| dc.description.abstract | One of the greatest accomplishments of public health has been the significant reductions in harms due to smoking. Although tobacco cigarettes have typically been the product with the highest prevalence of use, there remains a significant number of youth that use other products such as e-cigarettes, cigarillos or little cigars (CLCs), cigars, smokeless tobacco (SLT), and hookah. Past research has focused on tobacco cigarette smoking behaviours and has neglected investigating the use of other tobacco and nicotine products. The objectives of this dissertation were (1) to examine the ability of current susceptibility measures to predict the use of other tobacco and nicotine products, (2) to identify latent classes of tobacco and nicotine product use, and (3) to identify latent trajectory groups for the use of each product. Four manuscripts addressed these objectives for six tobacco products (i.e., tobacco cigarettes, e-cigarettes, CLCs, cigars, SLT, and hookah) using longitudinal data from students in Ontario that participated in the COMPASS study from 2013-2016.
The first manuscript calculated the sensitivity, specificity, positive and negative predictive values of the susceptibility to smoking construct for the use of each tobacco and nicotine product. Results indicated that the sensitivity of the construct was moderate while the specificity was high. The positive predictive value was variable, depending on the prevalence of the product, while the negative predictive value was very high. Similar values were calculated for each measure of the susceptibility construct.
The second manuscript identified student-level sociodemographic and behavioural characteristics of non-smoking youth at baseline that used each tobacco and nicotine product one- and two-years later. Given that the first manuscript provided evidence for the predictive validity of the susceptibility construct, this manuscript included susceptibility to future smoking as a predictor in the models. Baseline susceptibility to future smoking was strongly associated with the use of each tobacco product and e-cigarette at one- and two-year follow-up. Additionally, students that had friends that smoked cigarettes or who reported binge drinking at baseline had higher odds of reporting the use of each product at follow-up.
The third manuscript identified tobacco and nicotine product use clusters for three consecutive years using latent class analysis. At baseline, a three-class model was identified as best [(1) non-current users; (2) current tobacco cigarette, CLC, and e-cigarette users; (3) current polyproduct users], while a four-class model was identified one-year [(1) non-current users; (2) current e-cigarette users; (3) current dual tobacco cigarette and CLC users; (4) current polyproduct users] and two-years later [(1) non-current users; (2) current dual tobacco cigarette and e-cigarette users; (3) current tobacco cigarette, CLC, cigar, and e-cigarette users; (4) current polyproduct users]. Results of the multinomial regression models indicate that students that reported having friends that smoked cigarettes, binge drinking, and using marijuana were more likely to be classified into a current use class relative to a non-current use class.
The final manuscript identified latent trajectory groups for the use of each tobacco and nicotine product using latent trajectory analysis. Given the results of the first two manuscripts, measures of susceptibility to future smoking were included when identifying trajectory groups. Consistent across all products, five groups of users were identified: (1) non-susceptible non-users, (2) non-susceptible puffers, (3) stable low intenders, (4) escalating experimenters, and (5) consistent current users. Across all tobacco and nicotine products, students had the highest probability of remaining in the same group over time, although some transitions in group membership were evident. Results of the multinomial logistic regression models indicate that across all products, students that reported having friends that smoked cigarettes, binge drinking, and using marijuana were more likely to be classified into any other trajectory group relative to the non-susceptible non-users group. Additionally, students that reported a higher school connectedness score and eating breakfast every day in a usual school week were less likely to be classified into any other trajectory group relative to the non-susceptible non-users group.
This dissertation fills an important gap with respect to our knowledge of other tobacco and nicotine product use among youth in Canada. The findings of this dissertation have implications for research and practice and highlight the need for inclusive tobacco control programming, particularly with respect to school-level prevention and cessation programs. Given than many youth reported using more than one tobacco or nicotine product and commonly reported binge drinking and using marijuana, multi-substance use programs are needed. In addition, given that students transition into and out of tobacco and nicotine product use throughout secondary school, consistent programming may be needed to discourage the initiation and escalation of tobacco and nicotine product use throughout adolescence. | en |
