Article Text

Download PDFPDF

Relation between national-level tobacco control policies and individual-level voluntary home smoking bans in Europe
  1. Amy K Ferketich1,
  2. Alessandra Lugo2,
  3. Carlo La Vecchia2,
  4. Esteve Fernandez3,4,5,
  5. Paolo Boffetta6,
  6. Luke Clancy7,
  7. Silvano Gallus8
  1. 1The Ohio State University College of Public Health, Columbus, Ohio, USA
  2. 2Department of Clinical Sciences and Community Health, University of Milan, Milan, Italy
  3. 3Tobacco Control Unit, Cancer Prevention and Control Program, Institut Català d'Oncologia—ICO, L'Hospitalet de Llobregat, Barcelona, Spain
  4. 4Cancer Control and Prevention Group, Institut d'Investigació Biomèdica de Bellvitge-IDIBELL, L'Hospitalet de Llobregat, Barcelona, Spain
  5. 5Department of Clinical Sciences, Universitat de Barcelona, Barcelona, Spain
  6. 6Institute for Translational Epidemiology and Tisch Cancer Institute, Ichan School of Medicine at Mount Sinai, New York, New York, USA
  7. 7TobaccoFree Research Institute Ireland, Dublin, Ireland
  8. 8IRCCS—Istituto di Ricerche Farmacologiche ‘Mario Negri’, Department of Epidemiology, Milan, Italy
  1. Correspondence to Dr Amy K Ferketich, Division of Epidemiology, The Ohio State University College of Public Health, 310 Cunz Hall, 1841 Neil Avenue, Columbus, OH 43210, USA; ferketich.1{at}osu.edu

Abstract

Background Little is known about the relationship between national tobacco control policies and implementation of private home smoking bans.

Objective To determine the relationship between the Tobacco Control Scale (TCS), a score measuring national-level strength of tobacco control policies, and the prevalence of in-home smoking bans and beliefs on other tobacco control policies, among the Member States (MS) of the European Union (EU) that participated in the Pricing Policy And Control of Tobacco in Europe (PPACTE) project.

Methods A face-to-face representative survey, based on 18 056 individuals aged ≥15 years, from 18 European countries—including 16 EU MS—was conducted in 2010. Multilevel logistic regression models were fit to examine the relationship between the TCS score and in-home smoking ban prevalence and beliefs that other policy approaches are useful.

Results In 2010, the TCS scores ranged from 32 in Austria and Greece to 77 in England. The TCS score correlated with the prevalence of in-home smoking bans (rsp=0.65). A 10-unit increase in the TCS score significantly increased the odds of in-home smoking ban (OR=1.33; 95% CI 1.01 to 1.76). The odds of believing that providing cessation services (OR=1.21), raising prices (OR=1.01) and extending bans is useful (OR=0.93) were not significant.

Conclusions Government tobacco control policies are positively related to the individual-level tobacco policy of having an in-home smoking ban.

  • Public opinion
  • Public policy
  • Denormalization
View Full Text

Statistics from Altmetric.com

Introduction

Individual tobacco control policies have been linked to a number of positive outcomes. Clean indoor air laws are associated with an improvement in indoor air quality,1 a decrease in the cotinine concentration among non-smoking employees,2 ,3 fewer pulmonary symptoms among non-smoking employees,1 a reduction in the rate of acute myocardial infarction,4–7 a decrease in the population smoking prevalence,8 and an increased success in smoking cessation.9 ,10 Moreover, raising tobacco taxes, banning advertising of tobacco and enforcing age restrictions on tobacco have all been associated with reductions in smoking prevalence.11–13

Tobacco control policies change the social norms of groups of individuals.14 Strong policies denormalise smoking, resulting in a lower prevalence of smoking.15–17 An even stronger approach, however, is for a nation to have a set of policies focused on strengthening the tobacco control environment. This was the goal of the WHO's Framework Convention on Tobacco Control (FCTC), which urges national governments to develop, implement and enforce a set of measures aimed at increasing tobacco control.18 Included in the list of options were higher cigarette taxes, bans on sales to minors, bans on advertising, passage of clean indoor air laws, procedures to ban illegal trade of cigarettes and public education campaigns. While 178 parties have signed and ratified the WHO FCTC, not all have implemented the entire set of provisions.19

