Support for and likely impacts of endgame measures in the Smokefree Aotearoa Action Plan: findings from the 2020–2021 International Tobacco Control New Zealand (EASE) surveys

Despite public health efforts, an estimated 363,000 New Zealanders still smoke.¹ Smoking is the leading cause of preventable death in Aotearoa New Zealand. Its impact is substantially greater among certain population groups, including among those with socio-economic disadvantage, and among Māori.^2,3 In 2011, in response to a Māori Affairs Select Committee report, the New Zealand Government adopted a goal to reduce smoking prevalence and tobacco availability to minimal levels by 2025 (the Smokefree Aotearoa 2025 Goal).^4,5

While a range of tobacco control interventions were introduced over the subsequent decade, these were largely “business-as-usual” approaches. In December 2021, the New Zealand Government introduced the Smokefree Aotearoa 2025 Action Plan (Smokefree Action Plan),⁶ to reach a goal of <5% daily smoking prevalence for all groups of New Zealanders by 2025. The plan included three world-first “endgame” measures: 1) mandating very low nicotine cigarettes (VLNCs), 2) a substantial reduction in the number of retailers where tobacco can be sold, and 3) introduction of a “smokefree generation” by disallowing the sale of smoked tobacco products to people born on or after a certain date. Other supporting interventions included increased resources for mass media campaigns to promote smoking cessation and discourage smoking uptake in young people.⁶

In January 2023, smokefree legislation came into force that would implement the three world-first measures in the action plan (the Smokefree Environments and Regulated Products [Smoked Tobacco] Amendment Act [SERPA Act]). The reduction in retailers was due to be introduced in July 2024, mandated VLNCs in April 2025 and the smokefree generation in January 2027.⁷ However, in February 2024, New Zealand’s new Government partially repealed the smokefree legislation, stopping the implementation of all three of these measures.

The International Tobacco Control (ITC) Policy Evaluation Project is an international cohort study, conducted in over 30 countries. It aims to measure the impacts of public health policies to reduce the adverse impacts of smoking.⁸ The purpose of this study is to use data from the New Zealand arm of the study to understand 1) the degree of support for the Smokefree Action Plan measures among people who smoke or recently quit smoking, and 2) anticipated responses to the introduction of a retailer reduction and VLNCs among people who smoke. This information is important for establishing the degree of acceptability of the measures and estimating their likely impacts.

Methods

Study design, sampling and recruitment

Data were analysed from Waves 3 and 3.5 of the New Zealand arm of the ITC study (also known as EASE: Evidence for Achieving Smokefree 2025 Equitably). This is an ongoing prospective cohort and repeat cross-sectional study that surveys people who currently smoke or quit smoking within the last 2 years.⁸ Survey waves are conducted every 12–18 months, and participants lost to follow-up are replenished by new participants.

Participants are eligible to take part if aged ≥18, living in Aotearoa New Zealand, and:

• currently smoke cigarettes or tobacco at least monthly, and have smoked at least 100 cigarettes in their lifetime, or
• previously smoked at least monthly, have smoked at least 100 cigarettes in their lifetime and quit smoking within the past 24 months.

Wave 3 was conducted online from 8 November to 24 December in 2020 and from 1 February to 27 February in 2021. It included participants from Wave 2 who agreed to participate in follow-up surveys and replenishment participants recruited through an online panel and social media. The sampling scheme was designed to ensure adequate statistical precision and explanatory power for priority population groups, aiming to recruit equal numbers of Māori, Pacific peoples and Non-Māori-Non-Pacific participants (i.e., 533 participants in each group), and 400 participants aged 18–24 years. We undertook active recruitment targeting these groups through posts on the University of Otago Pacific Islands Centre Facebook page, two local/community Facebook groups in areas with large Māori and Pacific populations (Porirua and South Auckland) and targeted paid social media advertisements.

Wave 3.5 was an interim online survey with a shorter questionnaire, conducted online from 8 June to 26 July in 2021. We only invited participants from Wave 3 for this survey, with no replenishment of participants.

