Appendix C: The evidence

This appendix lists evidence statements from two reviews provided by external contractors (see appendix A) and links them to the relevant recommendations (see appendix B for the key to quality assessments). The evidence statements are presented here without references – these can be found in the full review (see appendix E for details). It also sets out a brief summary of findings from the economic appraisal.

Evidence statement ES8 indicates that the linked statement is numbered ES8 in the review 'School-based interventions to prevent the uptake of smoking among children and young people: effectiveness review'. Evidence statement QR3 indicates that the linked statement is numbered QR3 in the review 'Facilitators and barriers to the delivery of school-based interventions to prevent the uptake of smoking among children: a systematic review of qualitative research'.

The reviews and economic appraisal are available on the NICE website. Where a recommendation is not directly taken from the evidence statements, but is inferred from the evidence, this is indicated by IDE (inference derived from the evidence) below.

Recommendation 1: evidence statements QR1, QR5

Recommendation 2: evidence statements ES1, ES7, ES13, ES19, ES21, ES26, QR4, QR6

Recommendation 3: evidence statements ES11, QR3; IDE

Recommendation 4: evidence statement QR4

Recommendation 5: IDE

Evidence statements

Evidence statement ES1

There is evidence from 27 studies that provided usable data for meta-analysis that interventions may be effective. Meta-analysis of 27 randomised controlled trials (RCTs) demonstrated a significant intervention effect for school-based intervention for preventing uptake of smoking among children. There was moderate statistical heterogeneity between the trial results.

Applicability: most of the studies took place outside of the UK It is not clear if these findings are directly applicable to the UK.

Evidence statement ES7

There is moderate evidence indicating that multi-component interventions incorporating both school and community components (with or without an additional family component) are ineffective in preventing the uptake of smoking compared to usual education. Five RCTs provided evidence comparing a multi‐component intervention that incorporates both school and community components to usual education (three [+] USA), one [-] Australia, one [-] UK). Four of the studies (two [+] USA, one [-] Australia, one [-] UK) found no significant difference between the multi-component intervention group and the usual education group during a maximum follow-up between 6 months (one study [‐] UK).and 5 years (one study [+] USA). One study ([+] USA) found no difference at 3-year follow-up and small, marginally significant positive or negative intervention effects (depending on the school component) at 4-year follow-up.

Evidence statement ES11

It is not clear whether effectiveness of school-based smoking prevention programmes depend on the status of the person delivering it. There is conflicting evidence whether peer-led programmes produced most effective intervention effects on smoking initiation. It is important to note that a peer-led programme may be differentially effective based on how leaders are selected and how groups are formed, and may be curriculum-dependent. There is some evidence that teacher-led, health educator-led, and peer-led programmes tend to be equally effective.

Seven RCTs examined whether effectiveness of school-based smoking prevention programmes depend on the status of the person delivering it.

Three other studies provided evidence that peer-led interventions tend to enhance smoking prevention programmes. For example, results from one (+) USA RCT showed a marked suppression in the onset of both experimental and regular smoking among those students exposed to the resistance training with peer involvement. Similarly, one (-) USA RCT found that a cognitive-behavioural approach when carried out by peer-leaders and when additional boosters are provided can reduce tobacco use. Yet one (+) USA RCT provided evidence that a peer-led programme will be differentially effective based on how leaders are selected and how groups are formed, and this effect may be curriculum dependent.

In one RCT ([-] USA), there was no statistically significant difference in regular smoking rates among students taught by health educators and those taught by adult teachers assisted by older teens. One (++) UK RCT found that the effect of ASSIST intervention was much the same for peer supporters and non-peer supporters. Similarly, one (-) Australia RCT confirmed non-superiority of peer-led programmes to teacher-led programmes. However, this result was gender-specific.

Both the teacher-led and peer-led programmes reduced, to about the same degree, the uptake of smoking by girls while only the teacher-led programme appeared to be effective in boys. One (++) Canada RCT provided evidence that teachers and nurses were equally effective providers regardless of delivery method. While, one (-) USA RCT reported that students exposed to interactive health educator-led interventions were less likely to use tobacco compared to those not exposed to health educator-led instruction.

Applicability: most of the studies were conducted in the USA. It is not clear if these findings are directly applicable to the UK since the interventions under investigation are specific to USA. Furthermore, demographics of the participants are different from those in the UK. Only one (++) UK study is likely to be directly applicable.

ASSIST intervention model

A Stop Smoking In Schools Trial (ASSIST) was a randomised controlled trial involving the use of peer supporters to encourage year eight students (aged 12 and 13) not to smoke. Initially, all year eight students were asked to fill in a questionnaire to nominate the peer supporters. They were asked questions such as, 'who do you respect in year eight in your school?'; and 'who are good leaders in sports or other group activities in year eight in your school?' Those nominated took part in a 2-day external training session led by trainers experienced in youth work and a public health specialist. Following the training, the peer supporters spent 10 weeks having informal conversations about smoking with their class mates. These took place during travel to and from school, break and lunchtime and after school. The peer supporters logged the conversations in a simple diary. The trainers had four follow-up meetings with them to help solve any problems and monitor the diaries.

