Effectiveness of Smoking-Cessation Intervention in All of the Smokers at a Worksite in Japan

Transcription

1 Industrial Health 2000, 38, Original Article Effectiveness of Smoking-Cessation Intervention in All of the Smokers at a Worksite in Japan Takashi KADOWAKI 1 *, Makoto WATANABE 1, Akira OKAYAMA 2, Kayoko HISHIDA 3 and Hirotsugu UESHIMA 1 1 Department of Health Science, Shiga University of Medical Science 2 Department of Hygiene and Preventive Medicine, Iwate Medical University 3 Health Information Total Service Co., Ltd. Received June 16, 2000 and accepted September 8, 2000 Abstract: In Japan, the prevalence of smoking among males and females was 56.1% and 14.2%, respectively, in Male smoking prevalence was exceedingly high as compared to those in other industrialized countries. We conducted a randomized controlled intervention study on smoking cessation for all smokers in a worksite regardless of their willingness to quit smoking. All of the male smokers in a radiator manufacturing factory (n=263) were randomly allocated to an intervention group (n=132) or a control group (n=131). Subjects in the intervention group received individual counseling by a doctor, and those who signed a Smoking Cessation Declaration underwent a fivemonth intervention. Subjects in the control group received equivalent delayed intervention for four months. The cessation rate after the original intervention was 12.9% (17/132) and 3.1% (4/131) in the intervention and control groups, respectively (p=0.003). Among those who once succeeded in quitting, 48.6% (18/37) maintained cessation at the long-term survey. Overall, the cessation rate was 8.4% (22/263) and the prevalence of smoking among males significantly decreased from 62.9 to 56.7% (p=0.038). As a conclusion, intervention in all smokers at a worksite regardless of their willingness to quit is effective and impacts the overall smoking rate. Key words: Health education, Intervention study, Occupational setting, Randomized controlled trial, Smoking cessation, Worksite Introduction Tobacco use is an important cause of premature mortality and disability worldwide 1), and a serious problem particularly in developing countries, whose smoking prevalence has been increasing 1). Smoking is expected to kill 10 million people annually worldwide by 2030, and over 70% of these deaths will be in the developing world 2). China has followed a similar pattern of male cigarette smoking to that in the United States, with a delay of about 40 years 3, 4). The same problem holds true to Japan; the prevalences of smoking were 56.1% in men and 14.5% in women in ). Accordingly, the lung cancer mortality rate has been increasing and this is *To whom correspondence should be addressed. now the leading cause of death among malignant neoplasms in Japanese males, outweighing stomach cancer 5), while in the US, the lung cancer mortality rate is beginning to decrease in accordance with a decrease in the prevalence of smoking 6). In such a situation, an integrated strategy to reduce the prevalence of smoking is urgently necessary. There have been several reports on smoking intervention from Japan 7, 8), but they have chiefly been on a voluntary basis with a limited sample size. We should focus not only on smokers willing to quit smoking, but all the smokers regardless of their willingness to quit. In order to determine the effectiveness of intervention in all smokers regardless of their willingness to quit smoking, we conducted a randomized controlled intervention trial in an occupational setting.

