& Research Tobacco Research Advance Access published January 22, 2010 NicotineNicotine & Tobacco Original Investigation Secondhand smoke drift: Examining the influence of indoor smoking bans on indoor and outdoor air quality at pubs and bars Emily Brennan,1 Melissa Cameron,1 Charles Warne,1 Sarah Durkin,1 Ron Borland,1 Mark J. Travers,2 Andrew Hyland,2 & Melanie A. Wakefield1 1 2 Cancer Council Victoria, Carlton, Australia Roswell Park Cancer Institute, Buffalo, NY Corresponding Author: Melanie Wakefield, Centre for Behavioural Research in Cancer, Cancer Council Victoria, 1 Rathdowne Street, Carlton Vic 3053, Australia. Telephone: (+61) 3-9635-5046; Fax: (+61) 3-9635-5380; E-mail: melanie.wakefield@cancervic.org.au Received August 24, 2009; accepted December 8, 2009 Abstract Introduction: This study aimed to examine the influence of indoor smoking bans on indoor and outdoor air quality at pubs and bars and to assess whether secondhand tobacco smoke (SHS) drifts from outdoor smoking areas to adjacent indoor areas. Methods: Data were covertly collected from a convenience sample of 19 pubs and bars that had at least 1 indoor area with an adjacent semi-enclosed outdoor eating/drinking area. Using TSI SidePak Personal Aerosol Monitors, concentrations of SHS (PM2.5) were measured concurrently in indoor and outdoor areas before and after implementation of the indoor smoking ban. Information was collected about the number of patrons and lit cigarettes and about the enclosure of outdoor areas. Results: Indoor PM2.5 concentrations reduced by 65.5% from preban to post-ban (95% CI 32.6%–82.3%, p = .004). Outdoor exposure to PM2.5 also reduced from pre-ban to post-ban by 38.8% (95% CI 3.2%–61.3%, p = .037). At post-ban, indoor concentrations of PM2.5 were positively associated with outdoor concentrations. After adjustment for covariates, a 100% increase in geometric mean (GM) outdoor PM2.5 was associated with a 36.1% (95% CI 2.4%–80.9%) increase in GM indoor PM2.5 exposure (p = .035). Discussion: Indoor smoking bans are an effective means of improving indoor and outdoor air quality in pubs and bars, although the air quality of smoke-free indoor areas may be compromised by smoking in adjacent outdoor areas. These findings require consideration in efforts to ensure adequate protection of the health of employees and patrons at hospitality venues. Introduction Recognizing the substantial health risks associated with secondhand tobacco smoke (SHS) exposure (U.S. Department of Health and Human Services, 2006) and consistent with the recommendations of Article 8 of the World Health Organization’s Framework Convention on Tobacco Control (FCTC; World Health Organization, 2003), many nations have implemented comprehensive smoke-free legislation in indoor public places, including hospitality venues. These policies substantially reduce indoor SHS exposure (Hyland, Travers, Dresler, Higbee, & Cummings, 2008) and lead to significant improvements in health outcomes (Centers for Disease Control and Prevention, 2009; Eagan, Hetland, & AarØ, 2006; Khuder et al., 2007; Sargent, Shepard, & Glantz, 2004). SHS exposure is of particular concern for people employed in the hospitality industry as they experience some of the highest levels of occupational SHS exposure, and these workplaces are typically among the last to be regulated for indoor smoking (Davis, 1998; Eagan et al.; Eisner, Smith, & Blanc, 1998; Howard, 2004; Wakefield et al., 2003). Hospitality workers and patrons also continue to be exposed to SHS in outdoor and “quasi-outdoor” areas (outdoor areas bound by some combination of walls and/or a roof) at hospitality venues where smoking is typically unregulated. There is evidence that outdoor SHS levels are comparable to indoor concentrations under certain conditions (Klepeis, Ott, & Switzer, 2007; Travers, Higbee, & Hyland, 2007), although outdoor concentrations are more sensitive to wind conditions and the proximity of smokers and therefore do not accumulate as readily (Klepeis et al.; Repace, 2005). There are also indications that SHS from outdoor and quasi-outdoor areas may drift into adjacent indoor areas, thereby affecting indoor air quality at smoke-free venues. For instance, evaluation of the indoor smoking bans in Ireland indicated that bars with an