Tobacco control policies and socio-economic inequalities in smoking cessation Bosdriesz, J.R.

Transcription

1 UvA-DARE (Digital Academic Repository) Tobacco control policies and socio-economic inequalities in smoking cessation Bosdriesz, J.R. Link to publication Citation for published version (APA): Bosdriesz, J. R. (2017). Tobacco control policies and socio-economic inequalities in smoking cessation: Evaluating natural experiments General rights It is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons). Disclaimer/Complaints regulations If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Ask the Library: or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. You will be contacted as soon as possible. UvA-DARE is a service provided by the library of the University of Amsterdam ( Download date: 19 Oct 2018

2 9 Tobacco control policy and socioeconomic inequalities in smoking in 27 European countries Jizzo Bosdriesz Karien Stronks Marc Willemsen Anton Kunst Published in Drug and Alcohol Dependence 2016;165:79-86

3 Chapter 9 ABSTRACT Background Higher Tobacco Control Scale (TCS) scores, in the early 2000s were associated with higher smoking cessation rates across Europe, both among lower and higher educated people. We aimed to assess if this association held in recent years. Methods Repeated cross-sectional Eurobarometer surveys were used, in 27 European countries from (study sample = 73,617 adults). We used multilevel regression to model associations between the TCS (ranging from 0 to 100, quantifying: tobacco price, smoke-free laws, mass-media campaigns, advertising bans, warning labels, and cessation support) and both smoking cessation and cigarettes smoked per day. We modelled associations according to respondents education and occupation, with adjustment for age, sex, and survey-wave. Results We found no association between the TCS and smoking cessation for lower or middle educated respondents, but we did find an association for higher educated respondents (OR: 1.13, 95%CI: ). For smoking intensity, we observed no associations with the TCS for lower educated respondents (beta: 0.04, 95%CI: ) but we did observe significant associations for middle (beta: -0.25, 95%CI: ) and higher educated respondents (beta: -0.27, 95%CI: ). Associations were observed for both manual and non-manual classes, but not for those not working for pay. Of the TCS domains, none were associated with smoking cessation for lower educated respondents, but five were for higher educated respondents. Conclusions Associations between tobacco control policies and smoking cessation were found mostly among higher socioeconomic groups. This underlines the need for specific tobacco control policies that explicitly focus on reaching low socio-economic groups. 184

4 Tobacco control and inequalities in smoking in Europe INTRODUCTION While it has long been known that 1 in 2 smokers will die from regular smoking, recent studies show it might even be 2 out of 3 smokers. 1 3 Combined with persistently high prevalence rates globally, this means that smoking is a long way from relinquishing its position as a major threat to global health. Many steps have been taken to combat this deadly addiction, the ratification of the Framework Convention on Tobacco Control (FCTC) being the foremost achievement. 4 The FCTC provides minimum requirements on a broad range of evidence-based tobacco control measures for signatory countries to implement. Many European countries have been among the first to implement evidence-based tobacco control policies. 5 However, although these policies have contributed to reducing smoking rates in the European population as a whole, 6 there have been inequalities in the decline of trends in smoking. 7,8 Inequalities by socio-economic status (SES) are especially worrying, as the smoking prevalence among those with a low SES is highest and is declining less rapidly, compared those with a higher SES. In a previous study, we found that whilst trends in the low and high SES groups were similar in Europe during the 1980s and early 1990s, there was a strong widening of inequalities from 2000 onwards. 8 This study did not directly test the role of tobacco control policies during this period, but the results suggest that these policies have not been able to reduce the socio-economic inequalities, and they might even have contributed to the widening of inequalities. Schaap et al. tested the association between national level tobacco control policies, measured by the Tobacco Control Scale (TCS) and smoking cessation in 18 European countries in the early 2000s. 9 The TCS is a quantitative measure of the extent to which comprehensive tobacco control policies have developed. This scale contains six domains such as price, smoke-free laws, advertising bans, and cessation support. Schaap et al. found that countries with a higher TCS score had higher quit ratios, and that this association was similar for low and high SES groups. The increasing inequalities during later years suggest that tobacco control may have benefitted higher educated people more than their lower educated counterparts. Many more tobacco control policies have been introduced in these subsequent years. It is therefore important to assess whether the previously found association between tobacco control policy and socio-economic inequalities in smoking in the EU still held for the subsequent 10 years

