Interventions for preventing multiple risk behaviours in young people

Background

Health risk behaviours, such as smoking and drug use, can group together during the teenage years, and engagement in these multiple risk behaviours can lead to health problems such as injury and substance abuse during childhood and adolescence, as well as non-communicable diseases later in life. Currently, we do not know which interventions are effective in preventing or decreasing these risky behaviours among children and young people.

Search methods and selection of studies

We carried out thorough searches of multiple scientific databases to identify studies that looked at ways of preventing or decreasing engagement in two or more risk behaviours, including tobacco use, alcohol use, illicit drug use, gambling, self-harm, sexual risk behaviour, antisocial behaviour, vehicle-risk behaviour, physical inactivity, and poor nutrition, among young people aged eight to 25 years. We divided these studies into groups (individual-level, family-level, and school-level studies) according to whether researchers worked with individuals, families, or children and young people in schools, respectively. We specifically looked at "gold standard" studies - randomised controlled trials that aimed to examine two or more behaviours of interest.

Main results

In total, 70 studies were eligible for inclusion in this review. Half included populations without any consideration for risk status, and half focused on higher-risk groups. Most were conducted in the USA or in high-income countries. On average, studies examined the effects of interventions on four behaviours, most commonly alcohol, tobacco use, drug use, and antisocial behaviour.

We found that for multiple risk behaviours, school-based studies for all young people are more beneficial than a comparator for preventing tobacco use, alcohol use, and physical inactivity, and that they may also be beneficial in relation to illicit drug use and antisocial behaviour. Findings were weaker for cannabis use, sexual risk behaviour, and unhealthy diet. Evidence suggests that certain school-based programmes could have a beneficial impact on more than one behaviour. In contrast, we did not find strong evidence of beneficial effects of interventions for families or individuals for the behaviours of interest, although caution must be applied in interpreting these findings because we identified fewer of these studies. Last, we found seven studies that reported increased levels of engagement in risk behaviours among those receiving the intervention compared to those given the control.

Overall, reviewers judged the quality of the evidence to be moderate or low for most behaviours examined using standardised criteria, with one behaviour found to have very low quality evidence. In part, this was due to concerns around how some studies were conducted, which could have introduced bias.

Conclusions

Our findings suggest that school-based interventions offered to all children that aim to address engagement in multiple risk behaviours may have a role to play in preventing tobacco use, alcohol use, illicit drug use, and antisocial behaviour, as well as in improving physical activity, among young people, but not in the other behaviours examined. We did not find strong evidence of benefit of interventions for families or individuals. Concerns around reporting of studies and study quality highlight the need for additional robust, high-quality studies to further strengthen the evidence base in this field.

Authors' conclusions: 

Available evidence is strongest for universal school-based interventions that target multiple- risk behaviours, demonstrating that they may be effective in preventing engagement in tobacco use, alcohol use, illicit drug use, and antisocial behaviour, and in improving physical activity among young people, but not in preventing other risk behaviours. Results of this review do not provide strong evidence of benefit for family- or individual-level interventions across the risk behaviours studied. However, poor reporting and concerns around the quality of evidence highlight the need for high-quality multiple- risk behaviour intervention studies to further strengthen the evidence base in this field.

Read the full abstract...
Background: 

Engagement in multiple risk behaviours can have adverse consequences for health during childhood, during adolescence, and later in life, yet little is known about the impact of different types of interventions that target multiple risk behaviours in children and young people, or the differential impact of universal versus targeted approaches. Findings from systematic reviews have been mixed, and effects of these interventions have not been quantitatively estimated.

Objectives: 

To examine the effects of interventions implemented up to 18 years of age for the primary or secondary prevention of multiple risk behaviours among young people.

Search strategy: 

We searched 11 databases (Australian Education Index; British Education Index; Campbell Library; Cumulative Index to Nursing and Allied Health Literature (CINAHL); Cochrane Central Register of Controlled Trials (CENTRAL), in the Cochrane Library; Embase; Education Resource Information Center (ERIC); International Bibliography of the Social Sciences; MEDLINE; PsycINFO; and Sociological Abstracts) on three occasions (2012, 2015, and 14 November 2016)). We conducted handsearches of reference lists, contacted experts in the field, conducted citation searches, and searched websites of relevant organisations.