assigned outdoor smoking area had higher postban airborne nicotine concentrations, suggesting that outdoor SHS was migrating indoors (Mulcahy, Evans, Hammond, Repace, & Byrne, 2005). This outdoor to indoor drift may particularly be a problem when the outdoor area directly adjoins the indoor area. It may also be problematic when the outdoor doi: 10.1093/ntr/ntp204 © The Author 2010. Published by Oxford University Press on behalf of the Society for Research on Nicotine and Tobacco. All rights reserved. For permissions, please e-mail: journals.permissions@oxfordjournals.org 1 Secondhand smoke drift area is an enclosed space since the presence of overhead roofs and covers have been shown to increase SHS concentrations in outdoor dining areas by around 50% (Cameron et al., 2009). In the state of Victoria, Australia, smoking restrictions in hospitality venues have been gradually implemented over the past 10 years, with bans in indoor areas of pubs and bars implemented in July 2007 (Tobacco [Amendment] Act, 2000, 2005; Tobacco [Further Amendment] Act, 2001). This legislation stipulates that outdoor dining and drinking areas are not required to be smoke-free unless the area has a roof covering at least half of the notional roof area and walls covering at least 75% of the notional wall area (Tobacco [Amendment] Act, 2005). Under this legislation, roofs are defined as any permanent or temporary structure that impedes upward airflow. Walls are defined as any structure that impedes lateral airflow, and the definition of the notional wall area is the total wall surface that would exist if all walls were continuous, were at the perimeter of the roofed area, and were all of a height equal to the lowest height of the roof. Anticipating that the provision of an outdoor smoking area would be important for attracting and retaining patrons once the bans were implemented, many Victorian hospitality venues developed or renovated their existing outdoor areas in accordance with the new regulations (Australian Hotels Association, 2006, 2008; Lucas & Houston, 2007). As such, many of the outdoor areas in Victoria that permit smoking are highly enclosed with substantial wall and roof cover. To ensure adequate protection against the risks associated with SHS exposure, the recommendations and supporting guidelines of Article 8 of the FCTC advocate that jurisdictions begin to give greater consideration to the implementation of smoking bans in outdoor and quasi-outdoor areas. The guidelines also advise that there is a need for scientific evidence to further demonstrate and quantify the health hazards associated with SHS exposure in various outdoor settings (World Health Organization, 2003, 2009). Consequently, our study of pubs and bars had three objectives. First, we sought to corroborate previous research findings, indicating that indoor smoking bans significantly improve indoor air quality (Hypothesis 1). Second, we sought to evaluate the influence of indoor smoking bans on outdoor air quality and we hypothesized that there would be an increase in outdoor exposure to SHS given the expected increase in the number of smokers using these areas at post-ban (Hypothesis 2). Third, extending the findings of Mulcahy et al. (2005) in Ireland, we also sought to assess whether SHS migrates from outdoor smoking areas to adjacent indoor areas and we predicted that indoor SHS concentrations at post-ban would be positively associated with outdoor SHS concentrations at post-ban (Hypothesis 3). Methods Venue sample We selected a purposive sample of 20 pubs and bars located within a 7-km radius of the center of Victoria’s capital city. Eligible pubs and bars had at least one indoor area with an adjacent semi-enclosed outdoor eating/drinking area that was connected by a direct access (i.e., a doorway or large open window, as opposed to a corridor or staircase). In this study, semi-enclosed 2 outdoor areas were a defined area bound by walls (any permanent or temporary structure that impeded lateral airflow) and/or by a roof (any permanent or temporary structure that impeded upward airflow). By this definition, tables on the footpath outside venues were not considered semi-enclosed outdoor areas. There were no size requirements (minimum or maximum) for indoor and outdoor areas to be