The Role of Sex and Age on Pre-drinking: An Exploratory International Comparison of 27 Countries

Abstract

Aims

This exploratory study aims to model the impact of sex and age on the percentage of pre-drinking in 27 countries, presenting a single model of pre-drinking behaviour for all countries and then comparing the role of sex and age on pre-drinking behaviour between countries.

Methods

Using data from the Global Drug Survey, the percentages of pre-drinkers were estimated for 27 countries from 64,485 respondents. Bivariate and multivariate multilevel models were used to investigate and compare the percentage of pre-drinking by sex (male and female) and age (16–35 years) between countries.

Results

The estimated percentage of pre-drinkers per country ranged from 17.8% (Greece) to 85.6% (Ireland). The influence of sex and age on pre-drinking showed large variation between the 27 countries. With the exception of Canada and Denmark, higher percentages of males engaged in pre-drinking compared to females, at all ages. While we noted a decline in pre-drinking probability among respondents in all countries after 21 years of age, after the age of 30 this probability remained constant in some countries, or even increased in Brazil, Canada, England, Ireland, New Zealand and the United States.

Conclusions

Pre-drinking is a worldwide phenomenon, but varies substantially by sex and age between countries. These variations suggest that policy-makers would benefit from increased understanding of the particularities of pre-drinking in their own country to efficiently target harmful pre-drinking behaviours.

INTRODUCTION

Pre-drinking (also known as pre-loading, pre-partying or pre-gaming) is most commonly defined as the consumption of alcohol in domestic settings prior to attending licensed premises (Foster and Ferguson, 2014). Often motivated by the higher cost of alcohol in licensed venues, many people also choose to pre-drink to achieve rapid intoxication, or to facilitate socializing with friends (Foster and Ferguson, 2014; Miller et al., 2016; O’Neil et al., 2016; Labhart and Kuntsche, 2017). The practice has become an issue of increasing global concern due to evidence linking pre-drinking with higher levels of alcohol use and intoxication (Hughes et al., 2008; Reed et al., 2011; Labhart et al., 2013), and increased risk of adverse alcohol-related consequences such as blackouts, assault, injury or arrest (Pedersen and LaBrie, 2007; Hughes et al., 2008; Paves et al., 2012; Labhart et al., 2013; Miller et al., 2016).

The majority of studies examining pre-drinking are from the United States (US; Foster and Ferguson, 2014) but the practice has also been reported in literature from the United Kingdom (UK; Hughes et al., 2008; McClatchley et al., 2014), Canada (O’Neil et al., 2016), several European countries (Hughes et al., 2011; Labhart et al., 2013; Østergaard and Andrade, 2014), Brazil (Santos et al., 2015b), Argentina (Pilatti and Read, 2018), Australia (Miller et al., 2016) and New Zealand (Cameron et al., 2018). However, several inconsistencies between studies preclude clear understanding of pre-drinking trends at the national and international levels (Labhart et al., 2017). First, the definition and measurement of the practice varies between countries; in the US, pre-loading and pre-partying generally refer to the act of drinking before attending an event (such as a fraternity or sorority party or a sports game), and as such, US-based studies on pre-drinking typically utilize samples of college students describing retrospective alcohol use, many of whom are under the legal drinking age of 21 (Paschall and Saltz, 2007; Zamboanga et al., 2010). Alternatively, research from other countries, such as the UK and Australia, describes pre-drinking as the act of consuming alcohol at a domestic residence before attending licensed venues (such as nightclubs or bars), with studies in these countries often relying on street-intercept surveys with nightlife goers (Foster and Ferguson, 2014; Labhart et al., 2017). Second, the lack of consistency in the timeframes on which the prevalence figures are based, ranging from ‘on the same night’ up to ‘in the past year’, have prevented sound comparisons across studies and countries (Labhart et al., 2017). Consequently, simple trends such as the effect of sex and age on pre-drinking remain largely unknown in many countries and at the international level, resulting in uncertainty for policy-makers aiming to implement evidence-based strategies targeting risky pre-drinking behaviours among specific groups of the population.