Both surveys were implemented by research company Research New Zealand. Full details of the sampling and survey methods are available in the ITC Technical Report.⁸

Data collection and measures

Measures of ethnicity, age, gender, evidence of financial hardship and smoking status were collected. Ethnicity questions were based on the New Zealand Census questions. Smoking status was defined as a “person who smokes daily not intending to quit”, “person who smokes daily intending to quit”, “person who smokes less than daily but at least monthly” or “person who recently quit smoking”. Wording of the survey questions relating to smoking status and financial hardship is shown in the Table 1 legend.

Wording of questions assessing support for policy measures and expected behaviours if policies are introduced are given in Textbox 1 and the results Tables. The question about support for a smokefree generation was included only in Wave 3.5; all other questions are reported from the Wave 3 survey.

View Textbox 1, Table 1–3, Figure 1.

Data analysis

Data analysis was conducted in R 4.1 (R Institute, Vienna, Austria), using the survey package⁹ to conduct analyses on weighted data, accounting for complex survey design. Weighting was conducted using raked weight calculations drawing on ethnicity, gender, age group and region, with weights calibrated based on population estimates from the New Zealand Health Survey (for survey years 2018–2019 and 2019–2020, combined). These weights permitted estimates to be applicable to the Aotearoa New Zealand population of people who smoke or who have recently quit smoking.

We report prevalence of outcome measures for key demographic and smoking-related sub-groups as weighted percentages with 95% confidence intervals (95% CI). Missing and refused answers were excluded. We estimated support for measures both excluding and including “Don’t know” answers. In the results section we present support results with “Don’t know” answers excluded, as this directly addresses the question of support from participants who expressed an opinion about support or opposition to the smokefree measures. The corresponding analyses of support measures including “Don’t know” as a valid response option are presented in Figure 1 and the Appendices. Anticipated actions are presented with “Don’t know” responses excluded.

To compare groups, we present marginally standardised percentages and absolute differences (with 95% CI) that adjust for potential confounding from the following covariates:¹⁰ smoking status and quit intention, prioritised ethnicity, gender, age group and financial hardship. Marginal standardisation and differences for multinomial outcomes (more than two levels) were conducted in Stata 17 (Statacorp, College Station, TX).

Prioritised ethnicity was used for weighting (participants classified as: Māori [including people who also identified as Pacific peoples], Pacific peoples [excluding people who also identified as Māori] or Non-Māori-Non-Pacific). However, the results are reported using a modified total response ethnicity approach¹¹ to report estimates for Māori and Pacific peoples (relative to an exclusive non-Māori/non-Pacific category). This is to ensure appropriate representation of Māori and Pacific participants. For the analysis using modified total ethnicity, groups included Māori (including people who also identified as Pacific peoples), Pacific peoples (including people who also identified as Māori) or Non-Māori-Non-Pacific (people who do not identify as Māori or Pacific peoples). For reporting of patterning by ethnicity, two separate analyses were run to produce estimates for total Māori (relative to the mutually exclusive Non-Māori-Non-Pacific group) and for total Pacific peoples (relative to the mutually exclusive Non-Māori-Non-Pacific group).

Data for participants reporting “Other” for their gender were excluded from marginally adjusted estimates and differences, as there was an insufficient number to allow for inclusion as a category in the multivariable models (n=18 at W3, n=5 at W3.5).

Ethics

Ethical approval was obtained prior to participant recruitment from the University of Otago Human Ethics Committee (20/020) and University of Waterloo Research Ethics Board (REB #42549). All participants provided consent for participating in the surveys.

Results

Survey participants

Participant characteristics are shown in Table 1 (unweighted percentages to describe the participant profile). In Wave 3, there were 1,230 participants; 80.7% currently smoked and 19.3% had recently quit smoking. In Wave 3.5, there were 615 participants (50% retention from W3); 64.1% currently smoked and 35.9% had recently quit.

Support for the repealed measures

Support for each of the measures is summarised in Figure 1, including values for when “Don’t know” answers were included.

Support for the mandated VLNC policy, among those who expressed support or opposition, was 75.0% (Figure 1, Table 2). Support for retailer reduction was 35.2% and support for a smokefree generation was 82.7% (Figure 1, Table 2). When “Don’t know” answers were included, support was lower, particularly for VLNCs at 60.5% (Figure 1).

