Removing seasonal confectionery from prominent store locations and purchasing behaviour within a major UK supermarket: Evaluation of a nonrandomised controlled intervention study

Background
The proportion of energy from free sugars and saturated fat currently exceeds the UK-recommended intake across all age groups. Recognising the limits of reformulation programmes, the government in England has announced their intention to introduce legislation to restrict the promotion of foods high in free sugars, salt, and saturated fats in prominent store locations. Here, we evaluated a grocery store intervention to remove seasonal confectionery from prominent locations within a major UK supermarket.

Methods and findings
A nonrandomised controlled intervention study with interrupted time series (ITS) analysis was used. Data were analysed from 34 intervention stores located in 2 London boroughs and 151 matched control stores located elsewhere in the UK owned by the same retailer. Stores were matched based on store size and overall sales during the previous year. Between 15 February 2019 and 3 April 2019 (before Easter), stores removed free-standing promotional display units of seasonal confectionery from prominent areas, although these products were available for purchase elsewhere in the store.

Store-level weekly sales (units, weight (g), and value (£)) of seasonal chocolate confectionery products were used in primary analyses, with data from 1 January 2018 to 24 November 2019. Secondary outcomes included total energy, fat, saturated fat, and sugars from all in-store purchases. Multivariable hierarchical models were used to investigate pre/post differences in weekly sales of confectionery in intervention versus control stores. ITS analyses were used to evaluate differences in level and trends after intervention implementation.

Over a preintervention baseline period (15 February 2018 to 3 April 2018), there were no significant differences in sales (units, weight, and value) of all chocolate confectionery between intervention versus control stores. After intervention implementation, there was an attenuation in the seasonal increase of confectionery sales (units) in intervention stores compared to control (+5% versus +18%; P < 0.001), with similar effects on weight (g) (+12% versus +31%; P < 0.001) and value (£) (−3% versus +10%; P < 0.001). ITS analyses generally showed statistically significant differences in the level at the point of intervention (P ranges 0.010 to 0.067) but also in the trend afterwards (P ranges 0.024 to 0.053), indicating that the initial difference between intervention and control stores reduced over time. There was a significant difference in level change in total energy sold, adjusted for the total weight of food and drink (kcal/g, P = 0.002), and total fat (fat/g) (P = 0.023), but no significant changes in saturated fat or sugars from total sales in ITS models. There was no evidence that the main results varied across store deprivation index. The limitations of this study include the lack of randomisation, residual confounding from unmeasured variables, absolute differences in trends and sales between intervention versus control stores, and no independent measures of intervention fidelity.

Conclusions
Removal of chocolate confectionery from prominent locations was associated with reduced purchases of these products, of sufficient magnitude to observe a reduction in the energy content of total food purchases. These results from a “real-world” intervention provide promising evidence that the proposed legislation in England to restrict promotions of less healthy items in prominent locations may help reduce overconsumption.

Trial registration
https://osf.io/br96f/.