The Tobacco Control Scale (TCS) score was designed to quantify the strength of tobacco control policies of the European Union (EU) Member States (MS).20 The TCS score has been associated with attitudes towards policies, smoking prevalence and concerns about second-hand smoke in other studies.21 ,22 Little, however, is known about the relationship between national-level tobacco control policies and implementation of private home smoking bans. To the best our knowledge, only one ecological study evaluated the issue.23 Analysing aggregate data on a 2009 Eurobarometer, no relation was found between the TCS score and the prevalence of voluntary smoking in private venues (houses and cars).23

Private places, such as cars and homes, are rarely included in comprehensive tobacco control policies. In the USA, some municipalities have banned smoking in cars when children are present and others have banned smoking in multiunit housing systems.24 ,25 In-home smoke exposure affects children most, since they spend most of their time in the home compared to other places.26 Thus, children benefit from in-home smoking bans. However, numerous studies have demonstrated a strong association between in-home smoking bans and smoking cessation among adults. Mills et al27 performed an extensive review of the literature and concluded that in-home smoking bans encourage smokers to make a quit attempt and remain abstinent.

The data for this study were obtained from the EU MS that participated in the Pricing Policy And Control of Tobacco in Europe (PPACTE) project.28 ,29 The study objective was to determine the relationship between the TCS score and the prevalence of in-home smoking bans and beliefs on other tobacco control policies using cross-sectional data from a pan-European survey.

Methods

Data source

The PPACTE project was designed to evaluate the effectiveness of tobacco pricing policy in Europe.28 ,29 In 2010, computer-assisted personal interviews were conducted in 18 European countries (see table 1) by Doxa—the Italian branch of the Worldwide Independent Network/Gallup International Association (WIN/GIA). In each country, Doxa enrolled a sample of around 1000 participants, representative of the country-specific general population aged 15 years and older with respect to age, gender, geographic area and socioeconomic status. In total, 18 056 individuals completed the questionnaire that included items on demographics (age, gender, education level), tobacco use behaviour (ever use and current use, age at initiation) and evaluation of tobacco control policies.

Table 1

Prevalence (%) of home smoking bans and attitudes towards tobacco control policies, overall and by country

In several countries (Albania, Croatia, Hungary, Italy, Poland and Romania) a multistage methodology was used. In the first stage, the primary unit of selection was a geographic area or voting centre. In the second stage, households or municipalities were selected. In the last stage, respondents were chosen randomly, in order to be representative of the population in terms of sex, age, geographic area and socioeconomic characteristics. In those countries where adult respondents had been selected from electoral rolls, a quota method was used to select respondents aged 15–17. For other countries (Austria, England, Finland, France and Ireland) a quota method was used for the selection of the entire sample, stratifying the population according to selected variables including age, sex, and alternatively geographic area and/or occupation. For the other countries (Bulgaria, the Czech Republic, Greece and Latvia), a stratified random method or a simple random method was used. In the data processing phase, statistical weights were generated to balance the sample of respondents to assure representativeness of the corresponding national populations in terms of age, sex and socioeconomic characteristics. Full details on the survey methodology and participation rates, as well as individual-level characteristics, are reported elsewhere.28 ,29 A copy of the English version of the questionnaire is available online.29

Measures

The primary dependent variable was the presence of a complete in-home smoking ban. The PPACTE questionnaire included the question, “At your home, where can people (including anyone living in the household and guests) smoke?” The response options included, “everywhere; in some specific indoor areas (eg, in the kitchen, in the bathroom); only outside.” The ‘only outside’ option indicated that there was a complete in-home smoking ban. The other dependent variables of interest were the extent to which other tobacco control efforts were thought to be useful. Respondents were asked, “To control and limit tobacco use, the government or the national political decision makers could adopt several strategies. How useful do you assess each one?...Free psychological or pharmacological support for smoking cessation, including nicotine replacement therapy (patches, gums, etc), bupropion and varenicline; raising the price of cigarettes; extending smoking bans.”

