Associations of social media use with smoking and e-cigarettes: a national longitudinal study

Abstract

Background: Social media use is high among children and young people and might influence health behaviours. We examined social media use and use of tobacco and e-cigarettes in the UK.

Methods: We used data from participants aged 10-25 years from the UK Household Longitudinal Study (January 2015-January 2022). Participants were asked: "On a normal weekday, that is Monday to Friday, how many hours do you spend chatting or interacting with friends through a social website or app like that?". Specific social media platforms were not specified. Responses were none, less than 1 h, 1-3 h, 4-6 h, 7 h or more. Outcomes were current tobacco smoking and e-cigarette use. Generalised Estimating Equation (GEE) logistic regression models investigated associations of social media use with tobacco and e-cigarette use, and fixed effects analyses investigated changes in social media use with uptake of both products. Models included possible confounders such as age, sex, household income, ethnicity (White vs non-White) and use of tobacco or e-cigarettes by others within the home. All participants gave written informed consent.

Findings: The analytic sample included 10 808 participants with 27 962 observations (mean age 15·7 years [SD 3·8], 5080 [47%] male, 5728 [53%] female, and 7868 [73%] White). Current tobacco smoking was reported at one or more timepoints by 929 (8·6%) participants, and current e-cigarette use by 270 (2·5%) participants. In adjusted GEE models, all levels of social media use were associated with greater odds of current smoking than no use. This association was particularly apparent at higher levels of use adjusted odds ratio [aOR] 3·11, 95% CI 2·41-4·03 for ≥7 h use vs no use), with similar associations for e-cigarettes (aOR 3·04, 2·11-4·40 for ≥7 h use vs no use). Fixed effects analyses also found increased use of social media to be associated with increased uptake of both products (eg, changing to using social media for ≥7 h/day was associated with >2 times the odds of taking up tobacco smoking [aOR 2·33, 1·28-4·24]).

Interpretation: These analyses suggest an association between social media use and e-cigarette and tobacco use. Potential pathways include promotion of these products on social media. Further research with details on specific platforms would be useful as well as with longer follow-up time.

Funding: Cancer Research UK.