5 Chapter 9 We aim to assess whether tobacco control policy was associated with socioeconomic inequalities in smoking across the EU in the period 2006 to Because tobacco control policies can decrease consumption of tobacco not only by increasing smoking cessation, but also by decreasing smoking intensity (the number of cigarettes smoked per day), we will examine these two outcomes separately. In addition to associations with the total tobacco control score, we will also evaluate six components of tobacco control policy (e.g. price of tobacco). To do this, we will use recent data on smoking in all 27 EU countries from one single survey, and we will apply a multilevel design to assess the relationship of smoking with the TCS. METHODS Population The study population consists of respondents from the Eurobarometer surveys that were conducted on behalf of the European Commission to monitor public opinions in the Member States on various topics including smoking. We have selected the Eurobarometer survey waves that most closely matched the years for which data on our main predictor, the TCS, were available, 2006, 2009 and The Eurobarometer surveys were administered face-to-face, in the appropriate language. Respondents were selected by random multi-stage probability sampling. The sampling method took into account the geographical and demographic distributions of each population, such that all samples are nationally representative. From each sample cluster, an initial address was randomly selected, and subsequent addressed were selected by following a random route procedure. Each wave of the survey consisted of around 1,000 respondents from each of the (at that time) 27 member states of the EU, which resulted in an initial study sample of 80,623 respondents for all three waves combined. From this, 4,321 respondents younger than 20 were excluded. In addition 1,061 more respondents were excluded because of missing values on smoking status and 1,624 more respondents were excluded because of missing values on educational level. This resulted in a final sample of 73,617 respondents. Variables Smoking status was assessed by the question Which of the following applies to you? Those who answered You smoke packed cigarettes, You smoke roll-up cigarettes, and You smoke cigars or a pipe were classified as current smokers. 186

6 Tobacco control and inequalities in smoking in Europe Those replying You used to smoke but you have stopped were categorised as former smokers, and those replying You have never smoked were never smokers. From these data, we derived the prevalence of current smokers (current smokers as a proportion of respondents) and the quit ratio (former smokers as a proportion of current plus former smokers). The intensity of smoking was measured among current smokers with the question How many cigarettes a day do you smoke?. This number was measured in 9 categories, ranging from 1-4, 5-9 through >40. We have recalculated this variable to a continuous one, taking the mean number per category (3, 7, 12, 17, 22, 27, 32, 37, and 45). To measure SES, we used both educational level and occupational level. Educational level was measured by the question How old were you when you stopped full-time education? The answers were categorized into three groups: 15 years or younger was defined as low education, 16 to 19 years old was defined as middle education and 20 years or older was defined as high education. As we only included those aged 20 years and older, we assigned those still studying to the high education group. The Eurobarometer classified occupational class into 19 categories. We have merged these into three categories: manual (farmers, fishermen, craftsmen, skilled manual workers, servants), non-manual (professionals such as lawyers, accountants or doctors; business owners, middle and general management, desk functions, supervisors), and other (doing housework, students, unemployed, retired). The main predictor was the tobacco control scale (TCS), as developed by Joossens and Raw in This scale ranges from 0 to 100 and quantifies tobacco control policy on six domains: price of tobacco, smoke-free areas, tobacco control budget, bans on tobacco advertising, health warning labels, and cessation support. The scores for these domains have been allocated on the basis of their effectiveness in reducing smoking, estimated from scientific evidence and experts opinions. Tobacco control experts rated their countries performance on each of the policy measures. The TCS has later been calculated for the years 2007, 2010, and However, the point allocation has been changed between these years to include newer types of policies, making the original TCS scores difficult to compare. To make the scores directly comparable, we have recalculated all scores using the 2013 scoring system. 13 We used the raw data for all of the separate measures provided in the 2005, 2007, and 2010 TCS reports to calculate the new scores These detailed scores are shown in Appendix Table 1. For the main analyses, we used the overall score, and for additional analyses we used the separate scores on each of the six domains. For the analyses per domain, we recalibrated the scales, so that they all 187 9

