In this chapter, we provide an overview of our key findings, and locate these findings within the rapidly developing literature on young people’s e-cigarette and tobacco use. The primary aim of our study was to evaluate impacts of TPD on young people’s e-cigarette use. However, in line with our evaluation and integration framework (see Chapter 2, Figure 2), we present findings chronologically, beginning with understanding the context and role of e-cigarettes before the TPD regulations, moving into implementation, before focusing on mechanisms of change and impacts. This frames impacts of the legislation as the end of the story the integration of our multiple data sources enables us to tell. For consistency with the presentation of our results chapters (see Chapters 3–5), we follow the same ordering in discussing our findings in this chapter.
The role of e-cigarettes in renormalising smoking prior to the implementation of the Tobacco Products Directive regulations
In our study, we found little evidence of change in decline for ever use of tobacco during the period when e-cigarettes were gaining popularity as cessation devices16 and young people’s experimental e-cigarette use was increasing.39 There was a slight slowing in the decline in regular smoking post 2010; however, this was mirrored in other substances, reflecting a broader slowing of decline in adolescent substance use behaviours, rather than being specific to tobacco. Measures of ever use for cannabis and alcohol indicated a slowing in decline while equivalent measures of tobacco did not, suggesting that disruptions to declines in tobacco lagged behind other substance use behaviours. We also found that young people’s attitudes had hardened against smoking, during the time that e-cigarette use was gaining traction in the UK.
Although at the time of initial publication our analyses was one of the first of its kind, numerous studies have since concurred with these findings across varying international contexts. Studies from Taiwan55 and New Zealand56 find that young people’s smoking rates fell as fast, or faster, during the emergence of e-cigarettes. An analysis of NYTS data in the USA found continuing decline in tobacco use following the emergence of e-cigarettes,57 with ever use declining at a faster rate post 2011. In Italy, where tobacco smoking remains high compared with other European countries,107 data from the Global Youth Tobacco Survey in 2010, 2014 and 2018 showed a decline in ever smoking among young people aged 13–15 years (from 46% to 41%) and smoking in the past 30 days (from 8% to 5%). From the UK, data from the YTPS (see Appendix 2, Figures 10 and 11) also finds a continued decline in tobacco use and perceived norms during the emergence of e-cigarettes.
FIGURE 10 Changes in tobacco use and perceived smoking prevalence within YTPSs. (Figure not reproduced here; see the original article.)
FIGURE 11 Percentage of ever and regular e-cigarette users in 2014 and 2016 within YTPSs. (Figure not reproduced here; see the original article.)
Young people’s perceptions of e-cigarettes and the normalisation of vaping
In group interviews prior to the full implementation of the TPD regulations, young people tended to differentiate between e-cigarettes and tobacco, preferring the word ‘vapes’ to ‘e-cigarettes’. Consistent with similar qualitative research with younger children,62,63 young people tended to interpret adult use of e-cigarettes as an attempt to quit smoking, rather than as an endorsement of smoking behaviour. Disapproval of smoking remained strong, although approval of e-cigarette use by young people was more nuanced. Experimental and social use in contexts such as performing tricks at parties were largely approved of, but regular use outside these social contexts was largely disapproved of, and was seen as ‘something the smokers do’. Our qualitative findings of dual acceptance of experimental vaping but disapproval of regular use have been observed in international qualitative studies.108 Similar to a number of previous qualitative studies,36,37 young people in our sample emphasised flavours as key drivers of e-cigarette use. The internet was described as playing an important role in young people’s exposure to e-cigarettes both in terms of perceived access to devices and media exposure to use, as many young people perceived that peers bought devices online and described online trick culture as forming a key driver of social use of e-cigarettes, particularly among boys.108 Hence, although there was no evidence of the renormalisation of smoking, initial qualitative data provided evidence that some forms of e-cigarette use itself were becoming normalised.
