Evaluation of the introduction of a healthy food and drink policy in 13 community recreation centres on the healthiness and nutrient content of customer purchases and business outcomes: An observational study.

Introduction: This observational study assessed the introduction of a comprehensive healthy food and drink policy across 13 community organisation managed aquatic and recreation centres in Victoria, Australia, and the associated changes on business outcomes, and the healthiness of purchases. The policy, based on state government guidelines, mandated that food and drink availability be based on healthiness classification: 'red' (limit) <10%, and 'green' (best choice) >50%, and the remainder 'amber' (choose carefully).

Methods: Six years of monthly sales data were split into three periods, prior to (1/01/2013-31/12/2014), during (1/01/2015-31/12/2016) and post (1/1/2017-31/12/2018), policy implementation. Using point-of-sale data, food and drink nutrient content, and state guidelines, items were classified as 'red'/'amber'/'green'. Linear models with Newey West standard errors were fitted to compare the mean value of outcomes between post- to pre-policy implementation periods, for each outcome and centre; and were pooled using random effect meta-analyses.

Results: Comparing post- to pre-policy implementation periods, total food sales did not change (mean percentage difference: -3.2% (95% confidence interval (CI) -21% to 14%), though total drink sales declined -27% (CI -37% to -17%). The mean percentage of 'red' foods sold declined by -15% (CI -22% to -7.7%), 'amber' food sales increased 11% (CI 5.5% to 16%). 'Green' food sales did not change (3.3%, CI -1.4% to 8.0%). The mean percentage of 'red' drinks sold declined -37% (CI -43% to -31%), 'amber' and 'green' drink sales increased by 8.8% (CI 3.6% to 14%) and 28% (CI 23% to 33%), respectively. The energy density and sugar content (percentage of total weight/volume) of both food and drinks decreased.

Conclusions: This study has shown that the implementation of a policy to improve the health of retail food environments can result in a shift towards healthier purchases. Sales revenue from foods did not decline, though revenue from drinks did, indicating future research needs to explore mitigation of this.