7 Chapter 9 range from 0 to 10. We assigned the TCS scores per country to all respondents from that country, for the given year. The TCS scores of 2005, 2007 and 2010 were allocated to the Eurobarometer respondents of 2006, 2009 and 2012 respectively. Statistical analysis For descriptive purposes, the smoking prevalence rate for the total population and for each country separately was age-standardized by the direct method, 14 using the new EU 27 standard population. 15 In addition, we calculated the education and occupational ratios, by dividing the standardised prevalence rates for the high groups (high education and non-manual occupation) by the outcomes for the low groups (low education and manual occupation respectively). To analyse the association between tobacco control and the two measures of smoking behaviour, we used multilevel regression models. We used three-level regression models, logistic regression for smoking cessation as the outcome variable, and linear regression models for intensity of smoking as the outcome variable. In these models, individual respondents were the first level, country was the second level, and survey wave was the third level. In Model 1, the total TCS score and educational level were included as the main predictors. Model 2 was the same, except that occupational level was a main predictor instead of educational level. In both Model 1 and Model 2, we controlled for age, sex, survey wave, and the country-agespecific prevalence of ever-smoking. The latter variable was included to control for the historic pattern of smoking uptake within the country. To test for differences by SES (Table 4), Model 1 included a term for the interaction between the TCS and education, and Model 2 assessed the interaction between the TCS and occupation. Lastly, for analyses per domain of the TCS, we used the exact same model specification, but with the domain specific scores instead of the total TCS score as predictor. For the logistic regression models, the regression coefficient corresponding to the TCS was transformed into a measure that represents the odds of quitting smoking associated with 10 points increase in the TCS score. For the linear regression, the outcome represented the increase in the number of cigarettes smoked per day associated with 10 points increase in the TCS score. Because the three surveys waves were combined, the coefficient for the TCS represents both comparisons between countries and over time. 188

8 Tobacco control and inequalities in smoking in Europe Table 1 Characteristics of the study population for all three survey waves combined, per country Age Education 1 Occupation 2 Country N % 3 M 4 (sd) L M H M NM O Austria 2, Belgium 2, Bulgaria 2, Cyprus 1, Czech Republic 2, Denmark 2, Estonia 2, Finland 2, France 2, Germany 4, Greece 2, Hungary 2, Ireland 2, Italy 2, Latvia 2, Lithuania 2, Luxembourg 1, Malta 1, Netherlands 2, Poland 2, Portugal 2, Romania 2, Slovakia 2, Slovenia 2, Spain 2, Sweden 2, United Kingdom 3, TOTAL 73, Age in years when finished/stopped full time education: low (L) = 15 years, middle (M) = years, high (H) = 20 years. 2 Occupational classes: manual (M) = farmer/fisherman/manual worker, non-manual (NM) = professional/white collar/executive, other (O) = business owner/ craftsman/retired/housewife/student/military/unemployed. 3 Percent female respondents. 4 Mean. 9 RESULTS Table 1 displays the characteristics of the study population per country. 56.2% of the sample is women; the average age is The low educated group is smaller (21.1%) than the middle (45.5%) and high educated (33.3%) groups. Similarly, the 189

9 Chapter 9 manual occupation group (14.1%) is smaller than the non-manual group (31.5%) and the other/non-employed group (54.5%). Table 2 Smoking behaviour for all three survey waves combined, per country Smoking prevalence Quit ratio 1 Smoking intensity 2 Country Pop. 3 Ed. ratio 4 Occ. ratio 5 Pop. Ed. ratio Occ. ratio Pop. Ed. ratio Occ. ratio Austria Belgium Bulgaria Cyprus Czech Republic Denmark Estonia Finland France Germany Greece Hungary Ireland Italy Latvia Lithuania Luxembourg Malta Netherlands Poland Portugal Romania Slovakia Slovenia Spain Sweden United Kingdom EU TOTAL Quit ratio = former smokers / (former smokers + current smokers). 2 Mean number of cigarettes smoked per day. 3 Pop. = the age-standardized population mean of 2006, 2009 and 2012, for men and women combined. 4 Ed. ratio = the value for the high education group divided by the value for the low education group. 5 Occ. ratio = the value for the non-manual occupation group divided by the value for the manual occupation group. The national age-standardized rates of smoking measures are shown in Table 2. The EU average prevalence of current smoking (i.e. smoking at time of surveying) was 27.5%; the extremes are 15.0% in Sweden and 40.2% in Greece. The educational 190

