Review of: Mantani A et al, J Medical Internet Res 2017;19(11):e373

Mobile phones have allowed the introduction of guided, self-help cognitive behavioral therapy (CBT) for depression with enhanced accessibility, efficiency, and affordability. Several meta-analyses suggest that computers can augment face-to-face psychotherapy and even work on their own through self-guided programs. Most of those studies involved patients with mild to moderate depression, which leaves open the question of how well this approach would work in more severe cases.

This study tested a self-guided mobile app in patients with moderate to severe depression who had not responded to at least 1 antidepressant trial. The Japanese app, called ­Kokoro, used cartoon characters to present concepts from CBT, including self-monitoring, behavioral activation, and cognitive restructuring.

The authors randomized 164 patients to an intervention group (medication switch plus Kokoro app) and control group (medication switch only). Although the treatments were not blinded, the outcomes were assessed with blinded raters.

After 9 weeks, the intervention group showed greater improvement in the Patient Health Questionnaire-9, the primary outcome measure (p < 0.001). Rates of remission (18% vs 10%) and response (32% vs 18%) were also greater, and the magnitude of the benefit compared favorably with the effect sizes seen in antidepressant ­trials.

In the second phase of the study, both groups were given access to the app for an additional 2 months. After that time, both groups had similar depression scores. The intervention group maintained their gains, and the control group caught up.

Most patients stayed engaged with the 8-session app, but that engagement was not entirely self-driven. Each week, participants received a brief, personalized email congratulating them on their progress.

TCPR’s Take
This study demonstrates significant benefits for this CBT app in difficult-to-treat depression. Its strengths include a randomized controlled design, blinded ratings, and high levels of engagement and completion. The main limitation is the lack of blinding in the treatment arm, which makes it difficult to rule out a placebo effect. By making changes to medications in both groups at the start of the trial, the authors attempted to minimize expectancy effects.

For clinicians, the main limitation may be the inaccessibility of the Japanese-language app, a common problem in this type of research. Most of the available mental health apps are untested, and most of the tested apps are not available. A reasonable substitute is Intellicare, a suite of CBT-based apps made free through NIMH funding. These apps looked promising in an uncontrolled study, and a large, randomized controlled trial of Intellicare is nearing completion (intellicare.cbits.northwestern.edu).