Article Text

Download PDFPDF
Impact of standardised packaging in the UK on warning salience, appeal, harm perceptions and cessation-related behaviours: a longitudinal online survey
  1. Crawford Moodie1,
  2. Catherine Best1,
  3. Sara C Hitchman2,
  4. Nathan Critchlow1,
  5. Anne-Marie MacKintosh1,
  6. Ann McNeill2,
  7. Martine Stead1
  1. 1 Institute for Social Marketing and Health, Faculty of Health Sciences and Sport, University of Stirling, Stirling, UK
  2. 2 National Addictions Centre, Institute of Psychiatry, King's College London, London, UK
  1. Correspondence to Dr Crawford Moodie, Institute for Social Marketing and Health, Faculty of Health Sciences and Sport, University of Stirling, Stirling FK9 4LA, UK; c.s.moodie{at}stir.ac.uk

Abstract

Introduction In the UK, since 20 May 2017, tobacco companies must sell cigarettes and rolling tobacco in standardised packs.

Methods Three waves of a longitudinal online survey with smokers (≥16) before standardised packaging (wave 1 (W1): April to May 2016) and after standardised packaging (wave 2 (W2): September to November 2017; wave 3 (W3): May to July 2019). Of the 6233 smokers at W1, 4293 responded at W2 and 3175 at W3. We explored smokers’ response to warning salience, appeal (appeal, quality, value, satisfaction and taste compared with a year ago), harm (harmfulness compared with a year ago, harm compared with other brands and whether some brands have more harmful substances), and quit plans, attempts and quitting.

Results Compared with W1, the proportions noticing warnings first on packs, and rating cigarettes/rolling tobacco less appealing and worse value than a year ago, were higher at W2 and W3. Disagreeing that some brands contain more harmful substances was higher at W2. Interactions between social grade and survey wave for warning salience, and each appeal and harm outcome, were non-significant. Smokers switching from not noticing warnings first at W1 to noticing warnings first at W2, or who had a lower composite appeal score at W2, were more likely to plan to quit and to have made a quit attempt at W2. Smokers who switched to disagreeing that some brands contain more harmful substances at W2, after giving a different response at W1, were more likely to quit at W3.

Conclusions Standardised packaging appears to be having the intended impacts.

  • packaging and labelling
  • public policy
  • surveillance and monitoring

Data availability statement

Data are available upon reasonable request. Data will be available on project completion and when planned articles are submitted.

Statistics from Altmetric.com

Request Permissions

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.

Introduction

Belgium became the 15th country to require tobacco products to be sold in standardised (or plain) packaging in January 2021. The objectives of standardised packaging (SP) are to discourage uptake, encourage cessation, discourage relapse and reduce exposure to tobacco smoke.1–3 It is expected that SP will help meet these objectives by increasing the salience of the on-pack warnings, reducing the appeal of tobacco products and reducing misperceptions of harm as a consequence of pack design.1–3 Since May 2017, cigarettes and rolling tobacco in the UK must be sold in drab brown packs with pictorial warnings on 65% of the main display areas (previously text warnings covered 43% of the pack front and pictorial warnings 53% of the reverse) (see figure 1). We explore whether SP in the UK increased warning salience and reduced appeal and harm perceptions among smokers aged 16 and over, and the impact, if any, on cessation-related behaviours.

Figure 1

Fully branded and standardised packs in the UK.

For warning salience, a survey in England in the last 6 months of the (12-month) transition period, when both standardised and fully branded packs were sold,4 5 found that smokers using SP were more likely than those who had never used SP to notice warnings often.6 A longitudinal survey in England found that the proportion of smokers usually noticing warnings first on packs increased from 18% at the start of the transition period to 45% up to 12 months post-SP.7 A longitudinal survey in the UK and Norway (where SP was introduced but the warnings, which had been on packs for several years, were unchanged) found that noticing warnings on packs increased in the UK post-SP but not in Norway.8 In Australia, cross-sectional tracking surveys and longitudinal surveys found an increase in warning salience among smokers up to 15 months post implementation.9–12

For appeal, a longitudinal survey in England found that smokers were four times more likely to ‘not at all’ like the look of their pack following the introduction of SP than they were prior to it being implemented.7 There was also a decrease in the proportion reporting how much brands differed in terms of prestige, but an increase in reporting that the quality of their cigarettes was high.7 In France, smokers were approximately three times less likely to indicate that they enjoyed the look of their pack post-SP than they were pre-SP.13 In Australia, smokers using SP during the transition period were more likely than those using fully branded packs to rate their cigarettes lower in quality and less satisfying than a year before.14 Cross-sectional tracking surveys show that up to 12 months post-SP there was a significant increase in negative pack perceptions among smokers,9 10 with packs rated as having lower appeal and cigarettes considered lower quality, lower value and less satisfying than a year before.10

