Article Text
Abstract
Background Ontario, Canada prohibited menthol tobacco product sales beginning 1 January 2017. We measured retail sales of menthol cigarettes and possible substitute products before and after policy implementation in Ontario.
Methods We licensed retail scanner data for tobacco product sales in Ontario and British Columbia (BC), a comparison province without a menthol tobacco policy at that time. We assessed changes in per capita unit sales (per 1000 people) from pre-policy (January–June 2016) to post-policy (January–June 2017) periods. Classification of cigarettes as menthol or non-menthol, or having menthol-suggestive descriptors (‘green’, ‘blue’, ‘silver’ and ‘fresh’), was based on scanner data.
Results Ontario menthol cigarette sales decreased 93%, from 596 to 40 packs per capita compared with a 2% decrease (696 to 679 packs per capita) in BC. Menthol capsule cigarette sales remained low in Ontario (<1% of total cigarette sales) but rose sixfold in BC. Although cigar sales data were unavailable, substitution appeared minimal; sales of non-menthol cigarettes increased 0.4% in Ontario (11 470 to 11 519 packs per capita) while vaping product sales decreased. Ontario had a larger increase in sales of cigarettes with menthol-suggestive descriptors (11% increase) than BC (3% increase). In Ontario, nearly all (>99%) pre-policy sales of cigarettes with ‘green’ menthol-suggestive descriptors were menthol cigarettes, but post-policy, 94% of ‘green’ cigarettes sold were non-menthol.
Conclusions Ontario’s menthol policy was associated with a decrease in retail sales of cigarettes classified as menthol, with little evidence of product substitution. Understanding changes in sales of cigarettes with menthol-suggestive descriptors would be informative.
- public policy
- non-cigarette tobacco products
- surveillance and monitoring
Data availability statement
No data are available.
Statistics from Altmetric.com
Introduction
Menthol in cigarettes and other tobacco products has been linked with increased tobacco product initiation, particularly among youth and young adults, and with greater dependence compared with non-menthol tobacco products.1 2 Policies have been implemented to restrict menthol tobacco product sales in multiple countries and in subnational jurisdictions in the USA and Canada.3 4 Additionally, in April 2021, the US Food and Drug Administration announced a commitment to ban menthol as a characterising flavour in cigarettes and to ban menthol and other characterising flavours in cigars.5 On 1 January 2017, 9 months before Canada implemented a prohibition on the manufacture and sale of cigarettes, blunt wraps and most cigars containing menthol (Tobacco and Vaping Products Act, R.S.C. 2018, c. L-9, sec. 69 (Canada)), the province of Ontario implemented a sales restriction that prohibited the sale of menthol tobacco products including cigarettes, cigars weighing less than 6 g, smokeless tobacco and hookah (Making Healthier Choices Act, R.S.O. 2015, c. P.7 (Canada)) (we use ‘sales restriction’ or ‘policy’ rather than ‘ban’ to describe this province-level policy that prohibits the sale of some menthol tobacco products). The Ontario policy exempted pipe tobacco and cigars weighing more than 6 g. The policy did not apply to vaping products, which are not classified as tobacco products in Canada.6
The intent of Ontario’s menthol tobacco product sales restriction was to reduce youth tobacco use initiation and to reduce the prevalence of tobacco use (H.R. 84, 2015 Leg., 41st Sess. at 4351 (Ontario 13 May 2015)), and lawmakers cited data related to menthol’s role in youth initiation and dependence in their support for the legislation (H.R. 37, 2014 Leg., 41st Sess. at 1751 (Ontario 3 December 2014)). In 2015, prior to the menthol sales restriction, cigarette smoking prevalence among those 15 years and older was 13.0% in Canada and 11.3% in Ontario.7 Youth and young adult smokers in Canada more often reported menthol use than older smokers (15.1% of Canadian smokers aged 15–19 years and 19.5% of those aged 20–24 years reported past 30-day menthol use in 2015, compared with rates below 12% among adults aged 25 years and older).8 After the menthol restriction, Ontario menthol smokers aged 16 years and older were more likely than non-menthol smokers to report making a quit attempt and quitting smoking, both 1 month after the menthol restriction and 1 year later.9 10 Similarly, a survey of smokers in seven Canadian provinces found that menthol smokers were more likely to quit smoking than non-menthol smokers after menthol sales prohibitions were implemented at the provincial or federal level.11
Reports by tobacco manufacturers to Health Canada indicate reduced wholesale menthol cigarette sales following implementation of Ontario’s menthol policy.12 However, studies have not yet explored how the policy affected retail sales, which could corroborate or conflict with wholesale data trends. Retail sales data can provide information about retail availability of tobacco products and some evidence of population-level consumption and substitution effects and potential unintended consequences of the policy. For example, researchers in Canada identified the retail availability of menthol-flavoured capsule cigarettes following the announcement of the Ontario menthol restriction.13 These products contain a crushable capsule in the filter that can impart different flavours, and are relatively new to the global market.13–15 Capsule cigarettes are subject to Ontario’s menthol restriction if they are menthol flavoured.
