Brief school‐based interventions and behavioural outcomes for substance‐using adolescents
Abstract
Background
Adolescent substance use is a major problem, in and of itself and because it acts as a risk factor for other problem behaviours. As substance use during adolescence can lead to adverse and often long‐term health and social consequences, it is important to intervene early on in order to prevent progression to more severe problems. Brief interventions have been shown to reduce problematic substance use among adolescents and are especially useful for individuals who have moderately risky patterns of substance use. Such interventions can be conducted in school settings. This review set out to evaluate the effectiveness of brief school‐based interventions for adolescent substance use.
Objectives
To evaluate the effectiveness of brief school‐based interventions on reducing substance use and other behavioural outcomes among adolescents compared to another intervention or assessment‐only conditions.
Search methods
We searched 10 electronic databases and six websites on evidence‐based interventions, and the reference lists of included studies and reviews, from 1966 to March 2013. We also contacted authors and organisations to identify any additional studies.
Selection criteria
We included randomised controlled trials that evaluated the effects of brief school‐based interventions for substance‐using adolescents.
The primary outcomes were reduction or cessation of substance use. The secondary outcomes were engagement in criminal activity and engagement in delinquent or problem behaviours related to substance use.
Data collection and analysis
We used the standard methodological procedures outlined by The Cochrane Collaboration, including the GRADE approach for evaluating the quality of evidence.
Main results
Six studies involving 1139 participants were included in this review. Overall the quality of evidence was moderate in the information provision comparison, and low or very low in the assessment only comparison. Reasons for downgrading the quality included risk of bias of the included studies, imprecision and inconsistency. Our findings suggested that compared to information provision only, brief interventions (BIs) did not have a significant effect on any substance use (three studies, 732 participants, standardised mean difference (SMD) ‐0.06; 95% confidence interval (CI) ‐0.20 to 0.09) or delinquent‐type behaviour outcomes among adolescents (two studies, 531 participants, SMD ‐0.26; 95% CI ‐0.54 to 0.02). When compared to assessment‐only controls, BIs had some significant effects on substance use and delinquent‐type or problem behaviours, but high levels of heterogeneity existed between studies and it was not always possible to pool the results. When the comparison was with assessment‐only conditions, studies of individual interventions that measured BI effectiveness reported significantly reduced substance use in general and in two studies reduced frequency of alcohol use specifically. When the data were pooled, BIs reduced cannabis frequency (SMD ‐0.22; 95% CI ‐0.45 to ‐0.02) across three studies (n = 407). Cannabis quantity was also reduced by BIs in comparison to assessment only (SMD ‐60.27; 95% CI ‐66.59 to ‐53.95) in one study (n = 179). However, the evidence for studies that compared brief interventions to assessment‐only conditions was generally of low quality. Brief interventions also had mixed effects on participants' delinquent or problem behaviours.
Authors' conclusions
There was limited quality evidence that brief school‐based interventions were more effective in reducing substance use than the assessment‐only condition, but were similar to information provision. There is some evidence for the effectiveness of BI in reducing adolescent substance use, particularly cannabis, when compared to assessment only. However, it is premature to make definitive statements about the effectiveness of brief school‐based interventions for reducing adolescent substance use. Further high quality studies examining the relative effectiveness of BIs for substance use and other problem behaviours need to be conducted, particularly in low‐ and middle‐income countries.
PICOs
Plain language summary
Can brief interventions delivered in schools reduce substance use among adolescents?
Adolescents worldwide are known to use both legal and illegal substances to different degrees, which can lead to other problems. Brief interventions are short programmes that aim to help reduce or stop substance use, in this case among adolescents. We conducted a Cochrane systematic review to look at the effectiveness of school‐based brief interventions for substance use and substance‐related problem behaviours among adolescents. The studies compared brief intervention programmes with two major kinds of comparison or control groups, 1) an information provision group and 2) an assessment‐only group where participants received no intervention. Results of the review were mixed as we found studies that showed that brief intervention had no effects on substance use as well as studies where the brief intervention significantly reduced substance use and other problem behaviours. The pattern of results indicated that participants who received a brief intervention generally did better in reducing their substance use than participants who received no intervention at all. However, participants who received a brief intervention did no better in reducing their substance use than participants who received information only interventions. Across these intervention programmes, the evidence of benefit we found for a general reduction in substance use (especially cannabis, and less so for alcohol) was of limited quality. In view of the risk of bias, variation in study findings and imprecision, we are unable to make definitive statements about the effectiveness of these interventions.
Authors' conclusions
Summary of findings
| Brief intervention compared to information provision for substance‐using adolescents | ||||||
| Patient or population: Substance‐using adolescents | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Estimate effect | No of Participants | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| Information provision | Brief intervention | |||||
| General substance use outcomes | The standardised mean substance use outcomes in the intervention groups was | SMD ‐0.06 (‐0.20 t0 0.09) | 732 | ⊕⊕⊕⊝ | Included both alcohol and drug use outcomes | |
| Alcohol frequency | The standardised mean alcohol frequency in the intervention groups was 0.03 standard deviations lower | SMD ‐0.03 (‐0.20 t0 0.14) | 527 | ⊕⊕⊕⊝ | Number of days of alcohol use | |
| Alcohol quantity | The standardised mean alcohol quantity in the intervention groups was 0.09 standard deviations lower | SMD ‐0.09 (‐0.26 to 0.08) | 527 | ⊕⊕⊕⊝ | Number of standard alcohol units | |
| Tobacco frequency | See comment | See comment | Not estimable | 326 | ⊕⊕⊕⊝ | Number of days cigarettes used |
| Cannabis quantity | See comment | See comment | Not estimable | 326 | ⊕⊕⊕⊝ | Number of joints (cannabis cigarettes) |
| Cannabis frequency | The mean cannabis frequency in the intervention groups was | SMD ‐0.02 (‐0.19 to 0.15) | 531 | ⊕⊕⊕⊝ | Number of days cannabis use | |
| All behavioural outcomes | The mean all behavioural outcomes in the intervention groups was | SMD ‐0.26 (‐0.54 to 0.02) | 531 | ⊕⊕⊝⊝ | Interactional Problems Score | |
| *The basis for the assumed risk (e.g. the mean control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the estimate effect of the intervention (and its 95% CI). The estimate effects for certain outcomes were not estimable due to only one study assessing the specific outcome, or extremely high levels of heterogeneity making effects across studies difficult to compare. | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Risk of bias (‐1): In two of the three studies, there was uncertainty about the risk of bias 2 Imprecision (‐1): The confidence intervals contained the null value of zero and the upper or lower confidence limit crosses an effect size of 0.5 in either direction | ||||||
| Brief intervention compared to assessment only for substance‐using adolescents | ||||||
| Patient or population: Substance‐using adolescents | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Estimate effect | No of Participants | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| Assessment only | Brief intervention | |||||
| General substance use outcomes | See comment | See comment | Not estimable | 407 | ⊕⊕⊕⊝ | Included both alcohol and drug use outcomes |
