<<

Natural, Organic, Additive- and Pure on Packs in 14 Countries

Kevin Welding, PhD Ryan David Kennedy, PhD Meghan Bridgid Moran, PhD Joanna E. Cohen, PhD Katherine Clegg Smith, PhD

Objectives: The industry uses terms like “light” and “mild” to mislead consumers to be- lieve its product is less harmful. US tobacco brands have marketed products with terms (eg, nat- ural) that inaccurately convey reduced harm. Little is known about how these terms are being used elsewhere. This study aims to describe the frequency of “natural,” “organic,” “additive-free,” and “pure” descriptors on cigarette packs in 14 low- and middle-income countries (LMICs). Meth- ods: Cigarette packs were purchased as part of TPackSS (Tobacco Pack Surveillance System) be- tween 2013 and 2017 from 14 LMICs where the majority of tobacco users live. Packs were coded for “natural,” “organic,” “additive-free,” and “pure” descriptors. Results: Packs containing at least one of the 4 sets of descriptors were found in 12 of 14 countries. “Natural” terms were most com- monly identified, present on packs from 10 countries. “Organic” terms were least common, only found in 2 countries. Overall, 2.8% of the 5576 country-year-unique packs contained at least one of the studied descriptors. Conclusions: The is using potentially misleading terms on packs in LMICs. These problematic associations can be resolved by following labeling guidelines to prevent misleading or deceptive terms.

Key words: tobacco; ; policy; content analysis; LMICs Tob Regul Sci.™ 2019;5(4):352-359 DOI: https://doi.org/10.18001/TRS.5.4.5

he World Health Organization’s Framework and “fine” to circumvent bans on other terms,9-12 Convention on (FCTC), as well as light-colored tobacco pack colors, which Article 11 (Packaging and labeling of tobac- are likely to be perceived to have lower risk.13- Tco products) recognizes the importance of the pack- 16 Terms that convey a product is natural or free age as a promotional vehicle for tobacco companies from additives have also been used by the tobac- and requires the removal of potentially misleading co industry to describe their products. The use of packaging information, including the terms “light” term “natural” has been used by American tobacco and “mild.”1-8 companies to describe aspects of for over However, even when these descriptors are banned, 100 years. One initial use of the term was to de- tobacco companies use a variety of other tactics to scribe the tobacco growing process or tobacco with convey inaccurate reduced risk of cigarette brands phrases in advertising such as “naturally ripened by or brand variants to consumers.2 For example, the the sun.”17 Later, the term natural was used to char- tobacco industry has used terms such as “smooth” acterize aspects of the physical cigarette such as the

Kevin Welding, Senior Biostatistician, Johns Hopkins Bloomberg School of Public Health, Department of Health, Behavior & Society, Institute for Global Tobacco Control, Baltimore, MD. Ryan David Kennedy, Assistant Professor, Johns Hopkins Bloomberg School of Public Health, Department of Health, Behavior & Society, Institute for Global Tobacco Control, Baltimore, MD. Meghan Bridgid Moran, Assistant Professor, Johns Hopkins Bloomberg School of Public Health, Department of Health, Behavior & Society, Baltimore, MD. Joanna E. Cohen, Professor, Johns Hopkins Bloom- berg School of Public Health, Department of Health, Behavior & Society, Institute for Global Tobacco Control, Baltimore, MD. Katherine Clegg Smith, Professor, Johns Hopkins Bloomberg School of Public Health, Department of Health, Behavior & Society, Institute for Global Tobacco Control, Baltimore, MD. Correspondence Dr Welding; [email protected]