included in the sample. Data collection procedure Each venue was visited before (pre-ban) and after (post-ban) the indoor smoking ban was implemented in July 2007. Pre-ban data collection occurred over 6 weekends in March and early April 2007, and post-ban data were collected over 5 weekends in November and early December 2007. The mean daily minimum and maximum temperatures were similar in March (minimum, 15.2 °C and maximum, 25.8 °C) and November (minimum, 14.3 °C and maximum, 24.4 °C). Venues were visited on Friday and Saturday nights between approximately 7 p.m. and midnight when venues were expected to be at their busiest. Each venue was visited by two pairs of researchers (an indoor pair and outdoor pair) to allow concurrent collection of indoor and outdoor air quality data. At the beginning of each venue visit, a 5-min sample of air quality data was collected from an area on the footpath away from the venue and from crowds and smokers. Providing a measure of the ambient air quality in the environment surrounding the venue, these data were included in all the statistical analyses conducted in this study in order to minimize the influence of ambient sources of particulate matter on the levels attributable to SHS. Upon entering the venue, researchers were instructed to choose a location that met several criteria: Each research pair had to be situated within 5 m of the main access (doorway or large window) between the indoor and outdoor areas, indoor and outdoor pairs were required to be in view of one another, and all researchers had to be able to sit or stand as though they were regular patrons visiting the venue. Air quality and observational data were collected for 30 min at each venue, after which time researchers left the venue. The data collection procedure was identical at pre-ban and post-ban, and at preban a sketch of the venue was made, indicating the position of the air quality monitor, so that at post-ban researchers could locate the monitor as close as possible to the pre-ban location. Because a key objective of this study was to measure air quality in pubs and bars under natural conditions, venue proprietors were not informed about the study or our visit to their venue, and several measures were taken to ensure that data collection was undetected by owners and patrons (Petticrew et al., 2007). Air quality data The concentration of particulate matter smaller than 2.5 m in diameter (PM2.5) served as the indicator of air quality. Air quality data were collected using two TSI SidePak AM510 Personal Aerosol Monitors. Before each data collection session, the SidePaks were prepared according to the procedure described by Hyland et al. (2008). A standard calibration factor of 0.32 was applied to the raw data to correct for the properties of SHS (Hyland et al.; Travers et al., 2007). Because two separate SidePaks were used to allow concurrent data collection, an additional adjustment factor was applied to the raw data from each monitor to account for the mean difference between the two monitors, as calculated by a series of experiments comparing Nicotine & Tobacco Research the monitors in various environments (across measurements, the difference between the monitors was less than 10%). Data were recorded by the SidePaks at 30-s intervals, with each 30-s data point being an average of the previous thirty 1-s measurements. To facilitate unobtrusive data collection, a piece of tubing was attached to the inlet of the SidePak, and the SidePak was placed in a camera-like carry bag with the tubing protruding from the bag by only 1 inch. The bag containing the SidePak was either carried over the researcher’s shoulder or placed on a table for the entire data collection period. Observational data Observational data were collected for both indoor and outdoor areas. At 5-min intervals (0–25 min), researchers recorded counts of the number of patrons and the number of lit cigarettes used to create a measure of the active smoking prevalence at each 5-min timepoint (active smoking prevalence = number of patrons with lit cigarettes/total number of patrons) and whether or not the main access between indoors and outdoors was open. Further observational data collected once per venue included whether there were windows or doors open in the indoor area (additional to the main access) and whether a kitchen was operating, as cooking sources also produce PM2.5 particles (Hyland et al., 2008). Additional observational data for outdoor areas included the presence of overhead covers (including trees, umbrellas, shade cloth, building awnings, and roofing) and estimates of the proportion of overhead and wall space that was enclosed. Using the estimates of enclosed overhead space and wall space, the degree of enclosure at each venue was calculated and categorized according to one of three levels of outdoor enclosure: (a) low, roof <50% covered and any level of wall enclosure; (b) intermediate, roof ≥50% covered and walls <75% enclosed; and (c) high, roof ≥50% covered and walls ≥75% enclosed. Statistical analysis Air quality data collected at each indoor and outdoor location were averaged across the 30-min recording period. A measure of the average ambient concentration of PM2.5 was also obtained for the environment around each venue, taking the average of the ambient data collected simultaneously by the two SidePaks prior to entry. All statistical analyses used log-transformed mean PM2.5 concentrations due to the positively skewed distribution of the data. Geometric means (GM) and geometric standard deviations (GSD) are reported in addition to arithmetic means (mean; SD) and minimum and maximum PM2.5 concentrations. Bivariate and multivariate linear regressions were used to test the hypotheses and to examine the association between active smoking prevalence and PM2.5 exposure levels. In the final multivariate models, covariates were retained when bivariatelevel associations existed or there were substantive reasons for their inclusion. Log-transformed mean ambient PM2.5 concentration was a covariate in all analyses. To adjust for clustering between repeated measures at each venue, we used robust SEs. All tests were two tailed. Due to violation of the indoor smoking ban, one venue was excluded from the analytic sample, reducing the sample size to 19. It is likely that this violation of the ban was unintentional and is attributable to the layout of the venue, which features a large and direct opening without any doors or barriers to distinguish the indoor and outdoor areas. Results Across venues (n = 19), the mean number of patrons was similar before (30.1) and after the ban (26.4), t(18) = 1.07, p = .297. As expected, the mean prevalence of active smoking (number of patrons with lit cigarettes/total number of patrons) in indoor areas declined significantly from pre-ban (4.7%) to post-ban (0.0%). Indoor PM2.5 concentrations were associated with the mean number of patrons with lit cigarettes indoors after adjusting for ambient exposure levels and time (pre-ban or post-ban) such that an increase of one in the mean number of lit cigarettes was associated with an 854.3% increase in GM indoor PM2.5, t(18) = 4.79, p < .001. Mean active smoking prevalence in outdoor areas did not change significantly from pre-ban (6.2%) to post-ban (7.3%), t(18) = −0.86, p = .401. The mean number of lit cigarettes outdoors was associated with PM2.5 concentrations (after adjusting for covariates) such that an increase of one in the mean number of lit cigarettes related to an increase in GM outdoor PM2.5 of 23.9%, t(18) = 2.22, p = .040. The number of venues at which the main access was open between indoors and outdoors was consistent at pre-ban (n = 15) and post-ban (n = 16), as was the number of indoor areas with an additional open window or door (pre-ban n = 2 and postban n = 4). The number of venues that had a kitchen operating was higher at post-ban (n = 14 compared with n = 6), while there were no differences in the number of venues with each level of outdoor enclosure. Nine venues had low enclosure (<50% roof coverage), two at pre-ban and one at post-ban had intermediate enclosure, and eight venues at pre-ban and nine at post-ban had high enclosure (≥50% roof coverage and ≥75% wall coverage). Impact of the smoking ban on indoor air quality Table 1 shows that there was a 71.6% reduction in the unadjusted indoor GM PM2.5, from 61.3 mg/m3 at pre-ban to 17.4 mg/m3 at post-ban. After adjusting for covariates, ban implementation was associated with a significant reduction in GM PM2.5 of 65.5% (Table 2). Impact of the