Modeling individual behavioral changes : the case of tobacco uptake in a cohort of school children

The aim of this thesis was to elucidate preventable determinants of smoking and smokeless tobacco use in adolescence in the frame of a longitudinal study (BROMS cohort). The study population encompasses 3019 children recruited from 5th grade classes in 96 schools in Stockholm, Sweden during the academic year 1997-1998. A study of developmental trajectories of tobacco use between the ages of 11 and 17 revealed that snus use, once initiated, is best described through three patterns, mainly differing in the speed of progression. The best description of cigarette smoking is provided by four patterns, distinguished by age at onset and speed of progression. Several traditional risk factors for tobacco use measured at baseline, such as gender, parents’ and friends’ tobacco use, and prevalence of tobacco use among classmates influenced individual probabilities of belonging to a particular trajectory. Characteristics of the class, rather than of the school were associated with smoking initiation between 5th and 6th grade. A decreased risk of smoking uptake was associated with exposure to short anti-tobacco education before the 5th grade, compared to no education. Problematic interpersonal relations in the class were associated with an increased risk of smoking initiation, compared to positive interpersonal relations. Parents’ tobacco use was associated with adolescents’ current use of either cigarettes or snus. Prevalence of current smoking was lower among children whose fathers used snus than among those whose fathers smoked. Maternal influences on youths’ tobacco use were stronger than paternal influences. Sixth grade students’ knowledge of tobacco properties and effects was not associated with future use. Despite this lack of association, a clear link was found between access to information sources and knowledge about tobacco. Nearly two thirds of tobacco users in the middle-high levels of the compulsory school used cigarettes as first tobacco product, while one out of five took up snus first. After initial use, experimentation with the other product occurred to greater extent for “cigarette starters” than for “snus starters”. Whereas the existence of the so called “gateway effect” of snus use, i.e. induction of cigarette smoking, could not be excluded, the effect in this cohort was likely modest, accounting at the most for 6% of the final smoking prevalence