Understanding Barriers to Diabetes Self-Management Using Momentary Assessment and Machine Learning (Preprint)

Persons with diabetes must perform many self-management tasks each day to obtain optimal control of their blood glucose. Psychosocial and contextual factors impact the ability to perform those tasks. Ecological momentary assessment (EMA) uses technology-mediated approaches to monitor and assess psychosocial and contextual variables that may impact self-management. To utilize EMA data in applied settings, however, feasible methods are needed to automate prioritization of the many factors that can impact health behaviors. This study uniquely applies machine learning algorithms to demographic and EMA-generated psychosocial data to predict self-management in adolescents with type 1 diabetes (T1D). The results suggest certain domains of factors more accurately predict on self-management than others and have promise for prioritization in future research. Results have implications for scaling up this combination of assessment and analytic approaches in population health.