Effect of a small increase in off-premises trading hours on alcohol sales in Norway: A stepped-wedge cluster-randomized controlled trial.

Abstract

Evidence of the effect of limiting off-premises alcohol trading hours is still scarce. This study tested the effect of a small extension in trading hours on alcohol sales in alcohol monopoly outlets in Norway. The extension of trading hours was implemented within a stepped-wedge cluster-randomized trial design. Eligible state monopoly outlets (n = 229) were clustered into trade districts (n = 62), which were block-randomized to one of three sequences regarding date of implementation: 1 September 2020 (n = 21 districts, 82 outlets), 1 December 2020 (n = 21 districts, 73 outlets) and 1 March 2021 (n = 20 districts, 74 outlets). Outcomes were followed-up for a 1-year period. Study participants were state monopoly outlets in urban and rural trade districts in all parts of Norway. Monthly alcohol sales in litres of pure alcohol per trade district and per outlet were measured from March 2020 to March 2022 (primary outcome). We applied a linear mixed-effect model with two-way fixed effects within a difference-in-difference framework. As a robustness check we considered the effects of cross-border trade and effects in subgroups of outlets. Trading hours in monopoly outlets were extended by 1 hour on Saturdays. The extension was permanent. Pre-intervention periods and not-yet-treated units served as control conditions. We did not find a statistically significant effect of the small extension in trading hours on monthly alcohol sales (i) per trade district [average treatment effect: -185.5litres, 95% confidence interval (CI) = -1159.9, 788.9] and (ii) per outlet (-35.3litres, 95% CI = -142.1, 72.0). These findings were consistent across estimation methods and model specifications. There is no clear evidence that a small extension in off-premises trading hours affected alcohol sales in monopoly outlets in Norway.