https://orcid.org/0000-0001-5216-3897
Abstract
For cigars sold individually without packaging, including many premium cigars, the US Food and Drug Administration (FDA) proposed that retailers display six warning statements on a sign at the point-of-sale (POS).
To examine the potential effectiveness of cigar warning signs, we conducted a between-subjects online experiment. Participants were 809 U.S. adults who reported using cigars (78% ever large cigar use, 49% past 30-day large cigar use) recruited from a probability-based panel. Participants viewed an image of a cigar store countertop with randomization to one of four conditions: (1) no warning sign, (2) a sign with six FDA proposed text-only warnings, (3) a sign with six novel text-only warnings, or (4) a sign with six novel text + image warnings. We used analysis of variance (ANOVA) models and post hoc Tukey tests to examine the results.
The FDA-proposed text-only warning sign was perceived as less effective in discouraging participants from smoking cigars (M: 3.26, SD: 1.39; scale range: 1–5, where five indicates higher discouragement) compared with the novel text-only warning sign (M = 3.38, SD = 1.40) and the novel text + image warning sign (M = 3.65, SD = 1.34). The novel text + image warning sign increased discouragement from smoking cigars versus the FDA-proposed text-only warning sign (p = .02) and decreased the perceived satisfaction of smoking cigars versus no warning sign (p = .04). In a sensitivity analysis, the novel text + image warning sign decreased the perceived satisfaction of smoking cigars (p = .01), decreased cigar purchase intentions (p = .03), decreased the urge to smoke (p = .03), and increased discouragement from smoking cigars (p = .006) compared with all other study conditions.
Results provide new evidence that policymakers, such as the FDA, could use when proposing POS warning signs for cigars.
The US FDA proposed that retailers display a warning sign at the POS for cigars sold individually without packaging. We conducted an online experiment concerning the potential effectiveness of this regulatory policy with people who use cigars recruited from a probability-based panel. Results provide the first evidence that the FDA-proposed text-only warning sign was perceived as less effective than other types of warning signs and that adding images could potentially increase the effectiveness of warning signs. These findings are particularly relevant for premium cigars, which are often sold individually in brick-and-mortar retail settings.
Introduction
Cigars can cause multiple cancers, cardiovascular disease, lung disease, and other adverse health effects.1 The risks from diverse types of cigar products (e.g. premium, cigarillo, and filtered little cigar) are not inherently different from one another, but the prevalence and patterns of use differ.2 In 2016, the US Food and Drug Administration (FDA) extended its regulatory authority over cigars.2 While the FDA proposed that most cigar products include required warning statements on the product packaging, some cigars are sold individually and not in product packages.3 Therefore, the FDA proposed that retailers display six warning statements (Supplementary Table A) on a sign posted at the point-of-sale (POS) within three inches of the cash register where payments are made.3 This regulation may be particularly relevant for brick-and-mortar retailers selling premium cigars, which for this study, refers to large, handmade cigars that are often sold individually.2 Although the implementation of warning labels for cigars has been vacated because of a court decision,4 retailers may voluntarily comply with the cigar POS warning sign regulation. No published research, to the best of our knowledge, has examined POS warning signs for cigars sold individually without packaging.
Methods
Participants
NORC at the University of Chicago recruited a final sample of 809 participants for our study in March 2022. This experiment was part of a larger study examining the impact of larger, pictorial warning labels on little cigar and cigarillo packages. To be eligible for participation, adults had to be a part of the AmeriSpeak probability-based panel, be 18 years or older, speak English, live in the United States, and report past 30-day use of little cigars or cigarillos. We chose to include people who used little cigars or cigarillos in this experiment for two reasons. First, the warnings proposed by the FDA were created for all cigar products. Therefore, people who use any cigar product will likely be exposed to these warnings, including potentially at the POS. Second, research shows that many people who use cigars concurrently use multiple types of cigars.5 Therefore, we anticipated that many people who used little cigars or cigarillos would use or be at risk of using traditional large or premium cigars. NORC randomly selected U.S. households using area probability and address-based sampling, with a known, non-zero probability of selection from the NORC National Sample Frame. The AmeriSpeak panel provides sample coverage of approximately 97% of the US household population. A total of 840 people were eligible for the study (12.5% of those completing the screener), and of those, 809 completed the study and met the standards of data quality review (i.e. did not speed through the survey or skipped most of the questions).
Procedures
We preregistered this online between-subjects experiment at aspredicted.org (#88507). Participants viewed an image of a cigar store countertop with randomization to one of four conditions: (1) no warning sign (n = 213), (2) a warning sign with the six FDA-proposed text-only warnings (n = 194), (3) a warning sign with six novel text-only warnings developed and tested by our study team (n = 191), or (4) a warning sign with six novel text + image warnings developed and tested by our study team (n = 211) (Figure 1). We selected the novel warning statements and images because they performed well in previous experiments conducted by our study team (Clark S, et al., unpublished data, March 2022).6 The selected images were either graphic or depicted personal suffering. The University of North Carolina at Chapel Hill Institutional Review Board approved the study.