Menthol smokers might react to a sales restriction by substituting other products, such as non-menthol cigarettes or vaping products,16 which were not restricted by Ontario’s menthol policy. Cigarettes are more commonly used in Canada than are other tobacco products; only 2.5% of Canadians 15 years and older reported past 30-day use of cigars or cigarillos in 2015; prevalence rates of hookah/waterpipe and smokeless tobacco were less than 1%.8 Analysis of wholesale sales data found a small increase in non-menthol cigarette sales in Ontario after the menthol restriction, although this increase did not fully offset the decrease in wholesale menthol cigarette sales.12
Studying policy effects on retail product sales can provide more granular insights into consumer purchases, including possible substitution behaviour, than can be inferred from data on the movement of tobacco product inventories from manufacturers to wholesalers. Retail sales data can also detect and quantify sales of non-menthol cigarettes with menthol-suggestive descriptors. Researchers have documented menthol cigarette brand replacements in Canada, with tobacco companies promoting non-menthol cigarettes in packages that use text and colours associated with menthol or potential reduced risk, such as ‘green’, ‘blue’ and ‘smooth’.17 18 This relabelling is reminiscent of tobacco manufacturers’ use of colours to replace ‘light’ and ‘mild’ descriptors on cigarettes after the 2009 Family Smoking Prevention and Tobacco Control Act banned use of misleading descriptors in the USA.19 Similarly, analyses of tobacco product sales in the USA have documented an increase in non-cigarette tobacco products with concept flavour names,20 making policy enforcement more challenging21 and mitigating the effect of flavoured tobacco policies.22 Absent from the literature is any quantification of menthol-suggestive descriptor cigarette sales after implementation of a menthol sales policy.
This paper explores changes in retail sales of tobacco products before and after implementation of the policy in Ontario, as well as coincident changes in a non-policy comparison province. We measured changes in sales of menthol cigarettes, capsule cigarettes, smokeless tobacco and roll-your-own (RYO) tobacco. We explored possible quitting behaviour among smokers by tracking sales of non-prescription nicotine replacement therapy (NRT) products (eg, nicotine patches, nicotine gum), and possible product substitution by measuring changes in sales of vaping products and non-menthol cigarettes, smokeless tobacco and RYO tobacco. We also investigated a potential unintended consequence of the policy by measuring changes in menthol-suggestive descriptors on cigarette packs.
