Effects of a Restricted Work-Site Policy on Employees Who Smoke

Janet Brigham, PhD, Janet Gross, PhD, Maxine L. Stitzer, PhD, and Linda J. Felch, M4 Introduction example, in the study by Hatsukami et al.,6 smokers who reduced the number of Many institutions throughout the smoked by 50% or reduced the have moved toward restrict- yield of their cigarettes by about ing smoking in public places, particularly 50% experienced some craving and with- since the Surgeon General's 1986 report drawal discomfort but reported signifi- on involuntary smoking.1 Several studies cantly less discomfort than total abstain- have indicated that the incidence of ers. On the other hand, West et al.7 found public smoking changes when restricted- that switching to an ultra-low-yield ciga- smoking policies are implemented. Still- rette did not result in any appreciable man et al.2 reviewed empirical evidence of reporting of withdrawal distress, except reduction in smoking at the Johns Hop- for increased hunger. The mechanisms kins medical institutions and concluded thought to cause low-level withdrawal that implementing a smoke-free policy in symptoms under conditions of restricted a large medical center can decrease visible smoking are reduction in blood nicotine smoking and passive exposure to levels and habit change. smoke. A previous report by Becker et al.3 In a related vein, research on re- indicated that public smoking at the Johns stricted smoking in habitual smokers has Hopkins Children's Center was virtually suggested that smokers may adjust their eliminated in 1987 by implementation of a smoking behavior to offset effects of smoke-free policy. Recently, Borland et restricted access or reduced al.4 reported that Australian smokers yield.8'9 Benowitz et al.10 found that when exposed to a mandatory work-site smok- smokers reduced their consumption from ing ban reduced their smoking by about an average of 37 to 5 cigarettes per day seven cigarettes per day while working (i.e., a sixfold reduction), the toxicity per inside and by about five cigarettes per day cigarette increased threefold because of overall with very little compensatory, or more puffs per cigarette, longer breath increased, smoking observed outside of holding, and deeper inhalation, whereas work after enforced reduction at work. the cumulative daily exposure to and Of interest to policymakers and carbon monoxide dropped by only about work-site managers are the potential one half. This tendency for smokers to negative side effects of daytime absti- alter their smoking style (e.g., by increas- nence, which may affect the smoker's ing inhalation or taking more frequent behavior (e.g., symptoms of the tobacco withdrawal syndrome), as well as the Janet Gross, Maxine L. Stitzer, and Linda J. overall impact on smoking exposure (e.g., Felch are with the Johns Hopkins University compensatory changes in rates ofsmoking School of Medicine, Baltimore, Md. At the time of this study, Janet Brigham was with the outside the work place). Researchers on Johns Hopkins University School of Medicine; have found that some currently she is with the Western Psychiatric of the symptoms of the tobacco with- Institute, University of Pittsburgh, Pitts- drawal syndrome that typically accom- burgh, Pa. are to a Requests for reprints should be sent to pany tobacco abstinence5 present Maxine L. Stitzer, PhD, Behavioral Pharmacol- lesser extent in individuals who reduce ogy Research Unit, 5510 Nathan Shock Dr, cigarette consumption or switch to a Baltimore, MD 21224. lower-nicotine-yielding cigarette.6'7 For This paper was accepted August 3, 1993.

American Journal of Public Health 773 Brigham et al.

