Adaptive Messaging to Support Depression Self-Management

The primary goal of this prospective randomized controlled trial is to pilot all trial procedures for a larger, fully-powered trail. The investigators will enroll 100 participants out of consideration for feasibility. Among individuals with elevated depression symptoms, the investigators will compare the effects of 8 weeks of an adaptive text messaging intervention (n=40), 8 weeks of a non-adaptive text messaging intervention (n=40), and a psychoeducation-only control group (n=20). The investigators will include a 6-month post-treatment follow-up to verify that treatment gains are maintained. The study will enroll individuals who have at least moderate symptoms of depression as indicated by Patient Health Questionnaire-9 (PHQ-9) scores greater than or equal to 10. The investigators will recruit participants through digital methods (e.g., advertising on social media platforms like Facebook and Instagram). Recruitment advertisements will direct potential participants to a centralized recruitment website that provides additional information about the IRB-approved protocol. Participants will represent likely users of a mobile-based depression intervention. They will also reflect the racial and ethnic diversity of the United States. Interested individuals will confirm eligibility by completing an online screening survey. Participants will self-report their specific recruitment source as part of this survey and will provide contact information. If eligible participants confirm interest, study staff will obtain informed consent. Computer-generated randomization will be created prior to the start of the study to avoid allocation bias. Random block sizes of 5 and 10 will be allocated in a 2:2:1 ratio to adaptive messaging, non-adaptive messaging, and the psychoeducation-only control group. Both messaging arms will receive 8 weeks of daily text messaging. Messages will prompt learning and applying cognitive and behavioral self-management strategies. Whereas the non-adaptive messaging system will randomly select daily messaging styles (representing a variety of self-management strategies and message tones), the adaptive messaging system will apply reinforcement learning to select daily messaging styles to maximize user engagement. Text messages will also be used to solicit feedback such as message ratings to inform the reinforcement learning algorithm (if applicable). The psychoeducation control group will receive a few brief text messages per week providing weblinks to access psychoeducation information. Administration of study measures will occur as part of the research protocol via REDCap online at baseline, weeks 4 and 8, and 3-month and 6-month follow-up. The primary outcomes is depression symptom severity (PHQ-9). Secondary outcomes are: anxiety severity (GAD-7) and suicidality (DSI-SS). The investigators will additionally assess subjective and objective engagement with the intervention (messaging arms only). The investigators will collect qualitative data with a subset of 10 participants in each messaging arm via user interviews. Using an Intention to Treat paradigm, the investigators will use longitudinal linear mixed-effects models to test our hypothesis that: H1: Relative to the psychoeducation-only control, messaging interventions will reduce depression (H1a) and anxiety and suicidal ideation (H1b). H2: Relative to non-adaptive messaging, adaptive messaging will reduce depression (H1a) and anxiety and suicidality (H2b). H3: Adaptive messaging will reduce depression relative to non-adaptive messaging by producing greater objective and subjective engagement.