| Alcohol frequency | See comment | See comment | Not estimable | 407 | ⊕⊕⊝⊝ | Number of days of alcohol use |
| Alcohol quantity | See comment | Not estimable | 179 | ⊕⊕⊝⊝ | Number of standard alcohol units | |
| Tobacco frequency | See comment | SMD ‐0.03 (‐0.61 to 0.01) | 162 | ⊕⊕⊝⊝ | Number of days cigarettes used | |
| Cannabis frequency | The mean cannabis frequency in the intervention groups was | SMD ‐0.22 (‐0.43 to 0.02) | 407 | ⊕⊕⊝⊝ | Number of joints (cannabis cigarettes) | |
| Cannabis quantity | See comment | Not estimable | 179 | ⊕⊕⊝⊝ | Number of days cannabis use | |
| All behavioural outcomes | See comment | See comment | Not estimable | 245 | ⊕⊝⊝⊝ | Interactional Problems Score |
| *The basis for the assumed risk (e.g. the mean control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the estimate effect of the intervention (and its 95% CI). The estimate effects for certain outcomes were not estimable due to only one study assessing the specific outcome, or extremely high levels of heterogeneity making effects across studies difficult to compare. | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Risk of bias (‐1): There was no allocation concealment in the study | ||||||
Background
Description of the condition
Substance use among adolescents refers to the use of licit substances (including alcohol and prescription or over‐the‐counter medicines) and illicit drugs (cannabis, heroin, cocaine, amphetamines, methaqualone, hallucinogenic drugs). Globally, alcohol and cannabis (after tobacco) are the most commonly‐used substances among young people, and binge drinking and cannabis use are becoming increasingly normalised (United Nations 2003). Middle and secondary or high school is an especially high‐risk period for the initiation of substance use as adolescents transition from one type of schooling to another and face numerous challenges (National Institute on Drug Abuse 2003). School surveys conducted in different regions of the world, such as Europe (Hibell 2009), Australia (White 2006a, White 2006b), the USA (Johnston 2013) and South Africa (Reddy 2010), have reported a high prevalence of alcohol use among young people as well as high levels of other drug use. For example, a study of adolescent drug use across 35 European countries reported that 90% of school students reported lifetime alcohol use, while 19% had engaged in illicit drug use (Hibell 2009). An Australian school survey similarly indicated that 86% of students reported lifetime use of alcohol (White 2006a), 18% reported lifetime use of cannabis, and 17% reported lifetime inhalant use (White 2006b). In the USA national Monitoring the Future survey, lifetime and past 30 day use of alcohol was 44% and 26% respectively, while lifetime and past 30 day use of cannabis was 25% and 15% respectively (Johnston 2013). In addition, the national Youth Risk Behaviour Survey (YRBS) in the USA (Eaton 2012) reported that the lifetime prevalences for alcohol, cannabis, prescription drugs and inhalants were 71%, 40%, 21% and 12% respectively. The South African national YRBS found lifetime prevalence rates of 50% for alcohol use, 13% for cannabis use and 12% for inhalants or prescription drug use (Reddy 2010).
These high rates of substance use among adolescents are cause for concern, not only because the early initiation of substance use is a risk factor for substance abuse and dependence in later life (Winters 2008) but also because of its association with increased morbidity and mortality among young people. For example, the most recent Global Burden of Disease study found that alcohol (7%) and illegal drugs (2%) were two of the main risk factors for incident disability‐adjusted life‐years for youth aged 10 to 24 years (Gore 2011).
It is important to intervene early with adolescents who use substances as substance use is often associated with a number of other problem behaviours including withdrawal from school involvement, drinking and driving, violent behaviour and general delinquency. These kinds of behavioural outcomes have been consistently associated with adolescent substance use in studies throughout the world (Feldstein 2006; Hallfors 2006; Plüddemann 2010; Storr 2007). For example, the YRBS in the USA found that 8.2% of high school students had driven a car or vehicle after alcohol use (Eaton 2012) in the past 30 days, while in South Africa, this was reported to be 25.9% (Reddy 2010). Studies also show that substance use can play a role in criminal behaviour. Youth offenders in a recent study (Leoschut 2007) reported that they committed crimes in order to finance their drug habit. Some also reported that substance use gave them the courage to commit their crimes, or an excuse if they were apprehended. Ward 2007 also suggested that when young people are under the influence of substances they may not be able to monitor their behaviour as well as when they are sober.
Adolescents who become involved with the legal system partly due to substance use are more likely to associate with deviant networks and be disadvantaged in terms of education and employment. They are also more likely to be involved in criminal activity during adulthood (Bernburg 2003). Adolescents involved in the criminal justice system often have more psychiatric problems (Corneau 2004; Lanctôt 2007) and are more in need of drug treatment in adulthood than their peers who are not involved in the criminal justice system. For example, Corneau 2004 estimated that 12% of institutionalised adolescents need drug treatment as adults. Furthermore, substance‐using adolescents who are involved in the criminal justice system are more likely to have negative interpersonal relationships, including violent intimate partner relationships (Lanctôt 2007). If an intervention can take place early on with these adolescents it may be able to prevent the development of some of these negative consequences.
Description of the intervention
Brief interventions (BIs)
Brief interventions (BIs) are targeted, time‐limited, low threshold services that aim to reduce substance use and its associated risks, as well as prevent progression to more severe levels of use and potential negative consequences (Babor 2007). In general, BIs are delivered in person and provide information or advice, increase motivation not to use substances, and teach behaviour change skills with the aim of reducing substance use. However, the way that BIs have been defined and delivered has varied in the literature, such as the number of sessions provided, the length of the intervention sessions, and the format of delivery (Young 2012). It is thus important to recognise common elements used to define BI. One such component is the screening of potential participants. Although screening has formed part of BIs in other settings, it often does not take place in schools, although there are a few exceptions (Hallfors 2006). A second common element of BIs is their short length, as they generally last between one (Moyer 2002) and five intervention sessions (Tevyaw 2004).
In addition to advice‐giving, the common elements of successful BIs are referred to by the acronym FRAMES, and include provision of the following:
-
Feedback on behaviour and its consequences to the client;
-
Responsibility for change as the responsibility of the individual;
-
Advice for change;
-
Menu of options for change;
-
Empathy;
-
Self‐efficacy for change (Bien 1993).
These kinds of interventions were developed based upon the theoretical assumption that people are not always ready to change their patterns of substance use. In such cases, straightforward advice‐giving is of limited use and the adolescents need to recognise for themselves that their behaviour is problematic and identify their own reasons for wanting to change their behaviour. The development of this brief method was guided by a number of principles, it should be useable in time‐limited consultations; the training of practitioners should take between 12 and 15 hours; interviewers should be able to raise the subject of behaviour change in a sensitive and respectful manner; and the method itself should be flexible, meaning that it can be used with individuals at various stages of readiness to change (Rollnick 1995). Most BIs rely on principles of motivational interviewing (Winters 2007a), or brief motivational enhancement therapy (Tevyaw 2004), which focuses on building adolescents' readiness to change their behaviours. This technique provides personalized feedback on substance use together with a motivational interviewing counselling style (Miller 2002).
