http://orcid.org/0000-0002-5610-3098
Abstract
Aims: Smoking is associated with significant health and economic burden globally, including an increased risk of many leading causes of mortality and significant impairments in work productivity. This burden is attenuated by successful tobacco cessation, including reduced risk of disease and improved productivity. The current study aimed to show the benefits of smoking cessation for workplace productivity and decreased costs associated with loss of work impairment.
Materials and methods: The data source was the 2011 Japan National Health and Wellness Survey (n = 30,000). Respondents aged 20–64 were used in the analyses (n = 23,738) and were categorized into: current smokers, former smokers, and never smokers. Generalized linear models controlling for demographics and health characteristics examined the relationship of smoking status with the Work Productivity and Activity Impairment questionnaire (WPAI-GH) endpoints, as well as estimated indirect costs.
Results: Current smokers reported the greatest overall work impairment, including absenteeism (i.e. work time missed) and presenteeism (i.e. impairment while at work); however, after controlling for covariates, there were no significant differences between former smokers and never smokers on overall work impairment. Current smokers and former smokers had greater activity impairment (i.e. impairment in daily activities) than never smokers. Current smokers reported the highest indirect costs (i.e. costs associated with work impairment); however, after controlling for covariates, there were no significant differences between former smokers and never smokers on indirect costs.
Limitations and conclusions: Smoking exerts a large health and economic burden; however, smoking cessation attenuates this burden. The current study provides important further evidence of this association, with former smokers appearing statistically indistinguishable from never smokers in terms of work productivity loss and associated indirect costs among a large representative sample of Japanese workers. This report highlights the workplace benefits of smoking cessation across productivity markers and cost-savings.
Introduction
Smoking is associated with significant health and economic burden globally, including over 5 million deaths annually, 600,000 deaths due to exposure to second hand smoke, and an increased risk of several leading causes of mortalityCitation1,Citation2. Nearly a quarter of the global adult population are current smokersCitation3,Citation4, the majority of which live in low and middle income countriesCitation1,Citation2. The association between tobacco use and morbidity has been well-documented in JapanCitation5,Citation6. Even though smoking prevalence has declined in Japan over the past decade, 19.3% of the population reported being current smokers in 2013 (32.2% of men and 8.2% of women)Citation7, and a recent study estimated that smoking was still associated with nearly half a million annual deaths among adults over 45Citation8.
In addition to its strong association with mortality, tobacco use is associated with significant societal consequences in the form of direct costs (or those incurred from healthcare resource utilization) and indirect costs (or those associated with work productivity loss due to poor health, including work productivity impairment due to health)Citation2. A 2010 report found that the total societal cost of smoking in Japan in 2005 was estimated to be 6,362 billion Yen. This estimate included 2,366 billion in losses in labor due to smoking, 1,768 billion in health costs, 1,560 billion in losses in labor due to smoking time, 476 billion in excessive care costs, and 191 billion in facilities and environmental costsCitation9. Globally, several studies have documented an association between smoking status and increased sick leave and lost work productivityCitation10,Citation11. One German study estimated annual productivity costs to be in excess of €9.6 billionCitation12. A study from China found tobacco use to be associated with higher rates of presenteeism or impairment experienced while at work among 1,500 survey respondents at a petrochemical companyCitation13.
Presenteeism due to smoking can occur for multiple reasons, including a number that have been investigated in Japanese study cohorts. A Japanese study reported that smoking status was associated with increased workplace sleepiness among medical residentsCitation14. In addition, current and former male smokers working in the Japanese manufacturing sector are at a significantly increased risk of occupational injury at work relative to non-smokersCitation15, while smokers in the telecommunication industry were found to be at higher risk of poor mental healthCitation16. More specifically, smoking has been associated with depressive symptomatology among a large sample of male Japanese employeesCitation17.
