摘要：Objectives. We identified the most effective mix of school-based policies, programs, and regional environments associated with low school smoking rates in a cohort of Canadian high schools over time. Methods. We collected a comprehensive set of student, school, and community data from a national cohort of 51 high schools in 2004 and 2007. Hierarchical linear modeling was used to predict school and community characteristics associated with school smoking prevalence. Results. Between 2004 and 2007, smoking prevalence decreased from 13.3% to 10.7% in cohort schools. Predictors of lower school smoking prevalence included both school characteristics related to prevention programming and community characteristics, including higher cigarette prices, a greater proportion of immigrants, higher education levels, and lower median household income. Conclusions. Effective approaches to reduce adolescent smoking will require interventions that focus on multiple factors. In particular, prevention programming and high pricing for cigarettes sold near schools may contribute to lower school smoking rates, and these factors are amenable to change. A sustained focus on smoking prevention is needed to maintain low levels of adolescent smoking. Adolescent tobacco use remains a major public health priority to reduce the disease burden of smoking-related illness. In addition to the health consequences of tobacco use that typically begin in adolescence, the total financial costs of smoking in the United States are estimated at $167.8 billion annually 1 and in Canada, $17 billion annually. 2 A significant challenge in addressing the problem is that youth smoking behaviors occur in many complex environments, including the home, school, community, and wider policy context. An ecological model provides a framework for understanding the complex influence and interactions of the broader environment on individual health behaviors. 3 Focusing on broader environmental approaches to adolescent tobacco control may serve to decrease opportunities for youths to obtain tobacco and foster social norms that discourage youths from smoking. To inform the development and implementation of more effective approaches to adolescent tobacco control, researchers need to identify the multiple influences of school, neighborhood, and community characteristics on adolescent smoking. 4–6 School-based approaches to addressing adolescent tobacco use have been the focus of much research and public health policy. Despite this, findings from studies examining the effectiveness of school-based policies 7–21 and prevention programs 22–24 on student smoking remain mixed. Previous research has shown that smoking rates vary across schools, irrespective of individual factors, suggesting a school effect on smoking rates. 8 School-level determinants have been found to explain between 4% and 40% of the variation in smoking across middle and high schools. 7,25,26 Moreover, factors in a school’s community, including the implementation of strategies such as educational programs, policies, and municipal regulations are likely to influence adolescent smoking. A study examining the impact of broader tobacco control policies (e.g., clean air policies, cigarette taxes) over a 15-year period found an association with lower adolescent smoking. 27 Ecological models have suggested that to better understand the influence of school contextual variables on smoking rates, the relationships among school-based tobacco control policies, programs, and other characteristics of the broader school and community environment need to be explored simultaneously. An integrated approach such as this would help determine the most effective mix of policy and program factors that support low school smoking rates. Most smoking outcomes have been assessed at the individual level. Although the purpose of school-based tobacco control policies and programs is to influence individual smoking behavior, they are also designed to modify the school setting by enhancing nonsmoking norms. Therefore, investigating and understanding smoking outcomes at the school level is important. Studying school-level smoking behavior will help to identify the relative importance of various factors in students’ school environments that have an impact on school-level smoking prevalence. For the relatively immobile high school student, local areas are very important for accessing goods and services, including those detrimental to health. 28 As such, school- and community area–targeted programs can be informed to increase capacity for change. We designed the Project Impact study to examine the relationship between adolescent smoking and the broader school and community environment. Previous findings from this study have suggested that significant differences exist in tobacco retailer variables between schools with high and low smoking rates, 6,25 that smoking rates vary across schools, and that policy characteristics can explain some of this variation. 6,14,29 Evidence has also shown that community characteristics are related to smoking rates. 6 However, evidence is lacking regarding how multiple factors, across school and community environments, work in combination and over time to influence adolescent smoking at the school level. The purpose of this study was to follow a cohort of Canadian schools over time to determine the influence of school-based tobacco control policies, programs, the areas surrounding schools (neighborhoods), and regional environments (municipality) on school smoking prevalence. We sought to identify factors within the schools and their communities that are associated with school smoking rates. Findings will be useful to policymakers in informing decision making that advances tobacco control.