Relevance for adolescents
BIs have been identified as useful for individuals who have moderately risky patterns of substance use (Barry 1999). This makes this type of intervention relevant for use with adolescents who for the most part have not yet developed substance dependence. BIs seem to be better suited for those adolescents who are less set in a delinquent lifestyle and who are not institutionalised (Brunelle 2000). Tevyaw 2004 characterises BI methods as accepting adolescents as individuals, instead of confronting them and their behaviour or lecturing them as their teachers, parents and other authority figures may do. BIs could therefore be a more effective strategy for building rapport and a collaborative therapeutic relationship with adolescents than other confrontational forms of interacting with adolescents. Furthermore, the methods are seen as a cost‐effective alternative to traditional, lengthier treatment of adolescents who use substances (Tevyaw 2004).
Ideal conditions: what we do and do not know
BIs have traditionally been used in healthcare and substance abuse treatment settings (Bien 1993), but studies have suggested that their use could be expanded to other settings, such as schools (Winters 2007a). There are a number of advantages of school‐based BIs for substance‐using adolescents. Firstly, students usually are not dependent on substances yet, but a number of these adolescents may exhibit mild or moderate use, which makes them good candidates for BI. Secondly, research has shown that BIs can be conducted during school or in after‐school hours, making the intervention very accessible to students. Finally, the growing volume of BI material on how to conduct BI sessions means that they can often be run by staff available to schools, and not only by health professionals (Winters 2007a). There is also some research that suggests that BIs may work in other settings as well, such as family interventions for school‐going adolescents in terms of alcohol, tobacco and marijuana use (Spoth 2001). Recently, research has also suggested that web‐based BI programmes may be useful in reducing substance use in young adults (for example Bingham 2010). Despite the promise of school‐based BI programmes, meta‐analyses of school‐based interventions have not yet been conducted.
How the intervention might work
The goals of BIs are to assess substance use in adolescents, provide advice on these behaviours, facilitate behaviour change with regards to substance use, and motivate the adolescents to receive further treatment if necessary (Bien 1993). The primary focus of these types of interventions is to systematically target problematic behaviours (Tevyaw 2004), using a motivational interviewing framework.
The theoretical basis for BIs is grounded in client‐centred therapy, behavioural therapy, and the transtheoretical model of behaviour change. The transtheoretical model of behaviour change argues that readiness for change develops along a series of stages rather than as a fixed event that either occurs or does not occur. These steps are pre‐contemplation, contemplation, preparation, action, and maintenance, and individuals usually move between these stages before reaching termination (Prochaska 1993). From this perspective, motivation is seen as a state that can be altered rather than a trait that is inherent and cannot be changed. Since BIs are typically organised around a developmental theory of normative and non‐normative patterns of substance use, this is an appropriate theoretical orientation for a behaviour change strategy aimed at adolescents (Winters 2007a).
Why it is important to do this review
Brief interventions are recognised as an appropriate treatment for adolescents who use substances, yet there have only been a few reviews of the effectiveness of BI for adolescent substance use. Tait 2003 conducted a systematic review of 11 studies of BIs for adolescent substance use and found that BI was effective in reducing alcohol use among adolescents, but not in reducing polysubstance use. Only two of these studies were conducted in schools (with one conducted by nurses over the telephone), and these two studies showed only moderate effect sizes of between 0.377 and 0.52. BIs also did not have a significant effect on drinking in the last seven days. In their review of brief motivational interventions among adolescents, Tevyaw 2004 reported significant reductions in alcohol‐related problems such as drinking and driving, traffic violations and, to a lesser extent, reductions in drinking rates. While the reviewed studies were conducted in a number of settings, including emergency rooms and colleges, not many of these settings included high schools. Furthermore, existing reviews were conducted a number of years ago and have not been updated. It is useful to re‐examine the evidence in an updated review.
There are no existing reviews that examine the effectiveness of BIs for reducing substance use among high school (or the equivalent of high school) students specifically. Furthermore, there are no existing reviews that address BIs for substance use as a primary outcome and its behavioural outcomes (for example problem behaviours) as secondary outcomes. The current review would be the first to do so.
Objectives
To evaluate the effectiveness of brief school‐based interventions, compared to another intervention or assessment‐only, on reducing substance use and other behavioural outcomes among adolescents.
Methods
Criteria for considering studies for this review
Types of studies
We included randomised controlled trials that evaluated the effects of BIs on substance use as well as on other behavioural outcomes associated with adolescent substance use. Studies that recruited adolescents from anywhere else other than an educational setting were excluded.
Types of participants
Participants were adolescents under the age of 19 who were attending high school, secondary school, or a further education training college that provided alternative schooling or vocational training for adolescents between the ages of 16 and 18 years, and who used or misused alcohol or other drugs, or both, but did not meet the criteria for substance dependency. In addition, participants had faced negative behavioural consequences due to their substance use.
Types of interventions
Experimental intervention
The intervention should have been labelled as a BI, but could also have been defined as motivational interviewing, brief skills‐orientation, motivational enhancement, or other specific types of BIs, which were up to four sessions long and used BI principles to facilitate change. The focus should have been on building the individual's motivation to change. The BIs could have been offered as a stand‐alone option, integrated with other intervention efforts, or as a precursor to other treatments. Only BIs that were offered to individuals in a face‐to‐face modality were included in this review.
Control intervention
The control could have been no intervention, placebo, assessment only, or other types of interventions or education.
Types of outcome measures
Primary outcomes
-
Abstinence or reduction of substance use behaviour.
The outcome measures could have been self‐reported measures, including dichotomous and continuous outcomes. In addition, substance use could have been measured with standardised measures of substance use that are appropriate for adolescents such as the Alcohol Diagostic Interview (ADI), Adolescent Drug Abuse Diagnosis (ADAD), Adolescent Drug Involvement Scale (ADIS), Adolescent Alcohol and Drug Involvement Scale (AADIS), and Personal Experience Inventory (PEI), which are generally self‐report measures.
Any biological testing could also have been included, such as urinalysis for drug use and breathalyser tests for alcohol use.
Secondary outcomes
-
Engagement in criminal activity (such as theft, drug and alcohol crimes, property crimes) related to substance use.
-
Engagement in delinquent‐type behaviours (such as drinking and driving, aggression and fighting, bullying, carrying weapons to school, buying and selling drugs, gang involvement, truancy, suspension and expulsion, and disobeying rules in general) related to substance use.
The secondary outcomes refer to problem behaviours related specifically to delinquent‐type behaviours.
It was not expected that the included BIs would have adverse effects on the primary or secondary outcomes.
Search methods for identification of studies
Included studies were published from 1966 onwards, the year that BIs were first introduced.
Electronic searches
Relevant trials were obtained from searching the following sources:
1) Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library (2010, Issue 3) which includes the Cochrane Drug and Alcohol Group Specialised Register;
2) PubMed (January 1966 to March 2013);
3) PsycINFO (January 1966 to March 2013);
4) ERIC (Education Resources Information Centre) (January 1966 to March 2013);
5) ISAP (the Index of South African Periodicals), Social Science Index (January 1966 to March 2013);
6) Academic Search Premier (January 1966 to March 2013);
7) EMBASE (1974 to March 2013);
8) LILACS (2004 to March 2013)
9) Alcohol and Alcohol Problems Science Database (1972 to March 2013);
10) Social Science Citation Index (January 1966 to March 2013).