While smoking results in many negative consequences, smoking cessation is associated with benefits for individuals, employers, and society at large, both globally and in JapanCitation18. Benefits of quitting smoking include decreased risk of disease, as well as reduced direct and indirect costs, including improved workplace productivityCitation19,Citation20. In addition to lower costs to society and employers, quitting can also be associated with reduced costs for individuals, including possible reductions in health and life insurance premiums, lower healthcare costs due to smoking-related illnesses, as well as money saved from no longer purchasing tobacco productsCitation21,Citation22. This is a critical area of need, as evidence suggests that fewer Japanese smokers attempt tobacco cessation compared to those in other countries. Evidence suggests that nearly 43% of adults who smoked in the US attempted to quit in 2012Citation23, considerably higher than results of a Japanese study of nearly 2,000 adult smokers that reported only 23% made an attempt to quit in the past year, and only 13% of those who attempted to quit maintained a 6-month abstinence periodCitation24.
Smoking continues to exert a notable burden in Japan. Therefore, it is critical that research continues to explore the direct and indirect costs associated with smoking and the alleviation of these costs through smoking cessation. Importantly, a paucity of research exists documenting the impact of smoking on work productivity and the benefits of quitting, particularly in certain regions such as Japan. The current study provides novel and unique insight into the association between smoking status and work productivity in Japan, utilizing large-scale representative survey methodology. Furthermore, by examining the work productivity indirect costs incurred by former smokers and never smokers, we can use cross-sectional data to examine the effects of smoking cessation on workplace productivity and work-related costs. More specifically, if we find that former smokers and never smokers show similar patterns of work productivity loss and indirect costs, this would provide evidence that stopping smoking is associated with workplace benefits like those for individuals who have never smoked before.
Methods
Sample
This project includes data from the 2011 National Health and Wellness Survey (NHWS) from Japan (n = 30,000). The NHWS is a self-administered, internet-based questionnaire from a sample of individuals aged 18 or older. A stratified random sample, with strata by gender and age, was implemented to ensure that the demographic composition of the sample is comparable to that of the corresponding adult population based on: data from the 2010 International Data Base of the US Census Bureau for JapanCitation25, which is based on the official Japanese census. Online recruitment is supplemented by offline recruitment to reach the elderly that may not have internet access. Several peer-reviewed publications have previously compared the NHWS with other governmental sourcesCitation26–28. This was an analysis of existing data from 2011 NHWS, and sample size was pre-determined. The sample for the current study only included all respondents aged 20–64 (n = 23,738).
Measures
The primary independent variable for this project was smoking status. Respondents were asked if they ever smoked, if they selected yes then they were asked what their current smoking status was. Respondents were then categorized into the following groups based on their responses to the two smoking questions: never smoked, former smoker, and current smoker.
Demographics and health characteristics
The following demographic and health characteristic variables were examined for differences between groups: for age, sex, education, income, body mass index (BMI) category, alcohol use, and comorbidity burden (measured via the Charlson Comorbidity Index (CCICitation29)).
Work productivity and activity impairment
Work productivity loss and activity impairment were measured via the Work Productivity and Activity Impairment-General Health scale (WPAI-GHCitation30). This instrument provides data for absenteeism (the percentage of work time missed because of one’s health problems in the past 7 days), presenteeism (the percentage of impairment experienced while at work in the past 7 days because of one’s health problems), and overall work productivity loss (an overall impairment estimate that is a combination of absenteeism and presenteeism). Only respondents who reported being full-time or part-time employed (via the labor force participation question) provided data for absenteeism, presenteeism, and overall work impairment. Activity impairment (the percentage of impairment in daily activities because of one’s health problems in the past 7 days) was collected for all respondents.
Indirect costs
In this article, indirect costs were defined as the monetary value of the above-mentioned work productivity loss. Estimated costs were calculated by using estimated overall wages/salaries in Japanese Yen for each respondent from the Ministry of Health, Labor, and WelfareCitation31. The number of hours missed in the last week because of one’s health (absenteeism) and the number of hours of productivity loss (e.g. 20% presenteeism for 50 h worked =10 h productivity loss) missed in the last week because of health impairment while at work (presenteeism) were each multiplied by the hourly wage rates to arrive at total lost wages per respondent for 1 week. Costs were then annualized by multiplying weekly estimates by 50 (work weeks in a year). Presenteeism costs and absenteeism costs were then added to calculate indirect costs.
