Employee Attitudes and Smoking Behavior at the City of Hope National Medical Center Smoke-Free Campus

JN046_Jrnl_40609Lin.qxd 6/20/06 11:27 PM Page 535

535

Original Article

David Lin,a Douglas C. Stahl, PhD,b David Iklé, PhD,b and Frederic W. Grannis, Jr, MD,b Stanford and Duarte, California

Key Words formation may be useful to NCCN and other medical centers in as- Smoke-free campus, smoke-free workplace, tobacco, cigarette, sessing current and planning future campus smoking policies. (JNCCN smoking cessation, environmental tobacco smoke, tobacco control, 2006;4:535–542) prevention, questionnaire Involuntary (second-hand) smoking is estimated to cause Abstract 3000 lung cancer deaths per year and as many as 62,000 Since 1989, City of Hope National Medical Center (COH), located in deaths annually from coronary artery disease.1 In addi- Duarte, California, and a member of the National Comprehensive Cancer Network, has prohibited smoking anywhere on the 100-acre tion to adverse health effects, tobacco use also has vari- campus. Because little published information is available on smoke- ous negative influences on the workplace, including free campuses (SFCs), we investigated the attitudes of COH em- increased employee disability, absenteeism, reduction in ployees toward the SFC and attempted to answer the question of job performance, increased risk for injury, increased health whether a difference in employee smoking behavior occurred and care costs to employers, increased workplace fires, and was attributable to the SFC policy. An anonymous survey was sent 2–11 to all 2787 campus employees using both Web-based data entry and reduced employee wages. scannable paper forms. Employees with network access were con- Recognizing that tobacco smoke exposes workers to tacted by e-mail and those without computers were contacted in per- serious health risks, a concerted effort has been made son. Respondents were asked questions regarding their attitudes since the 1980s to pass smoke-free workplace (SFW) toward the SFC and about their smoking history and current smok- legislation. ing behavior. A total of 1356 responses (48.7%) were received. One hundred (7.4%) respondents were current smokers, 242 (17.8%) The severity of these measures range from minimal were ex-smokers, and 1014 (74.8%) were nonsmokers. Smokers and or no restrictions, smoking banned in some work areas, ex-smokers smoked a mean of 2.4 fewer cigarettes (95% CI, 1.8 to smoking permitted in restricted areas only, and 100% 3.1; P < .0001) on workdays than on days off, with evidence of a smoke-free environments. By 1999, 70% of U.S. em- small amount of compensatory smoking (one cigarette per day). Of ployees worked in environments where smoking was pro- the smoking employees, 61.6% believed that the SFC reduced their 12 cigarette consumption, and 42.2% quit smoking while employed at hibited inside workplace buildings. The percentage COH between 1989 and 2002. Of all respondents, 92.6% supported of Californians working in SFWs increased from 35% the COH SFC policy. High acceptance was consistent across gender, in 1990 to 93.4% in 1999.13 Both tobacco industry and ethnicity, job type, and educational level. COH’s SFC policy is strongly public health research indicates that SFWs result in fewer supported by employees and may decrease cigarette consumption cigarettes smoked by workers and increase the percent- and facilitate smoking cessation among smoking employees. This in- age of smokers who quit. According to subpoenaed documents from the tobacco industry, research has led to the conclusion that From aStanford University, Stanford, California, and bCity of Hope National Medical Center, Duarte, California. SFWs are very effective in reducing the number of ciga- Submitted July 12, 2004; accepted for publication rettes smoked by customers and that these environments May 8, 2005. This study was funded with generous contributions from Mrs. foster smoking cessation (SC) attempts: Harriet Bodle and the City of Hope National Medical Center. Correspondence: Frederic W. Grannis, Jr, MD, Associate Professor, Total prohibition of smoking in the workplace strongly af- Department of Thoracic Surgery, City of Hope National Medical fects industry volume. Smokers facing these restrictions Center, 1500 East Duarte Road, Duarte, CA 91010. E-mail: fgrannis@ coh.org consume 11%–15% less than average and quit at a rate

536 Original Article

Lin et al.

