Emotion regulation and insomnia as candidate processes of change in a digital mental health intervention.

BACKGROUND: Digital mental health interventions can improve distress and well-being, but the processes statistically accounting for these effects remain unclear. We examined emotion regulation and insomnia as candidate processes in university students. METHODS: Data came from a three-arm randomized trial of the ETUCARE program, an eight-week multicomponent program for French university students. The ITT dataset contained 1064 participants (online: n = 460; hybrid: n = 235; control: n = 369); 947 contributed to FIML estimation. Parallel mediation estimated indirect effects through emotion regulation and insomnia. A ≥70%-adherence sample provided a per-protocol sensitivity analysis, and an exploratory cross-lagged model examined longitudinal associations. RESULTS: ETUCARE assignment was associated with improved emotion regulation (β = .166, p < .001) and reduced insomnia (β = -.142, p < .001). Both showed significant indirect effects on distress and well-being: better emotion regulation and lower insomnia were associated with lower distress and higher well-being. Combined pathways accounted for 33.6% of the total association with distress. A constrained model was compatible with similar absolute magnitudes but opposite directions across outcomes, without establishing equivalence. ITT and per-protocol estimates were closely aligned. CONCLUSIONS: These findings are consistent with a process-based account of digital interventions in which emotion regulation and sleep represent candidate shared processes statistically associated with changes across multiple domains of mental health. The findings do not establish causal mechanisms or temporal precedence. Nevertheless, targeting emotion regulation and sleep may represent a promising strategy for scalable digital interventions. Future experimental research should clarify their temporal dynamics and identify the most active intervention components.