Stimuli for the experimental conditions. Participants saw an image of a cigar store countertop at the point-of-sale with randomization to one of four conditions: (1) no warning sign, (2) a warning sign with six Food and Drug Administration proposed text-only warnings, (3) a warning sign with six novel text-only warnings developed and tested by our study team, or (4) a warning sign with six novel text + image warnings developed and tested by our study team. * Denotes that images were shown to participants, but we are not able to publish the images for copyright reasons. These images were: (A) a side-by-side image of a younger person experiencing the aftermath of a stroke and a close-up image of damaged brain tissue (stroke and blood clots), (B) a side-by-side image of a person with a large scar from bowel surgery with a colostomy bag filled with waste and a close-up of a colon cancer tumor (colon cancer) (C) a side-by-side image of a man with a laryngectomy stoma and a close-up of a tumor inside a person’s mouth (pharyngeal and throat cancer), (D) a side-by-side image of a person with a scar extending down their neck and a close-up image of an esophageal tumor (esophageal cancer), (E) an image containing two pairs of lungs where one is healthy and one is diseased and damaged (lung cancer and lung damage) and (F) an image showing blood in a toilet (bladder cancer and blood in urine). If you would like to see the images, please contact the corresponding author.
Measures
Before the experiment, we asked participants about ever and past 30-day cigar use using questions adapted from the Population Assessment of Tobacco and Health (PATH) Study. To assess large cigar use, we provided the description “Large, traditional cigars contain tightly rolled tobacco that is wrapped in a leaf and is bigger than cigarillos and little cigars.” We also included three premium cigar brands as examples (i.e. Macanudo, Romeo y Julieta, and Arturo Fuente) and a picture of a premium cigar.
To introduce the experiment, we asked participants to imagine that they were at a shop that sells large cigars and instructed them that they would see an image of a counter where the cigars were sold. After participants viewed the cigar store countertop from their assigned condition, we adapted survey items from previous studies and asked participants about their (1) smoking satisfaction using three items from a validated scale (Cronbach’s alpha = 0.97),7 (2) purchase intentions,8 (3) urge to smoke,9 and (4) cigar harm perceptions.10 For participants assigned to see a warning sign (conditions 2–4), we assessed how much the warning sign would discourage them from smoking cigars using one item from a perceived message effectiveness scale.11 All outcome measures asked about “large cigars” because (1) this is often the language used to describe these products,12 and (2) it was similar to measures used in the PATH Study. All outcome measures can be seen in Supplementary Table B. After the experiment, we assessed demographics.
Analysis
For the main analysis, we used analysis of variance (ANOVA) models and post hoc Tukey tests to compare the effects of experimental conditions on outcomes (e.g. smoking satisfaction scale). Because this was an experiment and participants were randomly assigned to each condition (all p-values > .05 in the randomization checks), we did not control for demographic variables or cigar use in the analyses. We hypothesized that novel text + image warning sign would be more impactful than all other experimental conditions and that any of the warning sign conditions would be more impactful than the no warning sign condition. We conducted two additional analyses that were not preregistered, including: (1) a sensitivity analysis where we examined the impact of the novel text + image warning sign versus all other experimental conditions combined into a single comparison group using an independent groups t-test, and (2) a moderation analysis where we examined whether past 30-day large cigar use moderated the effects of experimental condition on outcomes. We conducted analyses in SAS (version 9.4) and used two-tailed tests and a critical alpha of 0.05.
Results
Participant Characteristics
The mean age of participants was 44.1 (SD: 15.3), and about half of the sample identified as female (51.9%) (Supplementary Table C). About one-fifth of the sample identified as black non-Hispanic (19.2%) or Hispanic (17.2%). More than three-quarters of the sample reported ever using large cigars (77.5%), and almost half of the sample (49.2%) reported past 30-day large cigar use.
Main Analysis
The FDA proposed text-only warning sign was perceived as less effective in discouraging participants from wanting to smoke cigars (M = 3.26, SD = 1.39) compared with the novel text-only warning sign (M = 3.38, SD = 1.40) and the novel text + image warning sign (M = 3.65, SD = 1.34) on a scale of 1–5, where five indicates higher discouragement (Table 1). The novel text + image warning sign decreased the perceived satisfaction of smoking cigars compared with no warning sign (p = .04) and increased discouragement from smoking cigars compared with the FDA-proposed text-only warning sign (p = .02). There were no other significant differences between conditions on outcomes.