that is 84% higher than average. Only 6.4%–10.3% ferences in SC compared with workers in non–smoke- of smokers face total workplace prohibition but these free environments (quit ratio of 0.506 vs. 0.377). restrictions are rapidly becoming more common…. Longo et al.23 noted that 5 million Americans Milder workplace restrictions, such as smoking only would be directly affected by smoking bans in in designated areas, have much less impact on quitting American hospitals. The same group studied work- 14 rates and very little effect on consumption. place smoking ban effects on SC rates in smoke-free The SFW reduces exposure of nonsmokers to in- hospital workers versus workers in non–smoke-free voluntary inhalation of tobacco smoke. In the mid- environments over 3 years, finding that 26% of 1990s the California Tobacco Survey showed that 2.2 employees in a SFW quit versus 14% in non-SFWs million California nonsmokers were exposed to to- (P = .02).24 The Community Intervention Trial for bacco smoke at work, including 9.3% in SFWs versus Smoking Cessation showed that SFW employees were 51% at work sites with smoking area restrictions. 25% more likely to make serious attempts to quit and Younger smokers, men, Hispanics, and persons with maintain cessation. Continuing smokers smoked 2.75 lower educational levels had higher exposure to fewer cigarettes per day.25 environmental tobacco smoke (ETS).15 In Finland, SFWs may also affect primary smoking preven- exposure to ETS declined, tobacco consumption 26 diminished, and nicotine concentration in the blood tion. Farkas et al. documented reduction in initiation of smokers diminished after national SFW legisla- of smoking among adolescents who worked in SFWs tion.16 In New Zealand, nicotine exposure in employ- compared with those who worked in non-SFWs. ees varies directly with smoke-free status.17 Studies in These studies suggest that SFWs improve health California measuring pulmonary function before and among nonsmokers and smokers, reduce smoking rates after smoking was prohibited in bars and taverns among employees, facilitate higher chances of SC, documented improvement in respiratory health of and affect primary smoking prevention. workers.18 The positive effects of an SFW might be mitigated Studies have proven that SFWs not only protect by the employee’s ability to smoke immediately out- nonsmokers, but also encourage reduction and cessa- side the workplace. We were unable to identify any pre- tion among smokers. The Institute for Health Policy vious studies of the effects of smoke-free campuses Studies at the University of San Francisco School of (SFC) on employee smoking behavior. We performed Medicine surveyed more than 11,000 people who a retrospective questionnaire study to investigate worked indoors. Prevalence of smoking was lower at whether the SFC at the City of Hope (COH) has an SFWs (13.7% vs. 20.6%). Cigarette consumption was effect on employee smoking and SC rates. approximately 21% lower among employees working in non-SFWs: “The smoke-free workplace is an effective public health measure for decreasing smoking Methods prevalence and cigarette consumption among continuing smokers.”19 This study was conducted at COH, a National After implementation of an SFW policy, workers Comprehensive Cancer Network (NCCN) member at Johns Hopkins University School of Medicine cancer center hospital and research campus in Duarte, smoked 4 cigarettes fewer each day and had signifi- California. The campus has more than 100 buildings cantly reduced nicotine and carbon monoxide levels spread out over 110 acres. COH has had an official SFC during work hours.20 Workers at the Centers for Disease policy since 1989. Employees and visitors are prohib- Control and Prevention (CDC) had an overall smok- ited from smoking anywhere on campus grounds, both ing prevalence of only 11.1%. Daily cigarette con- inside and outside of buildings. The study was ap- 21 sumption decreased after implementation of an SFW. proved by the COH Institutional Review Board. In SFWs in Toronto, employees were less likely to smoke at work and reduced the number of cigarettes Participants that they smoked, although SC rates remained un- The study population consisted of all 2787 employees changed.22 Hospital employees at the University of working on the COH Duarte campus, including those Missouri-Columbia had statistically significant dif- hired through outside contractors.

