What is Binge Drinking: Definition, Facts, Statistics and Effects

“Binge drinking” is a very debated term. It can historically be defined in two main different ways, with the discrepancy mainly regarding the length of the episode in question, but also according to the actual quantity of alcohol consumed. The first definition, and the older one, of the term refers to an extended episode of heavy drinking to the point of experiencing a hangover, usually implying more than one single day (WHO, 1994).

The other, and more common use nowadays, is to explain a situation of heavy drinking in a much shorter period of time, mainly several hours. There were some attempts to try to define the concept more objectively, for example by identifying this behaviour with the drinking of 5 or more drinks successively (Cahalan & Room, 1974). Another slightly different definition of this term is consuming half of the suggested weekly quantity of alcohol in one single session (Bennet et al.’s, 1990), which can be quantified as at least 10 units for men and 7 for women.

According to the National Institute on Alcohol Abuse and Alcoholism, binge drinking is interpreted as an event leading to a Blood Alcohol Concentration (BAC) count of 0.08% (0.8g/L) or higher, which happens after 5 drinks for men and 4 for women in two hours-time (NIAAA). The problem with this technical definition and the reason why it has been largely criticized, is due to the fact that there is also some subjectivity to be taken into account, such as a person’s alcohol capacity, a person’s drinking habits, as well as the differences in tolerance related to the different gender, age and also occasion.

The basic idea that people have of “binge drinking” is the one of drinking so much to be intoxicated with the will of getting drunk, but actually, the problem with this explanation is that no one is assessing the speed of the drinking and the concentration of alcohol present in the blood, which can surely be considered more objective measures of intoxication.

After the comment upon the different uses of the term, for the purpose of my paper I will always refer to “binge drinking” with the consumption of 5 drinks or more in a single short episode, because I want to take into account the concept of heavy drinking but considering it as something not only occasional but also restricted to few hours, which is usually a night, since the sample I will analyze is the adolescents, for whom the latter definition is surely much more appropriate and more frequently happening for many different reasons, that I will analyze deeper throughout the paper, as celebrating something, the will to have fun or the desire to try something new also due to peer pressure, the will to feel older or the hope to forget for a few hours about the issues and to reduce stress.

Diffusion and Importance of the Phenomenon

According to the World Health Organization for the European Region (WHO, 2014), alcohol is the sixth cause of ill health and premature death in high income countries and the proof stays in the fact that Europe has the highest level of consumption of alcohol in the world, which is mainly driven by some very high percentages of consumption in Eastern and Central Europe. Going deeper, the percentage of adults who drink alcohol is roughly 70% and the individual average yearly consumption of pure alcohol is around 10.7 litres, with the male portion consuming in proportion more than women, reaching an average male quantity of 19.4 litres per year.

Moreover, the use of alcohol among adolescents is another major public health concern in many European countries. According to the Health Behaviour in School-aged Children (HBSC), which is a WHO cross-national study, in the WHO European Region one fifth of the young population aged 15 and over shows ‘binge’ use of alcohol for recreational purposes and the rate at which this happens is the highest in the world (HBSC, 2016).

Among adolescents, according to the different ages the percentage of alcohol use is very different. As it can be seen in the graphs below, at the age of 11, the presence of weekly drinking and drunkenness is quite low, but as the age increases to 13 or 15, the percentage of this phenomenon increases significantly.

Alcohol can harm through both short-term use, which is often related to intense use as binge drinking episodes and is considered as the acute effects of drinking, and long-term use, depicted as the chronic effects. Alcohol is a toxic substance which can harm any organ or system of the human body and can lead to many different underestimated but serious diseases, such as an increasing risk of appearance of a wide variety of cancers, cardiovascular diseases, lung diseases, gastrointestinal conditions, it can affect the developing fetus with an intensified risk of premature birth and low birth weight and it can also impact the development of the brain.

This latter concern is even more worrying when taking into consideration drinking alcohol during adolescence, because by negatively affecting the brain, it causes severe consequences for the cognitive function and the emotional and social development. When analyzing the impact on health of alcohol and decide to measure it through the Disability-Adjusted Life Years (DALYs), alcohol is responsible for 2% of female and 12% of male premature death and disability, which makes this drug dependence one of the highest risk factors for the people in the European Union.

One DALY can be referred to as one year of healthy life lost, and the sum of DALYs of the whole population can be considered as a measurement gap between the current health status and the ideal situation in which the population lives to an advanced age without diseases and disabilities. Alcohol can also harm other people than the drinker, and this can happen for example, through violence, domestic violence and traffic accidents due to the loss of the clarity of thought.

When taking a deeper insight into the consequences, it can be better understood the seriousness of this phenomenon. The use of such a substance causes 17000 deaths per year due to road traffic accidents which is one third of all road deaths, 2000 homicides, 10000 suicides which are one sixth of all suicides, 27000 accidental deaths, 50000 deaths because of cancer, 45000 deaths from liver cirrhosis and 200000 episodes of depression (,). It is quite straightforward that the greater in magnitude and frequency of the episodes of heavy drinking, the greater the negative impact on life and the greater the probability of incurring in one of the previously mentioned diseases or conditions.

Among young adolescents the incurring of diseases and disabilities is quite low, but as they grow the rate increases strongly, mainly related to road accidents, unsafe sex, psychological problems and alcohol use (Gore et al., 2011). Suicides, homicides and other accidents are highly linked to the consumption of alcohol and since the 80% of young deaths are connected to these causes and not to the appearance of cancers or infections, the presence of powerful interventions will prevent them to drink so much and often and will strongly reduce the mortality rate among young adults (Carpenter and Dobnik, 2011).

Data and Sample

I use data coming from the fifth data collection wave (2011) of the European School Project on Alcohol and other Drugs (ESPAD) survey, which is the largest cross-national research database on the use of substances among 15 and 16 year-old adolescents in the European Union. In order to reach this work, there was also the support from other entities, such as the Swedish Ministry of Health and Social Affairs, the Portugal European Monitoring Centre for Drugs and Drug Addition (EMCDDA) and the Pompidou Group at the Council of Europe, which funded also the participation of the researchers coming from central and eastern Europe in the annual Project Meetings.

The aim of this project was to collect comparable data on this topic in as many European countries as possible and the long-term objective is to monitor trends and compare those trends within and between countries, this is why the survey is repeated every four years, with the 1995 as the starting year. In order to collect the data, it was established a common methodological protocol including also a master questionnaire in the early 1990s and for pragmatic reasons it was conducted among students in classrooms.

The participation of the students is voluntary and anonymous with teachers or assistants that act as survey leaders and the results are presented only in a comprehensive way including all the involved countries. All the steps are conducted in a standardized way, from the data collection period, which is usually spring, to the capture, from the cleaning to the delivery and the final analysis of the collected data, in order to have results that are as much comparable as possible.

The first wave of the ESPAD survey, in 1995, contained information coming from 26 countries, while the sixth wave, the one of 2011, includes data from 36 countries, with the subsequent collection of results from other three countries in autumn of 2011. Consequently, we can safely say that now ESPAD survey covers more or less the whole European continent, with information in the different waves coming from 40 countries.

The samples of 2011 can be considered nationally representative, with the exception of four cases: in Belgium the survey was conducted only in the Flanders, which is the Dutch-speaking part, in Bosnia and Herzegovina it encompasses only the Republic of Srpska, in Germany five out of the sixteen Bundesländer are included and in the Russian Federation only the city of Moscow was incorporated.

The overall validity of the results is considered to be high in most countries, albeit we have to take into account the different cultural contexts in which the students have answered to the questionnaire. Where the size of the country in terms of population allowed for, the national sample size was usually close to 2400 participating students, but as it can easily be understood, for some smaller countries, that number was higher than the whole population of students of the desired age, so it would be impossible to reach that sum.

Measurement of Variables

The dependent variable chosen is the dummy binge drinking. The initial idea for the dependent variable was binge drinking (5+ drinks), which comes from the following question ‘How many times (if any) have you had five or more drinks on one occasion in the last 30 days?’. Subsequently, by looking at the descriptive statistics and at the histogram of that variable (Table 1), it was evident that it was extremely right-skewed and the possible attempts to try to reduce it where quite unsatisfactory, so the only possible choice was to transform it to a dummy variable and compute a logistic regression instead of the linear one.

The new dummy variable has value 0 in the case of the absence of binge drinking and 1 for the presence of it, regardless of how many times it occurred in the last month. By doing this, I decided to focus my attention not on the frequency of such phenomenon among my sample, but just on the presence or not. Nevertheless, for the purpose of my analysis and having considered that my sample is composed of adolescents, analyzing the presence or absence of such heavy drinking is sufficient and satisfying because the fact that a 15-year-old adolescent in the last 30 days incurred even only in one episode of such behavior is in any case already something to worry about.

Independent and Control Variables

The first independent variable that I want to study because of the findings in previous studies is the Family well off, which according to the questionnaire, indicates the perception of well-being of one’s family compared to the other ones and is measured on a scale from 1 – very much better off, to 7 – very much less well off. The second couple of variables that is of my interest was Mother and Father education, which at the beginning I thought could be highly correlated with the family well off, but by looking at the Variance Inflation Factors (VIF) coefficients, which is an instrument to detect multicollinearity between independent variable, I saw that it was not the case.

Furthermore, in order to avoid the possibility of correlation between the two variables of education, I decided to create only one single variable called Education, which measures the highest level of education between the two parents, with a scale going from 1 – completed primary school or less, to 5 – completed college or university and furthermore, there was a country, the Federation of Bosnia and Herzegovina, which didn’t have valid values, so I put that country’s results as system missing. Moreover, another important independent variable to consider was the Average grade at school at the end of last term, for which Hungary hadn’t have values, so I put them as missing.

In addition, the variable has outcomes that were spread from 1 to 345 with categories from 1 – majority 10s, to 7 – majority 4s, so I decided to calculate the outliers, with the formula [mean ± 3.5ẟ] and after having calculated the confidence interval, which was [-6.0175;10.835], I decided to leave only values going from 1 to 10, and put as system missing all the values greater than 10. Furthermore, another independent variable chosen was Friends get drunk, which indicates the number of friends estimated to get drunk and is measured on a scale from 1 – none, to 5 – all the friends.

Also this variable has problems regarding one country, which in this specific case is Netherlands, for which I put its outputs as system missing. At the beginning I wanted also to include in my analysis the variables Mother and Father on drunk, but when looking at the frequency tables, it was clear that they had roughly 60% of missing values, surely too much to be included in my regression.

In addition, there were three variables that were extremely right skewed according to their histograms, consequently, to avoid this problem I transformed them into dummy variables. The first one measures how many days the adolescent in question missed school in the last 30 days because he/she decided to skip or “cut” and ranges from 1 – none, to 6 – 7 days or more and what I did was to create a dummy with value 0 – never, and 1 – skipped some days.

The second variable is Daily smoking, which measures how frequently the young adult smoked cigarettes in the last 30 days, with 1 – indicating not at all smoked, and 7 – more than 20 cigarettes per day. I transformed it into the variable Dummy smoking, which has value 0 – no smoking and 1 – smoking, without indicating the frequency of such activity. The third is Cannabis lifetime, which evaluates the number of occasions in which the boy/girl used cannabis during his/her life and is measured on a scale from 1 – never, to 7 – 40 or more. Also in this situation I created a dummy variable, which has the following values, 0 – no cannabis use and 1 – cannabis use.

Another aspect I wanted to analyze was the relationship that the adolescent has with the people around him, to better understand whether the environment in which he/she lives is comfortable for him/her, therefore I included in my regression also other three variables considering the satisfaction of the relationship with mother, with father and with friends, which are measured on a scale ranging from 1 – very satisfied, to 6 – no such person. Again because of the distribution of the variables I transformed them into three dummies with value 0 – not satisfied, corresponding to the values from 3 to 6 in the initial variables, and 1 – satisfied, standing for the values 1 and 2.

Finally, I added also some control variables, which are sex and country. The former is a dummy and has the following categories, 0 – male, and 1 – female. For the latter, I created as many dummy variables as the number of country to add them to my regression analysis I put the country “Albania” as my reference category, so in the end not included it in my analysis.