We developed a detailed search strategy for each database. The search strategy combined the subject search with the Cochrane Highly Sensitive Search Strategy (CHSSS) for identifying randomised trials in PubMed, sensitivity maximising version (2008 revision), as referenced in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).
The subject search utilised a combination of controlled vocabulary and free text terms based on the search strategy for searching PubMed. This search strategy was adapted as appropriate for the other databases (see Appendix 1; Appendix 2; Appendix 3; Appendix 4; Appendix 5; Appendix 6; Appendix 7 for all searches). No language restrictions were applied.
Searching other resources
We contacted relevant authors and searched citations in all relevant papers to obtain information on potential additional randomised controlled trials. In addition, the authors searched for other unpublished studies and assessed relevant conference proceedings for additional references. The following websites were also searched:
-
http://nrepp.samhsa.gov/;
-
http://sbirt.samhsa.gov/core_comp/brief_int.htm;
-
http://motivationalinterview.org;
-
Current Controlled Trials (http://www.controlled‐trials.com/);
-
Clinical.Trials.gov;
-
Trialsjournal.cm.
Data collection and analysis
Selection of studies
All the papers from the electronic searches were assessed by TC and BM to identify potentially eligible studies, and the full texts retrieved. Selection from this initial search was based on information derived from the title, abstract and keywords.
These included randomised controlled trial or clinical trial and substance use, alcohol use, drug use (and related terms), alcohol or drug use or substance use reduction strategies (and related terms), problem behaviours (including but not limited to aggression, fighting, suspension, expulsion, weapon‐carrying), interventions, school staff or settings or both (and related terms).
If the title, abstract and keywords did not provide enough information to make an informed decision with regards to inclusion of the paper, the full text of the paper was obtained.
The full texts of potentially relevant studies were assessed for inclusion by TC and BM. While there was no disagreement on the inclusion of studies, a third review author was on hand (JL) to resolve any disagreements.
Data extraction and management
Data were independently extracted by two independent authors (TC and BM) using a piloted data extraction form based on the Cochrane Collaborative Drugs and Alcohol Review Group's extraction form and subsequently entered in the Cochrane Collaboration software (Review Manager 5.1) for analysis (data extraction form available on request from Carney). Data on the following information were extracted from studies: study design and method, allocation process, participant data, intervention, and outcomes. When there was missing information from the original studies on outcomes or other important information, we contacted the corresponding author via e‐mail in order to request additional data. Certain statistics were not readily available in the articles, and if authors were not able to provide this information to the authors we calculated them from existing data using the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) for guidance.
Assessment of risk of bias in included studies
Two review authors independently assessed potential biases resulting from the trial design. Any discrepancies between the review authors were resolved by discussion. Quality assessment was based on the following aspects of methodology: random sequence generation, allocation concealment, blinding of outcome assessor, completion of outcome data, selective outcome reporting, and additional biases (Higgins 2011). Firstly, a description of the various areas of possible biases was entered into a data extraction form. Secondly, these were assigned a judgement of 'low', 'high' or 'unclear' risk of bias. If the two review authors struggled to make a judgement, we contacted the author of the article in an attempt to gain more information about the particular bias domain, and only if it was still unclear was it assigned a judgement of 'unclear'.
Blinding of personnel and participants was impossible due to the type of intervention delivered, and therefore this item relating to risk of performance bias was deactivated in the risk of bias table (see Appendix 8). Only the blinding of the outcome assessor was assessed in this review. For other risks of bias, the following were examined:
-
appropriateness of the statistical tests used in data analysis;
-
compliance with the intervention(s);
-
validity and reliability of outcome measures.
For a detailed description of the criteria used to assess risk of bias, please see Appendix 8.
Measures of treatment effect
We compared the outcomes of the experimental and control groups at different follow‐up appointments. Originally we planned to categorise the findings into short‐term follow‐up appointments (one to three months), medium‐term follow‐up appointments (four to 11 months) and long‐term follow‐up appointments (12 months and longer). However, this was not possible because there were very few studies that could be included. For studies that reported on more than one follow‐up period, we used the last follow‐up period. Dichotomous outcome measures were assessed by calculating the risk ratio (RR) with the 95% confidence interval (CI), while for continuous outcome measures the standardised mean difference (SMD) with 95% CI was the treatment measure used as the summary statistic. Commonly in meta‐analysis the studies assess the same outcome but measure it in a variety of ways, so the same outcome may be measured with different scales (Higgins 2011). If standard deviations for the mean values were not provided, we used the standard errors that were provided and used the calculation in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) to change them to standard deviations.
The quality of evidence was rated using the GRADE approach (Balshem 2011). The GRADE approach defines the quality of evidence for outcomes as the extent to which one can be confident that the estimate of an effect is close to the quantity of interest. The following criteria were used to rate the quality of evidence for selected outcomes: 1) methodological quality, which considers the risk of bias within studies; 2) directness of evidence, by addressing population, intervention and outcome of interest; 3) inconsistency (significant heterogeneity which cannot be explained); 4) imprecision (if the 95% CI includes appreciable benefit or harm); and 5) publication bias, which involves study sample size and whether or not studies were commercially funded.
Unit of analysis issues
The analysis of clinical trials needs to take into account the level at which randomisation occurred. While this can be on an individual basis, cluster‐randomised trials have groups of individuals (for example schools, community) as opposed to individuals as the unit of analysis. The review authors originally planned to measure the intra‐cluster correlation coefficient (ICC) in these studies and then use the ICC to measure the design effect, which is an inflation factor that is used to increase the statistical power of the study (Campbell 2000). However, the authors of the cluster‐randomised trials used the Huber‐White estimator of variance to control for the effects of clustered recruitment, and further calculations were therefore not necessary. While the review authors had decided to use a conversion rate of 4.29 (30 days/7) where outcomes across studies used different measurement times other than monthly frequency, it was unnecessary to do any additional conversions as the measures in the studies were of monthly use (for example frequency of use, quantity of use).
Dealing with missing data
We contacted the original investigators of the included studies up to three times to request any missing data (missing studies, outcomes, summary data, individuals, and study‐level characteristics). We needed to decide whether the data were missing at random (not related to the actual data) or not missing at random (related to the actual data). When study data were assumed to be missing at random, only the available data were analysed. For data that were not missing at random, this needed to be addressed by performing a sensitivity analysis or, if this was not possible, by replacing missing data with specified values (Higgins 2011). The imputation of missing data with specific replacement values was not needed for the studies included in this review.
Assessment of heterogeneity
We assessed the extent of heterogeneity across the studies using the Chi² (X²) test and I² statistic and looking at whether the P values were statistically significant (Higgins 2011), with a P value of 0.10 or less showing significant heterogeneity.
Assessment of reporting biases
We planned to use funnel plots (plots of the effect estimate from each study against the sample size or effect standard error) in an attempt to assess any publication bias. More specifically, we planned to examine the funnel plots for asymmetry as an indication of publication bias. However, asymmetrical funnel plots are not always caused by publication bias and publication bias does not always cause asymmetrical funnel plots (Higgins 2011). This was not possible for the current review because less than 10 studies were included.