352 Welding et al filter, the menthol and the tobacco. 17 chase in a country by visiting a mix of vendor types More recently, the cigarette brand Natural Amer- located in high-, middle-, and low-income neigh- ican Spirit has been marketed using the labels borhoods in the most populous city and then 2 (3 “natural,” “organic,” and “additive-free.”18,19 Smok- in India and 4 in China) of the remaining cities in ers of Natural American Spirit cigarettes have 22 the top 10 most populous in each country. Data times greater odds than smokers of other brands to collectors visited 12 neighborhoods in each city (falsely) believe that their brand of cigarette is less and visited up to 4 vendors within a neighborhood harmful than other brands.20 US- and Canadian- to find additional unique packs. A walking proto- based research has found that consumers perceive col that included alternating vendor types was used these specific terms to convey reduced risk.17,21-24 to increase the chances of finding unique packs in Research has found that tobacco marketing that as- each locale. More information on sampling and sociates the product with environmentally friendly data collection methods are outlined extensively practices (sometimes described as greenwashing), elsewhere.30 Between 2015 and 2017, TPackSS re- or that conveys a product is additive-free through turned to 9 of the 14 countries (Bangladesh, , other terms, such as ‘tobacco and water’, conveys China, India, Indonesia, Philippines, , Thai- inaccurate reduced risk to consumers.25,26 This effect land, and Vietnam), where tobacco control policies is known as a ‘health halo’ and occurs when con- related to packaging or labeling had changed, to sumers use one attribute about a product (ie, that it collect a new sample of packs employing the same is ‘natural’) to make inferences about the product’s sampling principles. healthfulness, even when it is unrelated.27,28 In the United States (US), the tobacco industry Coding Procedures has been using these problematic tactics to mislead Coding the packs included a thematic approach consumers about product risk. However, there is to coding all lexical content other than brand name little understanding about how prevalent the use or health warning label on the outside of the pack- of these types of descriptors are in non-US settings. aging. Where necessary, lexical content was trans- Such monitoring can alert policymakers to poten- lated to English. All packs were first double-coded tially problematic marketing tactics requiring regu- by a pair of trained coders for the presence of any lation. To address this need, in the current study, natural or organic related references. The percent we sought to identify the frequency of “natural,” agreement between coders was 99.5% and the “organic,” “additive-free,” and “pure” (and related prevalence and bias adjusted kappa was 0.991. This terms) on cigarette packs from 14 low- and mid- subset of packs was then further coded to separate dle-income countries. We reviewed cigarette packs the presence of terms categorized as (1) natural from collected through the Tobacco Pack Surveillance those of (2) organic. Lexical content categorized as Study (TPackSS), which systematically collects to- ‘natural’ included the specific term ‘natural’ and bacco packs in the low- and middle-income coun- versions of the term (eg, ‘naturalness’, ‘naturally’). tries with the greatest number of smokers.29 Similarly, the ‘organic’ category included the use of specific term ‘organic’ and variations. All packs METHODS were also double-coded for the presence of terms Sample “additive-free,” and/or “no added flavors,” and/or In 2013, cigarette packs were collected from 14 “pure.” The percent agreement between coders was countries (Bangladesh, Brazil, China, Egypt, India, 99.3% and the prevalence and bias adjusted kappa Indonesia, Mexico, Pakistan, Philippines, Russia, was 0.987. This subset of packs was then coded to Thailand, , , and Vietnam) using a separate the presence of terms categorized as (3) ad- standardized protocol.30 These countries were se- ditive-free from those of (4) pure. Lexical content lected because, at the start of data collection, they categorized as ‘additive-free’ included the specific were among the low- and middle-income countries terms ‘no additives,’ ‘additive-free,’ or other state- with the greatest number of tobacco users. The ments claiming the product did not contain addi- primary goal in sample construction was to collect tives. Lexical content categorized as ‘pure’ included one of every unique cigarette pack available for pur- the specific term ‘pure’ and variations like ‘purity’.

Tob Regul Sci.™ 2019;5(4):352-359 DOI: https://doi.org/10.18001/TRS.5.4.5 353 Natural, Organic, Additive-free and Pure on Cigarette Packs in 14 Countries

Table 1 Presence of the Lexical Terms, “Natural,” “Organic,” “Additive-free,” or “Pure” on Cigarette Packs Collected in Low- and Middle-income Countries – Reported by Country Sample Size Any Category Natural Organic Additive-free Pure Country (N) % (n) % (n) % (n) % (n) % (n)

Bangladesh 424 2.83% (12) 1.42% (6) 0.00% (0) 1.42% (6) 0.00% (0)

Brazil 269 0.74% (2) 0.74% (2) 0.00% (0) 0.00% (0) 0.00% (0)

China 1191 1.18% (14) 0.76% (9) 0.08% (1) 0.34% (4) 0.25% (3)