For harm, a survey in England in the transition period found that smokers using SP were more likely to think about the risks of smoking because of the look of the pack than smokers who had never used SP.6 A longitudinal survey in England found that the proportion of smokers indicating that their brand was no different in harmfulness, and no smoother or harsher, than other brands, did not change post-SP.7 In France, surveys with smokers and non-smokers found an increase in the perceived dangerousness of smoking and the fear of the consequences of smoking within a year of SP being introduced.15 In Australia, during the transition period, smokers using SP did not differ from smokers using fully branded packs in terms of frequency of thoughts about harms or perceived exaggeration of harms.14 Compared with smokers interviewed pre-SP, those interviewed in the year after SP was mandatory were more likely to believe that brands do not differ in harmfulness, but there was no change in perceived exaggeration of harms or the belief that brand variants do not differ in perceived harmfulness compared with a year before.10

During the transition periods in the UK and Australia, smokers using SP were more likely than those using fully branded packs to indicate that they thought about quitting because of the look of the pack,6 thought about quitting at least once a day in the past week14 and rated quitting as a higher priority.14 Smokers in Australia were interviewed twice, a month apart, with those completing the follow-up in the late transition period more likely to intend to quit, and those followed up in the early transition period or post-SP more likely to have reported having made a quit attempt, than those followed up pre-SP.16 Additional analysis focusing only on smokers recruited post-SP found that several baseline measures suggesting improved warning effectiveness, and reduced appeal, increased the likelihood of quitting-related outcomes.17 For instance, smokers at baseline who noticed warnings first on packs, disliked the look of their pack or reported lower satisfaction from their cigarettes compared with a year ago were more likely to report daily thoughts about quitting at follow-up.17 Smokers attributing much more motivation to quit to warnings at baseline were more likely to have attempted to quit at follow-up.17

We explored whether SP in the UK increased warning salience and reduced appeal and misperceptions of harm and, for those smokers for whom this was the case, whether this was associated with quit plans, attempts and quitting. As socioeconomic differences are often overlooked in SP research,18 we also explored whether any change in warning salience, appeal and harm across waves differed by social grade.

Methods

Design and sample

The ‘Adult Tobacco Policy Survey’ is a longitudinal online survey following cigarette smokers recruited pre-SP (April to May 2016) and followed up 4–6 months post-SP (September to November 2017) and 24–26 months post-SP (May to July 2019). To be eligible for inclusion at wave 1 (W1), participants had to be 16 or over and report smoking cigarettes (factory made and/or hand rolled) in the last 3 months. The sample was recruited from the online panel of market research company YouGov. All W1 participants were eligible for inclusion at future waves. Of the 6233 cigarette smokers at W1, 4293 (69%) participated at wave 2 (W2) (3629 cigarette smokers, 607 ex-cigarette smokers, 36 non-cigarette smokers, 7 cigarette smokers who had not smoked in the past 3 months, 14 missing data on smoking status) and 3175 (51% of the W1 sample) at wave 3 (W3) (2412 cigarette smokers, 700 ex-cigarette smokers, 44 non-cigarette smokers, 6 cigarette smokers who had not smoked in the past 3 months, 13 missing data on smoking status). At each wave, participants received a small incentive. There was an information page at the start of the survey, with consent required.8

Measures

Demographics

Information on gender, age, ethnicity, education, income and social grade was captured. Age was coded as ‘16–24’, ‘25–39’, ‘40–55’ and ‘56 and over’, with ethnicity recoded into ‘White British’, ‘White non-British’, ‘Other ethnic group’ and ‘Not stated’. Highest educational qualification obtained was coded as ‘High school’, ‘Technical, trade school, A levels, or community college’, ‘University degree or higher degree’ and ‘Don’t know or prefer not to say’. Annual household income was recoded as ‘under £30 000’, ‘£30 000–£44 999’, ‘£45 000 and over’ and ‘Don’t know or prefer not to answer’. Social grade was determined by occupation of the main income earner within the household using the National Readership Survey,19 with grades A, B and C1 (signifying upper class and middle-class grades) and C2, D and E (working-class grades) recoded into ‘ABC1’ and ‘C2DE’.