The TCS score, a score measuring the strength of tobacco control policies at a country level in all the EU MS, was first described by Joossens and Raw in 2006,20 and then updated in 2007,30 in 201031 and most recently, in 2013.32 The TCS score ranges from 0 (low implementation of tobacco control strategies) to 100 (high implementation), according to the extent to which six tobacco control interventions have been adopted by a country. These interventions, identified by the World Bank,33 include the following six components: (1) tobacco price increases through taxation; (2) clean indoor air laws; (3) public information campaigns; (4) bans on advertising and marketing of tobacco products; (5) warning labels on tobacco products; and (6) availability of tobacco dependence treatments. The 2010 TCS score is based on legislation present on 1 January 2011, tobacco prices on 1 July 2010 and the tobacco control budget in 2009.31

Statistical analysis

All of the analyses used weighted data to account for the fact that the within country surveys were complex probability surveys, and to generate estimates representative of various country populations. Moreover, to calculate results for the whole sample of countries, we applied a weighting factor, with each country contributing in proportion to its population aged 15 years or above, obtained by Eurostat.34

Descriptive statistics were calculated to examine the 2010 version of the TCS score by country, as well as the prevalence of in-home smoking bans, and beliefs about the usefulness of providing cessation assistance, raising cigarette prices, and extending smoking bans. Correlations were performed by means of Spearman rank-correlation coefficients (rsp).

We next examined the presence of in-home smoking bans and beliefs about the usefulness of these policy approaches by gender, age, education level and smoking status of the respondent. Multilevel logistic regression models, after adjustment for sex, age, level of education and smoking status (fixed effects) and country (random effect) were fit to examine the relationship between a 10-unit change TCS score and the presence of in-home smoking bans and the extent to which each potential policy approach was believed to be useful to control tobacco use. Whenever the multilevel analysis did not converge, an unconditional logistic regression model was fit, including the covariates sex, age, level of education and smoking status. ORs and the corresponding 95% CIs were estimated from these models. The same multilevel models were fit to derive ORs for complete home-smoking bans, according to the strength of the implementation of tobacco control policies at a national level, using the median split of the 2010 TCS score and, in turn, its six components.

No TCS score is available for Albania and Croatia, since they were not EU MS in 2010. Albania and Croatia were therefore included in the descriptive statistics but excluded from the modelling. Thus, the correlations and models were based only on 16 European countries.

Results

Item-level missing data ranged from 1.1% (in-home smoking ban) to 5.8% (support for smoking cessation). Thus, of the 18 056 individuals aged 15 years and older who completed a questionnaire, the analyses were based on a range from 17 010 to 17 865 individuals.

Table 1 contains descriptive information about the 2010 version of the TCS score, the prevalence estimates of in-home smoking bans and the extent to which the three policy options are thought to be useful to control tobacco use. The TCS scores ranged from a low of 32 in Austria and Greece to a high of 77 in England. On average, 62.2% of respondents indicated that their home had a complete smoking ban; however, there was variability across countries, from a low of 30.6% in Croatia to a high of 93.2% in Finland. Similarly, there was a considerable variability between countries in the extent to which individuals believed that the three tobacco control policies were useful. Overall, 73.4% of individuals believed that offering cessation support was helpful, but only 35.7% of individuals in Hungary held that belief. In contrast, 92% of respondents in Portugal held that belief. Just over half of respondents believed that the other two policy approaches would be useful. Raising cigarette prices was thought to be useful by 55% overall, with a low of 18.9% in Hungary to a high of 69% in Albania. Extending smoking bans was thought to be useful by 56.5% of all participants, ranging from 37% in Latvia and 83.4% in Albania. A significant correlation was found between the TCS score and the prevalence of in-home smoking bans (rsp=0.648; p=0.007; figure 1). After excluding, in turn, potential outliers, such as the UK, Finland and Greece, the correlation coefficients did not substantially change and the p values remained significant. On the contrary, none of the correlations between TCS score and the belief that the three policy approaches are useful reached statistical significance.

Figure 1

Prevalence (%) of complete in-home smoking ban and Tobacco Control Scale (with Spearman correlation coefficient), in 16 European countries. Pricing Policy And Control of Tobacco in Europe, 2010.