Analyses
All analyses were carried out using SPSS version 20. Demographic and health characteristic differences were examined by smoking status. Means and standard deviations are reported for continuous variables, and frequencies and percentages are reported for categorical variables. For categorical variables, Chi-square tests were used to determine significant differences, while one-way ANOVAs were used for continuous variables when comparing the smoking groups. Following initial comparisons, generalized linear models were used to assess the association of smoking status with work productivity and activity impairment after controlling for covariates. Covariates included: age, gender, education, income, BMI category, alcohol use, and CCI. Adjusted means, standard errors, and 95% confidence intervals are reported. P-values are considered statistically significant if p < .05. (for unadjusted results, see Appendix 1).
Results
Initial Group comparisons: demographics and health characteristics
There were significant differences among smoking status groups for all demographic and health characteristic variables: age, sex, education, annual household income, BMI, alcohol use, and CCI (see ). Current smokers were older than never smokers, but younger than former smokers. Current smokers had the greatest proportion of males, lowest proportion of those with college education, greatest proportion of alcohol drinkers, and greatest proportion of those on the labor force compared with never and former smokers. Never smokers had the greatest proportion of women, greatest proportion of those with college education, smallest proportion of those who drink alcohol, and fewer respondents participating in the labor force.
Table 1. Demographics and health characteristics by smoking status.
Multivariable analyses: WPAI controlling for age, gender, income, education, BMI category, alcohol use, and CCI
After controlling for covariates, never smokers and former smokers did not differ on work productivity loss, but former smokers had 6% (p = .011) greater activity impairment than never smokers (see ). Additionally, current smokers had 15% (p = .114) greater absenteeism, 7% (p = .012) greater presenteeism, 7% (p = .008) greater overall work impairment, and 10% (p < .001) greater activity impairment than never smokers. Current smokers had greater presenteeism (7%) (p = .026) and overall work impairment (7%) (p = .030) than former smokers. There were no significant differences across groups for absenteeism.
Table 2. Adjusted means examining effect of smoking status on WPAI-GH and costs after controlling for demographic and health characteristics.
Regarding costs, never smokers and former smokers did not differ on work productivity loss costs after controlling for covariates (see ). However, current smokers had significantly greater presenteeism costs than both never smokers (12%) (p < .001) and former smokers (10%) (p = .007) and greater indirect costs than never smokers (12%) (p < .001) and former smokers (10%) (p = .012). There were no significant differences across groups for absenteeism costs. (Note: for unadjusted and adjusted cost estimates that account for age and gender, see Appendix 2).
Discussion
Smoking cessation can not only lessen the risk of many of the leading causes of death, but, for employers, it has also been associated with improved workplace performance and reduced indirect costs. The current study provides important insight into this association among a large representative sample of Japanese workers. Results suggest that former smokers appear statistically indistinguishable from never smokers in terms of their work productivity loss and the associated indirect costs. This finding is supported by a recent study in which no differences were found for absenteeism between former smokers and never smokers in China, the US, and Europe, and no differences between former and never smokers for presenteeism and overall work impairment in ChinaCitation32. Given the improvement after quitting smoking in work productivity loss and associated costs, smoking cessation programs may be important considerations for the workplace setting.
As a group, current smokers tended to drink alcohol, have lower rates of college education, and be part of the labor force, compared to never smokers. After controlling for confounding factors, current smokers had significantly higher rates of activity impairment, overall work impairment and presenteeism, confirming the significant burden tobacco use exerts among workers. Current smoking status was also associated with higher presenteeism costs and indirect costs. In contrast, former and never smokers endorsed similarly low rates of work and activity impairment, providing additional evidence of the benefits that smoking cessation can potentially bring both employers and employees alike.
Previous research has found higher absenteeism for current smokers compared with never smokers and former smokers in the US, Europe, and ChinaCitation32. In contrast to previous studies, there were no significant differences across smoking status groups for absenteeism and absenteeism costs after controlling for confounding factors. It is possible that this finding may be the result of the study’s cultural context, in which it may be less socially acceptable in Japan for employees to take time off, and, thus, the impact of smoking status was attenuatedCitation33,Citation34. This remains a potential line of future research and requires further investigation.
Limitations and future directions
The results provide novel insight into the impact of smoking status on work productivity among Japanese employees; however, it is important to note study limitations. First, to participate in an online survey, respondents need to be relatively healthy. Therefore, our sample may over-represent healthier individuals, therefore under-reporting the indirect costs associated with smoking. This may impact how representative our survey population is of the smoking population in Japan.
