Some text in this chapter has been reproduced from Moore et al.101 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, including for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The text below includes minor additions and formatting changes to the original text. Short-term analyses, as published in Moore et al.,101 are reported alongside the long-term models that include data for 2019 to enable side-by-side comparison of estimated change using shorter- and longer-term data.
This chapter addresses the following research questions:
- Has regulation of e-cigarettes via the TPD interrupted growth in young people’s e-cigarette use?
- How do young people perceive risks and social norms surrounding e-cigarettes (and how do these change over time as products become TPD compliant)?
- How do young people interpret and respond to the presence or absence of health warnings on e-cigarette packets?
- To what extent, and in what ways, do young people continue to interact with e-cigarette marketing (e.g. via online marketing) after the prohibition of cross-border advertising?
- Did the rate of decline in young people’s smoking change after additional regulations of tobacco and e-cigarettes were introduced May 2016?
The chapter begins with analysis of qualitative interviews with young people in England, Scotland and Wales in 2017 and 2018, focussing on change over time in perceptions of e-cigarettes. The qualitative interviews are augmented with quantitative data on processes discussed by young people, using items added to the SHRN survey in Wales in 2017 and 2019. Data sources are compared to corroborate or challenge one another. We then present analyses of change over time in e-cigarette use before and after the introduction of TPD regulations. Finally, analysis of long-term trends in tobacco smoking, and change at the point of intervention, extends the quantitative analyses presented in Chapter 3.
Sample description
As indicated in Table 9, in 2017, more participants took part in Wales, with an approximately equal division of boys and girls in each country and an average group size of three or four young people. In 2018, participants were more equally distributed across the three nations, with additional schools recruited in England and Scotland. Although in Scotland and Wales there remained an even split of girls and boys taking part in group interviews, in England there was a tendency for more participation among girls than among boys.
| 2017 | 2018 | |||||
|---|---|---|---|---|---|---|
| England | Scotland | Wales | England | Scotland | Wales | |
| Number of schools | 2 | 2 | 3 | 3 | 4 | 4 |
| Number of interviews | 7 | 4 | 10 | 15 | 14 | 12 |
| Number of participants | 27 | 12 | 37 | 47 | 48 | 53 |
| Percentage male/female | 52/48 | 50/50 | 46/54 | 33/67 | 48/52 | 49/51 |
Young people’s perceptions of e-cigarettes following implementation of the Tobacco Products Directive
Changes in perceived prevalence and norms for e-cigarette use
In 2017, a large majority of young people across all groups had stated that e-cigarette use had recently risen rapidly among their age group and was now perceived as more common than tobacco smoking. Experimental use of e-cigarettes was widely viewed as unproblematic. Trying an e-cigarette was driven by various factors, including sharing with peers, fun and the appeal of flavours:
Yeah and you can get them like scented and stuff like that and so it entices young people.
I just liked the different flavours. ’Cos my friend had jam donut, another friend had gummy bear flavour and Heisenberg which is a minty flavour which is quite nice.
However, regular e-cigarette use was judged negatively. When asked to consider who regular users were among their school cohort, some young people felt that regular users came from the same groups as smokers, characterised as ‘outsiders’ or disengaged young people. Although experimentation with e-cigarettes was very evident, it was also a discussion point in many groups that this behaviour was a trend that seemed to have peaked and was likely to decrease among their cohort going forward:
Like the fidget spinners, and then I think as soon as fidget spinners die it’s going to be the same kind of, ‘Oh why are you still using them?’ kind of approach, the same as vaping.
To explore this perceived peak in use, perceptions of prevalence were discussed in repeat interviews with young people in 2018. The repeat interviews appeared to confirm initial indications of e-cigarettes as a ‘fad’ that had peaked, with young people losing interest over time:
... you’ll see things of it occasionally but I don’t see anyone doing it just to show off now, any more.
E-cigarettes were now described as being more commonly used by young adults who were using them to quit smoking. However, some interviewees argued that young people younger than themselves were still attracted by flavours and features that had drawn interviewees to try them previously. In 2018, approval of e-cigarette use was more variable than in 2017, with increased likelihood of peer censure associated with e-cigarettes being seen as an outdated fad:
They just go – [mocking tone] I’ve got a vape, I’m sick, look at me!
I know some of the kids in our year they’re like, oh they just made jokes about it, oh you’re vaping, and think it’s a bit chavvy, and some, they think it’s cool.
When initially discussing reasons for e-cigarette use in 2017, very few interviewees cited quitting smoking as a driver for young people’s e-cigarette use. However, in the 2018 follow-up interviews, cessation was often cited as a reason for young people using e-cigarettes. Regular use of e-cigarettes as a cessation aid was largely approved of, as opposed to regular use of e-cigarettes in non-smokers:
I think that if it’s used to actually try and give up smoking then I see a huge advantage to it, but obviously if you’re videoing yourself doing tricks and stuff that’s not trying to give it up [smoking], is it?
Perceived family and peer reactions to e-cigarette use
At both time points, most participants reported that their parents would strongly disapprove of them using e-cigarettes or tobacco, with almost all participants suggesting that they would anticipate a worse parental reaction to tobacco. Some interpreted this as parental fear of tobacco being more clearly associated with moving on to other substances such as cannabis, while others suggested it being due to harms of smoking being better known than for vaping:
My parents would not be happy. Because they know the damage it does to your body ... My mum’s a nurse and she knows all about it, and it wouldn’t be very approved of in my family.
In 2017, young people described some expectation of negative peer reaction to both smoking and e-cigarette use, but at lower levels than that expected from parents. Most young people suggested that peers would be more likely to censure smoking than use of e-cigarettes. In 2018, some young people expressed a general decrease in disapproval of ‘social’ smoking relative to their previous year in school, associated with maturation, although disapproval of regular smoking remained strong:
I think it is more socially accepted like, now, like in our year to smoke than it was last year. Like you see a lot more people smoking than before especially at parties.
Survey data from Wales in 2017 and 2019 concurred with perceptions of high perceived parental disapproval of tobacco and e-cigarettes, with stronger disapproval of tobacco than e-cigarettes. Among young people who did not report current use of tobacco or e-cigarettes, a clear majority said that if they were to start, then their parents would try to stop them. At both time points, young people were more likely to report that their parents would try to stop them smoking than try to stop them using e-cigarettes, with a larger proportion of young people saying that their parents would try and persuade them to stop using e-cigarettes (see Figure 3).
FIGURE 3 Percentage of non-smoking/non-vaping 13- to 15-year-olds reporting that were they to take up smoking/vaping then their parents would try to stop them or persuade them to stop. The question was asked of a random subsample of young people in the 2017/19 SHRN surveys (2017, n = 11,442; 2019, n = 11,197; smoking, n = 8535; vaping, n = 8329). (Figure not reproduced here; see the original article.)
Among young people who reported current use of tobacco or e-cigarettes, half reported that their parents did not know. Among young people whose parents did know about their smoking or vaping, most reported that their parents’ either tried to stop them or tried to persuade them to stop (see Figure 4).
FIGURE 4 Percentage of current users of e-cigarettes and tobacco reporting parental responses to their use. The question was asked of of a random subsample of current users in the 2017/19 SHRN surveys (2017, n = 989/993; 2019, n = 744/706, repectively). (Figure not reproduced here; see the original article.)
Risk perception
In 2017, increased visibility of e-cigarettes in recent years had been widely cited as instrumental in increasing acceptability of e-cigarettes and reducing risk perceptions. Nevertheless, young people frequently cited mechanical risks (e.g. malfunctioning devices), potentially dangerous ingredients in liquids and, most commonly, as-yet unidentified harms:
Because they know what’s in the cigarette, like in tobacco cigarettes but they don’t know what’s fully in the e-cigarettes or the vapes.
