Friends and family matter Most: a trend analysis of increasing e-cigarette use among Irish teenagers and socio-demographic, personal, peer and familial associations

In Ireland, in both the 2015 and 2019 waves of the ESPAD surveys, students were surveyed using a two-stage probability sampling technique: first at national level (a stratified random sample) and then at school level (systematic random sampling of classes). In the first stage at national level, a stratified random sampling of schools in Ireland was carried out. The schools were stratified according to geographic region and school type (secondary, vocational, community/comprehensive), religious affiliation (Roman Catholic, Church of Ireland, inter-denominational), gender (all-boys, all-girls, mixed), and school-level disadvantage status (Delivering Equality of Opportunity in Schools: DEIS vs. non-DEIS).

In the second phase, within schools in the sample, a systematic random sampling technique was used to identify the target sample of those born in 1999 (for the 2015 wave) and 2003 (for the 2019 wave), to produce a nationally representative sample of students who turned 16 years old during the year of data collection, as per ESPAD criteria. Only students present on the day of the survey were sampled, yielding a response rate of 86% for the 2015 sample and 80% for the 2019 sample. The total valid sample for this study was 3421 16-year-olds, comprising 1472 students (born in 1999) in the 2015 sample, and 1949 students (born in 2003) in the 2019 sample. Completed survey data were entered manually into SPSS v22 as they appeared in the survey. Full accounts of the sampling and data cleaning procedure have been reported elsewhere [2, 22]. All anonymised data are retained on a secure server and accessible through request to the corresponding author.

A full list of the questions, and answer categories, from which variables were drawn is in Additional files 1 & 2 (ESPAD questionnaires 2015 and 2019).

First, Pearson’s chi-square test was used to compare the demographic, personal, peer and familial variables between 2015 and 2019 and to determine whether differences in the variables between the two waves were statistically significant (Table 1). Then, e-cigarette ever-use and current-use were examined using a multivariable logistic regression model with e-cigarette ever-use (Table 2) and current-use (Table 3) as the dependent variable, analysed for all and by gender. To assess potential multicollinearity, all variables in the study were adjusted by the dependent variable (e-cigarette ever- and current-use) using Spearman’s correlation coefficient and variance inflation factor (VIF) as appropriate between variables, and a VIF < 5 was used to detect multicollinearity. All analyses and results are presented as adjusted odds ratios (AOR) and 95% confidence intervals. A p-value of .05 was used as a cut-off for significance. All statistical analyses were conducted using STATA version 16.

Prevalence of e-cigarette ever-use and current-use among Irish 16-year-olds rose significantly between 2015 and 2019. Prevalence of both is significantly higher for boys than for girls in both waves (Fig. 1). The overall prevalence of ever-use of e-cigarettes increased from 23 to 37%, with boys increasing from 26 to 43.2% and girls from 19.9 to 31.6%. Current-use increased overall from 10.1 to 18.1%, from 11.6 to 22.9% in boys and from 8.6% to 13.6 in girls. Most of the other factors included in our model are statistically unchanged between 2015 and 2019 (Table 1).

Pearson’s chi-square analyses (Table 1) of risk factors associated with e-cigarette use (socio-demographic, personal, peer and familial characteristics) show very few statistically significant differences (p < .05) between the 2015 and 2019 ESPAD waves.

Notable findings concern the relationship with tobacco use when first using e-cigarettes and reasons reported for trying e-cigarettes. In 2015, 33.3% had never used tobacco when they first used e-cigarettes and this rose to 66.7% in 2019. Also, the number who said that they were regular tobacco users when they first used an e-cigarette decreased from 14.9 to 9% between waves.

Reasons for trying e-cigarettes also changed, with significantly fewer reporting trying e-cigarettes in order to quit smoking in 2019 (3.4%) than in 2015 (17.3%). Trying e-cigarettes out of curiosity increased from 63.1% in 2015 to 66.3% in 2019, and those who used it because of friends increased from 21.4% in 2015 to 28.8% in 2019.

Table 2 shows the results of the multivariable logistic regression analysis of e-cigarette ever-use in the 2015 and 2019 ESPAD Ireland surveys, reported as adjusted odds ratios (AOR, 95% CI) for the total sample, and separately for boys and girls.

The odds of ever-using e-cigarettes increased significantly for the whole sample between 2015 and 2019 (AOR 2.29, 95% CI:1.89–2.78), and the increase was more pronounced for girls (AOR 2.67, 95% CI:2.02–3.54) compared with boys (AOR 2.04, 95% CI:1.55–2.68). Use of combustible tobacco (both ever and current smoking) was associated with increased odds of ever-using e-cigarettes, with ever-smoking having higher odds for girls (AOR 1.56, 95% CI:1.12–2.18) and current smoking higher odds for boys (AOR 2.60, 95% CI:1.71–3.93).