Stakeholders views on the regulation of tobacco and e-cigarettes
Long-standing declines in young people’s smoking were largely attributed by policy stakeholders to increasingly strong regulation of tobacco. Although views on e-cigarettes are often characterised into binary pro and anticamps,31 policy stakeholders perceptions of likely impact on public health were nuanced. Stakeholders described simultaneously positive views about the potential role of e-cigarettes in smoking cessation, alongside anxieties about potential harms of e-cigarettes for young people, differing to some extent in the emphasis placed on risks and benefits, rather than occupying binary positions. Within the TPD, marketing restrictions and moves to notification of new products were largely seen as proportionate in achieving goals, such as limiting appeal to non-smokers. However, device regulations were seen as potentially needing some refinement going forward, with some concerns about the evidence base behind elements such as the nicotine warning86 and refill size restrictions. Stakeholders argued that reviews following exit from the EU might offer opportunities to refine elements of the TPD regulations, but that departure from the EU was unlikely to substantially impact the TPD regulations or the trajectory of UK tobacco control action more broadly. Although elements of e-cigarette marketing were regulated via the TPD, concerns have continued to be expressed that e-cigarettes continue to be marketed to young people via their packaging, and that loopholes enable companies to give free samples of e-cigarettes to those underage, so long as these are not being sold.109 Hence, there are a number of key areas in which research into the refinement of existing regulations may be warranted.
Interviews with specialist and non-specialist e-cigarette retailers highlighted differences in awareness and perceptions of TPD legislation. Non-specialist retailers were less aware of regulations and were more ambivalent about e-cigarettes. Many specialist retailers acknowledged the importance of some regulation to limit appeal of e-cigarettes by young people, but voiced anxieties about the direction of travel for regulation. In particular, there were concerns that TPD signalled a movement towards regulating e-cigarettes as if they were ‘the same as tobacco’. Although EU regulations do not go so far as Food and Drug Administration communications in classing e-cigarettes themselves as tobacco products,60 they are, nevertheless, regulated under a broader banner of a ‘tobacco products’ directive. There was a common view that moving further towards regulating e-cigarettes as if they are tobacco might undermine their relative advantage over tobacco, leading to public perception of e-cigarettes and tobacco as the same. Retailers expressed suspicion that tobacco companies had a role in shaping e-cigarette regulation to reduce the relative advantage of a competing product. Retailers commonly positioned themselves as advocates of cessation110 and, hence, potential allies of public health,111 often expressing opposition to the tobacco industry and its associated interests. Nevertheless, a small number of retailers described reliance on tobacco industry actors for knowledge about the TPD because of a perceived lack of communication from authorities. Involvement of tobacco representatives in supporting retailers’ compliance with regulations resonates with findings from Scotland on implementation of plain packaging regulations.112 TSOs in our study also indicated that a lack of resource to support or enforce TPD regulations, and limited consistent communication for retailers on what was TPD compliance, meant that trade bodies stepped in to fill this gap. The relationship between the tobacco and e-cigarette industry has become increasingly complex, as tobacco companies have invested in e-cigarette companies or developed their own products.113 In some cases, it is possible that limited official communication on expectations of the TPD created a vacuum that enabled industry actors to further alliances with elements of the e-cigarette industry.
In line with findings from a post-implementation review in England,114 compliance with TPD regulations was largely seen as having been good by most stakeholders, and in our audits. However, consistent with young people’s perceptions that e-cigarettes were commonly accessed by peers via online purchase, additional perceived unintended consequences with particular implications for young people included displacement of sales of unregulated products onto the internet.
Young people’s interactions with elements of Tobacco Products Directive regulations and change in vaping perceptions
Following full implementation of the TPD regulations, young people reported limited interaction with e-cigarettes within their packaging and, hence, little exposure to nicotine warnings. Although an increased proportion of young people perceived e-cigarettes and tobacco as equally harmful in the years following the TPD regulations, given the apparently limited exposure to warning labels, the EVALI outbreak may offer a more plausible explanation for this sudden change in risk perceptions among young people.73 In 2018 interviews, e-cigarettes were seen by many young people as a fad that was beginning to lose its appeal, a finding that has been observed in international contexts independent of the TPD. Drawing on data from Norway, Tokle115 highlighted the similarity of e-cigarettes and ‘fidget spinners’, which was also cited by young people in our interviews. In qualitative interviews, most young people recalled little exposure to marketing, and it is plausible that marketing restrictions played a role in reducing the ability of industry to maintain young people’s interest in e-cigarettes.83 However, by contrast to interviews, survey data indicated that most young people reported that they had seen some form of advertising in the past month, with point of sale the most common source. One recent study, which compared exposure to e-cigarette marketing among 16- to 19-year-olds in England, Canada and the USA from 2017 to 2019, found that exposure remained common and increased in all countries.116 However, this occurred to a far smaller extent in England, perhaps indicating partial displacement of marketing to unregulated channels, although not to a sufficient extent to overcome effects of restrictions relative to less regulated environments.
