Background To increase knowledge of smoking-related health risks and provide smoking cessation information at the point of sale, in 2009, New York City required the posting of graphic point-of-sale tobacco health warnings in tobacco retailers. This study is the first to evaluate the impact of such a policy in the USA.
Methods Cross-sectional street-intercept surveys conducted among adult current smokers and recent quitters before and after signage implementation assessed the awareness and impact of the signs. Approximately 10 street-intercept surveys were conducted at each of 50 tobacco retailers in New York City before and after policy implementation. A total of 1007 adults who were either current smokers or recent quitters were surveyed about the awareness and impact of tobacco health warning signs. Multivariate risk ratios (RR) were calculated to estimate awareness and impact of the signs.
Results Most participants (86%) were current smokers, and the sample was 28% African–American, 32% Hispanic/Latino and 27% non-Hispanic white. Awareness of tobacco health warning signs more than doubled after the policy implementation (adjusted RR =2.01, 95% CI 1.74 to 2.33). Signage posting was associated with an 11% increase in the extent to which signs made respondents think about quitting smoking (adjusted RR =1.11, 95% CI 1.01 to 1.22).
Conclusions A policy requiring tobacco retailers to display graphic health warning signs increased awareness of health risks of smoking and stimulated thoughts about quitting smoking. Additional research aimed at evaluating the effect of tobacco control measures in the retail environment is necessary to provide further rationale for implementing these changes.
- graphic health warnings
- smoking cessation
- public policy
- advertising and promotion
Statistics from Altmetric.com
- graphic health warnings
- smoking cessation
- public policy
- advertising and promotion
The impact of exposure to tobacco advertising and promotion in the retail environment on adult and youth smoking behaviours is well characterised in the literature.1–3 More specifically, exposure to retail cigarette advertising and promotion among youth is associated with higher rates of youth smoking initiation and lifetime smoking.1 ,4 Exposure among adults is associated with increased impulse purchases of cigarettes that can lead to lower rates of successful cessation.3 ,5 Extant research suggests that characteristics of the retail environment, the primary source of tobacco industry expenditures on tobacco product marketing, are making it easier for adults and youth to smoke.6
Smokers' behaviours are also influenced by their understanding of the health risks of smoking.7 Smokers who perceive greater smoking-related health hazards are more likely to consider quitting and to quit smoking successfully.7 Although the harmful effects of smoking have been known for decades, there remain significant gaps in smokers' understanding of these risks.8 ,9 One survey found that 15% of US smokers believed that smoking did not cause heart disease.7 In New York City (NYC), almost 50% of smokers do not think that they are at greater risk of heart attack than non-smokers of the same age (unpublished data: NY State Adult Tobacco Survey, 2007). These gaps in risk awareness are widest among lower socioeconomic groups.10
To increase knowledge of smoking-related health risks and provide smoking cessation information at the point of sale (POS), in 2009, the NYC Board of Health required the posting of graphic POS tobacco health warnings in tobacco retailers. NYC was the first jurisdiction in the USA to post such signs. Only a limited number of countries post health warnings about smoking in retailers, including Australia, Brazil, Canada, Chile, New Zealand and Uruguay.2 ,11–13 The regulation was based in part on evidence that graphic cigarette pack warnings are more effective than text-only warnings in increasing levels of awareness and perceived effectiveness among smokers, defined as stimulating thoughts about quitting and the health risks of smoking as well as preventing smoking.14 The regulation required that all licensed tobacco retailers which sold tobacco products face-to-face to customers post the health warning signs. Retailers could post one small sign at each cash register or one large sign where tobacco products were displayed. Failure to post signs could result in a fine of up to $2000. Based on qualitative testing of design and concepts, the NYC Department of Health and Mental Hygiene developed signage that contained graphic images of the short-term and long-term health consequences of smoking and information about how to obtain smoking cessation assistance (the New York State Smokers' Quitline phone number and website, and 3-1-1, NYC's general information line). Three signs depicted graphic images of lung cancer, stroke and tooth decay and were available in English, Spanish, Russian and Chinese (figure 1). Signs were distributed to retailers for display, along with education materials about how to post, beginning in December 2009. Tobacco manufacturers, retailers and trade associations sued NYC in June 2010 over the requirement to post warning signs, and the law was overturned in December 2010.15 NYC is currently appealing this decision. While the City never enforced the signage requirement due to the legal challenge and litigation is still pending at the time of writing, the signs were not revoked and retailers could choose to display or not display the signs. The objective of our study was to assess the awareness and impact of the POS health warning signage. We hypothesised that, following implementation of the signs, current smokers and recent quitters surveyed would be (1) more aware of the signs, (2) more likely to think about smoking-related health risks, (3) more likely to think about quitting smoking or to stay quit and (4) less likely to purchase cigarettes. This study is the first to evaluate the impact of graphic tobacco health warning signs at the POS in the USA.
