Evaluating Cigarette Pack Insert Messages with Tips to Quit

Abstract

Objectives:

Canada is the only country that currently uses cigarette pack inserts to communicate health messages to smokers, including tips to quit. Messages about strategies for quitting smoking are also central to the US Food and Drug Administration’s (FDA) Every Try Counts (ETC) campaign. This study assessed US smokers’ responses to Canadian and ETC-based messages formatted for pack inserts.

Methods:

US adult smokers (N = 524) were recruited from an online consumer panel and rated 8 insert messages: 4 based on Canadian inserts and 4 based on ETC. Participants randomly viewed each message accompanied by an image of either a person or a symbolic representation of the topic. Participants rated the perceived effectiveness (PE) of each message. Paired t-tests were used to assess mean differences in PE across topics, image types, and quit intentions.

Results:

ETC messages were consistently rated as more effective than Canadian messages regardless of quit intentions. Image types did not significantly influence PE.

Conclusions:

Messages from ETC are perceived as more effective than messages used in Canada. The FDA has the authority to communicate with smokers through inserts and should consider adopting inserts to promote smoking cessation.

Prominent pictorial health warning labels (PHWLs) that illustrate the harms of tobacco use are a fundamental policy for addressing the growing burden of tobacco use around the world.¹ Since Canada was the first country to implement PHWLs in 2001, more than 100 countries have adopted this policy.² A little-known feature of the Canadian policy is its inclusion of inserts placed or printed inside of packs that contain messages with tips to quit (ie, self-efficacy messages) and information on the benefits of quitting (ie, response efficacy messages). This approach of raising fear about a health issue with PHWLs and enhancing efficacy with inserts is consistent with both communication theories and meta-analytic evidence about ways to maximize behavior change.^3-7 Observational studies in Canada suggest that these inserts promote self-efficacy to quit, and reading them is associated with sustained smoking cessation.⁸ The US Food and Drug Administration (FDA) is scheduled to implement PHWLs in 2021 and has regulatory authority to communicate with smokers through inserts.⁹ Brief experiments with US smokers found that self-efficacy messages based on Canadian inserts are less effective than response efficacy messages,^10,11 which suggests that self-efficacy messages delivered through inserts could be improved.

Every Try Counts (ETC) is FDA’s point-of-sale campaign uses advertisement in convenience stores and gas stations that targets smokers with recent, unsuccessful quit attempts.¹² ETC messages about the importance of learning from prior quit attempts and taking small steps towards quitting were tested in focus groups with key subpopulations of US smokers who want to quit.¹³ These topics are not emphasized in Canadian insert messages. Furthermore, ETC content is not formatted for inserts, although they can be easily adapted. Additionally, the type of image used in efficacy messages, whether the image is of a person or a symbolic representation of the message, can influence its effectiveness.¹¹ This study aimed to determine the relative effectiveness of ETC- and Canadian-style text messages and visual images, designed for use in cigarette pack inserts.

METHODS

Sample

In May of 2019, 524 US adult smokers were recruited from the Qualtrics online panel. Eligible participants were between 18 and 64 years old, smoked at least 10 cigarettes per day, and had smoked at least 100 cigarettes in their lifetime. Additionally, sample quotas were used for intention to quit within the next 6 months (≥40%) and educational attainment (≥30% high school education or less). Participants who incorrectly answered an attention check question in the middle of the survey were excluded from the sample. Participants received standard compensation from the panel provider. The protocol was approved by the IRB at the University of South Carolina (Pro00083728).

Study Design and Protocol

In a single session, participants viewed and rated 8 inserts. Four messages were selected from those that smokers rated as most to helpful and motivating for quit in studies that adapted Canada’s prior inserts or those pretested during development of their next round of inserts for 2022.^10,11,14 The other 4 messages were derived from ETC advertisements, which we reviewed on the official ETC website and in cessation messages sent through the campaign’s text message services.^12,13,15 All messages are shown in Table 1. Word counts for Canadian topics ranged from 38-62 and the Flesch-Kincaid Grade Level ranged from 3.3-6.5. Word counts for ETC topics ranged from 52-56 and the Flesch-Kincaid Grade Level ranged from 2.4-6.7. The research team consulted a legal scholar to ensure that the topics would be legally defensible in the US regulatory context and worked with a graphic designer to format the inserts. For each of the 8 message topics, we created 2 versions: one with an image of a person and one with a symbolic representation of the message. Participants viewed and evaluated each of the 8 insert topics in random order. For each message, participants were randomized to view one of the 2 executions (ie, image of a person or a symbolic image).

