Tobacco Use Patterns Among Youth in the United States: Sociodemographic and Tobacco-Related Correlates

Background and Purpose: Cigarette smoking has declined in the past decade, yet use among middle school and high school students remains a concern. Although extensive research on adolescent tobacco use has been conducted, relatively little research has examined tobacco use patterns among adolescents, especially given the recent proliferation of new forms of tobacco use (e.g. e-cigarettes). For example, some students may use only one form of tobacco, whereas other students may use multiple forms. Moreover, frequency of use may also vary with some youth using sporadically and other youth using more frequently.   This is important, as risk factors for use may be different among students who use different types of tobacco. The purpose of this study is to examine heterogeneity in tobacco use patterns among middle and high school students based on six forms of tobacco use (cigarettes, cigars, chewing tobacco, pips, e-cigarettes, and hookah pipes) and three levels of use status (no use, any lifetime use and past month use).

Methods: To evaluate the tobacco use patterns among youth, we analyzed data from the National Youth Tobacco Survey (NYTS) from 2011.  The NYTS is a nationally representative survey of youth in middle and high schools conducted annually focused on tobacco use and contextual factors related to use. We estimated latent class analysis models with covariates. Latent class indicators were 3-level ordinal indicators of use of the abovementioned six different tobacco types. Covariates were male gender, dummy-coded race/ethnicity, exposure tobacco advertising, healthcare professional advice about use, exposure to tobacco use, and parental disapproval. We estimated one- to five-class models and evaluated models based on information criteria (e.g. AIC, BIC), classification quality (i.e. entropy) and substantive meaningfulness.

Results: Our study showed that a four-class model was the best fit to the data with the following classes: “Very low risk” (77.6%), “lifetime experimenters” (13.2%) “multiple-type past-month users” (5.5%), and “multiple-type lifetime users (3.6%)”. Compared with low risk users, persons in all the other classes were more likely to be exposed to tobacco use in cars and in their homes, had greater access to tobacco, and had lower levels of parental disapproval.  Youth in high school and males were more likely to be in all higher risk groups.  Latino youth were more likely to be lifetime experimenters (OR=1.88) and African American youth were less likely to be lifetime multiple tobacco users (OR=.21)

Conclusions and Implications: From LCA analysis, use of a single type of tobacco is uncommon, as most youth who were currently using one tobacco type, also used other types.  Rather than being specific to one class, tobacco access was associated with each at-risk pattern of use. Future research should consider temporal factors in use patterns with the goal of understanding how use of different tobacco types (e.g. e-cigarettes) may influence future patterns in adolescence and young adulthood. As options for tobacco use expand among youth, research studies should consider whether certain types of tobacco use patterns are problematic, with the distal goal of informing prevention efforts with youth.