The Development and Systematic Evaluation of an AI-Assisted Just-in-Time-Adaptive-Intervention for Improving Child Mental Health

PROJECT SUMMARY/ABSTRACT Early childhood mental health problems constitute a significant public health concern with wide-ranging impacts on functioning both concurrently and later in life. Although childhood mental health is influenced by a variety of factors, the quality of relationships with caregivers plays a critical role. Critical, coercive, and conflictual parent- child interactions have been consistently linked with increased risk of externalizing and internalizing symptoms, whereas supportive and nurturing relationships have been shown to confer protective effects. Early intervention of maladaptive family relationships is thus crucial for preventing or offsetting negative developmental trajectories in at-risk children. A variety of therapeutic methods have been developed and employed to foster positive parent- child relationships and improve child mental health, including parent training/education, in-person therapy, home visiting, school curriculums, and web programs. However, systematic obstacles interfere with the accessibility, generalizability, and acceptability of these traditional appointment- and module-based approaches. Furthermore, limitations in the family-centered flexibility, individual responsiveness, and broad availability of these services render them inadequate to address the unique needs of at-risk populations who would benefit from more readily accessible and inexpensive 24-hour support that is provided in real time and real life—when and where support is needed most. Not surprisingly, research finds that roughly half of the families who do participate in traditional appointment- and module-based mental health services fail to show sufficient symptom improvement. Just-in- time adaptive interventions (JITAIs), in contrast, utilize smartphones, wearables, and artificial intelligence (AI) to identify and respond to psychological and behavioral processes and contextual events as they unfold in everyday life. Although JITAIs have the potential to transform the way people receive mental health support, barriers to their successful, wide-scale implementation remain. Using pilot data collected from smartphones and wearables, our interdisciplinary team of psychologists and engineers used AI to build machine learning algorithms to detect psychological states and contextual events, such as ongoing moods and relationship conflict, in couples. In the current project, we propose developing and testing a JITAI to provide opportune supports to families in dynamic response to contextual events and shifting psychological states to amplify attachment bonds, regulate emotion, and intervene in maladaptive parent-child interactional patterns. Building on our prior research, we will (1) build software to unobtrusively capture real-time data from commercially-available mobile devices, (2) use machine learning to develop algorithms to automatically monitor psychological and behavioral processes relevant to child mental health, (3) launch a JITAI to provide as-needed intervention, and (4) carry out a micro-randomized clinical trial to test the efficacy, acceptability, and safety of our JITAI for decreasing child internalizing and externalizing symptoms. Our project will contribute to the development of technology ecosystems and service delivery models with the power to meaningfully transform the accessibility and dynamic responsiveness of mental health care.