Parental monitoring was significantly associated with ever-use for all (AOR 3.96, 95% CI:2.54–6.18), and was more important for boys (AOR 5.42, 95% CI:2.72–10.79) compared with girls (AOR 3.33, 95% CI:1.84–6.03). Truancy and being in a blended family were associated with increased odds for boys but not for girls.

Table 3 shows the results of the multivariable logistic regression analysis of e-cigarette current-use in the 2015 and 2019 ESPAD Ireland surveys, reported as adjusted odds ratios (AOR, 95% CI) for the total sample, and separately for boys and girls.

The odds of teenagers being e-cigarette current-users increased significantly between 2015 and 2019 (AOR 2.41, 95% CI:1.85–3.12); the increase was greater for girls (AOR 3.11, 95% CI:2.10–4.61) compared with boys (AOR 1.96, 95% CI: 1.37–2.82). Being a current smoker increased the odds of current-use e-cigarette (AOR 1.78, 95% CI:1.23–2.55), significantly so for boys (AOR 2.13, 95% CI:1.30–3.51). The more friends young people have who smoke, the greater their odds of e-cigarette current-use. Those who reported that “most or all” of their friends smoked had greater odds of e-cigarette current-use (AOR 5.45, 95% CI:3.65–8.14), and this was more pronounced for boys (AOR 5.90, 95% CI:3.31–10.52) than for girls (AOR 5.31, 95% CI:3.01–9.37).

As with ever-use of e-cigarettes, decreased parental monitoring was also associated with increased odds for teenagers’ current-use of e-cigarettes. The odds were significantly increased (AOR 4.48, 95% CI:2.83–7.11) in young people who say that their parents “usually don’t know” where they are on a Saturday night (reference group “parents always know”). Odds were higher for boys (AOR 5.50, 95% CI:2.85–10.61) than for girls (3.50, 95% CI:1.79–6.84).

From the outset, boys in our trend analyses were more likely to be both ever- and current- users of e-cigarettes. This is in line with many other studies [1, 20]. Various theories have been offered to explain gender and substance use including tobacco and e-cigarettes, such as Connell’s (2005) influential construct of hegemonic masculinity and how it puts men at risk of harmful health behaviours and consequences that can be destructive for them [29], including for teenage boys [30, 31], and Butler’s [32] consequential theory of gender performativity - that gender is not an essential, biologically determined quality or an inherent identity, but is repeatedly performed, based on, and reinforced by, societal norms, this repeated performance of gender being also performative - applied to smoking by women in Australia by Gilbert and colleagues [33]. They argued that smoking is “a gender act that can be internalised and which, when repeatedly performed by women in gender-appropriate ways, constructs a ‘feminine’ gender identity” [32, 33]. Such theories and how they relate to our findings on gendered e-cigarette use are outside the scope of our data. We raise them here to acknowledge that our findings have a broader and deeper context within discourses on gender and substance use [31].

Boys have higher prevalence of e-cigarette use but the rate of increase in this study is significantly greater for girls, and this was particularly pronounced for current-use, with the trend analysis showing girls having more than 50% higher odds (AOR 3.11, 95% CI 2.10–4.61) than boys (AOR 1.96, 95% CI 1.37–2.82) of being e-cigarette current-users in 2019 compared with 2015. This gendered pattern of substance use showing initial high male use, with female use subsequently over-taking that of males reflects historical patterns of women’s and men’s tobacco use, driven, in part at least, by the tobacco industry’s gendered marketing, and exploitation of social change and social disruption [34‐36], such as the post-war targeting of women by the tobacco industry “as an equality and freedom issue” [36]. The latter comprised advertising and marketing by the industry, specifically and successfully targeted to women and girls, a market identified as a large untapped lucrative reservoir [34‐36]. E-cigarette advertising and direct and covert marketing uses strikingly similar techniques to those used previously by the cigarette industry [37] - featuring young, attractive models, sponsorship of sports events and parties, product placement, and direct payments to social media influencers [37]. We add support to Kong et al. (2017) who observed that, while boys in the U.S. appear to have greater use of e-cigarettes, girls may be at increased risk if e-cigarettes are targeted to them “as it has been for cigarettes” and we join in calling for further research on gender differences in e-cigarette use, particularly in gendered rates of increase, and on the role of industry advertising and marketing, including the gendered nature of such activities on the internet [1]. We recommend that insights about gender, from emerging theories and historical developments such as those mentioned above, be incorporated into both policy-making and health education programmes that are intended to reduce children’s e-cigarette use.