Analyses that excluded “Don’t know” values are presented in Table 2. For analyses with “Don’t know” answers included, please see Appendix Table 4.

As outlined in Table 2, all three measures had significantly greater support from people who recently quit smoking (compared to people who currently smoked) and from people who smoked less than daily (compared to people who smoked daily and intended to quit). There was lower support for VLNCs and retailer reduction among Māori compared with Non-Māori-Non-Pacific.

Support for mandated VLNCs and a smokefree generation was lower among people aged 18–24 compared to those aged ≥45; however, a substantial majority supported both these measures in all three age groups. People aged 25–44 were also less likely to support a smokefree generation compared to people aged ≥45. People aged 25–44 were more likely to support a retailer reduction compared with people aged ≥45.

Support for a smokefree generation was higher in females compared to males (absolute marginal difference [aMD] 8.4%, Appendix Table 1). There was no clear evidence of any other differences in support for the three policy measures by ethnicity, age, gender or financial hardship. (Table 2 and Appendix Table 1).

Support for increased media campaign spending and Aotearoa New Zealand’s Smokefree goal

Support for increased media campaign spending to reduce youth uptake of cigarette smoking was 92.6% among those who expressed support or opposition (Figure 1). Support for media campaign spending to encourage smoking cessation was 69.0% (Figure 1). For detailed analyses by smoking status, ethnicity, age, gender or evidence of financial hardship see Appendix Tables 2, 3 and 5.

Overall support for the Smokefree Aotearoa goal of less than 5% daily smoking prevalence by 2025 was 56.7% (95% confidence interval 52.8, 60.5) when “Don’t know” answers were excluded. When “Don’t know” answers were included, support was 52.1% (95% confidence interval 48.3, 55.9, “Don’t know” was 8.1%, oppose was 39.8%). For details, including support by smoking status, ethnicity, age, gender and evidence of financial hardship see Appendix Tables 2, 3 and 5.

Anticipated response to mandated VLNCs and retailer reduction

As outlined in Table 3, in response to the introduction of mandated VLNCs, 18.4% of people who smoke thought that they would reduce the amount they smoked, 13.0% thought they would quit smoking entirely and 14.3% thought they would switch to vaping.

In response to a reduction in retailer availability, 21.6% of people who currently smoke thought they would reduce the amount they smoked, 12.3% thought they would quit smoking entirely and 12.9% thought they would switch to vaping.

As demonstrated in Table 3, most of the responses varied by smoking status. For both measures, people who smoked daily and did not intend to quit smoking were more likely to report that they would “carry on smoking like I do now” than people who smoked daily and intended to quit and people who smoked less than daily.

In response to the introduction of VLNCs, men (compared to women) and people aged 18–24 years (compared to 45 and over) were more likely to report that they would “carry on smoking like I do now, but find a way to get the cigarettes or tobacco I want to smoke.” People aged 18–24 and people aged 25–44 were less likely to “carry on smoking like I do now, with the cigarettes or tobacco that were available” compared to people aged 45 and over. People aged 25–44 were also more likely to
“reduce the amount I smoke” compared to people aged 45 and over. Values are available in Appendix Tables 6 and 7.

In response to a retailer reduction, people aged 18–24 and people aged 25–44 years were less likely to “quit smoking entirely” compared to people aged 45 and over. People aged 25–44 were more likely to “reduce the amount I smoke” compared to people aged 45 and over. People with evidence of financial hardship were less likely than those not in financial hardship to report that they would “quit smoking entirely.” Values are available in Appendix Tables 8 and 9.

There were no major differences in anticipated responses by ethnicity (Appendix Tables 6 and 8).

