This review is the first of its kind to examine the impact of early interventions on adolescent substance use and behavioural outcomes. To the best of our knowledge, there are no other reviews that measure the effect that brief interventions have on substance use and behavioural outcomes. This review therefore adds valuable evidence for interventions that address two risk behaviours simultaneously. This is important for policy makers for a number of reasons. First, all the interventions included in the meta-analysis reported significant reductions in drug and alcohol use, providing strong evidence in support of the effectiveness of early interventions for adolescent substance use. This indicates the effectiveness of treating adolescents early as their substance use progresses, that is, before they need specialised treatment  or face unwanted consequences such as incarceration.
Second, the majority of the early interventions included in this systematic review were brief in nature. This suggests that interventions do not need to be lengthy to be effective with adolescents. This is important for policy makers as there are cost implications associated with lengthier interventions. Brief interventions are among some of the most cost-effective types of behavioural treatments , and may be especially useful in low- and middle-income countries where health care systems are burdened and have few resources available for intensive behavioural interventions. To ensure that adolescents are reached early, they should be targeted while still at school. Furthermore, screening for identification of appropriate adolescents should be conducted in a number of places. This could possibly further take some of the burden off more specialised treatment providers, and provide services to adolescents who do not have access to the health care system .
Third, this review provides evidence for the format best suited to delivering these interventions to at-risk adolescents. When subgroup analyses were conducted to compare the effectiveness of interventions delivered in individual versus group format and those delivered over single versus multiple intervention sessions, a clear pattern of findings emerged which favoured the delivery of interventions in an individual-format and over multiple sessions. For example, interventions delivered in an individual format across multiple sessions had a much greater effect on the frequency of alcohol use and the frequency of cannabis use than single-session interventions or those delivered in a group-format.
Similarly, the results of the overall meta-analysis found that the interventions had a significant effect on problem and criminal behaviours related to substance use, and subgroup analyses found that interventions delivered in an individual-format and over multiple sessions had significant, albeit relatively small, effects on behavioural outcomes. In contrast, interventions delivered in group formats or in single sessions did not have the desired effect on behavioural outcomes . This is in keeping with findings from a previous study  that found group interventions to be harmful for high-risk adolescents, for both short- and longer-term outcomes such as convictions for criminal activities and psychiatric problems. Further, a single intervention session may not be long enough to change relatively entrenched behaviours. While these findings of the subgroup analysis suggest that effective early interventions for addressing adolescent substance use and related problem behaviours should be in an individual-format and delivered over multiple sessions, this needs to be further researched as the effect sizes (particularly for behavioural outcomes) remain small and few studies have directly examined behavioural outcomes for substance use interventions delivered in group formats. In low- and middle-income countries especially, where resources are often limited, group-format interventions are cost effective and providing a one-on-one intervention in schools may not always be feasible or affordable. Furthermore, there was only one group-format intervention that met the inclusion criteria for this review, which cannot be generalised to all types of group interventions. Clearly this is an area that warrants more research in low- and middle-income countries before policy recommendations can be made.
It should be noted that the small effect sizes obtained could be due to several factors, including measurement issues. For instance, some studies used measures that included multiple consequences of substance use that were not necessarily related to the behavioural outcomes of interest, and these could not be separated out in the meta-analysis (for example, getting in to trouble with family ). Also, the time frames of the follow-up measurements may have influenced the results. Brief interventions can have immediate effects on substance use outcomes but other behavioural outcomes are often more distal outcomes of these interventions. It may be necessary to conduct research that includes longer follow up periods to accurately assess whether the intervention has an impact on other behavioural outcomes.
Fourth, the review is useful as it provides clear guidelines regarding the strength of interventions and also points to interventions that seem more likely to be effective than others. This will be of value to policy makers who are looking for a strong evidence base to guide the adoption of one intervention vis-à-vis others. Furthermore, some of the studies clearly seemed more effective than others, for example the outcome effect size for Winter et al’s Teen Intervene were generally consistently larger than that of the other studies. Teen Intervene has been registered with the U.S. National Registry of Evidence-based Programs and Practices since 2007 . Of the nine interventions included in this systematic review, it is the only intervention that included a session with the adolescents’ parents. Previous research has indicated that the involvement of key family members in interventions with substance-using adolescents who also face other problems (such as truancy and involvement in justice-diversion programs) is useful and may lead to sustained outcomes . This review clearly suggests that Teen Intervene is a powerful intervention for adolescents who use substances. However, it may need to be expanded to include a stronger focus on delinquent-type behaviours (and not just legal consequences associated with substance use).
Findings from this study should be considered in the light of some limitations. First, there was substantial heterogeneity between the studies. In this systematic review, the length of follow up appointments and study quality, especially in defining and reporting outcomes, varied substantially across studies. This highlights the need for standardised outcome measures for substance use and other behavioural outcomes, as well as guidelines for the choice of follow up periods for intervention studies. Another weakness was that only two of the studies reported an intention to treat analysis, and while one of the remaining studies had an extremely low loss to follow-up rate , the remaining results could have been somewhat biased if participant follow-up was related to their response to treatment.
A final limitation is that all the included studies are from developed country settings (primarily the USA or UK) and therefore it is unclear whether these findings can be generalised to low- and middle-income countries. Although there is some promising intervention work on substance use and other problem behaviours from developing regions, none of these studies met the inclusion criteria for this review as they were mainly descriptive, non-experimental studies . This highlights the need for more intervention research in developing regions that address the interlinked risks of substance use and problem behaviours. A recommendation for further work is, therefore, to test the cultural applicability of the recommended intervention from this review (Teen Intervene) to a developing country setting such as South Africa.
Third, as we did not expect to find a large number of intervention studies that addressed substance use and behavioural outcomes, the inclusion and exclusion criteria for this review were not as rigorous as they could have been as studies did not necessarily have to be randomized controlled trials. Future systematic reviews that are not as exploratory in nature should aim to include only randomized controlled trials in so that stronger evidence of intervention effectiveness can be provided.
Finally, results were difficult to compare across studies for two reasons: (a) few studies were found that examined behavioural outcomes directly and; (b) the five studies that included these outcomes used different measurement tools. This clearly highlights the need for future studies that provide integrated interventions that directly target substance use as well as substance-related behavioural outcomes within the context of the intervention.