Approximately 70% of all mental health problems appear before 25 years of age and often become chronic when not treated or ineffectively treated [1]. Such data raise questions about elevated depression rates in youth [2] as exemplified in the National Survey on Drug Use and Health (N=611,880), which found an increase in the depressive episode rates by 63% from 2009 to 2017 [2]. Youthful online engagement makes online intervention delivery attractive, particularly with possible reductions in costs, geographic barriers, and access inequities [3-6].

Cognitive behavioral therapy (CBT) is the best-validated psychotherapy [7], and in recent years has been coordinated with mindfulness meditation (CBT-M) following strong evidence of the combined efficacy [8,9]. Online CBT-M research with student and adult populations has yielded psychometric and neurophysiological [10-20] benefits in single-arm and randomized controlled trials (RCTs). These results join a growing world literature supporting online CBT efficacy, exemplified in a meta-analysis of 3876 RCT participants indicating that online CBT was significantly more effective than control conditions in reducing depressive symptoms (Hedges g=0.27) [21]. Individual RCTs have shown online CBT to be equally effective to in-person CBT in studies with large effect sizes, along with substantial remission rates for major depressive disorder [21].

These findings motivated a focus on assessing online CBT in patients concurrently receiving standard psychiatric care to examine whether online CBT and psychiatric treatment as usual (TAU) was superior to psychiatric TAU. Control participants received pharmacotherapy only when deemed appropriate by treating psychiatrists, and the TAU comparison accounted for the standard use of and response to antidepressant medications.

The key behavioral intervention in this study was online access to 24 CBT workbooks and 56 mindfulness instruction videos that supported metacognitive change and autonomic balance [22], which have been linked to improved mood and reduced anxiety [23]. Navigation coaching was supplied by students who were pursuing graduate degrees (MSc, MA, PhD) in kinesiology and health science, education, and psychology. Their group training (prior to and during the study) took place at a seminar (for 1.5 hours weekly) that focused on reviews of CBT and mindfulness-based clinical research supplemented by anonymized case discussions. One (cumulative) hour of coaching was provided to each participant weekly during 24 weeks (which included text-message exchanges with participants), and each coach received 1-hour weekly sessions of one-to-one supervision.

Navigation coaching has been increasingly applied to support the adoption of evidence-based, health-related behaviors as demonstrated in adherence to cancer screening, and exercise and diet regimens [24]. Our experiences with navigation coaching include assisting patients with type 2 diabetes to reduce hemoglobin A1c blood levels [18-20] and assisting individuals in undertaking colorectal cancer screening [25]. Positive outcomes suggest that navigation coaching can be applied to the treatment of depression in assisting the use of CBT and mindfulness methods to address depressive symptoms [26]. Past successes with text messaging–assisted interventions (with critical medical outcomes) [18-20] further influenced the emphasis on text messaging between navigator coaches and patients.

Our study objective was to assess whether online CBT-M with weekly interactions with a coach navigator and standard psychiatric care was superior to standard psychiatric care alone (as workbooks and videos were never provided without coach navigator assistance). We hypothesized that intervention participants would demonstrate significant improvements in primary outcomes when compared to the waitlist controls who received only standard psychiatric TAU.

The online CBT-M intervention was beneficial, given significant between-group differences in depression (BDI-II, QIDS), anxiety (BAI), pain (BPI), and mindfulness (FFMQ–Nonjudging subscale). Other notable between-group observations involved a 9% dropout rate in the intervention group that significantly differed from the 61% dropout rate in the TAU control group (the 61% dropout rate was estimated at ~14% above the mean for CAMH TAU). This difference suggests that the intervention had positive effects on participant retention. The intervention sample included a large subgroup with severe depression (n=10 participants, defined as severe by a BD-II score>29), 50% of whom were in remission (BDI<14) at the final (6-month) assessment. Of the 6 participants who exhibited moderate depression, 5 achieved remission, and of the 3 study participants with mild depression, 2 achieved remission.

Given the CBT-M intervention, it was notable that between-group differences were found in the Nonjudging subscale of the FFMQ that assesses the excess self-critical thinking associated with distress [43]. Intervention group participants engaged in significantly less self-critical self-judgment at the 6-month follow up than the TAU controls. Although the study sample size did not allow for mediation analyses [46], the between-group difference observed suggests that the mindfulness component of the CBT-M intervention was likely involved in the modification of depressogenic cognitions [2]. The between-group differences also appear linked to the self-acceptance emphasis in the CBT-M interventions employed.

Significant between-group differences were found in self-reported chronic pain as indicated in the BPI scale. The inclusion of a pain assessment reflects recent findings about the high comorbidity prevalence in depression with respect to chronic pain [47] and the possible efficacy of behavioral pain reduction methods [48,49]. Despite this study’s findings, the behavioral intervention literature on reductions in chronic pain remains sparse, and additional targeted studies in populations with pain and mental health difficulties are warranted.

Although attrition in psychiatric treatment has been linked to early improvements associated with medication changes [50], this explanation does not seem to apply in the current trial, as 4 of the 5 total control participants who discontinued after the 3-month midterm measures received no pharmacotherapy or no modifications in the prestudy pharmacotherapy established. The control participants who dropped out following the baseline assessment (before midterm, 3-month assessment; n=9) did not receive medication initiation or modification. The attrition difference also did not appear to be based on more severe baseline depression as the mean BDI-II depression score at baseline for TAU control participants (BDI-II=27.0) reflected milder depression symptoms than those of intervention participants (mean BDI-II=30.14).

Significant study strengths included the control comparison with a standard-care psychiatry intervention, delivered at the same institution, independently versus in combination with the experimental behavioral intervention. This resulted in detailed records of how pharmacological and behavioral interventions interacted, assisting estimations of independent and combined benefits. The study further controlled for the intervention-related placebo effects observed in 35%-40% of RCT participants exposed to TAU conditions [7]. This was also a necessary control for medication effects, given that individuals treated for depression show improvement with antidepressants alone [7]. In this study, CBT-M effects were clearly additive to TAU effects. Although the TAU-only group attrition rate can be seen as a study limitation, there were demonstrated associations between the CBT-M intervention and retention (ie, lower attrition in the experimental group) that indicated retention benefits associated with the behavioral treatment.

In recent meta-analyses focused on CBT-M delivery for depression, multiple CBT modalities have been assessed, notably individual, group, telephone-based, and guided self-help, all of which appear to be significantly more effective than waitlist and care-as-usual control conditions, and unguided self-help [51]. These analyses reflect the investigative search for the most cost-effective CBT delivery. In the context of current meta-analyses, our intervention can be characterized as combining telephone-based with guided self-help (online), with results that show significantly better outcomes than care-as-usual controls.

Key limitations of our study include a lack of participant blinding and the limited power associated with a small sample size. However, the HDRS assessment was undertaken by a rater blinded to group allocation. Although the between-group differences on the HDRS were trending toward significance (P=.09), they were not statistically significant. A final limitation is that the study psychiatrists administering TAU to control and intervention participants were not blind to which participants were in the intervention versus control groups, and this might have led to biased treatment.

Future studies comparing CBT-M and standard-care psychiatry would benefit from larger sample sizes, more complete blinding, and extended follow up after intervention conclusion (eg, 6-12 months). Despite these limitations, the results indicate that online CBT-M combined with TAU psychiatric treatment was an effective treatment for major depressive disorder and led to significantly greater reductions in BDI-II scores than TAU psychiatry alone.