Second, work productivity estimates are self-reported and, therefore, may be subject to recall bias. Respondents may have over- or under-reported these estimates, for example. However, the impact of recall bias on our data may be relatively minimal due to the short timeframe (past 7 days) in which respondents are asked to recall this information. Health characteristics such as smoking status reported by participants also are not independently verified. Future research should corroborate self-reported smoking status data with physician- and/or employer-reported data to verify our findings. In addition, indirect costs were estimated based on median wages by age and gender, but do not reflect individual respondents’ occupations. Future research could include a calculation of indirect costs that reflects respondents’ occupations.
Third, whereas this report provides important preliminary evidence of a potential longitudinal association between tobacco use and work productivity, the cross-sectional nature of the study methodology precludes causal inferences from being drawn. Moreover, some proportion of quitters in the study may have relapsed, but we did not analyze the duration of time since quitting. Igarashi et al.Citation35 found that a longer duration of time since quitting smoking is associated with greater health benefits. However, we can conclude that quitting smoking is associated with workplace benefits, given that former and never smokers are statistically indistinguishable in terms of their work productivity loss and costs incurred. Previous research by Wacker et al.Citation43 also found a strong association between smoking status and indirect costs, with current smokers incurring the highest indirect costs. Despite this limitation due to the cross-sectional nature of the data, additional research is warranted to explore this link over time. Bockerman et al.Citation36 examined the long-term association between smoking and labor market outcomes. They found that smoking predicted lower lifetime earnings. Additionally, Baker et al.Citation32 examined the benefits of quitting smoking within a US sample. They found that former smokers who quit between 0–4 years prior showed lower absenteeism, presenteeism, and activity impairment. Future studies should more thoroughly investigate the length of time respondents quit smoking and frequency of current and/or past smoking on productivity using a longitudinal framework.
Smoking cessation programs, including pharmacotherapy and behavioral interventions, have been found to both promote a healthy workforce and to be cost-effective when implemented in work-place settingsCitation37–40. A recent Japanese simulation-based study encouraged implementation of pharmacotherapy interventions for smoking cessation, stating that such programming could promote greater success among quitters, and would be associated with cost-savings once indirect and direct costs were consideredCitation35. Further, a randomized controlled trial of a smoking cessation counseling intervention at a worksite in Japan reported positive results, reducing smoking rates by 8.4%Citation41. Unfortunately, despite this evidence base, the majority of employers globally fail to cover tobacco cessation programs in the workplaceCitation42. Few individuals successfully quit, and further supportive programming is needed, particularly in the workplace setting. For example, smoking cessation programs in the workplace may allow easier access to information about smoking cessation and treatment. The potential for the workplace to provide the ideal setting for ongoing cessation programming should be fully explored, particularly given that benefits are found across health, work productivity, and cost outcomes.
Conclusions
Smoking and tobacco use exerts a tremendous health and economic burden throughout the world. Evidence demonstrates that successful tobacco cessation attenuates this burden, including reducing the risk of disease, improving productivity, and decreasing associated costs. The burden of illness is significantly reduced; in fact, in terms of work productivity loss and associated costs, former smokers appear indistinguishable compared with never smokers. These findings highlight the workplace benefits of smoking cessation.
Transparency
Declaration of funding
This study was funded by Pfizer Inc.
Declaration of financial interests/other interests
KS and RY are employees and stockholders of Pfizer Inc. NMF and JV were paid employees of Kantar Health at the time of conducting and reporting upon this study. Kantar Health received funding from Pfizer for conducting this study and the development of the manuscript. AI and RG received funding from Pfizer Inc. for other research. JME peer reviewers on this manuscript have no relevant financial or other relationships to disclose.
Supplemental material
Download Zip (46.6 KB)Acknowledgments
Errol J. Philip provided editorial support for manuscript preparation as a paid consultant to Kantar Health.