Data synthesis
A meta‐analysis was performed as there were more than two individual trials with comparable intervention methods and outcomes that could be analysed. Random‐effects models were used based on the fact that we expected different types of interventions to be included in the review and combined in the meta‐analysis (such as interventions of different duration and using different follow‐up measures).
Subgroup analysis and investigation of heterogeneity
Although the authors originally planned to conduct subgroup analyses for studies with low and unclear risk of bias and, if possible, for different ages, gender and school grades for adolescent study participants, this was not possible. There was only a small number of studies included in the meta‐analysis, and the results were not reported by these variables of interest.
Sensitivity analysis
The authors decided that if there was significant unexplained heterogeneity and more than 10 studies were included in the analysis, they would perform a sensitivity analysis to consider if the following had an impact on effect size:
1. studies conducted in settings other than traditional high or secondary schools (e.g. alternative high schools, reform school);
2. studies which utilised quasi‐experimental designs (as long as an experimental and a control group were included);
3. studies which had attrition rates of more than 20%.
Since only six studies were included in the review, these sensitivity analyses were unnecessary.
Results
Description of studies
See Characteristics of included studies; Characteristics of excluded studies.
Results of the search
Our search strategy retrieved 1220 potentially relevant references of which 183 were duplicate references. We screened a total of 1037 references (see Figure 1). We assessed the abstracts and where required obtained the full texts of the journal articles. One article needed to be translated in its entirety, from Portuguese to English, because the abstract in English did not provide enough information. At this stage 1009 abstracts were excluded as they were clearly not relevant for the review. This left 28 full‐text articles on randomised controlled trials (RCTs) that were reviewed. One of the articles was obtained ahead of press from the primary author. A further 22 RCTs were excluded. The reasons for exclusion are summarised in the Characteristics of excluded studies table.The six remaining RCTs (reported in seven separate articles) met our inclusion criteria and are described in detail in the Characteristics of included studies table.
Study flow diagram.
Included studies
We identified six studies (reported in seven articles) that were published between 2004 and 2011 for inclusion in this review. A total of 1139 participants were included in these studies at the start of the studies (with an additional 162 who were followed up in McCambridge 2004). The number of participants that remained in the studies at the follow‐up appointments was 1120. All six studies were RCTs, of which two were cluster‐RCTs (McCambridge 2004; McCambridge 2008). All interventions were provided on a face‐to‐face individual basis.
Types of comparison
-
Brief intervention versus information provision, three studies, 732 participants (McCambridge 2008; Walker 2011; Werch 2005)
-
Brief intervention versus assessment only, three studies, 407 participants (McCambridge 2004; Winters 2007b; Winters 2012)
Location
All of the studies were based in educational settings. Three were based in public secondary schools (Werch 2005; Winters 2007b; Winters 2012), while two were based in further education colleges which provided alternative schooling and training for 16 to 18 year old adolescents (McCambridge 2004; McCambridge 2008). These two studies were conducted in the United Kingdom while the remaining three were conducted in the USA (Werch 2005; Winters 2007b; Winters 2012).
Length and description of intervention
The six interventions met the criteria for brief interventions. Participants received some or all of the following: screening, motivational interviewing, information provision and discussion, brochures, and follow‐up appointments. Three of the studies provided participants with a single BI session (McCambridge 2004; McCambridge 2008; Werch 2005) while the other three studies held two intervention sessions with the adolescent participants (Walker 2011; Winters 2007b; Winters 2012).
Screening and outcomes measures
All six of the studies used self‐report measures. Some used established screening and diagnostic tools such as the Global Appraisal of Individual Needs Interview (GAIN‐I) (Walker 2011), Alcohol Use Disorders Identification Test (AUDIT) (McCambridge 2004; McCambridge 2008), Timeline Followback Interview (TLFB) (Winters 2007b; Winters 2012), Severity of Dependence Scale (SDS) (McCambridge 2008) and Substance Use Disorder Manual of the Adolescent Diagnostic Interview (ADI) (Winters 2007b; Winters 2012) to measure substance use. Others used substance use questionnaires, such as the Alcohol Beverage Youth Survey (Werch 2005). A combination of instruments was also used to measure alcohol behaviours. There was consistency regarding the measures of alcohol and cannabis frequency (number of days used) and quantity (number of units used). The Fagerström Test was also used in one study to measure nicotine dependence (McCambridge 2008).
Measures of behavioural outcomes were less clear and seemed to ask about the consequences of the participants' drug use more generally. McCambridge 2008 used a measure that assessed interactional problems, and was adapted from its original use for adolescents who had alcohol problems to include those who used drugs. Walker 2011 used the Marijuana Problem Inventory to measure problem behaviours associated with cannabis use. Two of the other studies (Winters 2007b; Winters 2012) used the Personal Consequences Scale, which measured legal, health, motor vehicle, social and family problems experienced due to substance use.
Length of follow‐up
The trials differed in terms of follow‐up. While some of the trials conducted short‐term follow‐up appointments, such as Winters 2007b at one month and McCambridge 2004 and Walker 2011 at three months, they also conducted longer‐term follow‐ups, such as six month and 12 month follow‐ups respectively. One trial only had one medium‐term follow‐up at four months (Werch 2005), while the remaining four studies also reported six month follow‐ups. For the data analysis, we selected the last follow‐up appointment for analysis.
Secondary population group
Two of the trials reported a secondary population group, which was the parents of the adolescents who used substances (Winters 2007b; Winters 2012). This made up a third experimental group, where both adolescents and parents received the intervention. While these secondary population groups were considered important, they were not compared in the meta‐analysis as the other studies only had one experimental group with adolescents as their population, and there were no other interventions that worked with parents to compare the outcomes with.
Excluded studies
We excluded 22 potentially eligible studies that were obtained and read in full. Three of the studies were excluded because the length of the interventions did not fit the criteria for brief intervention, while another five were prevention studies and not early intervention studies. Nine of the studies were not school‐based; they were either based at college level or in the community. Finally, some studies were excluded for methodological reasons: one study was a pilot and had no control group, in one study there was no randomisation, and one article described a survey and did not contain any information about interventions.
Risk of bias in included studies
Figure 2 provides a summary of the risk of bias assessments for all the studies. A summary of the risk of bias for each study and in each area is provided in Figure 3.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Allocation
Generation of randomisation sequence
Sequence generation was judged as adequate in four of the studies (McCambridge 2008; Walker 2011; Winters 2007b; Winters 2012). The other two studies (McCambridge 2004; Werch 2005) referred to random allocation but this was not clarified and we were not able to contact the authors for further information. The level of bias was therefore found to be unclear.
Concealment of allocation
Concealment of allocation was adequate in two of the studies (McCambridge 2004; McCambridge 2008). It was unclear in two of the studies (Walker 2011; Werch 2005) and once again contacting the authors proved unsuccessful. In the remaining two studies, allocation concealment did not take place (Winters 2007b; Winters 2012). Communication with the authors revealed that the reason was that it was believed that it would negatively affect study participation.