Evidence statement ES13

There is clear evidence that the addition of booster sessions enhanced effectiveness of main programmes.

Four studies (one [++] and three [-]) analysed effectiveness of booster sessions. Evidence from one (++) USA study suggests that addition of booster sessions significantly enhanced the effectiveness of the main programme and was more effective than the delayed programme controls. One (-) USA study found that boosters can be an effective tool for maintaining or increasing the effectiveness of smoking prevention programmes. One (-) USA study revealed that addition of booster sessions to cognitive-behavioural approach can reduce tobacco use. Another (-) USA study showed that continued intervention students reported significantly less smoking than lapsed intervention and continued control students.

Applicability: all four studies were conducted in the USA. It is not clear if the findings are directly relevant to the UK.

Evidence statement ES19

There is conflicting evidence of differential effect of intervention according to the sex of the target audience. There is moderate evidence that sex is an important predictor of post-test smoking, but direction of effect (either in male or female student) is inconclusive. Furthermore, association of sex with smoking prevalence depends on how the outcome was measured. One recent study ([+] UK) found no significant difference in effectiveness of school-based intervention among male and female students.

Another study ([++] USA) provided no evidence of Hutchinson Smoking Prevention Project impact on the prevalence of daily smoking, either for girls or for boys. Three studies (one [++] Canada; one [+] Canada and one [-] USA) demonstrated that the intervention was more effective among male students; while only one study ([-] Australia) found that both teacher-led and peer-led programmes reduced the taking up of smoking by girls to about the same degree.

There was also conflicting evidence from nine studies whether sex was an important predictor of post-test smoking. Only one study ([-] The Netherlands) provided evidence that sex was not associated with post-test smoking. Two studies (one [+] USA and one [-] USA) found that female students were more likely than male students to have reported smoking at follow-up and only one study ([-] Australia) found that boys were less likely than girls to have reported smoking at follow-up. Yet, three studies (two [-] USA; and one [-] The Netherlands) revealed that males were more likely to be a smoker than their female counterparts. Another two studies (one [+] USA and one [+] Italy), demonstrated that compared to male students, female students were less likely to have used tobacco.

Applicability: most of the studies were conducted in the USA. It is not clear if these findings are directly applicable to the UK since the interventions under investigation are specific to the USA. Furthermore, demographics of the participants are different from those in the UK. Only one study is likely to be directly applicable.

Evidence Statement ES21

There is moderate evidence that ethnicity is an important predictor of smoking behaviour, such that white students were less likely to be smokers. Similarly, there is moderate evidence that the observed association between race and smoking behaviour depended on how the outcome was measured.

Four studies (two [+] USA and two [-] USA) specifically studied whether ethnicity is an important factor in predicting post-test smoking among students exposed to school-based smoking prevention programmes. Only one study ([-] USA) demonstrated no association between ethnicity and smoking status. However, three studies found that ethnicity was an important factor in predicting post-test smoking behaviour. For example, one study ([+] USA) provided evidence that white students were less likely to be classified as smoker. Two studies (one [+] USA and one [-] USA) revealed that ethnicity affects smoking prevalence depending on how the outcome was measured. One multi-country study ([-] EU) in six European countries, provided evidence that in The Netherlands there was differential significant effects for adolescents with a Dutch and non-Dutch origin. The Dutch ESFA programme was effective for non-native adolescents with fewer new weekly smokers compared to new weekly smokers in the control group. An opposite effect was found in native Dutch adolescents with more new weekly smokers in the experimental compared to new smokers in the control group.

Applicability: none of the studies were conducted in the UK. It is not clear if the USA/EU findings are directly applicable to the UK since the school-based prevention programmes under investigation are specific to USA. Furthermore, demographics of the participants are different from those in the UK.

Evidence statement ES26

In one RCT, engagement with the intervention (reported programme interesting/very interesting and useful) was shown to be related to follow-up smoking status; those engaging being less likely to be smokers at 1 year.

Evidence statement QR1

Delivery context: evidence from two UK (one [++], one [+]), one Canadian (++) and three American (all [+]) qualitative studies suggests that aspects of the delivery context of school-based interventions act as barriers or facilitators to effective delivery. The main facilitators were:

  • timing the intervention to suit (that is, not conflict with) school-assessment schedules

  • timing the intervention to include multiple sessions over the course of a school year

  • reinforcing smoking prevention messages in school curricula until school leaving age

  • delivering school-based prevention interventions as part of a wider tobacco control strategy

  • involving key partner organisations in design and delivery (such as the school nursing service and universities).

The main barrier was delivering the intervention in a setting where teachers and other school staff are smokers.

