Is the Public Ready for a Tobacco-Free Ireland? A National Survey of Public Knowledge and Attitudes to Tobacco Endgame in Ireland

Sampling, recruitment and fieldwork

Sampling, recruitment, and data collection were conducted by an Irish-based market research company (IPSOS MRBI) from February 15^th to 28^th 2022. The target population was defined as members of the Irish general public aged 15 years or older including tobacco/e-cigarette users and non-users. Based on previous similar research in New Zealand it was hypothesised that prevalence of support for TFI may lie between 50-75%.³³ A precision-based sample size was calculated based on the conservative assumption that 50% of the public reported support for TFI; a sample size of 784 was sufficient to measure this proportion with a 95% confidence interval (CI) of +/-3.5%; therefore a sample of size 1,000 was deemed more than sufficient for the purpose of the study.⁵¹

The market research company recruited a sample of 1,000 participants via random-digit-dialling using records of known mobile and landline prefixes provided by the Commission for Communications Regulation. The mobile:landline sample ratio was 85:15 and population coverage was estimated to be in excess of 99%. Quota sampling to reflect age, gender, region and social class enabled minor post-hoc weighting to adjust for generalisability where required. To obtain 1,000 complete responses, 3,386 individuals were contacted, equating to a response rate of 29.5% of those contacted. Participants were excluded if they were non-fluent in English or if they did not complete the survey in its entirety.

The survey consisted of 35 questions delivered via computer-aided telephone interviewing (CATI) as part of an omnibus survey (See questionnaire, supplementary file).⁵² Prior to administration, trained field interviewers were briefed on survey content including definitions of key terms and piloting was conducted. Informed verbal consent to participate was sought and secured prior to interviewing by trained interviewers. Questions were rotated and average completion time was 24 minutes. Fieldwork quality assurance procedures and data checking post-collection optimised data accuracy. Data were inputted via pre-coded responses (drop-down menus). To monitor interview quality, test interviews were monitored by supervisory staff. Ethical approval was granted by the Royal College of Physicians of Ireland Research Ethics Committee.

Measures

Independent variables included demographic characteristics and tobacco/e-cigarette use behaviours (See data dictionary, supplementary file). To allow comparisons of key views among those at highest risk of tobacco-related harm, two variables (smoking status and e-cigarette use) were combined to produce the binary composite variable (tobacco/e-cigarette use); for this variable, those who responded ‘don’t know’ (n=6) were excluded from comparative analysis.

The principle dependent variables were support for the TFI goal and component tobacco endgame measures, which were dichotomised (‘support’/’no support’). ‘Support’ in this context was defined as agreement (‘strongly agree’/’somewhat agree’); ‘no support’ was defined as absence of support for such measures – aligning with previous literature this included those who responded ‘neither agree nor disagree’/’somewhat disagree’/’strongly disagree’/’don’t know’.^{24 34 44} Binary variables were also derived for views on government and Health Service Executive (HSE) action to tackle tobacco-related harm and for opinions on perceived TFI goal achievability.

Analysis

Data were analysed using IBM SPSS Statistics Version 26.0.⁵³ Frequency-based weights were applied throughout all analyses. Descriptive statistics were used determine counts and proportions of categorical demographic variables. Prevalence of public knowledge and attitudes to the TFI goal and views on government and HSE action to combat tobacco-related harm are presented as weighted prevalence estimates with 95% confidence intervals. Pearson’s Chi-squared test was used to compare differences in responses between tobacco/e-cigarette users and non-users. Prevalence of support for tobacco endgame measures was analysed among the total sample and sub-analysed by participant tobacco/e-cigarette use. Participant views regarding underlying pre-conditions and an acceptable timeframe for a tobacco sales phase-out were also analysed. Univariate logistic regression was used to assess associations between TFI goal support and respondent characteristics. Multivariable logistic regression modelling was then used to assess factors associated with TFI goal support, adjusting for relevant confounders.^31-33,47 Confidence intervals were calculated for unadjusted and regression-adjusted odds ratios. Statistical significance was determined at the 0.05 level.