tional status. The self-reported amount of TABLE 1 -Baseline Characteristics of Groups (n = 67) smoking was significantly higher in the control group (t (60) = 2.02, P < .05). Group However, the groups did not differ on the biological indices of tobacco exposure, Restricted (n = 34) Control (n = 33) which included expired-breath carbon monoxide and saliva cotinine and nico- Mean age, y (SD) 38 (9.5) 37 (10.4) tine. The groups also differed with respect Mean weight, lb (SD) 176 (43) 173 (48) to the total withdrawal score and the score Occupational status,a % on the Depression-Dejection subscale of 1 12 10 the Profile of Mood States. In both cases, 2 30 29 the control subjects had higher scores, 3 42 26 4 15 35 indicative of greater distress. Gender, % male 27 30 Procedures Mean scores on self-report (SD) Tension-Anxiety 7 (6.2) 10 (8.0) All subjects participated in data Depression-Dejection* 5 (6.6) 10 (11.0) collection sessions at the research labora- Anxiety-Hostility 7 (8.6) 7 (7.7) Vigor 17 (5.4) 16 (7.8) tory after their work shifts. Experimental Fatigue 8 (6.6) 8 (6.7) (restricted smoking) subjects attended a Confusion-Bewilderment 4 (4.4) 5 (4.8) laboratory data collection session once Friendliness 19 (3.7) 19 (4.7) per week for 4 weeks before and 4 weeks Total Mood Disturbance 13 (31.8) 25 (39.3) after implementation of the smoking ban. Total withdrawal score 11 (7.1) 15 (9.6) Control subjects attended sessions at their Smoking characteristics, mean (SD) work site once weekly for 8 consecutive Years smoked 19 (8.6) 18 (10.4) weeks. Informed consent, which had been No. of cigarettes/day* 22 (10.7) 27 (10.0) approved by the Francis Scott Key Medi- Carbon monoxide, ppm 26 (14.0) 29 (12.6) Saliva nicotine, ng/mL 566 (764) 595 (640) cal Center Institutional Review Board, Saliva cotinine, ng/mL 232 (136) 282 (121) was obtained from all subjects after the study was explained at the first session. al = low status and 4 = high status. Subjects were paid $100 for successfully *P < .05. completing the study. Measures puffs) could offset the potential health pants were between 18 and 65 years old At each study contact, behavioral, advantages of reducing the number of and were not seeking treatment to quit biological, and subjective report data were cigarettes smoked at work. The extent to smoking. Forty-one smokers were re- collected. engage in compensatory which smokers cruited from the Francis Scott Key Medi- Smokng amount. Smoking amount smoking behavior and increase their expo- cal Center, the site of the smoking ban, was assessed via self-report and by collec- sure to toxic constituents of tobacco as a through advertisements, flyers, and refer- tion of cigarette butts for designated time result of restrictions on workplace smok- rals. Before analysis, four dropped out of periods. To determine the pattern of ing is unknown. the study, and three were omitted from smoking during the work shift, subjects the the Among hospitals joining the analysis because ofincomplete compli- were asked at each lab visit to recollect smoke-free movement recently was Fran- ance with the study procedures. The and report the number of cigarettes cis Scott Key Medical Center, part of the comparison control group was composed smoked at their workstation, at other sites institutions in Johns Hopkins medical of 51 individuals recruited through news- inside the building, and at any sites Baltimore, Md. This institution imple- paper advertisements at several nearby outside the building for the time period of mented a hospital-wide smoking ban on hospitals that had no restricted smoking the most recent work shift. To estimate event as an July 1, 1989. We used this policy. Of those initially enrolled, 10 24-hour smoking amount and toxic expo- to examine the opportunity biological, failed to complete the study, 1 quit sure to tobacco constituents, subjects subjective, and behavioral impact of a smoking during the study, and 7 were collected their cigarette butts for the smoking policy restriction on individual excluded from analyses because their 24-hour period before the lab session. smokers, including objective measures of mean number of cigarettes per day before Three separate plastic collection bags tobacco smoke exposure. We compared the smoking ban exceeded the maximum reflected three intervals of the day: (1) these effects with those in a control group number observed in the restricted-smok- between the end of work and bedtime the of smokers whose work-site smoking ing group. Consequently, 67 participants previous day, (2) between arising and policy remained unrestricted. were included in the final analysis of the going to work the day of the lab session, data. and (3) during work the day of the lab Methods Table 1 shows the demographic, session. The number of cigarette butts per smoking, and psychological characteris- interval and the weight of the butts from Subjects tics of the two groups at baseline. The each interval were assessed. Thus, for the Ninety-two smokers were recruited groups were similar with respect to mean period of the work shift, both self-report for the present study. All study partici- age, distribution of gender, and occupa- and butt count data were available. Butt

