Efficacy of a Digital Cognitive Behavioral Intervention Embedded in Routine Care Compared to Treatment as Usual in Adolescents and Young Adults with Moderate Depressive Symptoms: A Randomized Controlled Trial

Objective The number of adolescents and young adults (AYA) being identified with depressive symptoms is increasing. Unfortunately, there is a paucity of behavioral health (BH) care options for this population, and obtaining care is more challenging in underserved areas. Digital cognitive behavioral interventions (dCBIs) have been effective in decreasing self-reported depressive symptoms in pediatric patients. We examined the effectiveness of augmenting therapy with dCBI as compared to therapy alone in a multi-site randomized control trial (RCT). Method A multi-site RCT was conducted among patients, aged 16-22 years, being treated for depressive symptoms within BH collaborative care provided in pediatric primary care settings in Pittsburgh PA, Boston MA and San Diego CA. Patients were randomized to treatment as usual (TAU) or TAU+dCBI. Data obtained from the dCBI app included the number and type of sessions completed, and engagement with an asynchronous coach within the dCBI. Outcomes of interest were scores on the Patient Health Questionnaire (PHQ-9), the Children's Depression Rating Scale Revised (CDRS-R), the Generalized Anxiety Disorder (GAD-7), the Satisfaction with Life Scale (SWLS), and Children's Global Assessment Scale (CGAS), obtained at baseline, 6 and 12 weeks. Intent-to-treat analyses used linear mixed effects models to test the effect of treatment group, adjusting for gender, race, and other treatment (antidepressants, behavioral therapy) at baseline. Additional per protocol analyses examined treatment differences when the analysis was restricted to engaged users. Sensitivity analyses further examined the effect among those without other behavioral health treatment at baseline. Results 185 participants were randomized and enrolled: 73 in TAU and 112 in TAU+dCBI. Mean age 17.8 (1.66), 57% white, 75% female, and 48% with public insurance. Among participants randomized to TAU+dCBI, the mean number of techniques used was 5.5 + 5.7. Sixty-one percent (n=68) used at least 3 techniques and were considered engaged users. In intent-to-treat (n=112) and per protocol (n=107) analyses, there were no statistically significant differences between TAU and TAU+dCBI on any of the outcome measurements. However, among those without other treatment at baseline (n=36), those randomized to TAU+dCBI had a significantly greater decrease in CDRS-R over 12 weeks relative to TAU. In this subgroup of 36, the estimated within-person change for TAU+dCBI was d (95% CI) = 0.86 (0.59, 1.13) versus only 0.19 (-0.20, 0.58) for TAU (β = -0.21, 95% CI=-0.35, -0.06, p=0.01, for the group-by-time interaction term). Conclusion We did not find a significantly different change in depressive symptoms between patients randomized to TAU+dCBI versus TAU. However, sensitivity analyses suggested that the app may be more effective among subgroups without other treatment options for behavioral health. As only 60.7% of those randomized to the TAU+dCBI study arm engaged with the app, future studies should include ways to promote engagement. The potential effects of dCBI should be further explored as apps offer a convenient, non-stigmatizing means of accessing therapy techniques to supplement traditional therapy. Given the digital acceptance and familiarity among AYA, this could be a promising means of providing needed access to therapy at a time when therapy options are scarce. Clinical trial registration information Study to Compare the Use of a Behavioral Health App Versus Care and Usual for 16-22 Year Olds with Depression; https://clinicaltrials.gov/study/NCT05159713