Original Article 537

City of Hope Smoke-Free Campus

Smoking Survey using Chi-squared tests for proportions and 2-sample E-mail messages were sent to all COH employees who and paired t tests for continuous measures, as appro- had addresses, and articles about the survey request- priate. All 2-tailed tests of hypotheses were performed ing the anonymous participation of all employees were at the .05 level of significance. All analyses were per- posted in a campus publication 3 times at weekly in- formed using JMP Version 5 statistical analysis software tervals. The e-mail message had a link to a Web page (SAS Institute, Cary, North Carolina). form that contained the survey questions and responses. The questionnaire can be viewed at URL Results http://www.cityofhope.org/smokingsurvey.pdf. Participants completed the survey online and The target population consisted of the 2787 employ- responses were entered directly into a database ees and contractors working on the COH campus. (Microsoft SQL Server, Microsoft Corp., Bellevue, Responses totaled 1356, for a response rate of 48.7%. Washington). Employees without e-mail access filled Of the surveys, 1277 (94.2%) were submitted online by way of the Web-based questionnaire, and the re- out paper survey forms in either English or Spanish maining 79 (5.8%) were completed on the scannable and faxed their responses directly into the study data- forms and submitted by way of fax. Of the scannable base using an optical character recognition system forms, 47 (59.5%) were completed in English and 32 (TELEform Software Suite, Cardiff Software, Inc, (40.5%) in Spanish. Some questionnaires had missing Vista, California). We attempted to increase the per- data. Of the respondents, 358 (26.4%) were male and centage of responses by offering participants entry into 997 (73.5%) were female, with ages ranging from 14 to a lottery drawing for tickets to musical and sporting 77 years (mean ± SD = 41 ± 11). General demograph- events. To ensure anonymity, each Web form and pa- ics obtained from personnel records of survey respon- per survey contained a unique code number to be used dents and COH employees are presented in Table 1. by participants to claim lottery prizes. Evidence was found of selection effects that may in- Smoking Status troduce bias and weaken conclusions drawn from the Study participants were classified into 3 groups. We results. For example, Asian-American employees and defined a smoker as an individual who had smoked service workers had lower rates of participation than more than 100 cigarettes and had smoked within the Caucasian and administrative/support employees. past 30 days; an ex-smoker as one who had smoked Smoking Status more than 100 cigarettes but had not smoked within One hundred (7.4%) respondents were current smok- the past 30 days; and a nonsmoker as one who smoked ers, 242 (17.8%) were ex-smokers, and 1014 (74.8%) fewer than 100 cigarettes and had not smoked within were nonsmokers. Of the nonsmokers, 745 (54.9%) the past 30 days. had never smoked a cigarette and 269 (19.8%) smoked Nonsmokers were asked only demographic ques- fewer than 100 cigarettes in their lifetime but never tions, whereas smokers and ex-smokers were asked a became regular smokers. Of 611 respondents who series of questions about their smoking history and smoked at least one cigarette, 342 (56.0%) went on to their opinions regarding the SFC and its effect on their become smokers and ex-smokers. smoking behavior. In particular, they were asked to Smoking History estimate the number of cigarettes they smoked on a The mean age at initiation of smoking was 18.2 years typical workday and the number they smoked on a with a standard deviation of 6.3 years. This age dis- typical day off. tribution is similar to that of smoking initiation de- Statistical Analysis scribed for the U.S. population.27 This retrospective, descriptive, exploratory study was Ex-smokers smoked for a mean and 95% CI of designed to evaluate smoking behavior and beliefs in 12.3 (range, 10.9–13.6) years. The mean age of SC relation to the SFC policy at COH. Results are sum- for ex-smokers was 31.7 (range, 30.3–33.1) years, and marized using standard descriptive statistics for cate- 207 (88.8%) of 233 ex-smokers quit with no assis- goric and continuous variables. Where appropriate, tance. Fewer than 13.7% of ex-smokers had any form 95% confidence intervals (CIs) are reported. Tests of of SC intervention, such as counseling, nicotine hypotheses comparing selected groups were performed replacement therapy, or bupropion.

538 Original Article

Lin et al.

