Causal Relationship Between ENDS Use and Subsequent Cigarette Initiation Among Adolescents: a Propensity Score Analysis Using Data of the PATH Study

• Activity Number: 523 - Contributed Poster Presentations: ENAR
• Type: Contributed
• Date/Time: Wednesday, July 31, 2019 : 10:30 AM to 12:20 PM
• Sponsor: ENAR
• Author(s): Shu Xu* and Bin Liu and Yifan Xu and jiarui he and Raymond Niaura and Donna L. Coffman
• Companies: New York University and New York University and New York University and new york university and New York University and Temple University
• Keywords: inverse propensity score weighting; marginal structural model ; average treatment effect; PATH

Abstract

Increasing e-cigarette among adolescents is a critical public health concern in the U.S.. The current study purses a causal inference of the association between the use of electronic nicotine products (ENDS). We carefully defined the average treatment effect as “What difference in the probability of cigarette use would we expect to observe if all adolescents in our study had used e-cigarette, compared to if all of them did not?” The current study includes 1072 adolescents from a national representative longitudinal data, the Population Assessment of Tobacco Use and Health study. A marginal structural model via inverse propensity score weighting assessed the association between ENDS and subsequent cigarette smoking. We included 94 covariates from Wave 1, ENDS exposure at wave 2, and cigarette use (outcome) at wave 3. We computed inverse propensity score weights, and assessed balance by computing standardized mean difference in each Wave 1 covariates across ENDS use groups. Missing data were imputed (m = 5) using MICE in the mice R package. Results implied that ENDS use was associated with greater risk for subsequent cigarette smoking initiation in the following year.