Longitudinal Transitions in Initiation, Cessation, and Relapse of Smoking and E-Cigarette Use Among US Youth and Adults

IntroductionLongitudinal surveys provide data to estimate transition probabilities between cigarette smoking, e-cigarette use, and dual use of both, facilitating projections of future use and the impact of policies. MethodsWe fit a continuous time Markov multi-state model for youth (ages 12-17y) and adults ([&ge;]18y) in Waves 1-4.5 of the Population Assessment of Tobacco and Health (PATH) longitudinal survey and estimated smoking and e-cigarette transition frequencies, including initiation, cessation, and relapse. We validated transition frequency results in a microsimulation model by projecting smoking and e-cigarette use prevalence over time. ResultsThere was more volatility in smoking and e-cigarette use among youth than among adults. For youth never smokers, annual smoking initiation among never/current/former e-cigarette users occurred in 0.4% (95% CI 0.2-0.6%)/8.8% (7.0-10.7%)/3.1% (2.1-4.2%), and current e-cigarette users were more likely to quit e-cigarettes than to initiate smoking (absolute difference in annual probability 46.5%, 38.7-54.2%). For adult current smokers, annual smoking cessation among never/current/former e-cigarette users occurred in 22.6% (20.9-24.3%)/14.5% (11.5-17.4%)/15.1% (12.1-18.2%). For adult current dual users, 14.5% quit smoking and 49.5% quit e-cigarettes annually. For adult former smokers, annual smoking relapse among never/current/former e-cigarette users occurred in 17.7% (15.8-19.6%)/29.3% (23.8-34.7%)/32.8% (27.1-38.6%). Using these transition probabilities in a microsimulation model accurately projected smoking and e-cigarette use prevalence at 12 and 24 months compared to PATH empirical data (root-mean-square error <0.7%). DiscussionPATH Waves 1-4.5 contain sufficient data to generate smoking and e-cigarette use transition frequency estimates for youth and adults in a microsimulation model. E-cigarette use among youth is especially volatile. IMPLICATIONSThis study estimates longitudinal transitions between nine cigarette smoking and e-cigarette use states among both youth and adults. Youth smoking and e-cigarette use was more volatile than that of adults. Smoking initiation, cessation, and relapse differed by e-cigarette use status, but it is not clear that e-cigarette use itself caused these differences. A tobacco and nicotine policy model parametrized with our estimates accurately predicted smoking and e-cigarette use prevalence over two years, enabling future work to project the outcomes and cost-effectiveness of policies targeting smoking and e-cigarette use.