Blinding
This review reports on psychological interventions such as motivational interviewing, where it was not possible to blind the participants or staff who worked on the study to the intervention. Performance bias therefore could not be assessed for these studies. Moreover the risk of performance bias can actually influence the outcomes if they are self reported and not objective. Findings for the risk of bias for blinding of outcome assessment indicated that there were no systematic differences between how outcomes were determined between the groups and the reported and unreported findings in two of the studies (McCambridge 2004; McCambridge 2008). In four studies there was insufficient information to evaluate the risk of bias (Walker 2011; Werch 2005; Winters 2007b; Winters 2012) in terms of blinding, and the authors could not be contacted.
Incomplete outcome data
All six of the studies were reported to have low risk of bias because there were either low rates of attrition or factors associated with attrition were identified and controlled for in both groups in the original analysis.
Selective reporting
Four of the studies were free of selective reporting, and reported on all prespecified outcomes (McCambridge 2008; Walker 2011; Winters 2007b; Winters 2012). One of the trials did not report on all longer‐term outcomes as the findings were no longer significant (McCambridge 2004), and the sixth study did not report all outcomes (Werch 2005).
Other potential sources of bias
No other sources of bias (appropriateness of statistical tests used in data analysis; compliance with the intervention(s); validity and reliability of outcome measures) were identified.
Effects of interventions
See: Summary of findings for the main comparison Brief intervention compared to information provision for substance‐using adolescents; Summary of findings 2 Brief intervention compared to assessment only for substance‐using adolescents
We could not combine the effects across studies for some of the outcomes due to high levels of heterogeneity. While a meta‐analysis of results across one study was not possible (see summary of findings Table for the main comparison; summary of findings Table 2) the effect of the intervention compared to the control group is reported below.
1. Comparison of brief intervention (BI) to information provision
Any substance use: three studies (McCambridge 2008; Walker 2011; Werch 2005) with a total of 732 participants reported on any substance use. A meta‐analysis of the three studies showed no significant difference between the two groups, with a standardised mean difference (SMD) of ‐0.06 (95% CI ‐0.20 to 0.09). There was no significant statistical heterogeneity (I² = 0%, Chi2 = 0.05 , P = 0.97). See Analysis 1.1.
Alcohol frequency: two studies (McCambridge 2008; Werch 2005) with a total of 527 participants reported on alcohol frequency. There was no significant effect as the SMD was ‐0.03 (95% CI ‐0.20 to 0.14). There was no statistical heterogeneity between the two studies (I² = 0%, Chi2 = 0.20 , P = 0.65). See Analysis 1.2.
Alcohol quantity: two studies (McCambridge 2008; Werch 2005) including 527 participants considered this outcome. There was no significant effect (SMD ‐0.09; 95% CI ‐0.26 to 0.08). Heterogeneity between the two studies was not significant (I² = 0%, Chi2 = 0.15, P = 0.69). See Analysis 1.3.
Tobacco frequency: one study (McCambridge 2008) with 179 participants reported on frequency of tobacco use. The SMD was ‐0.12 (95% CI ‐0.34 to 0.10). See Analysis 1.4.
Cannabis quantity:McCambridge 2008 (with a total of 326 participants) found a SMD of ‐0.08 for cannabis quantity (95% CI ‐0.3 to 0.13). See Analysis 1.5.
Cannabis dependence: two studies (McCambridge 2008; Walker 2011) with a total of 531 participants reported on the cannabis dependence outcome. The SMD was ‐0.03, which was not significant (95% CI ‐0.02 to 0.14). Heterogeneity was not significant (I² = 0%, Chi2 = 0.24, P = 0.62). See Analysis 1.6.
Cannabis frequency: two studies (McCambridge 2008; Walker 2011) involving 531 participants measured cannabis frequency and found a SMD of ‐0.02. This reduction in cannabis frequency was not statistically significant (95% CI ‐0.19 to 0.15). Heterogeneity between the two studies was once again not significant (I² = 0%, Chi2 = 0.0, P = 0.94). See Analysis 1.7.
All behavioural outcomes: the information pooled in the meta‐analysis included consequences and problems associated with alcohol or cannabis use, or both, such as drug selling, drug‐related crime and arrests for being intoxicated. Two studies (McCambridge 2008; Walker 2011) with a total of 531 participants at entry measured behavioural outcomes, which were secondary outcomes for this review. The SMD was ‐0.26 (95% CI ‐0.54 to 0.02), which was close to reaching statistical significance in favour of the experimental group, and heterogeneity was not significant (I² = 62%, Chi2 = 2.62, P = 0.11). See Analysis 1.8.
2. Comparison of brief intervention (BI) to assessment only
Any substance use: three studies (McCambridge 2004; Winters 2007b; Winters 2012) with a total of 424 participants at entry measured any substance use outcomes. Given the heterogeneity of the methods of reporting study results, a summary meta‐analysis was not possible. In the McCambridge 2004 study (n = 179) there was a significant SMD between the intervention and control group of ‐0.61 (95% CI ‐0.91 to ‐0.31). Winters 2007b (n = 53) and Winters 2012 (n = 192) found significant SMDs of ‐0.58 (95% C: ‐1.13 to ‐0.03) and ‐0.39 (95% CI ‐0.71 to ‐0.08) respectively. See Analysis 2.1.
Alcohol frequency: the alcohol frequency outcome was measured by three studies (McCambridge 2004; Winters 2007b; Winters 2012) involving 424 participants in total. Given the heterogeneity of the methods of reporting study results a summary meta‐analysis was not possible. In the McCambridge 2004 study the SMD of ‐0.16 was not significant (95% CI ‐0.45 to 0.14). The other two studies had statistically significant SMDs, as the SMD for the Winters 2007b study was ‐1.17 (95% CI ‐1.76 to ‐0.59) and ‐0.82 for the Winters 2012 study (95% CI ‐1.14 to ‐0.50). See Analysis 2.2.
Alcohol quantity: one study (McCambridge 2004) with 179 participants in total measured alcohol quantity. There was not a significant SMD between the group that received the intervention and the group that received an assessment only (SMD ‐0.27; 95% CI ‐0.57 to 0.02). See Analysis 2.3.
Tobacco frequency: only McCambridge 2004, which had a total of 162 participants at 12 month follow‐up, reported on the tobacco frequency outcome for a comparison between BI and the control assessment only. The SMD of ‐0.30 was not significant (95% CI ‐0.61.13 to 0.01). See Analysis 2.4.
Cannabis frequency: three studies (McCambridge 2004; Winters 2007b; Winters 2012) reported on cannabis frequency. A meta‐analysis of the 407 participants showed that the mean difference between the groups was statistically significant with a SMD of ‐0.22 (95% CI ‐0.43 to ‐0.02). There was no heterogeneity between studies (I² = 0%, Chi2 = 0.95, P = 0.75). See Analysis 2.5.
Cannabis quantity:McCambridge 2004 (n = 179) found a large, significant SMD of ‐60.27 for cannabis quantity (95% CI ‐66.59 to ‐53.95). See Analysis 2.6.
Cannabis dependence: one study (Winters 2012) with a total of 189 participants looked at cannabis dependence as an outcome. The SMD of ‐0.26 between the BI and control groups was not significant (95% CI ‐0.57 to 0.06). See Analysis 2.7.
