File(s) under permanent embargo
Child and adult snack food intake in response to manipulated pre-packaged snack item quantity/variety and snack box size: a population-based randomized trial
journal contributionposted on 2019-01-01, 00:00 authored by J A Kerr, P W Jansen, F K Mensah, K Gibbons, T S Olds, J B Carlin, S A Clifford, D Burgner, Lisa GoldLisa Gold, L A Baur, M Wake
Objectives: Snacks contribute to overconsumption of energy-dense foods and thence obesity. Previous studies in this area are limited by self-reported data and small samples. In a large population-based cohort of parent–child dyads, we investigated how modification of pre-packaged snack food, i.e. (a) item quantity and variety, and (b) dishware (boxed container) size affected intake. Methods: Design: Randomized trial nested within the cross-sectional Child Health CheckPoint of the Longitudinal Study of Australian Children, clustered by day of visit. Sample: 1299 11–12 year olds, 1274 parents. Exposure: 2 × 2 manipulation of snack box container size and item quantity/variety: (1) small box, few items, (2) large box, few items, (3) small box, more items, (4) large box, more items. Procedure: Participants received a snack box during a 15 min break within their 3.5 h visit; any snacks remaining were weighed. Outcomes: Consumed quantity (grams) and energy intake (kilojoules). Analyses: Unadjusted linear regression. Results: Children who were offered a greater quantity and variety of snack items consumed considerably more energy and a slightly higher food mass (main effect for energy intake: 349 kJ, 95% CI 282–416, standardized mean difference (effect size) 0.66; main effect for mass: 10 g, 95% CI 3–17, effect size 0.17). In contrast, manipulating box size had little effect on child consumption, and neither box size nor quantity/variety of items consistently affected adults’ consumption. Conclusion: In children, reducing the number and variety of snack food items available may be a more fruitful intervention than focusing on container or dishware size. Effects observed among adults were small, although we could not exclude social desirability bias in adults aware of observation.