Recently, there has been extensive use of Electronic Nicotine Delivery Systems (ENDS) by young people in the United States (US). These products have the potential to significantly reduce the proportion of young adults who are addicted to cigarette smoking, either by influencing the proportion who escalate to daily cigarette smoking or by increasing the proportion who quit under the age of 35 years. Previous research has demonstrated that there are two groups of escalators to daily cigarette smoking: 1) early escalators who are daily smokers before age 20 years, and 2) later escalators (become daily smokers at 21+ years). This proposal will explore the long-term role of ENDS in changing the escalation and quitting of cigarette smoking using the first four waves of the national Population Assessment of Tobacco and Health (PATH) longitudinal study. The study used a four stage stratified area probability design and has enrolled ~26,000 people under the age of 35 years, half of whom were female. The public use data for the first two waves are available and include longitudinal weighting variables. Wave 3 is expected to be publicly available in Spring 2018, and Wave 4 one is expected year later. The weighted response for Wave 2 was 87% for 12-17 year olds, and 83% for adults. With these PATH data, we will explore two scientific premises: 1) that widespread promotion of ENDS has resulted in a new generation of nicotine users who rely on ENDS usage and who did not escalate to daily cigarette smokers when becoming adults; and 2) that the diffusion of ENDS among young tobacco users will result in lower nicotine dependence that, in turn, will be associated with higher rates of successful discontinuance of cigarette smoking before age 35 years. A significant methodological weakness in such complex longitudinal analyses is related to the consideration of a large number of time-varying confounding variables that are prognostic for changes in tobacco use behaviors and are also associated with the outcome of interest. Some epidemiological methods allow the achievement of covariate balance between different exposure groups. With these methods, ongoing differences in outcomes of interest are directly attributable to ENDS usage, for example. In this context, we leverage Propensity Score Matching (PSM), and Marginal Structural Models (MSMs) to adjust for time-varying confounding. We will also explore Targeted Maximum Likelihood Estimation (TMLE) techniques to optimize the performance of PSM. We will conduct separate analyses for early escalators, late escalators, and quitters before age 35 years. This proposal will answer an urgent public health question in regards to the long-term potential benefits associated with ENDS usage, while also providing an example to the tobacco research field of state-of-art epidemiological methods for complex longitudinal analyses.
Electronic Nicotine Delivery Systems (ENDS) appear to be changing the product pattern through which young people become nicotine dependent. Using the first four waves from the national Population Assessment of Tobacco and Health (PATH) Study, we will investigate whether ENDS use changes the pattern of escalation and early quitting among those less than 35 years in the United States.