Table 1 General Demographics of Survey cigarettes fewer on workdays, corresponding to a dif- Respondents and Entire Population ference in tobacco consumption between work and days off of 20.7% (P < .0001). This result is consistent of Employees with employee opinions that COH’s SFC decreases Respondents All Employees tobacco consumption. Of the smoker and ex-smoker N% N % respondents, 106 (61.6%) asserted that they smoked Gender less because of the smoke-free policy at COH. Current Male 358 26.4 780 28.0 smoking employees estimated that they would smoke Female 997 73.6 2007 72.0 a mean of 2.17 (range, 1.7–2.6) cigarettes more if they Ethnicity were allowed to smoke on campus in a designated area. African- 65 4.8 165 5.9 Only 1 (0.6%) employee said that the SFC increased American cigarette consumption. Employees may have smoked Asian- 238 17.6 673 24.1 fewer cigarettes while on campus because it took them American a mean of 3.31 (2.6–4.0) minutes each way to walk to Caucasian 691 51.1 1302 46.7 the edge of campus to smoke. Hispanic 312 23.1 625 22.4 Our data also suggest that COH’s SFC facilitates Other 46 3.4 11 0.4 SC. Of 173 employees who smoked at some time dur- Higher Level ing their employment, 73 (42.2%) quit smoking while of Schooling working on the SFC. Of these ex-smokers, 30 High School 336 24.9 – – (41.1%) believed that the SFC helped them quit or less smoking. College 659 48.8 – – Among current smokers, 69 (69.0%) expressed a Post-graduate 355 26.3 – – desire to quit. Moreover, 64 (64.6%) of 99 respond- Job Classification ing smokers stated that one potential effect of an SFC Knowledge 720 53.2 1697 60.9 would be to make SC easier. In addition, 54 (54.6%) Administra- 487 36.0 609 21.9 smokers believed that an on-campus CS program tive support would be conducive to CS. Only 12 (12%) of smok- Service 80 5.9 283 10.2 ers expressed no desire to quit. Student and 66 4.9 187 6.7 The COH smoke-free policy was a “good idea” other according to 1249 (92.6%) respondents. This high acceptance was consistent across gender, ethnicity, job type, and educational level. Only 28 (2.1%) re- The responses showed that on average, men spondents said that the policy was a “bad idea.” Although smoked more and for longer durations than women. the highest level of objection came from smokers, only The significant differences in smoking behavior be- 12 (12%) disagreed with the smoke-free policy. Fifty- tween men and women are highlighted in Tables 2 five (31.8%) smokers and ex-smokers stated that they and 3. Table 4 shows the inverse relationship between would have smoked outside of their workplace building on campus if this had been allowed. Of the smoking and educational level observed in the study employees who were smokers, 14.9% violated the subjects. As educational level increased, the percent- SFC policy on one or more occasions. age of smokers and ex-smokers decreased (P = .0059). Compensatory Smoking Effects of a Smoke-Free Campus One serious concern is that employees might smoke The smoking behavior of employees at COH is de- more during off-hours to compensate for the number scribed in Table 4. To evaluate the effect of COH’s of cigarettes that they could not smoke while on cam- SFC on employee cigarette consumption, we com- pus. To test this hypothesis, we divided the day into 2 pared the number of cigarettes that smokers and ex- strata: hours asleep and hours awake. The average smokers smoked on workdays with the number of American adult gets approximately 7 hours of sleep per cigarettes they smoked on days off over a 24-hour pe- night, and is therefore awake for 17 hours.28 In our riod. The mean difference was 2.4 (range, 1.8–3.1) study, we observed a mean of 11.6 (range, 10.3–12.9)

Original Article 539

City of Hope Smoke-Free Campus

Table 2 Smoking Behavior of Male and Female Survey Respondents Male Female Variables Mean N SD Mean N SD P value* Total years smoked 15.6 102 10.98 13.13 240 11.27 .0005 Average number of cigarettes 15.3 99 9.6 10.8 237 12.3 .0019 smoked per day Average number of cigarettes 2.9 51 2.9 1.8 121 2.2 .0168 smoked during 8 work hours Age of smoking initiation 18.28 102 7.85 17.99 240 5.5 .7347

*The P value is from the t test of the hypothesis of equality of means between the 2 groups. Significant values are shown in bold. ␣ = 0.05 was used.