774 American Journal of Public Health May 1994, Vol. 84, No. 5 Work-Site Smoldng weight was used as an indirect measure of comparisons ofpreban and postban inter- behavior outside the building. Before the smoking intensity.9 Like butt length, butt actions were conducted by using Tukey's ban, the restricted group smoked an weight indicates the amount ofmaterial in Honestly Significant Difference (HSD) average of 0.83 (SD = 1.0) cigarettes per the cigarette left unsmoked. Butt length test. day outside the building compared with provides an indirect measure of smoking 2.51 (SD = 1.7) cigarettes after the ban intensity and relative toxicity because tar Results (q = 6.7, P < .01). The ban on smoking and carbon monoxide delivery increase had no effect on behavior at other logarithmically as cigarettes are smoked Work-Site Smoking locations within the building. Restricted to shorter butt lengths.11 Before weighing, Amount and intensity. Based on the subjects reported a mean of 0.84 all butts were desiccated until the mois- self-reported smoking amount, consump- (SD = 1.1) cigarettes per day before the ture content in the desiccator was below tion of cigarettes during work hours by ban and a mean of 0.45 (SD = 1.2) 20%. subjects in the restricted group dropped cigarettes per day after the ban at other Biological exposure. At each lab ses- significantly as a function of the ban (F [1, building sites. Co,ptr6l subjects did not sion, subjects provided saliva samples, 65] = 15.7, P < .0001). Restricted sub- change their smoking patterns over time which were then analyzed for nicotine and jects reported an average of 7.57 (meannuftiiber of cigarettes at worksta- cotinine (a metabolite of nicotine with a (SD = 4.7) cigarettes per day during tion = 6.8; mean at other inside loca- half-life of 18-20 hours) by using gas study weeks 1-4 (before the ban) vs 3.64 tions = 1.2; mean outside building = 1.8). chromatography (analyses conducted by (SD = 4.9) cigarettes per day during Before implementation of the smok- Labstat, Inc, Kitchener, Ontario, Canada). study weeks 5-8 (after the ban) (Tukey's ing ban, 82% of the subjects in the Breath carbon monoxide samples, which HSD q = 9.15, P < .001). None of the restricted group reported ever smoking at reflected recent exposure to tobacco smok- subjects in the restricted group quit their workstations. After implementation ing, were measured by using a Vitalo- smoking during the study. The smoking of the ban, this figure decreased to 18%. graph breath analyzer (Lenexa, Kans). pattem for control subjects did not differ The percentage of control subjects smok- Subjective report. Three subjective over time. The mean rate of smoking was ing at the workstation during these two measures were used to assess withdrawal 10.02 (SD = 2.9) cigarettes per day before time periods was 88% and 85%, respec- symptoms associated with smoking reduc- the ban vs 9.53 (SD = 4.8) cigarettes per tively. The percentage of subjects who tion and the impact of those symptoms on day after the ban. Reported smoking rates reported going outside the building to participants' work. A withdrawal symp- were highly similar to those determined smoke during the workday increased in tom questionnaire adapted from that with the butt count procedure for this the restricted group from 68% to 97% as a used by Hughes and Hatsukamis assessed time interval. Although there was a small function of the ban. The percentage of subjects' ratings of 18 symptoms (0 = none discrepancy between self-report and butt control subjects reporting smoking out- to 3 = severe), many of which character- count with respect to absolute numbers of side the building during these two time ize the DSM-III Tobacco Withdrawal cigarettes, there was in both cases a periods changed only slightly from 55% to Syndrome. This measure has been widely four-cigarette-per-day reduction in smok- 49%. used in smoking research and has good ing during work hours in the restricted Biological exposure. The repeated- psychometric properties.12 The Profile of group as a function ofthe ban, vs less than measures ANOVA revealed significant Mood States,13 a widely used 65-item a one-cigarette-per-day change in the interactions on both saliva nicotine