DESCRIPTION OF THE PROGRAMME

The St-Louis du Parc Heart Health Project was implemented from 1992 to 1997, among children aged 9–12 years in an inner city neighbourhood served by the St-Louis du Parc Local Community Health and Social Service Centre (CLSC) in Montreal, Canada. The population of this low income territory is multiethnic.²¹ The St-Louis du Parc Heart Health Project was based on the hypothesis that a comprehensive programme of interventions targeting both children and significant adults (family, youth workers, teachers, nurses), for multiple risk factors (sedentary behaviour, obesity, and smoking) in multiple sites (schools, community organisations, recreation centres, and homes) and using diverse strategies outlined in the Ottawa Charter for Health Promotion^22,23 (healthy public policies, community development, supportive environments, and personal skills development) would have an impact on children’s cardiovascular risk factor profiles (physical activity, non-smoking, and diet). The programme was implemented in eight elementary schools, four recreation organisations, four cultural organisations, and the CLSC.^22,24,25

Smoking related interventions

The programme comprised 25 interventions that targeted children, significant adults, and their environments.²⁴ Three interventions focused specifically on smoking: a school based anti-smoking curriculum, an extracurricular programme, and enforcement of a law prohibiting the sale of tobacco to minors. The analysis reported here was restricted to the school based curriculum because it was the only intervention to which all subjects were exposed.

Of 20 hours devoted annually to the heart health school curriculum, 5–8 hours were dedicated to smoking prevention. The smoking prevention curriculum was developed in 1991 based on the results of published meta-analyses on smoking prevention in schools,²⁶ the 1991 recommendations for youth tobacco prevention programming of the National Cancer Institute,²⁷ and an examination of existing school based prevention programmes (that is, the Waterloo^28,29 and Know Your Body programmes³⁰). Based on social learning theory and geared to age specific stages of cognitive development, the content of the programme aimed to: provide information about the effects of tobacco on health; develop non-smoking attitudes; and consolidate or develop skills in peer pressure resistance and decision making. It also aimed to stimulate schools to adopt smoke-free policies. A health educator delivered the programme during the first year in grades 4, 5, and 6. Subsequently teachers were asked to administer the curriculum, although some continued to ask that the health educator teach it. The health educator was therefore present in participating schools for the full five years of the project.

EVALUATION METHODS

The impact of the programme was assessed in a quasi-experimental study design in which eight experimental schools were matched to 16 comparison schools on poverty and language indices. Data were collected in a total of 11 195 questionnaires. The response proportions in the longitudinal data at one year were 71.6% in intervention and 60.5% in comparison schools. They were 74.0% and 57.9% respectively in comparison schools in the two year follow up.²⁰

Smoking status was measured in two items adapted from previous research.³¹ Students were categorised as never-smokers if they reported no smoking at all during their lifetime. Ever-smokers included subjects who reported any smoking during their lifetime, even just a puff. Current smokers included subjects who reported smoking one or more times in the past year.

An implementation evaluation was conducted through surveys of fourth, fifth, and sixth grade teachers in the experimental schools, at three time points during the project. This evaluation assessed the role of organisational factors, personal characteristics (such as smoking status), and implementation strategies on the implementation of the programme.²⁵

In order to assess the relation between intensity of exposure to the programme and the strength of the reverse effects on smoking, each school was categorised according to the intensity of the programme actually delivered, based on implementation evaluation and administrative information about: the number of smoking curriculum hours given; the amount of time the health educator was present in the school; the number of interventions conducted targeting parents and teachers; and the extent of local community groups’ collaboration. Low intensity schools included those with less than three hours of the smoking curriculum annually and no more than one smoking prevention activity outside the classroom.

RESULTS

Detailed descriptions of the study methods and the main results have been reported elsewhere.²⁰ For this paper, we stratified schools by intensity of programme delivery to examine if the relation between exposure and outcome differed according to intensity.

Compared to low intensity schools, the reverse impact of the programme was more pronounced in high intensity schools (table 1). In low intensity schools, the programme had no effect on smoking initiation (tendency for never-smokers to become ever-smokers). However, the St-Louis du Parc Heart Health Project increased both smoking initiation and maintenance (tendency for current smokers to retain that status) in high intensity schools, as well as maintenance in low intensity schools.

