Background
The use of tobacco and other drugs account for almost 5 percent of the global burden of disease in terms of disability-adjusted life years [
1], with the prognosis that tobacco use will result in one billion deaths in the 21st century [
2]. In addition, substance use disorders are at their peak in young people aged 16 to 25 years [
3], at a time when many young people are attending university [
4]. A recent study demonstrated that the prevalence of tobacco use in college students was high, with 26.2 percent of students using any tobacco product in the 30 days prior to sampling [
5]. In this study [
5], smoking cigarettes was the most common method of tobacco use, with 18.6 percent of students smoking cigarettes in the previous 30 days. Other nonalcohol substance use is also prevalent, with a recent study indicating that 9.4 percent of first-year students met criteria for cannabis use disorder [
6]. Drug use in young people is typically initiated at the age of, or just prior to, commencement of study at university. For example, in Australia, the mean age of initiation of tobacco use is 16 years and of cannabis is 18.5 years [
7]. Early intervention for substance use offers the potential to prevent the development of clinically significant problems. The university setting is therefore an ideal environment in which to provide both broad-scale preventative and treatment approaches [
8] for tobacco and other drug use disorders in this group.
The provision of screening and brief intervention has the potential to reduce substance use among university students [
9]. However, the high clinical load often experienced at university health clinics may limit the feasibility of face-to-face screening and intervention [
10]. Students may also be reluctant to seek help from counseling centers in person [
11], with Australian research indicating that young people are particularly unlikely to seek help for drug and alcohol use disorders [
12]. A prevalence study conducted in the United States found that rates of substance use were similar between students and nonstudents, but that students who experienced substance use disorders (including alcohol) were less likely to seek help for these problems than young adults in the community (OR = 0.52; 95% CI = 0.30–0.90) [
13]. As a result, very few university students will access appropriate care [
14] for tobacco and other drug use problems.
Barriers to treatment for substance use problems include cost, difficulty accessing facilities, and stigmatization [
15]. The development of online interventions has the potential to circumvent these barriers and to provide a higher level of scalability at minimal marginal cost per user [
16]. This is important in the university context, where traditional campus mental health services [
17] can be more time consuming for the therapist and less cost-effective than distal interventions [
18]. An important benefit of online interventions is “24/7” availability, allowing access either at times of high motivation to change behavior or during periods of increased risk of relapse [
19,
20]. Publically delivered telephone interventions also utilizing additional methods such as coaching and quit packs, are also often accessible outside regular work hours for this purpose [e.g., Quit for Life, USA (24 hours), Quitline, Australia (8 am–8 pm, Monday–Friday). Additionally, technology-based interventions are easily disseminated, relatively inexpensive, and highly relevant to university populations who are familiar with the use of the internet for health-related problems, particularly information-seeking [
21-
23].
Computer-delivered interventions appear promising in reducing symptoms of other types of mental health problems [
24] as well as for alcohol use [
25] in university populations. Previous reviews have examined internet/computer-based tobacco interventions [
26-
29] or other drug use interventions with college students [
30] and in schools [
31]. In addition, previous studies have employed a narrow definition of technology confined to the internet or computers and excluded other types of technology (e.g., telephone). Therefore, no reviews have examined both tobacco and other drugs and used a broad definition of technology, including both the internet and other types of technology (e.g., SMS, telephone). The current study systematically reviewed published randomized trials of technology-based interventions evaluated in a tertiary setting for tobacco and other drug use (excluding alcohol).
Discussion
The current systematic review identified 12 randomized trials detailing 20 technology-based interventions targeting tobacco or other drug use in tertiary students. The majority of papers (n = 9) targeted tobacco use, with eight of these targeting smoking. The meta-analysis conducted on the subset of tobacco studies reporting abstinence demonstrated that the interventions increased the rate of abstinence by 1.5 times that of controls. The duration of this abstinence varied between 7 and 30 days (one study did not report required duration). Only two studies targeted marijuana use, and one study used a multifaceted approach targeting stress, marijuana, alcohol, and tobacco use. Outcomes for the marijuana and multi-targeted study were mixed. Overall, three-quarters of the interventions were delivered using computers or the internet, with a minority using telephone or SMS technology.
The current review and meta-analysis indicates that technology-based interventions are promising for reducing tobacco use in tertiary students. Our findings are similar to those by Myung et al. [
26] reporting on 22 internet- or computer-based smoking cessation programs that yielded an abstinence rate 1.5 times higher than the control group (RR, 1.44; 95% CI, 1.27–1.64). Considering internet-based interventions alone, Civljak and colleagues [
53] reported benefits, especially where the information is appropriately tailored and employs frequent automated contacts. However, the internet-based programs did not show consistent effects. Overall, this suggests the use of either proximal technology-based interventions or more intensive/tailored internet programs.
The majority of interventions for tobacco use in the meta-analysis were compared with usual care control conditions, indicating that additional, tailored content may increase abstinence in this group as suggested above. Indeed, the addition of tailoring to several usual care (standard) interventions [
52,
59], specifically to the young adult age group, was associated with moderate effects. The intervention with the strongest effect size was the most intensive, including a 30-week program with access to a website plus personalized follow-up emails from peer coaches [
60]. This intervention also included a chance to win a $3000 prize for all participants. This raises the possibility that more intensive interventions may result in higher abstinence rates. Nevertheless, even brief written information was found to be effective in some studies [
52,
61], albeit not others [
64]. While age-tailored content appears useful, further research is needed to determine which other information is most effective in increasing abstinence in this population.
