Authors: John W. Ayers, Adam Poliak, Nikolas T. Beros, Michael Paul, Mark Dredze, Michael Hogarth, Davey M. Smith
Source: American Journal of Preventative Medicine, Volume: 67, Issue: 1, DOI: 10.1016/j.amepre.2024.01.016, June 2024
Type of Publication: Article in a Periodical
Abstract:
Introduction
The evidence hierarchy in public health emphasizes longitudinal studies, whereas social media monitoring relies on aggregate analyses. Authors propose integrating longitudinal analyses into social media monitoring by creating a digital cohort of individual account holders, as demonstrated by a case study analysis of people who vape.
Methods
All English language X posts mentioning vape or vaping were collected from January 1, 2017 through December 31, 2020. The digital cohort was composed of people who self-reported vaping and posted at least 10 times about vaping during the study period to determine the (1) prevalence, (2) success rate, and (3) timing of cessation behaviors.
Results
There were 25,112 instances where an account shared at least 10 posts about vaping, with 619 (95% CI=616, 622) mean person-days and 43,810,531 cumulative person-days of observation. Among a random sample of accounts, 39% (95% CI=35, 43) belonged to persons who vaped. Among this digital cohort, 27% (95% CI=21, 33) reported making a quit attempt. For all first quit attempts, 26% (95% CI=19, 33) were successful on the basis of their subsequent vaping posts. Among those with a failed first cessation attempt, 13% (95% CI=6, 19) subsequently made an additional quit attempt, of whom 36% (95% CI=11, 61) were successful. On average, a quit attempt occurred 531 days (95% CI=474, 588) after their first vaping-related post. If their quit attempt failed, any second quit attempt occurred 361 days (95% CI=250, 474) after their first quit attempt.
Conclusions
By aligning with standard epidemiologic surveillance practices, this approach can greatly enhance the usefulness of social media monitoring in informing public health decision making, such as yielding insights into the timing of cessation behaviors among people who vape.
Computer Science