and self-report questionnaire, was used to control group. breath carbon monoxide measures, which measure overall mood disturbance. This The average butt weight per cigarette reflected changes in work-shift smoking questionnaire yields factor scores on sub- smoked during work was calculated from behavior. As shown in Figure 1, average scales of Tension-Anxiety, Depression- the returned butts for this period. There saliva nicotine measured at the end of the Dejection, Anger-Hostility, Vigor, Fatigue, were no effects on butt weight found from work shift in the restricted group de- Confusion-Bewilderment, Friendliness, the repeated-measures analysis of vari- creased from 623 ng/mL before the ban to and Total Mood Disturbance. A work ance (ANOVA). In the restricted group, 354 ng/mL after the ban (q = 5.11, attitude and performance questionnaire the mean butt weight was 0.36 g P < .01), whereas there was no change in developed by the investigators solely for (SD = 0.11 g) before the ban and 0.35 g the control group. There was also a purposes of this study was designed to (SD = 0.09 g) after the ban. For the significant decrease in breath carbon elicit subjects' ratings (on a scale of 0 to control group, the means were 0.38 g monoxide in the restricted group from 10) regarding their concentration ability (SD = 0.06 g) and 0.36 g (SD = 0.08 g), 25.6 to 20.2 ppm (q = 5.43, P < .01) vs no and productivity at work, relations with respectively. change in the control group. coworkers, work enjoyment and satisfac- Smoking location. An analysis of Smoking during Off-Work Hours tion, and overall sense ofwell-being. self-reported amount of smoking at the variouswork-shift locations showed signifi- Amountand intensity. Butt count data StatisticalAnalysis cant interactions for smoking at the revealed that the number of cigarettes We analyzed the data with a two- workstation (F [1, 65] = 41.6, P < .0001) smoked per day during off-work hours did factor, mixed-model analysis of variance and for smoking outside the building (F not change significantly in either group as using SPSS-X for the Macintosh. The [1, 65] = 19.6,P < .0001). Restricted sub- a function of the ban (mean number of grouping factor was restricted smoking jects reported an average of 5.89 cigarettes [restricted group] = 12.56 pre- (smoking ban) vs control (unrestricted (SD = 4.8) cigarettes per day at the ban [SD = 5.7] vs 11.44 postban smoking); the repeated-measures, within- workstation before the ban compared [SD = 6.4]; mean number of cigarettes subject factor was time: preban (weeks with 0.69 (SD = 2.2) cigarettes after the [control group] = 11.99 preban [SD = 4.4] 1-4) vs postban (weeks 5-8). All tests of ban (q = 12.25, P < .01). The reverse vs 11.14 postban [SD = 4.8]). Nor did the significance were two tailed. Posthoc pattern was observed for their smoking mean butt weight of cigarettes collected

May 1994, Vol. 84, No. 5 American Journal of Public Health 775 Brigham et al.

and after the ban were 0.36 g (SD = 0.12 and a significant (3.23-point) decrease in Nicotine g) and 0.33 g (SD = 0.14 g), respectively. total withdrawal score (allPvalues < .05). 800S The mean weight of butts collected after Only scores on the craving-for-cigarettes 700 - work was highly similar to the mean and urges-to-smoke items were signifi- 600- weight of butts collected before work. cantly higher in the restricted group vs the E soo. Biological exposure. Cotinine levels control group during weeks 5-8 after the c 400- were significantly higher in the control ban (P < .01). [1, 65] = 13.7,P < .0001), 300 groupoverall (F Profile of Mood States. The Total and there was a significant interaction Mood Disturbance score, shown in Table *AmnI Pre Post between group and time (F [1, 65] = 17.0, 2, had a significant interaction (F [1, P < .0001). As shown in Figure 1, coti- 65] = 5.39, P < .02) with an increase in Co nine, which reflects overall nicotine expo- the restricted group scores and a decrease 35 - sure for the previous several days, de- in the control group scores. However, 30.- creased by 15% in the restricted group these within-group changes were not from 211 to 180 ng/mL (q = 3.61, P > .05) statistically significant in posthoc testing. 