Across interviews, the internet featured strongly as a perceived source of exposure to e-cigarettes. Notably, however, according to survey data, the internet was the only source in which young people’s exposure decreased in the years following TPD regulations, and this perhaps reflected diminishing interest in e-cigarettes, with fewer young people actively seeking e-cigarette-related material on the internet. Despite concerns expressed by both young people and retailers, quantitative data from Wales after the implementation of TPD regulations indicated that few young people reported online purchase, with quantitative data providing little support for the hypothesis that young people had switched to online purchase following new regulations. The most common routes of supply included via peers and other informal supply chains, with a minority of young people obtaining e-cigarettes directly from retailers. This contrasts with findings from the International Tobacco Control four-country study, in which young people were equally likely to report obtaining e-cigarettes via informal or commercial supply routes, although, notably, the study focused on 16- to 19-year-olds and, hence, included a significant number of young people who could legally buy e-cigarettes themselves.117
Change in young people’s use of e-cigarettes following Tobacco Products Directive regulations
As reported in the years prior to the TPD regulations,39 and consistent with qualitative reports of widespread approval of experimentation with e-cigarettes, experimentation with e-cigarettes had been growing rapidly, including among never smokers. However, our primary statistical analyses of data from Wales found that, at both short- and long-term follow-up, the odds of ever e-cigarette use decreased by 4% per month following the TPD regulations, relative to the baseline trend. Although there was some evidence of emerging inequality in short-term analysis, with smaller changes in minority ethnic groups, for example, these differences did not follow through to long-term analysis, where very similar estimates were observed for all subgroups. Analyses of data from England and Scotland were consistent with the pattern observed in Wales of growth in e-cigarette use prior to TPD regulations and, subsequently, plateauing or declining thereafter. Regular e-cigarette use also followed a pattern of pre-regulation increase followed by post-regulation plateauing or decline. Consistent with qualitative reports of less widespread approval of regular use, regular use remained rare among never smokers, at 1% or lower in all countries and years. Hence, concerns regarding regular use of e-cigarettes acting as a gateway to nicotine addiction for a new generation of tobacco-naive young people are not reflected in data from Great Britain,39 pre or post regulation.
Some international divergence has emerged in relation to young people’s use of e-cigarettes, and it is likely that this reflects differences in regulatory landscapes.77 In particular, our findings contrast with a large growth in e-cigarette use by young people around the same time in the USA.67,76 This was largely attributed to high-strength devices whose entry to UK markets was delayed by the TPD and, because of the TPD, were limited to lower levels of nicotine strength than US versions. Our findings on the plateauing of e-cigarette use are, however, largely consistent with other UK national surveys monitoring trends in e-cigarette use and smoking overtime. Hammond et al.69 found no change in vaping in England in the post-TPD period from 2017 and 2018 among slightly older adolescents, despite substantial growth in the same period in both the USA and Canada. In ASH smoke-free Great Britain surveys118,119 of 11- to 18-year-olds, the percentage of young people having used e-cigarettes had increased over time from 4% in 2013 to around 10% by 2015, but the percentage remained relatively stable thereafter. Although young people’s accounts indicated that interest in e-cigarettes was waning, and interaction with many elements of the TPD regulations was limited, it is plausible that marketing restrictions coupled with new reporting requirements and product regulations inhibited the ability for new products to be introduced rapidly and promoted in ways that appeal to young people.116 More recent data from the USA indicate that although young people’s e-cigarette use did initially grow rapidly, it is now declining.70
Changes in young people’s use of tobacco after the implementation of Tobacco Products Directive regulations
Tobacco use continued to decline during the emergence of e-cigarettes up to 2015, which, as our data confirms, was also the period of most rapid growth in young people’s experimentation with e-cigarettes. However, the most recent data sets indicate some plateauing in rates of both ever and regular smoking across nations, during a period beyond regulation and in which young people’s experimentation with e-cigarettes was no longer growing. The only nation in which there was some indication of continued decline in smoking was England. As noted elsewhere,26 however, the English survey is smaller and has seen more substantial drop-offs in its response rates than have the Welsh or Scottish surveys and, hence, it is unclear whether this represents divergence between nations or an artefact of methodological differences.