Egypt 58 0.00% (0) 0.00% (0) 0.00% (0) 0.00% (0) 0.00% (0)

India 230 1.74% (4) 0.43% (1) 0.00% (0) 0.43% (1) 0.87% (2)

Indonesia 467 6.00% (28) 6.00% (28) 0.00% (0) 0.00% (0) 0.21% (1)

Mexico 134 5.22% (7) 5.22% (7) 0.00% (0) 0.00% (0) 0.00% (0)

Pakistan 382 3.14% (12) 0.52% (2) 0.00% (0) 2.36% (9) 0.26% (1)

Philippines 251 0.00% (0) 0.00% (0) 0.00% (0) 0.00% (0) 0.00% (0)

Russia 1004 5.98% (60) 2.99% (30) 0.40% (4) 1.89% (19) 2.09% (21)

Thailand 237 0.42% (1) 0.00% (0) 0.00% (0) 0.42% (1) 0.00% (0)

Turkey 308 0.65% (2) 0.00% (0) 0.00% (0) 0.65% (2) 0.00% (0)

Ukraine 324 0.93% (3) 0.93% (3) 0.00% (0) 0.00% (0) 0.00% (0)

Vietnam 297 4.38% (13) 3.03% (9) 0.00% (0) 0.67% (2) 0.67% (2)

Total 5576 2.83% (158) 1.74% (97) 0.09% (5) 0.79% (44) 0.54%(30) (All Countries)

Note. Because a cigarette pack can include more than one lexical term the third column is not a simple summation of the 4 categories

The resulting 4 categories, referred to as “natural,” Brazil, China, Egypt, Indonesia, India, Mexico, “organic,” “additive-free,” and “pure,” are reported Pakistan, Philippines, Russia, Thailand, Turkey, by country where the pack was purchased. Ukraine, and Vietnam. Coding results are presented as the number and The analysis identified cigarette packs where mar- proportion of packs with any of the terms by coun- keting appeals included terms in one of 4 catego- try and by cigarette brand. ries: natural, organic, additive-free, or pure. Packs (N = 158, 2.8%) from all countries except Egypt RESULTS and Philippines used terms in at least one of the 4 A total of 3240 country-unique packs were pur- categories. Russia and Indonesia were the countries chased in 2013, and 2336 country-unique packs with the highest frequency (6%) of these terms. between 2015 and 2017 for a total of 5576 coun- China and Russia were the only countries to have try-year-unique packs analyzed from Bangladesh, packs with each of the lexical terms of interest.

354 Welding et al

Table 2 Presence of the Lexical Terms, “Natural,” “Organic,” “Additive-free,” or “Pure” on Cigarette Packs Collected in Low- and Middle-income Countries – Reported by Cigarette Brand

Total Packs Any Breakdown of Any Category Packs Brand in Sample Category Family Natural Organic Additive-free Pure (N) % (n) % (n) % (n) % (n) % (n)

Dunhill 142 20% (29) 38% (11) 0% (0) 62% (18) 0% (0)

Cigaronne 15 67% (10) 20% (2) 0% (0) 0% (0) 100% (10)

Richmond 23 39% (9) 100% (9) 0% (0) 0% (0) 0% (0)

A Mild 10 70% (7) 100% (7) 0% (0) 0% (0) 0% (0)

Pall Mall 68 10% (7) 100% (7) 0% (0) 0% (0) 0% (0)

Pepe 7 100% (7) 0% (0) 0% (0) 100% (7) 14% (1)

Treasurer 7 100% (7) 0% (0) 0% (0) 43% (3) 100% (7)

Benson & Hedges 61 10% (6) 100% (6) 0% (0) 0% (0) 0% (0)

Salem 14 36% (5) 100% (5) 0% (0) 0% (0) 0% (0)

Esse 159 3% (4) 75% (3) 0% (0) 0% (0) 25% (1)

Joe Black 3 100% (3) 100% (3) 0% (0) 0% (0) 0% (0)

Araya 4 50% (2) 100% (2) 0% (0) 0% (0) 0% (0)

Aroma Rich 11 18% (2) 0% (0) 100% (2) 0% (0) 0% (0)

Bohem 9 22% (2) 100% (2) 0% (0) 0% (0) 0% (0)

