300 RESEARCH PAPER Socially cued smokig i bars, ightclubs, ad gamig veues: a case for itroducig smoke-free policies L Trotter, M Wakefield, R Borlad... Tobacco Cotrol 2002;11:300 304 See ed of article for authors affiliatios... Correspodece to: Lisa Trotter MA, Research ad Evaluatio Maager, 1 Rathdowe Street, Carlto Victoria 3053, Australia; Lisa.Trotter@cacervic.org.au Received 5 September 2002 ad revisio requested 3 July 2002. Accepted 8 September 2002... Objective: To assess smokers perceived effects of smokig bas i bars, ightclubs, ad gamig veues o their smokig behaviour. Desig: Cross sectioal structured iterview telephoe survey of a radom sample of smokers. Settig: Populatio survey i Victoria, Australia. Participats: The sample comprised 597 smokers ad aalyses were coducted o 409 smokers who reported patroisig bars, ightclubs or gamig veues at least mothly. Outcome measures: Two outcomes studied were socially cued smokig ad readiess to quit as a result of restrictios o smokig i social veues. Respodets were idetified as socially cued smokers if they reported attedig bars, ightclubs or gamig veues at least mothly ad said that they smoke more i these veues. The potetial ifluece of bas i social veues o readiess to quit was measured by askig respodets if they would be more or less likely to quit smokig if smokig were baed i hotels, licesed bars, gamig veues, ad ightclubs. Results: Of all adult smokers, 69% atteded bars, ightclubs or gamig veues at least mothly. Of these smokers, 70% reported smokig more i these settigs (socially cued smokers) ad 25% idicated they would be likely to quit if smokig were baed i social veues. Compared to smokers ot likely to quit if there were bas, smokers likely to quit were more likely to be socially cued (odds ratio (OR) 2.64), to be cotemplatig or preparig to quit (OR 2.22), to approve of bas i social veues (OR 2.44), ad to be aged uder 30 years (OR 1.73). Compared with smokers ot socially cued, socially cued smokers were more likely to be uder the age of 30 years (OR 6.15), more likely to believe that there is a safe level of cigarette cosumptio (OR 2.25), ad more likely to have previously made a quit attempt (OR 2.60). Coclusios: These fidigs suggest that bas o smokig i bars, ightclubs, ad gamig veues could reduce cigarette cosumptio ad icrease quittig amog smokers who frequetly patroise these settigs. These beeficial effects are likely to be strogest amog youger smokers. T he effect of restrictios o smokig i the workplace ad at home has bee well documeted i the literature, but there has bee little study of the effects of smokig policies o smokig behaviour i social settigs such as bars, ightclubs, ad gamig veues ( social veues ). Smokig restrictios i the workplace ad at home have bee foud to cotribute to reduced cosumptio, 12 itetios to quit, relapse prevetio, 3 ad possibly icreased cessatio, 45 as well as sed a message that smokig is socially uacceptable. 6 Although the mai purpose of smokig restrictios i public places is to reduce exposure to evirometal tobacco smoke, it may also have the effect of reducig tobacco cosumptio ad ecouragig quittig. The literature o situatioal iflueces o smokig relapse supports the otio that restrictios o smokig i veues such as bars ad ightclubs may remove the social pressure to smoke. Social situatios exert a powerful ifluece o smokig relapse, with aroud a quarter of relapse crises occurrig i settigs such as bars ad restaurats. 78 Shiffma 7 explaied this as beig caused by exposure to other smokers ivolvig both direct offers of cigarettes ad idirect pressure through observatio of other smokers as well as the ifluece of alcohol weakeig willpower to avoid smokig. Aother possible cosequece of restrictios o smokig i social veues is the prevetio or delay of smokig uptake. A relatively ew lie of evidece has begu to suggest that bas i public places may reduce smokig amog youg people. 910 Give that the people who patroise social veues are mostly youg, smokig bas i these settigs might also serve to iterrupt the process of progressio from experimetatio with smokig to log term tobacco depedece. Although may of the factors that ifluece the uptake of smokig are preset at a very youg age ad school aged childre are ofte targeted for itervetio, the period after leavig school has bee idetified as a critical time for trasitio to regular smokig. 