cigarettes per day per smoking employee. Assuming Discussion that a smoker smokes at a constant hourly rate and The effectiveness of an SFW may be mitigated by the does not wake up during the night to smoke, a smok- employee’s ability to step outside the building to ing employee would smoke 0.682 cigarettes per hour smoke. In an SFC, however, where smoking is pro- and 5.46 cigarettes in an 8-hour workday. Because em- hibited not only inside buildings but also between ployees reported that they smoke 2.1 (range, 1.7–2.5) and outside of buildings, workers must leave the cam- cigarettes while on campus, we estimate the SFC re- pus in order to smoke. COH has had an official SFC duces cigarette consumption by 3.36 cigarettes dur- policy since 1989. The physical setup of the COH ing an 8-hour workday. Normally, in the remaining 9 campus theoretically makes it difficult for employees hours that an employee is awake and not at work, we to smoke on breaks because they must walk long dis- would expect a total cigarette consumption of 6.14 tances to do so. cigarettes. When compensatory smoking was evident, Although smoking restriction is more compre- however, we would expect the employee to smoke hensive on an SFC than an SFW, no additional more often to compensate for the 3.36 cigarettes not reduction of smoking was obvious on the COH cam- smoked while at work. Therefore, the employee would pus. The 2.4 decrease in cigarette consumption theoretically smoke a total of 9.5 cigarettes. However, within COH’s SFW is comparable to that from ear- our study showed a mean cigarette consumption of lier medical SFW studies. In Fichtenberg and 7.1 (range, 6.1–8.0) during off-work hours. Thus, we Glantz’s29 meta-analysis, this figure ranged between conclude that the SFC at COH reduced employee 2.2 and 5.4 (mean, 3.1) cigarettes per workday. Our cigarette consumption with evidence of minimal study, however, was conducted at a medical center compensatory smoking (1 cigarette per day). where tobacco consumption is lower than the

Table 3 Additional Smoking Behavior of Male and Female Survey Respondents Male Female Variables Count N % Count N % P value* Number whose smoking 30 102 29.41 66 240 27.50 .038 behavior was affected by COH’s smoke-free policy Number of ex-smokers who 57 74 77.03 148 168 88.10 .016 quit with no assistance Number of smokers who want 18 28 64.29 51 72 70.83 .031 to stop smoking

*The P value is from the z test of the hypothesis of equality between proportions between the 2 groups. Significant values are shown in bold. ␣ = 0.05 was used.

540 Original Article

Lin et al.

Table 4 Smoking Status and Education* tion. This decrease protects workers, patients, and School Smoker Ex-Smoker Non-Smoker visitors from involuntary inhalation of tobacco smoke. Completed N (%) (%) (%) Also, with 2.4 fewer cigarettes smoked per day, smok- High School 336 11.3 19.6 69.1 ers are effectively cutting their cigarette consump- & below tion by 20.7%, thereby reducing tobacco industry College 659 6.8 17.9 75.3 revenue. This reduction in tobacco consumption can Post- 355 4.8 15.8 79.4 be expected to decrease the future risk for disease graduate caused by tobacco products, including coronary artery disease, chronic obstructive lung disease, and *Smoking status was significantly related to education (P = .0059). lung cancer. Facilitating Smoking Cessation A CDC report stated that 70% of current smokers national average. A 2000 CDC study reported that wanted to quit in the year 2000.33 Our study mirrors more than half of U.S. adults smoked more than 15 these data, with 69% of COH’s current smokers con- cigarettes per day and 15.3% smoked 25 or more cig- templating SC. Only 12% of current COH smokers are arettes per day.30 In comparison, the mean number content with their smoking behavior. of cigarettes smoked daily in the COH population Of employees who were regular smokers, 42.2% was 10.4 (range, 9.0%–11.8%); fewer than 19.4% have quit and maintained abstinence while working of employee respondents smoked more than 15 cig- at COH, 41.1% of whom stated that the SFC played arettes (range, 9.5% –23.7%) per day; and only 4.1% a role in the success of their SC. This suggests that in smoked 25 or more cigarettes per day (range, addition to reducing tobacco consumption, COH’s 1.6%–10.0%), which was significantly different from SFC provides an environment that facilitates SC. This the CDC percentages. A disproportionately lower phenomenon has also been described in a Duke prevalence of smoking occurred in the COH em- University Medical Center study in which a 22.5% ployee population. For example, at Duke University SC rate was observed 15 months after a SFW policy Medical Center, 23% of employees smoke compared was instituted.34 Most (64.6%) COH smokers believed with 7.4% at COH.31 that the SFC at COH would make it easier if they This low prevalence of smoking is probably ex- tried to quit. plained by the high percentage of employees who are physicians, scientists, administrator/managers, college Attitudes Toward a Smoke-Free Campus graduates, and females, and by the location of the med- More than 90% of COH employees agreed with the ical center in the state of California. Marked dispari- SFC policy. Even among current smokers, only 12% ties are present in the prevalence of smoking among of the one group whose habits would be most directly different job categories. For example, nationwide only affected disagreed with the smoke-free policy. In the 5.4% of physicians and 9% of administrator/managers Duke study, 75.8% of subjects agreed with the smoke- smoke, whereas 39.2% of “blue collar” and 34.5% of free policy.35 service workers smoke.32 Smoking is less common in Reflections on Response Rate and Methods 2 women, college graduates, and Californians. Because Use of computerized survey techniques, including tobacco use is initially lower at COH, any decrease in e-mail announcements, HTML questionnaire forms, the number of cigarettes smoked may be correspond- optical character recognition systems, and encryption ingly smaller. of survey completion receipts that allowed collection Future prospective studies performed before and of lottery prizes yielded responses from almost half of after transition from a non-SFW and an SFW to an all employees. SFC at medical and non-medical institutions in different geographic areas will be required to determine if a quantitative difference in reduction of smoking Conclusions exists between SFWs and SFCs. Cigarette consumption on workdays compared with A decrease in tobacco consumption yields sev- days off decreased by a mean of 2.4 cigarettes per day eral significant benefits for the entire COH popula- and the prevalence of smoking decreased by 3.3% per