smoking ban on outdoor air quality Contrary to predictions, unadjusted GM PM2.5 exposure in outdoor areas reduced by 31.1% from 19.0 mg/m3 at pre-ban to 13.1 mg/m3 at post-ban (Table 1). After adjustment for ambient PM2.5, implementation of the smoking ban was associated with a 38.8% reduction in GM outdoor PM2.5 exposure (Table 2). There was no evidence of an association between outdoor air quality and the level of enclosure of the outdoor area at either the bivariate or multivariate level. Association between outdoor air quality and indoor air quality at post-ban As predicted, indoor air quality was associated with outdoor air quality at post-ban. When adjusting for ambient PM2.5 levels and the presence of an open or closed door (i.e., the main access) between indoor and outdoor areas, a 100% increase in GM outdoor PM2.5 exposure was associated with a 36.1% rise in GM indoor PM2.5 exposure (Table 2). 3 Secondhand smoke drift Table 1. Indoor and outdoor PM2.5 exposure levels at pre-ban and post-ban Indoor PM2.5 exposure (mg/m3) GM (GSD) M (SD) Minimum Maximum Ambient PM2.5 GM (GSD) Outdoor PM2.5 exposure (mg/m3) Pre-ban Post-ban % Reductiona Pre-ban Post-ban % Reductiona 61.3 (3.1) 102.7 (98.7) 6.4 338.1 4.5 (1.7)b 17.4 (2.4) 25.8 (30.7) 2.7 136.3 5.3 (1.7)b 71.6 74.9 19.0 (2.6) 32.1 (43.3) 3.6 162.0 13.1 (2.4) 18.8 (17.8) 3.7 73.7 31.1 41.4 Note. One venue was excluded from the analytic sample. At this venue, pre-ban indoor PM2.5 = 40.1, post-ban indoor PM2.5 = 249.7, pre-ban outdoor PM2.5 = 14.0, and post-ban outdoor PM2.5 = 295.5. When included in sample (n = 20), pre-ban indoor PM2.5 (GM) = 60.0, post-ban indoor PM2.5 (GM) = 19.9, pre-ban outdoor PM2.5 (GM) = 18.6, and post-ban outdoor PM2.5 (GM) = 15.3. GM = geometric mean; GSD = geometric standard deviation. a Unadjusted for covariates. b The same measure of mean ambient PM2.5 exposure serves as a covariate in analyses examining both indoor and outdoor exposure levels. However, contrary to expectations that an open door facilitates outdoor to indoor SHS drift, this multivariate regression also indicated that having an open access between indoors and outdoors was associated with lower levels of GM indoor PM2.5 exposure compared with at those venues where the door was closed (Table 2). Descriptive analyses comparing door-closed and dooropen venues at post-ban were used to investigate this unexpected finding. Outdoor mean active smoking prevalence levels were higher at those venues where the door was closed (n = 3, 11.7%) than at those venues where the door was open (n = 16, 6.4%), t(17) = 2.17, p = .044. At post-ban, the outdoor GM PM2.5 concentration was 7.3 mg/m3 at door-closed venues and 14.6 mg/m3 at door-open venues, t(17) = −1.28, p = .217, while the indoor GM PM2.5 concentration was 27.5 mg/m3 at door-closed venues and 16.0 mg/m3 at door-open venues, t(17) = 1.01, p = .326. Discussion Consistent with previous evaluations of indoor smoke-free policies, this study demonstrated that the smoking ban in Victorian pubs and bars reduced both indoor and outdoor exposure to SHS. This study also demonstrates the potential for substantially greater exposure levels than those observed here, in finding that GM indoor and outdoor SHS concentrations may be expected to increase by almost 900% and 25%, respectively, for every increase of one in the mean number of lit cigarettes. Our concurrent measurement of indoor and outdoor air quality also indicated that SHS may drift from outdoors to indoors as post-ban indoor concentrations were positively and significantly associated with SHS concentrations in the adjacent outdoor areas. Extending the findings of Mulcahy et al. (2005) and consistent with earlier studies that observed SHS drift between partitioned smoking and nonsmoking indoor areas (Cains, Cannata, Poulos, Ferson, & Stewart, 2004; Pion & Givel, 2004), the findings from this observational study indicate that the benefits of indoor smoking bans may be lessened when smoking is permitted in adjacent outdoor and quasi-outdoor areas. Outdoor to indoor SHS drift may be partly attributable to the high degree of enclosure of many of the outdoor smoking Table 2. Relation between indoor smoking ban and indoor/outdoor PM2.5: Multivariate linear regression analyses Hypothesis 1: Impact of smoking