Smoking prevalence and purchasing of menthol cigarettes since the menthol flavour ban in Great Britain: a population-based survey between 2020 and 2023

Study design

Data for this study were drawn from the Smoking Toolkit Study, an ongoing monthly population-based survey including demographic and smoking-related questions [17]. This study includes data collected between October 2020 and March 2023 from England, Scotland, and Wales. Prior to the analysis, the study protocol was published on the Open Science Framework (https://osf.io/s8mjr/). The manuscript followed the Strengthening the reporting of observational studies in epidemiology (STROBE) statement [18]. Data collection was conducted by a market research company using a combination of random location and quota sampling. Anonymised data were provided to the research team. The data collection method changed in March 2020 from face-to-face to telephone surveys. Studies showed similar results when comparing the two data collection methods [19–21].

Outcome variables and covariates

The primary outcome measures were the prevalence of menthol cigarette smoking as a proportion of all adults who smoke and specifically young adults smoking (18-24 years) and, for each purchasing source, the proportion of individuals smoking menthol cigarettes who states that they purchased cigarettes through this source. First, participants were grouped according to whether they currently smoked cigarettes. Then, they were further classified based on whether they smoked menthol cigarettes (or for sensitivity analysis, any flavoured cigarettes). All variables are listed in Table 1 and are based on self-report.

Analysis

The analysis was conducted in RStudio (version 2022.07.2, R version 4.2.1). For each variable included in the analysis, the number and percent of missing values are reported in the supplement (Table S1). All values that the interviewer noted as “Don’t know” or “Refused” were assumed to be missing. The study is based on a complete case analysis. For the first research question, the weighted prevalence of smoking menthol cigarettes as a proportion of all adults in Great Britain who smoke cigarettes was assessed for each quarter between October 2020 and March 2023. Prevalence was reported for quarters rather than months due to small samples in 18-to-24-year-olds, in Scotland and in Wales. Data were weighted using raking to match the population of Great Britain [25]. The weighted prevalence was plotted over time. Further, a logistic regression model with a restricted cubic spline function (3 knots at beginning, middle, and end of time series) was fitted to assess weighted prevalence as a proportion of people smoking menthol cigarettes over time. Prevalence ratios (PR) and corresponding 95% compatibility intervals (CI, using bootstrapping with n=2000 replicates (Cole et al., 2020; Efron & Tibshirani, 1994; Hawkins & Samuels, 2021; Rafi & Greenland, 2020)) were calculated for the change in prevalence comparing the first quarter (Q4 2020) to the last quarter (Q1 2023). Additionally, the weighted prevalence was assessed for young adults aged 18 to 24 years in Great Britain. For the second research question, the weighted prevalence of smoking menthol cigarettes as a proportion of those who smoke cigarettes was assessed for each nation (England, Scotland, and Wales) over time (in quarters). The weighted prevalence of smoking menthol cigarettes was plotted over time for each nation and PRs and corresponding 95% CIs (using bootstrapping with n=2000 replicates) calculated for the change in prevalence in each nation across the study period. Additionally, logistic regression including an interaction term between time and nation was used to calculate the PR ratio (unadjusted and adjusted for age, gender, social grade) of smoking menthol cigarettes among those who smoke cigarettes in Scotland and Wales compared to England (i.e., dividing the PR of Scotland or Wales, respectively, by the PR of England). Corresponding 95% CIs were computed using bootstrapping (n=2000 replicates).

In England, the question about purchasing sources of cigarettes was only asked once per quarter since April 2022 (i.e., in April 2022, July 2022, October 2022, and January 2023). Therefore, months in which data were only collected in Scotland and Wales were excluded from the analysis for the third research question. The weighted proportion of people who smoked menthol cigarettes stating that they purchased cigarettes by the various sources (not mutually exclusive) was computed for the entire study period (excluding the above-mentioned months). Further, weighted proportions of licit, illicit, and cross-border purchases were assessed. The weighted proportion for each purchase source was compared between those who smoked menthol cigarettes and those who did not smoke flavoured cigarettes using Chi-square statistics and Cramer’s V as a measure of the effect size (following the interpretation by Cohen ²⁶, categorising Cramer’s V into small, medium, and large effect sizes). Further, purchase sources were compared between nations. In sensitivity analyses, the research questions were assessed including all people who stated that they smoked flavoured cigarettes (menthol or some other flavour) instead of just those who stated that they smoked menthol cigarettes. Further, the prevalence of smoking menthol cigarettes among all adults and specifically young adults in all of Great Britain and the prevalence of smoking menthol cigarettes among all adults separately in the three nations were assessed.

Sources of purchasing

Table 4 shows the sources of purchase among those who smoked menthol cigarettes compared to those only smoking non-flavoured cigarettes (unweighted data in supplement, Table S5). The main sources were newsagents/off license/corner shops and supermarkets. There were no noteworthy differences between the two groups. Most participants reported purchasing through licit sources (for menthol cigarette smoking: 93.9%, 95% CI: 92.2-95.5; for non-flavoured cigarette smoking: 93.5%, 95% CI: 92.7-94.2). Illicit sources of purchase were reported by 14.8% (95% CI: 12.2-17.3) of those smoking menthol cigarettes and 12.5% (95% CI: 11.5-13.5) of those smoking only non-flavoured cigarettes. Cross-border purchases were reported by 11.5% (95% CI: 9.2-13.8%) of participants smoking menthol cigarettes and 9.9% (95% CI: 9.0-10.8) of participants smoking only non-flavoured cigarettes.

Table 5 shows the sources of purchasing cigarettes among those who smoked menthol cigarettes in England, Scotland, and Wales (unweighted data in supplement, Table S6). All differences between nations were non-significant. The main sources of purchase for all nations were small shops (i.e., newsagent, off license, or corner shop) and supermarkets.