In order to identify deficiencies in programme design that might explain the reverse effects, all programme components were submitted to a critical appraisal using three sets of current guidelines for the design and implementation of school based smoking prevention programmes: those of the National Cancer Institute,^27,32 of the Institute of Medicine,²¹ and the Centers for Disease Control and Prevention,^33,34 as well as those identified in a literature review^{1,2,4,7,8,10,16,26,35–53} (table 2). According to this analysis, the St-Louis du Parc Heart Health Project was consistent with several key guidelines: it used a comprehensive approach; it was culturally adapted; it used relevant content and methods of integrating the content; it had recommended intensity levels; it included booster sessions; it was appropriately timed with respect to students’ ages and grades; it used teachers as well as health professionals to deliver interventions; and it included teacher training. However, two criteria were only partially respected. First, sustained activity to strengthen school smoking policies was not undertaken because of overwhelming resistance from teachers. Second, support for smoking cessation efforts (by students) and teachers was provided on an ad-hoc individual basis, rather than through systematic cessation programmes.

DISCUSSION

The St-Louis du Parc Heart Health Project conformed to most recommendations for effective prevention of tobacco consumption among primary school children, yet had the opposite effect. Consultation and reflection among tobacco experts resulted in several possible explanations for the reverse effects. In the following section we first discuss the importance of the social (parents, instructors, siblings) and physical (smoke-free area bylaws, law forbidding the sale of tobacco products to minors) environments on smoking initiation. Next, we discuss the number and quality of anti-smoking interventions in schools, including the presence of health educators, frequency of anti-tobacco messages, adaptation of messages to the developmental stages of children, and finally the cultural adaptation of anti-smoking activities.

Physical and social environment

Comprehensive school based tobacco prevention programmes such as the St-Louis du Parc Heart Health Project generally intend to operationalise a socio-ecological paradigm, on the premise that children’s behaviour is strongly influenced by their physical and social environments. This perspective holds that effective school based smoking prevention must take into consideration the prevailing smoking norms and policies, siblings, parents’ and teachers’ attitudes and behaviour, participants’ socio-demographic characteristics, and tobacco industry tactics, in interventions that are targeted to specific participants and reinforced by multi-component tobacco prevention and control efforts.^29,54

Several environmental components of the St-Louis du Parc Heart Health Project were not implemented as planned. For example, systematic enforcement of tobacco control policies in experimental schools was not undertaken because of negative reactions by parents⁵⁵ and teaching staff, so that teachers continued to smoke on school premises during the entire five year programme. The sociophysical environment in experimental schools was therefore inconsistent with the anti-smoking messages delivered through the curriculum. In addition, interventions aimed at stopping the sale of single cigarettes to minors in local conveniences stores were not supported by all of the schools’ parent committees,⁵⁵ contributing to continued easy access to tobacco products. These factors may have worked at cross purposes with the in-class intervention. In a study of 32 Quebec communities, Frohlich^56,57 showed that primary school students’ smoking initiation is associated with characteristics of their social environment and local resources for tobacco restrictions, so that the more prevalent smoking is among high status children and the more ambiguous the rules and laws in the community, the more children will be sensitised to the inconsistency and turn toward smoking. Recent results from the US state of Oregon,⁵⁸ Wales,⁵⁹ and Spain⁶⁰ suggest an association between school policy strength, policy enforcement, and smoking behaviours of high school students—prevalence of smoking is lower in schools where student smoking restrictions are enforced.

Our consultations carried out with tobacco experts reinforced the importance of interventions driven by a comprehensive ecological model. However, it was suggested that for environmental interventions to be effective, they should be on a larger scale than that afforded by a neighbourhood level intervention, probably under provincial (state, district) leadership to ensure better coordination of actions. Community and other non-governmental organisations should be encouraged to play a major role in local anti-smoking interventions.²³ For example, although the law forbidding the sale of tobacco products to minors is part of a provincial action plan, its application is local in that every neighbourhood is responsible for implementation in convenience stores in its territory. In our case, the law had not been implemented provincially, which led to its inconsistent application locally. Moreover the experts also emphasised that more understanding is needed of how the interactions among the various social and physical environmental factors influence children’s smoking behaviour, especially among younger children.

Differential reverse effects

We sought explanations as to why the strongest reverse effects on smoking were observed in the four St-Louis du Parc schools where programme intensity was greatest. While involved in all eight experimental schools, the health educator was much more intensively involved in the four schools that showed the greatest negative impacts. Our implementation evaluation indicated that several teachers in these four schools preferred that the smoking content of the curriculum be delivered by the health educator. It is possible that the reverse effects were attributable, at least in part, to the educator. He could have been perceived to be pro-smoking by the children or he could inadvertently have passed pro-smoking messages. Our results evoke those of the Alert project,¹⁶ where schools in which the anti-smoking curriculum was delivered by an adult health educator in conjunction with teen leaders showed higher levels of continued smoking among baseline smokers in intervention compared to control schools.