11 12 The role of situatioal ifluece may be greatest early o whe smokig amog youg people is ot so iflueced by addictio. 13 Recreatioal veues that allow smokig expose youg people to cotexts i which smokig may be viewed as the orm. This may ecourage their progressio to more regular smokig. Pierce ad others 14 15 foud that exposure to smokers distiguished adolescets who progressed to established smokig from those who remaied experimeters. Further, the tobacco idustry is actively promotig tobacco i bars ad ightclubs which may cotribute to smokig uptake ad relapse back to smokig for those tryig to quit. Studies of tobacco idustry documets 16 17 have foud that bar promotios are geerally targeted at a youg adult audiece ad characterise cigarettes as beig part of a glamorous lifestyle that icludes attedace at ightclubs ad bars. This fidig is cosistet with the evidece from a recet study of tobacco promotios that foud a large icrease i etertaimet focused promotios mostly bar-club ad evet promotios. 18 May jurisdictios i Australia, Caada, the USA, ad other coutries have legislated smokig restrictios i public places. With few exceptios, smokig policies apply to restaurats but do ot exted to bars, ightclubs or gamig veues. The potetial effects of smokig restrictios i bars, ightclubs,
Socially cued smokig i bars 301 ad gamig veues o smokig behaviour may be better uderstood by examiig the behaviour, beliefs, ad opiios of smokers who frequetly patroise these veues. METHOD Data were draw from a telephoe survey coducted i November 2000 i the state of Victoria, Australia. The survey was coducted by a large market research compay. Oly respodets aged 18 years ad over were iterviewed. A total of 417 smokers were iterviewed from a radom sample of 2000 Victorias, of which smokers comprised 21%. A survey of 1000 Victorias coducted by aother compay usig the same method ad questioaire was used to boost the umber of smokers. Oly smokers were iterviewed ad they comprised 18% of this sample ( = 180). Thus, 597 smokers were iterviewed i total. The samplig frame for the survey was derived from a curret CD-ROM telephoe directory database. The data collectio occurred over a period of 10 days, icludig weeked days ad weekday eveigs. Respose rates have bee calculated as the proportio of all telephoe cotacts made that resulted i a useable iterview. There were 2000 useable iterviews (31%) out of 6488 umbers cotacted ad, for the boost sample, there were 1001 useable iterviews (46%) out of 2173 umbers cotacted. Part of the reaso for the differetial respose rates may have bee that the booster survey was shorter for o-smokers, comprisig oly a questio o tobacco use. There was o other obvious reaso for the differece i respose rates; however, a compariso of frequecy distributios betwee survey samples showed o differeces i sample compositio. Variables Smokig status was measured usig questios developed by a atioal expert committee coveed by the Australia Istitute of Health ad Welfare. 19 Smokers were defied as those who reported curret smokig either daily, weekly or less tha weekly. Socially cued smokers were defied as smokers who reported goig to either bars (described as hotel or licesed bar), ightclubs or gamig veues (described as gamig veue to play poker machies) ( social veues ) at least mothly ad reported that they smoke more i these veues. Not socially cued smokers were defied as smokers who go to social veues at least mothly ad do ot report that they smoke more i these veues. Smokig behaviour was described i terms of cosumptio 20 (< 5 or > 5 cigarettes per day) ad time to first cigarette for the day 21 (first cigarette for the day < 29 or > 29 miutes after wakig). These variables were combied to provide a idicatio of icotie depedece. Those cosidered addicted either smoked more tha five cigarettes per day or smoked their first cigarette of the day less tha 30 miutes after wakig. Smokers were desigated as beig i precotemplatio if they idicated that were ot plaig to quit i the ext six moths 22 ad smokers who did so respod were