Smoking and cessation-related behaviours

Participants were asked whether they smoked cigarettes (factory made or hand rolled) daily, smoked cigarettes but not every day (weekly, monthly, within the last 3 months, not in the last 3 months), did not smoke cigarettes at all but used some other tobacco product(s) or had quit. All participants at W1 had smoked cigarettes in the past 3 months. Participants were asked ‘Are you planning to quit smoking?’, with response options (Within the next month, Between 1 and 6 months from now, Sometime in the future, Beyond 6 months, Not planning to quit, Don’t know) collapsed into ‘Planning to quit’ versus ‘Not planning to quit’, with ‘Don’t know’ excluded. Participants were asked how many attempts they had made to quit smoking in the past 12 months (‘No attempt’, ‘1 attempt’, ‘2 attempts’, ‘3 or more attempts’ and ‘Don’t know’), collapsed into ‘One or more attempts’ and ‘No attempt’. Quitting at W3 was based on reported quitting when asking about smoking status.

Nicotine dependence

The Heaviness of Smoking Index (HSI)20 combines cigarettes smoked per day (10 or fewer=0, 11–20=1, 21–30=2, 31 or more=3) and time to first cigarette on the days that they smoke (within 5 min=3, 6–30 min=2, 31–60 min=1, 61 min or more=0). Missing cases were included as a missing category.

Warning salience

Participants were asked ‘When you look at a pack of cigarettes/rolling tobacco what do you usually notice first—the warning labels, or other aspects of the pack, such as branding?’ Responses were ‘Warnings labels’, ‘Other aspects of the pack, such as branding’ and ‘Don’t know’. To examine change in warning salience between waves, we compared ‘Warning labels’ to any other response. To examine the association between change in warning salience between W1 and W2 and planning to quit and quit attempts at W2, and quitting at W3, we generated an indicator variable for those who moved from not noticing warnings first at W1 to noticing warnings first at W2.

Appeal

Participants were asked ‘Compared to the cigarettes/rolling tobacco you were smoking a year ago, how would you rate your current cigarettes/rolling tobacco in terms of…?’ ‘Value for money’, ‘Quality’ and ‘Taste’. Response options for each were ‘Better’, ‘About the same’, ‘Worse’ and ‘Don’t know’. They were also asked ‘Compared to the cigarettes/rolling tobacco you were smoking a year ago, how would you rate your current cigarettes/rolling tobacco in terms of whether they are…?’ ‘Appealing’ and ‘Satisfying’. Responses were ‘More’, ‘About the same’, ‘Less’ and ‘Don’t know’. To examine change between waves, we compared ‘Worse’ or ‘Less’ versus any other response. To examine the association between reduced appeal and planning to quit and quit attempts at W2, and quitting at W3, we created a change score for each appeal variable. ‘Worse’ or ‘less’ was scored 3, ‘about the same’ 2 and ‘better’ or ‘more’ 1. The response in W2 was subtracted from that in W1 to create a change score ranging from −2 to +2 with a positive score indicating evaluation becoming more negative over time. To examine the magnitude of the association between change across all appeal variables and cessation-related behaviours, a composite appeal measure was created, with scores summed across the five appeal change variables to create a score from −10 to 10.

Harm

Participants were asked: (1) ‘Compared to the cigarettes/rolling tobacco you were smoking a year ago, how would you rate your current cigarettes/rolling tobacco in terms of harmfulness?’, with response options ‘Higher’, ‘About the same’, ‘Lower’ and ‘Don’t know’; (2) ‘Is your usual/current brand of cigarettes/rolling tobacco a little less harmful, no different, or a little more harmful, compared with other brands?’, with response options ‘Little less harmful than other brands’, ‘No different’, ‘Little more harmful than other brands’ and ‘Don't know’; (3) ‘To what extent do you agree or disagree with the following statement: Some brands of cigarettes contain more harmful substances than other brands?’, with response options ‘Strongly agree’, ‘Agree’, ‘Neither agree nor disagree’, ‘Disagree’, ‘Strongly disagree’ and ‘Don’t know’.

When examining change between waves, we compared the proportion responding ‘Higher’ versus any other response for the first question, ‘No different’ versus any other response for the second question and ‘Strongly disagree or disagree’ versus any other response for the third question. To examine the association between change in perceived harm between W1 and W2 and planning to quit and quit attempts at W2, and quitting at W3, we generated three indicator variables. The first was for participants who moved to considering their usual brand as having higher harmfulness than a year ago at W2 after giving a different response at W1. The second was for participants who moved to considering their current/usual brand no different in harmfulness from other brands at W2 after giving another response at W1, and the third was for participants who disagreed that some cigarette brands contain more harmful substances at W2 after giving another response at W1.