The plateauing observed for tobacco use was also reflected in other substances, as has been reported elsewhere.120 Indeed, disruption to long-standing declines in tobacco use appears to have lagged somewhat behind those for other substances, with ever use of tobacco beginning to plateau more recently than equivalent measures of both alcohol or cannabis, which had already begun to plateau from 2011 to 2015. Recent data from the USA indicate that growing cannabis use may act as a driver of nicotine dependence and sustained tobacco use among young people, with growing co-occurrence of tobacco and cannabis use, and dependence strongest among combined users.121 Although there were insufficient data points to model continued change in attitudes towards smoking in post-legislation models, data from SALSUS and the SDDU survey indicate that the proportion of young people who report that it is OK to try smoking rose from 2015 to 2018 in Scotland and remained flat from 2016 to 2018 in England. Data published by the SDDU and SALSUS teams indicate that this change is also mirrored in recent increases in the acceptability of drug use and alcohol consumption,94,95 again indicating broad social trends in terms of attitudes towards substance use, rather than being specific to tobacco. Nevertheless, this supports a conclusion that the suite of tobacco regulation implemented in 2016, including the TPD regulations and plain packaging for tobacco cigarettes, may not have been sufficient to maintain or increase the downwards trajectory in young people’s use of tobacco. Further action will be needed to reach young people who continue to take up smoking if UK nations are to achieve goals of becoming smoke free by 2030.122 Given the clustered nature of adolescent risk behaviours, and their often shared trajectories over time, it is perhaps unlikely that the goal of becoming smoke free by 2030 will be achieved through tobacco-specific intervention alone and, instead, it may require an increased emphasis on engaging with the broader context of adolescent risk behaviour and identifying and modifying common determinants.123 One possible interpretation might be that there is an interaction between regulation of e-cigarettes and tobacco, with regulation of e-cigarettes leading young people who might have used e-cigarettes towards tobacco instead. Although such a conclusion is highly tentative, future research could usefully explore the effects of tobacco regulation on e-cigarette uptake, and vice versa. For example, following completion of this study, flavoured tobacco was banned in the EU as a final stage of the TPD; however, flavoured e-cigarettes remain available in many countries, including the UK. Conversely, flavoured e-cigarettes were regulated in the USA, although flavoured tobacco cigarettes remain available. Whether or not these respective legislative decisions lead young people to switch from flavoured tobacco or e-cigarettes to the alternative deserves attention.
Recommendations for future research
Recommendations for research, not ordered in terms of priority, include the following:
- Continued population-level monitoring of trends over time in young people’s tobacco cigarette and e-cigarette use and perceptions remains vital.
- Harmonised international data sets may be particularly valuable in enabling changes following new policies to be compared with an external comparator.
- As regulatory divergences have emerged between nations, international comparisons of recent trends in e-cigarette use and smoking may provide insight into impacts of different regulatory environments on young people’s smoking and e-cigarette use.
- There are high levels of confusion on the harms of e-cigarettes relative to tobacco among young people. Future research should seek to understand how to reinforce the role of e-cigarettes as cessation devices for smokers, while discouraging use by non-smokers and acknowledging ambiguity and unknown long-term harms.
- Although young people’s e-cigarette use is no longer continuing to grow, a substantial proportion of young people continue to use e-cigarettes. Hence, research into how exposure can be further reduced, without reducing the effectiveness of e-cigarettes as cessation aids for smokers, remains important.
- Given the prominent role of the internet and video-sharing platforms in young people’s exposure to e-cigarettes, further understanding of how young people interact with these platforms is an important area for future research.
- Although much has been achieved in reducing young people’s use of tobacco, there is now tentative evidence from our secondary analysis that this decline may be stalling, particularly in the devolved nations, and this is reflected in other substances and is not unique to tobacco. Hence, it is vital to understand how tobacco-specific interventions, and interventions that address common causes of multiple adolescent risk behaviours, can enable further reduction in the prevalence of tobacco use among young people.
Our study demonstrates the value of complex mixed-method designs in understanding the implementation, mechanisms, context and effects of policies within complex and ever-changing contexts. Engagement with rapid developments in theory-driven and complexity-informed evaluation approaches might be valuable in future similar studies.
Strengths and limitations
This study represents, one of the first robust evaluations of the introduction of e-cigarette regulations, as well as being one of the first studies to robustly examine the renormalisation hypothesis. The quantitative elements of the study benefited from use of large historical data sets across all three UK nations. Quantitative findings are contextualised and explanations for estimated trends generated through qualitative work with a diverse range of stakeholders.