desigated as beig i cotemplatio/preparatio. To measure the potetial ifluece of bas i social veues o readiess to quit, respodets were asked if smokig were baed i hotels, licesed bars, gamig veues ad ightclubs, would you be more or less likely to quit smokig altogether? with respodets able to idicate more likely, less likely or o differece. Respodets were also asked do you thik that there is a safe umber of cigarettes that you ca smoke before it affects your health?. Respodets were asked if they approve or disapprove of the govermet baig smokig i bars, ightclubs or gamig veues. A variable was costructed that combied resposes for each of these variables ito two categories: approval/either approve or disapprove or disapproval. Demographic iformatio was also collected, icludig age Table 1 Frequecy distributio of patroage at social veues (=597) Bars (%) Nightclubs (%) Gamig veues (%) More tha weekly 13.4 2.7 2.8 At least oce a week 16.6 5.4 5.4 At least oce every 2 weeks 11.6 5.2 5.4 At least oce a moth 18.1 8.2 9.7 Less tha oce a moth 25.3 24.5 31.0 Never 14.6 54.1 45.7 Do t kow 0.5 0 0 (< 29 or > 29 years), sex, ad educatio (tertiary educatio more or secodary school ad lower less ). Statistical aalyses All aalyses were udertake usig the statistical package SPSS Versio 10.0.7. Covetioal χ 2 tests were used to test for associatios betwee likelihood of quittig if bas were i place, type of smoker (socially cued or ot socially cued), ad the variables of iterest. A sigificace level of 0.05 was adopted. All variables were iitially icluded i logistic regressio aalysis predictig key outcomes. Because of missig data, this led to a reductio i sample size, so the aalyses were repeated usig oly the sigificatly related variables (ad those almost sigificat) to maximise sample size. I o case did this affect the variables icluded ad the data reported are o the maximal sample. A compariso of the age ad sex distributio for the survey with Australia Bureau of Statistics populatio estimates for Victoria i 2000 23 idicated that the sample was represetative of the populatio except that wome (survey 59.4% v populatio 50.1%) ad people aged 60 years ad older (19.8% v 12.1% populatio) were over represeted i the sample. The sample was weighted by age ad sex accordig to populatio cesus data to estimate the proportio of smokers i Victoria who atted social veues mothly, smoke more i these veues, ad are likely to quit if there were bas i social veues. However, weightig procedures were ot used i the χ 2 ad logistic regressio aalyses. Research questios We sought to determie: (1) to what extet smokers who frequetly patroised social veues were likely to quit smokig if there were bas i social veues; (2) the characteristics of smokers who idicated they were likely to quit if there were bas i these social veues; ad (3) the characteristics of social smokers. RESULTS The proportio of smokers i Victoria who atted social veues mothly was estimated to be 69.4% (95% cofidece iterval (CI) 65.5% to 73.2%). Of this group, 70.1% (95% CI 65.5% to 74.6%) smoke more i social veues ad 25.4% (95% CI 21.1% to 29.9%) are likely to quit if smokig were baed i social veues. A criterio used to select the sample for further aalyses was frequecy of patroage at social veues. As ca be see i table 1, bars were visited at least mothly by 60% of smokers, ightclubs were visited at least mothly by 21% of smokers, ad gamig veues were visited at least mothly by 23% of smokers. The 409 (69%) smokers who reported at least mothly patroage of at least oe of these veues costituted the sample for the remaider of the aalyses. Table 2 shows the demographic characteristics, smokig behaviour, ad opiios of at least mothly ad less tha
