The impact of promoting revised UK low-risk drinking guidelines on alcohol consumption: interrupted time series analysis

Abstract

The UK’s Chief Medical Officers revised the UK alcohol drinking guidelines in 2016 to ≤ 14 units per week (1 unit = 10 ml/8 g ethanol) for men and women. Previously, the guideline stated that men should not regularly consume more than 3–4 units per day and women should not regularly consume more than 2–3 units per day. To evaluate the impact of promoting revised UK drinking guidelines on alcohol consumption. Interrupted time series analysis of observational data. England, March 2014 to October 2017. A total of 74,388 adults aged ≥ 16 years living in private households in England. Promotion of revised UK low-risk drinking guidelines. Primary outcome – alcohol consumption measured by the Alcohol Use Disorders Identification Test – Consumption score. Secondary outcomes – average weekly consumption measured using graduated frequency, monthly alcohol consumption per capita adult (aged ≥ 16 years) derived from taxation data, monthly number of hospitalisations for alcohol poisoning (International Statistical Classification of Diseases and Related Health Problems, Tenth Revision: T51.0, T51.1 and T51.9) and assault (International Statistical Classification of Diseases and Related Health Problems, Tenth Revision: X85–Y09), and further measures of influences on behaviour change. The Alcohol Toolkit Study, a monthly cross-sectional survey and NHS Digital’s Hospital Episode Statistics. The revised drinking guidelines were not subject to large-scale promotion after the initial January 2016 announcement. An analysis of news reports found that mentions of the guidelines were mostly factual, and spiked during January 2016. In December 2015, the modelled average Alcohol Use Disorders Identification Test – Consumption score was 2.719 out of 12.000 and was decreasing by 0.003 each month. After the January 2016 announcement, Alcohol Use Disorders Identification Test – Consumption scores did not decrease significantly (β = 0.001, 95% confidence interval –0.079 to 0.099). However, the trend did change significantly such that scores subsequently increased by 0.005 each month (β = 0.008, 95% confidence interval 0.001 to 0.015). This change is equivalent to 0.5% of the population moving each month from drinking two or three times per week to drinking four or more times per week. Secondary analyses indicated that the change in trend began 6 months before the guideline announcement. The secondary outcome measures showed conflicting results, with no significant changes in consumption measures and no substantial changes in influences on behaviour change, but immediate reductions in hospitalisations of 7.3% for assaults and 15.4% for alcohol poisonings. The pre-intervention data collection period was only 2 months for influences on behaviour change and the graduated frequency measure. Our conclusions may be generalisable only to scenarios in which guidelines are announced but not promoted. The announcement of revised UK low-risk drinking guidelines was not associated with clearly detectable changes in drinking behaviour. Observed reductions in alcohol-related hospitalisations are unlikely to be attributable to the revised guidelines. Promotion of the guidelines may have been prevented by opposition to the revised guidelines from the government's alcohol industry partners or because reduction in alcohol consumption was not a government priority or because practical obstacles prevented independent public health organisations from promoting the guidelines. Additional barriers to the effectiveness of guidelines may include low public understanding and a need for guidelines to engage more with how drinkers respond to and use them in practice. Current Controlled Trials ISRCTN15189062. This project was funded by the National Institute for Health Research (NIHR) Public Health Research programme and will be published in full in Public Health Research; Vol. 8, No. 14. See the NIHR Journals Library website for further project information.