Changes in Exposure to Secondhand Smoke and ... - Hogrefe eContent

2 downloads 0 Views 616KB Size Report
Abstract: Aims: This study examines exposure to secondhand smoke and smoking behavior in the German general population before and after the introduction ...
SUCHT, 56 (5), 2010, 373–384

S. Müller et al.: Secondh and Smoke SUCHTand , 56Smoking (5) © 2010 Behavior Verlag Hans Following Huber,Smoke-Free Hogrefe AG ,Laws Bern

Themenschwerpunkt

Changes in Exposure to Secondhand Smoke and Smoking Behavior in the General Population After the Introduction of New Smoke-Free Laws in Germany Veränderungen in der Passivrauchexposition und im Rauchverhalten nach Einführung der Nichtraucherschutzgesetze in Deutschland Stefanie Müller, Ludwig Kraus, Daniela Piontek, and Alexander Pabst IFT Institut für Therapieforschung, Munich, Germany Abstract: Aims: This study examines exposure to secondhand smoke and smoking behavior in the German general population before and after the introduction of new smoke-free laws in 2007 and 2008. Methods: Data came from the 2006 and 2009 German Epidemiological Survey of Substance Abuse (ESA). A propensity-score-matched subsample of n = 7,412 subjects between 18 and 64 years was used for the analysis. We employed multinomial logistic regression to examine changes in exposure to secondhand smoke at work, during leisure time, and at home among nonsmokers between 2006 and 2009. Logistic and ordinary least square regression were used to address changes in cigarette smoking prevalence and number of cigarettes smoked among smokers. Results: Exposure to secondhand smoke at work and during leisure time was substantially reduced in 2009 compared to 2006. The number of cigarettes smoked decreased in the same time period. No changes were found for exposure to secondhand smoke at home and cigarette smoking prevalence. However, sensitivity analysis revealed a moderate stability of the results for exposure to secondhand smoke at home and number of cigarettes smoked. Conclusions: Results indicate that smoke-free laws constitute an important step forward in the protection of nonsmokers in Germany. However, our findings suggest that specific preventive measures are still needed to reduce cigarette smoking prevalence. Keywords: secondhand smoke, smoking, smoke-free legislation, tobacco policy

Zusammenfassung: Ziele: Ziel dieser Studie ist die Erfassung von Veränderungen in der Allgemeinbevölkerung nach Einführung der Nichtraucherschutzgesetze zwischen 2007 und 2008. Methode: Datengrundlage sind die Erhebungen von 2006 und 2009 des Epidemiologischen Suchtsurveys (ESA). Für die Analysen wurde eine auf Grundlage von propensity scores ausbalancierte Substichprobe von n = 7,412 Personen im Alter zwischen 18 und 64 Jahren verwendet. Veränderungen zwischen 2006 und 2009 in der Passivrauchexposition von Nichtrauchern am Arbeitsplatz, während der Freizeit und zuhause wurden mittels multinomial logistischer Regression untersucht. Veränderungen in der Zigarettenrauchprävalenz und der Anzahl gerauchter Zigaretten innerhalb der letzten 30 Tage wurden mit logistischer bzw. linearer Regression erfasst. Ergebnisse: Die Passivrauchexposition am Arbeitsplatz und während der Freizeit verringerte sich zwischen 2006 und 2009. Auch die Anzahl gerauchter Zigaretten nahm ab. Passivrauchexposition zuhause und Zigarettenrauchprävalenz blieben unverändert. Die Sensitivitätsanalyse zeigte jedoch für die Passivrauchbelastung zuhause und die Anzahl gerauchter Zigaretten nur eine moderate Stabilität. Schlussfolgerung: Nichtraucherschutzgesetze stellen einen wichtigen Schritt für den Nichtraucherschutz in Deutschland dar. Die Ergebnisse deuten jedoch auch darauf hin, dass die Reduktion der Zigarettenrauchprävalenz spezifische Präventionsmaßnahmen erfordert. Schlüsselwörter: Nichtraucherschutzgesetze, Passivrauchen, Rauchen, Tabakpolitik

