A randomized controlled trial to assess the potential efficacy, feasibility and acceptability of an m-health intervention targeting parents of school aged children to improve the nutritional quality of foods packed in the lunchbox ‘SWAP IT’

Twelve Catholic primary schools in the Hunter New England region, NSW, Australia participated in a 2x2 factorial cluster randomized controlled trial. Participants included children aged 5–12 years (Kindergarten to Year 6) and their parents. The paper reports on the lunchbox component, with data collected at baseline and immediately post-intervention (10 weeks). Schools were randomly allocated to four groups: no intervention, physical activity intervention only, lunchbox intervention only, or both; the design allowed separate assessment of the lunchbox intervention.

Results: Among 1,915 lunchbox observations, mean energy in lunchboxes did not differ significantly between lunchbox intervention and controls at follow-up (−118.39 kJ; 95% CI −307.08 to 70.30; p=0.22). Mean lunchbox energy from recommended foods increased by 79.21 kJ (95% CI 1.99 to 156.43; p=0.04). Percentage energy from recommended foods rose by 4.57 percentage points (95% CI −0.52 to 9.66; p=0.08). App engagement included 71% of parents aware of the intervention, 55% reporting healthier swaps, and 84% finding pushed content helpful; pushed messages ranged from 387 to 1,550 views per week (mean 1,025). Conclusions: The SWAP IT lunchbox intervention delivered via a school mobile app was feasible, acceptable, and showed potential to improve the nutritional quality of foods packed in lunchboxes. These findings support a larger randomized trial to confirm longer-term effects and explore public-health impact if scaled.

Pilot study not powered for clinical outcomes; short duration with a single post-intervention follow-up (10 weeks); not powered to detect interaction effects; imbalance in baseline socio-economic/remoteness between arms; variable app engagement; outcome assessment relied on lunchbox observations and self-reported acceptability; generalizability to other settings is unknown.