10 Tobacco control and inequalities in smoking in Europe ratio (relative risk for higher educated respondents compared to lower educated respondents) for smoking prevalence is 0.65 for the whole EU. This ratio is 1.08 in Portugal, which means that smoking is 8% more prevalent among higher educated respondents compared to lower educated respondents, whilst this ratio is below 1 in all other countries. The average quit ratio is 47.0, indicating that 47% of eversmokers in the EU had quit smoking. The national quit ratios ranged from 32.1 in Latvia to 67.3 in Sweden. The educational ratio for the EU on average is 1.38, meaning that the quit ratio is 38% higher among higher educated respondents, compared to lower educated respondents. This ratio is 0.98 in Portugal and 0.97 in Romania, and above 1 for all other countries. Lastly, the EU average number of cigarettes smoked per day (by current smokers) is 16.0, varying from 12.7 in Slovakia to 21.6 in Cyprus. The EU average educational ratio is 0.84, and below 1 in most countries, indicating that the number of cigarettes smoked per day is lower among higher educated respondents. In Figure 1, the TCS is plotted against the quit ratio and smoking intensity. Each country is represented by two dots, one for each SES group. Associations with the TCS are represented by separate lines for lower and higher educated respondents and for respondents with a manual and non-manual occupation. The TCS is positively correlated with the quit ratio, more so among the higher educated respondents than among lower educated respondents. The correlations for both occupation groups are about equally strong. The TCS and intensity of smoking seem mostly uncorrelated, with the expected negative correlation only among those with a manual occupation. Table 3 shows the results of regression analyses of the association between the TCS and both smoking outcomes, adjusted for age, sex, education, occupation, and survey wave. We found associations of the TCS with the quit ratio in the expected direction, but they are not statistically significant for either smoking cessation (OR: 1.03, 95%CI: ) or intensity of smoking (beta: -0.16, 95%CI: ). This analysis also shows that both education and occupation are strong predictors of smoking cessation and intensity of smoking. 9 Table 4 shows the results of the tests on interaction between TCS and education or occupation. The OR of 1.13 (95%CI: ) indicates that a ten-point increase in the TCS score is associated with a 13% higher quit ratio in higher educated respondents, compared to lower educated respondents. The difference between the middle and lower educated respondents, and the difference between manual and 191

11 Chapter 9 Figure 1 Scatter plots of national, age-standardized quit ratio and intensity of smoking for all three survey waves combined, by educational level and occupational class, plotted against the TCS scores. 192

12 Tobacco control and inequalities in smoking in Europe 9 Figure 1 Continued 193

13 Table 3 Associations between total TCS score with smoking behaviour Smoking cessation Smoking cessation Smoking intensity Smoking intensity Model 1 Model 2 Model 1 Model 2 OR 95% CI OR 95% CI beta 95% CI beta 95% CI TCS Age Male 1.00 reference 1.00 reference 0.00 reference 0.00 reference Female Low ed reference reference - - Middle ed High ed Manual occ reference reference Non-manual occ Other occ Ever smoking Survey wave Model 1 (logistic regression) is adjusted for age, sex, ever smoking, survey wave, and education. Model 2 (logistic regression) is adjusted for age, sex, ever smoking, survey wave, and occupation. Model 3 (linear regression) is adjusted for age, sex, survey wave, and education. Model 4 (linear regression) is adjusted for age, sex, survey wave, and occupation. 1 Age in years when finished/stopped full time education: low 15 years, middle = years, high 20 years. 2 Occupational classes: manual = farmer/fisherman/manual worker, non-manual = professional/white collar/executive, other = business owner/craftsman/retired/housewife/student/military/ unemployed. All models show outcome measures associated with a 10 point increase in TCS score. Significant results (p 0.05) are highlighted in bold. Chapter 9 194