DOI: 10.1024/0939-5911/a000048

SUCHT, 56 (5) © 2010 Verlag Hans Huber, Hogrefe AG, Bern

374

S. Müller et al.: Secondhand Smoke and Smoking Behavior Following Smoke-Free Laws

Introduction This paper examines changes in exposure to secondhand tobacco smoke and smoking behavior in Germany between 2006 and 2009. The objective is to evaluate the effects of smoke-free laws implemented within this period. Secondhand tobacco smoke (SHS) is a mixture of the smoke emitted from the burning end of a tobacco product and the smoke-filled air exhaled by a smoker (U.S. Department of Health and Human Services, 2004). There is conclusive evidence that SHS exposure contributes to the development of cancer, cardiovascular disease, and respiratory conditions and is particularly dangerous for children (Deutsches Krebsforschungszentrum, 2005; IARC Handbooks of Cancer Prevention, 2009; IARC Monographs on the evaluation of the carcinogenic risk of chemicals to humans, 2004; U.S. Department of Health and Human Services, 2006). In Germany, in 2005 an estimated 3,300 deaths were attributable to SHS exposure (Keil et al., 2005). Because there is no safe limit for tobacco smoke, the goal of smoke-free legislation is to entirely eliminate involuntary exposure to SHS (IARC Handbooks of Cancer Prevention, 2009). Research to date shows that smoke-free legislation is followed by substantial reductions in exposure to SHS (e.g., Fong et al., 2006; for an overview see IARC Handbooks of Cancer Prevention, 2009). In addition, although evidence is less conclusive, smoking bans have been found to be associated with lower smoking prevalence (Fichtenberg & Glantz, 2002; Wakefield et al., 2000) and more consistently with a reduction in the number of cigarettes smoked among smokers (Fichtenberg & Glantz, 2002; Gallus et al., 2006). Moreover, beneficial health effects have been observed after the implementation of smoke-free legislation. Short-term improvements in acute respiratory illnesses following smoking bans have been reported among employees in the hospitality sector (Eagan, Hetland & Aarø, 2006; Eisner, Smith & Blanc, 1998), and early findings indicate declines in hospital admissions for acute myocardial infarction linked to restrictions on smoking (Khuder et al., 2007; Pell et al., 2008). Since the effectiveness of smoke-free legislation is well established, measures to protect the population from exposure to SHS are an integral part of all effective tobacco control policies (Hanewinkel, 2008; World Health Organization, 2005). While several European countries started to introduce complete smoke-free environments in all types of institutions (e.g., government facilities, health care, and educational premises) and other types of places (e.g., public transport, indoor offices, restaurants, bars, and pubs) in 2004 (IARC Handbooks of Cancer Prevention, 2009), Germany’s smoke-free legislation was very lenient until 2007 (Joossens & Raw, 2006). In 2006, German legislation banned smoking in workplaces exempting the hospitality sector (§ 5 ArbStättV) with the result of very high SHS exposure to staff and guests in bars and restaurants (Bolte et al., 2008; Schneider et al., 2008). SUCHT, 56 (5) © 2010 Verlag Hans Huber, Hogrefe AG, Bern

Criticism concerning incomprehensive smoking bans was raised by leading health authorities in Germany (Deutsches Krebsforschungszentrum, 2005) and the European Union (Handelsblatt, 2006), pressuring the government as well as the federal states to take action. In 2007, a statewide bill improving the protection of nonsmokers throughout Germany was introduced. The bill banned smoking in buses, trains, train stations, taxis, and federal government buildings while still providing the option for separate smoking rooms (§ 1 BNichtrSchG). The federal states passed laws prohibiting smoking in public, educational, and health premises, in cultural and sports venues as well as in the hospitality sector. However, exceptions could be made, especially in bars and clubs (Deutsches Krebsforschungszentrum, 2009). In 2008, the Federal Constitutional Court allowed smoking in one-room pubs smaller than 75 square meters if minors are not admitted, no food is served, and the venue is clearly labeled as a “smoking bar.” In addition, discotheques may create smoking sections in side rooms if minors are not admitted and there is no dance floor in the smoking section (Bundesverfassungsgericht, 2008). To date, little is known about the effects of the new smoke-free laws on exposure to SHS and smoking behavior in Germany. One study indicates substantially improved air quality in selected hospitality venues in two federal states before and after the legislation (Deutsches Krebsforschungszentrum, 2007). With regard to smoking behavior, a study in the general population found no differences in smoking prevalence and the number of cigarettes smoked following the introduction of smoke-free legislation (Anger, Kvasnicka & Siedler, 2010). A comprehensive assessment of the effects of smoking bans in Germany is important to help policymakers provide adequate protection for nonsmokers. This study examines changes in exposure to SHS among nonsmokers as well as cigarette smoking prevalence and the average number of cigarettes smoked among smokers in the German general population after the introduction of new smoke-free laws using cross-sectional data from 2006 and 2009.

Methods Sample Data came from the nationwide, cross-sectional 2006 and 2009 Epidemiological Survey of Substance Abuse (ESA; Kraus & Baumeister, 2008; Kraus & Pabst, 2010). Participants from the German adult population aged 18 to 64 years were selected using a two-stage probability sampling design. In the first stage, communities were selected proportional to the population size. In the second stage, participants were randomly drawn from the residents’ registration office. Additionally, the 2006 and 2009 samples were disproportionally drawn in terms of age and birth cohort, respectively.

S. Müller et al.: Secondhand Smoke and Smoking Behavior Following Smoke-Free Laws

A mixed-mode procedure was used in both surveys. In 2006, all participants received postal questionnaires in a first step. In the case of nonresponse, participants could answer the questionnaire by telephone after the third reminder. In 2009, questionnaires could be answered by mail, telephone, and online. Out of all eligible subjects, 7,912 (2006) and 8,030 (2009) participated, corresponding to response rates of 44.9% in 2006 and 50.1% in 2009, respectively. Data of all surveys were weighted to represent the distribution of age and gender in the general population of Germany in a given year. Records with implausible values in cigarette consumption (> 200 cigarettes per day: n = 1) or number of children at home (> 60 children at home: n = 6) were discarded. Cases with missing values in one of the outcome variables were excluded as well (cigarette smoking prevalence within the last 30 days: n = 121, quantity frequency index of cigarettes within the last 30 days among smokers: n = 130, SHS exposure at work among employed subjects: n = 309, SHS exposure during leisure time: n = 335, SHS exposure at home: n = 227). Thus, the analytical sample consisted of n = 14,813 (92.9%) valid cases, thereof n = 7,099 (89.7%) in 2006 and n = 7,714 (96.1%) in 2009.