For instance, while the majority of studies report that males are more likely to engage in pre-drinking and to drink more alcohol while pre-drinking than females (Paschall and Saltz, 2007; Hughes et al., 2008; LaBrie and Pedersen, 2008; Pedersen et al., 2009; Read et al., 2010; Wahl et al., 2013; Santos et al., 2015a; 2015b), there is also some contradictory evidence. A study of US college students by Read and colleagues (Read et al., 2010) found that, although males consumed more drinks than females while pre-drinking, higher blood alcohol concentrations were recorded among female participants, and a study among New Zealand undergraduate students found a larger effect of pre-drinking on blood alcohol concentration for women than for men (Cameron et al., 2018). A study comparing pre-drinking behaviours across four European countries found that females in the UK were significantly more likely to pre-drink than males (Hughes et al., 2011), while several other studies have reported no significant differences in pre-drinking between sexes (Pedersen and LaBrie, 2007; Reed et al., 2011).

With relation to age, the majority of studies report that pre-drinking is most common among young adults, especially those aged between 18 and 21 (Paschall et al., 2007; Read et al., 2010; Zamboanga et al., 2010; Reed et al., 2011). Several US-based studies have reported that underage drinkers (aged under 21 years) were significantly more likely to pre-drink than legal drinkers (Paschall et al., 2007; Read et al., 2010; Zamboanga et al., 2010); a logical finding considering that underage drinkers are not able to legally purchase alcohol in public settings. However, as with sex, several studies present conflicting findings, such as a study conducted among bar attendees in the US (mean age 24.7 years; Clapp et al., 2009) which reported that while younger participants had more heavy drinking episodes, older participants were more likely to pre-drink (Reed et al., 2011). It should however be noted that almost all studies in this area focus on young adults, restricting de facto potential investigations in other age groups.

Between-country comparisons in this area may allow for further insight into the impact of embedded drinking culture on pre-drinking patterns, particularly related to age and sex, as well as a comparison of effective policy responses targeting risky drinking behaviours. However, we are aware of only two studies that have compared pre-drinking behaviour between countries (Hughes et al., 2011; Labhart et al., 2017). The first is our previous study, which compared the percentage of pre-drinkers across 25 countries but did not include sex and age differences (Labhart et al., 2017). The second, conducted by Hughes and colleagues (2011), is the only study, to our knowledge, that has investigated the influence of sex on pre-drinking behaviour between countries: the Netherlands, Slovenia and Spain and the UK. The authors reported significant differences between countries in the percentage of female and male pre-drinkers on the night of the interview; while a higher percentage of Spanish males reported pre-drinking compared to Spanish females (60.5% vs. 59.0%), a greater percentage of females (compared to males) reported pre-drinking among respondents from the Netherlands (58.8% vs. 54.5%), Slovenia (38.2% vs. 32.6%), and especially the UK (70.3% vs. 51.0%). In addition, the authors noted significantly higher median grams of alcohol consumed and blood alcohol content among males than females in all four countries (with this latter difference significant in all countries except Spain; Hughes et al., 2011). With the majority of research in this area based in the US (Foster and Ferguson, 2014), these studies provide unique and valuable insight into pre-drinking trends in a number of other countries. We are unaware of any studies comparing age-related patterns of pre-drinking between countries.

In this exploratory study, using data from a large global survey, and extending upon previous research by the authors (Labhart et al., 2017), we aim to model the impact of sex and age on the percentage of pre-drinking in different countries, with the purpose of informing focused policy responses aiming to reduce alcohol-related harms.

METHODS

Sample

Data for this study is drawn from the 2015 Global Drug Survey (GDS), an annual, anonymous, online survey (Barratt et al., 2017) accomplished in collaboration with media partners from around the world (see acknowledgments). The GDS uses a cross-sectional design to recruit a non-probability sample of people who self-select to complete the online survey investigating past use of alcohol and other drugs (Barratt et al., 2017). The study received ethical approval from Kings College London (PNM/14/15-18), The University of Queensland (2017001452/11671/001), and The University of New South Wales (HC17769). All participants provided informed consent before commencing the survey.

From an initial sample of 101,311 individuals who completed the GDS between November 2014 and January 2015 in 174 different countries, we excluded 3,456 records (3.4%) due to data capture glitches, duplicate entries, reporting no alcohol or other psychoactive drug use, or reporting the use of a fake drug. To allow complete case analysis and sample comparisons across countries, we then removed all records for respondents who did not fulfil the following inclusion criteria: aged between 16 and 35 years (due to the small size of the sample aged ≥36 years when stratified by country, age and sex); indicated sex as male or female; consumed alcohol in the past 12 months; indicated they were a ‘nightlife goer’ by providing a valid answer to the question ‘when you go on a night out in your country how safe do you generally feel: ‘in bars and pubs’ or ‘in nightclubs’’ (Bellis et al., 2015); and answered the pre-drinking question (see Measures section).

