RESUMO
OBJECTIVE: To identify predictors of smoking initiation among non smoking Tunisian school children; and to propose efficient antismoking strategies in order to prevent smoking initiation. METHODS: It was a cohort study surveying prospectively for four years pupils attending schools in Sousse city in Tunisia. 441 non smoking pupils aged 13-15 years attending secondary schools in Sousse. Data were collected by a self administered questionnaire during class session. Two ways cross tabulation, univariate and multivariate logistic regression analyses were the main analytical methods. RESULTS: 57.1% of the surveyed population were girls, 42.9 % were boys. 63% had at least one of their peers who smoked. Before 1999, 16.6% had already tried to smoke, 29.5% had already experienced alcohol. In 2003, smoking prevalence was 17%. 4.8% were girls; 33.3% of boys; p<0,001). 69.9 % of these smokers declared that they would carry on smoking during the following five years. Predictors which were highly associated with smoking initiation were previous experimentation with alcohol and tobacco, having a smoking best friend, lack of sensitization from the part of the school, believing that smoking makes one feel cool and that tobacco shouldn't be forbidden in public places. CONCLUSION: Intervention programs should target young children to avoid experiencing the first cigarette. Multidisciplinary management including community and school based intervention highlighted by mass media campaigns may provide schoolchildren with skills to resist smoking peers prompts to adopt unhealthy habits such as smoking.