Measures Outcome Measures Exposure to SHS was assessed by asking respondents how often they stayed in rooms where smoking occurred at work, during leisure time, and at home. Answer categories were (almost) daily, two to three times a week, once a week, once a month, less than monthly, and never. For the sake of parsimony, these categories were combined into (almost) daily, one to three times a week, and (almost) never. Smoking behavior was assessed in terms of cigarette smoking prevalence and average number of cigarettes smoked per day within the last 30 days. Cigarette smoking prevalence was defined as a) having smoked more than 100 cigarettes in one’s life and b) having smoked cigarettes within the last 30 days. The number of cigarettes smoked by smokers reflects the average number of cigarettes smoked per day and was measured using a quantity-frequency index. Frequency of cigarette use was assessed by respondents’ indications on how many days within the last 30 days they had smoked. Quantity was measured by the average number of cigarettes smoked on a typical day. The average number of cigarettes was obtained by multiplying the two measures and by dividing the result by 30. Cigarette smoking prevalence was chosen instead of smoking prevalence including cigarettes, cigars, cigarillos, and pipe to ensure comparability of indicators for smoking behavior (cigarette smoking prevalence and average number of cigarettes). The absolute difference between cigarette smoking prevalence and

375

smoking prevalence is 0.2%, indicating that smoking prevalence is nearly identical to cigarette smoking. Nonsmokers, however, were defined as not having smoked more than 100 cigarettes, cigars, cigarillos, or pipe fillings in their life, and not having smoked within the last 30 days.

Covariates To control for possible confounding influences, age, gender, education, inflation-adjusted equivalence income, marital status, children at home, nationality, and region were used as covariates in all analyses. Age was grouped into three categories of 18–29, 30–49 and 50–64 years. Education was categorized into participants with less, equal, and more than 10 years of education. Income was grouped into less than 1000 Euros, 1000 to 2000 Euros, and more than 2000 Euros per month. Marital status comprised the categories single, married, widowed, and divorced. The number of children at home was grouped into households with no children, with one to two children, and three or more children. Nationality indicated German vs. other nationalities and region indicated West vs. East Germany. For the computation of propensity scores, also profession (blue collar, white collar, self-employed, student, miscellaneous), employment (full time, part time, not employed, miscellaneous), federal state, physical and mental health (measured on a 5-point scale from very good to very bad) and number of chronic diseases were used. In addition, 30-day prevalence of prescription drug use and at least weekly use as well as problematic use of prescription drugs within the last 12 months as indicated by the Kurzfragebogen zum Medikamentengebrauch (KFM; Watzl, Rist, Höcker & Miehle, 1991) were included. Moreover, quantity (in grams of ethanol) and frequency of alcohol consumption and binge drinking within the last 30 days were used for the computation of propensity scores. Furthermore, 30-day prevalence of cannabis use, problematic cannabis use within the last 12 months measured by the severity of dependence scale (SDS, Gossop & Darke, 1995) and 30-day prevalence of illegal drug use (other than cannabis) were employed. In order to assess changes in exposure to SHS and smoking behavior after the introduction of the new smoke-free laws, we used an indicator of survey year 2006 (coded 0 “before the introduction”) and 2009 (coded 1 “after the introduction”).

Statistical Analyses All analyses were conducted using survey procedures of Stata 10.1 SE software package (StataCorp LP, College Station, TX, USA) to adjust for the complex sampling design. For descriptive statistics, Pearson χ² tests with SUCHT, 56 (5) © 2010 Verlag Hans Huber, Hogrefe AG, Bern

376

S. Müller et al.: Secondhand Smoke and Smoking Behavior Following Smoke-Free Laws

SUCHT, 56 (5) © 2010 Verlag Hans Huber, Hogrefe AG, Bern

S. Müller et al.: Secondhand Smoke and Smoking Behavior Following Smoke-Free Laws

Rao/Scott correction were used for comparisons between survey years. To address missing values in covariates, we implemented multiple imputation using chained equations (Royston, 2005). The multiple imputation procedure produced five samples out of which one was randomly selected. As a result, we were able to use data from all 14,813 subjects with valid outcome data.

Propensity Score Matching In our study, the effect of the introduction of smoke-free laws on exposure to SHS and smoking behavior was evaluated using observational data. Unlike in experimental studies, participants were not randomly assigned to the control (not exposed to smoke-free laws, survey year 2006) or the intervention group (exposed to smoke-free laws, survey year 2009). Therefore, large differences in covariates may exist that could result in a biased intervention effect estimate (D’Agostino, Jr., 1998). Propensity scores, defined as the conditional probability of being treated given the covariates, can be used to balance covariates in the two groups, and therefore minimize this bias. Propensity scores were estimated using logistic regression with survey year as criterion and covariates as predictors. For exposed subjects in 2009, a one-one matched sample of control subjects from the 2006 data set was selected. The technique used was Mahalanobis metric matching including the propensity score within propensity score calipers (Rosenbaum & Rubin, 1985). The caliper width was 0.2 of a linear propensity score standard deviation, and the variables included in the metric were age, gender, education, inflation-adjusted equivalence income, marital status, number of children at home, nationality, and region as well as the propensity score. As a result of matching, 7,412 subjects (50.0%) could be selected for the analyses. Changes in effect size indicate that individuals surveyed in 2009 resembled those surveyed in 2006 with regard to most covariates. Table 1 shows the distribution of the most important sociodemographic characteristics in the original unmatched sample and the matched sample (full table available on request). As indicated by the smaller standardized differences (Austin, 2009), balance of covariates between survey years in the matched sample was improved in all but three variables. However, differences in the number of children at home, nationality, and region were small. Differences in number of children at home and income were still existent.