Finally, to have sufficient data for sex comparisons, countries with a sample size of less than 90 people who indicated pre-drinking were also excluded, with the exception of Colombia which had a sample size of 89; these data was included in analyses as the authors were unable to find published literature on pre-drinking behaviour in Colombia. This minimum country-specific sample size was based on the overall percentage of pre-drinkers (62.6%; see below) and allowed a 10% absolute error with 5% type I error (Chow and Liu, 2013). In total, a further 33,370 records (32.9% of total) were removed prior to analysis as some respondents provided missing information on more than one inclusion criteria. Unlike our previous study of 25 countries (Labhart et al., 2017), we separated the UK into the three countries of England, Scotland, and Wales, to gain greater insight into pre-drinking practices within the separate countries. The final sample for analysis comprised 64,485 respondents from 27 countries (see Fig. 1).

Fig. 1.

Descriptive statistics for pre-drinking by country. Note: Data are presented, by country, in ascending order of percentages of pre-drinkers. Dot point and T-bar represent percentage of pre-drinkers; light grey bar represents percentage of male respondents; dark grey bar represents age (centre line for mean, range for standard deviation). The solid red line depicts overall percentage of pre-drinkers, the dotted red line represents the overall percentage of male respondents, and the dashed red line represents the average age of respondents. Country sample size is in parentheses.

Measures

A measure of pre-drinking was created from the question: ‘When you arrive at the first bar or nightclub you are visiting on a night out, would you usually: ‘be completely sober (i.e. have had no alcoholic drink or drugs)’, ‘have already consumed a bit of alcohol (but not be drunk)’ or ‘have already consumed a lot of alcohol (feeling drunk)’?’ (Labhart et al., 2017) All questions referred to the past 12 months. Individuals who responded ‘completely sober’ were coded no to pre-drinking (i.e. 0) and individuals who responded positively to consuming either a bit or a lot of alcohol before going out were coded yes to pre-drinking (i.e. 1).

Statistical analysis

Logit models and predictive margins were used to estimate the age- and sex-specific percentages of pre-drinking both for the whole sample (see model 1; allowing for random effects for each country) and for individual countries (see model 2). We calculated 95-percent confidence intervals (CI) using the delta-method. For model 1, to account for large variation in sample sizes within countries, and the possibility of country-specific influence on age and sex and pre-drinking behaviour, Stata’s mixed-effects models using QR decomposition (meqrlogit) was utilized (StataCorp, 2017).

$$log\left(\frac{{\Pi }_{ij}}{1-{\Pi }_{ij}}\right)={\beta }_0+{\beta }_a{X}_{aij}+u_j$$

(model 1)

j = 1,…,27 countries and i = 1,…,n_j respondents within country j; ${\beta }_a$⁠, coefficient of X_a; X_a, a combination of covariates sex (as a binary variable) and age (as a continuous variable).

$$log\left(\frac{{\Pi }_i}{1-{\Pi }_i}\right)={\beta }_0+{\beta }_a{X}_{ai}+{\epsilon }_i$$

(model 2)

By country: i = 1,…,n respondents; ${\beta }_a$⁠, coefficient of X_a; X_a, a combination of covariates sex (as a binary variable) and age (as a continuous variable).

The covariates X_a, shown in model 1 and 2, represent the full range of multivariate models, which includes an interaction term between sex and age, and accounts for non-linear relationships between age and pre-drinking behaviour; in this instance the quadratic and cubic form of age. These two forms of age are explored to best account for the typical ‘inverted U-shaped’ curve of age as a function of drinking patterns across the life-course (i.e. increasing up to a certain age, then decreasing; Kuntsche and Gmel, 2013). For Model 1, several alternative models were estimated, ranging in decreasing order of complexity, including the interaction between age (cubic form) and sex (see Fig. 3); the main effects of age (cubic form) and sex (see Supplementary Figure 5); the interaction between age (quadratic form) and sex (see Supplementary Figure 6); the main effects of age (quadratic form) and sex (see Supplementary Figure 7); the interaction between age (linear form) and sex (see Supplementary Figure 8); and the main effects for age (linear form) and sex (see Supplementary Figure 9). Using maximum likelihood ratio tests, the most robust model (see Fig. 3) was selected on the grounds of parsimony, and depicts the full sample’s pre-drinking behaviour with respects to age and sex. Since pre-drinking behaviours might vary across countries, the same procedure was also repeated for each country separately (see Fig. 4; all models are available upon request to the authors). All data preparation and analyses were undertaken using Stata V15 (StataCorp, 2017).