All behavioural outcomes: two studies (Winters 2007b; Winters 2012) with a total of 245 participants across studies measured behavioural outcomes related to delinquent‐type or criminal behaviours, which were secondary outcomes for this review. Given the heterogeneity between the two studies, a summary meta‐analysis was not possible. The SMD was significant for the earlier study (SMD ‐1.16; 95% CI ‐1.74 to ‐0.57) (Winters 2007b) but not for the later study (SMD ‐0.21; 95% CI ‐0.52 to 0.1) (Winters 2012).
McCambridge 2004 reported on these behaviours using dichotomous outcomes. At three month follow‐up, participants in the control group were found to be almost twice as likely to have sold drugs to friends (RR = 0.38; 95% CI 0.23‐0.66). See Analysis 2.8.
Discussion
Summary of main results
The findings of this review indicate that, compared to an assessment‐only control, brief intervention (BI) has a significant effect in terms of 'any substance use', alcohol frequency, cannabis frequency, cannabis quantity and behavioural outcomes. For all these studies, change was in the directions of the experimental condition (for example, quantity and frequency of use were reduced). Unfortunately the statistically significant amount of heterogeneity in these comparisons made pooling results difficult for some outcomes, and the outcomes for tobacco frequency, cannabis quantity and alcohol quantity were from individual studies only. In addition, the quality of the outcomes related to substance use was low, except for the 'any substance use' outcome, which was of moderate quality. The quality of the behavioural outcomes was very low. When comparing BI to the information‐only control, the effects were generally not significant while the quality of evidence was higher overall (moderate for all except two of the outcomes).
We were interested in assessing whether BI is more effective than assessment only or information provision in reducing alcohol and other drug use. As indicated above, BI had a significant effect on some of the outcome variables of interest compared to the assessment‐only control group but these studies were generally of low quality, making conclusions difficult. The results for BI in comparison to information provision indicated that BI did not have a significant effect on any of these outcomes.
BIs did not reduce problem behaviours for the experimental groups in comparison to both control comparisons (that is information provision and assessment only), although the Winters 2007b study indicated that the intervention had a significant effect on these behavioural outcomes. Other findings from the McCambridge 2004 study, although not included in the meta‐analysis, also indicated that BIs led to reductions in drug‐selling to friends.
Overall completeness and applicability of evidence
A small number of studies (n = 6) were included in this review. These studies covered a narrow age range, with the mean age range being from 15.4 to 18 years old, and three of the six studies reported a mean age of 17 to 18 years old (McCambridge 2004; McCambridge 2008; Werch 2005). This makes it somewhat difficult to generalise the results to students who are in early adolescence, and who will be at a different phase of social and cognitive development. Also, the minimum legal drinking age in the United Kingdom (UK) is 18 years (International Center for Alcohol Policies 2010), where some of the studies were conducted (McCambridge 2004; McCambridge 2008); in comparison to the legal age of 21 years in the USA. Alcohol use is more acceptable and therefore may be more common among this age group. It is possible that adolescents of different ages might need interventions that are tailored specifically for their age group, which the NICE guidelines for school‐based interventions for alcohol advise (NICE 2007).
While the secondary outcomes were described as those that measured criminal, delinquent or problem behaviours, it was difficult to disentangle these behaviours from other interactional and social behaviours in the results as many studies used scales with specific psychometric properties (McCambridge 2004; McCambridge 2008; Walker 2011; Winters 2007b; Winters 2012) that made looking at a single item difficult. It was difficult to take out the problem behaviour items specifically. Additional research using rigorous methods measuring an array of these outcomes is needed before more specific generalisations and recommendations about whether BIs for substance use can also curb delinquent behaviours can be made.
Only three published programmes (adapted motivational interviewing, 'Teen Intervene', brief experimental alcohol beverage‐tailored programme) were used or adapted for use in the six studies included in this review (McCambridge 2004; Winters 2007b; Werch 2005). This might limit the generalisability of the findings further because there was not a large variety in the interventions that were delivered. These interventions were developed either in the USA or the UK, which might limit the applicability of the evidence to students in schools in these particular developed country settings. We are uncertain how effectively they could be applied in countries with different cultural and social norms.
Quality of the evidence
We used the GRADE approach to assess the quality of evidence for each outcome. See summary of findings Table for the main comparison; summary of findings Table 2.
The quality of the evidence for the outcomes varied from very low to moderate. For some of the outcomes the evidence was downgraded on account of risk of bias, unexplained significant heterogeneity, and imprecision. Overall, the quality of the evidence seemed to be higher for the comparison BI versus information provision and lower for the comparison BI versus assessment only. For the first set of comparisons, namely BI with information, in two of the three studies we did not have enough information to be certain about the risk of bias in a number of areas, and therefore the quality of the evidence was downgraded to moderate. In addition, for two of the outcomes the imprecision of the results meant that the confidence intervals were wide and crossed the 0.5 mean difference in either direction. The quality of the evidence was therefore further downgraded. For the BI compared with assessment only, there were issues with risk of bias across all of the outcomes. There was also imprecision in five of the outcomes, which led to further downgrading of the evidence quality. For the behavioural outcomes, there was a large amount of unexplained heterogeneity in addition to the issues with risk of bias and imprecision, meaning that the quality of the evidence was downgraded to very low.
Potential biases in the review process
We believe that we have identified all the studies that focused on the effect of BIs on general substance use as a primary outcome, and behavioural outcomes as secondary outcomes, that met our study design and participant inclusion criteria up to March 2013. We used a comprehensive search strategy designed with assistance from the Alcohol and Drug Review Group, and ensured that there was independent assessment for inclusion eligibility, risk of bias and data extraction. We also tried to search for possible unpublished literature but were not very successful. The small possibility does exist that unpublished randomised controlled trials (RCTs) were excluded from the review. We also took into account the fact that although journal articles have strict word or page limits, there is certain information that is vital to have when completing such a review. With one exception, all authors were very responsive and were able to provide additional information when requested. Certain questions about the risk of bias remained unanswered in the studies where we could not contact the authors. We applied strict criteria in the process of grading the evidence and were transparent about the judgements that led to the decisions on how the studies were rated for the various outcomes.
Agreements and disagreements with other studies or reviews
We found a significant effect size for most comparisons between BI and assessment‐only control conditions. Tait 2003's review reported similar treatment sizes, although the studies included in their review were conducted in multiple settings while the studies included in this review were conducted in educational settings only. Similarly, although the Jensen 2011 review looked at motivational interviewing only and included studies that were again conducted in a number of settings, our effect sizes were comparable with their range of results. The meta‐analysis on alcohol use among adolescents by Watchel 2010, although conducted in clinical settings (and not school settings), found that motivational interviewing (one form of brief intervention) was partially successful, with the most encouraging results being those related to harm minimisation (looking at harms associated with drinking).
In addition, the BIs included in the Watchel 2010 review were not particularly effective in reducing secondary behavioural outcomes such as health, legal and social harms. Unfortunately our effect sizes were not directly comparable to the previous studies as ours included standardised mean differences while the other studies included Cohen's d for their effect sizes.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Comparison 1 Brief intervention versus information provision, Outcome 1 Any substance use outcomes.