Original Article 541

City of Hope Smoke-Free Campus

year on COH’s SFC. A very high level of support 2. Osinubi OY, Slade J. Tobacco in the workplace. Occup Med for the policy exists among employees. This informa- 2002;17:137–158. 3. Ryan J, Zwerling C, Orav EJ. Occupational risks associated with cig- tion may be useful to other NCCN institutions in as- arette smoking: a prospective study. Am J Public Health 1992;82:29–32. sessing current and planning future campus smoking 4. Ryan J, Zwerling C, Jones M. Cigarette smoking at hire as a pre- policies. dictor of employment outcome. J Occup Environ Med 1996;38: 928–933. Limitations of Study 5. Parrott S, Godfrey C, Raw M. Costs of employee smoking in the Because no control group was included, this study workplace in Scotland. Tob Control 2000;9:187–192. does not allow conclusions as to whether an SFC 6. Weis WL. Can you afford to hire smokers? Pers Adm 1981;26:71–78. offers additional reduction of smoking or facilitation 7. Weis WL. Profits up in smoke. Personnel J 1981;60:162–165. of SC compared with an SFW. Because the study 8. Pugsley MW. Nonsmoking hiring policies: Examining the status of smokers under Title I of the Americans with Disabilities Act of 1990. occurred in California, where the prevalence of smok- Duke Law J 1994;43:1089–1114. ing is low, and at a medical center with a high per- 9. Bloch H, Pupp R. The impact of smoking on health, productivity, centage of employees known to have a low prevalence and earnings: a new collective bargaining frontier. In: Musacchio, of smoking (e.g., females, health care workers, ad- RA, Hough DE, eds. Socioeconomic issues of health. Chicago (IL): American Medical Association, 1981; 35–50. ministrators, managers), these results may not be ap- 10. Levine PB, Gustafson TA, Velenchik AD. More bad news for smok- plicable to other types of business or educational ers? The effects of cigarette smoking on labor market outcomes campuses. Differences in smoking behavior may ex- National Bureau of Economic Research. NBER Working Paper Series, ist between current and ex-smokers that can be dis- Working Paper, 5270. Cambridge (MA); September 1995. 11. Bertera RL. Effects of behavioral risks on absenteeism and health- sected out in a larger study group. care costs in the workplace. J Occupational Med 1991;33:1119–1124. We did not assess differences in employee smok- 12. Shopland DR, Gerlach KK, Burns DM, et al. State-specific trends in ing behavior among employees working day shifts and smoke-free workplace policy coverage: the current population survey those working night shifts. This question might be ad- tobacco use supplement. J Occup Environ Med 2001;43:680–686. dressed in a future study. Further prospective studies 13. Gilpin EA, Farkas AJ, Emery SL, et al. Clean indoor air: advances in California. Am J Public Health 2002;92:785–791. to determine the impact of SFCs should be performed 14. Hieronimus, J (Philip Morris Tobacco Company). Restrictions; bans, at medical and non-medical campuses to assist in as- workplace smoking restrictions; consumption. January 22, 1992. 5 sessing current and planning future campus smoking pages. Bates No.: 2023914280/4284. Available at: http://www.pm- policies. docs.com/getallimg.asp?if=avpidx&DOCID=2023914280/4284. Accessed September 11, 2002. 15. Borland R, Pierce JP, Burns DM, et al. Protection from environmental tobacco smoke in California. The case for a smoke-free work- Acknowledgments place. JAMA 1993;269:1388. 16. Heloma A, Jaakkola MS, Kahkonen E, et al. The short-term impact The references to tobacco industry documents are of national smoke-free workplace legislation on passive smoking and courtesy of Anne Landman’s collection of selected tobacco use. Am J Pub Health 2001;91:1416–1418. documents at http://tobaccodocuments.org/landman/. 17. Al-Delaimy W, Fraser T, Woodward A. Nicotine in hair of bar and The following individuals from the City of Hope restaurant workers. N Z Med J 2001;114:80–83. National Medical Center Departments of Biostatistics 18. Eisner MD, Smith AK, Blanc PD. Bartender’s respiratory health after establishment of smoke-free bars and taverns. JAMA 1998;280: and Bioinformatics provided valuable technical assis- 1909–1914. tance to this study: Michael Mora, Projects Manager; 19. Woodruff TJ, Rosbrook B, Pierce J, et al. Lower levels of cigarette Tremon White, MBA, Application Development consumption found in smoke-free workplaces in California. Arch Lead; Chris Lanning, Operations Lead; Monica Intern Med 1993;163:1485–1493. Youngblood, Operations Specialist – Data Collection 20. Brigham J, Gross J, Stitzer ML, et al. Effects of a restricted work-site smoking policy on employees who smoke. Am J Public Health 1994; Analyst; Enrique Orazco, Lead DBA; Lei Chen, 84:773–778. Programmer/Analyst; and Eric Huang, Database 21. Emont SL, Zahniser SC, Marcus SE, et al. Evaluation of the 1990 Administration Lead. Centers for Disease Control and Prevention smoke-free policy. Am J Health Promot 1995; 9:456–461. 22. Pederson LL, Bull SB, Ashley MJ. Smoking in the workplace: do smoking patterns and attitudes reflect the legislative environment? References Tob Control 1996; 5:39–45. 1. State-specific prevalence of current cigarette smoking among adults 23. Longo DR, Brownson RC, Johnson JC, et al. Hospital smoking bans and the proportion of adults who work in a smoke-free environment. and employees smoking behavior: results of a national survey. JAMA MMWR Morb Mortal Wkly Rep 2000;49:978–982. 1996;275:1252–1257.