There were differences in the presence of in-home smoking bans and the belief that the three policy approaches were useful (table 2). In general, women, older individuals, those with a greater level of education and non-smokers were more likely to live in a home with a smoking ban than their counterparts (p<0.001). Beliefs about the usefulness of the three policy approaches were not different by age, gender or education level; however, non-smokers had a higher prevalence of endorsing the usefulness of these three approaches compared to smokers.

Table 2

Prevalence (%)* of individuals aged ≥15 years in 18 European countries with complete home smoking ban and that consider useful (quite to very useful) different tobacco control policies, according to different individual-level characteristics

Table 3 presents the results from the multilevel logistic regression analyses with the TCS score (continuous variable; increment of 10-unit) as the primary predictor and in-home smoking bans and belief that the three policy approaches are useful as the four outcomes. After controlling for age, gender, education level, and smoking status of the participants, and taking into account clustering within countries, there was a significant association between TCS score and in-home smoking bans. A 10-unit increase in the TCS score significantly increased the odds of having an in-home smoking ban (OR=1.33; 95% CI 1.01 to 1.76). The odds of believing that providing cessation services (OR=1.21; 95% CI 0.95 to 1.55), raising prices (OR=1.01; 95% CI 0.99 to 1.04) and extending bans is useful (OR=0.93; 95% CI 0.78 to 1.10) were not significant.

Table 3

Adjusted regression* coefficients, with corresponding ORs and 95% CI, for TCS score predicting each belief about the policy approach

The proportion of voluntary in-home smoking bans was more frequent in countries with stronger tobacco control policies (OR=2.28; 95% CI 1.20 to 4.32). Compared to countries with a weak tobacco control environment, those with stronger TCS components had systematically more smoke-free homes (table 4). The corresponding ORs were not significant, however, for advertising bans and availability of tobacco dependent treatments.

Table 4

ORs and 95% CIs for complete versus non-complete home-smoking bans, according to the strength of the implementation of tobacco control policies at a national level, using the median split of the 2010 TCS score and its six components

Discussion

The TCS score was correlated with the prevalence of in-home smoking bans using data from 16 EU MS that participated in the PPACTE project. This result held in the multilevel regression model that controlled for demographic variables and smoking status. Moreover, the result was consistent whether the TCS score was analysed as a continuous variable or dichotomised at the median. Importantly, four of the individual contributions to the TCS scale (pricing, public campaigns, smoking bans and health warnings) were significantly related to in-home smoking bans. Thus, the result is not disproportionately influenced by the one component related to clean indoor air laws. This finding supports previous cross-sectional studies that found a higher prevalence of in-home smoking bans in jurisdictions with clean indoor air laws.35–37 In a prospective study, Mons et al38 found that the prevalence of in-home smoking bans increased in European countries after clean indoor air laws were enacted, whereas there was no change overtime in the absence of a clean indoor air law. The TCS scale includes more information about the tobacco control environment than simply the presence of clean indoor air laws; thus, there is now evidence that a strong set of government policies proposed by the WHO FCTC, which may indirectly denormalise tobacco use, is related to an individual-level policy. However, we must exert caution in concluding any causal relationship between the two measures because our analyses were ecological or based on cross-sectional data.

Fewer than two of three respondents, across all countries, reported living in a home that is smoke free. Not surprising, fewer smokers compared to non-smokers lived in homes with smoking bans, but still the prevalence among non-smokers was less than three of four. The strong correlation between the TCS score and the prevalence of in-home smoking bans suggests that as European countries adopt more tobacco control policies and denormalise tobacco use, more homes will be smoke free in the future. It is also possible that there is an effect if neighbouring countries adopt stronger tobacco control policies. Hovell et al39 examined the influence of socially reinforcing contingencies on adoption of in-home smoking bans among Mexican adults living in the USA (San Diego) and Mexico (Tijuana and Guadalajara). They found that while the prevalence of smoking bans was highest in San Diego (91%), residents of Tijuana, which is just over the border with California, were significantly more likely to have an in-home smoking ban (66%) than residents of Guadalajara (39%), which is hundreds of miles away from the border. They concluded therefore that a strong tobacco control environment in one country may transfer to another through shared social factors. However, as the distance between the social networks increases, the less impact the policies in one country may have on the other.

