(PS12-C51) Behavioral Activation as a Potential Treatment Mechanism of Improvement in Depression and Anxiety Among Users of a Digital Mental Health App

Background: Digital mental health interventions (DMHIs, such as apps) have proliferated in an attempt to scale treatment and improve access to care for common mental health conditions such as depression and anxiety. Although the evidence base for DMHIs is growing, questions about treatment effects and mechanisms remain, and many commercially available apps remain untested. This work aimed to examine the potential treatment mechanism of behavioral activation (BA) as a mediator of improvement in depression and anxiety among users of a popular DMHI.

Method: Participants were 280 adults (Mage = 32.47 [9.32] years) enrolled in a fully-remote randomized controlled trial testing the intervention effects of a DMHI on symptoms of depression and anxiety. Participants were randomized to either the DMHI treatment group (Sanvello; n = 79) or waitlist control (WLC; n = 201). Sanvello is a leading commercially available DMHI that employs cognitive behavioral therapy strategies to help users manage their anxiety, depression, and stress over successive self-guided modules. Previously presented main outcomes demonstrated improvement in depression and anxiety among those randomized to use Sanvello, relative to those in WLC, over 8 weeks. Here we tested BA (assessed by the Behavioral Activation for Depression Scale [BADS]) as a potential mediator for the relationships between treatment group and symptoms of depression and anxiety. Linear mixed effects models examined BADS course over 8 weeks between groups. Separate mediation models tested for change in depression (PHQ-9) and anxiety (GAD-7) outcomes. BADS change score was calculated as the change from baseline to week 4 follow-up. Changes in depression and anxiety were calculated as the change in PHQ-9 and GAD-7 scores from week 4 to week 8 follow-up assessments to account for the temporal requirement of a change in the mediator (BADS) prior to change in outcome (PHQ-9 and GAD-7).  

Results: Linear mixed effects models found that BADS improved over the course of 8 weeks for the treatment group relative to WLC. Subsequent mediation testing revealed that change in BADS score did not mediate relationships between treatment group and change in depression or anxiety outcomes at 8 weeks after controlling for change during the first 4 weeks of treatment. 

 



Conclusions: This study extends our understanding of the role of behavior change in psychotherapeutic change in DMHI, which is crucial for optimizing treatment outcomes. Contrary to our hypothesis, behavior change did not mediate therapeutic change in our sample of DMHI users relative to waitlist control. However, group differences observed in BADS improvement over 8 weeks remains an important finding given the purported role of activation and reduced avoidance in the treatment of both depression and anxiety. This may center the relevance of behavioral activation (BA) as an evidenced-based treatment for both depression and anxiety. An important limitation of the parent RCT is the lack of representation in the sample, limiting generalizability. In order to effectively scale acceptable and effective treatment, developers and researchers need to consider methods of engaging diverse and underserved populations.