Methods
We obtained customised retail scanner sales data from The Nielsen Company (Nielsen) for tobacco and nicotine products in two Canadian provinces: Ontario and British Columbia (BC), with BC selected as the comparison area to Ontario because it is the most populous Canadian province that did not have a menthol restriction during the study period. Both provinces had pre-existing smoke-free laws and bans on retail tobacco displays, and neither province implemented other tobacco control policies during the study period. Patterns of cigarette smoking prevalence in the years leading up to the policy were relatively similar for the two provinces.8 Nielsen collected scanner data from a sample of retailers in each province and applied proprietary weighting methods to project province-wide sales by all retailers in specified channels. Cigarette and RYO tobacco data represent aggregate sales from food, department, gas and convenience stores; data on vaping products and NRT represent aggregate sales from grocery, pharmacy, mass merchandiser, warehouse clubs, gas and convenience stores. Smokeless tobacco data represent aggregate sales from gas and convenience stores. Cigar sales data for the provinces were not available from Nielsen. For available product categories, we obtained sales records for each province during two periods: the first 24 weeks of 2016 in aggregate (10 January–25 June; pre-policy); and the first 24 weeks of 2017 in aggregate (8 January–24 June; post-policy). We focused on the 24 weeks immediately after policy implementation to avoid overlap with Canada’s federal menthol restriction, which was implemented in October 2017. We compared post-policy sales to the first 24 weeks of 2016 to control for seasonal effects and possible pre-policy hoarding by consumers. Data were provided at the Universal Product Code (UPC) level, allowing for product-specific analysis and investigation of UPC-aggregated sales by characteristics such as flavour and package colour. Nielsen coded product characteristics based on product packaging.
We classified cigarettes and RYO tobacco as menthol or non-menthol as reported by Nielsen. We included clove cigarettes with menthol cigarettes, as Ontario’s policy restricted menthol and clove at the same time; less than 0.01% of cigarette sales during pre-policy and post-policy periods were clove. We used Nielsen-provided product descriptors to classify smokeless tobacco and vaping products as menthol, tobacco/unknown or other flavour. In cases where the flavour categorisation was unclear (eg, ‘Shoot the Moon’), brand-specific internet searches were conducted to identify manufacturer-provided or retailer-provided flavour descriptions or product images that indicated the appropriate flavour categorisation, when available. The tobacco/unknown category includes products with no flavour description available in the data, products with flavour descriptions indicative of tobacco or no flavour (eg, ‘tobacco’, ‘original’), and products with flavour descriptions that could not be linked to a characterising flavour by name or using internet searches. We classified products as ‘other flavour’ if they had flavour descriptions indicative of a characterising flavour other than menthol or tobacco (eg, ‘cherry’), or were described as having a characterising flavour during internet searches. We classified mint-flavoured tobacco products as ‘other flavour’ because Ontario’s policy restricted sales of flavoured tobacco products, including mint, 1 year earlier than the menthol sales restriction.
Separately, we identified capsule cigarettes as those having one or more of the following descriptors in the Nielsen dataset, based on information found from internet searches of capsule cigarettes: ‘Convertible’, ‘Capsule’, ‘Click’, ‘Crush’, ‘Duo’, ‘Switch’, ‘Burst’, ‘Hybrid’, ‘Unison’, ‘Aqua Filter’, ‘Synchro’, ‘Snap’ or ‘Boost’. We analysed capsule cigarettes by the Nielsen-reported menthol/non-menthol designation.
We also identified cigarettes with menthol-suggestive descriptors, which may or may not be classified as menthol in the Nielsen data or explicitly labelled as menthol on the product package. We considered products as having menthol-suggestive descriptors if the Nielsen product characteristics included one or more of the following terms: ‘Green’, ‘Blue’, ‘Silver’, ‘Black’, ‘White’, ‘Smooth’ or ‘Fresh’, since these were found to relate to menthol cigarettes in a previous analysis18 or in the present data. For example, in the current dataset, the product characteristic ‘green’ sometimes has a flavour descriptor of ‘menthol’ and other times does not. We also searched the dataset for the following product descriptors potentially related to menthol: ‘Ice’, ‘Blast’, ‘Winter’, ‘Cool’, ‘Polar’, ‘Arctic’, ‘Freeze’, but they were not present in this dataset. We analysed sales of cigarettes with menthol-suggestive descriptors by the Nielsen-reported menthol classification to assess changes from pre-policy to post-policy periods. Additionally, because green packaging is commonly associated with menthol,17 23 we conducted a focused analysis of product sales with the characteristic ‘green’, a subset of menthol-suggestive descriptors. These products were analysed overall and by the Nielsen-reported menthol classification.
