Mediators and mechanisms of change in digital health interventions for physical activity behaviour: a systematic review.

Abstract

Digital health interventions are used to promote physical activity (PA), but the mechanisms through which they work remain unclear. This systematic review synthesised evidence on mediators and mechanisms of change in digital PA interventions for adults, using the Theoretical Domains Framework (TDF) to organise mechanisms. Six databases were searched from inception to 5 March 2025 and updated on 25 February 2026. Eligible studies were experimental or quasi-experimental trials in adults that evaluated a digital intervention, included a PA outcome, and reported at least one mediation analysis for the intervention-mediator, mediator-PA, total intervention-PA, or indirect pathway. Twenty-four articles representing 23 trials met the inclusion criteria, and the updated search identified no new eligible studies. Sample sizes ranged from 51 to 1,166 (median = 242). Digital interventions produced a small total effect on omnibus PA (Path C: median r = .21, IQR = .19-.28). Across TDF domains, Path A effects were generally small to moderate, with behavioural regulation showing the clearest significance pattern. Path B associations were small and inconsistent, and Path AB indirect effects were uniformly small (median r ≈ .01-.06), with no TDF domain showing consistent mediation. Digital interventions may increase PA modestly, but stronger mediation research is needed.