302 Trotter, Wakefield, Borlad Table 2 Characteristics of those who patroise social veues at least mothly ad less tha mothly At least mothly Less tha (=409) mothly (=188p Value Percetage of total 597 68.5 31.5 Age (% <30) 594 33.5 9.0 0.000*** Sex (% female) 597 49.1 69.7 0.000*** Educatio (% less) 595 50.5 63.1 0.004** Nicotie depedece (% yes) 597 49.9 54.8 0.265 Made quit attempts (% yes) 591 79.3 81.2 0.588 Safe umber of cigarettes (% yes) 597 14.4 12.2 0.470 Bas (% approve/either) 579 64.7 75.8 0.008** Stage of chage (precotemplatio) 597 51.8 58.5 0.128 This differs from the level reported i the text because this is a uweighted figure. Table 3 Differeces betwee those likely or ot likely to quit if bas Likely (=100) Not likely (=309) p Value Odds ratio (95% CI) p Value Percetage of total 409 24.4 75.5 Age (% <30 years) 395 47.0 29.2 0.001** 1.73 (1.03 to 2.88) 0.037** Sex (% female) 398 45.0 50.7 0.326 Educatio (% less) 397 49.0 50.5 0.795 Nicotie depedece (%yes) 398 44.0 52.3 0.148 Made quit attempts (% yes) 394 84.8 77.3 0.109 Safe umber of cigarettes (% yes) 398 19.0 12.8 0.123 Socially cued smoker (yes) 398 85.0 64.8 0.000*** 2.64 (1.40 to 5.00) 0.003** Bas (%approve/either) 388 79.8 58.8 0.000*** 2.44 (1.39 to 4.30) 0.002** Stage of chage (precotemplatio) 398 35.0 56.4 0.000*** 2.22 (1.35 to 3.67) 0.002** This differs from the level reported i the text because this is a uweighted figure. CI, cofidece iterval. mothly patros. The results show that smokers who patroise social veues at least mothly were more likely to be youger, male, have higher educatioal attaimet, ad lower approval of bas tha smokers who patroise social veues less tha mothly. Factors associated with itetio to quit smokig if social veues became smoke-free were ivestigated (table 3). The results show that smokers who are likely to quit if there were bas i social veues were likely to be youger, socially cued (that is, smoke more i social veues), express greater approval of bas, ad be cotemplatig or preparig to quit, compared to those ot likely to quit if there were bas. Variables foud to be sigificatly associated with a icreased perceived likelihood of quittig i the bivariate aalyses were etered ito a logistic regressio aalysis ad 386 cases were icluded i the fial aalysis. Table 3 shows a model cotaiig four variables which provide a sigificat fit to the data (χ 2 = 44.16, df = 4, p = 0.000). Compared to smokers ot likely to quit i respose to smokig bas, smokers likely to quit were two ad a half times more likely to be socially cued (that is, to smoke more i these veues), twice as likely to be cotemplatig or preparig to quit, twice as likely to approve of bas i social veues, ad oe ad a half times more likely to be aged uder 30 years. Sice beig a socially cued smoker was strogly associated with itetio to quit if social veues became smoke-free, further aalyses were coducted to ivestigate the characteristics of this group. Table 4 idicates that compared to others, socially cued smokers were youger, had a lower idicatio of depedece, had previously tried to quit, ad believed there is a safe level of cigarette cosumptio. Variables foud to be sigificatly associated with membership of the socially cued smoker category i the bivariate aalyses were etered ito a logistic regressio aalysis ad 402 cases were icluded i the fial aalysis. Table 4 shows a model cotaiig three variables which provide a sigificat fit to the data (χ 2 = 54.17, df = 3, p = 0.000). Compared with smokers ot socially cued, socially cued smokers were six times more likely to be uder the age of 30 years, two times more likely to believe that there is a safe level of cigarette cosumptio, ad two ad a half times more likely to have previously made a quit attempt. Because socially cued smokig ad age are highly correlated we did additioal aalyses relatig predictors to age, but foud othig to suggest the results were due to residual cofoudig. DISCUSSION The fidigs from this study suggest that the smokig behaviour of a large proportio of smokers, especially youg smokers, may be iflueced by the impositio of smoke-free policies i bars, ightclubs, ad gamig veues. Overall, 69% of smokers report patroisig social veues at least mothly. The majority (70%) of smokers who frequetly patroise social veues report that they smoke more i these settigs (socially cued smokers) ad, to the extet that this is true, are likely to reduce their cosumptio overall if smokig were baed i social veues. Further, a quarter of smokers who frequetly patroise social veues reported that they would be more likely to quit smokig altogether if smokig was baed i hotels, licesed bars, gamig veues, ad ightclubs. The geeralisability of these results eeds to be cosidered i the light of modest respose rates. We suspect o-respodets might be more likely to atted social veues (thus beig harder to reach for surveyig), thereby possibly uderestimatig the proportio of smokers who atted social veues at least mothly.
