Hazardous alcohol use among university students

University students have been found to have considerably riskier drinking patterns than their non-student peers [9–11]. For example, in New Zealand, a random sample of students (response rate 82%) from one large public university were found to have a prevalence of hazardous drinking (a score of 8 or higher on the Alcohol Use Disorders Identification Test; AUDIT[12]), which was double that of their peers in the general population [9]. The prevalence of harmful drinking, as indicated by a score of 15 or higher on the AUDIT, was three times higher than that among their non-student peers.

Previous research on e-SBI

On the basis of extensive development research conducted at a university student health service in New Zealand [13, 14], a first randomised controlled trial involving 104 participants was conducted in 2002. Students who screened positive for hazardous drinking (AUDIT ≥ 8) at an initial assessment, conducted electronically in the reception area, were randomly assigned to a leaflet-only control group or to receive an assessment and personalised feedback intervention, delivered entirely via the Internet [15]. This took an average of 15 minutes and was completed during the waiting time to see medical staff. Participants were followed up after six weeks and again after six months. Relative to controls, those who received the intervention drank 26% less alcohol after six weeks and had 24% fewer problems six months later [16].

These promising results were confirmed in the next trial of e-SBI, in which 1,010 students were screened over a 3 week period in the reception area and 576 were randomised to intervention or control groups. Participants were re-assessed 6 and 12 months later, with 85% retention. The effects were similar to those of the first trial [17]. At six months, relative to controls, patients receiving e-SBI reported significantly lower drinking frequency (-21%), lower total consumption (-23%), and fewer academic problems (-24%). Encouragingly, intervention effects endured. At 12 months, significant differences in total consumption (-23%; equivalent to 3.5 standard drinks per week) and academic problems (-20%) remained, and AUDIT scores were 2.2 points lower than those of controls [17].

This reduction in AUDIT score was estimated to be equivalent to an absolute risk reduction of 9% (95% CI 3% to 14%) in diagnoses of alcohol abuse and dependence [18]. Assuming the program could be implemented with 50% of the New Zealand student population, this equates to 1,424 cases of alcohol use disorders prevented per year, a significant public health benefit. A clear limitation of the primary care based delivery of e-SBI is that in many universities the student health service does not provide healthcare to the majority of the student population. Accordingly, to realise the population-level benefits of e-SBI, a pro-active case finding approach would be required.

To address this limitation, a third trial was conducted at an Australian university, where we sought to determine whether an e-SBI program called THRIVE (Tertiary Health Research Intervention Via E-mail) could be delivered on the basis of a universal screening program, i.e., by-passing the primary care setting. In addition to circumventing the problems of interfacing with a busy primary care service, the approach takes advantage of the economy of scale that can be achieved with the Internet, making it possible to offer assistance to thousands of students at low cost, including many who would not routinely come into contact with health services of any kind.

We invited 13,000 17-24 year-old students to complete a web-based AUDIT and 7,237 responded [19]. A third (n = 2,435) scored in the hazardous/harmful range (≥ 8) and were randomised to THRIVE or screening alone, and 2,050 (80%) completed at least one follow-up assessment. Intervention, delivered immediately following the assessment, consisted of 10 minutes of web-based assessment and personalised feedback. After one month, participants receiving intervention drank significantly less often (-11%), smaller quantities per occasion (-7%) and consumed a lower volume of alcohol overall (-17%), than did controls. At six months, intervention effects persisted for drinking frequency (-9%) and volume of alcohol consumption (-11%) [20].

Overall, the effects seen in the New Zealand trials were replicated but were somewhat smaller in the THRIVE trial. Nonetheless, given the reach of the intervention delivered on the basis of universal screening, THRIVE has greater potential to produce a population effect than primary-care based delivery of e-SBI.

The need for large effectiveness trials

We are at the stage when there is a pressing need for large effectiveness trials. There being several recent trials conducted among university students, however, much of the research has been conducted in conditions which would generalise poorly to practice, and, with one exception, a study whose results are yet to be published [21], there have been no large, multi-site effectiveness trials. Trialling the intervention across multiple sites has the advantages of testing the robustness of its effects across different student drinking cultures, which national surveys have shown to vary in levels of consumption [22] and exposure to alcohol outlets [23] and promotion [24].

In addition to being firmly rooted in the experience of three previous clinical trials, the proposed study is based on extensive experience in surveying university student drinking via the Internet. In 2005 a web survey (response rate 65%) was conducted with large random samples of students from six New Zealand university campuses [22]. The survey was repeated in 2007 at eight campuses (response rate 68%). New Zealand is now uniquely placed in having up-to-date national student alcohol consumption data (from the 2005 and 2007 student surveys) on which to base the proposed intervention.

Responsiveness to Māori health

New Zealand was founded on the basis of a treaty between the indigenous (Māori) peoples and those, largely from Great Britain [as it was then], who were part of the colonial expansion of European peoples. The Treaty of Waitangi, signed in 1840, and recognised as the founding document of New Zealand, grants to Māori the rights and privileges of British subjects.

The alcohol research undertaken among New Zealand university students has been conducted within a Māori health or Treaty framework. Accordingly, we adopted an Equal Explanatory Power model [25] for the national student surveys. We sought to invite equal numbers of Māori and non-Māori students from each campus in order to maximize the explanatory power of the study for Māori, who have traditionally been poorly served by population surveys, despite bearing a considerably greater burden of alcohol-related harm than non-Māori [26]. We have maintained this approach in the trial design, described below.

