EVIDENCE BASED PUBLIC HEALTH POLICY AND PRACTICE Large scale food retailing as an intervention for diet and health: quasi-experimental evaluation of a natural experiment

Objectives: To assess the effect on fruit and vegetable consumption, self reported, and psychological health of a ‘‘natural experiment’’—the introduction of large scale food retailing in a deprived Scottish community. Design: Prospective quasi-experimental design comparing baseline and follow up data in an ‘‘intervention’’ community with a matched ‘‘comparison’’ community in Glasgow, UK. Participants: 412 men and women aged 16 or over for whom follow up data on fruit and vegetable consumption and GHQ-12 were available. Main outcome measures: Fruit and vegetable consumption in portions per day, poor self reported health, and poor psychological health (GHQ-12). Main results: Adjusting for age, sex, educational attainment, and employment status there was no population impact on daily fruit and vegetable consumption, self reported, and psychological health. There was some evidence for a net reduction in the prevalence of poor psychological health for residents who directly engaged with the intervention. Conclusions: Government policy has advocated using large scale food retailing as a social intervention to improve diet and health in poor communities. In contrast with a previous uncontrolled study this study did not find evidence for a net intervention effect on fruit and vegetable consumption, although there was evidence for an improvement in psychological health for those who directly engaged with the intervention. Although definitive conclusions about the effect of large scale retailing on diet and health in deprived communities cannot be drawn from non-randomised controlled study designs, evaluations of the impacts of natural experiments may offer the best opportunity to generate evidence about the health impacts of retail interventions in poor communities.