25 - and increased by 13.5% in the control Two interactions on the Profile of Mood = 4.62, 20 - group from 296 to 336 ng/mL (q States factors of Tension-Anxiety P < .01). Subjects were categorized ac- (P < .03) and Depression-Dejection 15 - cording to whether their average cotinine (P < .002) reflected significant preban Pre Post levels increased or decreased from the group differences. In both cases, the preban to the postban period. In the control group had higher scores, indicat- Cotinine restricted group, 71% of the subjects ing greater symptom reporting. The re- 400- showed a reduction in cotinine levels, with stricted group showed no significant * a mean decrease of 55.9 ng/mL (range = changes in factor scores between the 300 - 0.7-120.6 ng/mL). In the control group, preban and the postban periods, although 33% of the subjects showed a reduction in scores on the Fatigue factor decreased 200 - cotinine levels, with a mean decrease of significantly over time in the control group 39.8 ng/mL (range = 8.1-137.7 ng/mL). (P < .05). 100 - In the restricted group, 29% of the Work atituides andperfonnance. Four Pre Post subjects showed an increase in cotinine of the five measures of work satisfaction Time 0- cordro levels, with a mean increase of 28.4 and the overall well-being score had | Restricted ng/mL (range = 7.4-83.6 ng/mL). In the significant interactions in the repeated- control group, 67% ofthe subjects showed measures ANOVAs (P < .03). Posthoc Note. Bars reflect the standard deviation. an increase in cotinine levels, with a mean analyses shown in Table 2 revealed that The preban point represents the aver- age of four weekly data collections increase of 80.5 ng/mL (range = 1.3-227 restricted group members noted small but during weeks 1-4; the postban point ng/mL). A chi-square test on the number significant decreases on self-reports of represents the average of four weekly of with increasing vs decreasing data collections during weeks 5-8. Par- subjects concentration, (P < .05), productivity ticipants in the restricted group were cotinine levels was significant (X2 [1] = (P < .01), relations with coworkers (P < banned from smoking at the work site 9.32, P < .002). .05), and overall well-being (P < .05). during the postban period. Asterisks indicate differences between the means Control subjects showed no changes over according to Tukey's HSD. Both nicotine Subjective Reports time on the work-productivity indices. and CO levels changed significantly Only the overall well-being item had a across the time periods in the restricted Withdrawal symptom questionnaire. group (P < .01). significant between-group difference after Table 2 shows the mean values and the ban, with the restricted group report- standard deviations for the withdrawal ing a lower score (q = 4.39, P < .05). FIGURE 1-Saliva nicotine, items with significant interaction terms, expired breath carbon significant changes from the preban to the monoxide (CO), and postban period in the restricted group, Discussion saliva cotinine levels differences are shown for the and significant group during preban and postban the postban period. Restricted subjects This study showed reduced daytime time periods. reported significant increases in 4 of 18 smoking after implementation of a work- withdrawal symptom items (craving for place smoking ban. Specifically, smoking cigarettes, difficulty concentrating, depres- during work hours was reduced by an sion, and increased eating) after imple- average of four cigarettes per day. This from the before-work and after-work mentation of the smoking ban. However, reduction is consistent with the self-report assessment periods differ as a function of the magnitude of these changes across the data from Johns Hopkins Hospital smok- the smoking ban implementation. For the two time periods was less than or equal to ers2 and with the results reported for light restricted group, the mean butt weight per 0.5 points, and the total withdrawal score and moderate smokers in the Australian cigarette for cigarettes smoked outside did not increase significantly. Control Public Service.4 The present findings are workbefore the ban was 0.33 g (SD = 0.11 subjects, on the other hand, reported also consistent with a recent report of a g) vs 0.34 g (SD = 0.13 g) for the period significant decreases over time on five cross-sectional study by Kinne et al.,14 after the ban. For the control group, the withdrawal items (urges to smoke, irritabil- which found that men employed in work mean outside-of-work butt weights before ity, restlessness, impatience, headaches) sites with restrictive smoking policies