We standardised unit sales such that each analytical unit equates to the most frequently occurring package size by product type within the dataset, following previously documented methods.24 25 The unit of standardisation in this study may differ from other similar studies since it relies on the most frequently occurring package sizes purchased, which may differ for numerous reasons, such as consumer preferences, tobacco industry marketing/supply changes and regional differences. One analytical cigarette unit or RYO tobacco unit equals a single pack of 20 cigarettes or the equivalent amount of RYO tobacco. For smokeless tobacco, one analytical unit equals 40 g of chewing tobacco or 34 g of moist snuff. For vaping products, one analytical unit equals one bottle of e-liquid (10 mL), one prefilled cartridge, one disposable vaping product or one rechargeable kit. For NRT, one analytical unit equals one 105-piece package of gum or lozenges, one seven-count package of patches, or one bottle of nicotine spray or liquid.
We aggregated UPC-level standardised unit sales within each province and period and used 2016 Canadian Census data26 to calculate provincial per capita sales by dividing unit sales by total population and multiplying by 1000. We also calculated the proportion of sales (‘market share’) by province, product and flavour category for each period. For Ontario and BC, we analysed the per cent change in per capita standardised unit sales and market share by product type and flavour category between the first 24 weeks of 2016 and the first 24 weeks of 2017. All analyses were conducted using Stata V.15.1.
Results
In Ontario, menthol cigarette retail sales decreased 93.2% from 596 packs per capita in the pre-policy period to 40 packs per capita in the post-policy period (table 1). Unit sales of menthol cigarettes in BC decreased 2.3% during the same time. The market share of menthol cigarettes dropped from 4.9% to 0.4% in Ontario and remained constant at 6.0% in BC. Per capita sales of non-menthol cigarettes increased by 0.4% in Ontario. Non-menthol cigarette per capita unit sales decreased by 3.2% in BC. Total cigarette sales decreased by 4.2% in Ontario and by 3.2% in BC.
Capsule cigarette sales per capita were lower in Ontario than in BC both before and after the policy was implemented (table 1). Total capsule cigarette market share stayed constant at approximately 1% in Ontario in the pre-policy and post-policy periods, while capsule cigarette market share in BC increased more than sixfold, from 1.9% to 12.2% of all cigarettes sold. The market share of menthol-flavoured capsule cigarettes in BC increased from 5.2% in the pre-policy period to 13.3% in the post-policy period.
Per capita sales of cigarettes with menthol-suggestive descriptors were higher in Ontario than in BC during both study periods. Sales increased by 2.7% in Ontario and decreased by 2.6% in BC after implementation of the Ontario policy. Sales of cigarettes with menthol-suggestive descriptors that were classified as menthol by Nielsen decreased 94.7% in Ontario and 10.6% in BC.
Total unit sales of cigarettes with ‘green’ in the description (a subset of menthol-suggestive descriptors) decreased 2.5% in Ontario from pre-policy to post-policy periods, and the distribution of sales by menthol status differed across the periods. One hundred per cent of sales of cigarettes with ‘green’ in the description were classified as menthol in the pre-policy period in Ontario, compared with 5.9% classified as menthol in the post-policy period. In BC, 100% of ‘green’ pack sales were menthol in both study periods.