Study design and study participants
Cross-sectional street-intercept surveys conducted before and after signage implementation were used to assess knowledge of smoking-related health risks, purchase behaviours, and awareness and attitudes about NYC's POS signage. A research firm was contracted to survey current smokers and recent quitters as they exited tobacco retail establishments. Baseline surveys were conducted in November 2009; follow-up surveys were conducted in August and September 2010, about 9 months following signage distribution.
Ten randomly selected tobacco retailers were selected in each of the five boroughs from a universal database of approximately 11 000 licensed tobacco retailers in NYC, which were categorised into 15 different store types. Our study targeted the two largest categories (grocery/supermarkets and pharmacies), which together comprise about 75% of all tobacco retailers. A total of seven grocery/supermarkets and three pharmacies were sampled in each borough, which approximated the underlying proportion of these two store types in the underlying retailer sample. One retailer included in the first survey went out of business by the time of the second survey and was replaced with another retailer of the same type in the immediate vicinity.
Ten exiting customers per retailer were surveyed. A total of 1007 surveys (504 at baseline and 503 at follow-up) were conducted. Data collectors approached adults exiting tobacco retailers to assess their interest in and eligibility for the survey. Participation was voluntary and verbal consent was obtained. Inclusion criteria included being aged 18 years or older, a current smoker or recent quitter (within last 6 months) and an NYC resident. Across both survey waves, 3200 potential respondents were approached with a 43% refusal rate. Among those screened, 817 (26%) were deemed ineligible, for an overall response rate of 32%. The majority (86%) of those screened were ineligible due to smoking status. All potential respondents determined to be eligible during the screening process participated in the survey.
The interview, approximately 5 to 7 min in length, was administered to eligible participants in English or Spanish. All participants who completed the interview received an $8 Metrocard as compensation. The Health Department's Institutional Review Board determined that this study was exempt research.
Demographic characteristics included in the analysis were age, sex, race/ethnicity, language spoken most often at home, education and income. Current smoking status was assessed through two questions: “Have you smoked at least 100 cigarettes in your entire life?” and “Do you now smoke cigarettes every day, some days or not at all?”. Recent quitters did not currently smoke but had smoked within the last 6 months. Cigarette consumption among current smokers was measured as the number of cigarettes smoked per day (CPD) (1–10, 11–20 and 21+ CPD).
Awareness of signs was assessed through two questions: “During this visit to the store, did you notice health warning signs about smoking posted in the checkout area when you were paying for your purchases?” (prompted response options were ‘Yes, noticed and read’; ‘Yes, noticed but did not read’; ‘No’; ‘Don't know’) and “In the past 30 days, when you were paying for your purchases in stores, how often did you notice health warning signs about smoking posted in the checkout area?” (prompted response options were ‘Very often’; ‘Often’; ‘Sometimes’; ‘Rarely’; ‘Never’).
In each survey wave, a measure of confirmed awareness was used to identify those respondents who noticed the signs during the current visit and/or last 30 days and recalled at least one of the elements featured in the signs in an open-ended question. The signage elements classified as confirmed included cancer, mouth diseases, stroke, graphic images and cessation information including tips to stop smoking and available cessation services (311 or the NYS Smokers' Quitline). The elements classified as unconfirmed included: addiction warnings, legal age of sale signs and pregnancy-related warning signs, all of which are also posted at the POS. Unconfirmed elements also included text-based warnings on cigarette advertisements and packs, as well as television, print, outdoor and online anti-tobacco advertisements. Although NYC's graphic warning signs were not yet posted at the time of the baseline survey, respondents who indicated awareness of signs were coded as having confirmed or unconfirmed awareness to ensure consistency of the awareness measures across the surveys. Signage impact was assessed through the following questions: “During this visit to the store or over the past 30 days, to what extent did these health warning signs (a) make you think about the health risks of smoking; (b) make you think about quitting smoking (among current smokers)?” or (c) “help you to quit or stay quit (among recent quitters)?”(prompted response options were ‘A lot’; ‘Somewhat’; ‘A little’; ‘Not at all’; ‘don’t know’). Participants were also asked: “During this visit to the store, did these warning signs about smoking stop you from purchasing cigarettes when you were about to buy them?”.