In neither the ecological nor multilevel analyses did we observe an association between the TCS score and beliefs that providing cessation services, raising cigarette prices or extending smoking bans would be useful to reduce smoking. It is possible that people who live in countries that have strong tobacco control policies in place believe that it is not possible to extend them.

The TCS score has been associated with attitudes towards policies, smoking prevalence and concern about secondhand smoke in other studies.21 ,22 Martinez-Sanchez et al reported that the TCS score was positively correlated with attitudes towards smoking bans in restaurants and bars and negatively correlated with smoking prevalence in 27 European countries.22 In contrast to our findings, they did not find a significant relationship between the TCS score and self-reported exposure to in-home smoking. Possibly the difference was due to how in-home smoke exposure was measured in their study; individuals were asked if someone they lived with smoked inside the home. In PPACTE, participants were asked about in-home smoking rules, which would include smoking by household residents as well as guests. Willemsen et al22 reported a strong correlation between the TCS score and concern about secondhand smoke exposure among 27 European EU MS. Their finding, along with ours, suggest a pathway whereby strong tobacco control policies raise the awareness of the harmful effects of secondhand smoke, which then prompt individuals to implement in-home smoking bans. A community-level intervention study in Egypt provides some support for this hypothesis.40 That study found that communities that received an educational campaign addressing the harmful effects of secondhand smoke were more knowledgeable and were more likely than control communities to actively reduce their secondhand smoke exposure using various means, including implementing in-home smoking bans.

The limitations of this study are similar to that of most survey-based studies, which include self-reported smoking status, a simple assessment of attitudes towards policy approaches, the potential for over-reporting of in-home smoking bans due to social desirability, and bias due to non-response. Another limitation is the ecological analysis we performed to examine the correlation between the TCS score and the prevalence of in-home smoking bans; such analyses are susceptible to the ecological fallacy. However, results from the multilevel analysis add strength to this study because we could confirm the ecological finding in a sample of individuals in the same countries.

In conclusion, these results provide convincing evidence that a strong set of government policies, that indirectly denormalises tobacco use, is related to an individual-level tobacco policy. Thus, environments that severely restrict smoking in public do not appear to have the unintended consequence of increasing smoking in the home environment. In fact, these results, in combination with other studies on the topic35 ,41 suggest that a strong tobacco control environment influences an individual's decision about banning smoking in the home.

What this paper adds

What is already known on this subject

  • Prior research suggested that clean indoor air laws were associated with in-home smoking bans.

What important gaps in knowledge exist on this topic

  • We did not know whether other national tobacco control policies were associated with in-home smoking bans.

What this study adds

  • In this study of 16 European Union Member States, we found that strong national-level, comprehensive tobacco control policies, measured by the Tobacco Control Scale (TCS), are associated with in-home smoking bans.

References

View Abstract

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:

Footnotes

  • Contributors SG and AKF had the original idea of the work. SG, EF, PB and CLV gave substantial contribution to the conception and design of the work. SG provided the survey data. AL conducted the data analysis. AKF drafted the work in collaboration with SG. All other authors gave substantial contributions to the interpretation of data for the work, and critically revised the draft manuscript for important intellectual content. All the authors approved the final version of the manuscript, and agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

  • Funding The project ‘Pricing Policies and Control of Tobacco in Europe (PPACTE)’ was funded by the European Commission Seventh Framework Programme Grant Agreement HEALTH-F2- 2009-223323. The work of SG, AL and CLV is partially supported by the Italian Foundation for Cancer Research (FIRC) and the Italian League Against Cancer (LILT), Milan. The content is solely the responsibility of the authors and does not necessarily represent the official views of their institutions.

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval The study protocol of the PPACTE survey was approved by the Institutional Review Board of the Istituto di Ricerche Farmacologiche Mario Negri. The procedures for recruitment of subjects, informed consent, data collection, storage and protection (based on anonymous identification code) were all in accordance with the current country specific legislations.

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

  • Data sharing statement Individuals should contact Dr Silvano Gallus to discuss use of the data reported in this study.

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.