Changes in Outcome Measures The propensity score matched sample was used to evaluate differences in the outcome variables between survey years. With regard to exposure to SHS at work, during

377

leisure time, and at home, multinomial logistic regression was used to assess changes between survey years using (almost) never as the base category. The resulting relative risk ratios (RRR) indicate the proportionate change in the relative risk of being exposed (almost) daily/1–3 times a week rather than (almost) never. An adjusted Wald test was used to assess the joint significance for the predictor survey year. With regard to cigarette smoking prevalence, logistic regression was employed to assess changes before and after the introduction of new smoke-free laws. Changes in the average number of cigarettes smoked per day were examined using ordinary least square regression. As the distribution of number of cigarettes was skewed to the right, a logarithmic transformation was used to approach a normal distribution (Tabachnick & Fidell, 2001). Coefficients for continuous variables can be interpreted as the expected proportional change in the outcome variable per proportional change in the predictor (Gelman & Hill, 2007). Note that the number of respondents differs across the outcome measures. For SHS exposure at work, during leisure time and at home, only nonsmokers were included. In addition, for SHS exposure at work, only employed nonsmokers were used. Changes in the number of cigarettes smoked were evaluated in smokers only, while the whole sample was used for the analysis of cigarette smoking prevalence.

Sensitivity Analysis Different specifications of the propensity score produce different matched samples which include only a part of the original study participants (50.0%). Therefore, it is important to examine sensitivity of the results to different specifications of the propensity score, i.e., different sample compositions (Guo & Fraser, 2009; Rozen, 2005). In this study, an adapted version of Hirano and Imbens’ (2001) method for specifying propensity score variables on predetermined t values was used. First, critical t values of t = 0, t = 1, t = 2, t = 4, t = 8, and t = ∞ for inclusion of a covariate in the logistic regression were determined. Second, for each of the 23 covariates, a simple logistic regression model was run resulting in a specific t value. Third, while all 23 covariates met the requirement of t = 0 (original specification), for the model specification of t > 1 only covariates exceeding t-values of 1 were included in the model. For the last model t = 8 (W t = ∞), no covariates were included resulting in a propensity score reflecting the probability of being treated. Fourth, the matching procedure based on each of the six specifications of the propensity score was executed, and the original statistical analyses were run in the resulting matched samples (five new matched samples plus original matched sample). Sensitivity of the results to the specifications of the propensity score was evaluated by means of the significance of the coefficients for survey year relSUCHT, 56 (5) © 2010 Verlag Hans Huber, Hogrefe AG, Bern

378

S. Müller et al.: Secondhand Smoke and Smoking Behavior Following Smoke-Free Laws

ative to the coefficients obtained in the original matched sample.

and results for the number of cigarettes smoked in only 1 out of 5 samples.

Results

Discussion

Descriptive Results

This study investigated changes in exposure to SHS and smoking behavior in Germany after the introduction of new smoke-free laws in 2007 and 2008. Results show that exposure to SHS among nonsmokers at work and during leisure time was substantially reduced in 2009 compared to 2006. The number of cigarettes smoked decreased in the same time period. No changes could be found for SHS exposure at home and cigarette smoking prevalence. However, sensitivity analysis revealed a moderate stability for the results for SHS exposure at home and the number of cigarettes smoked. Caution is suggested in inferring causal effects of smoking bans on exposure to SHS and smoking behavior from these results. First, using observational data to assess the effects of smoke-free legislation may result in biased intervention effect estimates due to differences in covariates. These differences were substantially reduced in the propensity-score-matched data sets (Table 1). Thus, biases introduced by observed covariates are unlikely, but biases caused by unobserved covariates cannot be ruled out. Second, given that our results are based on cross-sectional data from 2006 and 2009, changes might have been influenced by other factors than smoke-free legislation. Between 2006 and 2009 cigarette prices rose slightly due to an increase in value added tax from 16% to 19%; also, advertising bans were introduced and sales to minors were forbidden. In addition, there is indication for a overall declining trend (Lampert & List, 2010). Nevertheless, we believe that results can be interpreted as indications of the effects of smoke-free legislation keeping in mind the existence of possible alternative explanations.

To illustrate the effects of matching on outcome measures, Table 2 shows descriptive statistics for the outcomes in the original as well as in the matched sample before and after the implementation of smoke-free laws. Exposure to SHS among nonsmokers was reduced in all domains and in both samples between survey years. An exception was being exposed 1–3 times a week to passive smoking at home in the matched sample. In both survey years, the vast majority of respondents reported to have (almost) never been exposed to SHS. With regard to smoking behavior, both cigarette smoking prevalence and average number of cigarettes smoked per day declined across survey years in the unmatched and matched sample. However, after matching, the reduction in the number of cigarettes was not significant.

Changes in Outcome Measures Table 3 shows results for changes in SHS exposure between survey years in the matched sample. After the introduction of smoke-free laws, the odds of being exposed to SHS as compared to being (almost) never exposed showed a substantial decline at the workplace and during leisure time (Fat work(2, 2,951) = 44.3, p < .001; Fduring leisure time(2, 5,128) = 90.5, p < .001). For example, in 2009, the odds of being (almost) daily exposed to SHS during leisure time decreased by more than two thirds compared to 2006. For exposure to SHS at home, no changes between survey years could be observed (Fat home(2, 5, 128) = 2.0, ns). Table 4 shows results for changes in smoking behavior between 2006 and 2009. The average number of cigarettes smoked per day decreased after the introduction of smokefree laws. In contrast, smoking prevalence did not change between survey years.