14 Tobacco control and inequalities in smoking in Europe non-manual occupation respondents are small and not statistically significant. For smoking intensity, the association with the TCS is stronger for the middle (beta: , 95%CI: ) and higher educated respondents (beta: -0.27, 95%CI: ) compared to lower educated respondents. The association between the TCS and smoking intensity is strongest for manual occupation respondents (beta: , 95%CI: ), and similar for non-manual respondents, with no association for the other occupation respondents (beta: 0.62, 95%CI: ). Table 4 Interaction between total TCS score and socio-economic status Smoking cessation 3 Smoking intensity 4 OR 95% CI Beta 95% CI Model 1 Education (low) reference 0.00 reference Education (middle) Education (high) TCS (for low education) TCS x Education (middle) TCS x Education (high) Model 2 Occupation (manual) reference 0.00 reference Occupation (non-m) Occupation (other) TCS (for manual occupation) TCS x occupation (non-m) TCS x occupation (other) Age in years when finished/stopped full time education: low 15 years, middle = years, high 20 years. 2 Occupational classes: manual = farmer/fisherman/manual worker, non-manual (non-m) = professional/white collar/executive. 3 Models are adjusted for age, sex, ever smoking prevalence, and survey wave. 4 Models are adjusted for age, sex, and survey wave. Significant results (p 0.05) are highlighted in bold. 9 Table 5 shows the results of the regression analyses per domain of the TCS. For the total population, none of the six domains is significantly associated with the quit ratio. In terms of SES differences, all six domains, except health warning labels, have a positive, significant association with smoking cessation among the higher educated respondents. In contrast, significant associations are not observed for any domain among the lower educated respondents, and only for cessation support among middle educated respondents. For smoking intensity overall, there was one 195

15 Chapter 9 Table 5 Associations between components of the TCS and smoking behaviour, per educational group Smoking cessation Smoking intensity Total population 1 OR 95% CI beta 95% CI Price Smoke-free laws Public information campaigns Advertising bans Health warning labels Cessation support services Total TCS Low education 2 OR 95% CI beta 95% CI Price Smoke-free laws Public information campaigns Advertising bans Health warning labels Cessation support services Total TCS Middle education 2 OR 95% CI beta 95% CI Price Smoke-free laws Public information campaigns Advertising bans Health warning labels Cessation support services Total TCS High education 2 OR 95% CI beta 95% CI Price Smoke-free laws Public information campaigns Advertising bans Health warning labels Cessation support services Total TCS Models are adjusted for all subscales, age, sex, education, ever smoking prevalence, and survey wave. 2 Age in years when finished/stopped full time education: low = 15 years, middle = years, high = 20 years. Models are adjusted for all subscales, age, sex, ever smoking prevalence, and survey wave. Significant results (p 0.05) are highlighted in bold significant association in the total population: a higher score on the smoke-free laws domain is associated with a lower number of cigarettes smoked per day (beta: -0.15, 196

16 Tobacco control and inequalities in smoking in Europe 95%CI: ). This association is also found in the middle and higher, but not in the lower educated respondents. DISCUSSION We found an association between tobacco control policies and smoking cessation among higher educated respondents. Five out of six of the separate policy domains were significantly associated with smoking cessation in higher educated respondents. Moreover, tobacco control policies were significantly associated with a decrease in the intensity of smoking among the middle and higher educated respondents. No significant associations were found in the lower educated respondents for either smoking outcome, not for total tobacco control and not for any of the separate policy domains. Evaluation of data The smoking prevalence rates derived from previous Eurobarometer survey waves have been compared to national survey estimates. 16,17 We compared the 2012 Eurobarometer survey results with another international survey. 18 The correlations between national prevalence rates estimated on the basis of the Eurobarometer and on the basis of the other survey were 0.78 for men and 0.65 for women. Details are shown in Appendix Table 2. The average difference in smoking prevalence rates between the surveys was 0.2 percentage points for men, and 1.1 percentage points for women. Thus, overall, estimates of national prevalence rates based on recent Eurobarometer surveys do not strongly differ from estimates based on other sources. However the prevalence estimates for some countries varied more strongly, and should be interpreted with caution To measure smoking cessation levels, we used the quit ratio. The quit ratio is a cumulative measure that represents changes in smoking cessation in a population over longer periods of time. Data on the age at which respondents had quit smoking, were not available in the Eurobarometer. Associations with recent tobacco control policies could have been identified with greater accuracy if we would have data on smoking cessation in recent years only. 9 Long-term changes in smoking prevalence are determined not only by smoking cessation, but also by smoking initiation. Eurobarometer surveys focus on adults and do not provide information on changes in smoking initiation rates in recent 197