RESULTS

Figure 1 presents respondents’ descriptive statistics, across the 27 countries, for the percentage of pre-drinkers, the percentage of males, and the mean and standard deviation (SD) of respondents’ ages. The three horizontal lines in Fig. 1 reflect the overall percentage of the sample for percentage of pre-drinkers (62.6%; solid red-line), percentage of male respondents (58.9%; dotted red line) and average age of respondents (24.2 years, standard deviation (SD) 4.8; dashed red line).

Greece had the lowest percentage of respondents who indicated pre-drinking (17.8%), whilst Ireland had the highest percentage (85.6%). Fig. 1 also highlights a vast range in the percentage of males across the 27 countries, ranging from 31.5% of Canadian respondents to 79.0% of respondents from Greece. There was also variation in age trends; the country with the youngest mean age reporting pre-drinking was the Netherlands (21.96 years; SD 4.01), and the country with the oldest mean age was Portugal (26.3 years; SD 4.91).

Fig. 2 separates the information presented in Fig. 1 by sex. Between countries, with small variation, the percentages of pre-drinkers for males and females follow a similar pattern across the 27 countries, although, within countries, the percentage is typically lower for female pre-drinkers. A notable exception is Canada, where the percentage of females who reported pre-drinking (83.0%) is 8.5 percentage points greater than males (74.6%). The greatest difference in pre-drinking percentages between sexes is seen in Spain, where 80.0% of males report pre-drinking, compared to 60.1% of females.

Fig. 2.

Descriptive statistics for pre-drinking split by sex and by country. Note: Data is presented, by country, in ascending order of percentages of pre-drinkers as seen in Fig. 1. Dot points represent percentage of pre-drinkers; dark grey bar represent age (centre line for mean, range for standard deviation). By figure, the dotted red line represents the overall percentage of respondents and the dashed red line represents the average age of respondents. Country sample size is in parentheses.

Figure 3 shows the predicted probability of pre-drinking modelled against the interaction of sex by the cubic form of age after accounting for the random effects by country. The figure depicts an increase in the probability of pre-drinking for both sexes as age increases from 16 years. For males, the probability of reporting pre-drinking behaviour increases until 20 to 21 years of age and then decreases with age, although the rate of decrease is not constant across ages 21 and 29. For females, while the probability of reported pre-drinking similarly increases until 20 years of age and then decreases with age, the rate of this decrease is different, decreasing at a greater rate after the age of 31.

Fig. 3.

Predicted probability of pre-drinking by age (in years) separated by sex. Note: Age is fitted as a cubic term and includes an interaction term with sex. Country random effects are accounted for.

What is also notable in Fig. 3 is the difference (depicted as the gap between the two lines) in the probability of pre-drinking, between males and females across the ages. The estimated percentage change is greatest for the youngest respondents, and reduces among older respondents. That is, for each year increase in respondents’ age, the probability of indicating pre-drinking is greater for males than it is for females compared to the previous age group. However, this rate decreases with age, until respondents are approximately 31 years of age, at which time it swaps for males and females.

Figure 4 shows the predicted percentage of pre-drinking for each of the 27 countries, accounting for age and sex. In Colombia, Mexico and Poland, a quadratic term of age (‘inverted U-shaped’) was observed, whereas for Belgium, Denmark, Finland, Greece, Italy, Portugal, Spain and Sweden a simple main effect (i.e. a roughly linear decrease) for age was the best form. Across most of the remaining countries, the age pattern in cubic form was observed, where pre-drinking expectedly increased and then decreased with age, but then appeared to level out at approximately 30 years of age. In Brazil, Canada, England, Ireland, New Zealand, and the United States, the predicted probability of pre-drinking actually appears to increase after the age of 30.

Fig. 4.

By country: predicted percentage of pre-drinking accounting for age and sex characteristics.

