The effect of depression on treatment outcome in social anxiety disorder: an individual-level meta-analysis

ABSTRACT

Social anxiety disorder (SAD) is highly comorbid with depression. In the present meta-analysis, we conducted the first individual-level examination of the association between pre-treatment depression and improvement in social anxiety symptoms during treatment. We identified eligible studies on cognitive behavior therapy (CBT) and pharmacotherapy for SAD and contacted authors to obtain individual-level data. We obtained these data from 41 studies, including 46 treatment conditions (n = 4,381). Our results showed that individuals who had high levels of depression at pre-treatment experienced greater decreases in social anxiety symptoms from pre- to post-treatment, but not at follow-up. When analyzing treatment modalities (individual CBT, group CBT, internet-delivered CBT, and pharmacotherapy), we found that depressive symptoms were associated with better post-treatment outcomes for individual CBT and internet-delivered CBT, but not for pharmacotherapy or group CBT. Our findings suggest that depression does not negatively affect treatment outcome in SAD and may even lead to improved outcomes in some treatment formats. Clinical implications of these findings are discussed.

KEYWORDS:

• Social anxiety disorder
• depression
• individual-level meta-analysis
• cognitive behavior therapy
• pharmacotherapy

Disclosure statement

No potential conflict of interest was reported by the authors.

Public health significance statement

This meta-analysis demonstrates that depression does not negatively affect treatment outcome in SAD and in fact, may even lead to improved outcomes in psychological treatments for the disorder.

Notes

1. We did not request the complete primary datasets from all authors. This is due to the difficulty to obtain such data (see Wicherts et al., 2006) for an empirical examination). Instead, based on their individual-level data, authors of studies included in the meta-analysis calculated the correlation between pre-treatment depression and change in social anxiety symptoms for each treatment condition. We obtained these correlations and aggregated them meta-analytically. This strategy is commonly referred to as a two-stage individual participant data (IPD) meta-analysis (Burke et al., 2016).

2. Some studies included in our meta-analysis reported an exclusion of individuals with high levels of depression. To examine whether such exclusion criteria affected the depression–outcome relationship, we coded studies for this practice. Specifically, for each study we coded whether that study excluded individuals due to high depression or did not exclude individuals due to high depression (creating a dichotomous variable).

3. We also examined whether depression at pretreatment could have affected dropout rates. We found that pretreatment depression did not significantly predict dropout at post-treatment (B = 0.32, SE = 1.97, t = 0.16, p = .871, 95% CI = −3.65 to 4.30)

4. We also examined whether depression at pretreatment could have affected dropout rates at follow-up. We found that pretreatment depression did not significantly predict dropout at follow-up (B = 6.62, SE = 8.22, t = 0.80, p = .447, 95% CI = − 12.82 to 26.07).