Longitudinal e-Cigarette and cigarette use among US Youth in the PATH Study (2013-2015)

Cassandra A. Stanton, Maansi Bansal-Travers, Amanda L. Johnson, Eva Sharma, Lauren Katz, Bridget K. Ambrose, Marushka L. Silveira, Hannah Day, James Sargent, Nicolette Borek, Wilson M. Compton, Sarah E. Johnson, Heather L. Kimmel, Annette R. Kaufman, Jean Limpert, David Abrams, K. Michael Cummings, Maciej L. Goniewicz, Susanne Tanski, Mark J. TraversAndrew J. Hyland, Jennifer L. Pearson

Research output: Contribution to journalReview articlepeer-review


Background: Evidence is accumulating that youth who try Electronic Nicotine Delivery Systems (ENDS, e-cigarettes) may go on to try cigarettes. This analysis examines the bidirectional patterns of ENDS and cigarette use among US youth over one year and uses propensity score matching (PSM) to examine frequency of ENDS use on changes in cigarette smoking. Methods: Our analysis included 11 996 participants who had two waves of available data (Wave 1 [W1] 2013-2014; Wave 2 [W2] 2014-2015) drawn from the longitudinal Population Assessment of Tobacco and Health Study. Cross-sectional weighted prevalence estimates are reported for cigarettes and ENDS. We used PSM to estimate the likelihood of ENDS use at W1 and to draw matched analytic samples, then used regression (logistic or linear) models to examine the effect of W1 ENDS use on W2 cigarette smoking. All statistical tests were two-sided. Results: In weighted analyses, 69.3% of W1 past-30-day cigarette smokers exhibited past-30-day smoking at W2; 42.2% of W1 past-30-day ENDS users were using ENDS at W2. W1 ever use of either product was similarly associated with W2 new use of the other product. Unweighted PSM models indicated W1 cigarette-naïve ENDS use was associated with W2 ever-cigarette smoking (n = 676; adjusted odds ratio = 3.21, 95% confidence interval [CI] = 1.95 to 5.45, P <. 001); W1 ever-ENDS use did not affect change in cigarette frequency at W2 (n = 1020, beta = 0.31, 95% CI = -0.76 to 1.39, P =. 57); 1-5 days ENDS use compared with ever, no past-30-day ENDS use was associated with a statistically significant decrease of W2 smoking days (n = 256, beta = -2.64, 95% CI = -4.96 to -0.32; P =. 03); and W1 6+ day ENDS users did not show a decrease in frequency of cigarette smoking. Conclusions: Ever-ENDS use predicts future cigarette smoking, and frequency of ENDS use has a differential impact on subsequent cigarette smoking uptake or reduction. These results suggest that both cigarettes and ENDS should be targeted in early tobacco prevention efforts with youth.

Original languageEnglish (US)
Pages (from-to)1088-1096
Number of pages9
JournalJournal of the National Cancer Institute
Issue number10
StatePublished - Oct 1 2019


  • Adolescent
  • Child
  • Electronic Nicotine Delivery Systems
  • Female
  • History, 21st Century
  • Humans
  • Male
  • Prevalence
  • Propensity Score
  • Public Health Surveillance
  • Smoking/epidemiology
  • Tobacco Products
  • United States/epidemiology
  • Young Adult

ASJC Scopus subject areas

  • Oncology
  • Cancer Research


Dive into the research topics of 'Longitudinal e-Cigarette and cigarette use among US Youth in the PATH Study (2013-2015)'. Together they form a unique fingerprint.

Cite this