Nevertheless, there were some important limitations. Although there are few viable alternatives for measuring substance use behaviour at scale, outcome measures rely on self-report. The outcome measures are, therefore, subject to social desirability bias, which may change in magnitude over time as behaviours become more or less socially unacceptable. There were differences in timing of surveys across nations, and in questions used. Although a limitation insofar as these differences make direct comparisons across countries challenging, the fact that similar changes in trend are observed across nations despite these differences enhances confidences that these are real social changes. The infrequent nature of some data collections and the rapid escalation of tobacco control measures through the study period meant that adjusting models for all potential interventions as confounders was not feasible. Although our primary statistical analysis of Welsh data was adjusted for school-level clustering, for many secondary analyses this was not possible because of the removal of school identifiers from archived data sets. Hence, the statistical significance of some changes in trend may be overstated because of inability to inflate standard errors to account for non-independence of young people within schools. Given that the TPD regulations were supranational legislation and no ‘similar’ country without regulations but comparable historical data was identified, we had no external control group. Hence, we cannot rule out the possibility that plateauing would have occurred without legislation. The limited pre-legislation data points for our primary analysis is another limitation, which is unavoidable where evaluating emerging public health issues for which measurement has been only recently integrated into surveys. Capturing and assessing trends in e-cigarette use over a longer time period provides more power in statistical analyses, but also risks weakening causal inference by increasing the likelihood of events external to the legislation occurring and causing change in trends.99 Notably, as the study approached the long-term data collection, the EVALI outbreak occurred in the USA, which might offer an alternative explanation for changes in trend for e-cigarette use. The consistency of change in trend in short- and long-term analysis, however, provides some confidence that change occurred prior to EVALI and continued thereafter, rather than occurring after or being further catalysed by EVALI. This conclusion is reinforced by comparison against simpler analyses of data from England and Scotland, which in both instances extended to only 2018 (i.e. prior to the EVALI outbreak), yet showed clear evidence of slowed growth in e-cigarette use in the years beyond TPD regulations. Recent research has indicated that young people’s tobacco use increased during the first COVID-19 lockdown in the UK,124 likely as a means of dealing with stresses of the pandemic. Although future rounds of data will enable us to monitor trends in the longer term, our ability to attribute these trends to the TPD regulations, rather than to more recent events, such as the COVID-19 pandemic, will diminish with time.
There were also some deviations to protocol. Although we had planned to include tracking of mentions of e-cigarettes on social media, sites such as Facebook and Twitter, our qualitative data indicated that other forms of social media were more influential, such as video-sharing websites. Hence, although the internet was cited as an important source of exposure, we were not able to quantitatively track this in our study. Some external data collections in England and Scotland did not take place, meaning that some analyses are based on fewer data points than in our original protocol. In particular, for secondary analysis of smoking trends, although there was a longer baseline time series than for e-cigarettes, the reduced frequency of some surveys meant that there was fewer data available for post-implementation analyses than we had hoped, and findings regarding plateauing in young people’s smoking since 2016 should be treated with caution until further rounds of data are available.
Although we specified a number of secondary analyses to examine the specificity of changes in use of e-cigarettes and tobacco, it is perhaps possible that these variables might themselves have been impacted by external legislation brought in alongside the TPD regulations. Energy drink use, for example, might have been impacted by the recent sugar tax announced around the same time as the TPD regulations were implemented. It is possible that the observed co-occurring changes in trend for tobacco and cannabis use might have reflected increased use of e-cigarettes as devices to support use of cannabis.
Given that the study started during the transitional phase of TPD implementation, it was not possible to collect qualitative data prior to this and, hence, initial data on young people’s perceptions of the recent emergence of e-cigarettes might have already began to be influenced by new regulations. The choice to involve teachers in selection of pupils for interviews, although pragmatic when sampling from within schools, may have led to some biases in who was put forward for qualitative interviews. In 2016, when this study was designed, we made a decision to conduct single-gender group interviews because of the gendered nature of historical smoking trends. However, our own survey data have since illustrated the extent to which young people who do not identify as male or female, and who may have been excluded by this decision, are affected by tobacco use.96 Were we to design the study again, we would not repeat this decision. We also collected somewhat limited demographic information from young people and a broader range of demographic information would have allowed us to better understand the diversity of views represented within our data.
As reflected in updated MRC guidance for development and evaluation of complex interventions125 (published between completion of this study and the submission of this report), recent years have seen rapid developments in the accommodation of theory-driven approaches to evaluation, such as realist evaluation,126,127 into mainstream evaluation practice and has increased the influence of complexity-informed approaches and systems perspectives.88,128 Our integrated mixed-method design is a strength. However, it is possible that future similar studies may benefit from more explicit engagement with these advances, enabling fuller understandings of cause and effect within complex open systems where co-occurring interventions and epidemiological events collide.