To assess knowledge of smoking-related health risks, a list of health effects and diseases related and not related to smoking were read and participants indicated the extent to which they agreed that smoking cigarettes was the cause of each. In the survey, the knowledge of health effects question was placed after tobacco purchasing behaviours, and before awareness and impact of the signage, to minimise recall bias.
Student t tests were used to determine the statistical significance of changes in responses to questions between the baseline and follow-up survey. Differences in demographic characteristics, smoking behaviours, awareness and impact of signs, knowledge of smoking-related health risks and purchasing behaviours were examined. Differences in awareness and impact of signs were also examined by location of store at the borough level.
To assess differences across survey waves, adjusted RR (ARR) were calculated for the confirmed awareness measure and for the two impact measures which significantly increased from baseline to follow-up in univariate analysis (ie, extent to which warning signs made respondent think about the health risks of smoking and think about quitting smoking). Dependent variables were dichotomous indicators of: (1) noticing signs during the current visit or last 30 days with confirmed awareness (yes/no), (2) warning signs made respondent think about the health risks of smoking “a lot” (vs “somewhat”, “a little” or “not at all”) and (3) among current smokers, warning signs made respondent think about quitting smoking “a lot” (vs “somewhat”, “a little” or “not at all”). Each multivariate model adjusted simultaneously for the effects of age, race/ethnicity, sex, language, education, shopping frequency and cigarette consumption. Models were not adjusted for annual income level as income was correlated with education status. Covariates were included in the multivariate models based on a priori knowledge to adjust for potential confounders. CPD was found to be an effect modifier for the outcome “signs making respondents think about the health risks of smoking” and was controlled for by including an interaction term in this adjusted model. Data analyses were conducted using SAS V.9.2. The generalised linear model was used to compute RR ratios with PROC GENMOD in SAS V.9.2.
Overall, the majority of participants in our study were 25–44 years old (42%) or 45–64 years old (37%) (table 1). More than half (59%) of those surveyed were male, and there were similar percentages of non-Hispanic whites, African–Americans and Hispanic/Latinos (range of 27%–32%). The majority of participants (81%) were English speakers, and 54% had a high school education or less. About half (52%) reported an annual income of <$25 000. The only difference in demographic characteristics of participants in the two surveys was language: a larger proportion of English speakers were included in the baseline survey compared with the follow-up survey (83% vs 78%, p=0.03). Most participants (86%) were current smokers and about three-fourths (79%) smoked daily. Smokers were less likely to be light smokers (10 or fewer CPD) at baseline than at follow-up (55% vs 65%, p<0.01).
Awareness of tobacco health warning signs at the POS during the current visit to the store or in the past 30 days increased from 30% before to 67% after (p<0.01) the signs were posted (table 2). Among respondents who noticed signs, the extent to which the signs made them think about the health risks of smoking ‘a lot’ increased from 34% to 47% (p=0.01) pre/post signage implementation. Among current smokers who noticed signs, the extent to which the signs made them think about quitting smoking ‘a lot’ increased from 31% to 43% (p=0.02). There was no difference across surveys in the signs helping recent quitters to stay quit (p=0.55). Differences in signs prompting smokers not to purchase cigarettes achieved only borderline significance (15% to 8%, p=0.05). The awareness and impact measures did not significantly vary by store location (data not shown).
Multivariate logistic regression modelling showed that following signage implementation, awareness of tobacco health warning signs at the POS doubled (ARR =2.01, 95% CI 1.74 to 2.33, p<0.01). Signage implementation was also borderline significantly associated with making respondents think about the health risks of smoking ‘a lot’ (ARR =1.11, 95% CI 0.96 to 1.28, p=0.16). Finally, signage posting was associated with an 11% increase (ARR =1.11, 95% CI 1.01 to 1.22, p=0.04) in the extent to which signs made respondents think about quitting smoking ‘a lot’.
