Premium
Effect of extended trading hours on alcohol sales in Norway: study protocol for a stepped wedge cluster‐randomized trial
Author(s) -
Grøtting Maja Weemes,
Bergsvik Daniel,
Rossow Ingeborg
Publication year - 2022
Publication title -
addiction
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.424
H-Index - 193
eISSN - 1360-0443
pISSN - 0965-2140
DOI - 10.1111/add.15704
Subject(s) - monopoly , norwegian , randomized controlled trial , business , binge drinking , economics , medicine , poison control , suicide prevention , environmental health , microeconomics , surgery , linguistics , philosophy
Background and aims Norwegian alcohol policy measures include national restrictions on sales hours and a state monopoly on retail sales. A 1‐hour extension of sales hours on Saturdays in the monopoly outlets took effect from September 2020. We aim to evaluate whether increase in sales hours results in (1) an increase in alcohol sales in the monopoly outlets and (2) an increase in total alcohol sales, including substitution effects from beer sales in grocery stores. Design The extension of Saturday sales hours is implemented within a stepped wedge cluster‐randomized trial design. Block randomization of 62 of the 66 Norwegian trade districts allocated monopoly outlets to one of three sequences regarding date of implementation. Setting and participants A total of 228 of 335 in total Norwegian state monopoly outlets are eligible and included. Intervention The extension of sales hours is from 3 p.m. to 4 p.m. starting on the first Saturday in (i) September 2020, (ii) December 2020 or (iii) March 2021. Measurements Growth rates in monthly alcohol sales, measured in litres of pure alcohol, in eligible monopoly outlets (primary outcome) are obtained together with beverage‐specific sales and alcohol sales in grocery stores (secondary outcomes). The observation period is set to 72 months prior to and 24 months after implementation. Findings Power analyses indicate that this stepped wedge cluster‐randomized controlled trial has a power above 90%, even at a high significance level (α = 0.01) and with other conservative model specifications. The planned trial offers a rare opportunity to study possible causal effects of a relatively small change in a widely used alcohol policy measure.