Craven A 14 14% (2) 100% (2) 0% (0) 0% (0) 0% (0)

Djarum 58 3% (2) 100% (2) 0% (0) 0% (0) 0% (0)

Fantom 2 100% (2) 100% (2) 0% (0) 100% (2) 0% (0)

Jinsheng 25 8% (2) 100% (2) 0% (0) 0% (0) 0% (0)

L.A. 21 10% (2) 100% (2) 0% (0) 0% (0) 0% (0)

Lucky Strike 51 4% (2) 50% (1) 0% (0) 50% (1) 0% (0)

Mild Seven 39 5% (2) 0% (0) 0% (0) 0% (0) 100% (2)

Qi Pi Lang 14 14% (2) 100% (2) 0% (0) 0% (0) 0% (0)

Single Instances 546 8% (42) 64% (27) 7% (3) 31% (13) 21% (9)

Note. Because a cigarette pack can include more than one lexical term the third column is not a simple summation of the 4 categories.

Tob Regul Sci.™ 2019;5(4):352-359 DOI: https://doi.org/10.18001/TRS.5.4.5 355 Natural, Organic, Additive-free and Pure on Cigarette Packs in 14 Countries

Figure 1 Examples of Cigarette Packs with Each of the 4 Categories of Terms

Natural Organic Additive-free Pure

“Naturally sun “No additives were “Organic” “Pure lamina blend” ripened tobacco” used in the blend”

Bangladesh, 2013 Russia, 2015 Turkey, 2013 Indonesia, 2016

Note. Translation of “Harmanda herhangi bir katki maddesi kullanilmamistir” is “No additives were used in the blend.”

Table 1 outlines the counts for each category by Esse, are using these descriptors on 3% of their 159 country. packs and brands like (N = 334), Dav- The analysis identified 64 different cigarette idoff (N = 151), and Winston (N = 123) are not brands using one of the 4 categories; in the sample using these descriptors at all. It is also interesting of 158 packs, the brand (29 packs) ap- to note that some brands (eg, Pepe, Treasurer) are pears most frequently. Table 2 presents the counts using a descriptor on all their packs in the sample. for each category by brand. Whereas Dunhill has a The use of “natural” terms was the most com- large number of country-year-unique pack presen- mon out of all the descriptors, used in 10 of the tations (with 142 packs in the full sample) it does countries and identified on 97 packs (1.7%). The not appear to be the reason we see a high frequency frequency was highest in Indonesia (6%) and Mex- of descriptor use. Other brands with a large num- ico (5.2%). The most common references in Indo- ber of country-year-unique pack presentations, like nesia and Mexico were to natural cloves (9 out of