The per cent of respondents who agreed that two of the three health conditions depicted in the signs (gum or mouth disease and stroke) were caused by smoking cigarettes increased across the two survey waves (table 3). The proportion of participants who agreed that gum or mouth disease was smoking related increased from 88% to 93% (p<0.01). Similarly, the per cent that agreed that throat or mouth cancers was smoking related increased across survey wave (93% to 96%, p=0.02). The proportion that agreed that stroke was related to smoking also increased, from 83% to 89% (p=0.01). However, knowledge of lung cancer being associated with smoking did not differ across survey waves (96% vs 98%, p=0.06). While there were no changes in knowledge of health conditions not related to smoking (ie, control conditions including liver disease and diabetes), there was an increase in knowledge of heart disease and impotence among men being smoking related. However, these health conditions were not depicted on signs. Knowledge of heart disease being associated with smoking increased from 86% to 91% (p=0.01) and knowledge of impotence increased from 46% to 57% (p<0.01).
This paper provides the first known results of an evaluation of the awareness and impact of graphic tobacco health warning signs at the POS in the USA. Our results show that signage implementation was associated with a doubling in the awareness of health warning signs and an 11% increase in stimulating thoughts about quitting. The ARR for signs stimulating thoughts about the health risks of smoking achieved borderline significance. The signs did not help recent quitters to stay quit or stop smokers from purchasing cigarettes at the current visit to the store.
There was a high level of awareness of tobacco health warning signs at the POS in the baseline survey (30%), although signs were not yet posted. This finding may be related to the posting of other signs about tobacco at the POS, including the legal age of sale sign and the Surgeon General's warning that smoking causes lung cancer, which is present on cigarette advertisements. Respondents may also have been reacting to anti-tobacco media and education efforts in other venues, including text-based warnings on cigarette packs as well as television, print, outdoor or online anti-tobacco advertisements, which are common in NYC. The high level of baseline awareness may even have been due to social response bias. Regardless, NYC POS signs did have a measurable impact on increasing awareness of tobacco health warnings, despite the high baseline awareness. Finally, this finding is consistent with the high levels of awareness reported in a previous study of POS anti-smoking warnings in Australia.11
Research shows that there are significant gaps in health knowledge related to the risks of smoking.8 ,9 Although our results achieved only borderline significance in stimulating thoughts about the health risks of smoking, the signs did increase knowledge of the health effects depicted in the signs, including gum and mouth diseases and stroke. This is similar to results from other studies, which have demonstrated that large, colour graphic cigarette pack warnings increase knowledge of the health consequences of smoking.14 ,16
The increases in knowledge of heart disease, impotence and throat/mouth cancers being related to smoking were unexpected, as they were not featured on the POS warning signs. Other public education efforts occurring in different venues between waves could have increased knowledge of these specific health effects. Also, there was a borderline significant difference in income level across survey waves, with participants in the follow-up survey less likely to be low income. As income is correlated with educational status, this difference may be contributing to the increased knowledge of smoking-related diseases not featured in the signs. The POS signs may also be stimulating thoughts and awareness of smoking-related health risks more generally.