A significant interaction between age and sex was observed only in Denmark and the Netherlands. In Denmark, the predicted probability of pre-drinking was greater for females compared to males for those aged between 16 and 25 years, but after 25 years of age this substantially reduces at a greater rate among Danish females as age increases when compared to males. For respondents from the Netherlands, much like other countries, the probability of pre-drinking for males remains higher than for females, however, across the ages males’ patterns of pre-drinking does not mirror their female counterparts; after 31 years of age, males’ predicted probability for pre-drinking rises with age, whereas for females it still continues to reduce.

DISCUSSION

Expanding on our previous study’s comparisons of pre-drinking practices between countries (Labhart et al., 2017), this exploratory study indicates that international pre-drinking practices are likely to vary further by sex and age. Although our findings are based on a self-selected sample completing the Global Drug Survey, this study’s large sample, representing 64,485 respondents from 27 countries, provides valuable insight into the possible influence of the countries’ drinking culture on potentially harmful pre-drinking practices.

First, consistent with previous studies (Paschall and Saltz, 2007; Hughes et al., 2008; LaBrie and Pedersen, 2008; Pedersen et al., 2009; Read et al., 2010; Wahl et al., 2013; Santos et al., 2015a, 2015b), in all countries we found a higher percentage of pre-drinking among males compared to females, with two exceptions; Canadian females reported pre-drinking at 8.5 percentage points higher (overall) than Canadian males at all ages, and Danish females reported a higher probability of pre-drinking, but only up until 25 years of age. While these findings may reflect country-specific trends in alcohol consumption among females in Canada and Denmark (Bulloch et al., 2016; Edkins et al., 2017), it is also possible that these findings extend those of several other recent international studies reporting increases in binge drinking among young females (Grucza et al., 2009; Davoren et al., 2016), including findings of a meta-analysis of alcohol consumption among university students in Ireland and the United Kingdom (Davoren et al., 2016). Several authors (Pedersen and LaBrie, 2007; LaBrie and Pedersen, 2008; Paves et al., 2012) suggest that females may use pre-drinking as a means of matching the intoxication levels of their male counterparts, or that the relatively brief nature of pre-party events combined with a co-ed context may provide opportunities for females to attempt to match the brisker drinking pace of males (Paves et al., 2012). Alternatively, females might also favour other aspects of pre-drinking such as socializing with friends before attending nightclubs (Labhart and Kuntsche, 2017), or may consume relatively small amounts to overcome shyness once out (Kuntsche and Labhart, 2013). Future research may wish to examine the influence of social context on pre-drinking behaviour so as to provide further insight into sex-related differences in pre-drinking practices (Paves et al., 2012).

In terms of age, our results show that across the sample pre-drinking increases in probability between age 16 and approximately 20 or 21 years, reflecting findings from international studies showing that pre-drinking is most common among young adults aged below 21 years (Paschall et al., 2007; Read et al., 2010; Zamboanga et al., 2010; Reed et al., 2011). However, although pre-drinking probability largely decreased among our sample after the age of 21, we also found that levels of pre-drinking remained largely constant from age 30 onwards, suggesting that pre-drinking practices are not limited to university-age consumers of alcohol. Furthermore, in Brazil, Canada, Ireland, New Zealand, England and the United States, the probability of pre-drinking actually appears to increase between the ages of 30 and 35. While some of this up-turn may be an artefact of fitting a cubic function of age, in these cases the best model suggests that the decreasing trend for pre-drinking observed in the young respondents is not observed among respondents aged over 30 years. While we did not record the volume of alcohol consumed by the respondents while pre-drinking nor their motivations to do so, we might expect that young drinkers are more likely to pre-drink to save money and get intoxicated in unsupervised environments, while some adults in their early thirties may enjoy the opportunity to socialize with friends in a quiet environment (Ritchie et al., 2009). These findings suggest that policy responses aiming to reduce pre-drinking practices may benefit from campaigns targeted at a wider age group, and not only those of university age.

These age-related findings may also be influenced in part by countries’ varying minimum legal purchase age (MLPA; see Supplementary Table 1), although this might be only marginal considering that the prevalence of pre-drinking among respondents from the United States—the only country with a MLPA of 21—was found to be 63% for men and 66% for women (65% overall; see Fig. 1). It should also be noted that the influence of MLPAs on age-related trends in pre-drinking is also challenging to model due to within-country variations in MLPA based on alcohol content (e.g. European countries such as Belgium, Denmark, Finland or Switzerland) and state/province (e.g. Austria and Canada).