356

Welding et al

28 packs) and natural menthol (6 out of 7 packs), is evidence that removing misleading descriptors respectively. Organic terms were only used in Chi- may have little to no change on inaccurate re- na and Russia, on 5 different packs. For 3 of these duced harm perceptions about the product. Plain packs, the term was written in Russian or Manda- packaging may be more effective at reducing these rin. For the other 2 packs, both from Russia, the misperceptions.12,23,31,32 English word “organic” was written (for an exam- The tobacco industry’s use of additives was ini- ple, see Figure 1). The “additive-free” terms were tially highlighted in the 1981 US Surgeon Gen- identified in 8 countries and appeared on 44 packs. eral’s Report on “The Changing Cigarette.”33 This The highest frequency for these terms was found report raised the possibility that cigarette additives in Pakistan. The Pakistani finding is driven by the increased the health risks of and called dominance of Dunhill packs in the Pakistan sam- for their disclosure.17 The industry’s use of addi- ple of unique brand variants, and that brand’s mar- tives has been the focus of public health campaigns keting of no additive cigarettes. The use of “pure” to educate consumers about the dangers of these terms was identified on packs from 6 countries on constituents and to highlight the practices of the 30 packs. The highest frequency of these terms (N tobacco industry.34 This historical framing of the = 21, 2.1%) was found in Russia. This was pre- unnaturalness of cigarettes, specifically from the dominately used by 2 brands in Russia, Cigaronne inclusion of chemical additives, arguably created (10 packs) and Treasurer (7 packs). a marketing opportunity for tobacco manufactur- The 29 Dunhill packs using the studied terms ei- ers to present products that were additive-free, and ther used natural terms or made no additive claims, therefore, perceived as less harmful. but did not use both on the same pack. The Esse There are strengths and limitations associated (N = 4) and (N = 2) packs use a simi- with this research. The analysis fills an important lar strategy. There are 19 other brands with more gap in studies of marketing of tobacco products than one pack using any of the terms, and the ma- in LMICs, specifically marketing on the pack.35,36 jority (15 packs) use only one tactic. The other 4 Data collection occurred in a range of geographi- brands (Cigaronne, Pepe, Treasurer, Fantom) use cally and culturally diverse countries, and data col- more than one type of term on some of their packs. lection was systematic, extensive, and intended to produce as comprehensive a sample of distinct packs DISCUSSION available for purchase within a country as possible. The tobacco industry uses terms such as natural, TPackSS data collection is designed to maximize organic, additive-free, and pure on cigarette pack- the breadth of unique packs collected, but it is pos- aging sold in LMICs. There is evidence from the sible that some pack presentations available for sale US and Canada that these terms are interpreted by in a country were not collected. Packs are unique tobacco users to mean that these cigarettes have re- within country and year, so duplicate packs could duced risk or are less harmful.17,21-24 It is important exist since the entire sample has 23 country and to understand if these terms have the same con- year combinations. We do not incorporate data on veyed meaning about harm reduction in different market share for the brands, as this is not avail- languages and cultural contexts as they do in high- able for all brand variants across all settings. As a income countries. This study shows the presence result, we cannot consider the population exposure of ‘natural’ marketing on packs in LMICs, but not of these natural terms on purchased cigarette packs. consumers’ interpretation of such messages; there is Based on Euromonitor brand market share data, a need for studies to understand how these terms in- 14 out of the 64 brands were a top 10 brand in fluence smokers and potential smokers in LMICs. market share and used these descriptors on at least Article 11 of the FCTC requires parties to remove one brand variant (China-Huang Shan, Shuangxi; potentially misleading packaging information and India-Four Square; Indonesia-A Mild, Clas Mild, could be used to support policies to restrict use of Djarum, L.A., U Mild; Mexico-Benson & Hedges, these terms. Policies to restrict such labeling, in ac- Delicados; Pakistan-Dunhill; Russia-Bond Street; cordance with FCTC Article 11 guidelines, should Vietnam-, Thang Long). The presence of be developed, implemented, and monitored. There these terms in 12 out of the 14 countries studied is