2 INTERVENTION IN ALL SMOKERS IN A WORKSITE Subjects and Methods Study population We performed this intervention study in a radiator manufacturing factory that employed 542 (423 male and 119 female) workers in All the survey and intervention procedures were examined and approved by the Shiga University of Medical Science Ethics Review Board. We conducted a self-administered questionnaire survey along with a written informed consent, with a response rate of 100%. The questionnaire consisted of questions regarding their smoking status (current smoker/ex-smoker/non-smoker, number of cigarettes per day), history (years smoked, previous attempts at quitting, etc.), degree of nicotine addiction (Fagerstrom Tolerance Questionnaire 9) or FTQ), willingness to quit (not at all/in the future/before long/ currently trying to quit smoking), and how they feel about their smoking environment in the worksite. Through this baseline survey, 267 smokers (263 male and 4 female) were identified. We conducted the same procedure of intervention to all smokers, but female smokers were excluded from the analysis due to the small number of smokers. We randomly allocated the male smokers into an intervention group (n=132) and a control group (n=131) using a random numbergenerating software. Smoking cessation intervention The intervention group underwent five months of intervention beginning in January, First, we approached all of the smokers in the intervention group regardless of their willingness to quit. Each smoker was asked to visit the nurse s office to attend a face-to-face counseling session with a doctor during working hours. The doctor discussed the harmful effects of tobacco and the benefits of cessation, and encouraged each smoker to quit. Smokers underwent a measurement of carbon monoxide (CO) concentration in expiratory air with a Smokerlyzer 10, 11) during the counseling. If a smoker expressed that he would try to quit, he was advised to decide when he would start smoking cessation and signed a Smoking Cessation Declaration along with when to start in order to express his determination. Then the doctor additionally explained situations where he might want to re-start smoking, and how to deal with such situations. Active intervention continued for five months from when the subject signed the Declaration. For those who did not sign the Declaration, we surveyed their smoking status without active intervention. The five-month program included periodic visits to work stations by the doctor and an occupational nurse for 397 encouragement and leaflet distribution, and group discussions. We also conducted A Smoking Cessation Marathon program beginning in the fourth months of the intervention. Participants were encouraged to try to abstain from smoking for 42 days in order to seize an opportunity of complete cessation. This was conducted for the purpose of giving re-smokers chances to challenge quitting again, exploring new challengers, and letting those who had quit maintain smoking cessation. During the five-month intervention, we used the materials (counseling material, pamphlets and follow-up leaflets) in Individual Health Education Manual on Smoking Cessation 12) which was originally developed by a Ministry of Health and Welfare Study Group on Health Education Material Development for Preventing Lifestyle Related Diseases in We additionally distributed a stamp sheet to each of the Smoking Cessation Marathon participant to let him confirm daily cessation. After five months, we conducted the same selfadministered questionnaire survey in all of the workers to identify those who succeeded in smoking cessation (response rate: 94.1%). The smoking status of nonrespondents was confirmed individually by the occupational nurse. Successful smoking cessation was defined as self-reporting of quitting for more than one month and an expiratory CO concentration of less than 9 ppm. Those who succeeded in smoking cessation in both groups were asked to visit the nurse s office to undergo an expiratory CO concentration. Delayed intervention For the purpose of offering each smoker an opportunity to consider smoking cessation, we made an equivalent intervention in the control group for four months after the initial five-month intervention was completed. During this delayed intervention period, the original intervention group was left without any active intervention. Observation of long-term effects To observe the long-term sustained cessation rate (more than 12 months in the intervention group and 8 months in the control group) and to determine the impact of our intervention on the smoking rate in the entire baseline population, we confirmed the smoking status of all of the subjects at an annual health check-up in June Each subject underwent an expiratory CO concentration measurement to confirm one s smoking status. Those who had continued smoking cessation since our intervention additionally underwent a urine test of nicotine metabolites which remain longer than lung CO (NicCheck 13, 14) ).