Sensitivity and Moderation Analysis
In a sensitivity analysis, we found that the novel text + image warnings decreased the perceived satisfaction of smoking cigars (p = .01), decreased cigar purchase intentions (p = .03), decreased the urge to smoke (p = .03), and increased discouragement from smoking cigars (p = .006) compared with all other study conditions (Supplementary Table D). We also examined whether past 30-day large cigar use (n = 393, 49.2%) moderated the effects of randomly assigned conditions on outcomes but found no evidence of significant interactions.
Discussion
Hundreds of thousands of adults in the United States use premium cigars.2 Yet, little research has been conducted on regulatory policies that could prevent or reduce premium cigar use, including warning signs at the POS. In our experiment, we found that the FDA-proposed text-only warning sign was perceived as less effective than other warning signs tested and that adding images to warning signs decreased perceptions of satisfaction from smoking cigars and discouraged cigar use. These findings are important because many people who use premium cigars usually purchase cigars in person (with estimates of in-person purchasing around 75%).2,5 If the FDA decides to pursue cigar warning requirements again, our findings can be used to potentially strengthen the proposed text-only POS warning signs by adding images.
We found that the warning signs with text-only statements did not significantly outperform the condition with no warning sign, which suggests that the warning images primarily drove results and extends research demonstrating the importance of pictorial warnings.13,14 Indeed, from our sensitivity analysis, which integrated the text-only and no-warning control conditions, we found consistent results on smoking satisfaction, purchase intentions, urge to smoke, and discouragement from smoking cigars. These findings suggest that more research, including in-person studies, longitudinal studies, and studies using objective measures of attention, is needed to determine if the warning sign regulatory policy, as currently written and with text-only warning statements, is sufficient to change attitudes, beliefs, intentions, and behaviors. Such studies are critical for supporting FDA regulatory actions, including efforts to stave off industry litigation—which is important for U.S. policy implementation.
Our findings that adding images to POS warning signs decreased perceptions of satisfaction from smoking cigars and discouraged cigar use extends previous research, which has documented significant associations between exposure to POS cigarette warning signs and thinking about the risks of tobacco,15 experiencing negative emotions,16 perceiving the signs to be effective,16 higher quit intentions,17 and increased quit attempts.17 Our study also adds to the evidence base on the effectiveness of pictorial warnings, particularly for cigars, for which there has been little research.18 However, the effectiveness of any POS tobacco warning sign likely depends on several factors, including placement, size, context, and message theme, among others. Importantly, none of these previous studies evaluated six warning statements in one sign—as proposed by the FDA—or specifically examined cigars. Based on our findings, future studies could try to discover how many images and warnings at the POS are optimal.
Finally, we hypothesized that POS warning signs would increase harm perceptions because cigars are often advertised and promoted as less harmful than other tobacco products,2 and cigars have not been the focus of widespread communication campaigns as seen with other tobacco products, like cigarettes. However, we found no significant effects of warning sign conditions on harm perceptions. The lack of significant findings could be explained by a ceiling effect (i.e. the harm perceptions means were relatively high) or the result of a one-time experiment not being able to change misperceptions associated with cigar use harm.19 Alternatively, it is possible that the experiment did change harm perceptions of specific health risks mentioned in the warnings, but we were not able to capture this change since we measured general harm perceptions.
Strengths of our study include our use of a probability-based panel and an experimental design. Limitations include that this study was not conducted in an actual cigar shop so results may not generalize to real-world settings; this was a one-time experiment that precluded measurement of future behavior; and this study did not distinguish between people who use large cigars versus those premium cigars; Furthermore, it is possible that the result may have differed if the study population more accurately represented people who use premium cigars (i.e. more white males with higher levels of education and income).2 Finally, our study tested warning signs for cigars sold individually without packaging, which often includes but is not necessarily limited to premium cigars.
Conclusions
This experiment provides new evidence that the FDA-proposed text-only warning sign for cigars sold individually without packaging was perceived as less effective than other types of warning signs and that adding images to warning signs decreased perceptions of satisfaction, increased discouragement from smoking cigars, and likely improved the warning signs’ effectiveness. These findings are particularly relevant for premium cigars, which are often sold individually in brick-and-mortar retail settings.
Supplementary Material
A contributorship form detailing each author’s specific involvement with this content, as well as any supplementary data, are available online at https://academic.oup.com/ntr.
Funding
Research reported in this publication was supported by NCI and FDA Center for Tobacco Products (CTP) under Award Number R01CA240732. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH or the Food and Drug Administration.
Acknowledgments
None.
Supplement Sponsorship
This article appears as part of the supplement “Regulatory Research Advances on Premium Cigars,” sponsored by the Center for Coordination of Analytics, Science, Enhancement, and Logistics (CASEL) in Tobacco Regulatory Science (5U54DA046060) from the National Institute on Drug Abuse at NIH and FDA’s Center for Tobacco Products.
Data Availability
Data are available upon request to Dr. Adam O. Goldstein ([email protected]).
Comments