Sensitivity Analysis The sensitivity analysis showed perfect stability of the results for exposure to SHS at work and at home to different specifications of the propensity score. Results of the original sample could be replicated in all 10 analyses (5 samples × 2 outcomes). For the other outcomes, however, results were less robust. Original findings for cigarette smoking prevalence could be replicated in 4 out of 5 samples, results for exposure to SHS at home in 3 out of 5 samples, SUCHT, 56 (5) © 2010 Verlag Hans Huber, Hogrefe AG, Bern

Exposure to SHS The main finding of our study is a substantial decline in exposure to SHS among nonsmokers at work and during leisure time after the implementation of new smoke-free legislation. These results are in line with a study reporting substantially improved air quality in hospitality venues in two German federal states after legislation (Deutsches Krebsforschungszentrum, 2007). Our findings are also consistent with international studies showing significant reductions of self-reported SHS exposure in the general population (e.g., Haw & Gruer, 2007; Pickett, Schober, Brody, Curtin & Giovino, 2006) as well as in biomarkers among hospitality employees following smoking bans in the United States (e.g., Farrelly et al., 2005; Hahn et al., 2006) and Europe (e.g., Mulcahy, Evans, Hammond, Repace & Byrne, 2005; Semple, Creely, Naji, Miller & Ayres,

S. Müller et al.: Secondhand Smoke and Smoking Behavior Following Smoke-Free Laws

379

SUCHT, 56 (5) © 2010 Verlag Hans Huber, Hogrefe AG, Bern

380

S. Müller et al.: Secondhand Smoke and Smoking Behavior Following Smoke-Free Laws

SUCHT, 56 (5) © 2010 Verlag Hans Huber, Hogrefe AG, Bern

S. Müller et al.: Secondhand Smoke and Smoking Behavior Following Smoke-Free Laws

381

Table 4 Results of regression analyses for changes in smoking behavior between 2006 and 2009 Smoking behavior Cigarette smoking prevalence (30 day) (n = 7,444) (OR, 95% CI)

Average number of cigarettes per day1 (n = 2,336) (β, 95% CI)

Year (2009)

0.91 (0.81, 1.02)

–0.12 (–0.23, 0.002)*

Gender (male)

1.38 (1.23, 1.55)***

0.22 (0.10, 0.34)***

18–29

1.00

1.00

30–49

0.89 (0.76, 1.03)

0.39 (0.25, 0.53)***

50–64

0.51 (0.41, 0.63)***

0.50 (0.30, 0.69)***

Age

Education (years) < 10

1.00

= 10

0.82 (0.70, 0.95)*

–0.30 (–0.43, –0.17)***

1.00

> 10

0.47 (0.40, 0.55)***

–0.68 (–0.83, –0.53)***

Income (adjusted to inflation EUR/month) < 1000

1.00

1000–2000

0.81 (0.71, 0.92)**

–0.13 (–0.24, –0.01)*

1.00

> 2000

0.74 (0.60, –0.91)**

–0.41 (–0.65, –0.17)**

Marital status Married

1.00

1.00

Divorced

2.26 (1.81, 2.82)***

0.21 (0.02, 0.40)*

Widowed

1.67 (1.08, 2.57)*

0.27 (–0.13, 0.66)

Single

1.48 (1.25, 1.76)***

0.19 (0.04, 0.35)*

0

1.00

1.00

1–2

0.86 (0.74, 0.99)*

–0.16 (–0.30, –0.02)*

3+

0.76 (0.56, 1.02)

–0.12 (–0.46, 0.22)

Number of children at home

Nationality (German)

1.11 (0.88, 1.42)

0.002 (–0.15, 0.15)

Region (West) 0.85 (0.73, 0.99) 2.14 (1.88, 2.40) Notes. 1Intensity is log-transformed due to the skewed distribution; OR = odds ratios, CI = confidence interval. *p < .05, **p < .01, ***p < .001.

2007; Valente et al., 2007). Smoking bans are estimated to be the second most effective measure after taxes in reducing exposure to SHS (Joossens & Raw, 2006). Given the high levels of exposure to SHS in some public places before the implementation of smoking bans (Mons, Amhof & Pötschke-Langer, 2008; Schneider et al., 2008), our results indicate that the new smoke-free legislation constitutes an important step forward in the protection of nonsmokers at public places. Since legal bans at workplaces (§ 5 ArbStättV) and voluntary bans in public transport, educational and health premises were in place prior to smoke-free legislation, the substantial effect found in SHS exposure at work in the general population may seem surprising. However, laws were not very specific about the measures that should be taken to protect nonsmokers and the degree of implementation and effectiveness varied. The introduction of smoke-free laws may have improved existing measures reducing SHS exposure at all workplaces, not only at those in the hospitality sector. Moreover, the in-

troduction of smoke-free laws might have increased awareness of the hazardous health effects of passive smoking leading nonsmoking employees to avoid smoking rooms at their workplaces. Not being a direct target of smoking bans it is not surprising that exposure to SHS at home did not change after the introduction of smoking bans. Evidence for reductions in SHS exposure at home in the context of smoke-free laws is mixed, i.e., some studies reported an increase in smoking bans at home following smoke-free laws, while others did not find an effect (Borland et al., 2006; Galàn et al., 2007). Our results discourage the assumption of a negative side effect of smoking bans, i.e., that, if banned in public places, smoking will be transferred to the home increasing social costs since nonsmoking family members would now be exposed (Borland, Mullins, Trotter & White, 1999; Merom & Rissel, 2001). However, given that the sensitivity analysis revealed moderate stability of the effect to different specifications of the propensity score, results should be interpreted with caution. SUCHT, 56 (5) © 2010 Verlag Hans Huber, Hogrefe AG, Bern