References
- World Health Organization. Tobacco; 2015. http://www.who.int/mediacentre/factsheets/fs339/en/. Accessed April 2017
- US Department of Health and Human Services (USDHHS). The health consequences of smoking—50 years of progress. A report of the surgeon general Atlanta, GA: US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health; 2014. http://www.surgeongeneral.gov/library/reports/50-years-of-progress/. Accessed April 2017
- World Health Organization. Tobacco control country profiles; 2014. http://www.who.int/tobacco/surveillance/policy/country_profile/en/. Accessed April 2017
- Jamal A, Homa DM, O’Connor E, et al. Current cigarette amoking Among Adults — United States, 2005–2014. Centers for Disease Control and Prevention; 2015. https://www.cdc.gov/mmwr/preview/mmwrhtml/mm6444a2.htm. Accessed April 2017
- Shibuya K. Change in mortality and years of life lost attributable to tobacco in Japan, 1985 and 1995. Asia Pac J Public Health 1999;11:65-70
- Shibuya K. [Recent trends in mortality, years of life lost (YLLs), and disability-adjusted life years (DALYs) attributable to tobacco in Japan]. Nihon Eiseigaku Zasshi 2001;56:484-91
- Ministry of Health Labour and Welfare. 2013 National Health and Nutrition Examination Survey (in Japanese). Tokyo: Ministry of Health Labour and Welfare, 2015
- Zheng W, McLerran DF, Rolland BA, et al. Burden of total and cause-specific mortality related to tobacco smoking among adults aged >/= 45 years in Asia: a pooled analysis of 21 cohorts. PLoS Med 2014;11:e1001631
- Institute for Health Economics and Policy. Research on smoking cessation policy “Cost Estimation of Spending” report; Tokyo, Japan: IHEP; 2010
- Robroek SJ, van den Berg TI, Plat JF, et al. The role of obesity and lifestyle behaviours in a productive workforce. Occup Environ Med 2011;68:134-9
- Wegner C, Gutsch A, Hessel F, et al. [Smoking-attributable productivity loss in Germany—a partial sickness cost study based on the human capital potential method]. Gesundheitswesen 2004;66:423-32
- Prenzler A, Mittendorf T, von der Schulenburg JM. [Modelling of the costs of productivity losses due to smoking in Germany for the year 2005]. Gesundheitswesen 2007;69:635-43
- Yu J, Wang S, Yu X. Health risk factors associated with presenteeism in a Chinese enterprise. Occup Med (Lond) 2015;65:732-738
- Wada K, Sakata Y, Theriault G, et al. Associations of excessive sleepiness on duty with sleeping hours and number of days of overnight work among medical residents in Japan. J Occup Health 2007;49:523-7
- Nakata A, Ikeda T, Takahashi M, et al. Non-fatal occupational injury among active and passive smokers in small- and medium-scale manufacturing enterprises in Japan. Soc Sci Med 2006;63:2452-63
- Suwazono Y, Okubo Y, Kobayashi E, et al. A follow-up study on the association of working conditions and lifestyles with the development of (perceived) mental symptoms in workers of a telecommunication enterprise. Occup Med (Lond) 2003;53:436-42
- Ikeda T, Nakata A, Takahashi M, et al. Correlates of depressive symptoms among workers in small- and medium-scale manufacturing enterprises in Japan. J Occup Health 2009;51:26-37
- Avila-Tang E, Apelberg BJ, Yamaguchi N, et al. Modelling the health benefits of smoking cessation in Japan. Tob Control 2009;18:10-17
- U.S. Department of Health and Human Services. How Tobacco Smoke Causes Disease: The Biology and Behavioral Basis for Smoking-Attributable Disease: A Report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 2010
- Rasmussen SR, Prescott E, Sorensen TI, et al. The total lifetime health cost savings of smoking cessation to society. Eur J Public Health 2005;15:601-6
- American Cancer Society. Guide to quitting smoking; 2014. http://www.cancer.org/healthy/stayawayfromtobacco/guidetoquittingsmoking/guide-to-quitting-smoking-toc. Accessed April 2017
- American Cancer Society. When smokers quit – what are the benefits over time?; 2014. http://www.cancer.org/healthy/stayawayfromtobacco/guidetoquittingsmoking/guide-to-quitting-smoking-benefits. Accessed October 2014
- CDC. Quitting Smoking. 2016. https://www.cdc.gov/tobacco/data_statistics/fact_sheets/cessation/quitting/index.htm. Accessed April 2017