Socially cued smokig i bars 303 Table 4 Differeces betwee socially cued smokers ad ot socially cued smokers Not cued Cued (=287)(=122) p Value Odds ratio (95% CI) p Value Percetage of total 409 70.2 29.8 Age (% <30 years) 406 42.5 12.4 0.000*** 6.15 (3.32 to 11.42) 0.000*** Sex (% female) 409 51.2 44.3 0.198 Educatio (% less) 408 50.3 50.8 0.931 Nicotie depedece (% yes) 409 45.6 59.8 0.009** Made quit attempts (% yes) 405 82.3 72.1 0.020** 2.60 (1.47 to 4.59) 0.001** Safe umber of cigarettes (% yes) 409 17.1 8.2 0.019** 2.25 (1.06 to 4.76) 0.034** Bas (% approve/either) 397 64.7 64.7 0.994 Stage of chage (precotemplatio) 409 49.8 56.6 0.213 Likely to quit if bas (more) 398 30.6 12.5 0.000*** This differs from the level reported i the text because this is a uweighted figure. CI, cofidece iterval. Our estimate of possible quittig at 25% is higher tha that reported by Philpot ad colleagues 24 who foud that 11.5% of people iterviewed while queuig for admissio to bars ad ightclubs (a youger sample tha ours) said that adoptio of smoke-free policies i hospitality veues would prompt them to quit. Our questio was ot as strogly worded, so a greater level of agreemet would be expected. Furthermore, youger people i our sample were also more likely to agree that bas would icrease likelihood of quittig ad this is cosistet with Philpot ad colleagues youger sample. Regardless of the exact level, a sigificat miority of smokers, especially socially cued ad youger smokers, believe bas i these veues will help them to quit. It may be the case that smokers act differetly i practice, as opposed to what they say they would do i respose to smokig bas. Therefore, a priority for subsequet research equiry would be to coduct studies to determie how much quittig ad declie i cosumptio is actually geerated whe smokig bas are implemeted i such veues. I this respect, observatioal studies of smokig behaviour i smokig permitted ad restricted veues, or cohort studies of how frequet atteders of social veues may chage their smokig patters followig the itroductio of smoke-free policies, would be iformative. Socially cued smokers are six times more likely to be uder the age of 30 tha o-socially cued smokers. Thus, the itroductio of smoke-free policies i bars, ightclubs, ad gamig veues could act as a strategy for prevetig the uptake of regular smokig. This possibility has already bee raised by some tobacco cotrol advocates who refer to bars ad ightclubs as icotie classrooms (G Coolly, persoal commuicatio, 2 November 2001). Recet evidece of tobacco idustry marketig which targets youg people i bars ad ightclubs 16 18 also supports this suggestio. The group of socially cued smokers we idetified are likely to be sigificat beeficiaries of smoke-free policies i social veues. As a group, they are youg ad hold beliefs that low levels of smokig are ot particularly harmful. As a result it would seem that they thik what they are doig is safe. This might be so, i relative terms, if they were ot puttig themselves at risk of depedece ad the subsequet harmful log term use this etails. The fidigs from this study suggest that a reaso for strog tobacco idustry oppositio to smoke-free policies i bars, ightclubs, ad gamig veues may be because of their cocer at the possibility that it will ecourage cessatio ad remove a cotext where may youg people are iduced to try smokig. Further research to evaluate the effects of smokefree policies i these veues o smokig behaviour is required. What this paper adds Restrictios o smokig i the workplace ad at home reduce levels of smokig i adults. We sought to determie if this may also be the case for recreatioal veues such as pubs ad clubs. A cross sectioal survey foud that 69% of smokers report patroisig bars, ightclubs or gamig veues at least mothly, ad 70% of those who patroise social veues at least mothly report smokig more i these settigs (socially cued smokers). These people are aged uder 30 years, have made