382

S. Müller et al.: Secondhand Smoke and Smoking Behavior Following Smoke-Free Laws

Smoking Behavior

Conclusion

Smoke-free laws are not intended to reduce smoking behavior per se. Yet, clean indoor air laws may make smoking less attractive by reducing opportunities to smoke and by supporting social norms against smoking (Levy, Chaloupka & Gitchell, 2004). This rationale might explain the decline in the number of cigarettes smoked revealed in our study. This finding is corroborated by international evidence indicating that smoking restrictions tend to decrease cigarette consumption rather than smoking prevalence (Braverman, Aarø & Hetland, 2008; Galàn et al., 2007; Heloma & Jaakkola, 2003). However, the sensitivity analysis suggested a moderate stability of the effect on smoking intensity. Therefore, despite support from international evidence, this result should be interpreted with caution. A recent German longitudinal study in the general population did not find changes in either smoking prevalence or number of cigarettes smoked (Anger et al., 2010). Anger and colleagues (2010) used variation in the dates of introduction of smoking bans across federal states to examine effects on smoking behavior. Individual differences between subjects exposed and not exposed to smoking bans were contrasted, eliminating general trends from their results. Hence, while German studies suggest that smoking bans may not have an effect on smoking prevalence, evidence on smoking intensity is inconclusive. There is evidence for subgroup-specific effects of smoking bans on smoking behavior. Anger and colleagues (2010) found a decrease in both smoking prevalence and intensity among the subgroup of young and unmarried individuals. Going out more often, this subgroup is likely to be more affected by smoking bans than older, married individuals. Future studies should investigate subgroup-specific effects of smoking bans, for example, individuals with a low socioeconomic background. A limitation of this study is that, after propensity score matching, our results do not generalize to the German general population. However, the aim of this paper was to obtain unbiased intervention effect estimates rather than representative prevalence estimates. For the same reason, low response rates encountered in both surveys are not likely to threaten the validity of our results. In contrast, differences in response and missing rates and differences in modes of administration between surveys may have biased intervention effect estimates. However, nonresponse and mode effect analyses revealed only small differences in smoking prevalence estimates between nonrespondents and respondents as well as across administration modes in both surveys (Kraus & Baumeister, 2008; Kraus & Pabst, 2010). This indicates that our results are unlikely to have been unacceptably biased by nonresponse related to sociodemographic data or differences in administration modes.

Results of this study indicate that the new smoke-free laws implemented in 2007 and 2008 in Germany constitute an important step forward in the protection of nonsmokers at public places. Thus, objectives of smoking bans were achieved. However, as shown by previous research, there seems to be no effect on cigarette smoking prevalence, suggesting that specific preventive measures are needed to reduce cigarette smoking prevalence.

SUCHT, 56 (5) © 2010 Verlag Hans Huber, Hogrefe AG, Bern

Acknowledgments Funding of the Epidemiological Survey of Substance Abuse (ESA) was provided by the German Federal Ministry of Health (Grant No. 119-4914-8/32).

Declaration of Possible Conflicts of Interest There were no conflicts of interest related to the preparation of this paper.

References Anger, S., Kvasnicka, M., & Siedler, T. (2010). One last puff? Public smoking bans and smoking behavior. German SocioEconomic Panel Study (SOEP) papers on multidisciplinary panel data research at Deutsches Institut für Wirtschaftsforschung (DIW) Berlin. Austin, P. C. (2009). Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Statistics in Medicine, 28, 3083–3107. Bolte, G., Heitmann, D., Kiranoglu, M., Schierl, R., Diemer, J., Koerner, W., & Fromme, H. (2008). Exposure to environmental tobacco smoke in German restaurants, pubs and discotheques. Journal of Exposure Science and Environmental Epidemiology, 18, 262–271. Borland, R., Mullins, R., Trotter, L., & White, V. (1999). Trends in environmental tobacco smoke restrictions in the home in Victoria, Australia. Tobacco Control, 8, 266–271. Borland, R., Yong, H. H., Cummings, K. M., Hyland, A., Anderson, S., & Fong, G. T. (2006). Determinants and consequences of smoke-free homes: Findings from the International Tobacco Control (ITC) Four Country Survey. Tobacco Control, 15(Suppl. 3), iii42–iii50. Braverman, M. T., Aarø, L. E., & Hetland, J. (2008). Changes in smoking among restaurant and bar employees following Norway’s comprehensive smoking ban. Health Promotion International, 23, 5–15. Bundesverfassungsgericht. (2008). Beschluss vom 30. Juli 2008. 1 BvR 3262/07. D’Agostino, R. B., Jr. (1998). Propensity score methods for bias reduction in the comparison of a treatment to a nonrandomized control group. Statistics in Medicine, 17, 2265–2281.