In both rounds of qualitative interviews, mechanical risks and unknown harms were commonly referenced. In 2017, addiction to nicotine was also mentioned in around two-thirds of groups as a risk of e-cigarette use. However, in 2018, addiction was less frequently mentioned. Many young people who had used e-cigarettes said that they were unsure if it had contained nicotine or not, with use driven by flavour rather than nicotine.
The SHRN survey data indicated that in both 2017 and 2019 a minority of young people who had tried e-cigarettes reported that the last e-cigarettes they used contained nicotine (see Figure 5). At both time points, around one in seven young reported not knowing what was in the last e-cigarette they used (14% in 2017 vs. 13% in 2019). Survey data also indicated changing views on relative harms of tobacco and e-cigarettes during this time. In 2017, approximately one in three young people reported that e-cigarettes were as harmful to health as tobacco, whereas almost half reported that tobacco cigarettes were worse (of the remaining young people, 3% reported that e-cigarettes were worse and 14% said that they did not know). In 2019, almost half of young people reported that e-cigarettes were as harmful to health as tobacco cigarettes, whereas approximately one-third reported that tobacco cigarettes were worse (of the remaining young people, 5% reported that e-cigarettes were worse and 13% said that they did not know).
FIGURE 5 Percentage of current e-cigarette users’ (n = 8887/2880) reporting whether or not their last e-cigarette contained nicotine, and all young people’s (n = 12,824/9481) perceived relative harmfulness of tobacco and e-cigarettes in Wales in 2017 and 2019, respectively. (Figure not reproduced here; see the original article.)
Availability of e-cigarettes
In 2017, most young people agreed that they would find it easier to obtain e-cigarettes than tobacco cigarettes, with few young people suggesting that both were equally obtainable:
But, like, e-cigarettes, it seems to be that you can get hold of one quite easily. But cigarettes – you have to like, know somebody who’s old enough to go buy you some.
For most interviewees who had tried an e-cigarette, this had been through an informal supply route (e.g. a friend at a social event), although some interviewees reported awareness of someone at school who had bulk bought online with the intention to sell them on:
I was also one of the people who were making profit from them, people would ask me could you order one for me, and I was like OK well I’ll order them. I’d order 10 at once.
Few interviewees stated that e-cigarettes were available through retailers. Qualitative data indicated that when purchased face to face, this was through shops known locally to be willing to sell illicit products, rather than via specialist vape shops:
Like I’ve seen down in X market, they sell to someone who was 9.
Discussions a year later suggested that the landscape of obtaining e-cigarettes had changed. Many young people still felt that they could obtain e-cigarettes if desired, mostly through older peers or online purchase, but it was seen as more challenging to buy through school supply chains. This was associated with a perception of decreased use, which meant that fewer people were now selling e-cigarettes:
Int: Do you think it’s got easier or harder to get hold of vapes?
W1(1)M: I think it’s a bit harder because you don’t really see that many people do it anymore.
Int: ... people selling them, would that go on in school or after school?
W1(1)M: I don’t think it’s much in school. I know some people used to do it [sell vapes], but I don’t know anyone now who does it.
In 2018, tobacco was viewed as being as easy to obtain as e-cigarettes. This was attributed to factors such as increased age, which meant greater likelihood of having older peers or appearing old enough to self-purchase in shops, and also to increased smoking prevalence in the year group and, hence, a larger number of people to obtain tobacco from if desired. A significant number of young people were aware of someone at school who sold individual cigarettes, making cigarettes an affordable option:
I know in school obviously a lot of people do it in the corner, but some people sell them individually for 50 pence or maybe a £1 or something, and people do pay that money for them individually.
Quantitative data from Wales (items included in only the 2017 survey) concurred with the view that most young people who used e-cigarettes obtained the e-cigarettes via means other than a retailer, with the most commonly reported supply route used being peers (see Figure 6). Approximately one-third of young people reported buying e-cigarettes from, or being given e-cigarettes by, peers. The next most commonly selected option was ‘some other way’, not captured by our response options, followed by obtaining e-cigarettes via adults. Fewer than one in six young people reported purchasing e-cigarettes from shops, with approximately 7% reporting buying e-cigarettes via the internet. Among young people who reported online purchase, approximately one in three (n = 29, 35.0%) reported that they had changed where they purchased e-cigarettes from in the past 6 months.
FIGURE 6 The percentage of current e-cigarette-using young people reporting obtaining e-cigarettes via each supply route. The question was asked of a random sample of young people who reported having used an e-cigarette more than once: 2017 SHRN survey only (n = 3313). (Figure not reproduced here; see the original article.)
Exposure to warning information and e-cigarette marketing
The TPD introduced a mandated warning, stating ‘this product contains nicotine, which is a highly addictive substance’. In 2017, recognition of this warning among interviewees was low. In 2018, recognition slightly increased, although remained limited, with young people continuing to indicate limited exposure to e-cigarettes within their packaging:
Int: Have you seen any warnings on vape packets?
E4(1)F: No. Can you even get vape packets?
When asked to discuss the current warning, a large majority of young people suggested that it would likely be ineffective at discouraging e-cigarette use among their age group. Young people commonly assessed the warning in relation to more graphic and more well-known visuals present on tobacco packaging, deeming e-cigarette warnings mild by comparison:
Like if I was vaping and like I was like ‘oh it’s fine because look at how bad cigarettes were, like I’m fine’, because there’s no health warnings on any of the [e-cigarette] packets.
Brand awareness among young people was very low across both rounds of interviews. The most commonly reported avenue for seeing adverts in 2017 was vape shops, followed by online sources (i.e. streamed videos of e-cigarette users performing tricks). Exposure to advertising in public spaces, such as billboards, was seen as relatively common. In the 2018 interviews, however, few young people reported seeing any e-cigarette advertising.
Contrasting with these reports, approximately two-thirds of young people in SHRN surveys in Wales in both 2017 and 2019 reported exposure to at least one form of e-cigarette advertising in the past month. The most common location in which young people reported seeing advertising for e-cigarettes was in shops, followed by exposure to advertising on the internet. However, consistent with qualitative data on of reduced online engagement with e-cigarette-related material, the percentage of young people reporting internet-based advertising exposure declined from 39% to 34% by 2019. There were small increases in exposure to advertising on buses or in bus shelters (see Figure 7).
FIGURE 7 Percentage of young people reporting having seen adverts in each location in the past month in 2017 and 2019. The question was asked of a random sample of participants in the SHRN surveys (n = 13,503/10,109). (Figure not reproduced here; see the original article.)
Change in young people’s e-cigarette use after the Tobacco Products Directive
This section analyses change over time in young people’s e-cigarette use before and after the implementation of the TPD. The section begins with our primary segmented regression analysis of data from Wales, and is followed by analyses of yearly data from England and Scotland to examine the transferability of findings from our primary statistical analysis in Wales to other countries in the UK.
Descriptive cross country comparison of changes in young people’s use of e-cigarettes
Figure 8 provides a descriptive overview of the percentage of ever and regular users of e-cigarettes over time, by country, between 2013 and 2019. The black line indicates the beginning of the intervention period (i.e. TPD implementation).
FIGURE 8 Percentage of 13- to 15-year-olds reporting ever and regular use of e-cigarettes by year and country from 1998 to 2019. The black bars represent the beginning of one ‘intervention’ period (i.e. the introduction of TPD). (Figure not reproduced here; see the original article.)
In Wales and Scotland, a near doubling in prevalence of ever use of e-cigarettes is evident in pre-TPD surveys from 2013 to 2015, with growth from 16% to 30% in Wales and from 12% to 23% in Scotland. In Wales, where post-TPD data are available from both the 2017 and 2019 surveys, a smaller continued growth was observed from 2015 to 2017, with ever use reaching 35% by 2017, followed by a marginal decline from 2017 to 2019 to 32%. In Scotland, where only one post-TPD survey was available, a slight increase in prevalence of ever use of e-cigarettes was observed in 2018 (25%) relative to 2015. In England, the first measurement was in 2014, at which point 27% of young people reported having tried an e-cigarette, and this rose to 33% by the 2016 survey (conducted marginally after the implementation of TPD), with no further growth after this time point. Hence, in all nations, there was evidence of rapid growth in ever use in the years leading up to the TPD regulations, with plateauing or marginal decline after the intervention point. Similarly, for the percentage of young people reporting regular use (i.e. at least weekly use) of e-cigarettes, all data sets provided evidence of some growth in the period prior to the TPD regulations, but with plateauing or marginal decline between the two most recent survey periods.