3 398 T KADOWAKI et al. Statistical analysis We used tests on the equality of proportion, Fisher s exact tests for categorical variables, and t-tests or analysis of variance (ANOVA) for continuous variables. As to the willingness to quit smoking, we divided the respondents into a negative group (not at all/in the future) and a positive group (before long/currently trying to quit smoking). Non-respondents are categorized into negative group. As to the evaluation of nicotine dependence, we classified the FTQ score into low (score 0 3), medium (score 4 6) and high (score 7 ) nicotine dependence. Non-respondents and incomplete respondents to the FTQ questions were excluded from the FTQ analysis. In order to compare those who succeeded in smoking cessation and who did not after the delayed intervention, we combined the two groups and divided them into succeed and non-succeed groups. Those who were lost to followup were included in non-succeed group. Those who initially answered to be non-smokers but identified themselves as originally habitual smokers (n=3) were categorized as current smokers, though we did not conduct any intervention approach for them. We assumed that those who were lost to follow-up (retired, moved to some other branch, or left the company for some reason) did not change their initial smoking habit. If a subject was originally a habitual smoker and quit once during the intervention period before leaving the factory, we considered that he restarted habitual smoking after he left. However, in calculating a smoking prevalence, we included only those who remained there in order to show the actual smoking rate of those in the same setting at each step. Results Baseline survey The smoking prevalence was 62.9% and 3.4% among males and females, respectively. As described above, we excluded female smokers from our analysis due to the small number of subjects (n=4). The mean age of ex-smokers was significantly higher than those of other categories (p=0.003) (Table 1). Comparison between the intervention and control groups The two groups showed no significant difference with regard to the number of subjects, mean age, beginning age of habitual smoking, number of cigarettes consumed per day, FTQ score, and proportion of those who had any previous experience with smoking cessation (Table 2). Intervention Each subject was first asked to attend an individual counseling session. Among the 132 intervention subjects, 125 (94.7%) received the counseling, and 63 (47.7%) signed a Smoking Cessation Declaration. The counseling took significantly longer time for those who signed the Declaration (16.0 ± 4.1 minutes) than those who did not (13.0 ± 5.4 minutes) (p=0.0005). Among those who signed the Declaration, 71.4% (n=45) and 28.6% (n=18) endured to abstain from smoking for one day and one week, respectively. As for the Smoking Cessation Marathon, 19 subjects (4 subjects who had quit during our intervention, 14 re-smokers, and 1 new challenger) participated in the program. Intervention results The cessation rate at the five-month survey was 12.9% and 3.1% in the intervention and control groups, respectively (p=0.003), confirmed by the questionnaire (response rate: Table 1. Smoking status and age distribution of the male workers at baseline survey, a worksite in Shiga, Japan, 1997 Smoker Ex-smoker Never-smoker Total n (%) 263 (62.2) 59 (13.9) 101 (23.9) 423 (100.0) Mean age* (SD + ) 33.3 (9.6) 38.2 (9.2) 34.1 (9.8) 34.2 (9.7) Age distribution n n n n Age *p=0.003 in analysis of variance. + Standard deviation. Industrial Health 2000, 38,

4 INTERVENTION IN ALL SMOKERS IN A WORKSITE 399 Table 2. Comparison of baseline characteristics between the intervention and control groups after the random allocation, male, a worksite in Shiga, Japan, 1997 Intervention group Control group n=132 n=131 P-value Mean age (SD + ) 33.5 (10.1) 33.2 (9.2) Mean age of beginning habitual smoking (SD) 18.8 (3.8) 18.6 (2.3) Mean no. of cigarettes per day (SD) 19.3 (7.2) 19.6 (6.5) Mean FTQ ++ score (SD) 4.3 (1.9) 4.6 (2.0) No. of those with previous quit attempts (%) 59 (44.7) 59 (45.0) No. of those who always inhale the smoke deeply (%) 65 (49.2) 72 (55.0) Willingness to quit at baseline (%) Negative (Not at all / In the future) 124 (93.9) 122 (93.1) Positive (Before long / Currently trying) 8 (6.1) 9 (6.9) + Standard deviation. ++ Fagerstrom Tolerance Questionnaire. Table 3. Comparison between the intervention and control groups before and after the five month intervention period regarding smoking cessation rate, changes in mean no. of cigarettes per day, mean FTQ score, and willingness to quit smoking, male, a worksite in Shiga, Japan, 1997 Intervention group Control group Cessation rate* (%) (cessation/subjects) 12.9 (17/132) 3.1 (4/131) (p=0.003) Cessation rate by nicotine dependence measured by FTQ + score* (%) (cessation/subjects) Low (0 3) 18.6 (8/43) 7.3 (3/41) (p=0.125) Medium (4 6) 12.3 (8/65) 2.2 (1/46) (p=0.055) High (7 -) 5.0 (1/20) 0.0 (0/19) (p=0.323) Proportion of those who quit/reduced cigarette 30.3 (40/132) 17.5 (23/131) (p=0.016) consumption per day (%) Willingness to quit (%) baseline at five month p-value baseline at five month p-value Negative (Not at all/in the future)*** Positive (Before long/currently trying) Succeeded in cessation *Success in cessation means that a person had been completely abstaining from smoking for more than 4 weeks and that the expiratory CO concentration was less than 9 ppm at the five-month follow-up survey. Those who were lost-to-follow-up were considered not to change their original smoking habit. **Non/incomplete-respondents to FTQ questions were excluded from this analysis. ***Nonrespondents were categorized into the negative group. + Fagerstrom Tolerance Questionnaire. 