A meta-analysis of emotional regulation outcomes in psychological interventions for youth with depression and anxiety

Abstract

Difficulties in applying emotional regulation (ER) skills are associated with depression and anxiety symptoms, and are common targets of treatment. This meta-analysis examined whether improvements in ER skills were associated with psychological treatment outcomes for depression and/or anxiety in youth. A multivariate, random-effects meta-analysis was run using metafor in R. Inclusion criteria included studies that were randomized controlled trials (RCTs) of a psychological intervention for depression and/or anxiety in patients aged 14-24, were peer reviewed, were written in English, measured depression and/or anxiety symptoms as an outcome and measured ER as an outcome. Medline, Embase, APA PsycInfo, CINAHL and The Cochrane Library were searched up to 26 June 2020. Risk of bias (ROB) was assessed using the Cochrane Collaboration Risk of Bias 2.0 tool. The meta-analysis includes 385 effect sizes from 90 RCTs with total N = 11,652. Psychological treatments significantly reduced depression, anxiety, emotion dysregulation (k = 13, Hedges' g = 0.54, P < 0.001, 95% confidence interval (CI) = 0.30-0.78) and disengagement ER (k = 83, g = 0.24, 95% CI = 0.15-0.32, P < 0.001); engagement ER also increased (k = 82, g = 0.26, 95% CI = 0.15-0.32, P < 0.001). Improvements in depression and anxiety were positively associated with improved engagement ER skills, reduced emotion dysregulation and reduced disengagement ER skills. Sensitivity considered study selection and publication bias. Longer treatments, group formats and cognitive-behavioural orientations produced larger positive associations between improved ER skills and reduced symptoms. ER skill improvement is linked to depression and anxiety across a broad range of interventions for youth. Limitations of the current study include reliance on self-report measures, content overlap between variables and inability to test the directionality of associations.

Figure 1. Three ways of measuring improvements in ER over the course of psychological treatment in the present synthesis

Note: A more detailed review, with additional examples of measures, can be found in the Supplementary Information. References corresponding to scales: DERS, ERSQ, ERQ, COPE, PHLMS, AAQ, and RRS.

Figure 2. PRISMA figure depicting the flow of studies in the present synthesis using a comprehensive search strategy.

Note: Two of the 88 studies in the included RCTs section had an additional study reported within the respective paper. Hence, there are a total of 90 RCTs.

Figure 3. Overall and domain-specific Risk of Bias ratings for each of the 90 RCTs included in the meta-analysis.

Note: Only the 90 RCTs received ROB ratings as most of the non-RCTs were single-arm designs and would automatically be rated as having “some concern”. Overall ratings are provided with the study characteristics of RCTs in Supplementary Table 1. A full breakdown of ROB ratings per study can be found in Supplementary Table 4.