Change in e-cigarette use by country
Primary statistical analysis: ever e-cigarette use in Wales
As indicated in Table 10, the trend displayed in Figure 8, that is of large growth in ever use of e-cigarettes from 2013 to 2015, followed by much smaller growth from 2015 to 2017 (which incorporates the beginning of TPD implementation), and small decline thereafter was observed across subgroups. Ever use was slightly higher among boys, older young people and young people from lower SES families throughout, and substantially higher among ever smokers.
| Subgroup | Ever e-cigarette use, % (95% CI) | |||
|---|---|---|---|---|
| 2013 | 2015 | 2017 | 2019 | |
| All | 16.4 (14.3 to 18.7) | 30.1 (27.6 to 32.8) | 34.6 (33.2 to 36.0) | 31.7 (30.4 to 33.0) |
| Boys | 17.0 (14.2 to 20.2) | 30.5 (27.7 to 33.5) | 37.0 (35.5 to 38.5) | 32.8 (31.5 to 34.2) |
| Girls | 15.7 (13.1 to 18.6) | 29.7 (26.9 to 32.7) | 32.2 (30.6 to 33.8) | 30.5 (29.0 to 32.1) |
| 13-year-olds | 11.5 (9.4 to 14.0) | 22.0 (19.4 to 24.8) | 26.0 (24.7 to 27.4) | 22.6 (21.3 to 24.0) |
| 15-year-olds | 22.1 (19.1 to 25.4) | 39.8 (37.0 to 42.7) | 45.5 (43.7 to 47.3) | 42.5 (40.8 to 44.2) |
| High SES family | 14.5 (12.1 to 17.2) | 26.9 (24.1 30.0) | 33.4 (31.9 to 35.0) | 30.4 (28.9 to 31.9) |
| Low SES family | 17.7 (15.3 to 20.3) | 31.8 (29.3 to 34.5) | 35.8 (34.3 to 37.3) | 32.4 (31.0 to 33.9) |
| White | 16.3 (14.2 to 18.7) | 30.2 (27.6 to 32.9) | 34.9 (33.5 to 36.3) | 32.2 (30.9 to 33.4) |
| Ethnic minority | 16.9 (12.5 to 22.5) | 30.0 (24.2 to 36.5) | 33.2 (30.2 to 36.5) | 28.6 (25.1 to 32.3) |
| Ever smoker | 53.3 (48.9 to 57.7) | 83.9 (80.7 to 86.7) | 88.5 (87.5 to 89.5) | 88.5 (87.4 to 89.5) |
| Never smoker | 6.4 (5.1 to 8.1) | 18.4 (16.5 to 20.5) | 22.3 (21.2 to 23.4) | 19.9 (18.9 to 21.0) |
For short-term post-implementation segmented regression analysis, which disaggregated data sets by survey month, data were available for 12 time points between November 2013 and December 2017, representing 51,056 young people in Wales. Our longer-term analysis through to December 2019 included four further data points, with a total analytical sample of 91,687 young people included in primary statistical analysis. There was no evidence of a non-linear trend (quadratic term OR 1.00, 95% CI 0.999 to 1.001; p = 0.91) and, hence, a quadratic term was not included in our models. As indicated in Table 11, ever use of e-cigarettes increased significantly prior to TPD implementation, with an approximately 4% (i.e. an OR of 1.04) increase in odds of ever use of e-cigarettes with each month passed prior to legislation. The whole sample regression model is presented visually in a predicted probability plot in Report Supplementary Material 1, Figure 1. In both our short- and long-term models, there was no clear evidence of a step change in prevalence at the intervention point (indicated by the variable ‘level’), with direction and size of ORs varying between subgroups, and CIs wide. However, for the whole sample, and among all subgroups (with the exception of young people from ethnic minority backgrounds in short-term analyses), ORs for change in trend were in the direction of a negative change in trend, indicating that growth in e-cigarette use beyond implementation was significantly lower than predicted from the secular trend. ORs in both short- and long-term models indicated an approximately 4% reduction (i.e. an OR of 0.96) in odds of ever e-cigarette use each month beyond the TPD implementation date of May 2016, relative to baseline trend. In the short-term analysis, this was not statistically significant; however, in models including longer-term data, the point estimate remained unchanged but CIs narrowed and reached statistical significance. Hence, overall, data are consistent with a hypothesis of negative change in trend (i.e. significant disruption to secular growth) at the point of TPD implementation. A sensitivity analysis including all young people in the Welsh surveys (i.e. aged 11–16 years) was highly consistent with our primary statistical analysis.
| Sample | Short-term analysis (2013–17) | Long-term analysis (2013–19) | |||
|---|---|---|---|---|---|
| OR (95% CI) | p-value | OR (95% CI) | p-value | ||
| All (n = 51,056a;91,687b) | Time | 1.04 (1.03 to 1.05) | <0.001 | 1.04 (1.03 to 1.05) | <0.001 |
| Level | 1.09 (0.41 to 2.89) | 0.860 | 1.10 (0.93 to 1.29) | 0.252 | |
| Post slope | 0.96 (0.91 to 1.01) | 0.125 | 0.96 (0.95 to 0.97) | <0.001 | |
| Boys (n = 24,993; 45,008) | Time | 1.04 (1.03 to 1.05) | <0.001 | 1.04 (1.03 to 1.05) | <0.001 |
| Level | 1.88 (0.70 to 5.06) | 0.213 | 1.27 (1.05 to 1.53) | 0.013 | |
| Post slope | 0.94 (0.89 to 0.99) | 0.020 | 0.96 (0.95 to 0.97) | <0.001 | |
| Girls (n = 26,063; 46,679) | Time | 1.04 (1.03 to 1.05) | <0.001 | 1.04 (1.03 to 1.05) | <0.001 |
| Level | 0.64 (0.19 to 2.22) | 0.485 | 0.96 (0.80 to 1.16) | 0.664 | |
| Post slope | 0.98 (0.91 to 1.05) | 0.587 | 0.96 (0.95 to 0.97) | <0.001 | |
| 13-year-olds (n = 28,471; 50,619) | Time | 1.04 (1.02 to 1.05) | <0.001 | 1.04 (1.03 to 1.05) | <0.001 |
| Level | 1.49 (0.39 to 5.68) | 0.563 | 1.13 (0.91 to 1.40) | 0.257 | |
| Post slope | 0.94 (0.88 to 1.02) | 0.124 | 0.96 (0.95 to 0.97) | <0.001 | |
| 15-year-olds (n = 22,585; 41,068) | Time | 1.04 (1.03 to 1.05) | <0.001 | 1.04 (1.03 to 1.05) | <0.001 |
| Level | 0.85 (0.26 to 2.78) | 0.789 | 1.08 (0.89 to 1.31) | 0.459 | |
| Post slope | 0.97 (0.91 to 1.04) | 0.375 | 0.96 (0.95 to 0.97) | <0.001 | |
| White (n = 45,727; 81,810) | Time | 1.04 (1.03 to 1.05) | <0.001 | 1.04 (1.03 to 1.05) | <0.001 |
| Level | 1.19 (0.45 to 3.15) | 0.726 | 1.10 (0.94 to 1.29) | 0.223 | |
| Post slope | 0.95 (0.90 to 1.01) | 0.089 | 0.96 (0.95 to 0.97) | <0.001 | |
| Ethnic minority (n = 4528; 8536) | Time | 1.03 (1.01 to 1.05) | 0.001 | 1.03 (1.01 to 1.05) | 0.001 |
| Level | 0.56 (0.04 to 7.35) | 0.662 | 1.12 (0.66 to 1.90) | 0.685 | |
| Post slope | 1.00 (0.86 to 1.15) | 0.963 | 0.96 (0.94 to 0.98) | <0.001 | |
| High SES (n = 25,002; 38,898) | Time | 1.04 (1.03 to 1.05) | <0.001 | 1.04 (1.03 to 1.05) | <0.001 |
| Level | 1.76 (0.62 to 4.97) | 0.285 | 1.24 (1.02 to 1.51) | 0.029 | |
| Post slope | 0.94 (0.88 to 1.00) | 0.037 | 0.96 (0.95 to 0.97) | <0.001 | |
| Low SES (n = 25,167; 50,374) | Time | 1.04 (1.03 to 1.05) | <0.001 | 1.04 (1.03 to 1.05) | <0.001 |
| Level | 0.73 (0.23 to 2.28) | 0.592 | 1.07 (0.90 to 1.28) | 0.453 | |
| Post slope | 0.98 (0.92 to 1.04) | 0.500 | 0.96 (0.95 to 0.97) | <0.001 | |
| Never smoker (n = 40,703; 73,952) | Time | 1.06 (1.04 to 1.07) | <0.001 | 1.06 (1.04 to 1.07) | <0.001 |
| Level | 1.05 (0.39 to 2.83) | 0.917 | 1.01 (0.84 to 1.21) | 0.950 | |
| Post slope | 0.94 (0.89 to 0.99) | 0.022 | 0.94 (0.93 to 0.95) | <0.001 | |
| Ever smoker (n = 8746; 15,327) | Time | 1.07 (1.06 to 1.08) | <0.001 | 1.07 (1.06 to 1.08) | <0.001 |
| Level | 2.34 (0.36 to 15.38) | 0.376 | 0.95 (0.68 to 1.34) | 0.777 | |
| Post slope | 0.89 (0.80 to 0.98) | 0.024 | 0.93 (0.92 to 0.95) | <0.001 | |