The association between intervention intensity and increased smoking invites consideration of the effects of repeated anti-smoking messages on children in this age group. It could be that multiple messages about smoking from diverse sources piques children’s curiosity, rendering them more aware of and sensitive to smoking, and more interested in experimentation. At the same time, if these sensitised children are exposed to smoking cues and models in their social environment, they could become acutely aware of the contradiction between their environments (for example, visibly smoking teachers and parents, easy access to cigarettes) and anti-smoking in-class messages, and more motivated to resolve the contradiction (cognitive dissonance) in favour of the most normative solution—that is, by becoming more favourable to smoking.

Health education programmes delivered in school settings must take the readiness of children for certain types of information into consideration. Primary school children try to make sense of educational information by transposing it to their daily experiences. When the information provided contradicts their experience, or when it suggests dangers to their parents’ health,⁶¹ the resulting dissonance can create anxiety. Adaptive defence mechanisms such as denial, negation or rejection could block information integration, and render the programme ineffective. However, if the anxiety defence mechanisms result in transformation of the messages into their opposite in order to resolve the dissonance, the risk of reverse effects would be increased. A mechanism of this type may be responsible for the increased smoking initiation observed in non-smoking children in St-Louis du Parc Heart Health Project, and is consistent with our finding that smoking initiation risk is highest among children whose parents and siblings smoke,⁶² because these children would have the strongest need to resolve the dissonance.

One of the difficulties encountered when designing the programme in the early 1990s was the paucity of data on the effectiveness of programmes designed specifically for children aged 9–10 years. According to Piagetian stages of cognitive development and Kohlberg’s stages of moral judgement, fourth grade children (ages 9 and 10 years) are likely to reason quite differently from sixth grade children (ages 12 and 13 years).⁶³ These differences have implications for the design of anti-smoking interventions. In Piagetian terms, fourth grade children are generally in the concrete operations phase, where their reasoning is based on the organisation of concrete, tangible elements. At this stage, anti-smoking messages may be understood as simple instructions or orders not to smoke. In terms of moral reasoning, children at this age act to obey the wishes, as they perceive them, of significant adults. Messages from adults are taken very literally, with little possibility of conditionality. In the fifth and sixth grade, children are in transition toward the formal operations stage, and are able to begin imagining their reactions to hypothetical situations, such as someone offering them a cigarette, and to integrate abstract concepts such as the negative effects of smoking. Children in the formal operations stage are able to see both the good and bad sides of a behaviour, and develop a personal position in relation to abstract arguments.

The content of the St-Louis du Parc Heart Health Project was evaluated by a developmental psychologist and found to be appropriately designed in term of Piagetian and moral reasoning stages for grade 6 children, but less so for grade 4 children. The grade 4 curriculum focused on the effects of tobacco on the body and the environment, on myths about smoking, and on self assertion. In grade 4, the effects of tobacco on the body in terms of the abstract concept of “health” may have been too intangible for children to understand them. Similarly, 9 year olds may not have been able to fully benefit from the self assertion content if they were not able to comprehend the hypothetical situations proposed in the curriculum. This may have caused the younger children to focus most directly on the concrete topic of smoking, increasing their awareness and interest in it without adequately communicating the prevention messages. Indeed, analyses of the outcome data showed that the reverse effects on smoking initiation were especially strong, relative to controls, among grade 4 intervention children.²⁰

Few similar projects have targeted disadvantaged youth in multiethnic environments,²¹ and little is known about appropriate cultural adaptation of the content and method of delivery of smoking prevention programmes in these settings. The St-Louis du Parc Heart Health Project was developed to reach children with diverse family origins. To ensure cultural adaptation, the programme messages were pre-tested in several cultural groups and modified based on the results.²³ The teachers’ curriculum materials identified issues and strategies for communicating health messages to specific cultural groups, and all teachers attended a training session on multiculturalism offered by their school board. Nevertheless, it is possible that despite the precautions taken to adapt the programme content appropriately, inadequate adaptation led students to misinterpret the anti-smoking messages. It is also possible that children newly arrived to Canada are more vulnerable to smoking during an acculturation period; they may see smoking as normative and a route to becoming integrated into the dominant culture. The programme may have inadvertently encouraged smoking during this vulnerable period. The likelihood that this occurred may vary according to specific cultures and recency of immigration. Although ethnic composition of the samples was controlled in the study design and analyses, the intervention group included children of more recent immigrants and of somewhat different ethnic origins from children in the comparison group.²⁰ In addition, the ethnic mix of students evolved over the study period in both the intervention and comparison schools. It may be that the reverse effects are associated with ethnicity variables not measured or controlled in the study.