ban on indoor PM2.5 Smoking ban Window open Kitchen operating Log (ambient PM2.5) Constant Factor change in geometric mean indoor PM2.5 0.345 0.507 0.607 1.725 33.810 p Value .004 .017 .135 .051 <.001 95% CI 0.177–0.674 0.295–0.873 0.310–1.187 0.998–2.981 12.169–93.937 Hypothesis 2: Impact of smoking ban on outdoor PM2.5 Smoking ban Log (ambient PM2.5) Constant Factor change in geometric mean outdoor PM2.5 0.612 2.09 6.247 p Value .037 .012 <.001 95% CI 0.387–0.968 1.198–3.646 2.685–14.536 Hypothesis 3: Impact of outdoor PM2.5 on indoor PM2.5 at post-ban Log (outdoor PM2.5) Door open Log (ambient PM2.5) Constant Factor change in geometric mean indoor PM2.5 1.560 0.260 2.251 4.440 p Value .035 .006 .035 .014 95% CI 1.036–2.351 0.106–0.640 1.067–4.749 1.418–13.901 4 Nicotine & Tobacco Research areas included in this study. More than half of the venues had overhead covers occupying at least half of the total roof area, and the majority of these venues also had walls enclosing more than three quarters of the total wall space. Highly enclosed outdoor areas may reduce the possibility of SHS naturally dissipating outdoors (Cameron et al., 2009) such that it is forced to drift into the adjacent indoor space. However, potentially limiting this explanation of outdoor to indoor drift is the unexpected finding that post-ban indoor concentrations were significantly lower at venues where the door between indoors and outdoors was open compared with at those three venues where the door was closed. While our ability to explain this finding is limited, our examination of observational data indicated that the prevalence of active smokers was higher in the outdoor areas at which the door was closed compared with venues at which the door was open. It is therefore plausible that having the door closed may have been an attempt to minimize the amount of SHS drifting indoors. Alternatively, this finding may indicate that open doors increase the ventilation of indoor areas, leading to reduced PM2.5 exposure. Further research examining the association between outdoor smoking areas and indoor air quality could shed further light on these possibilities. An unexpected finding was the reduction in outdoor SHS concentrations. While it was anticipated that the indoor smoking ban would move smokers outdoors and increase the concentration of SHS in outdoor areas, we found no change in the prevalence of active smoking in outdoor areas and implementation of the ban was associated with reduced outdoor SHS. Given that this unexpected reduction is not attributable to changes in the degree of enclosure of outdoor areas, or to changes in overall patronage to these hospitality venues, one possible explanation may be the overall decrease in the prevalence of active smoking at these venues. While smoking prevalence in outdoor areas was not affected by the ban, the elimination of smokers from indoor areas means that there were fewer people overall smoking at these venues at post-ban. Considering the findings suggesting that SHS may have been drifting from outdoors to indoors at post-ban, it seems possible that there may also have been air exchange between the two areas at pre-ban such that indoor SHS was contributing to outdoor concentrations. If this was the case, then the reduction in outdoor SHS exposure may be the result of the overall reduction in active smoking prevalence associated with implementation of the indoor ban. We acknowledge that SHS is not the only source of PM2.5 particles, with additional contributions coming from diesel vehicles, heaters, and cooking sources, such as the kitchens that were operating at 6 venues at pre-ban and at 14 venues at post-ban. However, it is a strength of this study that the influence of additional PM2.5 sources was minimized by controlling for ambient PM2.5 concentrations in all analyses. A second potential study limitation is that we did not account for wind conditions in our examination of outdoor PM2.5 concentrations. Previous studies of outdoor SHS concentrations have demonstrated that exposure levels are particularly sensitive to wind speed and direction (Klepeis et al., 2007; Repace, 2005), and we recommend that future research use venue-specific wind measures to account for these effects. That said, the high level of enclosure