S. Müller et al.: Secondhand Smoke and Smoking Behavior Following Smoke-Free Laws

Deutsches Krebsforschungszentrum [German Cancer Research Center]. (2005). Passivrauchen – ein unterschätztes Gesundheitsrisiko [Passiv smoking – an underestimated health risk] (5). Heidelberg: Deutsches Krebsforschungszentrum. Deutsches Krebsforschungszentrum [German Cancer Research Center]. (2007). Erhöhtes Gesundheitsrisiko für Beschäftigte in der Gastronomie durch Passivrauchen am Arbeitsplatz [Increased health risk for employees in the hospitality sector due to passive smoking at work]. Heidelberg: Deutsches Krebsforschungszentrum. Deutsches Krebsforschungszentrum [German Cancer Research Center]. (2009). Tabakatlas Deutschland 2009 [German Tobacco Atlas 2009]. Heidelberg: Deutsches Krebsforschungszentrum. Eagan, T. M. L., Hetland, J., & Aarø, L. E. (2006). Decline in respiratory symptoms in service workers five months after a public smoking ban. Tobacco Control, 15, 242–246. Eisner, M. D., Smith, A. K., & Blanc, P. D. (1998). Bartenders’ respiratory health after establishment of smoke-free bars and taverns. The Journal of the American Medical Association, 280, 1909–1914. Farrelly, M. C., Nonnemaker, J. M., Chou, R., Hyland, A., Peterson, K. K., & Bauer, U. E. (2005). Changes in hospitality workers’ exposure to secondhand smoke following the implementation of New York’s smoke-free law. Tobacco Control, 14, 236–241. Fichtenberg, C. M., & Glantz, S. A. (2002). Effect of smoke-free workplaces on smoking behavior: Systematic review. British Medical Journal, 325, 188. Fong, G. T., Hyland, A., Borland, R., Hammond, D., Hastings, G., McNeill, A., . . . Driezen, P. (2006). Reductions in tobacco smoke pollution and increases in support for smoke-free public places following the implementation of comprehensive smoke-free workplace legislation in the Republic of Ireland: Findings from the ITC Ireland/UK Survey. Tobacco Control, 15(Suppl. 3), iii51–iii58. Galàn, D., Mata, N., Estrada, C., Díez-Ganán, L., Velázquez, L., Zorrilla, B., . . . Ortiz, H. (2007). Impact of the “Tobacco control law” on exposure to environmental tobacco smoke in Spain. BMC Public Health, 7, 224. Gallus, S., Zuccaro, P., Colombo, P., Apolone, G., Pacifici, R., Garattini, S., & La Vecchia, C. (2006). Effects of new smoking regulations in Italy. Annals of Oncology, 17, 346–347. Gelman, A., & Hill, J. (2007). Data analysis using regression and multilevel/hierarchical models. New York: Cambridge University Press. Gossop, M., & Darke, S. (1995). The Severity of Dependence Scale (SDS): Psychometric properties of the SDS in English and Australian samples of heroin, cocaine and amphetamine users. Addiction, 90, 607–614. Guo, S., & Fraser, M. (2009). Propensity score analysis: Statistical methods and applications. Thousand Oaks, CA: Sage. Hahn, E. J., Rayens, M. K., York, N., Okoli, C. T. C., Zhang, M., Dignan, M., & Al-Delaimy, W. K. (2006). Effects of a smokefree law on hair nicotine and respiratory symptoms of restaurant and bar workers. Journal of Occupational and Environmental Medicine/American College of Occupational and Environmental Medicine, 48, 906–913. Handelsblatt. (2006). EU kritisiert mangelnden Nichtraucherschutz [EU criticizes insufficient protection of nonsmokers]. Received from http://www.handelsblatt.com/politik/interna-

383

tional/eu-kritisiert-mangelnden-nichtraucherschutz-in-deuts chland;1063518 Hanewinkel, R. (2008). Tabakpolitik [Tobacco control policy]. Suchttherapie, 9, 93–102. Haw, S. J., & Gruer, L. (2007). Changes in exposure of adult nonsmokers to secondhand smoke after implementation of smokefree legislation in Scotland: National cross sectional survey. British Medical Journal, 335, 549. Heloma, A., & Jaakkola, M. S. (2003). Four-year follow-up of smoke exposure, attitudes and smoking behavior following enactment of Finland’s national smoke-free work-place law. Addiction, 98, 1111–1117. Hirano, K., & Imbens, G. W. (2001). Estimation of causal effects using propensity score weighting: An application to data on right heart catherization. Health Services and Outcomes Research Methodology, 2, 259–278. IARC Handbooks of Cancer Prevention. (2009). Evaluating the effectiveness of smoke-free policies (Vol. 13). Lyon, France: International Agency for Research on Cancer. IARC Monographs on the evaluation of the carcinogenic risk of chemicals to humans. (2009). Tobacco smoke and involuntary smoking (Vol. 83). Lyon, France: International Agency for Research on Cancer. Joossens, L., & Raw, M. (2006). The Tobacco Control Scale: A new scale to measure country activity. Tobacco Control, 15, 247–253. Keil, U., Becher, H., Heidrich, J., Heuschmann, P., Kraywinkel, K., Vennemann, M., & Wellmann, J. (2005). Passivrauch bedingte Morbidität und Mortalität in Deutschland [Morbidity and mortality attributable to passive smoking in Germany]. In Deutsches Krebsforschungszentrum (Hrsg.), Passivrauchen – ein unterschätztes Gesundheitsrisiko. Heidelberg: Deutsches Krebsforschungszentrum. Khuder, S. A., Milz, S., Jordan, T., Price, J., Silvestri, K., & Butler, P. (2007). The impact of a smoking ban on hospital admissions for coronary heart disease. Preventive Medicine, 45, 3–8. Kraus, L., & Baumeister, S. (2008). Studiendesign und Methodik des Epidemiologischen Suchtsurveys 2006 [Study design and methodology of the 2006 Epidemiological Survey of Substance Abuse]. Sucht, 54(Sonderheft 1), S6–S15. Kraus, L., & Pabst, A. (2010). Studiendesign und Methodik des Epidemiologischen Suchtsurveys 2009 [Study design and methodology of the 2009 Epidemiological Survey of Substance Abuse]. Sucht, 56, 315–326. Lampert, T., & List, S. M. (2010). Tabak – Zahlen und Fakten zum Konsum [Tobacco – Figures and facts about consumption]. In Deutsche Hauptstelle für Suchtfragen e.V. (Hrsg.), Jahrbuch Sucht 2010 (S. 48–68). Geesthacht: Neuland. Levy, D. T., Chaloupka, F., & Gitchell, J. (2004). The effects of tobacco control policies on smoking rates: A tobacco control scorecard. Journal of Public Health Management and Practice, 10, 338–353. Merom, D., & Rissel, C. (2001). Factors associated with smokefree homes in NSW: Results from the 1998 NSW Health Survey. Australian and New Zealand Journal of Public Health, 25, 339–345. Mons, U., Amhof, R., & Pötschke-Langer, M. (2008). Gesetzliche Maßnahmen zum Nichtraucherschutz in Deutschland – Einstellungen und Akzeptanz in der Bevölkerung [Legal measures of nonsmoker protection in Germany – Public attitudes and acceptance]. In J. Böcken, B. Braun & R. Amhof (Hrsg.), GeSUCHT, 56 (5) © 2010 Verlag Hans Huber, Hogrefe AG, Bern