93.4%) and the expiratory CO measurement of less than 9 ppm (Table 3). All the non-respondents were individually clarified of their smoking status by the occupational nurse. The cessation rates of the Smoking Marathon participants were 42.1% among all the 19 participants including those who had already quit (8/19), and 26.7% among those who smoked at the beginning of the Smoking Marathon program (4/15). At the five month survey, the proportion of those who reduced the number of cigarette consumption (including those who quit) was significantly larger in the intervention group than the control group (p=0.016) (Table 3). Those with lower FTQ score showed higher cessation rate in both groups, but in the medium FTQ score group the cessation rate showed higher tendency in the intervention group (p=0.055) (Table 3). As for the willingness to quit smoking, the proportion of those who expressed negative willingness significantly decreased in the intervention group, while the distribution didn t show a significant change in the control group (Table 3). Delayed intervention As to the delayed intervention for the original control group (n=131), there were 123 smokers left after excluding those

5 400 T KADOWAKI et al. Fig. 1. Flow chart of the whole study, the number of smoking subjects in the rectangles, and the number of those who were away from smoking at each stage in the ovals. All the male smokers (n=263) were divided into an intervention group (n=132) and a control group (n=131). Among those in the intervention group, 125 subjects received an individual counseling and 63 subjects signed on the Smoking Cessation Declaration. There were 17 and 4 subjects who succeeded in smoking cessation in the intervention and the control group respectively at the five-month survey. At the end of the delayed intervention for the control group, 19 subjects in the intervention group continued cessation, and 16 subjects in the control group quit smoking. In the long term effect, 22 subjects continued smoking cessation for one year (intervention group) or eight months (control group). who were lost to follow-up (n=4) and the subjects who had already quit smoking (n=4). After undergoing the four-month intervention equivalent to the original intervention, 16 subjects (13.0%) succeeded in smoking cessation at the end of the delayed intervention (Fig. 1). Comparison of successful and unsuccessful subjects A comparison of subjects in both groups who succeeded to those who did not succeed at quitting revealed that smokers were more likely to abstain from smoking if the number of cigarettes consumed per day was smaller and if the beginning age of habitual smoking was higher (Table 4). The proportion of those who signed the Smoking Cessation Declaration was significantly higher in the success group. Among those who had no intention of cessation at baseline (n=60) and signed the Declaration at the counseling session (n=19), the quit rate was 15.8% (n=3) (not shown in tables). As for age, there was no significant relationship between the age group and the smoking cessation rate through our intervention (not shown in tables). As for willingness to quit, there was a linear tendency between the willingness to quit and the cessation rate; 6.9% of those who had no intention to quit smoking succeeded in quitting, and the cessation rate significantly increased as the willingness became favorable toward cessation (p=0.002) (Table 5). Observation of long-term effects At the survey 18 months after the beginning of this study and eight months after completing the intervention in both groups, the overall cessation rate was 8.4%. In Figure 1, the numbers in ovals show the number of those who newly quit smoking or continued cessation at each survey. The cessation rate at the 18th month survey among those who had once succeeded in quitting smoking affected by our Industrial Health 2000, 38,