The two marijuana interventions were not effective at reducing or preventing marijuana use [
67,
68]. The results of a recent review were inconclusive regarding the effectiveness of prevention programs for marijuana in young people [
70]. The authors attributed this in part to the poor methodology of the included papers, where approximately half of the included 25 studies failed to deal appropriately with attrition or provide sufficient data to calculate effect sizes. In contrast, the quality of the two marijuana studies in the current review was relatively high. Nevertheless, neither reported a positive outcome. The authors of these papers suggest that future studies should target certain groups of students (i.e., those with family history of the substance, and those with higher contemplation of changing their marijuana use) [
68], as well as nonusers, and examine longer-term follow-up data (>1 month), where preventative effects may more likely be detected [
67]. Additionally, a previous paper suggested it was important to take into account participant preference for technology type in marijuana interventions (e.g., telephone vs. web-based [
71]).
Less than half of the papers included in the present review examined interventions that were delivered distally. Given that distally-delivered internet interventions may be perceived as less stigmatizing than traditional approaches to care [
18,
72] and that smoking [
73] and drug use [
74] are stigmatized among young people, internet interventions may hold greater appeal for students who may be concerned with stigma associated with seeking services on campus [
75]. Therefore, it is important that further studies investigate distal methods, such as the internet, to deliver interventions targeting tobacco and other drug use. In addition, only one study using mobile phones was identified in the current study. Although it was not found to be effective, more research on this device is needed. Young people aged 16–34 years are highly familiar with mobile technology and are the most likely age group to own a smart phone [
76]. Therefore, further investigation into the delivery of interventions for this group using mobile phone applications is particularly important, as a systematic review examining the use of mobile phones for smoking cessation in the general population reported sustained benefits for this approach [
77].
The majority of included studies targeted smoking, which is unsurprising given the high prevalence of smoking [
5] compared to other types of drug use [
6]. Tobacco use is the leading cause of preventable illness and premature death across the world [
78]. Given the high prevalence of smoking and its associated morbidity/mortality, even small improvements in cessation could have a major impact on public health. Despite this and the relevance and applicability of technology-based approaches to this population [
21-
23], very few technology-based programs for tobacco use, especially prevention programs, were identified in the current review. Further technology-based interventions on tobacco use are needed in this at-risk group.
Abstinence remains the gold standard of smoking cessation interventions, as exemplified by the “Russell Standard,” which recommends continuous abstinence for 6 months as the key outcome for cessation trials [
50]. None of the studies reported against such a stringent criterion. Four [
59,
60,
64,
65] reported an abstinence measure at 6 months, but all used short-term abstinence (e.g., 7 days) at this time point. Some studies in the present review chose to measure other outcomes such as intentions to quit, particularly for nontreatment interventions. Interventions for other substances may take a “harm reduction” approach [
19]. Under this approach, reducing the frequency of cannabis use, as sought by Lee and colleagues [
68] would be regarded as a credible outcome.
The majority of studies utilized self-report methods, although some combined these measures with chemical tests using cotinine [
59,
65] or carbon monoxide testing [
60]. However, many studies still rely on self-report. For example, An et al. [
60] used carbon monoxide testing, yet used self-report 30-day abstinence as the primary outcome variable, with the authors suggesting that their testing equipment was not sensitive enough to detect occasional smoking [
60]. Chemical detection methods can be problematic. Previous research [
79,
80] has documented high levels of inconsistency between self-reported behavior and biological verification. More sophisticated methods of detecting residual chemicals from tobacco smoking should be investigated [
60].
The studies displayed some methodological problems. All RCTs failed to report sufficient information about randomization concealment, and almost two-thirds failed to report adequate randomization methods. Encouragingly, the majority of studies employed ITT analyses, including almost all of the studies included in the meta-analysis. However, no studies met the criterion of blinding the “outcome assessors” to the allocated interventions. This is because in psychological interventions using self-report, the participant is the outcome “assessor” and is not able to be blinded to their condition [
81]. Additionally, as pointed out by Farrer et al. [
24], study quality criteria could be refined to more accurately assess the quality of internet-based research.
Limitations
There are limitations to the present review that require consideration. First, given the present review sought to evaluate the evidence relevant to interventions in tertiary student populations, the results may not be applicable more broadly to the general population. A second limitation is that the sample size of studies included in the meta-analysis and the detected effect size were relatively small. Additionally, the current review searched three databases, and it is possible that relevant journals may not be indexed by these databases. However, hand-searching previous reviews and key papers was utilized to address this issue [
82]. Finally, it is possible that the incorporation of published papers only may have biased the review, given that publication may be biased towards papers with positive outcomes [
83]. Likewise, this also applies to the inclusion of English-language studies only.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
AG drafted the initial form and revisions of this manuscript, coded the papers, and conducted basic analyses. RT provided input into the introduction and feedback on the analyses and manuscript content and conducted the meta-analysis. LF coded papers, drafted the method, and provided assistance with the meta-analysis as well as input into the manuscript. JC conducted the searches and coded the papers. KG provided input, feedback, and refinements to the paper. All authors were involved in the development of the review plan, provided comments and revisions, and agreed to the final manuscript.