384

S. Müller et al.: Secondhand Smoke and Smoking Behavior Following Smoke-Free Laws

sundheitsmonitor 2008. Gesundheitsversorgung und Gestaltungsoptionen aus der Perspektive der Bevölkerung (S. 181–209). Gütersloh: Bertelsmann Stiftung. Mulcahy, M., Evans, D. S., Hammond, S. K., Repace, J. L., & Byrne, M. (2005). Secondhand smoke exposure and risk following the Irish smoking ban: An assessment of salivary cotinine concentrations in hotel workers and air nicotine levels in bars. Tobacco Control, 14, 384–388. Pell, J. P., Haw, S., Cobbe, S., Newby, D. E., Pell, A. C. H., Fischbacher, C., . . . Borland, W. (2008). Smoke-free legislation and hospitalizations for acute coronary syndrome. The New England Journal of Medicine, 359, 482–491. Pickett, M. S., Schober, S. E., Brody, D. J., Curtin, L. R., & Giovino, G. A. (2006). Smoke-free laws and secondhand smoke exposure in U. S. nonsmoking adults, 1999–2002. Tobacco Control, 15, 302–307. Rosenbaum, P. R., & Rubin, D. B. (1985). The bias due to incomplete matching. Biometrics, 41, 103–116. Royston, P. (2005). Multiple imputation of missing values: Update of ice. The Stata Journal, 5, 527–536. Rozen, T. D. (2005). Childhood exposure to second-hand tobacco smoke and the development of cluster headache. Headache, 45, 393–394. Schneider, S., Seibold, B., Schunk, S., Jentzsch, E., PötschkeLanger, M., Dresler, C., . . . Hyland, A. (2008). Exposure to secondhand smoke in Germany: Air contamination due to smoking in German restaurants, bars, and other venues. Nicotine & Tobacco Research, 10, 547–555. Semple, S., Creely, K. S., Naji, A., Miller, B. G., & Ayres, J. G. (2007). Secondhand smoke levels in Scottish pubs: The effect of smoke-free legislation. Tobacco Control, 16, 127–132. Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics. Boston, MA: Allyn and Bacon. U.S. Department of Health and Human Services. (2004). The health consequences of smoking: A report of the Surgeon General. Atlanta, GA: Department of Health and Human Services, Center for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health.

SUCHT, 56 (5) © 2010 Verlag Hans Huber, Hogrefe AG, Bern

U.S. Department of Health and Human Services. (2006). Control of secondhand smoke exposure. Atlanta, GA: Department of Health and Human Services, Center for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. Valente, P., Forastiere, F., Bacosi, A., Cattani, G., Di Carlo, S., Ferri, M., . . . Zuccaro, P. (2007). Exposure to fine and ultrafine particles from secondhand smoke in public places before and after the smoking ban, Italy 2005. Tobacco Control, 16, 312–317. Wakefield, M. A., Chaloupka, F. J., Kaufman, N. J., Orleans, C. T., Barker, D. C., & Ruel, E. E. (2000). Effect of restrictions on smoking at home, at school, and in public places on teenage smoking: Cross sectional study. British Medical Journal, 321, 333–337. Watzl, H., Rist, F., Höcker, W., & Miehle, K. (1991). Entwicklung eines Fragebogens zur Erfassung von Medikamentenmißbrauch bei Suchtpatienten [Development of a questionnaire for abuse of prescription drugs in addicts]. In M. Heide & H. Lieb (Hrsg.), Sucht und Psychosomatik. Beiträge des 3. Heidelberger Kongresses (S. 123–139). Bonn: Nagel. World Health Organization. (2005). Framework convention on Tobacco Control. Geneva, Switzerland: World Health Organization.

Eingereicht: 11.06.2010 Angenommen: 17.09.2010

Stefanie Müller IFT Institut für Therapieforschung Parzivalstraße 25 D-80804 München Germany Tel. +49 89 3-608-0432 Fax +49 89 3-608-0449 E-Mail: [email protected]