A UK study successfully harnessed respected peers to prevent smoking, but not through classroom activities. Instead the 12-13-year-olds simply exerted their influence in normal social interactions with same-age school mates.
FINDINGS 223 Welsh and English secondary schools were approached in 2001. About half were prepared to consider participating in the featured study.1 Of these, 59 were randomly sampled and agreed Seven of the randomly selected schools did not agree. to participate. They were randomly selected to continue with their normal smoking education (the control schools) or additionally to host the ASSIST intervention.
To prepare for this, participating year 8 pupils (12-13 years old) nominated year-mates they respected or who were good leaders. In ASSIST schools only, the research project's trainers asked roughly the top 1 in 6 most nominated pupils to act as 'peer supporters', excluding only smokers who were unwilling to try to stop, or those about whom the school had serious concerns.2 Few were excluded The study approached the top 17.5% of pupils and recruited 16%. for these reasons. Nearly all the chosen pupils completed two days of out-of-school training with health promotion staff and youth workers.
Through interactive learning methods such as role play and discussion, the training aimed to equip the 835 peer supporters with the confidence, social skills and knowledge to talk informally with their peers about the effects of smoking and the benefits of not smoking. They were asked to do this for 10 weeks and to keep a log of their conversations, supported by four follow-up visits from the health promotion trainers.
In each set of schools well over 5000 pupils completed questionnaires about their smoking before and after the intervention and then about one and two years later. At first few (6%) admitted smoking in the past week, growing to 20% by the final follow-up. After adjusting among other things Differences between the sets of schools in size, private v. state, free school meals percentage, single or mixed gender, and language. for starting levels of smoking, at the one- and two-year points ASSIST schools had about a fifth fewer pupils Statistically significant at the first point, narrowly missing this at the second. who had smoked in the past week. At two years the unadjusted percentages of past-week smokers were about 19% in ASSIST schools and 22% in control schools.
Apart from (at least at one-year) a minimal impact among the few regular weekly smokers,2 the intervention was about as effective with pupils who had and had not smoked, with boys and girls, and regardless of whether the pupil had been a peer supporter. However, its impact faded slightly over the follow-up period, and was much greater in schools serving the small communities of the south Wales valleys than in towns and cities, where it failed to achieve a statistically significant impact.
IN CONTEXT The innovation in the featured study was to harness peer influence outside the classroom. Pupils could exercise judgement over who to approach, how and when, rather than teachers mandating who they interacted with and the programme they used. As might be expected of these socially advanced youngsters, they appear to have used this discretion to good effect. Feedback suggested they found it easier and thought it would be more profitable to talk to friends and to pupils who were not already regular smokers.2 Their activity levels tailed off over the 10 weeks, partly because they risked repeating the message to the same circle of contacts. For this reason, continuous intervention for more than a few weeks seems unrealistic, but, as the authors comment, impacts might have been greater had the intervention been implemented each year in each school grade.
Inevitably there are questions about whether the approach would work as well elsewhere. First, schools self-selected in to the study and were generally prepared to go along with their pupils' choices of peer leaders (thought important to their credibility) and to release these pupils for training, even when the teachers had misgivings about the choices made.3 On the other hand, school staff were not directly involved in delivering the intervention.
Secondly, impacts may be dependent on the socialisation patterns of the youngsters. The greater impact in the Welsh valleys was not due to their relative deprivation. As reflected in the proportion receiving free school meals, this was not a significant predictor of intervention impact. Instead it seems likely that in these circumscribed, stable communities, pupils were in contact not just at school but on the way there and back and during leisure time. Influential pupils at school may also have been influential out of school where smoking occurs.
The intervention benefited from external trainers and external venues. This and the nomination procedure makes it unlikely that peer leaders will be seen as having been co-opted by the school, an image unlikely to appeal to the youngsters most likely to flout adult convention by smoking. However, external help comes at a cost. Overall the intervention cost £4700 per school, a figure which would be slightly lower if local trainers were available.
It is not unusual for preventive interventions to curb smoking but not (or not as much) other forms of substance use,4 making it unsafe to assume that similar methods would work in respect of drinking and illegal drug use.
Though rarely resorted to, it seems wise to have retained a veto over peer leaders who were committed to smoking or over whom the school had serious concerns. Socially advanced teenagers likely to be chosen by their peers have also tended to be advanced in their experimentation with substance use and can have an influence opposite to that intended.
Other studies which have relied on peer-selected leaders have instead paired them with the pupils who nominated them (the so-called 'network' method) in small-group, classroom-based substance misuse prevention activities devised and overseen by adult educationalists. Typically these US studies favoured the network option over whole class teaching or randomly selected groups, but sometimes not to a statistically significant degree, and with important exceptions. See background notes and earlier Findings web and PDF Nuggets for detailed analyses.
One exception was when the minority of pupils who used substances most frequently choose like-minded friends and leaders.5 In another study the network method was preferable only when paired with a curriculum devised to reflect the cultural heritages of the largely Hispanic pupils.6 Similar variations may partly lie behind the findings of a meta-analysis which combined relevant studies.7 This concluded that peer-led school-based substance use prevention had a slightly better record than adult-led programmes, but also that the advantage is inconsistent, apparent in some circumstances but not in others.
PRACTICE IMPLICATIONS Given the potential for counterproductive impacts, if the featured study's method is tried some monitoring of outcomes would be advisable. It does have the great advantage of not occupying classroom teaching time, always in short supply. This means such initiatives can (as in the study) supplement rather than displace classroom drug education and leave this (as some teachers argue should be the case) to focus on education rather than prevention, but at the cost of funding external staff and venues. Such work could however be incorporated within existing youth work projects, reducing the costs, and usefully placing these projects in contact with the most influential youngsters in their areas. If the featured study's methodology is followed, unless this can be arranged for school holidays or weekends, schools will have to be prepared to release about 15% of their pupils for the training (thought to be the 'critical mass') and to trust the pupils as a whole to take the lead in selecting who is trained.
1 FEATURED STUDY Campbell R. et al. An informal school-based peer-led intervention for smoking prevention in adolescence (ASSIST): a cluster randomised trial. Lancet: 2008, 371, p. 1595–1602.
2 Audrey S. et al. It's good to talk: adolescent perspectives of an informal, peer-led intervention to reduce smoking. Social Science & Medicine: 2006, 63(2), p. 320–334.
3 Audrey S. et al. Commitment and compatibility: Teachers' perspectives on the implementation of an effective school-based, peer-led smoking intervention. Health Education Journal: 2008, 67(2), p. 74–90.
5 Valente T.W. et al. Peer acceleration: effects of a social network tailored substance abuse prevention program among high-risk adolescents. Addiction: 2007, 102, p. 1804–1815.
6 Valente T.W. et al. The interaction of curriculum type and implementation method on 1-year smoking outcomes in a school-based prevention program. Health Education Research: 2006, 21(3), p. 315–324.
7 Cuijpers P. Peer-led and adult-led school drug prevention: a meta-analytic comparison. Journal of Drug Education: 2002, 32(2), p. 107–119.
Last revised 07 August 2008
STUDY 2006 Choose peer education groups carefully
REVIEW 2015 Prevention of addictive behaviours
STUDY 2000 Education's uncertain saviour
NASTY SURPRISES 2004 Confident kids ... like to party