of many of the outdoor areas in this study may have limited the potential influence of wind conditions on the outdoor SHS exposure levels detected in this study. Third, the limited duration of data collection means that these data do not reflect the total exposure levels that will be experienced by employees of hospitality venues and we suggest that future research consider greater use of methods that more accurately capture hospitality worker exposure to SHS (Semple et al., 2007). Another aspect of the present study worth noting is our use of a convenience sample of venues meeting specific criteria for the layout of the venue, as this means that the observed exposure levels may not be representative of those present in all pubs and bars, particularly when the indoor and outdoor areas are not directly connected by a doorway or the outdoor area is not enclosed by walls or overhead covers. However, restricting our sample to venues meeting these criteria facilitated our examination of SHS drift, and this concurrent data collection has also allowed us to draw conclusions about the influence of the ban on active smoking prevalence and overall patronage to hospitality venues. Our study indicated no overall change in patronage between preban and post-ban and an overall reduction in active smoking prevalence at this sample of hospitality venues. These findings are consistent with economic data collected in other jurisdictions and with smokers’ reports about their own behavior, which all suggest that indoor smoking bans do not negatively affect the patronage and revenue of hospitality venues (Huang & McCusker, 2004; Lal, Siahpush, & Scollo, 2004; McCarthy, Durkin, Brennan, & Germain, 2008; Scollo, Lal, Hyland, & Glantz, 2003). This observational study adds to the substantial body of evidence illustrating that indoor smoking bans improve the air quality in hospitality venues. However, our study is also one of the first to demonstrate that the air quality of smoke-free indoor areas may be compromised by smoking in adjacent outdoor areas. While we do not advise that the present results are used to advocate outdoor smoking restrictions at the expense of other policies known to reduce SHS exposure (Eriksen & Cerak, 2008) and smoking prevalence (Davis, Wakefield, Amos, & Gupta, 2007; National Cancer Institute, 2008), in light of the evidence that there is no risk-free level of SHS exposure (U.S. Department of Health and Human Services, 2006), these findings must be considered to ensure the adequate protection of the health of employees and patrons at hospitality venues. Funding This work was supported by Quit Victoria, Cancer Council Victoria, and the Flight Attendant Medical Research Institute. Declaration of Interests None declared. Acknowledgments The authors thank the research assistants who collected the data. 5 Secondhand smoke drift References Australian Hotels Association. (2006). OURhotel—special local government edition, 2006, Retrieved 28 April 2009, from http:// www.aha.org.au/Documents/AHALGEdition.pdf Australian Hotels Association. (2008). How to responsibly manage indoor smoking bans. Information and case study booklet. OURhotel—special local government edition, 2008, Retrieved 28 April 2009, from http://www.aha.org.au/booklet2008.pdf Cains, T., Cannata, S., Poulos, R., Ferson, M. J., & Stewart, B. W. (2004). Designated “no smoking” areas provide from partial to no protection from environmental tobacco smoke. Tobacco Control, 13, 17–22. Cameron, M., Brennan, E., Durkin, S., Borland, R., Travers, M. J., Hyland, A., et al. (2009). Secondhand smoke exposure (PM2.5) in outdoor dining areas and its correlates. Tobacco Control, doi: 10.1136/tc.2009.030544. Centers for Disease Control and Prevention. (2009). Reduced hospitalizations for acute myocardial infarction after implementation of a smoke-free ordinance—City of Pueblo, Colorado, 2002–2006. mmwr Morbidity and Mortality Weekly Report, 57, 1373–1377. Davis, R. M. (1998). Exposure to environmental tobacco smoke: Identifying and protecting those at risk. Journal of the American Medical Association, 280, 1947–1949. Davis, R. M., Wakefield, M., Amos, A., & Gupta, P. C. (2007). The hitchhiker’s guide to tobacco control: A global assessment of harms, remedies and controversies. Annual Review