Sensitivity analyses

Additional sensitivity analyses including all participants who reported that they smoked flavoured cigarettes (instead of only those who reported smoking menthol cigarettes) and the prevalence among all adults (i.e., the proportion of individuals smoking menthol cigarettes among all participants) are presented in the supplement (Tables S7-S13 and Figures S1-S4). The first sensitivity analysis including all people who smoked any flavoured cigarettes did not show meaningful differences compared to the main analysis. In the second sensitivity analysis investigating prevalence of smoking menthol cigarettes as a proportion of all adults (or all 18-to-24-year-olds, respectively), the difference in the prevalence between all adults and 18-to-24-year-olds became more pronounced (see Table S12 and Figure S3). Also, the decrease in prevalence was greater among 18-to-24-year-olds when calculating the prevalence among all participants (PR=0.64, 95% CI: 0.53-0.77, modelled prevalence for Q4 2020: 5.7%, and for Q1 2023: 3.7%) compared to the prevalence among only those who smoked (PR=0.75, 95% CI: 0.63-0.89, modelled prevalence for Q4 2020: 25.7%, and for Q1 2023: 19.4%). For the between-nation comparison, the results for prevalence among all adults and prevalence among those who smoked were comparable.

Summary of findings

Despite being banned in 2020, one million adults continue to smoke menthol cigarettes in Great Britain. The prevalence of menthol cigarette smoking only decreased slightly and non-significantly among adults who smoke, from 16% at the end of 2020 to 14% at the beginning of 2023. During the same period, the prevalence among 18-to-24-year-olds dropped by a quarter from 26% to 19%. These figures show that despite the ban, menthol cigarette smoking remains common among people who smoke in Great Britain, used by roughly 1 in 7 adults who smoke and 1 in 5 among young adults who smoke. The only nation with a substantial decline in the prevalence was Wales, where it fell by two-thirds from 23% to 8%. Compared with England and Scotland, Wales started off with a higher prevalence, but by the beginning of 2023 it had the lowest prevalence. In contrast, the prevalence in England and Scotland remained relatively stable throughout the period, but in Scotland it was consistently lower than in England.

It is unclear why the trend in Wales differed from the other two nations. Potential contributors could include differences in government approaches and differences in purchasing sources. Since our data on purchasing sources did not differentiate whether people bought menthol cigarettes or accessories to mentholate their cigarettes, we may have missed differences in purchasing sources by nation which are specific to menthol cigarettes. Another explanation could be that the tobacco industry focused its marketing tactics for legal menthol accessories on larger urban areas, mainly in England, rather than less populous localities found in Wales. A further, more unlikely explanation is that people who smoked menthol cigarettes in Wales at the start of the study period were more likely to have quit smoking altogether. When comparing England and Scotland, a potential explanation for the lower overall prevalence in Scotland could be the display ban for tobacco-related accessories in Scotland. Whereas tobacco products (cigarettes, loose tobacco, cigars, etc.) are subject to a display ban in all three nations, only Scotland bans tobacco product accessories from being visible at the point of sale. This difference in legislation could mean that people living in Scotland may be less aware of menthol-flavoured accessories.

There were no noteworthy differences in purchasing sources between menthol and non-flavoured cigarettes; small shops and supermarkets were the most popular places to buy both kinds of cigarettes. There was no evidence that people smoking menthol cigarettes were particularly likely to obtain these from abroad or through illicit sources. Rather, people continued to buy menthol cigarettes in regular shops. This indicates that they either purchased menthol-flavoured accessories or cigarettes they perceived to be mentholated, but which are not labelled as such by the manufacturer.

Comparison to existing literature

Canada was one of the first countries to introduce a menthol ban and, in contrast to the UK and the European Union, it completely prohibited menthol and its analogues and derivatives in cigarettes nationwide from 2017 [28]. In a pre-post ban comparison, only 20% of people who reported smoking menthol cigarettes before the ban continued doing so afterwards, while 59% switched to non-menthol cigarettes and 22% stopped smoking entirely [29]. In a Dutch longitudinal study, people who smoked menthol cigarettes before the ban there (in March 2020) were asked again about their cigarette smoking behaviour in June-July 2021 [30]. The results showed that 33% of them reported still smoking menthol cigarettes as their usual brand, 40% switching to non-menthol cigarettes, 1% smoking cigarettes with unknown flavour, and 26% having quit smoking. These two studies demonstrate that in other countries people also continued smoking menthol cigarettes despite the ban.

Data from Canada also refute the hypothesis that menthol cigarette bans lead to a surge in illicit purchases [31]. Similarly, Laverty et al. [32] found that the implementation of standardised packaging in the European Union did not lead to an increase in the availability of illicit cigarettes. These findings strengthen the alternative hypothesis that people who continue menthol cigarette smoking after the ban either buy menthol-flavoured accessories or cigarettes that are perceived as being mentholated by consumers but not labelled as such by the manufacturers [13, 29, 30, 33, 34].

Limitations

This study has several limitations. It is cross-sectional and it is based on self-reported data. No data were available on menthol cigarette smoking prevalence and the purchasing sources for those smoking menthol cigarettes prior to the ban because the question about menthol smoking was only included in the survey after the legislation came into force. The purchasing sources are not specific to where menthol cigarettes were purchased (i.e., they refer to any cigarettes) and it is difficult to clearly classify some of the sources as licit or illicit (e.g., pub behind the bar). One should also note that all the sources of purchase for menthol cigarettes, including supermarkets, could in theory be illicit given the menthol ban in the UK. Further, the question about menthol cigarette smoking in the survey does not provide information about which type of menthol cigarette participants were smoking.