Discussion
The objective of this study was to examine alcohol use trends among NTNU students in a twelve-year perspective. We identified moderate reductions in the frequency of alcohol intake, the quantity of alcohol consumed, and the occurrence of binge drinking among students in the period 2007–19. The proportion of students with risky alcohol use as determined by the AUDIT scores fell by as much as 19 % between 2007 and 2019, but the alcohol use of over half of the students still poses a long-term risk.
Consuming five or more units of alcohol on the same occasion is classified as binge drinking (10). Frequent binge drinking increases the risk of negative alcohol-related consequences, including injuries, physical violence and sexual assault (15). The reduction in binge drinking trends found in this study is therefore positive from a public health perspective. Our findings related to binge drinking show that this is more common at NTNU (68 %) than among students in Bergen (61 %) (8) and students nationwide (student health and satisfaction survey, the SHoT study; 57 %) (13, 14). The difference must be viewed in the context of the higher average age of students in the Bergen study (24.9 years), as well as the lower proportion of male participants in both studies (8, 13). Previous studies have shown that the tendency to binge drink decreases with increasing age, and that men drink larger quantities and more frequently than women (5, 8, 11, 13).
The high proportion of students with risky alcohol use in our study is consistent with results from previous studies showing a high prevalence of risky alcohol use among students in Norway (8, 12, 13). Erevik et al. found that the prevalence of risky alcohol use among students in Bergen in 2015 was 53 % (8), while the SHoT survey in 2018 reported risky alcohol use in 43 % of students nationwide (13). In our data, the proportion was somewhat higher, at 56 % in 2019. Caution should be exercised when interpreting the difference in results between NTNU and Bergen students due to the variations in average age and sex distribution. The prevalence of risky alcohol use among students is considerably higher than in the general population, where 15 % have risky alcohol use according to a survey by the Norwegian Institute of Public Health. In the segment of the general population most comparable to the student population (16–24 years), the proportion was 34 % (5).
Based on the data from the SHoT study, Heradstveit et al. examined the extent and development of risky alcohol use among Norwegian students in the period 2010–18. Their results showed that the prevalence of risky/potentially harmful alcohol use has been stable over the past eight years (12). This is in contrast to our study, where the proportion of NTNU students with risky alcohol use fell significantly, by 19 % in the period 2007–19 (Figure 3).
Data from the Nord-Trøndelag Health Study (HUNT3) have shown that alcohol intake varies throughout the year, with the highest consumption in the summer and early autumn (17). It is reasonable to assume that there may also be a similar seasonal variation among students, as the autumn semester includes an introduction programme for new students (buddy week) and (in odd-numbered years) Norway's largest cultural festival 'UKA'. The survey in 2007 was conducted in the spring, while in 2019 the data were collected in the autumn. Assuming a seasonal variation with increased alcohol consumption in the autumn semester, the reduction in alcohol consumption among NTNU students may be greater than observed in this study. The timing of the data collection for the SHoT surveys also differed: autumn 2010 and spring 2018. This reverse seasonal variation may also have resulted in a smaller relative difference between the samples (12).
In the SHoT survey from 2010, two-thirds of the students reported problematic high alcohol consumption levels in student environments and called for more alcohol-free options. This proportion was lower in the SHoT survey from 2018 (13), which may suggest that this data also reflects a change in students' drinking habits and attitudes that is consistent with the trends observed in our study. Our results indicate that the prevalence of risky alcohol use among NTNU students, just like minors and the general adult population, has fallen over the last decade (5, 18). The reasons for this shift are unclear, as national measures such as taxation and reduced availability have remained the same. Changing attitudes in the student community and the correspondingly stronger focus on alcohol-free options may have contributed to the reduction.
The data from the SHoT survey in 2018 represent the most comprehensive survey of Norway's student population, with over 50,000 participating students, but the response rate was low (31 %) (13). Our study includes only 3.9 % of NTNU's student population in 2019 but has a high response rate (96 %). The difference in response rate is probably a result of methodological differences; the SHoT survey is conducted online, while the NTNU survey was via a printed questionnaire, which generally leads to higher participation (19).
The SHoT survey has the capability to capture the entire student population since it is conducted online, but its low response rate makes it vulnerable to sample bias. For example, selection bias can result in an overrepresentation of students with low to moderate alcohol consumption, as these may be more likely to respond. This represents a subgroup in our dataset that has changed little between the survey periods. Our own study only captures students who attend lectures. Our measures to ensure a representative sample among these students included attendance at various faculties and at different times of the day. Our study has a high response rate and demographic diversity. It is therefore reasonable to assume that a representative sample across the student population has been achieved.
The use of only two measurement points is a major methodological limitation in the study that restricts the opportunities for commenting on ongoing trends during the time period in question. Relevant findings therefore need to be interpreted with caution.