INTRODUCTION: Concurrent electronic nicotine delivery system (ENDS) and cigarette (dual) use is harmful. Identifying longitudinal trajectories of ENDS and cigarette use among dual users can help to determine the public health impact of ENDS and inform tobacco control policies and interventions.
OBJECTIVES: (1) To identify independent and joint trajectories of ENDS and cigarette use among wave (W) 1 adult dual users across W1 to W5 of the Population Assessment of Tobacco and Health (PATH) Study; and (2) identify W1 predictors of ENDS and cigarette joint trajectory group membership.
METHODS: We used group-based trajectory modelling to estimate independent and joint trajectories of ENDS and cigarette use from wave 1 (W1; 2013–2014) to wave 5 (W5; 2018–2019) among W1 adult established dual users of ENDS and cigarettes (n=545) from the PATH Study. We used multinomial logistic regression to identify W1 predictors of joint trajectories.
RESULTS: Two ENDS (early quitters=66.0%, stable users=34.0%) and three cigarette (stable users=55.2%, gradual quitters=27.3%, early quitters=17.5%) trajectories of W1 were identified. In joint trajectory analysis, 41.6% of participants were early ENDS quitters and stable cigarette users; 14.8% early ENDS quitters and gradual cigarette quitters; 14.6% stable ENDS users and stable cigarette users; 11.2% stable ENDS users and gradual cigarette quitters; 10.3% early ENDS quitters and early cigarette quitters; and 7.4% stable ENDS users and early cigarette quitters. Cigarette and ENDS use frequency, nicotine dependence, cannabis use and other non-combusted tobacco product use predicted trajectory group membership (p values <0.05).
CONCLUSIONS: Most dual users maintained long-term cigarette smoking or dual use, highlighting the need to address cessation of both products. Continued monitoring of trajectories and their predictors is needed, given ongoing changes to the ENDS marketplace.
Trajectories of ENDS and cigarette use among dual users: Analysis of waves 1 to 5 of the PATH Study
Tobacco Control, 33 (e1), e62-e68. doi: 10.1136/tc-2022-057405.