Selection criteria: 

We included randomised controlled trials (RCTs), including cluster RCTs, which aimed to address at least two risk behaviours. Participants were children and young people up to 18 years of age and/or parents, guardians, or carers, as long as the intervention aimed to address involvement in multiple risk behaviours among children and young people up to 18 years of age. However, studies could include outcome data on children > 18 years of age at the time of follow-up. Specifically,we included studies with outcomes collected from those eight to 25 years of age. Further, we included only studies with a combined intervention and follow-up period of six months or longer. We excluded interventions aimed at individuals with clinically diagnosed disorders along with clinical interventions. We categorised interventions according to whether they were conducted at the individual level; the family level; or the school level.

Data collection and analysis: 

We identified a total of 34,680 titles, screened 27,691 articles and assessed 424 full-text articles for eligibility. Two or more review authors independently assessed studies for inclusion in the review, extracted data, and assessed risk of bias.

We pooled data in meta-analyses using a random-effects (DerSimonian and Laird) model in RevMan 5.3. For each outcome, we included subgroups related to study type (individual, family, or school level, and universal or targeted approach) and examined effectiveness at up to 12 months' follow-up and over the longer term (> 12 months). We assessed the quality and certainty of evidence using the Grades of Recommendation, Assessment, Development and Evaluation (GRADE) approach.

Main results: 

We included in the review a total of 70 eligible studies, of which a substantial proportion were universal school-based studies (n = 28; 40%). Most studies were conducted in the USA (n = 55; 79%). On average, studies aimed to prevent four of the primary behaviours. Behaviours that were most frequently addressed included alcohol use (n = 55), drug use (n = 53), and/or antisocial behaviour (n = 53), followed by tobacco use (n = 42). No studies aimed to prevent self-harm or gambling alongside other behaviours.

Evidence suggests that for multiple risk behaviours, universal school-based interventions were beneficial in relation to tobacco use (odds ratio (OR) 0.77, 95% confidence interval (CI) 0.60 to 0.97; n = 9 studies; 15,354 participants) and alcohol use (OR 0.72, 95% CI 0.56 to 0.92; n = 8 studies; 8751 participants; both moderate-quality evidence) compared to a comparator, and that such interventions may be effective in preventing illicit drug use (OR 0.74, 95% CI 0.55 to 1.00; n = 5 studies; 11,058 participants; low-quality evidence) and engagement in any antisocial behaviour (OR 0.81, 95% CI 0.66 to 0.98; n = 13 studies; 20,756 participants; very low-quality evidence) at up to 12 months' follow-up, although there was evidence of moderate to substantial heterogeneity (I² = 49% to 69%). Moderate-quality evidence also showed that multiple risk behaviour universal school-based interventions improved the odds of physical activity (OR 1.32, 95% CI 1.16 to 1.50; I² = 0%; n = 4 studies; 6441 participants). We considered observed effects to be of public health importance when applied at the population level. Evidence was less certain for the effects of such multiple risk behaviour interventions for cannabis use (OR 0.79, 95% CI 0.62 to 1.01; P = 0.06; n = 5 studies; 4140 participants; I² = 0%; moderate-quality evidence), sexual risk behaviours (OR 0.83, 95% CI 0.61 to 1.12; P = 0.22; n = 6 studies; 12,633 participants; I² = 77%; low-quality evidence), and unhealthy diet (OR 0.82, 95% CI 0.64 to 1.06; P = 0.13; n = 3 studies; 6441 participants; I² = 49%; moderate-quality evidence). It is important to note that some evidence supported the positive effects of universal school-level interventions on three or more risk behaviours.

For most outcomes of individual- and family-level targeted and universal interventions, moderate- or low-quality evidence suggests little or no effect, although caution is warranted in interpretation because few of these studies were available for comparison (n ≤ 4 studies for each outcome).

Seven studies reported adverse effects, which involved evidence suggestive of increased involvement in a risk behaviour among participants receiving the intervention compared to participants given control interventions.

We judged the quality of evidence to be moderate or low for most outcomes, primarily owing to concerns around selection, performance, and detection bias and heterogeneity between studies.