Comparison 1 Brief intervention versus information provision, Outcome 2 Alcohol Frequency.
Comparison 1 Brief intervention versus information provision, Outcome 3 Alcohol Quantity.
Comparison 1 Brief intervention versus information provision, Outcome 4 Tobacco frequency.
Comparison 1 Brief intervention versus information provision, Outcome 5 Cannabis Quantity.
Comparison 1 Brief intervention versus information provision, Outcome 6 Cannabis Dependence.
Comparison 1 Brief intervention versus information provision, Outcome 7 Cannabis frequency.
Comparison 1 Brief intervention versus information provision, Outcome 8 All behavioural outcomes.
Comparison 2 Brief intervention versus assessment only, Outcome 1 Any substance use outcomes.
Comparison 2 Brief intervention versus assessment only, Outcome 2 Alcohol Frequency.
Comparison 2 Brief intervention versus assessment only, Outcome 3 Alcohol Quantity.
Comparison 2 Brief intervention versus assessment only, Outcome 4 Tobacco frequency.
Comparison 2 Brief intervention versus assessment only, Outcome 5 Cannabis frequency.
Comparison 2 Brief intervention versus assessment only, Outcome 6 Cannabis Quantity.
Comparison 2 Brief intervention versus assessment only, Outcome 7 Cannabis Dependence.
Comparison 2 Brief intervention versus assessment only, Outcome 8 All behavioural outcomes.
| Brief intervention compared to information provision for substance‐using adolescents | ||||||
| Patient or population: Substance‐using adolescents | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Estimate effect | No of Participants | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| Information provision | Brief intervention | |||||
| General substance use outcomes | The standardised mean substance use outcomes in the intervention groups was | SMD ‐0.06 (‐0.20 t0 0.09) | 732 | ⊕⊕⊕⊝ | Included both alcohol and drug use outcomes | |
| Alcohol frequency | The standardised mean alcohol frequency in the intervention groups was 0.03 standard deviations lower | SMD ‐0.03 (‐0.20 t0 0.14) | 527 | ⊕⊕⊕⊝ | Number of days of alcohol use | |
| Alcohol quantity | The standardised mean alcohol quantity in the intervention groups was 0.09 standard deviations lower | SMD ‐0.09 (‐0.26 to 0.08) | 527 | ⊕⊕⊕⊝ | Number of standard alcohol units | |
| Tobacco frequency | See comment | See comment | Not estimable | 326 | ⊕⊕⊕⊝ | Number of days cigarettes used |
| Cannabis quantity | See comment | See comment | Not estimable | 326 | ⊕⊕⊕⊝ | Number of joints (cannabis cigarettes) |
| Cannabis frequency | The mean cannabis frequency in the intervention groups was | SMD ‐0.02 (‐0.19 to 0.15) | 531 | ⊕⊕⊕⊝ | Number of days cannabis use | |
| All behavioural outcomes | The mean all behavioural outcomes in the intervention groups was | SMD ‐0.26 (‐0.54 to 0.02) | 531 | ⊕⊕⊝⊝ | Interactional Problems Score | |
| *The basis for the assumed risk (e.g. the mean control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the estimate effect of the intervention (and its 95% CI). The estimate effects for certain outcomes were not estimable due to only one study assessing the specific outcome, or extremely high levels of heterogeneity making effects across studies difficult to compare. | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Risk of bias (‐1): In two of the three studies, there was uncertainty about the risk of bias 2 Imprecision (‐1): The confidence intervals contained the null value of zero and the upper or lower confidence limit crosses an effect size of 0.5 in either direction | ||||||
| Brief intervention compared to assessment only for substance‐using adolescents | ||||||
| Patient or population: Substance‐using adolescents | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Estimate effect | No of Participants | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| Assessment only | Brief intervention | |||||
| General substance use outcomes | See comment | See comment | Not estimable | 407 | ⊕⊕⊕⊝ | Included both alcohol and drug use outcomes |
| Alcohol frequency | See comment | See comment | Not estimable | 407 | ⊕⊕⊝⊝ | Number of days of alcohol use |
| Alcohol quantity | See comment | Not estimable | 179 | ⊕⊕⊝⊝ | Number of standard alcohol units | |
| Tobacco frequency | See comment | SMD ‐0.03 (‐0.61 to 0.01) | 162 | ⊕⊕⊝⊝ | Number of days cigarettes used | |
| Cannabis frequency | The mean cannabis frequency in the intervention groups was | SMD ‐0.22 (‐0.43 to 0.02) | 407 | ⊕⊕⊝⊝ | Number of joints (cannabis cigarettes) | |
| Cannabis quantity | See comment | Not estimable | 179 | ⊕⊕⊝⊝ | Number of days cannabis use | |
| All behavioural outcomes | See comment | See comment | Not estimable | 245 | ⊕⊝⊝⊝ | Interactional Problems Score |
| *The basis for the assumed risk (e.g. the mean control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the estimate effect of the intervention (and its 95% CI). The estimate effects for certain outcomes were not estimable due to only one study assessing the specific outcome, or extremely high levels of heterogeneity making effects across studies difficult to compare. | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Risk of bias (‐1): There was no allocation concealment in the study | ||||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Any substance use outcomes Show forest plot | 3 | 732 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.06 [‐0.20, 0.09] |
| 2 Alcohol Frequency Show forest plot | 2 | 527 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.03 [‐0.20, 0.14] |
| 3 Alcohol Quantity Show forest plot | 2 | 527 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.09 [‐0.26, 0.08] |
| 4 Tobacco frequency Show forest plot | 1 | Std. Mean Difference (IV, Random, 95% CI) | Totals not selected | |
| 5 Cannabis Quantity Show forest plot | 1 | Std. Mean Difference (IV, Random, 95% CI) | Totals not selected | |
| 6 Cannabis Dependence Show forest plot | 2 | 531 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.03 [‐0.20, 0.14] |
| 7 Cannabis frequency Show forest plot | 2 | 531 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.02 [‐0.19, 0.15] |
| 8 All behavioural outcomes Show forest plot | 2 | 531 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.26 [‐0.54, 0.02] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Any substance use outcomes Show forest plot | 3 | Std. Mean Difference (IV, Random, 95% CI) | Totals not selected | |
| 2 Alcohol Frequency Show forest plot | 3 | Std. Mean Difference (IV, Random, 95% CI) | Totals not selected | |
| 3 Alcohol Quantity Show forest plot | 1 | Std. Mean Difference (IV, Random, 95% CI) | Totals not selected | |
| 4 Tobacco frequency Show forest plot | 1 | Std. Mean Difference (IV, Random, 95% CI) | Totals not selected | |
| 5 Cannabis frequency Show forest plot | 3 | 407 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.22 [‐0.43, ‐0.02] |
| 6 Cannabis Quantity Show forest plot | 1 | Std. Mean Difference (IV, Random, 95% CI) | Totals not selected | |
| 7 Cannabis Dependence Show forest plot | 1 | Std. Mean Difference (IV, Random, 95% CI) | Totals not selected | |
| 8 All behavioural outcomes Show forest plot | 2 | Std. Mean Difference (IV, Random, 95% CI) | Totals not selected | |
