Longitudinal Assessment of Association Between Tobacco Use and Tobacco Dependence Among Adults: Latent Class Analysis of the Population Assessment of Tobacco and Health Study Waves 1–4

Abstract

Abstract Introduction With increasing tobacco product varieties, understanding tobacco use (TU) profiles and their associations with tobacco dependence (TD) has also become increasingly challenging. Aims and Methods We aimed to identify TU profiles and their associations with TD over time, and to identify subgroups with high risk of TD. We included 3463 adult recent tobacco users who had complete TU and TD data across waves 1–4 of the Population Assessment of Tobacco and Health (PATH) study. We used a composite index of TD and a summed TD score from an established 16-item TD measure. We applied a latent class analysis to identify TU profiles based on participants’ usage of eight common tobacco product groups at each survey wave and to check the stability of the TU profiles over time. We then used generalized estimating equations regressions to evaluate the longitudinal TU–TD association, adjusting for potential confounders. Results We identified three distinct TU profiles that remained consistent across four survey waves: Dominant cigarette users (62%–68%), poly users with high propensity of using traditional cigarettes, e-cigarettes, and cigars (24%–31%), and dominant smokeless product users (7%–9%). Covariate-adjusted models showed that TD was significantly lower among the poly users and the dominant smokeless users, compared to that among the dominant cigarette users. Conclusions Both TU profiles and their associations with TD were stable over time at the population level. Poly users and smokeless product users were consistently associated with lower TD than cigarette-dominant users, suggesting the need for tailored tobacco cessation interventions for users with different TU profiles. Implications The finding of consistent TU profiles across four survey waves extends the current literature in capturing TU patterns in an evolving tobacco product landscape. The finding of the overall higher level of TD among the cigarette-dominant users compared to the other TU latent profiles (the Cig+eCig+Cigar dominant poly users and the dominant smokeless product users) can help identify high-risk groups for potential interventions. Our application of innovative statistical methods to high-quality longitudinal data from the PATH study helps improve the understanding of the dynamic TU–TD relationship over time.