Overall, estimates of change in trend were more consistent between subgroups in the long-term analysis than in the short term analysis. Reductions in the secular growth of ever use were greater for boys than for girls in the short-term analysis, with an estimated 6% reduction in odds per month for boys and a 2% reduction in odds per month for girls. In the short-term analysis, subgroup analyses by ethnicity and SES indicated no change in trend in ethnic minority groups; however, a significant change in trend was observed for children from higher SES families, although this trend did not persist in the long-term analysis.
At both time points, for both never smokers and ever smokers, the ORs suggested larger changes in trend for both groups than for the whole sample. In short-term analysis, odds declined by 6% per month post implementation for never smokers and by 11% per month post implementation for ever smokers. In long-term analysis, odds declined by 6% per month post implementation for never smokers and by 7% per month post implementation for ever smokers (see Table 11). As a post hoc investigation to understand these differences, a binary term for ever smoking was added to the whole group model. In short-term analysis, the negative change in trend post intervention increased in magnitude, and was statistically significant (OR 0.93, 95% CI 0.88 to 0.98). This same estimate was repeated at long-term follow-up, with CIs narrowed because of the larger sample size (OR 0.93, 95% CI 0.92 to 0.94). This is consistent with a conclusion that ever smoking operated as a time-variant confounder, and that in a priori models change in trend was underestimated because of changes in ever smoking occurring in the opposing direction across the time series.
Regular e-cigarette use in Wales
Data on regular e-cigarette use were not collected in Wales until 2015 (i.e. pre TPD), and were collected again in 2017 (i.e. short-term analysis post TPD) and in 2019 (i.e. long-term analysis post TPD) (n = 87,704). A small increase in prevalence of regular use to 2017, followed by a small decline to 2019, was consistent across all subgroups, with 2019 estimates marginally lower than 2015 estimates in all cases (see Table 12). ORs from logistic regression analyses of data for each of the 3 years indicate an estimated 15% increase in odds of regular e-cigarette use between 2015 and 2017, falling short of statistical significance (OR 1.15, 95% CI 0.97 to 1.36; p = 0.11). The odds of e-cigarette use in 2019 declined by approximately 16% relative to 2015 (OR 0.84, 95% CI 0.70 to 1.02; p = 0.09). Only for high SES groups did increase in regular e-cigarette use from 2015 to 2017 reach significance, whereas contrasts between 2015 and 2019 were significant for boys and Year 11 students, indicating significantly reduced use in these groups. Regular use of e-cigarettes among never smokers remained low throughout, at approximately 1% across the time series.
| Sample | Percentage (95% CI) | OR (95% CI); p-value | |||
|---|---|---|---|---|---|
| 2015 | 2017 | 2019 | 2017 vs. 2015 | 2019 vs. 2015 | |
| All (n = 87,704) | 4.2 (3.9 to 4.6) | 4.8 (4.6 to 5.1) | 3.6 (3.5 to 3.8) | 1.15 (0.97 to 1.36); 0.111 | 0.84 (0.70 to 1.02); 0.085 |
| Boys (n = 42,989) | 5.1 (4.5 to 5.7) | 6.1 (5.8 to 6.5) | 4.1 (3.8 to 4.4) | 1.23 (0.99 to 1.53); 0.056 | 0.79 (0.64 to 0.98); 0.035 |
| Girls (n = 44,715) | 3.5 (3.0 to 4.0) | 3.6 (3.3 to 3.9) | 3.2 (3.0 to 3.5) | 1.04 (0.82 to 1.31); 0.767 | 0.92 (0.71- to 1.19); 0.528 |
| 13-year-olds (n = 48,384) | 2.8 (2.4 to 3.2) | 3.6 (3.3 to 3.9) | 2.7 (2.5 to 2.9) | 1.29 (1.01 to 1.65); 0.041 | 0.95 (0.72 to 1.25); 0.709 |
| 15-year-olds (n = 39,320) | 6.0 (5.3 to 6.7) | 6.4 (6.0 to 6.8) | 4.8 (4.5 to 5.1) | 1.07 (0.87 to 1.31); 0.543 | 0.78 (0.63 to 0.98); 0.033 |
| High SES (n = 37,080) | 3.6 (3.1 to 4.1) | 4.6 (4.3 to 4.9) | 3.2 (2.9 to 3.5) | 1.28 (1.04 to 1.57); 0.019 | 0.88 (0.70 to 1.11); 0.276 |
| Low SES (n = 48,312) | 4.4 (3.9 to 5.0) | 5.1 (4.8 to 5.4) | 3.7 (3.5 to 4.0) | 1.16 (0.93 to 1.46); 0.193 | 0.83 (0.65 to 1.06); 0.138 |
| White (n = 78,113) | 4.0 (3.6 to 4.4) | 4.8 (4.5 to 5.0) | 3.5 (3.4 to 3.7) | 1.20 (1.01 to 1.44); 0.043 | 0.87 (0.71 to 1.07); 0.199 |
| Ethnic minority (n = 8277) | 5.9 (4.6 to 7.4) | 6.0 (5.2 to 6.9) | 4.6 (4.0 to 5.3) | 1.02 (0.67 to 1.58); 0.912 | 0.77 (0.54 to 1.09); 0.141 |
| Ever smoker (n = 14,375) | 19.7 (17.9 to 21.6) | 21.1 (20.1 to 22.2) | 17.5 (16.6 to 18.4) | 1.06 (0.89 to 1.27); 0.502 | 0.85 (0.69 to 1.03); 0.104 |
| Never smoker (n = 71,195) | 0.9 (0.7 to 1.1) | 1.2 (1.0 to 1.3) | 0.8 (0.8 to 1.0) | 1.29 (0.97 to 1.73); 0.081 | 0.93 (0.66 to 1.31); 0.664 |
E-cigarette use in Scotland and England
As indicated in Table 13, ever use of e-cigarettes in Scotland approximately doubled from 2013 to 2015 among all subgroups, with more modest increase thereafter. Odds of ever e-cigarette use were 2.5 times lower in 2013 than in 2015. Odds of ever e-cigarette use increased more marginally from 2015 to 2018, despite a longer gap between data points. Despite some variation in magnitude across subgroups, this pattern of substantially larger growth from 2013 to 2015 than from 2015 to 2018 is observed across all subgroup analyses. There was an almost threefold increase in ever use of e-cigarettes among never smokers from 2013 to 2015, although with marginal increase thereafter. Regular use, although remaining low at approximately 2%, grew substantially in relative terms from 2013 to 2015, with little further growth from 2015 to 2018. This pattern was observed across all subgroups, with significant growth between 2013 and 2015 in all subgroups, but no significant growth from 2015 to 2018 in any group. The only group for whom near significant growth in regular use was observed from 2015 to 2018 was never smokers, although, notably, a 45% increase in relative odds of regular use from 2015 to 2018 reflects an absolute growth from 0.2% to 0.3%, with regular use remaining very low in absolute terms. It was not possible to analyse differential change by ethnicity in Scotland, as the 2015 and 2018 SALSUS data sets did not include a variable on ethnicity. In England (see Table 14), although there was evidence of growing prevalence of ever use from 2014 until 2016, there was no further growth and, indeed, a marginal reduction in ever use to 31% in the whole sample in the 2 years following the TPD (i.e. from 2016 to 2018). A similar pattern was observed across subgroups defined by age and gender, although decline from 2016 to 2018 was significant for only ethnic minority participants. Regular use increased between 2014 and 2016, but fell between 2016 and 2018, with significant declines among the whole sample, and among boys and younger children. In both Scotland and England, regular e-cigarette use among never smokers remained below 1% throughout the time series. Hence, data from England and Scotland concur with findings from Wales that e-cigarette use grew in the period prior to the TPD regulations, but limited or no further increase occurred in the years beyond the TPD regulations. Results from weighted models were largely consistent with unweighted models (see Report Supplementary Material 1, Tables 7 and 8), although marginally significant declines in ever and regular use in England from 2016 to 2018 were reduced and no longer significant following application of weights.