Design

The study consists of two parallel, multi-site, double blind, individually randomised controlled trials. The trials are recruiting Māori and non-Māori students aged 17-24 years from seven of New Zealand's eight universities. Figure 1 shows the trial design.

Sampling

In cooperation with each university administration, we have drawn random samples of undergraduates aged 17-24 years. This was done separately for Māori and non-Māori participants on the basis of self-reported ethnicity in university enrolment forms. The aim was to have equal numbers of Māori and non-Māori enrolled, however, that was impossible to achieve on some campuses, given low enrolments of Māori. In such campuses, we have invited all Māori in the requisite age group to participate and we increased the numbers of Māori invited from other universities accordingly.

Recruitment

We have adapted a survey recruitment approach that has been used extensively and which has been described in detail elsewhere [27, 28]. Four weeks after the start of the first semester random samples of students aged 17-24 years will be sent an e-mail containing a hyperlink to a web questionnaire. Up to three reminder e-mails will be sent in the following weeks to non-respondents.

Instrument, randomisation, and intervention

Respondents will visit the study website and be presented with an introductory page followed by one page of five questions concerning (1) gender, (2) age, and the three items of the AUDIT-Consumption subscale [12]: (3) frequency of drinking; (4) number of standard drinks (defined as containing 10 g ethanol) consumed per typical drinking occasion; and (5) frequency of having six or more drinks per occasion. The AUDIT consumption subscale (AUDIT-C) is a valid screening tool, with specificity and sensitivity similar to the full AUDIT [29] and its psychometric performance has been previously evaluated in online studies with university students [30].

Upon clicking the submit button on this page, AUDIT-C scores will be calculated. Those who score ≥ 4 on the AUDIT-C will be randomly allocated by the web server to either the intervention or control group. The control group will be sent to a Thanks page at this point and advised that they will be contacted again by e-mail in second semester to complete a similar brief questionnaire.

The rationale for limiting the control group's alcohol questions to three items is that previous research shows that simply asking questions about alcohol consumption can act as intervention, producing reductions in self-reported drinking levels [31]. Indeed, such effects have been shown for the AUDIT alone [32].

Those in the intervention group will continue and be asked AUDIT items 4-10, and additional questions concerning the largest number of standard drinks consumed on one occasion in the last four weeks, the duration of the drinking episode in hours, and their body weight (for the purpose of estimating peak blood alcohol concentration). They will also complete the Leeds Dependence Questionnaire (LDQ; 10 items) [33]. The psychometric performance online of both the AUDIT and the LDQ has been confirmed in previous study with university students [34].

The intervention group will then receive personalised feedback consisting of: their AUDIT score and an LDQ score with an explanation of the associated health risk and information about how to reduce that risk; an estimated blood alcohol concentration (BAC) for the respondent's heaviest episode in the previous four weeks, with information on the behavioural and physiological sequelae of various BACs, and traffic crash relative risk; estimates of monetary expenditure per month; bar graphs comparing episodic and weekly consumption with that of other students and members of the general population of the same age and gender; and hyperlinks for help with drinking problems. Further web pages will be presented as options, offering facts about alcohol, tips for reducing the risk of alcohol-related harm, and where medical help and counselling support can be found. A demonstration version of the instrument can be viewed at http://ipru3.otago.ac.nz/limesurvey/.

Follow-up

A high follow-up rate is clearly vital to the success of e-SBI trials and several approaches have been adopted to maximise this. In the second semester, approximately five months later, all participants will be invited to complete a brief on-line follow-up assessment, with the chance to win a $500 supermarket voucher or an Apple iPad. The assessment will include questions concerning the frequency of drinking and amount consumed per typical drinking occasion, all with a reference period of the last 4 weeks. In addition, participants will be presented with the five questions comprising the Academic Role Expectations and Alcohol Scale (AREAS) [35], plus questions concerning their subjective experience of the study.

Blinding

Study participants will be unaware that they are involved in a trial, and that they have been randomised. They will have consented to complete a web survey of "the experiences of tertiary students with alcohol over time". The attention of the intervention group will not be drawn to the attempt to influence their behaviour. These decisions both serve to minimise bias relating to assessment reactivity and other aspects of study participation [36, 37] and are also congruent with the aspiration to undertake the study in conditions as close as possible to those in which it would be routinely delivered.

Sample size estimation

The same sample size will be sought separately for Māori and non-Māori students. The estimate is based on the following parameters and assumptions.

Outcomes

The primary outcomes are (1) volume of alcohol consumed per week, (2) frequency of drinking (occasions per week), (3) amount consumed per typical drinking occasion, (4) the proportion who exceed New Zealand recommended upper limits for: risk of acute harm (no more than 4 drinks per occasion for women; no more than 6 drinks per occasion for men); and (5) risk of chronic harm (no more than 14 drinks per week for women; no more than 21 drinks per week for men); and (6) scores on the AREAS [35].

Analysis

The primary outcomes will be analyzed with negative binomial regression for panel data, using the Stata xtnbreg procedure [41]. For the proportions of students exceeding recommended upper limits we will use generalized linear mixed models with the xtmelogit procedure [42, 43]. The results will be presented as readily interpretable risk ratios and odds ratios respectively. Participants will be analysed in the group to which they were randomized (intention to treat), with multiple imputation methods used for any missing follow-up data [44, 45].

Ethical approval

Ethical approval has been given by the Multi-region Ethics Committee, Ministry of Health (Ref: MEC/10/01/009).