Introduction and Objective: Self-reports of diabetes self-management are often distorted by inaccurate retrospective reporting of behavior. This study evaluated whether daily ecological momentary assessment (EMA) ratings validly capture diabetes self-management and examined bidirectional within-person associations with CGM metrics among diverse adults with T1D. Methods: Adults with T1D (N=176; mean age 40±15 years; 52% women; 41% Latino, 29% White, 15% Black) wore blinded CGM for 14 days, rated self-management each evening on a smartphone (response scale: 0-100), and completed retrospective questionnaires at baseline (a 3-item insulin adherence questionnaire and the Diabetes Self-management Questionnaire DSMQ, covering the last 1 and 2 months, respectively). Multilevel models tested associations between mean EMA ratings and questionnaire data. Dynamic structural equation models evaluated same- and next-day within-person associations with CGM metrics. Results: Higher mean EMA self-management was associated with DSMQ scores (r=0.45, p.001) and insulin adherence (r=0.64, p.001). Within persons, better EMA self-management was associated with lower same-day mean glucose (r=−0.32, p.001), higher time in range (TIR; r=0.25, p.05), less time 250 mg/dL (r=−0.33, p.05), and greater next-day glucose variability (β=0.05, p.05; standardized estimate=0.10). Further, greater daytime TIR predicted better next-day self-management (β=7.98, p.01; standardized estimate=0.09). Older age and personal CGM use were associated with significantly stronger same-day associations between self-management and glycemic quality. Conclusion: Daily EMA self-management ratings demonstrate strong validity and capture bidirectional links with glycemic regulation. Better self-management is linked to same-day glycemic quality and next-day variability, while greater TIR is associated with better next-day ratings of self-management. EMA assessment of self-management enables close pairing with contextual factors to inform precision interventions. Disclosure C. Hoogendoorn: None. R. Hernandez: None. S. Schneider: None. E. Pyatak: None. J.S. Gonzalez: None. Funding R01 DK121298
HOOGENDOORN et al. (Fri,) studied this question.