Are E-cigarette Users a Unique Group of Smokers? Latent Class Analysis of the National Youth Tobacco Survey

Abstract

We used data from the 2014 National Youth Tobacco Survey to examine patterns of cigarette smoking behavior and tobacco use. In light of the recent upsurge in e-cigarette use, we modeled current use and future intentions to use vape products along with combustible cigarette smoking and other tobacco products (i.e., cigars, cigarillos, chew, snuff, and dip). Latent class analyses indicated four discrete classes of smokers including a nominally involved class with very modest levels of tobacco product use, a class blending e-cigarette and cigars, a class of youth who predominantly use combustible cigarettes, and a group reporting indiscriminate use of almost all tobacco products excluding chew. Tests of invariance in item response probabilities and latent class proportions showed little variation across race and gender, albeit a new class of combustible cigarette and e-cigarette users emerged when examined by grade. Members of the heavy smoking and tobacco use class were more likely to be male, White, and older. Predictors of class membership included expectancies (perceived benefits of use), perceived harm (risk), media exposure, tobacco dependence, and the desire to quit. Findings are discussed in terms of characterizing risk among already smoking youth and how actionable prevention measures can be incorporated into existing universal and indicated programs that target reducing tobacco use and smoking behaviors.

Keywords: e-cigarettes; explanatory predictor; latent class analysis; multinomial logistic regression; smoking behavior; tobacco survey.