| E-cigarette use | Sample | Percentage (95% CI) | OR (95% CI); p-value | |||
|---|---|---|---|---|---|---|
| 2013 | 2015 | 2018 | 2013 vs. 2015 | 2018 vs. 2015 | ||
| Ever e-cigarette use | All (n = 78,422) | 11.8 (11.5 to 12.2) | 22.8 (22.3 to 23.4) | 24.7 (24.1 to 25.3) | 0.43 (0.41 to 0.45); <0.001 | 1.11 (1.07 to 1.16); <0.001 |
| Boys (n = 38,934) | 12.2 (11.7 to 12.7) | 25.0 (24.2 to 25.7) | 27.7 (26.9 to 28.6) | 0.40 (0.38 to 0.43); <0.001 | 1.17 (1.11 to 1.25); <0.001 | |
| Girls (n = 39,488) | 11.4 (10.9 to 11.9) | 20.8 (20.1 to 21.5) | 21.6 (20.9 to 22.4) | 0.47 (0.44 to 0.50); <0.001 | 1.05 (0.98 to 1.12); 0.152 | |
| 13-year-olds (n = 41,422) | 7.3 (6.9 to 7.7) | 15.2 (14.6 to 15.9) | 16.6 (15.9 to 17.3) | 0.44 (0.40 to 0.47); <0.001 | 1.10 (1.03 to 1.18); 0.006 | |
| 15-year-olds (n = 37,000) | 16.6 (16.0 to 17.2) | 31.6 (30.8 to 32.5) | 34.2 (33.2 to 35.1) | 0.43 (0.41 to 0.46); <0.001 | 1.12 (1.06 to 1.19); <0.001 | |
| High SES (n = 36,525) | 10.4 (9.9 to 10.9) | 21.1 (20.3 to 21.9) | 23.2 (22.4 to 24.0) | 0.41 (0.38 to 0.44); <0.001 | 1.14 (1.07 to 1.22); <0.001 | |
| Low SES (n = 41,823) | 13.0 (12.5 to 13.5) | 24.1 (23.4 to 24.9) | 26.2 (25.4 to 27.0) | 0.45 (0.42 to 0.48); <0.001 | 1.12 (1.06 to 1.19); <0.001 | |
| Ever smoker (n = 16,295) | 39.5 (38.4 to 40.6) | 71.6 (70.3 to 72.9) | 74.6 (73.2 to 75.9) | 0.26 (0.24 to 0.28); <0.001 | 1.16 (1.06 to 1.28); 0.001 | |
| Never smoker (n = 61,182) | 3.8 (3.5 to 4.0) | 10.5 (10.1 to 10.9) | 11.5 (11.0 to 12.0) | 0.33 (0.30 to 0.36); <0.001 | 1.11 (1.04 to 1.19); 0.002 | |
| Regular e-cigarette use | All (n = 78,422) | 0.5 (0.4 to 0.6) | 2.1 (2.0 to 2.3) | 2.2 (2.0 to 2.4) | 0.23 (0.20 to 0.28); <0.001 | 1.06 (0.93 to 1.20); 0.400 |
| Boys (n = 38,934) | 0.7 (0.6 to 0.8) | 3.0 (2.7 to 3.3) | 3.0 (2.7 to 3.4) | 0.22 (0.18 to 0.27); <0.001 | 1.03 (0.89 to 1.20); 0.691 | |
| Girls (n = 39,488) | 0.4 (0.3 to 0.5) | 1.3 (1.1 to 1.5) | 1.5 (1.2 to 1.7) | 0.27 (0.20 to 0.36); <0.001 | 1.11 (0.89 to 1.38); 0.356 | |
| 13-year-olds (n = 41,422) | 0.3 (0.2- to 0.4) | 1.3 (1.1 to 1.5) | 1.3 (1.1 to 1.5) | 0.25 (0.18 to 0.33); <0.001 | 1.03 (0.83 to 1.28); 0.792 | |
| 15-year-olds (n = 37,000) | 0.7 (0.6 to 0.9) | 3.1 (2.8 to 3.5) | 3.3 (3.0 to 3.7) | 0.23 (0.18 to 0.28); <0.001 | 1.07 (0.92 to 1.25); 0.392 | |
| High SES (n = 36,525) | 0.5 (0.4 to 0.6) | 1.9 (1.7 to 2.2) | 2.1 (1.8 to 2.4) | 0.23 (0.18 to 0.31); <0.001 | 1.11 (0.91 to 1.34); 0.300 | |
| Low SES (n = 41,823) | 0.6 (0.5 to 0.7) | 2.3 (2.0 to 2.6) | 2.4 (2.1 to 2.7) | 0.24 (0.19 to 0.30); <0001 | 1.03 (0.87 to 1.22); 0.702 | |
| Ever smoker (n = 16,295) | 2.1 (1.7 to 2.4) | 9.7 (8.9 to 10.6) | 9.4 (8.6 to 10.3) | 0.20 (0.16 to 0.24); <0.001 | 0.96 (0.83 to 1.11); 0.582 | |
| Never smoker (n = 61,182) | 0.1 (0.0 to 0.1) | 0.2 (0.2 to 0.3) | 0.3 (0.2 to 0.4) | 0.31 (0.18 to 0.54); <0.001 | 1.45 (0.97 to 2.15); 0.067 | |
| E-cigarette use | Sample | Percentage (95% CI) | OR (95% CI); p-value | |||
|---|---|---|---|---|---|---|
| 2014 | 2016 | 2018 | 2014 vs. 2016 | 2018 vs. 2016 | ||
| Ever e-cigarette use | All (n = 14,010) | 27.3 (25.7 to 29.0) | 32.5 (31.3 to 33.8) | 30.7 (29.5 to 31.9) | 0.75 (0.68 to 0.83); <0.001 | 0.92 (0.85 to 0.99); 0.036 |
| Boys (n = 6881) | 29.0 (26.6 to 31.4) | 35.2 (33.4 to 37.1) | 34.1 (32.4 to 35.8) | 0.72 (0.62 to 0.83); <0.001 | 0.93 (0.83 to 1.05); 0.239 | |
| Girls (n = 7129) | 25.6 (23.3 to 28.0) | 30.1 (28.4 to 31.8) | 27.3 (25.7 to 29.0) | 0.79 (0.68 to 0.92); 0.002 | 0.90 (0.80 to 1.01); 0.085 | |