of Public Health, 28, 171–194. Eagan, T. M. L., Hetland, J., & AarØ, L. E. (2006). Decline in respiratory symptoms in service workers five months after a public smoking ban. Tobacco Control, 15, 242–246. Eisner, M. D., Smith, A. K., & Blanc, P. D. (1998). Bartenders’ respiratory health after establishment of smoke-free bars and taverns. Journal of the American Medical Association, 280, 1909–1914. Eriksen, M. P., & Cerak, R. L. (2008). The diffusion and impact of clean indoor air laws. Annual Review of Public Health, 29, 171–185. Howard, J. (2004). Smoking is an occupational hazard. American Journal of Industrial Medicine, 46, 161–169. Huang, P., & McCusker, M. E. (2004). Impact of a smoking ban on restaurant and bar revenues—El Paso, Texas, 2002. mmwr Morbidity and Mortality Weekly Report, 53, 150–152. Hyland, A., Travers, M. J., Dresler, C., Higbee, C., & Cummings, K. M. (2008). A 32-country comparison of tobacco smoke derived particle levels in indoor public places. Tobacco Control, 17, 159–165. Khuder, S. A., Milz, S., Jordan, T., Price, J., Silvestri, K., & Butler, P. (2007). The impact of a smoking ban on hospital admissions for coronary heart disease. Preventive Medicine, 45, 3–8. 6 Klepeis, N., Ott, W., & Switzer, P. (2007). Real-time measurement of outdoor tobacco smoke particles. Journal of the Air & Waste Management Association, 57, 522–534. Lal, A., Siahpush, M., & Scollo, M. (2004). The economic impact of smoke-free legislation on sales turnover in restaurants and pubs in Tasmania. Tobacco Control, 13, 454–455. Lucas, C., & Houston, C. (2007, January 15). Pubs rush on outdoor areas to beat smoking ban. The Age, 3. McCarthy, M., Durkin, S., Brennan, E., & Germain, D. (2008). Smokefree hospitality venues in Victoria: Public approval, patronage and quitting behaviour, 2004–2007 (CBRC Research Paper Series No. 32). Melbourne, Australia: Centre for Behavioural Research in Cancer, Cancer Council Victoria. Mulcahy, M., Evans, D. S., Hammond, S. K., Repace, J. L., & Byrne, M. (2005). Secondhand smoke exposure and risk following the Irish smoking ban: An assessment of salivary cotinine concentrations in hotel workers and air nicotine levels in bars. Tobacco Control, 14, 384–388. National Cancer Institute. (2008). The role of the media in promoting and reducing tobacco—executive summary. In Tobacco control monograph No 19. Retrieved 10 February 2009, from http://www.cancercontrol.cancer.gov/tcrb/monographs/19/ index.html Petticrew, M., Semple, S., Hilton, S., Creely, K. S., Eadie, D., Ritchie, D., et al. (2007). Covert observation in practice: Lessons from the evaluation of the prohibition of smoking in public places in Scotland. BMC Public Health, 7, 204–211. Pion, M., & Givel, M. S. (2004). Airport smoking rooms don’t work. Tobacco Control, 13(Suppl. 1), i37–i40. Repace, J. (2005). Measurement of outdoor air pollution from secondhand smoke on the UMBC campus. Retrieved 10 February 2009, from http://www.repace.com/pdf/outdoorair.pdf Sargent, R. P., Shepard, R. M., & Glantz, S. A. (2004). Reduced incidence of admissions for myocardial infarction associated with public smoking ban: Before and after study. British Medical Journal, 328, 977–983. Scollo, M., Lal, A., Hyland, A., & Glantz, S. (2003). Review of the quality of studies on the economic effects of smoke-free policies on the hospitality industry. Tobacco Control, 12, 13–20. Semple, S., MacCalman, L., Atherton Naji, A., Dempsey, S., Hilton, S., Miller, B. G., et al. (2007). Bar workers’ exposure to second-hand smoke: The effect of Scottish smoke-free legislation on occupational exposure. Annals of Occupational Hygiene, 51, 571–580. Tobacco (Amendment) Act. (2000). 43/2000 Parliament of Victoria. Retrieved 10 February 2009, from http://www.legislation. vic . gov . au / Domino / Web_Notes / LDMS / PubStatbook . nsf / f932b66241ecf1b7ca256e92000e23be/F1D6219209D64A7DCA 256E5B00213E5A/$FILE/00-043a.pdf Tobacco (Amendment) Act. (2005). 45/2005 Parliament of Victoria. Retrieved 10 February 2009, from http://www. Nicotine & Tobacco Research legislation.vic.gov.au/Domino/Web_Notes/LDMS/PubStatbook. nsf/f932b66241ecf1b7ca256e92000e23be/55B5BBE3746510FC CA25705F0020F82D/$FILE/05-045a.pdf for Health Promotion, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. Tobacco (Further Amendment) