Heavy alcohol intake amongst the student population has implications for the individual, educational institutions, and wider society [2–4]. Students have been reported to drink at higher levels than non-student peers [5–9], making this an issue of public health concern given the negative social and health consequences of heavy intake [10] and the link with other unhealthy behaviours (e.g. cigarette smoking [11] and recreational drug use [12]).

Across the world it has been reported that university students' levels of alcohol consumption are higher than that of their non-university peers [5, 13, 14]. Within the United Kingdom, a review of studies measuring undergraduate drinking concluded that 52% of males and 43% of females reported drinking above the recommended limits of 21 units per week for men and 14 units per week for women [15]. The comparable figures for 16–24 year olds in the general population are 37% and 33% respectively [16]. Furthermore, 15% of a UK sample of 3075 students drank at hazardous drinking levels of 51 or more units per week for men and 36 or more units per week for women (within the study 1 unit (10 ml ethanol) was taken as 1/2 pint beer/one glass wine/one measure spirits) [17].

The alcohol consumption of university (i.e. college) students in the US has been extensively studied [4, 18] and it is generally agreed that, while the trajectories of groups of students differ during their time at university [19–21], the overall trend in consumption is one of increase from high school to university, followed by a decline post-graduation. In contrast with their North American peers, the drinking patterns of students from the UK, and Europe in general, have been less extensively investigated [22]. Unlike in the US, where the minimum purchase age is 21 years, United Kingdom students can legally purchase alcohol from the age of 18, thereby creating a different context for alcohol consumption. Despite research consistently highlighting high consumption levels within the UK student population [15] there is uncertainty regarding whether/how these alcohol consumption patterns change over academic stay.

The majority of UK studies that have considered year of study have utilised a cross-sectional approach, restricted their samples to medical and/or dental students and have shown contradictory patterns [8, 23]. The one available longitudinal study of alcohol consumption was also complicated by differences between medical and dental students [24]. There was no overall change in mean alcohol consumption among dental students but an increase among medical students, from second year to final year. In addition, the percentage of dental students drinking more than the recommended limits decreased while it increased in medical students. Academic performance of medical students at UK universities may be adversely affected by heavy alcohol intake [7, 25]. However, it is unclear what students think about how their alcohol consumption may be affecting their studies.

As levels of alcohol intake increase, so does the prevalence of a variety of risky behaviours, including unsafe sexual activity [4, 26], behaviour leading to injury [27, 28] and damage to property [29], violence [30] and illegal behaviour [28]. Increased frequency of injury and assault [31] inevitably leads to increased strain on care and emergency services, and links between alcohol consumption and hospital admissions are well established [32]. Heavy alcohol intake has also been linked to depression and liver damage [33]. In the city where the current study took place, it is estimated that alcohol misuse costs 275 million pounds per year, the majority of which is spent on policing, repairing damage and injury and alcohol-related illness treatment [34].

The present investigation aimed to examine the patterns of alcohol consumption of two cohorts of students over the course of their undergraduate education at a university in the UK, using both cross-sectional and opportunistic longitudinal study designs. It also investigated student perceptions of the consequences of their alcohol consumption.

The current observation that around 90% of students consume alcohol at least weekly is in line with the rates in the general UK population [41]. The percentage of students consuming alcohol above the recommended sensible limits is broadly in line with the UK research cited earlier [8, 9, 24]. Importantly, the present study found a significant reduction in the number of units per week consumed over the three-year undergraduate time span, in both the longitudinal and cross-section samples. Research in the US suggests that alcohol consumption peaks at around 21 years and then declines over time [42, 43]. In the present study the consumption of UK students was highest in their first year (mean age 19 years). While we do not have pre-university alcohol consumption data this timing difference is likely to be related to the differences in minimum purchase age between the two countries. Despite the apparent difference in age of peak consumption, the current results are in line with the US literature in suggesting a 'maturing out' of high levels of consumption for many students. By their final year, around 70% of students reported drinking within low levels of consumption.

The difference in outcome between the present study and, for example, those of File et al. [23] and Newbury-Birch et al. [24] could reflect changes in student drinking patterns since 1994/1995. Alternatively, the alcohol consumption of medical and/or dental undergraduates may not reflect that of the wider student population. In addition, the timing of data collection during the academic year may also influence reported alcohol consumption [15]. It should be noted that, as in the current study, Newbury-Birch et al [24] collected during the Spring while File et al.'s [23] data were collected during the week following Freshers' (introductory) week, a time when higher alcohol intake has been observed.

Despite the significant reduction across year of study, nearly a third of students were still drinking above recommended levels in Year 3 and those who reported drinking within the high risk category during year one were more likely to still be at high risk in subsequent years. Students who maintain high levels of alcohol consumption throughout their undergraduate studies might benefit from interventions to moderate this behaviour. Research from the US suggests that understanding the developmental/time trajectories of alcohol consumption is important in determining the potential health outcomes of differing patterns of consumption over time [44, 45]. While understanding the 'normative' pattern of behaviour can be helpful, this approach is limited in identifying problematic consumption patterns over time. Future research within the UK would therefore benefit from attention to trajectories of student consumption and patterns of binge drinking. This information could then be used to target interventions at students most at risk of negative health consequences.

In line with previous findings, males reported drinking more units per week on average than females [8, 23, 24]. However, there was no significant difference in the proportion drinking at low levels (as the limits are lower for females). This implies that in terms of potential health consequences female students are drinking at levels comparable to their male peers and suggests a need for interventions for both sexes.

The current study also shows that students are aware of the negative impact that alcohol consumption has on their studies and their finances. This finding has the potential to inform future intervention strategies, as it would suggest that the negative impact of alcohol on the financial and academic aspects of university life is more salient to students and therefore they may be more receptive to interventions emphasising these areas of their lives.

In terms of strengths and limitations, the current study is the first to investigate changes in alcohol consumption in a University-wide UK student population using a longitudinal design and including participants from all Faculties. The combination of longitudinal and cross-sectional samples, the latter including many participants, was a valuable methodological feature. However, it is important to note that 1197 participants (21%) in the cross-sectional sample responded to more than one of the three surveys meaning that the data for each year were not completely independent. Although the demographic characteristics of both the longitudinal and cross-sectional samples were representative of the student population as a whole, it is possible that the behaviour of students who chose repeatedly not to respond may be different from those that engaged. In addition, participants were asked only to record the number of units of alcohol they consumed per week. Research suggests it is common for individuals to underestimate personal consumption when asked to self-report an average value using standard alcoholic units [46], so it is likely that our results underestimate the absolute level of alcohol consumption within the student population. Future research might increase the accuracy of self-reported alcohol consumption by investigating in more detail the daily and weekly consumption levels of each person. This might be achieved by asking participants to record how many of each type of alcoholic drink was consumed over the last week using a retrospective daily diary approach.

The current study did not investigate when or where drinking occurred, the frequency of alcohol free days per week, or the occurrence of heavy episodic drinking. Given the potential health implications of heavy episodic drinking [3, 6] it will be necessary for future research to investigate how heavy episodic drinking changes across the undergraduate course of study. A further study limitation was the use of a single item scale to measure the perceived negative impact of alcohol. While this provided some crude baseline information future research would benefit from using a more comprehensive and standardised measure of negative consequences.

In conclusion, the observation that full-time undergraduate students' consumption of alcohol declined over the three-year course of undergraduate study is of value in directing the limited resources available for preventive interventions. Introductory weeks for first year students in UK Universities often have social events associated with, or centred on, alcohol. This is often a time when campaigns for drinking in moderation are also at their peak. Later identification, e.g. in semester two, of higher level drinkers would be of current and future benefit. However, the persistence of high levels of alcohol consumption in a large proportion of students suggests the need for effective preventative and treatment interventions for all year groups. Poor physical health, financial hardship and declining academic achievement are likely indicators of this need, albeit ones not exclusively related to problems with alcohol.