Advancing youth tobacco surveillance with novel methods.

Abstract

Tobacco surveillance in the United States is robust but cannot be quickly modified to capture newly identified tobacco products or behaviors. We present an example of a rapidly deployed nonprobability survey using social media recruitment that collected data on rapidly changing tobacco use behaviors. We recruited 15- to 17-year old current vapers from NY, USA, using targeted social media advertisements to complete the New York Adolescent Vaping Survey (NY AVS), which asked about vaping behaviors not addressed in existing probability surveillance surveys. We used the New York Youth Risk Behavior Survey (NY YRBS) to apply calibration weights to ensure that the distribution of the demographic characteristics accurately reflected the population distribution. We found systematic differences in demographic variable distributions between the probability-based NY YRBS and the convenience sample of the NY AVS that were reconciled in the weighting calibration. We found no statistically significant differences between the NY YRBS and NY AVS estimates after calibration for two outcome variables of interest. Recruiting a sample of adolescents using social media advertising to conduct a rapid survey on vaping provided valuable data that complemented traditional surveillance surveys; this approach could be used to fill future knowledge gaps in youth tobacco surveillance.