6 INTERVENTION IN ALL SMOKERS IN A WORKSITE 401 Table 4. Comparison between those who succeeded and non-succeeded in smoking cessation through our intervention, male, a worksite in Shiga, Japan, 1997 Succeeded Non-succeeded p-value n=37* n=226** Mean age at baseline (SD + ) 33.7 (9.3) 33.3 (9.7) Mean age of beginning habitual smoking (SD) 20.7 (5.4) 18.4 (2.5) <0.001 Mean no. of cigarettes per day (SD) 16.8 (7.6) 19.9 (6.6) Mean FTQ ++ score (SD) 3.9 (2.1) 4.6 (1.9) No. of those with previous quit attempts (%) 20 (54.1) 98 (43.4) No. of those who always inhale the smoke deeply (%) 17 (45.9) 120 (53.1) No. of those who declared to quit (%) 25 (67.6) 72 (31.9) <0.001 *Including subjects in both groups who succeeded in smoking cessation during the initial five months intervention and those in the control group who succeeded in smoking cessation during the following delayed intervention. **Including those who were lost-to-follow-up. + Standard deviation. ++ Fagerstrom Tolerance Questionnaire. Table 5. Distribution of the subjects according to the willingness to quit smoking at baseline and the cessation rate of each category, male, a worksite in Shiga, Japan, 1997 Willingness to quit at baseline Not at all In the future Before long Currently trying n* (Distribution among all subjects (%)) (49.8) (43.7) (5.0) (1.5) Succeeded in cessation** (Cessation rate in each category (%)) (6.9) (20.0) (23.1) (50.0) *Including those who were lost to follow-up. **Including subjects in both groups who succeeded in smoking cessation during the initial five months intervention and those in the control group who succeeded in smoking cessation during the following delayed intervention. p=0.002 in Fisher s exact test. intervention (those who quit during the original intervention period in both groups and during the delayed intervention period in the control group) (n=37) was 48.6% (n=18). The prevalence of smoking among males changed from 62.9% to 56.7% and the reduction was significant in one-sided test (p=0.038). In the comparison between those who continued and discontinued smoking cessation after once quit through our intervention, there was no significant difference in age, age of beginning habitual smoking, number of cigarettes per day, FTQ score, previous quit attempts, deep inhaling of smoke, and signing the Declaration. Discussion We clarified the effectiveness of smoking cessation intervention in all smokers at a worksite regardless of their willingness to quit through a randomized controlled trial. The smoking cessation rate was 12.9% in the intervention group at five months, and 8.4% of the subjects quit smoking in the long term. Among those who once quit smoking, 48.6% sustained cessation. Considering the study design, which aimed at all of the smokers instead of only volunteers, our result strongly suggests that approaching all of the smokers in an occupational setting is effective, considering that 13% is thought to be a reasonable benchmark for assessing the effectiveness of smoking cessation programs for voluntary participants 15). As an overall effect, the prevalence of smoking among males significantly decreased from 62.9% to 56.7%, which is a considerable decrease, given that the natural trend is a decrease of 5% in the past 10 years 16). As to nicotine dependence, those with lower FTQ score showed higher cessation rate in both groups. However, in the medium nicotine dependence group, which accounts for more than 40% of the smoking subjects, the cessation rate showed higher tendency in the intervention group. This suggests that intervention is especially effective for those

7 402 T KADOWAKI et al. with medium nicotine dependence who may experience hardship in smoking cessation. The results also showed that the intervention reduced the proportion of those with negative willingness toward smoking cessation; even if the subjects failed to quit smoking, they may potentially succeed in cessation in the next opportunity. Majority of smokers think that they can quit smoking by themselves; 68% of the male smokers answered that they do not need any support for smoking cessation according to a survey conducted in the same region 17). However, our results showed that professional support is actually effective. Intervention can either lead a smoker to successful cessation or improved willingness toward smoking cessation. It also provides a smoking cessation opportunity for those subjects who have never thought of quitting smoking; 6.9% of those who had no intention of quitting succeeded, which is a larger decrease than the natural trend. It has been reported that those who have tried to quit smoking previously are more likely to succeed in cessation in voluntary basis studies 18, 19). This indicates that by approaching all smokers, we can either help a subject quit initially or help him in future cessation attempts. Although our results showed no difference in succeeding cessation between those who had previous quit attempts and who did not, this may be due to our study design which was not a voluntary basis; our subjects included smokers with little willingness toward cessation who have failed to quit in the past. We could still succeeded in reducing the proportion of those with negative willingness toward smoking cessation. The results also suggest that continuous support should be necessary even after a person succeeded in short term smoking cessation; nicotine dependence did not affect resmoking in the long-term once subjects succeeded in cessation. The reason why occupational settings are desirable is that the main target of the primary prevention is the healthy working population. In a clinical setting, doctors can have a significant effect on a patient s smoking behavior 20), but