Views on Government Action to Tackle Tobacco-Related Harm

Most respondents (76.2%, 95% CI 73.6%-78.8%) believed the government should do more to tackle the harm done by smoking. There was a significant difference in this view between tobacco/e-cigarette users and non-users (68.4% versus 78.2, p=0.004). Less than half (42.2%, 95% CI 39.1%-45.3%) believed the government were doing enough to ensure the TFI goal is achieved.

Awareness and Attitudes to the TFI Goal

One third of respondents (34.0%, 95% CI 31.1%-36.9%) were aware of the TFI goal. Three-quarters (74.6%, 95% CI 71.9%-77.3%) supported the TFI goal; a significant difference in support was evident between tobacco/e-cigarette users and non-users (54.4% versus 79.4%, p<0.001). The majority of respondents believed the TFI goal was achievable (76.6%, 95% CI 74.0%-79.2%); while 16.5% of respondents considered the goal achievable by 2025, most (60.2%) considered it achievable beyond 2025.

Factors Associated with TFI Goal Support

Following adjustment, females (aOR 1.55, 95% CI 1.13-2.14, p=0.007), those of higher social class (aOR 1.48, 95% CI 1.03-2.12, p=0.035), and those of higher educational attainment (aOR 1.67, 95% CI 1.15-2.43, p=0.007) were significantly more likely to support the TFI goal than their comparative counterparts (Table 2). Compared to those aged 25-34 years, those aged 65 and older had almost 3 times higher odds of supporting the TFI goal (aOR 2.97, 95% CI 1.65-5.37, p<0.001) while non-tobacco/e-cigarette users remained significantly more likely to support the TFI goal compared to users (aOR 2.68, 95% CI 1.83-3.90, p<0.001).

Support for Tobacco Endgame Measures

Product-Focused Measures

The highest prevalence of support was evident for lowering the nicotine content in tobacco products (86.1%, 95% CI 84.0%-88.2%) and e-cigarettes (85.6%, 95% CI 83.4%-87.8%) to make the products less addictive; and for tighter regulation of tobacco products (79.0%, 95% CI 76.5%-81.5%) (Figure 1). Compared to tobacco/e-cigarette users, non-users were significantly more likely to support all six product-focused measures (Supplementary Table 1).

• Download figure
• Open in new tab

Figure 1: Percentage of Respondents who Supported Product-Focused and Institutional Structure-Focused Tobacco Endgame Measures (N=1,000)*

Results are weighted and may not sum to totals; Error bars represent 95% CI; *Statistically significant difference between groups on chi squared test.

User-Focused Measures

The highest support was found for banning tobacco product sales near playgrounds, schools and university campuses (78.2%, 95% CI 75.6%-80.8%) and raising the legal age of purchasing tobacco products to 21 years and older (70.6%, 95% CI 67.8%-73.4%) (Figure 2). Compared to tobacco/e-cigarette users, non-users were significantly more likely to support all six user-focused measures.

• Download figure
• Open in new tab

Figure 2: Percentage of Respondents who Supported User-Focused Tobacco Endgame Measures (N=1,000)*

Results are weighted and may not sum to totals; Error bars represent 95% CI; *Statistically significant difference between groups on chi squared test.

Supply-Focused Measures

Highest support was evident for phasing out tobacco products sales (82.8%, 95% CI 80.5%-85.1%), and for requiring shops selling tobacco products to display information that encourages users to quit (81.9%, 95% CI 79.5%-84.3%) (Figure 3). Compared to tobacco/e-cigarette users, non-users where significantly more likely to support seven of eight supply-focused measures. Of those who supported a complete phase-out of tobacco product sales, for 85.0% support was contingent on provisions for people currently-addicted; either increased government assistance to help people who smoke to quit (74.8%) or allowing smokers to continue to buy tobacco products using a licence (40.8%). Views were similar across tobacco/e-cigarette use status. Two thirds (66.7%) of those who supported a complete phase-out believed this should occur within the next 10 years.