Analysis

First, we examined whether warning salience, appeal and harm perceptions changed over survey waves. To do so, generalised estimating equations (GEE) were employed for each outcome, with wave as the independent variable. The analysis was adjusted for these baseline values: age group, gender, ethnicity, social grade, household income, HSI score and education. GEE used an exchangeable correlation structure and robust SEs. Analyses were not weighted as the sample is not a probability sample and frequency weighting to match baseline population proportions is not useful for a repeated measures analysis. To examine whether any change across waves differed by social grade, we added an interaction term between social grade and wave.

Second, we examined whether change in warning salience, appeal and harm between W1 and W2 was associated with cessation-related behaviours. These outcomes were evaluated using a series of logistic regressions with planning to quit at W2, quit attempts in the previous 12 months at W2 and having to quit at W3 as the dependent variables. The following were independent variables: change in warning salience; change in composite appeal score; change in each of the five appeal variables; and change in each of the three harm variables. Each was regressed against each of the cessation-related dependent variables. The analyses were adjusted for these baseline values: age group, gender, ethnicity, social grade, household income, HSI score and education. Where the dependent variable was planning to quit at W2 then the analysis was adjusted for planning to quit at W1, and where it was quit attempts at W2 the analysis was adjusted for quit attempts at W1.

Where the relationship between change in warning salience, appeal or harm and the cessation-related behaviours was significant, interaction between the independent variable and social grade was explored. Further details of the analysis are presented in sections 4 and 5 of the online supplemental file 1. As a sensitivity check, because questions on appeal, satisfaction, quality, taste and value asked participants to compare these attributes between the cigarettes they were smoking a year ago to those they were currently smoking, we restricted the analysis to brand loyal participants (those reporting smoking the same brand at W1 and W2).

Results

Sociodemographic characteristics of the sample across the three waves are shown in table 1, with cessation-related variables shown in table 2. Missing data rates for warning salience, appeal and harm variables are shown in online supplemental table 1.1.

Table 1

Sociodemographic characteristics of the sample

Table 2

Heaviness of Smoking Index, planning to quit and quit attempts by survey wave

Compared with W1, the proportion noticing the warning first on their pack, or who thought their cigarettes were less appealing and worse value than a year ago, was higher at W2 and W3 (table 3). Compared with W1, disagreeing that some brands contain more harmful substances than others was higher at W2. The interactions between social grade and wave for each outcome were not significant (see online supplemental tables 2.1–2.3 and online supplemental figures 1 and 2).

Table 3

Change across survey waves in warning salience, appeal, satisfaction, quality, taste and value by social group

A change from not noticing warnings first on packs at W1 to noticing warnings first at W2 was associated with higher odds of planning to quit and having made at least one quit attempt at W2, and increased quitting at W3 in an unadjusted model (although this no longer reached significance in an adjusted model) (table 4). A change to lower composite appeal for cigarettes from W1 to W2 was associated with higher odds of planning to quit and having made at least one quit attempt at W2. Among the five appeal measures, a change to lower appeal of cigarettes from W1 to W2 was associated with higher odds of planning to quit at W2, and a change to cigarettes being less satisfying than a year ago was associated with higher odds of having made at least one quit attempt at W2. Participants who indicated that their usual brand was more harmful than a year ago at W2, when they had not responded that way at W1, had higher odds of having made a quit attempt at W2. Participants who changed to disagreeing that some brands contain more harmful substances than others at W2 after giving another response at W1 were more likely to have quit smoking at W3. Where there were statistically significant relationships, potential interactions with social grade were investigated. There were no statistically significant interactions between the change scores and social grade (see table 4 and online supplemental tables 3.1–3.5).

Table 4

Association between change in warning salience, appeal and harm, and planning to quit, quit attempts and quitting

When the analysis was restricted to brand loyal smokers at W1 and W2 (table 5), composite appeal at W2 was no longer significantly associated with a quit attempt at W2, and disagreeing that some brands were more harmful at W2 was no longer significantly associated with quitting at W3.

Table 5

Change in composite score, warning salience, appeal and harm, and planning to quit, quit attempts and quitting for those who smoked the same brand at W1 and W2

Discussion

We found an increase in warning salience post-SP, consistent with past research,6–12 with the proportion of smokers noticing warnings first on packs remaining significantly higher at W3 (24–26 months post-SP) than at W1. The warnings on SP in the UK (and Australia) were new and covered a greater proportion of the pack than they did on fully branded packaging however; in the UK they also started from the top of the pack (rather than the bottom) and displayed pictorial images on the pack front and reverse (rather than just the reverse), with these images rotated annually (compared with not at all). Although it is not possible in this paper to determine whether increased warning salience was a function of the new warnings or SP, recent research highlights the benefits of introducing both simultaneously.7 8