Furthermore, examination of the tandem influence of sex and age on pre-drinking practices shows that the decrease in pre-drinking after age 30 may not be equal across sexes; the declining rate of pre-drinking is generally lower for females aged 31 to 35 years of age compared to males in the same age bracket. Our results add further impetus for the need for future research to investigate changing trends of alcohol consumption among females.

We also found markedly different patterns in age- and sex-related pre-drinking trends between countries. For example, while our previous study using this sample found the lowest percentage of pre-drinking among respondents from Greece (17.7%; Labhart et al., 2017) we now see that this low level of pre-drinking in Greece is likely maintained with age, and that pre-drinking practices appear to be nearly identical between sexes. A similar trend is observed in Portugal, except at a much higher level, where pre-drinking remains at approximately 60% across age groups. These findings may reflect the ‘Mediterranean’ pattern of alcohol use, where typically moderate amounts of alcohol are consumed daily with meals (Popova et al., 2007), as opposed to patterns of irregular episodes of binge drinking more common in Northern European countries (Popova et al., 2007; Kuntsche et al., 2015). An unusual age- and sex-related pattern of drinking is also noted in Denmark; females aged 16–25 were more likely to pre-drink than males in the same age group, but this relationship reverses after the age of 25, with females’ probability of pre-drinking also decreasing at a much greater rate. While it is difficult to ascribe reasons for this finding, a study among young Danish nightlife goers (mean age 23 years) found that pre-drinking was most common among younger females, who described that their primary motive for pre-drinking was to ‘get wasted/ out of control’ (Østergaard and Andrade, 2014). A further notable pattern of pre-drinking is observed among respondents from the Netherlands; although pre-drinking appears to primarily decrease with age, the predicted probability of pre-drinking increases among males after 31 years of age.

The observation of these country-specific trends in pre-drinking reminds that each country is unique, notably in terms of laws and policies (e.g. age limits for purchasing alcohol, on- and off-premise drinks prices and general drinking culture; Labhart et al., 2017). Although pre-drinking appears to be universally associated with increased levels of alcohol consumption and risks of negative consequences (Foster and Ferguson, 2014), our findings suggest that future research and policy aiming to reduce harmful pre-drinking practices may benefit from targeting both sexes, and a wider group of ages than only young adults. Furthermore, policymakers may benefit from local adaptation of interventions and policies from other countries with similar pre-drinking patterns in order to reduce the harms associated with pre-drinking such as Canada’s Safer Bars (Graham et al., 2003) or ‘Pre-thinking while you’re pre-drinking’ (Wells et al., 2009) programs.

Limitations

While our study’s large and diverse sample enables a unique comparison of age and sex trends in pre-drinking practices across 27 countries, some limitations to our findings should also be acknowledged. First, our analysis comprises an absolute measure of pre-drinking only, and does not account for the amount of alcohol that respondents consumed while pre-drinking. Second, the GDS comprises a non-probability sample that excludes individuals who do not use licit and/or illicit psychoactive drugs and who do not use the internet, and as such any indications of prevalence should be interpreted with caution (Barratt et al., 2017). However, our restriction of the analysis to those who attended a bar or nightclub in the previous year provides a more accurate estimate of pre-drinking as part of the practice of attending nightlife venues. Future research in this area would benefit from using representative country-level samples to confirm present findings, and/or from including measures of blood alcohol concentration and harmful consequences resulting from pre-drinking, so as to provide further insight into some of the possible reasons for these age- and sex-related differences in pre-drinking behaviours between countries.

CONCLUSION

This exploratory study adds to our understanding of age- and sex-related trends in pre-drinking within and across countries. While higher percentages of males engaged in pre-drinking practices compared to females, exceptions to this trend were found among respondents in Canada and Denmark. In several countries, although the probability of pre-drinking reduced after age 21, the practice often levelled out or even increased after the age of 30. Increased understanding of the considerable variations in age- and sex-related trends in pre-drinking trends between countries may benefit policy makers aiming to reduce harms related to pre-drinking.

CONFLICT OF INTEREST STATEMENT

Dr Adam Winstock is the founder and owner of the Global Drug Survey. Jason Ferris is the chief biostatistician of the Global Drug Survey. No other conflicts to declare.