In summary, two hypotheses were retained: (1) the physical and social environments to which subjects were exposed provided mixed messages so that there was no clear non-smoking norm; (2) the intervention may have encouraged tobacco use because of its novelty, intensity, content, or (lack of adequate) cultural adaptation. Alternatively the dynamic personality of the health educator may have encouraged children to smoke.

Standard guidelines: need for greater precision

Three groups of US experts published guidelines on the school based prevention of tobacco use by youth. First, in 1987 the National Cancer Institute brought a panel of experts together and this group published guidelines in 1989.³² In 1991,²⁷ the components required for a school based programme were published. Then, in 1994, while endorsing these directives, the Centers for Disease Control³³ added elements concerning policies, parental involvement, and support for students wishing to quit smoking. The Institute of Medicine’s committee²¹ on the prevention of nicotine use in children and youth stressed the importance of national leadership, a comprehensive approach, and consistent research. Finally, the Centers for Disease Control re-endorsed in 1999³⁴ the same recommendations as those published in 1994, stressing the importance of incorporating anti-tobacco programmes into a comprehensive school based programme, as well as the importance of combining school based efforts with community coalitions and national anti-tobacco campaigns.

Comparison of the St-Louis du Parc Heart Health Project to standard guidelines (table 2)^21,27,33,34 raised questions about several statements in these guidelines and their implications. Although they provide helpful overall design orientations, much is left to interpretation when programmes are actually operationalised. For example, the specific elements of effective parental involvement in smoking prevention are not defined. In addition, directives for achieving cultural adaptation are also not explicit. However, although we call for greater precision in these guidelines to facilitate operationalisation, it is clear that much more work is needed to develop the “evidence base” to achieve this precision.

Conclusion

Comprehensive school based smoking prevention programmes are a key strategy to reach children with important health messages. However, our experience with unanticipated reverse effects related to exposure to a state-of-the-art smoking prevention programme necessitates stepping back and carefully reviewing the paradigms underlying these programmes. We suggest that existing norms and consensus guidelines formulated in the early 1990s need to be revisited and questioned, to help identify the directions for school based interventions in the years to come. The critical appraisal undertaken with respect to the St-Louis du Parc Heart Health Project aims to broaden the debate on avenues of action targeting primary school children, so that future interventions and research can make a more effective contribution to eliminating smoking initiation and improving children’s health outcomes.

The results of this appraisal strongly suggest that school based interventions should begin by developing a clear and coherent social norm on the non-use of tobacco. Before implementing interventions aimed directly at children, school programmes should ensure that non-smoking legislation is implemented and enforced and that proximal environments (commercial and public spaces, business practices) are also operating in coherence with a non-use norm. Prevention programmes should systematically include cessation support for significant adults, most notably parents and teachers. Only then should direct within-school intervention be provided for children. The programme design should take key student characteristics into consideration, including sex, lifestyle, socioeconomic status, and ethnicity. It should also target children’s stages of cognitive and moral development appropriately.

This leads us to endorse the recommendations of the California Department of Health Services,⁶⁴ that prevention programmes “should not concentrate on youth as a primary target group. The goal is to build a social environment where families do not use tobacco, and adults are positive role models for youth” (page 20). According to these recommendations, school based tobacco prevention and education programmes should be established “only if certain stringent conditions are met: 1– only if a vigorous community-wide program addressing broad social norm change is also supported” (school-based programmes should coordinate with other local tobacco control activities and programmes in the community); “2– only if the programs consist of well-developed, sustainable interventions that reach kids at the school site and in classrooms; and 3– only if accountability mechanisms are built in to insure that the programs have school-site and teacher buy-in, and include intensive technical assistance and training for teachers” (page 20). We note, however, that recommendation would have precluded inclusion of our intervention and comparison schools from the St-Louis du Parc Heart Health Project.

Moreover, neighbourhood level interventions should be orchestrated to complement regional, provincial, and national programmes. In our case, a stronger provincial leadership on tobacco issues through policy and legislation would have enhanced coordination of actions, especially those directed to the creation of smoke-free environments and the limitation of access to tobacco products to children. Finally, school based health promotion programmes should be integrated into a broader, long term education policy,⁶⁵ as education is a strong predictor of healthy choices about tobacco.