Abstract
INTRODUCTION AND AIMS: This paper examines the impact of an alcohol labelling intervention on recall of and support for standard drink (SD) labels, estimating the number of SDs in alcohol containers, and intended and unintended use of SD labels. DESIGN AND METHODS: A quasi-experimental study was conducted in Canada where labels with a cancer warning, national drinking guidelines and SD information were applied to alcohol containers in the single liquor store in the intervention site, while usual labelling continued in the two liquor stores in the comparison site. Three waves of surveys were conducted in both sites before and at two time-points after the intervention with 2049 cohort participants. Generalised estimating equations were applied to estimate changes in all outcomes. RESULTS: Participants in the intervention relative to the comparison site had greater odds of recalling [adjusted odds ratio (AOR) 5.69, 95% confidence interval (CI) 3.02, 10.71] and supporting SD labels (AOR 1.49, 95% CI 1.04, 2.12) and lower odds of reporting using SD labels to purchase high strength, low-cost alcohol (AOR 0.65, 95% CI 0.45, 0.93). Exposure to the labels had negligible effects on accurately estimating the number of SDs (AOR 1.06, 95% CI 0.59, 1.93) and using SD labels to drink within guidelines (AOR 1.04, 95% CI 0.75, 1.46). DISCUSSION AND CONCLUSIONS: Evidence-informed labels increased support for and decreased unintended use of SD labels. Such labels can improve accuracy in estimating the number of SDs in alcohol containers and adherence to drinking guidelines.