in positive thoughts, but not negative thoughts, and several coping strategies precedes a change in symptom reduction and, therefore, at least partly support theoretical models of anxiety upon which the anxiety intervention is based.

Acknowledgments

This study was government funded. We report no potential conflicts of interest. We thank Marija Maric and David MacKinnon for their suggestions regarding the mediational analysis. We further like to thank all participating children, parents and therapists for their time and effort.

Notes

Note: Eta-squared is reported as effect size. Values with different subscripts differ significantly over time. ANOVA = analysis of variance; CSR = Clinical Severity Rating; RCADS = Revised Child Anxiety and Depression Scale; CATS-N/P = Children's Automatic Thoughts Scale–Negative/Positive; ACQ-C = Anxiety Control Questionnaire–Child; CCSC-R1 = Coping Strategies Checklist for Children–Revised; DPS = direct problem solving, PCR = positive cognitive restructuring, DS = distraction strategies, AS = avoidance strategies, SSS = support-seeking strategies; n/a = not available at T2.

^a n = 145.

¹We repeated all repeated measures ANOVAs including age as a covariate. There were no significant interaction effects between age and time and all results remained substantively unchanged.

Note: Model M → Y represents a model with unidirectional coupling from the mediator to the outcome variable, from in-treatment to posttreatment (γ_m1) or posttreatment to follow-up (γ_m2); model Y → M represents a model with unidirectional coupling from the outcome variable to the mediator, from in-treatment to posttreatment (γ_y1) or posttreatment to follow-up (γ_y2); No coupling represents a model with no dynamic coupling between the mediator and outcome variable. RCADS = Revised Child Anxiety and Depression Scale; CATS-N/P = Children's Automatic Thoughts Scale–Negative/Positive; ACQ-C = Anxiety Control Questionnaire–Child; CCSC-R1 = Coping Strategies Checklist for Children–Revised.

^a For these mediating variables the cognition-free RCADS was used as outcome variable.

Note: Numbers in parentheses represent critical ratios (CRs). CRs greater than 1.96 are salient.

*p < .001. ^∧ p < .01. ⁺ p < .05. Reported estimates are unstandardized.

²All analytic models were run including age as a covariate, and the pattern of results remained substantively unchanged (detailed results are available from the authors by request).