Assessing the Temporal Dynamics in Self-Regulated Learning Constructs as Predictors of Engagement in a Digital Health Intervention

Datos Bibliográficos

ID9614058
AutoresClaudia Liu (0000-0003-1359-2394, Deakin University, autor de correspondencia), Matthew Fuller-Tyszkiewicz (0000-0003-1145-6057, Deakin University), Jake Linardon (0000-0003-4475-7139, Deakin University), Hannah K Jarman (0000-0001-8225-4511, Deakin University), Mariel Messer (0000-0002-7186-7264, Deakin University)
Año2025
Volumen57
Número2
Páginas205-219
Fecha de publicación2025-09-01
Peer ReviewedSí
Open AccessSí
TipoARTICLE
RevistaBehavior Therapy (JOURNAL)
Identificadores de la revistaISSN: 0005-7894 • E-ISSN: 1878-1888
EditorialElsevier BV (PUBLISHER)
DOI10.1016/j.beth.2025.09.008
PMID41741095
OpenAlexW7083430721
IdiomaEN
Referencias citadas39

The promise of self-guided digital health interventions (DHIs) is consistently hampered by their poor engagement. While past research has identified various predictors of engagement, effect sizes have been small, and many have failed to replicate across studies. This poor predictive performance may, in part, be due to the field's overreliance on predictor assessments at pre-intervention (i.e., at baseline), which overlooks the possibility that those factors driving engagement can fluctuate across the user journey. Self-regulated learning (SRL) is a potentially relevant process underlying engagement with DHIs, as majority deliver their therapeutic content as self-guided psychoeducation materials. The present study therefore examined whether key SRL variables fluctuated across a 6-week randomized controlled trial of a dialectical behavior therapy DHI for adults with recurrent binge eating (n = 113), and also whether these changes influenced their predictive ability of three engagement outcomes: number of (1) modules and (2) weekly assessments completed, and (3) skill activities unlocked. Longitudinal analyses revealed considerable temporal variation in SRL variables over time, and that majority of this variance was attributed to within-person differences (intraclass correlations <.40). Correlation coefficients between SRL variables and engagement outcomes increased in strength with time, indicating that later measurements of SRL had greater predictive ability compared to their measurements at baseline. Broadly, our findings highlight the need for researchers to more carefully plan future DHI trials investigating engagement by factoring the temporal dimension of engagement and its predictors into study design

Correlation · Experience sampling method · Explained variation · Intervention (counseling) · Psychoeducation · Psychological intervention · Replicate · Variance (accounting) · Electrical and Electromagnetic Research · Geochemistry and Geologic Mapping · Geological Modeling and Analysis