Effect of snack-food proximity on intake in general population samples with higher and lower cognitive resource

Jennifer A. Hunter, Gareth J. Hollands, Dominique-Laurent Couturier, Theresa M. Marteau

Placing snack-food further away from people consistently decreases its consumption ("proximity effect"). However, given diet-related health inequalities, it is important to know whether interventions that alter food proximity have potential to change behaviour regardless of cognitive resource (capacity for self-control). This is often lower in those in lower socio-economic positions, who also tend to have less healthy diet-related behaviours. Study 1 aims to replicate the proximity effect in a general population sample and estimate whether trait-level cognitive resource moderates the effect. In a stronger test, Study 2 investigates whether the effect is similar regardless of manipulated state-level cognitive resource. Participants were recruited into two laboratory studies (Study 1: n = 159; Study 2: n = 246). A bowl of an unhealthy snack was positioned near (20 cm) or far (70 cm) from the participant, as randomised. In Study 2, participants were further randomised to a cognitive load intervention. The pre-specified primary outcome was the proportion of participants taking any of the snack. Significantly fewer participants took the snack when far compared with near in Study 2 (57.7% vs 70.7%, β = -1.63, p = 0.020), but not in Study 1 (53.8% vs 63.3%, X2 = 1.12, p = 0.289). Removing participants who moved the bowl (i.e. who did not adhere to protocol), increased the effect-sizes: Study 1: 39.3% vs 63.9%, X2 = 6.43, p = 0.011; Study 2: 56.0% vs 73.9%, β = -2.46, p = 0.003. Effects were not moderated by cognitive resource. These studies provide the most robust evidence to date that placing food further away reduces likelihood of consumption in general population samples, an effect unlikely to be moderated by cognitive resource. This indicates potential for interventions altering food proximity to contribute to addressing health inequalities, but requires testing in real-world settings. Both studies were registered with ISRCTN (Study 1 reference no.: ISRCTN46995850, Study 2 reference no.: ISRCTN14239872).