ABSTRACT
BACKGROUND: Cognitive-behavioural therapy (CBT) and counselling for depression (CfD) are recommended first-line treatments for depression. While they approach change differently, there is little understanding of the impact those approaches have on change during treatment. OBJECTIVES: This study aimed to identify whether CBT and CfD target different symptoms and explore the implications of modelling choices when quantifying change during treatment. METHODS: Symptom-specific effects of treatment were identified using moderated network modelling. This was a retrospective cohort study of 12 756 individuals who received CBT or CfD for depression in primary/community care psychological therapy services in England. Change was modelled several ways within the whole sample and a propensity score matched sample (n=3446). FINDINGS: CBT for depression directly affected excessive worry, trouble relaxing and apprehensive expectation and had a stronger influence on changes between suicidal ideation and concentration. CfD had a stronger direct influence on thoughts of being a failure and on the associated change between being an easily annoyed and apprehensive of expectation. There were inconsistencies when modelling change using the first and second appointments as the baseline. Residual score models produced more conservative findings than models using difference scores. CONCLUSIONS: CfD and CBT for depression have differential effects on symptoms demonstrating specific mechanisms of change. CLINICAL IMPLICATIONS: CBT was uniquely associated with changes in symptoms associated with anxiety and may be better suited to those with anxiety symptoms comorbid to their depression. When assessing change, the baseline should be the first therapy session, not the pretreatment assessment. Residual change scores should be preferred over difference score methods.