17 Chapter 9 years. Such changes can be monitored by multi-country surveys among youth, such as the ESPAD and HBSC surveys. 19,20 Interpretation of results Although we found an association between TCS and smoking cessation only for higher educated respondents, Schaap et al. 9 found such an association for both higher and lower educated respondents. These contrasting findings may be explained in two ways. Firstly, the data sources differed. Schaap et al. used smoking data from national surveys from 18 countries collected around 2000, and linked these to TCS scores from We used Eurobarometer data from 27 countries from 2006 to 2012, with TCS scores from 2007 to Secondly, the association between tobacco control and smoking cessation could have changed during the 2000s, remaining substantial only among those of higher SES. Since the early 2000s, the composition of tobacco control policies of European countries has changed, including a rise in smoke-free laws. 21 Several studies suggest that the effect of smokefree laws may be stronger among smokers of higher SES than among smokers of low SES. 22,23 Whereas we could not detect an association for lower educated respondents, we did find that a higher TCS score was significantly associated with a higher quit ratio, and lower intensity of smoking among higher educated respondents. More strikingly, our analyses for individual policy domains showed that all policy domains, except health warning labels, were positively associated with smoking cessation rates only among the higher educated respondents. These findings are mostly in line with expectations based on recent review studies which showed that many tobacco control policies are equity negative, i.e. more effective among those with a high SES. 22,24,25 Review studies report that raising taxes/prices is most likely to be equity positive, i.e. to decrease inequalities. 22,24,25 Unlike these reports, we did not observe taxation to be positively related to smoking cessation levels among the low educated. One possible explanation is that, during our study period, the price of tobacco, adjusted for Gross Domestic Product (GDP), hardly increased in most countries. 21 Note that tax increases might have been offset by pricing strategies or increases in GDP. As a result, our study assessed the impact of variations between countries rather than variations over time, and the former variations may be too small, or too much intertwined with other factors, to have a demonstrable impact on smoking cessation patterns. Another possible explanation is that raising tobacco taxes might have become less effective over time among those of lower SES. Most of the evidence 198

18 Tobacco control and inequalities in smoking in Europe that shows the positive effects of tax increases on smoking inequalities in European countries dates back to before Moreover, a US study shows the price elasticity of smoking decreased among those of lower SES from to from to The fact that the tobacco control policies as developed in European countries in the 2000s did not seem to reduce socio-economic inequalities in smoking is unsurprising given the fact that reducing inequalities was not an explicit goal of these policies. Most tobacco control policies implemented in this period were targeted at the total population or at population groups not defined in terms of SES (e.g. youth, bar personnel). One of the few policies implemented with the express aim to decrease inequalities by SES in smoking was the National Health Service s (NHS) smoking cessation support in the UK. 27 Studies showed that the main achievement of these NHS services was the high reach among low SES groups, which more than compensated for their lower quit rates compared to the higher SES groups. 28 Additionally, highly targeted mass media campaigns could also achieve a positive equity impact. Conclusions In our comparison across 27 countries, recent tobacco control policies were associated with smoking cessation among individuals of higher SES, but no association was found among those of lower SES. This finding underlines that in most instances these policies were aimed at reducing smoking in the general population, but not at decreasing inequalities in smoking. If these inequalities are to be reduced, future tobacco control policies should explicitly focus on reaching smokers from low socio-economic groups. Acknowledgements The authors used data from Eurobarometer surveys 66.2, 72.3 and 77.1 made available from the GESIS data archive, Cologne. They would like to thank Wim Busschers for advice on and assistance with statistical analyses. This study is part of the project Tackling socioeconomic inequalities in smoking (SILNE), which is funded by the European Commission, Directorate-General for Research and Innovation, under the FP7-Health-2011 program, with grant agreement number

