Estimating E-Cigarette Use Prevalence among US Adolescents UsingVaping-Related Online Search Trends

Abstract

Introduction: Adolescent e-cigarette use is a developing phenomenon. Greater surveillance of underage use is necessary to inform e-cigarette policy and mitigate adolescent e-cigarette use. Accurate prevalence estimates for adolescent e-cigarette use are provided by large national surveys. However, these surveys are costly and provide only annual estimates. To obtain more affordable estimates faster and more frequently, novel methods are required. Methods: Online search term popularity data were taken from Google Trends. Interest in vaping-related search terms were followed monthly from January 2011 to November 2020. Time-lagged zero-normalized cross-correlations were performed between the Google data and current (past 30 day) high-school e-cigarette use prevalence estimates from the National Youth Tobacco Survey (NYTS). The search interest data were then calibrated to the NYTS data to estimate adolescent e-cigarette use prevalence using online searches. Results: Maximum correlation coefficients of 0.979 for “vapes” and 0.938 for “vape” were obtained when search interest lagged use prevalence by one month, and 0.970 for “vape pen” when the lag was two months (p < 0.001 for all). Calibrating the search term data to NYTS provided a high-school current e-cigarette use prevalence estimate of 12.1–18.4% for November 2020, suggesting adolescent use of e-cigarettes has continued to decline since the NYTS estimate of 19.6% for January–March 2020. Conclusions: Online search trend data may provide reasonably reliable and more frequent estimates of adolescent e-cigarette use prevalence at substantially lower costs than traditional surveys. Such additional data may help to assess immediate impacts of policies and events.