Table 1.

Insert Images

Canadian-style Messages
	C1A- Active Support	C2A- Cravings Fade Away	C3A- Deal with Temptation	C4A- Feeling Stressed
Person Images
	C1B- Active Support	C2B- Cravings Fade Away	C3B- Deal with Temptation	C4B- Feeling Stressed
Symbolic Images
	Every Try Counts Messages
	ETC1A- Small Steps	ETC2A- Every Try Counts	ETC3A- Dealing with Cravings	ETC4A- Delay a Cig
Person Images
	ETC1B- Small Steps	ETC2B- Every Try Counts	ETC3B- Dealing with Cravings	ETC4B- Delay a Cig
Symbolic Images

Dependent Variable

Participant rated each message using 5 measures of perceive effectiveness (PE) adapted from a previously validated scale (ie, this message makes me stop and think, this message teaches me something new, this message makes me feel hopeful about quitting, this message helps me feel confident about how to quit smoking, and this message is one I would talk about with friends and family).^16,17 Response options used a 7-point Likert scale, and as internal consistency was high for each stimulus shown (range α = .94 - .97), responses were averaged for each of the 2 versions of the 8 insert topics.

Potential Covariates

Smoking-related variables included the number of cigarettes smoked per day and time to first cigarette, which were used to derive a heaviness of smoking index (HSI) which ranges from 0-6, with a higher score indicating higher physical addiction; intention to quit within the next 6 months (yes, no); and having made a quit attempt within prior 6 months (yes, no). Sociodemographic characteristics included sex (male, female), age (18-29, 30-45, 46-64), highest level of educational attainment (high school or less, more than high school), income level (< 10,000, 10,000-29,999, 30,000-44,999, 45,000-59,999, > 60,000, and no response), and racial or ethnic group (white, Black, Hispanic or Latino, Asian or Pacific Islander, American Indian, or other).

Statistical Analysis

Mean differences in PE between each of the 8 topics were assessed using paired t-tests, pooling data for imagery alternatives for each topic. Within each topic, mean differences in PE scores for the image type used (ie, symbolic or person) were assessed using unpaired t-tests. These analyses were repeated with the t-tests stratified by quit intention. We used a Bonferroni correction procedure to determine the level of statistical significance (eg, 28 tests for comparing topics, p < 0.0017; 8 tests for comparing images used within topics, p < 0.006). All analyses were conducted using Stata v.16.1 (StataCorp LLC, College Station, TX, USA).

RESULTS

The sample was mostly female (59%) and white (88% white, 4% Black, 4% Hispanic or Latino, 2% Asian or Pacific Islander, 1% American Indian, and 1% another race). The majority of the sample (66%) completed more than a high school degree. For annual income, 6% of participants made less than $10,000 per year, 22% made $10-000-29,999, 17% made $30,000-44,999, 13% made $45,000-59,999, and 40% made $60,000 or more. The mean heaviness of smoking index was 2.45 (SD = 0.05), indicating a low to moderate level of addiction. Finally, most participants (54%) were not intending to quit within 6 months and most (58%) had not attempted to quit within the last 4 months.

Perceived Effectiveness by Message Topic

Mean PE scores for ETC topics were all equal to or higher than the Canadian topics (see Figure 1). ETC3 (Dealing with Cravings) had higher PE scores than all topics (range mean difference = .24-.70, p < .0017) except ETC 4 (Delay a Cig). Similarly, ETC4 (Delay a Cig) ratings were higher than for all topics (range mean difference = .19-.65, p < .0017) except ETC 3 (Dealing with Cravings). PE scores for ETC2 (Every Try Counts) were higher than for C1 (Active Support; mean difference = .46, p < .0017), C2 (Cravings Fade Away; mean difference .33= , p < .0017), and C3 (Deal with Temptation; mean difference = .15, p < .0017), and ETC1 (Small Steps) was higher than C1 (Active Support; mean difference = .37, p < .0017) and C2 (Cravings Fade Away; mean difference = .24, p < .0017). C3 (Deal with Temptation) had a significantly higher score than C1 (Active Support; mean difference = .31, p < .0017) and C2 (Cravings Fade Away; mean difference = .17, p < .0017), as did C4 (Feeling Stressed) in comparison to C1 (Active Support; mean difference = .36, p < .0017) and C2 (Cravings Fade Away; mean difference = .23, p < .0017).

Figure 1. Combined PE Means by Topic for All Participants (N = 524).

