Electronic cigarette usage patterns: a case study combining survey and social media data

J Am Med Inform Assoc. 2019 Jan 1;26(1):9-18. doi: 10.1093/jamia/ocy140.

Abstract

Objective: To identify who were social media active e-cigarette users, to compare the use patterns from both survey and social media data for data triangulation, and to jointly use both datasets to conduct a comprehensive analysis on e-cigarette future use intentions.

Materials and methods: We jointly used an e-cigarette use online survey (n = 5132) and a social media dataset. We conducted analysis from 3 different perspectives. We analyzed online forum participation patterns using survey data. We compared e-cigarette use patterns, including brand and flavor types, ratings, and purchase approaches, between the 2 datasets. We used logistic regression to study intentions to use e-cigarettes using both datasets.

Results: Male and younger e-cigarette users were the most likely to participate in e-cigarette-related discussion forums. Forum active survey participants were hardcore vapers. The e-cigarette use patterns were similar in the online survey data and the social media data. Intention to use e-cigarettes was positively related to e-liquid ratings and flavor ratings. Social media provided a valuable source of information on users' ratings of e-cigarette refill liquids.

Discussion: For hardcore vapers, social media data were consistent with online survey data, which suggests that social media may be useful to study e-cigarette use behaviors and can serve as a useful complement to online survey research. We proposed an innovative framework for social media data triangulation in public health studies.

Conclusion: We illustrated how social media data, combined with online survey data, can serve as a new and rich information source for public health research.

Publication types

  • Research Support, N.I.H., Extramural
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adolescent
  • Adult
  • Consumer Behavior / statistics & numerical data
  • Datasets as Topic
  • Electronic Nicotine Delivery Systems / statistics & numerical data*
  • Female
  • Humans
  • Internet
  • Logistic Models
  • Male
  • Middle Aged
  • Public Health Surveillance / methods
  • Social Media*
  • Surveys and Questionnaires*
  • Young Adult