Review question
In this review we ask the question whether banning or restricting the advertising of alcohol in any form will lead to people drinking less alcohol. The form of the ban could include banning alcohol advertisements on television, the internet or billboards, or in magazines. We were also interested in the harms that banning advertisements may cause, such as reducing profits in the alcohol and advertising industries, and whether governments would lose taxes if alcohol purchases went down after a ban.
Background
The misuse of alcohol is a significant risk factor for ill health, injury (e.g. through violent behaviour or road traffic collisions), death and social problems around the world. Advertising to promote the drinking of alcohol is widespread. Banning or restricting the advertising of alcohol has been suggested as a possible way to lower the use of alcohol in the general public and to stop young people from starting drinking at an early age.
Study characteristics
The evidence we present is current to May 2014. We found four studies that evaluated the restriction or banning of alcohol advertising via any format. One was a small randomised controlled trial (RCT) that evaluated drinking behaviour in 80 young men in the Netherlands exposed to movies with either a high or low alcohol content combined with a commercial with either a neutral content (interpreted as a ban on alcohol advertising) or a high alcohol content. The other three studies were interrupted time series (ITS) studies. ITS studies are studies in which changes, usually in the general public, are measured at various points before, during and after an intervention such as a change in policy. Two of the three ITS studies evaluated what happened after an advertising ban was introduced by two different Canadian provincial governments. The third ITS study evaluated what happened after a ban was lifted after being in place for 50 years in another Canadian province. Each study evaluated a different category of ban (either partial or full).
None of the above studies were funded by the alcohol or advertising industries.
Key results
The data arising from the included studies did not show a clear effect either for or against the banning or restriction of alcohol advertising.
In the RCT, young men who watched movies with a low-alcohol content drank less than men who watched movies with a high-alcohol content. Young men exposed to commercials with a neutral content compared with those exposed to commercials for alcohol drank less. The trial was one and a half hours, so we do not know how long beyond the trial these effects lasted. The trial did not report on any harmful outcomes.
The results from the three ITS studies were inconsistent. We statistically combined the results of the two studies that assessed what happened after a ban was introduced. This showed an overall increase in beer consumption in the general population following the introduction of the ban, but the results were uncertain and could also be consistent with no difference or an overall decrease in alcohol consumption. The third ITS study, which evaluated the lifting of a total ban on all forms of alcohol advertising to a ban on spirits advertising only, also found uncertain results. None of the studies reported on any harms arising from the bans.
Quality of the evidence
Overall we judged the quality of evidence to be very low in the RCT. This was based on the fact that there were problems with the study methodology, the population included men only and the results were not very accurate. In the ITS studies, the quality was also judged to be very low due to problems with the study methodology and the results not being precise.
Conclusions
The review cannot recommend for or against banning alcohol advertising. Governments that are considering implementing alcohol advertising bans would be advised to implement the ban in a research environment and monitor the effects over time to build the evidence base.
There is a lack of robust evidence for or against recommending the implementation of alcohol advertising restrictions. Advertising restrictions should be implemented within a high-quality, well-monitored research programme to ensure the evaluation over time of all relevant outcomes in order to build the evidence base.
Alcohol is estimated to be the fifth leading risk factor for global disability-adjusted life years. Restricting or banning alcohol advertising may reduce exposure to the risk posed by alcohol at the individual and general population level. To date, no systematic review has evaluated the effectiveness, possible harms and cost-effectiveness of this intervention.
To evaluate the benefits, harms and costs of restricting or banning the advertising of alcohol, via any format, compared with no restrictions or counter-advertising, on alcohol consumption in adults and adolescents.
We searched the Cochrane Drugs and Alcohol Group Specialised Register (May 2014); CENTRAL (Issue 5, 2014); MEDLINE (1966 to 28 May 2014); EMBASE (1974 to 28 May 2014); PsychINFO (June 2013); and five alcohol and marketing databases in October 2013. We also searched seven conference databases and www.clinicaltrials.gov and http://apps.who.int/trialsearch/ in October 2013. We checked the reference lists of all studies identified and those of relevant systematic reviews or guidelines, and contacted researchers, policymakers and other experts in the field for published or unpublished data, regardless of language.
We included randomised controlled trials (RCTs), controlled clinical trials, prospective and retrospective cohort studies, controlled before-and-after studies and interrupted time series (ITS) studies that evaluated the restriction or banning of alcohol advertising via any format including advertising in the press, on the television, radio, or internet, via billboards, social media or product placement in films. The data could be at the individual (adults or adolescent) or population level.
We used the standard methodological procedures expected by The Cochrane Collaboration.
We included one small RCT (80 male student participants conducted in the Netherlands and published in 2009) and three ITS studies (general population studies in Canadian provinces conducted in the 1970s and 80s).
The RCT found that young men exposed to movies with a low-alcohol content drank less than men exposed to movies with a high-alcohol content (mean difference (MD) -0.65 drinks; 95% CI -1.2, -0.07; p value = 0.03, very-low-quality evidence). Young men exposed to commercials with a neutral content compared with those exposed to commercials for alcohol drank less (MD -0.73 drinks; 95% CI -1.30, -0.16; p value = 0.01, very-low-quality evidence). Outcomes were assessed immediately after the end of the intervention (lasting 1.5 hours), so no follow-up data were available. Using the Grading of Recommendations Assessment, Development and Evaluation approach, the quality of the evidence was rated as very low due to a serious risk of bias, serious indirectness of the included population and serious level of imprecision.
Two of the ITS studies evaluated the implementation of an advertising ban and one study evaluated the lifting of such a ban. Each of the three ITS studies evaluated a different type of ban (partial or full) compared with different degrees of restrictions or no restrictions during the control period. The results from the three ITS studies were inconsistent. A meta-analysis of the two studies that evaluated the implementation of a ban showed an overall mean non-significant increase in beer consumption in the general population of 1.10% following the ban (95% CI -5.26, 7.47; p value = 0.43; I2 = 83%, very-low-quality evidence). This finding is consistent with an increase, no difference, or a decrease in alcohol consumption. In the study evaluating the lifting of a total ban on all forms of alcohol advertising to a partial ban on spirits advertising only, which utilised an Abrupt Auto-regressive Integrated Moving Average model, the volume of all forms of alcohol sales decreased by 11.11 kilolitres (95% CI -27.56, 5.34; p value = 0.19) per month after the ban was lifted. In this model, beer and wine sales increased per month by 14.89 kilolitres (95% CI 0.39, 29.39; p value = 0.04) and 1.15 kilolitres (95% CI -0.91, 3.21; p value = 0.27), respectively, and spirits sales decreased statistically significantly by 22.49 kilolitres (95% CI -36.83, -8.15; p value = 0.002). Using the GRADE approach, the evidence from the ITS studies was rated as very low due to a high risk of bias arising from a lack of randomisation and imprecision in the results.
No other prespecified outcomes (including economic loss or hardship due to decreased alcohol sales) were addressed in the included studies and no adverse effects were reported in any of the studies. None of the studies were funded by the alcohol or advertising industries.