Two-thirds of NYC smokers reported trying to quit at least once during the past year but only about half of adults in NYC reported that it is easy for smokers to get information about smoking cessation services17 (unpublished data: NY State Adult Tobacco Survey, 2007–2008). Our signs contained cessation information; however, our results did not show an impact of the signs on helping recent quitters to stay quit or in preventing smokers from purchasing cigarettes. This could be because the follow-up surveys were conducted only 9 months following signage distribution to retailers. We anticipate that the warning signs could have had an effect on behavioural outcomes such as preventing relapse or cigarette purchases had the signs been posted for a longer time period. The signs did, however, stimulate thoughts about quitting smoking, which has been shown to be related to motivation to quit and successful quits.18–20 Other anti-tobacco education efforts, including graphic cigarette pack warnings, have had a demonstrated effect on stimulating thoughts about quitting as well.7 ,19 ,21 ,22
Evidence suggests that smoking-related interventions aimed at increasing awareness and knowledge as well as prompting behaviour change such as quit attempts generally have a larger impact on awareness and knowledge measures in comparison to behavioural change.23–25 Research also shows that while smoking interventions often have modest effects on immediate behavioural outcomes, they are more successful in altering long-term behaviour.26 While changing smokers' purchasing behaviours and preventing relapse is clearly more challenging, a large body of scientific evidence has demonstrated that graphic cigarette pack warnings can increase quit smoking success and decrease youth smoking initiation.14 ,27 While pack warnings influence smoking attitudes and behaviours after the cigarette purchase has already occurred, graphic depictions of the health consequences of smoking available to customers through POS signs may have a stronger impact on cigarette purchasing behaviours since they are visible before purchase. The ability of POS information to positively affect consumer purchasing has also been demonstrated by NYC's rule requiring calorie labelling on chain restaurant menus, which has led to fewer calories being purchased.28 ,29
There were limitations to this study. First, while signage posting was never enforced, some retailers (20%) did not have signs posted during the follow-up survey. While the absence of signs in stores likely influenced the level of awareness and impact of signs at the current visit to the store, these measures were also assessed within the last 30 days, which would capture the effect of signage viewed elsewhere. As this study employed a convenience sample based on the underlying distribution of retailers, the study sample may not be representative of the NYC adult smoker and recent quitter population. While the demographic characteristics of our sample were largely similar to those of current smokers and recent quitters based on data from a 2010 population-based survey of NYC adults, NYC smokers were significantly more likely to be aged 65 years and older, white and report higher education levels compared with participants in this study.17
The results may not be generalisable to other retailer types, as the study was restricted to two types of retailers (grocery/supermarkets and pharmacies). However, these two retailer types represent the majority of licensed tobacco retailers in the City. Given the unique urban environment of NYC, with almost 11 000 licensed tobacco retailers, these results may also not be generalisable to smaller cities or towns. Additionally, the high recall of signage at baseline, measured before the policy implementation, may contribute to underestimating the awareness and impact overall. Finally, while the knowledge of health effects question was placed in the survey before awareness and impact of the signage to minimise recall bias, this question order may stimulate better memory of signs. In future studies, multiple versions of the survey could be fielded to test for order effects.
The results described here highlight the effectiveness of graphic health warning signs at the POS in stimulating thoughts about quitting smoking and, to a lesser extent, the health effects of smoking. The US Food and Drug Administration's graphic cigarette pack warnings were required beginning in September 2012, although the requirement was struck down in federal court and is currently being appealed.30 If implemented in the future, the pack warnings could have a synergistic effect with the POS signs in terms of increasing awareness of the consequences of smoking and providing cessation information. Future studies could also assess intentions to seek quit-smoking assistance from smokers' quitlines as a result of this information being included on POS warning signs. Additional research aimed at evaluating the effect of tobacco control measures in the retail environment is necessary to provide further rationale for implementing these evidence-based changes.
What this paper adds
This study is the first to evaluate the impact of graphic tobacco health warning signs at the POS in tobacco retailers in the USA. Findings from street-intercept surveys conducted before and after signage implementation in NYC demonstrate that the signs were associated with increased awareness of tobacco health warning signs at the POS. Signage posting also had an effect on increasing knowledge of the health effects of smoking and stimulating thoughts about quitting smoking. These findings can be used to provide additional rationale for implementing policy changes in the retail environment.
We greatly appreciate the guidance and direction provided by Ann Pearson, JD, Kristen Keneipp, MPH, and Jamie Sokol, MPH, in the development and distribution of the NYC tobacco health warning signs. We also appreciate the oversight provided by Thomas A. Farley, MD, MPH, Thomas R. Frieden, MD, MPH, and Sarah Perl, MPH. Finally, we thank Global Strategy Group for conducting the street-intercept surveys on behalf of the New York City Department of Health and Mental Hygiene.
Funding This work was supported by the New York City Department of Health and Mental Hygiene, City Tax Levy. No outside funding was provided. Funded data collection and personnel time.
Correction notice This article has been corrected since it was published Online First. The conclusion section in the abstract has been amended.
Competing interests None.
Ethics approval Ethics approval was provided by the New York City Department of Health and Mental Hygiene.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.