To what extent does time spent on social media influence adolescent use of cigarettes and e-cigarettes: a longitudinal analysis of the UK Millennium Cohort Study

Abstract

BackgroundWe estimated the effect of social media use on adolescents aged 14 years and risk of cigarette, e-cigarette, and dual use when aged 17 years. Data was from the UK Millennium Cohort Study, which followed up approximately 19 000 children born between 2000 and 2002. MethodsRepresentative longitudinal data were collected at ages 14 years and 17 years. Directed acyclic graphs identified confounders (eg, demographics, mental health, in-person interactions, cognitive ability, risk-taking, antisocial behaviour, previous or current parental cigarette or e-cigarette use, and socioeconomic circumstances). The relationship between self-reported social media use per weekday (reference category: from 1 min to under 30 mins) and cigarette, e-cigarette, and dual use was examined using longitudinal analysis to estimate adjusted odds ratios (OR) or relative risk ratios (RRRs). A complete case sample was used; weights accounted for sample design and attrition. This study was a secondary data analysis of the UK Millennium Cohort Study (prospective longitudinal study). Ethical approval was received from a Research Ethics Committee at each study sweep. FindingsIn total, 6234 individuals (168 314 observations) were included. 5778 (92·7%) reported social media use, 1730 (27·8%) cigarette use, 1389 (22·3%) e-cigarette use, and 479 (7·68%) dual use. Social media use was associated with all outcomes in a dose-response manner. For cigarette use, ORs increased from 1·67 (95% CI 1·26–2·21) for 30 mins to 1 h, to 3·09 (2·43–3·91) for 2 h or longer of social media use. For e-cigarette use, ORs increased from 1·90 (1·41–2·55) for 30 mins to 1 h, to 3·34 (2·60–4·28) for 2 h or longer of social media use. For dual use, RRRs increased from 1·91 (1·16–3·15) for 30 mins to 1 h, to 4·26 (2·81–6·46) for 2 h or longer of social media use. For e-cigarette and dual use, associations were stronger for males than for females; the opposite was found for cigarette use. InterpretationAfter accounting for observed confounders and reverse causality, our findings suggest social media use, although only measured at one point in time, is associated with increased risk of cigarette, e-cigarette, and dual use. The greatest risk was observed in those who used social media for 2 h or longer. Given the potential health harms of social media use, guidance for parents and caregivers about safe social media use and regulation on time spent on social media is required. FundingMedical Research Council, Chief Scientist Office, Wellcome Trust.