Effects on alcohol consumption of announcing revised UK low-risk drinking guidelines: findings from a monthly cross-sectional survey

BackgroundHealth authorities publish alcohol consumption guidelines for low-risk drinking in most high-income countries but the effects of these guidelines on alcohol consumption are unclear. In January, 2016, the UK's Chief Medical Officers announced revised guidelines recommending that men and women should not regularly drink more than 14 units per week, a reduction to the previous guideline for men of 3–4 units per day. We aimed to evaluate the effect of announcing the revised guidelines on alcohol consumption.MethodsData were collected from March, 2014, to October, 2017, using the Alcohol Toolkit Survey, a monthly repeat cross-sectional survey of approximately 1800 adults (older than 16 years) resident in England. The survey uses a hybrid between random location sampling and quota sampling designed to generate a nationally representative sample, which selects random areas in England (about 300 households) from strata defined by area-level geographical and sociodemographic profiles. Participants provided verbal informed consent. The University College London ethics committee granted ethical approval for the Alcohol Toolkit Study and The University of Sheffield for the evaluation of the UK lower-risk drinking guidelines. The primary outcome is participants’ Alcohol Use Disorders Identification Test for Consumption (AUDIT-C) scores. Effects were estimated using segmented regression. Secondary analyses test for alternative breakpoints in the long-term trend and pulse effects. All analyses were preregistered in the ISRCTN registry, number ISRCTN15189062.FindingsAt baseline, 70·4% of the sample were drinkers and the mean AUDIT-C score was 2·8. The main analysis showed no significant step-change in AUDIT-C scores immediately following announcement of the guideline (β=0·001 [95% CI –0·079 to 0·099]; p=0·82) and the trend changed significantly such that scores increased by 0·005 each month (β=0·008 [0·001–0·015]; p=0·015). This finding was not robust as secondary analyses of alternative breakpoints suggested the change in behaviour began in June, before the new guidelines were announced. Secondary analyses also suggest that AUDIT-C scores reduced temporarily for 3 months (a pulse effect) after the announcement (β=–0·126 [–0·218 to –0·034]; p=0·007).InterpretationAnnouncing new UK drinking guidelines with no large-scale organised promotion did not lead to a substantial or sustained reduction in drinking or a downturn in long-term alcohol consumption behaviour. Well designed promotional campaigns might improve the effect of drinking guidelines on alcohol consumption. This study is limited by potential seasonal confounding—January is typically a light-drinking month, whereas December is a heavy-drinking month. We control for seasonal trends but this approach might be inadequate if seasonality varies substantially between years as our time series is relatively short.FundingNational Institute for Health Research Public Health Research Programme, School for Public Health Research, Cancer Research UK.