from the viewpoint of primary prevention, it is more desirable to quit smoking before having to see a physician. Community intervention is effective 21), but the majority of the smokers are male and it is hard for the community health-care providers to reach the working male subjects. The extent to which risk-reduction interventions are effective may depend on adherence of the subjects 22). In occupational settings, workers can readily reach occupational doctors or nurses within the worksite. For companies, smoking cessation is a sound economic investment according to a simulation analysis based on US data 23). There are several limitations in our study. First, 51 of the male workers (12.1%) were lost to follow-up during the 18-month study period due to national economic conditions. To be conservative in our analysis, we assumed that all of those who were lost to follow-up did not change their original smoking habit; even if a subject succeeded in smoking cessation, we considered them to be a smoker after he left the factory. Second, smoking cessation assessment by CO measurement may not be sufficient, because the half-value period of CO varies with the level of the individual s physical activity 14), and smokers may sometimes be classified as nonsmokers if measured after several hours from the last cigarette smoking. The reason why we introduced the urine test of nicotine metabolites which remain about 20 hours 14) in the long-term survey is to avoid misclassifying those who smoke only one or two cigarettes a day as a long-term result of our intervention. In conclusion, we showed the effectiveness of smokingcessation intervention in all smokers at a worksite. The results suggested that approaching all of the smokers impacts the overall smoking rate and encouraged intervention in all smokers regardless of their expressed willingness to quit. References 1) Peto R, Lopez AD, Boreham J, Thun M, Heath C Jr, Doll R (1996) Mortality from smoking worldwide. Br Med Bull 52, ) World Health Organization (1999) The Health Report , WHO, Geneva. 3) Liu BQ, Peto R, Chen ZM, Boreham J, Wu YP, Li JY, et al. (1998) Emerging tobacco hazards in China: 1. Retrospective proportional mortality study of one million deaths. BMJ 317, ) Niu SR, Yang GH, Chen ZM, Wang JL, Wang GH, He XZ, et al. (1998) Emerging tobacco hazards in China: 2. Early mortality results from a prospective study BMJ 317, ) Health and Welfare Statistics Association (1998) Journal of Health and Welfare Statistics in Health and Welfare Statistics Association, Tokyo. 6) Weiss W (1997) Cigarette smoking and lung cancer trends. A light at the end of the tunnel? Chest 111, ) Shimizu J, Kita Y, Kai K, Okayama A, Choudhury SR, Kawashima J, et al. (1999) Randomized controlled trial for smoking cessation among city office employees. Nippon-Koshu-Eisei-Zasshi 46, 3 13 (In Japanese). 8) Muto T, Nakamura M, Oshima A (1998) Evaluation of a smoking cessation program implemented in the Industrial Health 2000, 38,

8 INTERVENTION IN ALL SMOKERS IN A WORKSITE 403 workplace. Ind Health 36, ) Heatherton TF, Kozlowski LT, Frecker RC, Fagerstrom KO (1991) The Fagerstrom test for nicotine dependence: a revision of the Fagerstrom Tolerance Questionnaire. Br J Addict 86, ) Micro Smokerlyzer. Bedfont Scientific Ltd. UK. 11) Jarvis MJ, Belcher M, Versey C, Hutchison DC (1986) Low cost carbon monoxide monitors in smoking assessment. Thorax 41, ) Okayama A, Kita Y, Ueshima H, Kadowaki T (2000) Individual Health Education Manual on Smoking Cessation (Kinnenn-no-kobetsu-kennko-kyoikushidosha-manyuaru), Hoken-dojinsha, Tokyo (In Japanese). 13) NicCheck. DynaGen, Inc., USA. 14) Eswara AR, Nochur SV, Mossman DJ (1996) Detection of nicotine and its metabolism in urine. Am J Health Behav 20, ) Fisher KJ, Glasgow RE, Terborg JR (1990) Work site smoking cessation: a meta-analysis of long-term quit rates from controlled studies. J Occup Med 32, ) Kawane H (1998) The tobacco epidemic in Japan (letter). Thorax 53, ) Ueshima H, Kita Y, et al. (1997) Consciousness survey on tobacco and health , Hikone City Health Promotion Association, Hikone (In Japanese). 18) Ozasa K, Higashi A, Shiroeda O, Watanabe Y, Aoike A, Kawai K, et al. (1991) A study of smoking cessation education in a multiphasic screening setting. Nippon- Koshu-Eisei-Zasshi 38, (In Japanese). 19) Farkas AJ, Pierce JP, Zhu SH, Rosbrook B, Gilpin EA, Berry C, et al. (1996) Addiction versus stages of change models in predicting smoking cessation. Addiction 91, ) Ockene JK, Zapka JG (1997) Physician-based smoking intervention: a rededication to a five-step strategy to smoking research. Addict Behav 22, ) McAfee T, Sofian NS, Wilson J, Hindmarsh M (1998) The role of tobacco intervention on population-based health care: a case study. Am J Prev Med 14 (3 Suppl), ) Burke LE, Dunbar-Jacob JM, Hill MN (1997) Compliance with cardiovascular disease prevention strategies: a review of the research. Ann Behav Med 19, ) Warner KE, Smith RJ, Smith DG, Fries BE (1996) Health and Economic Implications of a worksite smoking-cessation program: a simulation analysis. J Occup Environ Med 38,