Negative mood symptoms are highly prevalent in early psychosis and strongly linked to long-term functional impairment and risk of relapse. Smartphone-based interventions offer a scalable and accessible solution for delivering personalized mental health support in daily life.
The goal of this micro randomized trial (MRT) is to evaluate the feasibility of using passive smartphone sensor data and machine learning to identify optimal moments for delivering smartphone app-based brief mental health interventions such as mindfulness exercises and psychoeducation, specifically moments when individuals are both most in need of support and most receptive to it.
The primary aim of this pilot study is to assess the feasibility and acceptability of the MRT study design, including recruitment and retention rates, EMA adherence, and participant feedback gathered through qualitative exit interviews to inform future design improvements. A secondary aim is to evaluate the feasibility of training an idiographic machine learning model using passive smartphone sensor data to detect moments of heightened negative affect. An exploratory aim is to estimate preliminary intervention effect sizes to inform a future fully powered clinical trial and to identify contextual factors associated with participant engagement.