776 American Journal of Public Health May 1994, Vol. 84, No. 5 Work-Site Smoldng smoked fewer cigarettes overall than did men employed in settings without restric- TABLE 2-Preban and Postban Mean Values for Questionnaire Items tive smoking policies. We found that the end decrease in smoking measured at the Restricted Group (n = 34) Control Group (n = 33) of the work shift was corroborated by reductions in the biological measures of Weeks 1-4 Weeks 5-8 Weeks 1-4 Weeks 5-8 salivary nicotine and expired breath car- Questionnaire Item Mean (SD) Mean (SD) Mean (SD) Mean (SD) bon monoxide, both of which reflect Withdrawal symptom recent smoking exposure. Further, ques- questionnaire tions regarding locations where employ- Difficulty concentratingC 0.4 (0.5) 0.7 (0.6)a 0.6 (0.7) 0.5 (0.6) the expected shift Increasedeating 0.3 (0.4) 0.7 (0.9)a 0.6 (0.6) 0.7 (0.7) ees smoked revealed Depressionc 0.3 (0.4) 0.5 (0.6)a 0.6 (0.6) 0.4 (0.6) from smoking at the workstation to Craving for cigarettesb,c 1.3 (0.2) 1.8 (0.3)a 1.5 (0.3) 1.2 (0.2) smoking outside the building as a result of Urges to smokeb,c 1.8 (0.7) 1.9 (0.6) 1.8 (0.5) 1.5 (0.6)a the smoking ban. TotalwithdrawalscoreC 10.1 (6.3) 12.1 (8.2) 13.3 (7.6) 10.1 (7.3)a Although it is clear that exposure Total Mood Disturbance 10.0 (25.9) 14.4 (31.2) 18.8 (31.4) 10.3 (27.1) during the work shift declined, it is scorec important to consider the impact of the Work attitudes and ban on overall smoking behavior and performance strik- Ability to concentratec 8.6 (1.3) 7.9 (1 .3)a 8.2 (1.4) 8.0 (1.5) tobacco smoke exposure. The most ProductivityC 8.7 (1.4) 8.1 (1.5)a 8.1 (1.5) 8.1 (1.5) ing finding was that smokers did not Relationswith 9.2 (0.7) 8.5 (1.2)a 8.7 (1.1) 8.7 (1.2) compensate for the pack per week reduc- coworkersc tion either by increased smoking outside Enjoyment of workc 7.6 (1.6) 7.0 (1.7) 7.0 (2.2) 7.2 (2.0) of work or by more intensive smoking. Overall well beingb,c 7.9 (1.4) 7.2 (1 .7)a 7.6 (1.5) 7.9 (1.5) Thus, a reduction in overall exposure aSignificant posthoc test for preban vs postban period. might be expected. Cotinine levels, which bSignificant group differences at postban period. reflect accumulated nicotine exposure CSignificant interaction in analysis of variance. over several previous days, declined in the restricted group from 211 to 180 ng/mL, a 15% statistically nonsignificant reduction. Because this reduction suggested an over- nine levels must be interpreted in light of tendency for regression to the mean. all trend in reduced exposure, we further observed changes in overall tobacco con- Restricted smokers reported some difficul- examined the data for changes in indi- sumption. The work-shift smoking reduc- ties with work productivity and a decline vidual subjects' cotinine levels. Although tion observed was equivalent to one pack in general well-being. Further, scores on a full 71% (n = 24) showed some cotinine of cigarettes per week (4 cigarettes/ the craving-cigarettes and urges-to-smoke reduction, only 38% (n = 13) showed a day x 5 days/week). In a one-pack-per- items were significantly higher for the reduction of 50 ng/mL or more. The day smoker, this would amount to a 14% restricted group than for the control average postban cotinine level observed reduction in smoking exposure over a group during weeks 5-8 after the ban. among those who had reduced cotinine 1-week period. In fact, this smoking Although the magnitude ofaverage change levels was 167.6 ng/mL (SD = 97.0 ng/ reduction is consistent with the observed on these measures was quite small, sub- mL). This level was only slightly lower 15% reduction in cotinine levels and also jects exhibited notable individual differ- than 180 ng/mL, the average postban consistent with the conclusion that there ences. The number of restricted smokers level for the entire restricted group, and were no compensatory changes in smok- who had a mean total withdrawal symp- indicated continued moderate levels of ing topography or increases in numbers of tom score of at least "moderate" (> 18) smoking exposure despite the trend to- cigarettes smoked outside of work. Al- increased by 56% from the preban to the ward reduction in many subjects. though this is a small reduction in expo- postban period (four subjects at preban The increased mean cotinine level sure for individual subjects, at the popula- and nine at postban). In contrast, for the observed for control subjects was unex- tion level, a 15% reduction in group, the number of subjects pected (particularly in its magnitude) and exposure overall could confer an appre- with moderate withdrawal scores re- was influenced by several individuals with ciable reduced health risk given that there mained steady over time: six and five, especially large increases (21% ofsubjects is strong evidence for a dose-response respectively. It would be interesting to had cotinine increases of > 100 ng/mL). relationship between smoking and health follow the pattern of the withdrawal It is unlikely that this observed increase effects."5 symptoms for a longer duration to deter- reflects a population tendency toward Smokers undergoing an abrupt work- mine whether there are any lasting effects increasing cotinine levels over time. It is site smoking ban showed statistically of work-site smoking restrictions. The possible that some