Per capita sales of assessed non-cigarette tobacco products were dramatically lower than cigarette product sales in both provinces, and nearly 100% of sales of RYO tobacco and smokeless tobacco were non-menthol in the pre-policy and post-policy periods (table 2). Per capita sales of RYO tobacco increased slightly from 45 to 49 units per capita in Ontario from pre-policy to post-policy, and remained considerably lower than in BC, where sales shifted from 549 to 457 units per capita. In both provinces, nearly 100% of all RYO tobacco sales in the pre-policy and post-policy periods were non-menthol. Smokeless tobacco sales per capita also experienced a decrease of approximately 13% in both provinces. No menthol-flavoured smokeless tobacco sales were observed in the data. Vaping product per capita unit sales were low and decreased from the pre-policy to post-policy periods in both provinces, from five to four units in Ontario and six to one unit in BC.
We also measured sales of NRT before and after policy implementation for both Ontario and BC. Per capita NRT sales increased by over 80% pre-policy to post-policy in both provinces, from 78 to 140 units in Ontario and 93 to 175 units in BC (data not shown).
Discussion
After Ontario’s menthol tobacco sales restriction went into effect, retail sales of menthol cigarettes in Ontario decreased substantially from pre-policy levels while sales of these products were relatively unchanged in BC, the comparison province. This suggests that policy-intended changes in retail sales of menthol cigarettes were associated with policy implementation in Ontario. These findings are consistent with an analysis of wholesale sales data12 and with self-reported data from a sample of Ontario menthol cigarette smokers.10 Using a complementary data source, this study expands the evidence base on the effectiveness of Ontario’s comprehensive menthol sales restriction. Although the increase in NRT sales is suggestive of higher rates of quitting behaviour in Ontario, we observed a similar increase in BC.
The divergent patterns of capsule cigarette sales in the two study provinces suggest that the Ontario policy may also have suppressed the surge of capsule cigarette sales seen in BC, even though the tobacco industry marketed menthol capsule cigarettes in Ontario around the time of the policy.18 The growth in capsule cigarette sales in BC, which might have been driven by industry marketing actions,13–15 18 highlights the evolving product landscape and the importance of ongoing surveillance of the marketplace and adoption of tobacco control policies to address emerging products.
This exploration of multiple tobacco product categories identified only minimal indication of potential substitution. Our results indicate that although per capita retail sales of menthol cigarettes decreased substantially, per capita sales of non-menthol cigarettes increased slightly, which is consistent with an analysis of wholesale sales patterns of menthol and non-menthol cigarettes in Ontario.12 We did not find evidence of shifts to non-menthol RYO tobacco, non-menthol smokeless tobacco or to vaping products (which were not subject to flavour or menthol restrictions). Although the current study did not measure cigar sales, a separate study found some evidence of association between menthol cigarette smoking pre-policy and post-policy past-year use of non-cigarette tobacco products, with those substitution effects being stronger for flavoured cigars than other tobacco products.9
We observed a small but measurable level of menthol tobacco sales in the post-policy period in Ontario, which might reflect actions by retailers to rid themselves of policy-restricted stock, even though retailers were not legally permitted to conduct sell-downs during 2017. If so, any retailer sell-down actions following policy implementation could introduce a conservative bias to our analysis, making the 93% decrease in menthol cigarette sales in Ontario even more compelling.
The observed changes in market share of products with menthol-suggestive descriptors or colours raise interesting questions about the potential unintended policy consequence of manufacturer use of colours to indicate cigarette characteristics (see Connolly and Alpert19). In particular, nearly all pre-policy period sales of cigarettes with ‘green’ descriptors in Ontario were menthol cigarettes, while nearly all post-policy sales of cigarettes with these descriptors were non-menthol; this might indicate changes to the products that were sold or in how the products were marketed. Other studies have reported the use of ‘green’ and other descriptors in place of ‘menthol’ descriptors, also noting that cigarette packs and marketing materials describe ‘non-menthol alternatives’.17 18 Addressing questions regarding menthol content could be conducted via review of ingredient lists and product testing for constituents and flavourings.
This study’s strengths include its use of a comparison province and inclusion of retail sales data for multiple tobacco products and NRT sales. We also rigorously applied systematic methods for assignment of products into flavour categories, similar to those used in other studies (see Gammon et al 20), and identification and classification of flavour and menthol package descriptors.