We agree with O’Leary et al. (2019) that, while the state and use of social media are ever changing, the potential to use social media as a form of promotion for healthy behaviours, especially among adolescents, will continue to offer promise [38]. Thus, we extend to the domain of tobacco and e-cigarette use their call for education interventions for teenagers [38].

Also, we draw attention to findings from ourselves and others regarding the potentially different online worlds inhabited by teenage girls (social media platforms) and teenage boys (gaming platforms) that have been identified [2, 39]. This leads us to speculate, for example, that boys may be targeted through gaming platforms and that girls’ rapidly increasing e-cigarette use may be related to their greater social media use. The scope within these parallel gendered domains for targeted marketing of e-cigarettes by industry merits further research and we also support calls for regulatory action to prohibit sponsored e-cigarette content on social media platforms used by youth [40, 41].

The link between cigarette smoking and e-cigarette use has been well-established [15‐17] and our findings support this, but with gender differences. Girls who had ever-smoked had higher odds (AOR 1.56, 95% CI: 1.12–2.18) of ever-using e-cigarettes. Boys who were current smokers had more than twice the odds of being e-cigarette ever-users (AOR 2.60, 95% CI: 1.71–3.93). Thus, differences in experimentation and continuation of both smoking and e-cigarette use appear to be gendered, pointing to different characteristics between the cigarette-smoking and e-cigarette using populations or to gender differences that require further exploration. We lend some support to the findings of Creamer et al. (2021) that, regarding psychosocial risk factors for cigarette smoking, e-cigarette users do not fit the traditional risk profile of cigarette smokers, and also require further research [19].

Adolescent peer social networks have been found to be important for health behaviour choices, with health behaviour similarity found to be driven by homophilic social selection and/or social influence [42]. Studies of adolescent social networks, including online networks, suggest that friends’ online behaviours are a viable source of peer influence [23]. Those with more peers who smoke have much higher odds of being ever-users of e-cigarettes, and this pattern was particularly strong for boys. Peer smoking was similarly implicated in e-cigarette current-use and, again, gender differences showed a somewhat stronger influence of peers on boys than on girls in relation to current-use of e-cigarettes. A review of 26 studies examining adolescents’ susceptibility to peer pressure to engage in risky behaviours identified two primary trends: one, that adolescent males appear to be more susceptible to peer influences that encourage risk-taking behaviours; and the other, that there are no consistent gender differences [43]. McCoy and colleagues conjecture [43] that, as attitudes about gender-appropriate behaviour shift across historical time, it may be that male and female teenage experiences are becoming increasingly similar, for example in experiencing comparable levels of deviant peer pressure around substance use in particular and also that differences across types of risky behaviours may “even out”, causing gender differences to disappear.

Gender being an incomplete explanation of the observed differences in teenage e-cigarette use, we draw attention to intersectionality [44] as a promising framework to achieve new understandings of teenage tobacco use. Although intersectionality has been examined in relation to adult smoking cessation (e.g., [45, 46]), there is little or no research to date on teenage tobacco use (and consequent health inequalities) that captures the complexity of “multiple aspects of identity” [41], employing analyses at the intersections of, for example, categories of gender, race, class/SES, disability, sexuality, and religion. More complex conceptual analyses are needed to generate new insights into this emergent and increasing problem of new tobacco product use by young people.

As our findings provide further support for the many studies that have found that peer smoking influences teenage e-cigarette use [20, 47‐50], we suggest that health education interventions that take account of peer influences remain important, particularly as higher levels of peer e-cigarette use [51, 52] and favourable e-cigarette peer norms [53] have been found by other researchers to be related to higher odds of personal e-cigarette use.

Parental monitoring was a separately important factor in explaining teenagers’ e-cigarette use and we add further support to our previous ESPAD findings [54, 55] - as well as findings from other studies [22, 56] - showing that lack of parental monitoring remains a significant predictor for all illicit substance use in the best-fitting models. However, our finding contrasts with that of Fotiou et al. [49] who reported that low parental monitoring correlates with tobacco but not with e-cigarette use. Our finding about parental monitoring was also gendered, being more significant for teenage boys than for teenage girls.

We highlight an urgent need for health education programmes that address the increasing trend of teenage e-cigarette use and recommend that such programmes acknowledge the important influences of friends and families.

We report on comparable nationally representative samples of teenagers from 2015 and 2019, and note significant gender differences. However, the quantitative methodology does not allow for more in-depth understanding of these gender differences nor why they occur. Thus, in the Discussion section, we offer some possible theories that might be tested in future research to explain the now well-established gender differences in e-cigarette use. Also, longitudinal and/ or qualitative approaches would provide greater insight into teenage e-cigarette use and the associations with the increased risk observed over time.