C1-C4 are Canadian-based messages. ETC1-ETC4 are messaged developed from FDA’s Every Try Counts campaign. See Table 1 for message content. The bar for each topic is labeled a-h from left to right on the X-axis. Superscripts beside a topic score at the topic of a bar indicate that that topic scored significantly higher (p < .0017) than the topic(s) listed in the superscripts. For example, C1 scored significantly lower than C4.

When data were stratified by quit intention, the pattern of results was the same, although the means were lower for those who did not intend to quit (see Figure 1). PE ratings for ETC3 (Dealing with Cravings) were higher than all topics except ETC4 (Delay a Cig) among participants intending to quit (range = .24-.69, p < .0017) or not (range = .24-.71). The same results were found for ETC4 (Delay a Cig) among participants intending to quit (range = .24-.69, p < .0017) or not ( range = .26-.62, p < .0017).

Perceived Effectiveness by Image Type

There were no statistically significant differences between image types for any topic, with mean differences being relatively small (range of mean difference = 0.01-0.31; p > 0.006).

DISCUSSION

This study found that US-based smokers rate ETC messages formatted for inserts as more effective than the messages used for inserts in Canada. The highest scoring ETC messages provided specific steps to take in order to quit (ie, tips to deal with cravings and tips to delay cigarettes). This suggests that ETC pre-testing¹³ has identified content that is likely to be effective when formatted for inserts, which the FDA has the regulatory authority to mandate.

We found no statistically significant differences between the types of images that accompanied the message topics we tested. Rather, the insert topics themselves appears to better account for smokers’ ratings of PE. This finding is consistent with results from a prior study on efficacy message inserts in which message topic was more important than type of imagery for determining PE.¹¹ That study and others,¹⁸ however, suggest that different types of images fit some message topics better than others. Because we only assessed self-efficacy messages with tips to quit and tested just 2 types of imagery per topic, our messages may not have had featured enough variety in topic or imagery to find a difference in PE. In the case of warning messages focused on specific health outcomes, the congruence of the text with the imagery that illustrates the health outcome appears to enhance the effect;¹⁸ however, in the case of more positively framed messages that aim to promote self-efficacy, the positive imagery tends to be more generic. Future research should assess and further evaluate congruence between imagery and efficacy enhancing text. Furthermore, while this study did not examine whether inclusion of any imagery enhances PE, prior research indicates that this is the case.¹⁰

Smokers who do not intend to quit rated each insert as less effective than smokers who intend to quit, a finding consistent with previous studies in which quit intentions are positively correlated with PE ratings of cessation-related communications.^18-21 However, both groups of smokers had similar patterns of response across message topics—ETC messages being rated as more effective than Canadian messages for both groups. This suggests that ETC message content may be effective for smokers who are not currently interested in quitting, even though these messages were designed for and pretested with smokers who want to quit.

There are several limitations to this study. While PE is predictive of cessation behaviors,²² the actual effectiveness of inserts inside cigarette packs may not substantially impact cessation behaviors if smokers do not read the messages once they take them out of packs. Our study involved forced-exposure, which may produce different responses than those once smokers are repeatedly exposed to them in natural settings. However, observational studies to evaluate Canada’s insert policy suggest that inserts motivate smokers to quit and that attention to inserts increased over time, even as attention to PHWLs decreased over time.¹⁰ Future research should assess the effectiveness of ETC and Canadian-based insert messages in more natural settings, such as randomized controlled trials that provide participants with cigarette packs with inserts and assess changes in cessation-related outcomes over time compared to control messages.^23,24

Another limitation is our convenience sample, which had limited racial/ethnic diversity, reducing the generalizability of the results. Indeed, disadvantaged smokers may benefit less from smoking cessation messages.²⁵ Future research should involve more disadvantaged smokers, who suffer disproportionately from smoking-related diseases, to better gauge the generalizability and potential public health impact of inserts. Nevertheless, this study provides preliminary evidence that ETC messages could be considered for inserts in future policy making.

IMPLICATIONS FOR TOBACCO REGULATION

This study highlights the potential effectiveness of ETC messages for use in pack inserts, adding to the growing body of literature on using cigarette packaging labeling policy to promote smoking cessation.^26-28 In 2012, the FDA adopted PHWLs—which have already been implemented in 125 countries²—but tobacco litigation has delayed implementation until 2021 by claiming that PHWLs violate their First Amendment rights.⁹ An insert policy may be more difficult for the industry to challenge because the messages are inside of packs, and thus would only be seen by smokers and would not interfere with brand information. While more research is warranted, our findings offer evidence that ETC messages reformatted as cigarette package inserts may be effective for promoting smoking cessation.