Tob Regul Sci.™ 2019;5(4):352-359 DOI: https://doi.org/10.18001/TRS.5.4.5 357 Natural, Organic, Additive-free and Pure on Cigarette Packs in 14 Countries important for advocacy efforts because it indicates globaltobaccocontrol.org/tpackss/about), the en- they are used in a range of cultural settings. As we tire TPackSS team, and the coders Femi Erinoso, did not code for semantic uses it is possible that Caleb Clawson, Deana Trimble, and Kara Robin- descriptors were used in contexts that might not son. This work was supported with funding from convey reduced harm to consumers. We cannot re- Bloomberg Philanthropies’ Bloomberg Initiative port anything about the meaning of these terms to to Reduce Tobacco Use (Bloomberg.org). Dr. Mo- consumers as this work does not include message ran’s effort is supported by NIDA and FDA Cen- testing. ter for Tobacco Products (CTP) (K01DA037903, Given what we know about the interpretation PI: Moran). The content is solely the responsibility of these terms in high-income countries, future of the authors and does not necessarily represent work should seek to incorporate analysis from the the official views of the Bloomberg Philanthropies, cultural context from which data were drawn. It NIH or FDA. is important to know if the descriptors examined in this study convey reduced harm in different References languages and whether there are other culturally 1. Wakefield MA, Morley C, Horan JK, Cummings KM. specific descriptors that convey reduced harm in The cigarette pack as image: new evidence from industry these countries. For comparison purposes, it could documents. Tob Control. 2002;11:i73-i80. 2. Hammond D, Dockrell M, Arnott D, et al. Cigarette pack be interesting to conduct a TPackSS collection of design and perceptions of risk among UK adults and unique pack presentations in a high-income coun- youth. Eur J Public Health. 2009;9(6):631-637. try. Future studies also could at whether ciga- 3. National Cancer Institute. Smoking and Tobacco Control rette packs that convey reduced harm are more Monograph No. 13: Risks Associated with Smoking Ciga- rettes with Low Tar Machine-Measured Yields of Tar and likely to be on packs intended for specific demo- Nicotine. Bethesda, MD: US Department of Health and graphic groups and whether the use of these po- Human Services, National Institutes of Health, National tentially misleading descriptors are correlated with Cancer Institute; 2001. smoking prevalence. 4. Ashley MJ, Cohen J, Ferrence R. ‘Light’ and ‘mild’ ciga- rettes: who smokes them? Are they being misled? Can J Public Health. 2001;92(6):407-411. IMPLICATIONS FOR TOBACCO 5. Cohen JE, Yang J, Donaldson EA. Impact of the removal REGULATION of light and mild descriptors from cigarette packages in Ontario, Canada: switching to “light replacement” brand Descriptors that inaccurately convey a product variants. Prev Med. 2014;69:120-125. as less harmful than others can mislead consum- 6. Shiffman S, Pillitteri JL, Burton SL, et al. Smokers’ beliefs ers and contribute to tobacco use initiation and about “Light” and “Ultra Light” cigarettes. Tob Control. 2001;10(Suppl 1):i17-i23. inhibit successful cessation. Policies to restrict such 7. Gilpin EA, Emery S, White MM, Pierce JP. Does tobac- labeling, in accordance with FCTC Article 11 co industry marketing of ‘light’ cigarettes give smokers guidelines (which includes restrictions on poten- a rationale for postponing quitting? Nicotine Tob Res. tially misleading descriptors and plain packaging), 2002;4(Suppl 2):S147-S155. should be developed, implemented, and monitored 8. Bekki K, Inaba Y, Kunugita N. WHO Framework Con- vention on Tobacco Control (FCTC) Article 11: packag- for adherence. ing and labelling of tobacco products. Nihon Eiseigaku Zasshi. Jap J Hyg. 2015;70(1):24-32. 9. Connolly GN, Alpert HR. Has the tobacco industry evad- Human Subjects Statement ed the FDA’s ban on ‘light’ cigarette descriptors? Tob Con- This paper did not include human subjects. trol. 2014;23(2):140-145. 10. King B, Borland R. What was “light” and “mild” is now “smooth” and “fine”: new labeling of Australian ciga- Conflict of Interest Statement rettes. Tob Control. 2005;14(3):214-215. All authors of this article declare they have no 11. Peace J, Wilson N, Hoek J, et al. Survey of descriptors on cigarette packs: still misleading consumers? N Z Med J. conflicts of interest. 2009;122(1303):90-96. 12. Yong H, Borland R, Cummings KM, et al. Impact of the Acknowledgements removal of misleading terms on cigarette pack on smok- ers’ beliefs about ‘light/mild’ cigarettes: cross-country This study would not be possible without the comparisons. Addiction. 2011;106(12):2204-2213. work of in-country collaborators (listed here: www. 13. Lempert LK, Glantz S. Packaging colour research by to-

