Optimizing Digital Behavioral Treatment for Co-occurring Insomnia and Depression

Although efficacious behavioral treatments like cognitive behavioral therapy for insomnia (CBT-I) and for depression (CBT-D) exist for insomnia and depression disorders, their impact on co-occurring symptoms can be modest and residual symptoms often remain. In addition, scalable versions of these interventions (like computer accessible or phone app versions) are needed for U.S. military Veterans who lack access to first-line interventions because of rural or other resource-limited environments. Finally, little is known about heterogeneity of treatment effects; that is, variation in which interventions are best for which individuals. This comparative effectiveness trial is designed to develop an an individualized intervention rule to match the right digital intervention to the right person with co-occurring insomnia and depression.

The study will use digital versions of CBT-I and CBT-D found to be efficacious for their intended target. A total of 1,160 participants who are Veterans with both insomnia and depression will be randomized with equal allocation to five study arms: two single interventions (CBT-I or CBT-D); two sequenced interventions (CBT-I+D or CBT-D+I); and a mood monitoring condition that will serve as the control group. Importantly, after approximately 400 participants have been randomized, we will end randomization to the mood monitoring condition and all subsequent participants (approximately 800) will then be randomized across the four remaining treatment conditions. Assessments occur at baseline, mid-treatment, post-treatment, at 3-month and 6-month follow-ups. The primary endpoint will be remission of both depression and insomnia at 3 months following the 12 week intervention period. All study activity will be conducted remotely including recruitment, treatment delivery and assessments, so that participants can be recruited nationwide.

The study will estimate average treatment effects on insomnia and depression remission at 3-months post-treatment (primary outcome) and 6-month maintenance of remission as well as treatment response. Significantly higher remission and response rates for sequenced rather than single interventions and lowest for the mood monitoring control condition are expected. Because variation in which interventions are best for which individuals are expected, study investigators will develop and evaluate an individualized intervention rule to determine an optimal approach for each participant in promoting 3-month remission of both disorders. A cutting-edge ensemble machine learning method will be used to do this.