542 Original Article

Lin et al.

24. Longo DR, Johnson JC, Kruse RL, et al. A prospective investigation 29. Fichtenberg CM, Glantz SA. Effect of smoke-free workplaces on of the impact of smoking bans on tobacco cessation and relapse. Tob smoking behavior: systematic review. Brit Med J 2002;325:188. Control 2001;10:267–272. 30. Centers for Disease Control. Reducing the health consequences of 25. Glasgow RE, Cummings KM, Hyland A. Relationship of worksite smoking: 25 Years of progress—a report of the Surgeon General. smoking policy to changes in employee tobacco use: findings from Rockville, MD: US Public Health Service; 1989. DHHS publication COMMIT. Community Intervention Trial for Smoking Cessation. CDC (89–8411). Tob Control 1997;2(suppl):44–48S 31. Stave GM, Jackson GW. Effect of a total work-site smoking ban 26. Farkas AJ, Gilpin EA, White MM, et al. Association between house- on employee smoking and attitudes. J Occup Med 1991;33: hold and workplace smoking restrictions and adolescent smoking. 884–890. JAMA 2000;284:717–22. 32. Nelson DE, Emont SL, Brackbill RM, et al. Cigarette smoking 27. Pierce JP, Fiore MC, Novotny TE, et al. Trends in cigarette smok- prevalence by occupation in the United States: a comparison be- ing in the United States. JAMA 1989;261:61–65. tween 1978 to 1980 and 1987 to 1990. J Occup Med 1994;36: 28. 1998 Omnibus Sleep in America Poll. National Sleep Foundation. 516–525. Washington DC 2001. Available at: www.sleepfoundation.org/pub- 33. Cigarette smoking among adults—United States, 2000. MMWR lications/1998poll.html#3. Accessed September 13, 2002. Morb Mortal Wkly Rep 2002;51;642–645.