control subjects re- significant increases in four of the most importance and increasing popularity of duced their smoking initially as a result of reliably reported tobacco withdrawal work-site smoking restrictions underscore entering a study about smoking behavior. symptoms: difficulty concentrating, crav- the need to educate employees and A larger sample size would have ing cigarettes, increased eating, and de- employers alike about the side effects of been helpful in assessing the statistical pression. The subjective report scores of abstinence during the work shift. significance of reductions in cotinine control subjects tended to decrease over In summary, implementation of a levels resulting from a work-site smoking time, either because of increases in total ban on indoor work-site smoking ban. In terms of understanding the clini- smoking exposure (as suggested by the resulted in a decrease in smoking during cal significance, however, changes in coti- cotinine data) or because of a nonspecific the work shift, as verified by reduced

May 1994, Vol. 84, No. 5 American Journal of Public Health 777 Brigham et al. nicotine and carbon monoxide levels tions in smokers. Br Med J. 1980;280:972- measured at the end of the work shift. The References 976. 1. The Health Consequences of Involuntary 9. Henningfield JE, Griffiths RR. Effects of reported reduction in numbers of ciga- Smoking: A Report of the Surgeon General. ventilated cigarette holders on cigarette rettes smoked was equivalent to about Rockville, Md: Office on Smoking and smoking by humans. Psychopharmacology. one pack per week. The cotinine analysis Health; 1986. 1980;68:115-119. suggested, however, that the smoking ban 2. Stillman FA, Becker DM, Swank RT, et al. 10. Benowitz NL, Jacob P, Kozlowski LT, Yu intervention may provide limited immedi- Ending smoking at the Johns Hopkins L. Influence of smoking fewer cigarettes on ate health benefits to smokers themselves medical institutions. JAMA. 1990;2264: exposure to tar, nicotine, and carbon 1565-1569. monoxide. N Engl J Med. 1986;315:1310- in terms of tobacco exposure reduction, 3. Becker DM, Conner HF, Waranch R, et al. 1313. while producing minor withdrawal discom- The impact of a total ban on smoking in the 11. Young JC, Robinson JC, Rickert WS. A fort. Another potential health benefit for Johns Hopkins Children's Center. JAMA. study ofchemical deliveries as a function of smokers that has received some support 1989;262:799-802. cigarette butt length. Beitr Tabakforschung 4. Borland R, Chapman S, Owen N, Hill D. Int. 1981;11:87-95. in the literature is that work-site smoking 12. Hughes JR, Gust SW, Skoog K, Keenan bans have the additional effect of motivat- Effects of work place smoking bans on cigarette consumption. Am JPublic Health. RM, Fenwick JW. Symptoms of tobacco ing a small number of smokers to quit 1990;80:178-180. withdrawal: a replication and extension. completely.16 However, this effect was not 5. Hughes JR, Hatsukami DK Signs and Arch Gen Psychiatry. 1991;48:52-59. observed during the relatively brief assess- symptoms of tobacco withdrawal.Arch Gen 13. McNair DM, Lorr M, Droppleman LF. EdITS Manual for the Profile of Mood ment period of this investigation. C] Psychiatry. 1986;43:289-294. States. San Diego, Calif: Educational and 6. Hatsukami DK, Dahlgren L, Zimmerman Industrial Testing Service; 1971. R, Hughes JR. Symptoms of tobacco 14. Kinne S, Kristal AR, White E, Hunt J. Acknowledgments withdrawal from total cigarette cessation Work-site smoking policies: their popula- This study was supported by two National versus partial cigarette reduction. Psycho- tion impact in Washington State. Am J Institute on Drug Abuse grants, USPHS pharmacology. 1988;94:242-247. Public Health. 1993;83:1031-1033. DA03893 and T32-DA07209. 7. West RJ, Russell MAH, Jarvis MJ, Feyera- 15. Reducingthe Health Consequences ofSmok- An earlier version of this article was bend C. Does switching to an ultra-low ing, 25 Years of Progress: A Report of the presented as a Citation Paper at the annual nicotine cigarette induce nicotine with- Surgeon GeneraL Rockville, Md: Office on meeting of the Society of Behavioral Medicine, drawal effects? Psychopharmacology. 1984; Smoking and Health; 1989. Chicago, Ill, April 18-21, 1990. 84:120-123. 16. Sorsensen G, Rigotti N, Rosen A, Pinney J, We would like to thank an anonymous 8. Russell MAH, Jarvis M, Iyer R, Feyera- Prible R. Effects of a worksite nonsmoking reviewer who provided insightful commentary bend C. Relation of nicotine yield of policy: evidence for increased cessation. and recommendations. cigarettes to blood nicotine concentra- Am JPublic Health. 1991;81:202-204.

778 American Journal of Public Health May 1994, Vol. 84, No. 5