358 Welding et al

bacco companies: the pack as a product characteristic. 25. Moran MB, Brown J, Lindblom E, et al. Beyond ‘natural’: Tob Control. 2017;26(3):307-315. cigarette ad tactics that mislead about relative risk. Tob 14. Agaku IT, Omaduvie UT, Filippidis FT, Vardavas CI. Regul Sci. 2018;4(5):3-19. Cigarette design and marketing features are associated 26. Moran MB, Pearson J. Real. Simple. Deadly. A pilot with increased smoking susceptibility and perception of test of consumer harm perceptions in response to Nat- reduced harm among smokers in 27 EU countries. Tob ural American Spirit advertising. Tob Regul Sci. 2019; Control. 2015;24(e4):e233-e240. 5(4):360-368. 15. Bansal-Travers M, O’Connor R, Fix BV, Cummings KM. 27. Epperson AE, Henriksen L, Prochaska JJ. Natural Ameri- What do cigarette pack colors communicate to smokers can spirit brand marketing casts health halo around in the US? Am J Prev Med. 2011;40(6):683-689. smoking. Am J Public Health. 2017;107(5):668-670. 16. Hammond D, Parkinson C. The impact of cigarette 28. Schuldt, JP, Muller D, Schwarz, N. The “fair trade” ef- package design on perceptions of risk. J Public Health. fect: Health halos from social ethics claims. Soc Psychol 2009;31(3):345-353. Personal Sci. 2012;3(5):581-589. 17. McDaniel PA, Malone RE. “I always thought they were 29. TPackSS – Tobacco Pack Surveillance System. Available all pure tobacco”: American smokers” perceptions of “nat- at: https://globaltobaccocontrol.org/tpackss/. Accessed ural” cigarettes and tobacco industry advertising strate- on November 23, 2018. gies. Tob Control. 2007;16(6):e7. 30. Smith, K, Washington, C, Brown, J, et al. The tobacco 18. Moran MB, Pierce JP, Weiger C, et al. Use of imagery and pack surveillance system: a protocol for assessing health text that could convey reduced harm in American Spirit warning compliance, design features, and appeals of to- advertisements. Tob Control. 2017;26(e1):e68-70. bacco packs sold in low-and middle-income countries. 19. O’Gara E, D’Silva J, Weiger C, et al. Restricting “natural” JMIR Public Health Surveill. 2015;1(2):e8. and “additive-free”: did the FDA’s agreement with Santa 31. Leas EC, Ayers JW, Strong DR, Pierce JP. Which ciga- Fe Natural Tobacco Company change advertising for rettes do Americans think are safer? A population-based Natural American Spirit? Tob Regul Sci. 2019;5(4)332- analysis with wave 1 of the PATH study. Tob Control. 338. 2017;26(e1):e59-e60. 20. Pearson JL, Johnson A, Villanti A, et al. Misperceptions 32. Leas EC, Pierce JP, Dimofte CV, et al. Standardised of harm among Natural American Spirit smokers: results cigarette packaging may reduce the implied safety from wave 1 of the Population Assessment of Tobacco of Natural American Spirit cigarettes. Tob Control. and Health (PATH) study (2013–2014). Tob Control. 2018;27(e2):e118-e123. 2017;26(e1):e61-e67. 33. US Department of Health and Human Services. The 21. Byron MJ, Baig SA, Moracco KE, Brewer NT. Adoles- Health Consequences of Smoking: The Changing Ciga- cents’ and adults’ perceptions of ‘natural’, ‘organic’ and rette: a Report of the Surgeon General. Rockville, MD: ‘additive-free’ cigarettes, and the required disclaimers. Tob US Department of Health and Human Services, Public Control. 2015;25(5):517-520. Health Service, Office on Smoking and Health; 1981. 22. Czoli CD, Hammond D. Cigarette packaging: youth 34. Campaign for Tobacco Free Kids. Tobacco Product Addi- perceptions of “natural” cigarettes, filter references, and tives – Essential Facts (Fact Sheet). Available at: https:// contraband tobacco. J Adolesc Health. 2014;54(1):33-39. www.tobaccofreekids.org/assets/global/pdfs/en/addi- 23. Pearson JL, Richardson A, Feirman SP, et al. American tives_en.pdf. Accessed on November 23, 2018. Spirit pack descriptors and perceptions of harm: a crowd- 35. Hoek J, Gendall P, Gifford H, et al. Tobacco branding, sourced comparison of modified packs. Nicotine Tob Res. plain packaging, pictorial warnings, and symbolic con- 2016;18(8):1749-1756. sumption. Qual Health Res. 2012;22:630-639. 24. Sanders-Jackson A, Tan ASL, Yie K. Effects of health-ori- 36. Moodie C, Angus K, Ford A. The importance of ciga- ented descriptors on combustible cigarette and electronic rette packaging in a ‘dark’ market: the ‘Silk Cut’ experi- cigarette packaging: an experiment among adult smokers ence. Tob Control. 2014;23:274-278. in the United States. Tob Control. 2018;27(5):534-541.

Tob Regul Sci.™ 2019;5(4):352-359 DOI: https://doi.org/10.18001/TRS.5.4.5 359