- Hagimoto A, Nakamura M, Morita T, et al. Smoking cessation patterns and predictors of quitting smoking among the Japanese general population: a 1-year follow-up study. Addiction 2010;105:164-73
- U.S. Census Bureau. International data base of the US Census Bureau for Japan. U.S. Census Bureau; 2010. https://www.census.gov/population/international/data/idb/informationGateway.php. Accessed April 2017
- Bolge SC, Doan JF, Kannan H, et al. Association of insomnia with quality of life, work productivity, and activity impairment. Qual Life Res 2009;18:415-22
- DiBonaventura MD, Wagner JS, Yuan Y, et al. Humanistic and economic impacts of hepatitis C infection in the United States. J Med Econ 2010;13:709-18
- Finkelstein EA, Allaire BT, Dibonaventura MD, et al. Direct and indirect costs and potential cost savings of laparoscopic adjustable gastric banding among obese patients with diabetes. J Occup Environ Med 2011;53:1025-9
- Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373-83
- Reilly MC, Zbrozek AS, Dukes EM. The validity and reproducibility of a work productivity and activity impairment instrument. Pharmacoeconomics 1993;4:353-65
- Ministry of Health Labour and Welfare. Wage and labour welfare statistics: office employment, basic survey on wage structure. Tokyo: Ministry of Health, Labour, and Welfare; 2014 [in Japanese]
- Baker CL, Flores NM, Zou KH, et al. Benefits of quitting smoking on work productivity and activity impairment in the United States, the European Union, and China. Int J Clinical Practice 2017;71:e12900
- Ministry of Land Infrastructure Transport and Tourism. Transport and tourism, Japan tourism agency, working team for well-distributed vacation time. Tokyo: Ministry of Land, Infrastructure, Transport and Tourism; 2010 [in Japanese]
- Ministry of Health Labour and Welfare. Wage and labour welfare statistics office employment, wage and labour welfare statistics division: general survey on working conditions. Tokyo: Ministry of Health, Labour and Welfare, 2010 [in Japanese]
- Igarashi A, Goto R, Suwa K, et al. Cost-effectiveness analysis of smoking cessation interventions in Japan using a discrete-event simulation. Appl Health Econ Health Policy 2016;14:77-87
- Böckerman P, Hyytinen A, Kapiro J. Smoking and long-term labour market outcomes. Tobacco Control 2015;24:348-53
- Centers for Disease Control and Prevention. Tobacco-use cessation. Workplace health promotion. CDCP: 2013. https://www.cdc.gov/workplacehealthpromotion/tools-resources/workplace-health/tobacco-use-cessation.html. Accessed April 2017
- Halpern MT, Shikiar R, Rentz AM, et al. Impact of smoking status on workplace absenteeism and productivity. Tob Control 2001;10:233-8
- Fitch K, Iwasaki K, Pyenson B. Covering smoking cessation as a health benefit: a case for employers; 2006. http://www.dfwbgh.org/events07/9-27-2007.pdf. Accessed April 2017
- Jackson KC II, Nahoopii R, Said Q, et al. An employer-based cost-benefit analysis of a novel pharmacotherapy agent for smoking cessation. J Occup Environ Med 2007;49:453-60
- Kadowaki T, Watanabe M, Okayama A, et al. Effectiveness of smoking-cessation intervention in all of the smokers at a worksite in Japan. Ind Health 2000;38:396-403
- The Henry J. Kaiser Family Foundation. 2013. Employer Health Benefits Survey; 2013 http://www.kff.org/private-insurance/report/2013-employer-health-benefits/. Accessed April 2017
- Wacker M, Holle, R, Heinrich J, et al. The association of smoking status with healthcare utilisation, productivity loss and resulting costs: results from the population-based KORA F4 study. BMC Health Serv Res 2013;13:278
Appendix 1
Table A1. Percent work productivity and activity impairment by smoking status.
Table A2. Annual indirect costs (Japanese Yen in thousands) overall for those in the labor force by smoking status.
Appendix 2
Table A3. Annual indirect costs (Japanese Yen in thousands), taking age and gender into account for those in the labor force by smoking status.
Table A4. Annual indirect costs (Japanese Yen in thousands), taking age and gender into account, for those in the labor force by smoking status, after controlling for demographic and health characteristics.