| 13-year-olds (n = 6940) | 19.4 (17.3 to 21.6) | 23.8 (22.2 to 25.4) | 21.9 (20.4 to 23.5) | 0.77 (0.65 to 0.90); 0.002 | 0.90 (0.80 to 1.03); 0.116 | |
| 15-year-olds (n = 7070) | 34.4 (31.9 to 36.9) | 41.2 (39.3 to 43.0) | 39.6 (37.8 to 41.4) | 0.74 (0.65 to 0.85); <0.001 | 0.93 (0.84 to 1.04); 0.184 | |
| White (n = 10,825) | 27.6 (25.8 to 29.5) | 32.9 (31.5 to 34.3) | 32.7 (31.3 to 34.2) | 0.77 (0.69 to 0.86); <0.001 | 0.98 (0.90 to 1.08); 0.736 | |
| Ethnic minority (n = 2623) | 25.3 (21.3 to 29.5) | 30.3 (27.1 to 33.5) | 24.8 (22.5 to 27.2) | 0.78 (0.60 to 1.02); 0.065 | 0.78 (0.64 to 0.94); 0.011 | |
| Ever smoker (n = 3051) | 69.0 (65.3 to 72.6) | 79.0 (76.7 to 81.2) | 80.9 (78.5 to 83.1) | 0.58 (0.46 to 0.71); <0.001 | 1.10 (0.90 to 1.35); 0.341 | |
| Never smoker (n = 10,775) | 14.3 (12.8 to 15.9) | 18.0 (16.8 to 19.2) | 18.2 (17.1 to 19.3) | 0.75 (0.64 to 0.86); <0.001 | 1.01 (0.90 to 1.13); 0.888 | |
| Regular e-cigarette use | All (n = 14,010) | 1.7 (1.2 to 2.2) | 3.4 (2.9 to 3.9) | 2.6 (2.2 to 3.1) | 0.46 (0.33 to 0.64); <0.001 | 0.76 (0.61 to 0.95); 0.015 |
| Boys (n = 6881) | 1.9 (1.3 to 2.8) | 4.9 (4.1 to 5.8) | 3.6 (2.9 to 4.3) | 0.37 (0.24 to 0.56); <0.001 | 0.71 (0.54 to 0.92); 0.011 | |
| Girls (n = 7129) | 1.4 (0.9 to 2.2) | 2.0 (1.5 to 2.6) | 1.7 (1.3 to 2.2) | 0.71 (0.42 to 1.20); 0.204 | 0.89 (0.61 to 1.31); 0.562 | |
| 13-year-olds (n = 6940) | 1.0 (0.5 to 1.7) | 2.4 (1.8 to 3.0) | 1.4 (1.0 to 1.8) | 0.41 (0.23 to 0.75); 0.004 | 0.57 (0.38 to 0.85); 0.006 | |
| 15-year-olds (n = 7070) | 2.3 (1.6 to 3.2) | 4.3 (3.6 to 5.2) | 3.9 (3.2 to 4.7) | 0.49 (0.33 to 0.73); <0.001 | 0.87 (0.67 to 1.14); 0.316 | |
| White (n = 10,825) | 1.7 (1.2 to 2.3) | 3.3 (2.8 to 3.9) | 2.9 (2.4 to 3.5) | 0.48 (0.33 to 0.69); <0.001 | 0.85 (66.8 to 1.09); 0.202 | |
| Ethnic minority (n = 2623) | 1.5 (0.6 to 3.1) | 2.9 (1.9 to 4.3) | 1.6 (1.0 to 2.5) | 0.53 (0.22 to 1.23); 0.139 | 0.59 (0.33 to 1.05); 0.074 | |
| Ever smoker (n = 3051) | 6.1 (4.4 to 8.2) | 12.3 (10.5 to 14.2) | 11.1 (9.4 to 13.1) | 0.44 (0.30 to 0.63); <0.001 | 0.86 (0.67 to 1.10); 0.231 | |
| Never smoker (n = 10,775) | 0.2 (0.1 to 0.6) | 0.6 (0.4 to 0.8) | 0.5 (0.3 to 0.7) | 0.41 (0.16 to 1.08); 0.071 | 0.88 (0.49 to 1.56); 0.652 | |
Specificity of change in trend: energy drink use in Wales
Energy drink use in Wales was analysed to examine whether changes in trend were specific to e-cigarettes or were also observed for another psychoactive substance (i.e. energy drinks) for which change could not be attributed to the TPD regulations. Descriptive data indicated a somewhat different pattern than for e-cigarette use, with little change in energy drink use from 2013 (53.5%) to 2015 (53.4%), a substantial drop between 2015 and 2017 (43.1%) and a more modest continued decrease from 2017 to 2019 (39.3%). In segmented regression analyses (see Table 15), although not significant, changes in trend of similar magnitude to those for e-cigarettes were also observed for energy drinks in short-term analyses, with a 4% reduction in the odds of energy drink use per month, relative to the baseline trend. However, the high degree of consistency from short- to long-term analyses observed with our e-cigarette models was not replicated for energy drink use, which, in long-term analyses, indicated a significant negative step change at the intervention point and a reduced, although marginally significant, change in trend. Hence, particularly in longer-term analyses, change in trend for energy drink consumption did not mirror change for use of e-cigarettes.