Compared with baseline, smokers considered their cigarettes less appealing and worse value compared with a year ago at the post-SP waves. For the appeal measures, smokers at W3 were asked to compare their cigarettes to a year earlier, and therefore the comparison at this point would be between SP. Nevertheless, for all five appeal measures the proportion of smokers rating their cigarettes negatively declined from W2 to W3. It may be that over time smokers become more accustomed to SP. It may also be that product innovation has, at least in part, helped increase appeal. Tobacco industry journals highlight the increased importance of the cigarette stick as a promotional tool,21–23 particularly in countries with SP.24 There has been a significant filter innovation in the UK, with flavour-changing capsule filters growing in popularity over the study period25 and the continued introduction of firmer and recessed filters, which smokers view favourably.26 Further research in markets with SP monitoring tobacco industry innovations and exploring smokers’ perceptions of these innovations is warranted.

Smokers are more likely than non-smokers to have erroneous perceptions about the harmfulness of smoking due to pack design.27 Whether SP can reduce these misperceptions among smokers is not clear given inconsistent findings.6 7 9 13 14 We found little evidence for a change in harm perceptions across waves, although there was an increase in the proportions disagreeing that some brands contain more harmful substances than others at W2. Consistent with research in Australia,10 no differences between higher (ABC1) and lower (C2DE) social grades for any of the harm, appeal or warning salience measures were observed between waves, suggesting that SP is not having a different impact among these groups.

Smokers who did not report noticing warnings first on packs at W1, but did at W2, were more likely to plan to quit and to have made a quit attempt at W2; they were also significantly more likely to have quit at W3 in an unadjusted model but not in an adjusted model, although the findings were in the same direction. Smokers with lower composite appeal of their cigarettes at W2 were more likely to plan to quit and to have made a quit attempt at W2. The findings are consistent with research in Australia, where it is the warnings on SP and, to a lesser extent, reduced appeal that appear most likely to help drive behaviour change.17 Despite few differences in harm perceptions over the three waves, those who disagreed that some brands were more harmful than others at W2, after giving a different response at W1, were more likely to have quit at W3. Concerns about harm have long been a key reason for quitting.28 This finding would seem to highlight the value of correcting the misperception that different brands carry different levels of harm. Packaging is central to this erroneous belief, with tobacco companies having engineered pack design, through the use of descriptors, colour and size, to give the impression that some brands (or variants) are less harmful than others. By banning potentially misleading pack descriptors, standardising pack colour and banning slimmer designs, SP may help expose this deception.

In terms of limitations, the findings are reliant on self-report. Our sample was drawn from an online panel and therefore excludes those who lack internet access or regular internet access, such as those on lower incomes.29 Attrition is an issue with longitudinal research, with almost half of the sample lost at W3. In addition, our W1 sample under-represented young smokers, who were also more likely to be lost at subsequent waves. Our findings may, therefore, not be generalisable to smokers across the UK.

In conclusion, SP in the UK appears to be achieving its intended impact of increasing warning salience and reducing appeal and some misperceptions of harm among smokers aged 16 and over, and by doing so helping increase cessation-related behaviours.

What this paper adds

  • There is a growing body of research exploring the core objectives of standardised packaging in countries where this policy has been introduced.

  • Using a longitudinal online survey in the UK, we found that warning salience increased and appeal decreased among smokers after standardised packaging.

  • Interactions between social grade and survey wave on warning salience, appeal and harm were non-significant.

  • For smokers for whom standardised packaging made the warnings more salient, or cigarettes less appealing, this was associated with planning to quit and quit attempts. For smokers for whom standardised packaging reduced the misperception that some brands contain more harmful substances than others, this was associated with quitting.

Data availability statement

Data are available upon reasonable request. Data will be available on project completion and when planned articles are submitted.

Ethics statements

Patient consent for publication

Ethics approval

The study received ethical approval from the University of Stirling.

Acknowledgments

The authors would like to thank YouGov for sample recruitment at each wave.

References

Footnotes

  • Twitter @cathbest, @n_critchlow, @kingsNRG

  • Contributors CM and SCH designed the study. CM and MS obtained funding. CB was responsible for data management and analysis. CM drafted the manuscript. All authors approved the final manuscript.

  • Funding The first two waves were funded by Cancer Research UK and the British Heart Foundation (grant number A18507). The third wave was conducted by the Public Health Policy Research Unit (PH-PRU), commissioned and funded by the National Institute for Health Research Policy Research Programme.

  • Disclaimer The views expressed are those of the authors and not necessarily those of the NHS, the National Institute for Health Research, the Department of Health and Social Care or its arm's-length bodies and other government departments.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.