previous quit attempts, ad believe there is a safe umber of cigarettes that ca be smoked before their health ca be affected. Further, 25% of smokers who frequetly patroise social veues report that they would be more likely to quit altogether if there were bas i these veues. These people are likely to be aged uder 30 years, cotemplatig or preparig to quit, ad approve of bas i social veues. These fidigs suggest that smokig restrictios i social veues may reduce smokig prevalece amog patros.... Authors affiliatios L Trotter, M Wakefield, Cetre for Behavioural Research i Cacer, The Cacer Coucil Victoria, Victoria, Australia R Borlad, VicHealth Cetre for Tobacco Cotrol, The Cacer Coucil Victoria REFERENCES 1 Chapma S, Borlad R, Scollo M, et al. The impact of smoke-free workplaces o decliig cigarette cosumptio i Australia ad the Uited States. Am J Public Health 1999;89:1018 23. 2 Owe N, Borlad R. Delayed compesatory cigarette cosumptio after a workplace smokig ba. Tobacco Cotrol 1997;6:131 5. 3 Gilpi E, White M, Farkas A, et al. Home smokig restrictios: which smokers have them ad how they are associated with smokig behaviour. Nicotie Tobacco Research 1999;1:153 62. 4 Farkas A, Gilpi E, Distefa J, et al. The effects of household ad workplace smokig restrictios o quittig behaviours. Tobacco Cotrol 1999;8:261 5. 5 Bieer L, Nyma A. Effects of workplace smokig policies o smokig cessatio: results of a logitudial study. Occup Eviro Med 1999;41:1121 7. 6 Borlad R, Mullis R, Trotter L, et al. Treds i evirometal tobacco smoke restrictios i the home i Victoria, Australia. Tobacco Cotrol 1999;8:266 71. 7 Shiffma S. Relapse followig smokig cessatio: a situatioal aalysis. Cosult Cli Psychol 1982;50:71 86. 8 Borlad R, Slip-ups ad relapse i attempts to quit smokig. Addictive Behaviors 1990;15:235 45. 9 Wakefield M, Chaloupka F, Kaufma N, et al. Effect of restrictios o smokig a home, at school, ad i public places o teeage smokig: cross sectio study. BMJ 2000;321:333 7. 10 Farkas A, Gilpi E, White M, et al. Associatio betwee household ad workplace smokig restrictios ad adolescet smokig. JAMA 2000;284:717 22.
304 Trotter, Wakefield, Borlad 11 Schofield P, Borlad R, Hill D, et al. Istability i smokig patters amog school leavers i Victoria, Australia. Tobacco Cotrol 1998;7:149 55. 12 Hill D, Borlad R. Adults accouts of oset of regular smokig: Iflueces of school, work, ad other settigs. Public Health Reports 1991;106:181 5. 13 Schofield PE, Pattiso PE, Hill DJ, et al. Youth culture ad smokig: Itegratig social group processes ad idividual cogitive processes i a model of health-related behaviours. J Health Psychol (i press). 14 Pierce, JP, Choi WS, Gilpi EA, et al. Validatio of susceptibility as a predictor of which adolescets take up smokig i the Uited States. Health Psychol 1996;15:355 61. 15 Choi WS, Pierce JP, Gilpi EA, et al. Which adolescet experimeters progress to established smokig i the Uited States. Am J Prev Med 1997;13:385 91. 16 Katz SK, Lavack AM. Tobacco-related bar promotios: Isights from tobacco idustry documets. Tobacco Cotrol 2002;11(supp I):i92 101. 17 Sepe E, Lig P, Glatz S. Smooth moves: bar ad ightclub tobacco promotios that target youg adults. Am J Public Health 2002;92:414 19. 18 Sepe E, Glatz SA. Tobacco promotios i alterative press: targetig youg adults. Am J Public Health 2002;92:75 8. 19 Australia Istitute of Health ad Welfare. Natioal health data dictioary. Versio 8.0 AIHW Catalogue No. HWI 18. Australia Istitute of Health ad Welfare: Caberra, 1999. 20 Shiffma S. Tobacco chippers idividual differeces i tobacco depedece. Psychopharmacology (Berl) 1989;97:539 47. 21 Fagerstrom K-O, Scheider N. Measurig icotie depedece: a review of Fagerstrom tolerace questioaire. J Behav Med 1989;12:159 81. 22 Prochaska JO, DiClemete CC, Norcross JC. I search of how people chage. Am Psychol 1992;47:1102 14. 23 Australia Bureau of Statistics. Populatio by age ad sex, Australia states ad territories, Catalogue No. 3201.0, Jue 2000. 24 Philpot S, Rya S, Torre L, et al. Effect of smoke-free policies o the behaviour of social smokers. Tobacco Cotrol 1999;8:278 81. DOONESBURY 2002 G.B. Trudeau. Reprited with permissio of UNIVERSAL PRESS SYNDICATE. All rights reserved