| Sample | Short-term analysis (2013–17) | Long-term analysis (2013–19) | |||
|---|---|---|---|---|---|
| OR (95% CI) | p-value | OR (95% CI) | p-value | ||
| All (n = 52,794a; 94,935b) | Time | 1.00 (1.00 to 1.01) | 0.856 | 1.00 (1.00 to 1.01) | 0.871 |
| Level | 1.38 (0.60 to 3.18) | 0.453 | 0.76 (0.68 to 0.85) | <0.001 | |
| Post slope | 0.96 (0.92 to 1.01) | 0.089 | 0.99 (0.99 to 1.00) | 0.014 | |
| Boys (n = 25,888; 46,682) | Time | 1.00 (0.99 to 1.00) | 0.285 | 1.00 (0.99 to 1.00) | 0.286 |
| Level | 1.32 (0.55 to 3.14) | 0.534 | 0.89 (0.78 to 1.02) | 0.092 | |
| Post slope | 0.97 (0.93 to 1.02) | 0.272 | 0.99 (0.99 to 1.00) | 0.167 | |
| Girls (n = 26,906; 48,253) | Time | 1.00 (1.00 to 1.01) | 0.183 | 1.00 (1.00 to 1.01) | 0.182 |
| Level | 1.46 (0.51 to 4.14) | 0.477 | 0.65 (0.57 to 0.74) | <0.001 | |
| Post slope | 0.95 (0.89 to 1.01) | 0.075 | 0.99 (0.98 to 1.00) | 0.012 | |
| 13-year-olds (n = 29,545; 52,561) | Time | 1.00 (1.00 to 1.01) | 0.261 | 1.00 (1.00 to 1.01) | 0.271 |
| Level | 1.25 (0.44 to 3.57) | 0.673 | 0.69 (0.61 to 0.78) | <0.001 | |
| Post slope | 0.96 (0.90 to 1.02) | 0.150 | 0.99 (0.98 to 1.00) | 0.008 | |
| 15-year-olds (n = 23,249; 42,374) | Time | 1.00 (0.99 to 1.00) | 0.352 | 1.00 (0.99 to 1.00) | 0.350 |
| Level | 1.53 (0.55 to 4.24) | 0.415 | 0.86 (0.75 to 0.99) | 0.037 | |
| Post slope | 0.96 (0.91 to 1.02) | 0.199 | 1.00 (0.99 to 1.00) | 0.219 | |
| White (n = 47,137; 84,455) | Time | 1.00 (0.99 to 1.01) | 0.936 | 1.00 (1.00 to 1.01) | 0.918 |
| Level | 1.38 (0.60 to 3.21) | 0.449 | 0.80 (0.71 to 0.89) | <0.001 | |
| Post slope | 0.96 (0.92 to 1.01) | 0.112 | 0.99 (0.99 to 1.00) | 0.022 | |
| Ethnic minority (n = 4787; 9009) | Time | 1.00 (0.99 to 1.01) | 0.807 | 1.00 (0.99 to 1.01) | 0.831 |
| Level | 2.54 (0.48 to 13.39) | 0.273 | 0.60 (0.48 to 0.76) | <0.001 | |
| Post slope | 0.92 (0.84 to 1.01) | 0.067 | 0.99 (0.98 to 1.01) | 0.409 | |
| High SES (n = 25,644; 40,018) | Time | 1.00 (0.99 to 1.01) | 0.993 | 1.00 (0.99 to 1.01) | 0.989 |
| Level | 1.53 (0.56 to 4.17) | 0.401 | 0.79 (0.69 to 0.90) | <0.001 | |
| Post slope | 0.96 (0.90 to 1.01) | 0.116 | 0.99 (0.99 to 1.00) | 0.032 | |
| Low SES (n = 25,836; 51,951) | Time | 1.00 (0.99 to 1.01) | 0.772 | 1.00 (0.99 to 1.01) | 0.754 |
| Level | 1.31 (0.54 to 3.16) | 0.550 | 0.82 (0.72 to 0.93) | 0.003 | |
| Post slope | 0.97 (0.92 to 1.01) | 0.172 | 0.99 (0.99 to 1.00) | 0.026 | |
| Never smoker (n = 41,457; 75,345) | Time | 1.00 (1.00 to 1.01) | 0.422 | 1.00 (1.00 to 1.01) | 0.433 |
| Level | 1.40 (0.60 to 3.30) | 0.437 | 0.71 (0.64 to 0.80) | <0.001 | |
| Post slope | 0.95 (0.91 to 1.00) | 0.054 | 0.99 (0.99 to 1.00) | 0.004 | |
| Ever smoker (n = 8805; 15,516) | Time | 1.00 (0.99 to 1.01) | 0.794 | 1.00 (1.00 to 1.01) | 0.795 |
| Level | 2.26 (0.63 to 8.07) | 0.210 | 0.79 (0.63 to 0.99) | 0.045 | |
| Post slope | 0.94 (0.87 to 1.00) | 0.065 | 0.99 (0.99 to 1.00) | 0.239 | |
Change in young people’s tobacco use following the Tobacco Products Directive regulations and the 2016 suite of tobacco regulation
Given the longer time series available for tobacco use, secondary analyses of change in trend for young people’s tobacco use drew on the integrated three-country data set used in Chapter 3, with additional data sets from 2016 onward. The analytical samples for smoking prevalence from 1998 to 2019 included 358,713 and 359,111 students for ever and regular smoking, respectively (i.e. 97% of all students within the target year groups).
As indicated in Figure 9, although decline continued during the period when e-cigarettes were growing in popularity (see Chapter 3) there is some evidence of a plateauing in the most recent data points. In Scotland and Wales, ever smoking falls until 2015, with little evidence of decline in subsequent years. However, there was tentative evidence of a continued decline in smoking in England from 2016 to 2018, having remained unchanged from 2014 to 2016. Similarly, although rates of regular smoking flattened in Scotland and Wales in the most recent data sets, there was some continued decline in England. In England, year group data were removed from data sets beyond 2014 and post-TPD analyses used age rather than year group. Differences between estimates for England where using year group (pre-legislation only) or age data are negligible.
FIGURE 9 Percentage of young people reporting ever or regular (weekly or more) smoking by country and year from 1998 to 2019. The black bars represent the beginning of two ‘intervention’ periods (i.e. emergence of e-cigarettes and then their regulation via the TPD). (Figure not reproduced here; see the original article.)
Odds ratios from segmented regression analyses in Table 16 indicate that for the sample as a whole, and among subgroups defined by gender and year group, there was a positive change in trend in both linear and quadratic models for ever and regular smoking. Hence, these models provide consistent evidence that the decline in young people’s smoking was slower in the years following the intervention point than it had been previously. Findings are consistent with a conclusion that the suite of tobacco regulation introduced in 2016 did not maintain or accelerate the downwards trend in young people’s smoking, and that this had begun to plateau. These conclusions were robust to a range of sensitivity analyses (see Report Supplementary Material 1, Table 1). Predicted probability plots for quadratic models of ever and regular smoking (see Report Supplementary Material 1, Figure 2) illustrate that trends for tobacco followed the opposite pattern to those observed for e-cigarette use, falling from 2011 to 2015 (as e-cigarette use was rising) but plateauing thereafter (as e-cigarette use was plateauing or falling).
| Sample | Ever smoked | Regular smoking | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Linear | p-value | Quadratic | p-value | Linear | p-value | Quadratic | p-value | ||
| Whole sample (n = 358,713/359,111) | Year | 0.91 (0.91 to 0.91) | <0.001 | 0.94 (0.93 to 0.95) | <0.001 | 0.93 (0.93 to 0.93) | <0.001 | 0.97 (0.96 to 0.99) | 0.005 |
| Year2 | 1.00 (1.00 to 1.00) | <0.001 | 1.00 (1.00 to 1.00) | <0.001 | |||||
| Level | 1.16 (1.10 to 1.23) | <0.001 | 1.15 (1.08 to 1.22) | <0.001 | 0.96 (0.87 to 1.07) | 0.471 | 0.94 (0.85 to 1.05) | 0.269 | |
| Post slope | 1.07 (1.05 to 1.09) | <0.001 | 1.09 (1.06 to 1.11) | <0.001 | 1.10 (1.06 to 1.15) | <0.001 | 1.13 (1.09 to 1.18) | <0.001 | |
| Boys subgroup (n = 179,174/179,494) | Year | 0.92 (0.92 to 0.93) | <0.001 | 0.92 (0.90 to 0.94) | <0.001 | 0.94 (0.93 to 0.95) | <0.001 | 0.95 (0.92 to 0.97) | <0.001 |
| Year2 | 1.00 (1.00 to 1.00) | 0.819 | 1.00 (1.00 to 1.00) | 0.696 | |||||
| Level | 1.17 (1.07 to 1.27) | <0.001 | 1.17 (1.07 to 1.27) | <0.001 | 0.98 (0.85 to 1.14) | 0.833 | 0.98 (0.84 to 1.14) | 0.808 | |
| Post slope | 1.05 (1.01 to 1.08) | 0.005 | 1.05 (1.01 to 1.08) | 0.009 | 1.09 (1.03 to 1.15) | 0.005 | 1.09 (1.03 to 1.16) | 0.005 | |