Evidence statement QR3

Peer interventions: there is evidence from three UK (two [++] and one [+]) and one American (+) study and one systematic review (++) that interventions that directly address peer smoking norms through involving young people in delivery can facilitate the successful implementation of school-based prevention interventions. The main facilitators to the delivery of peer interventions were:

  • nomination of peer supporters by fellow students

  • training for peer supporters delivered away from school and by external professionals

  • flexibility for peer supporters in how and when they deliver the intervention

  • adding 'value' to peer intervention by inclusion of other prevention education materials (such as videos) in schools

  • good communication between the external intervention development or research team and school staff.

Barriers to the delivery of peer interventions were:

  • teacher's concern about 'suitability' of some peer supporters selected by fellow students

  • peer norms and peer group structure can influence how much and when adolescents smoke, and can also influence the extent to which young people are receptive to prevention messages delivered by peers.

Evidence statement QR4

Delivery mechanisms: there is evidence from three UK (one [++], one [+] and one [-]) and three American (all [+]) qualitative studies that specific elements of the delivery mechanism for school-based prevention interventions can act as facilitators or barriers. Facilitators include:

  • delivery of the intervention by trusted external professionals (such as doctors)

  • delivery of the intervention by non-smoking teachers

  • delivery of the intervention by teachers with higher self-efficacy

  • involvement of parents in delivery (primarily delivery of supporting materials at home).

Barriers included:

  • delivery of the intervention by teachers who are reluctant to discuss parental smoking

  • delivery of the intervention by teachers who use outdated methods to communicate prevention messages.

Evidence statement QR5

Smokefree schools: there is evidence from one UK (+), one Canadian (++) and one American (-) study that the extent and enforcement of smokefree school policies can act as a facilitator or barrier to school-based smoking prevention. Facilitators included:

  • smokefree policies that include all internal areas and all school grounds

  • smokefree policies that applied to staff as well as pupils.

Barriers included:

  • existing designated smoking areas in school grounds or buildings

  • poor enforcement of smokefree policies.

Evidence statement QR6

Programme content: there is evidence from seven American (all [+]), one Canadian (++) and one UK (-) qualitative studies that specific elements of programme content can act as facilitators or barriers to the delivery of school-based prevention interventions. Facilitators include:

  • content that is innovative and interactive

  • content that includes role play

  • content that includes new material, such as on the cost of smoking

  • content that includes correcting misconceptions of high smoking prevalence among young people

  • content that is ethnically and culturally sensitive

  • content that is non-judgemental

  • content that included de-normalisation approaches (building on the Florida 'Truth' campaign approach, exposing the activities of the tobacco industry).

Barriers include:

  • content that included fear-based approaches to prevention

  • content that is too complex.

Expert testimony

Expert testimony to PHIAC (see appendix A) was used to inform the recommendations. Please refer to the PHIAC minutes for further details.

Cost-effectiveness evidence

Overall, school-based smoking prevention programmes were found to be cost-effective, although there was a lack of evidence on their long-term effects.

A modelling analysis was undertaken to explore whether a delay in the age of smoking uptake makes it more likely that someone will quit later in life. Effect sizes were based on 26 RCTs identified during the systematic review of effectiveness.

The outcome of the analysis suggests that a school-based smoking prevention programme may be cost effective at a threshold of £20,000 to £30,000 per quality adjusted life year (QALY) gained. This was the case when taking into account a range of factors. This includes the relationship between age of smoking initiation and probability of smoking in later life, the mortality of smokers compared to non-smokers, the health-related quality of life of people who smoke and their lifetime medical costs.

For further details, see 'School-based interventions to prevent the uptake of smoking among children and young people: cost-effectiveness model'.

Fieldwork findings

Fieldwork aimed to test the relevance, usefulness and feasibility of putting the recommendations into practice. PHIAC considered the findings when developing the final recommendations. For details, go to the fieldwork section in appendix B and online.

Fieldwork participants who are involved in smoking prevention activities in an educational setting were fairly positive about the recommendations and their potential to help prevent smoking among children and young people. Many participants stated that:

  • this was an important but neglected topic in schools

  • connections with the National Healthy Schools programme should be clarified and the guidance should use the terminology used by educators

  • more emphasis should be placed on the role of school staff in delivering interventions, rather than depending on outside experts

  • more emphasis should be put on integrating information on smoking into the general curriculum (that is, smoking prevention should take a cross-curriculum approach)

  • more clarity is needed on the recommendation about using a peer-led programme.

The young people who participated in the focus groups were positive about the recommendations, although their experiences of PSHE education varied. Many participants:

  • strongly approved of a 'whole school', joined-up approach to smoking prevention which also encourages those who are experimenting with smoking to quit

  • wanted to be directly involved in the design, delivery and evaluation of measures to prevent smoking

  • wanted to be able to discuss smoking in the context of their own lives and experiences.

  • National Institute for Health and Care Excellence (NICE)