• Download figure
• Open in new tab

Figure 3: Percentage of Respondents who Supported Supply-Focused Tobacco Endgame Measures (N=1,000)

Results are weighted and may not sum to totals; Error bars represent 95% CI; *Statistically significant difference between groups on chi squared test; ^Banned from smaller local shops, newsagents, off-licences and petrol stations.

Key Findings

While Ireland was early to adopt a tobacco endgame goal through the TFI policy in 2013, achievement by the 2025 target is increasingly unlikely and current legislative plans lack the innovation and audacity needed to end the epidemic tobacco in the short-term.⁵⁹ Public opinion has the potential to catalyse, shape and progress tobacco endgame.^11,16 This study found that that despite low TFI goal awareness, 74.6% supported the goal and 76.6% believed it was achievable. Although lower among tobacco/e-cigarette users, goal support was still high at over 50%. There was majority public support for 19 of 22 endgame measures and particular support was evident for product-focused measures. A key finding was that high public support for phasing-out tobacco sales was contingent on supports for currently-addicted users. In addition to tobacco/e-cigarette use status, variation in public support was also evident across age, gender and socio-economic status.

Knowledge and Attitudes to Tobacco Endgame in Ireland

High Irish public support for tobacco endgame among both tobacco/e-cigarette users and non-users was comparable to recent international findings.^30-32 Such support was grounded in realism since a majority (54.3%) considered it achievable by 2035. Irish awareness of the TFI goal was lower than comparable evidence from New Zealand,³³ suggesting scope to raise awareness. However, in this study, TFI goal awareness was not associated with support. Belief that government should do more to tackle smoking-related harm was higher than comparable findings from England and New Zealand, suggesting a mandate for policymakers to take action to progress endgame in Ireland.^30,33,47

Factors Associated with TFI Support

Consistent with international findings, females, non-tobacco/e-cigarette users and those aged 65 and older were more likely to support the TFI goal.³¹ Both higher educational attainment and higher social class were associated with increased goal support, indicating a positive association between goal support and socio-economic status. Against a context of widening inequalities in smoking in Ireland, this highlights the need for equitable endgame planning and communications. While belief the TFI goal was achievable was associated with increased goal support, contrary to previous literature, prior TFI goal awareness was not associated with increased support,³² suggesting support was high regardless of awareness.

Support for Potential Tobacco Endgame Measures

All product-focused and institutional structure-focused measures had majority support. Particularly high support for reducing tobacco product nicotine content to make them less addictive aligned with international findings and policy commitments,^{20,33-37,49,60,61} most notably in New Zealand and the USA,^60,61 as did requiring tobacco companies to pay for tobacco-related health costs.^30,38-41 Support for a ban on filters was weaker, particularly among users. Tobacco industry misinformation may encourage belief that filters protect users – conversely, filters reduce the harshness of tobacco smoke, increasing toxic inhalant exposure.⁶²

A majority of participants supported five of six user-focused measures and six of eight supply-focused measures. Mirroring national and most international literature, “Tobacco 21” was highly-supported by 70.6%,^{28,30,38-42,63} as were other youth-focused measures; e.g. banning sales near playgrounds, schools and universities.^{31,38-40,42,43} This finding is well-aligned with the priorities of non-governmental organisations in Ireland, which have already mobilised as an alliance to advocate for upward adjustment of age limits on tobacco retail.^18,19 An exception was the Tobacco-Free Generation policy, which had less support, however support was still higher than most international literature.^{33,38-40,45,64} Question phrasing, which had prohibitionist connotations, may have influenced findings.⁶⁵ Similarly, participants may have misunderstood unfamiliar endgame measures, suggesting the need to raise awareness of more complex proposals.

