(PS10-A17) Compliance in Ecological Momentary Assessment Studies for Youth with Psychological Symptoms: A Meta-analysis

Background: Ecological momentary assessment (EMA) is a clinical research tool allowing for the collection of in-vivo, real-time data and facilitating the study of youth psychopathology (Shiffman et al., 2008). While EMA has numerous methodological strengths relative to self- and parent-report measures and lab-based tasks (e.g., reduced retrospective recall biases, potential for implementation in underserved communities), repeated naturalistic assessments may increase participant burden and consequently reduce compliance rates (Hasselhorn et al., 2021). Meta-analyses have identified study design features (e.g., financial incentives) that improve EMA compliance (Wrzus & Neubauer, 2022). However, questions remain regarding how those design features impact compliance across diagnostic categories. In the current meta-analysis, we examine EMA compliance in youth with psychological disorders and symptoms and investigate which study designs are the most efficacious for whom. We will (1) describe EMA compliance rates, and (2) identify methodological components (e.g., survey number, study length) that maximize EMA compliance rates, in youth with different clinical conditions.

Method: The search strategy was developed in collaboration with a Biomedical Librarian and peer reviewed by a second librarian not otherwise associated with the project. Three electronic databases were used as information sources: PubMed, APA PsycInfo, and Web of Science. Title/abstract screening, full-text screening, quality assessment, and data extraction were conducted in Covidence. All data were double-coded by six coders and discrepancies were resolved via group consensus. The described search strategy was supplemented by additional strategies (e.g., scanning the references of included studies).

Results: We identified 891 potentially relevant articles and included 598 articles following title/abstract screening, with samples representing the full range of psychological disorders and symptoms (e.g., 18.0% substance use, 7.5% depression, 7.1% eating disorders, 5.5% anxiety, 4.4% suicidal ideation and self-injury, 4.0% schizophrenia and psychosis). Full-text screening and data extraction are ongoing. We will test associations between psychological disorders and symptoms (e.g., substance use, depression) and compliance (i.e., percent of prescribed surveys completed) via random-effect models in the R metafor package. Publication bias will be assessed (e.g., funnel plots). Conditional upon sufficient heterogeneity in effects (via Q tests), we will examine study design aspects as moderators. All statistical analyses will be completed in advance of the ABCT meeting.

Conclusion: With information on the intersection of EMA study design and sample characteristics, clinical researchers may be able to optimize in-vivo tools with youth psychiatric samples. Maximizing data collection while minimizing participant burden is particularly relevant for research in underserved communities. Taken together, findings will have implications for the broad and equitable implementation of an innovative psychopathology assessment tool, including in the context of assessing symptom change in real time over the course of intervention.