This study has some limitations. First, this analysis did not include cigar sales, as these data were not available. Second, Nielsen provided only an aggregate, projected sales volume for each UPC in the pre-policy and post-policy periods and no measure of estimation variance for these projections. Therefore, we were unable to test the significance of pre-policy to post-policy sales changes at the product level (eg, changes in total cigarette sales within province). However, because retail sales data are projected by Nielsen to represent the population of sales in the specified channels, observed changes in product sales are meaningful at the population level. Third, as with similar studies, retail scanner data were collected only from participating retailers using UPC scanners; data from tobacco specialty shops, vape shops, small groceries, First Nations retailers and online retailers are not available. Sales data may not represent changes in consumer buying preferences (eg, smokers may shift to purchasing through outlets without UPC scanners) or tobacco product use. Fourth, we could not determine whether colours listed in Nielsen product descriptions reflect packaging text and/or colour, although both are relevant to this assessment. Finally, we did not measure changes in illicit or cross-border sales, but other studies of Canadian menthol tobacco restrictions have not revealed a surge in illicit cigarettes11 27 and bordering provinces also experienced decline in wholesale sales of cigarettes from 2016 to 2017.27 28 In addition, cross-border sales may be relatively unlikely due to the highest population-dense areas (Toronto and Ottawa) being bordered primarily by the Great Lakes, and the Eastern provincial border is shared with Quebec (which already had a menthol restriction). Moreover, a previous study found no evidence of Ontario consumers purchasing menthol cigarettes across the national border with the USA.9
A substantial reduction in sales of menthol cigarettes was measured from before to after implementation of Ontario’s menthol tobacco product sales restriction, with no upsurge in capsule cigarette sales. There was some evidence that a small shift from menthol to non-menthol cigarettes occurred. Additional research could further examine policy-intended impacts, including long-term effects on consumer behaviour and population health. Future studies could assess sales of menthol products in BC and other provinces after implementation of Canada’s national menthol tobacco product ban. This evaluation of Ontario’s menthol tobacco sales restriction can inform tobacco regulatory considerations including the menthol cigarette and flavoured cigar product standards anticipated in the USA, as well as tobacco policy enforcement plans and future evaluation of similar policies. As jurisdictions regulate tobacco product sales, policy evaluation plays an important role in understanding the public health implications of such policies.
What this paper adds
What is already known on this subject
Evaluation studies of menthol tobacco sales restrictions have shown intended effects on wholesale sales, retailer compliance and consumer behaviour.
What important gaps in knowledge exist on this topic
Although tobacco manufacturer reports show decreased wholesale sales of menthol cigarettes after a 2017 menthol sales restriction in Ontario, Canada, it has not been previously examined whether these same patterns are observed in retail sales of menthol cigarettes and other menthol tobacco products. Additionally, no studies of menthol sales policies have yet quantified possible substitution of other tobacco products with retail sales data, including for vaping products, or potential unintended consequences, such as changes in sales of cigarettes with menthol-suggestive descriptors.
What this paper adds
Retail sales of menthol tobacco products declined substantially from before to after implementation of Ontario’s menthol tobacco sales restriction with minimal evidence of non-menthol cigarette substitution and no evidence of substitution of other tobacco products. Capsule cigarette sales seemed to be suppressed by the menthol policy. There is some indication, however, that cigarettes with menthol-suggestive descriptors are being described as non-menthol, reinforcing the importance of vigilant surveillance and policy enforcement.
Data availability statement
No data are available.
Ethics statements
Patient consent for publication
Footnotes
Contributors EMB, TR, DGG and AR conceptualised and designed the study. DGG and EMC conducted data analyses. EMB, TR, DGG and EMC drafted the article. All authors contributed to writing and editing, and approved the final version of the article.
Funding Funding for this study was provided by the US Food and Drug Administration (FDA), Centre for Tobacco Products (CTP) under contract HHSF223201110005B to RTI International.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.