Support for a user-licence to purchase tobacco (40.3%) and for restricting e-cigarette sales to prescription-only access (43.2%) was lower than most international estimates.^38-40,45 Substantial tax increases, a ban on tobacco use in public places and restricting tobacco retailers in terms of place, number and hours of sale, attracted minority support among tobacco/e-cigarette users. Support for raising tobacco taxation is likely to increase if revenue is used to fund healthcare initiatives,^33,42 therefore it would be useful to assess if hypothecated tax increases would be more acceptable to users.

While the survey did not explore why certain measures were unpopular, some with less support shared a focus on individual-level restrictions for users, which could be perceived as punitive. Moreover, support for a phase out of tobacco sales was contingent on support for current users. Such findings indicate public opinion aligns with endgame principles which emphasise action tackling systems-factors perpetuating the tobacco epidemic over individual-level factors, and the importance of pursuing measures which support rather than punish current users.¹⁰

Support for a phase-out was higher than international findings,^{31,35,41,45-48} and much higher than in four Northern European countries including Ireland in 2012,²⁹ suggesting support has increased in the past decade. This measure garnered higher support than less radical proposals e.g. restricting hours when tobacco can be sold. The term phase-out implies a gradual approach, potentially influencing findings.

Two-thirds of supporters believed a phase-out should occur within the next ten years –a higher proportion than international estimates, suggesting such a strategy could be acceptable in the medium-term.^46,66,67 Congruent with international literature, support was highly-contingent upon measures for currently-addicted users.^24,33,49,50 Requiring tobacco retailers to display information encouraging users to quit was also highly-supported, again emphasising the importance of quit supports. Conversely, congruent with findings from New Zealand, a majority of respondents did not support training retail staff to provide quitting advice at point-of-sale.³¹ Qualitative research, especially with current users, would be helpful to identify the most acceptable types of user supports.

Overall, support was significantly higher among non-tobacco/e-cigarette users for 20 of 22 endgame measures. Higher support among non-users for disruptive tobacco control measures is well-documented, particularly in Europe.^24,27,29,68 However, evidence suggests support for novel measures can increase during and post-implementation, as was the case with the Irish workplace smoking ban and plain packaging in Australia.^6,69 Therefore, comparatively lower support should not preclude implementation in the longer-term. However, further exploration of reasons for lower support for some measures is merited, particularly among users.

Strengths and Limitations

This is the first study in Ireland providing timely and comprehensive data on knowledge and attitudes of tobacco endgame among the general Irish population. It fills an important research gap and brings the public voice into policy formulation on this key public health challenge.^70,71 The cross-sectional study design was cost-efficient, time-efficient and appropriate for estimating prevalence of individual-level knowledge and attitudes.⁷² The large population-based sample afforded adequate statistical precision to estimate prevalence of key variables.

Notwithstanding measures to ensure sample representativeness, those not covered by the sampling frame (those without a registered mobile or landline number/without fluency in English) could not be included and no data were available on those who declined to participate; if such groups were systematically different to responders this could result in selection bias, specifically non-coverage and non-response bias respectively, reducing external validity. To mitigate such bias a post-hoc weighting strategy was employed. Still, comparisons with population estimates suggested smokers may have been under-represented and e-cigarette users over-represented, groups which had different views on the subject of the survey to the general population. However, weighted sample estimates for tobacco/e-cigarette use prevalence approximated corresponding population estimates.

Self-reported measurements may have introduced social desirability bias - stigma surrounding disagreement with tobacco control measures may have falsely overestimated true prevalence of support for endgame measures.⁷³ Moreover, interviewer-administered telephone interviews are associated with increased risk of social desirability bias compared to self-administered surveys.⁷³ Finally, although this study detected associations between TFI goal support and respondent characteristics, the cross-sectional nature of this study precludes causal inferences and omission of unmeasured variables from the final regression model (e.g. ethnicity) may have resulted in residual confounding.