| Girls subgroup (n = 179,539/179,617) | Year | 0.90 (0.89 to 0.90) | <0.001 | 0.95 (0.94 to 0.97) | <0.001 | 0.92 (0.92 to 0.93) | <0.001 | 1.00 (0.97 to 1.02) | 0.698 |
| Year2 | 1.00 (0.99 to 1.00) | <0.001 | 1.00 (0.99 to 1.00) | <0.001 | |||||
| Level | 1.16 (1.07 to 1.26) | <0.001 | 1.13 (1.05 to 1.22) | 0.002 | 0.95 (0.82 to 1.09) | 0.454 | 0.92 (0.79 to 1.06) | 0.227 | |
| Post slope | 1.09 (1.06 to 1.13) | <0.001 | 1.13 (1.09 to 1.16) | <0.001 | 1.12 (1.06 to 1.18) | <0.001 | 1.17 (1.10 to 1.23) | <0.001 | |
| 13-year-olds only (n = 187,347/187,534) | Year | 0.89 (0.89 to 0.89) | <0.001 | 0.94 (0.92 to 0.96) | <0.001 | 0.91 (0.90 to 0.92) | <0.001 | 1.00 (0.97 to 1.04) | 0.908 |
| Year2 | 1.00 (0.99 to 1.00) | <0.001 | 0.99 (0.99 to 1.00) | <0.001 | |||||
| Level | 1.21 (1.10 to 1.33) | <0.001 | 1.18 (1.07 to 1.30) | 0.001 | 0.98 (0.80 to 1.19) | 0.821 | 0.93 (0.76 to 1.14) | 0.472 | |
| Post slope | 1.08 (1.04 to 1.12) | <0.001 | 1.11 (1.07 to 1.16) | <0.001 | 1.11 (1.03 to 1.21) | 0.009 | 1.18 (1.08 to 1.28) | <0.001 | |
| 15-year-olds only (n = 171,366/171,577) | Year | 0.93 (0.93 to 0.93) | <0.001 | 0.94 (0.92 to 0.95) | <0.001 | 0.94 (0.93 to 0.94) | <0.001 | 0.97 (0.95 to 1.00) | 0.003 |
| Year2 | 1.00 (1.00 to 1.00) | 0.406 | 1.00 (1.00 to 1.00) | 0.005 | |||||
| Level | 1.12 (1.05 to 1.21) | 0.002 | 1.12 (1.04 to 1.21) | 0.002 | 0.95 (0.84 to 1.08) | 0.443 | 0.94 (0.83 to 1.06) | 0.330 | |
| Post slope | 1.06 (1.04 to 1.09) | <0.001 | 1.07 (1.04 to 1.10) | <0.001 | 1.09 (1.04 to 1.14) | <0.001 | 1.11 (1.06 to 1.16) | <0.001 | |
To examine whether change in trend was specific to tobacco or were observed across other substances, we provide estimates over time for cannabis use (see Table 17). Descriptive data (see Report Supplementary Material 1, Table 16) indicate plateauing or slight increase in ever use of cannabis, with segmented regression analyses showing a similar disruption to secular decline in ever cannabis use to that observed for smoking. Consistent with descriptive data, which provide some suggestion of growth in cannabis use, as opposed to flattening in tobacco use, changes in trend are consistently somewhat larger for cannabis use than for tobacco. We did not formally model whether or not this change in trend was also observed for alcohol use because of changes in how this question was asked from 2016 in England (which gave rise to an artificial stepped increase in estimates of alcohol use in England, and mean the English data cannot be treated as a single time series pre and post 2016). However, in Wales, there was only marginal further decline in ever drinking (from 54.3% in 2015 to 53.2% in 2019) and in Scotland there was an increase in ever drinking (from 44.5% in 2015 to 52.5% in 2018). Hence, plateauing in tobacco use observed in the most recent surveys appears not to be unique to tobacco, but is reflected in other substance use behaviours. In line with between-country differences in observed smoking trends, for smoking attitudes, although not formally modelled as only available in three post-legislation data sets, there were small declines beyond legislation in England in the percentage of young people reporting that it is OK for someone their age to try smoking (with 34.1% of young people saying that it is OK to try smoking in 2014 vs. 30.1% of young people in 2018), whereas in Scotland there was a small increase in the percentage of young people reporting that it is OK for someone their age to try smoking (from 27.0% of young people in 2015 to 30.5% of young people in 2018).
| Sample | Cannabis use | ||||
|---|---|---|---|---|---|
| Linear | p-value | Quadratic | p-value | ||
| Whole sample (n = 355,422) | Year | 0.92 (0.92 to 0.93) | <0.001 | 1.08 (1.06 to 1.10) | <0.001 |
| Year2 | 0.99 (0.99 to 0.99) | <0.001 | |||
| Level | 1.37 (1.27 to 1.47) | <0.001 | 1.28 (1.18 to 1.37) | <0.001 | |
| Post slope | 1.07 (1.04 to 1.11) | <0.001 | 1.17 (1.14 to 1.21) | <0.001 | |
| Boys (n = 177,463) | Year | 0.93 (0.92 to 0.93) | <0.001 | 1.06 (1.04 to 1.08) | <0.001 |
| Year2 | 0.99 (0.99 to 0.99) | <0.001 | |||
| Level | 1.35 (1.35 to 1.22) | <0.001 | 1.27 (1.15 to 1.40) | <0.001 | |
| Post slope | 1.06 (1.02 to 1.10) | 0.005 | 1.14 (1.09 to 1.19) | <0.001 | |
| Girls (n = 177,959) | Year | 0.92 (0.91 to 0.92) | <0.001 | 1.10 (1.07 to 1.13) | <0.001 |
| Year2 | 0.99 (0.99 to 0.99) | <0.001 | |||
| Level | 1.40 (1.26 to 1.56) | <0.001 | 1.29 (1.16 to 1.44) | <0.001 | |
| Post slope | 1.10 (1.05 to 1.14) | <0.001 | 1.21 (1.16 to 1.27) | <0.001 | |
| 13-year-olds (n = 184,231) | Year | 0.89 (0.89 to 0.90) | <0.001 | 1.13 (1.09 to 1.17) | <0.001 |
| Year2 | 0.98 (0.98 to 0.99) | <0.001 | |||
| Level | 1.41 (1.22 to 1.63) | <0.001 | 1.26 (1.09 to 1.46) | 0.002 | |
| Post slope | 1.11 (1.04 to 1.17) | <0.001 | 1.26 (1.18 to 1.34) | <0.001 | |
| 15-year-olds (n = 171,191) | Year | 0.93 (0.93 to 0.93) | <0.001 | 1.06 (1.04 to 1.08) | <0.001 |
| Year2 | 0.99 (0.99 to 0.99) | <0.001 | |||
| Level | 1.36 (1.25 to 1.47) | <0.001 | 1.28 (1.18 to 1.39) | <0.001 | |
| Post slope | 1.06 (1.03 to 1.10) | <0.001 | 1.14 (1.10 to 1.18) | <0.001 | |
For descriptive information for all variables overall and by subgroup, from 1998 to 2019, see Report Supplementary Material 1, Tables 9–19.
Chapter summary
Following the TPD regulations, e-cigarettes were increasingly described as something young people had tried, but were now losing interest in. Interactions with elements of the TPD regulations, such as device labelling, were described as limited, although risk perceptions for e-cigarettes relative to tobacco were increasing. Although some stakeholders in Chapters 3 and 4 emphasised emergence of e-cigarette markets on the internet as an unintended consequence of TPD, few young people described online purchase or having switched to purchasing e-cigarettes via the internet. In qualitative interviews, young people described little exposure to marketing, although, by contrast, surveys indicated that most young people continued to see some form of advertising in the years after implementation, with point of sale the most common source of exposure followed by the internet. Our analyses of survey data consistently indicated that young people’s experimentation with e-cigarettes, and regular use of e-cigarettes, had grown prior to TPD regulations, but did not continue to do so afterward, which is consistent with a hypothesised effect of the TPD regulations. However, secondary analyses of changes in trend for young people’s use of tobacco indicate that the suite of tobacco and e-cigarette regulation in May 2016 may not have been sufficient to maintain downwards pressure on young people’s smoking, with smoking rates beginning to plateau. Tobacco use exhibited the inverse trend to that observed for e-cigarettes, declining when e-cigarette use increased, but now plateauing as e-cigarette use begins to decline.