19 Chapter 9 Author contributions JRB and AEK conceived and designed the study. JRB prepared, analysed and interpreted the data and led the writing; and is the guarantor. MCW, KS and AEK interpreted the data and provided critical revisions. All authors have read and approved the final version of this paper. Conflicts of interests None declared. REFERENCES 1. Doll R, Peto R, Boreham J, Sutherland I. Mortality in relation to smoking: 50 years observations on male British doctors. BMJ. 2004;328(7455): Jha P, Ramasundarahettige C, Landsman V, et al. 21st-Century Hazards of Smoking and Benefits of Cessation in the United States. N Engl J Med. 2013;368(4): Pirie K, Peto R, Reeves GK, Green J, Beral V. The 21st century hazards of smoking and benefits of stopping: a prospective study of one million women in the UK. Lancet. 2013;381(9861): Clancy L. Progress in tobacco control. Health Policy (New York). 2009;91 Suppl 1:S3-S Warner KE, Mendez D. Tobacco control policy in developed countries: yesterday, today, and tomorrow. Nicotine Tob Res. 2010;12: Mackenbach JP, Karanikolos M, McKee M. The unequal health of Europeans: Successes and failures of policies. Lancet. 2013;381(9872): Giskes K, Kunst AE, Benach J, et al. Trends in smoking behaviour between 1985 and 2000 in nine European countries by education. J Epidemiol Community Health. 2005;59: Bosdriesz JR, Willemsen MC, Stronks K, Kunst AE. Socioeconomic inequalities in smoking cessation in 11 European countries from 1987 to J Epidemiol Community Heal. 2015;69: Schaap MM, Kunst AE, Leinsalu M, et al. Effect of nationwide tobacco control policies on smoking cessation in high and low educated groups in 18 European countries. Tob Control. 2008;17(4): Joossens L, Raw M. The Tobacco Control Scale: a new scale to measure country activity. Tob Control. 2006;15: Joossens L, Raw M. Progress in Tobacco Control in 30 European Countries, 2005 to European Cancer Leagues, Brussels; Joossens L, Raw M. The Tobacco Control Scale 2010 in Europe. European Cancer Leagues, Brussels; Joossens L, Raw M. The Tobacco Control Scale 2013 in Europe. European Cancer Leagues, Brussels; Ahmad O, Boschi-Pinto C, Lopez A, Murray C, Lozano R, Inoue M. Age Standardization of Rates: A New WHO Standard.;

20 Tobacco control and inequalities in smoking in Europe 15. Eurostat. Revision of the European Standard Population - Report of Eurostat s Task Force, Luxembourg; Bogdanovica I, Godfrey F, McNeill A, Britton J. Smoking prevalence in the European Union: a comparison of national and transnational prevalence survey methods and results. Tob Control. 2011;20:e Kulik MC, Eikemo TA, Regidor E, Menvielle G, Mackenbach JP. Does the pattern of educational inequalities in smoking in Western Europe depend on the choice of survey? Int J Public Health. 2014;59(4): Zatoński W, Przewoźniak K, Sulkowska U, West R, Wojtyła A. Tobacco smoking in countries of the European Union. Ann Agric Environ Med. 2012;19: Kuipers MAG, Monshouwer K, van Laar M, Kunst AE. Tobacco Control and Socioeconomic Inequalities in Adolescent Smoking in Europe. Am J Prev Med. 2015;49(5):e64-e Pförtner T-K, Hublet A, Schnohr CW, et al. Socioeconomic inequalities in the impact of tobacco control policies on adolescent smoking. A multilevel study in 29 European countries. Addict Behav. 2016;53: Bosdriesz JR, Willemsen MC, Stronks K, Kunst AE. Patterns of tobacco control policy progress in 21 European countries. Tob Regul Sci. 2015;1(3): Brown T, Platt S, Amos A. Equity impact of population-level interventions and policies to reduce smoking in adults: A systematic review. Drug Alcohol Depend. 2014;138: Regidor E, Pascual C, Giráldez-García C, Galindo S, Martínez D, Kunst AE. Impact of tobacco prices and smoke-free policy on smoking cessation, by gender and educational group: Spain, Int J Drug Policy. July Hill S, Amos A, Clifford D, Platt S. Impact of tobacco control interventions on socioeconomic inequalities in smoking: review of the evidence. Tob Control. 2014;23(e2):e89-e Hiscock R, Bauld L, Amos A, Fidler JA, Munafò M. Socioeconomic status and smoking: A review. Ann N Y Acad Sci. 2012;1248: Franks P, Jerant AF, Leigh JP, et al. Cigarette prices, smoking, and the poor: implications of recent trends. Am J Public Health. 2007;97(10): Bauld L, Judge K, Platt S. Assessing the impact of smoking cessation services on reducing health inequalities in England: observational study. Tob Control. 2007;16(6): Brown T, Platt S, Amos A. Equity impact of European individual-level smoking cessation interventions to reduce smoking in adults : a systematic review. Eur J Public Health. 2014;24(4):