Discussion

Among people who smoke or recently quit smoking there was strong support for mandated VLNCs (75%) and smokefree generation (83%) policies, as well as increased mass media expenditure. Support for a retailer reduction was the only measure with less than majority support (35%). People who smoked, particularly daily smokers without intent to quit, consistently demonstrated less support for the measures than people who had recently quit smoking. The findings were broadly in line with earlier findings from Wave 2 of the ITC NZ (EASE) Survey (conducted 2016–2017).¹²

The relatively low level of support for a retailer reduction aligns with other Aotearoa New Zealand-based studies.^12–15 A previous qualitative study among people who smoke found concern that a reduction in retailers could increase tobacco product prices, elevate stress due to changes in routine and reduce viability for local businesses.¹⁴

Around 50% of participants who smoked anticipated that they would reduce the amount they smoke, quit smoking completely or switch to vaping if either mandated VLNCs or substantial retailer reductions were introduced. The proportion anticipating these behaviour changes in response to mandated VLNCs (46%) was much greater than the proportion who stated they would try and obtain tobacco products they wanted to smoke (19%), presumably homegrown or illicit cigarettes or tobacco. These findings highlight that 1) many people who smoke anticipate the introduction of a retailer reduction or VLNCs would have helped them to reduce the amount they smoke or stop smoking, and 2) in contrast to arguments that these measures are likely to greatly increase the illicit market,¹⁶ only a minority of participants reported they would consider taking steps to obtain cigarettes with a higher nicotine content should VLNCs be introduced. However, the finding that some people who smoke would be likely to seek out illicit cigarettes or tobacco suggests that if a mandated VLNC policy is introduced additional actions to combat illicit trade should be introduced (such as increased resources for customs), as was planned in the Smokefree Aotearoa 2025 Action Plan.⁶

Importantly, there was no significant difference in anticipated responses to a retailer reduction or mandated VLNCs by ethnicity. Our findings are consistent with a recent study of anticipated responses to these measures in over 700 Māori who smoke.¹⁷ They also align with modelling that suggests the interventions could significantly reduce smoking prevalence for Māori and Pacific peoples.^6,18,19

A key strength of this study is that it provides results that are directly relevant to the three SERPA Act measures that were recently repealed, drawn from the people most affected by smoking. The sample is sufficiently large to provide robust indications of support and anticipated changes in response to the measures. It also allows us to evaluate differences in support by smoking status, intent to quit smoking and ethnicity.

Another strength is our presentation of data on support for the measures with and without “Don’t know” answers. This allows results to be compared to other studies assessing support for smokefree measures, regardless of whether they opt to include or exclude “Don’t know” answers.^12,20–25 Levels of support for the measures were largely similar, regardless of whether “Don’t know” answers were included or excluded, as “Don’t know” responses were rare. However, support for mandated VLNCs dropped substantially when “Don’t know” responses were included, reflecting the high percentage of “Don’t know” responses (19%). The high level of “Don’t know” responses for this policy likely reflects the unfamiliarity of VLNCs among people who smoke, as they have not been available in New Zealand. Of note, international studies have found at least 50% support for VLNC policies among participants in trials who had used VLNCs for several weeks.^20,21 Our findings emphasise the importance of assessing understanding of proposed policy measures and the need for public education.

One limitation of this study is that the recruitment target for Pacific participants was not reached, meaning that results for this group are less precise than for Māori or Non-Māori-Non-Pacific respondents.

Another limitation is that the study data were collected before the SERPA Act changes to include the three action plan measures were passed and subsequently repealed in early 2024. It is possible that responses to the survey questions may have changed in response to these events. At the time of writing, we are in the process of analysing data from Wave 4 (conducted in 2022) and recruiting for Wave 5 (September 2024), which will provide further insights. However, the results from Waves 3 and 3.5 align with findings from a population-based survey conducted in late 2023 in response to the news that the Government intended to repeal the three smokefree measures. Of those surveyed, support for retention of the three key measures was 68% for reduction in retailer numbers, 77% for mandated VLNCs and 65% for a smokefree generation.²⁶

Our findings call into question the Government’s decision to repeal the 2023 SERPA Act measures to reduce retailer numbers, mandate VLNCs and introduce a smokefree generation. The introduction of VLNCs and a smokefree generation were strongly supported by people who smoke or who have recently quit smoking, and retailer reductions by a majority of people who had recently quit. Anticipated responses to a reduction in retail numbers and VLNCs indicated that these measures had the potential to reduce smoking prevalence substantially and equitably.

View Appendices.

An erratum has been published for this article.