Optimizing the integration between human and digital support in a meditation app for depression and anxiety

Project Summary / Abstract The United States is experiencing a mental health crisis and most individuals who could benefit from treatment do not receive it. Mobile health (mHealth) technology has the potential to reduce the public health burden of depression and anxiety, two extremely common mental illnesses. Meditation apps are by far the most widely used mental health apps, but they are plagued by low rates of engagement and modest effect sizes relative to human-delivered interventions. The addition of small amounts of instructor support and digital prompts may increase the effectiveness of meditation apps while retaining the efficiency and scalability of the mHealth delivery format. There is a need for careful, programmatic research examining the effects of these components and identifying for whom and when they may be most beneficial. The proposed study will begin addressing this gap through a hybrid factorial and micro-randomized trial (MRT) to optimize augmentation of the Healthy Minds Program (HMP), a widely-used meditation app with promising empirical support. Participants downloading HMP with clinically elevated symptoms of depression and/or anxiety (n=688) will be randomly assigned via a 2X2X2 factorial design to one of eight conditions crossing introductory coaching (yes/no), on-demand email-based coaching (yes/no), and digital prompts (yes/no). Those in the digital prompts condition will be further micro-randomized daily via a MRT to receive or not receive a digital prompt encouraging them to engage with HMP. Psychological distress (composite of depression and anxiety) will be assessed at baseline, over the 4-week intervention period, and at 6-month follow-up. Key candidate mechanisms linking instructor support and digital prompts with outcomes will also be assessed. Aim 1a will evaluate the effect of introductory coaching, email-based coaching, and digital prompts on month 6 psychological distress (primary outcome) and candidate mechanisms (therapeutic alliance, HMP engagement). Aim 1b will identify baseline characteristics that moderate the effects of these components. Aim 2a will investigate whether receiving (vs. not receiving) a digital prompt on a given day reduces proximal (same day) psychological distress and daily HMP engagement. Aim 2b will investigate whether the proximal effects of receiving (vs. not receiving) a daily digital prompt are moderated by baseline and time-varying information. Aim 3 will examine whether therapeutic alliance and HMP engagement mediate the effects of introductory coaching, email-based coaching, and digital prompts on month 6 psychological distress. The proposed research will result in an optimized version of HMP that can be implemented at scale. Results will also clarify who is most likely to benefit from these components and under what conditions just-in-time prompting is most beneficial.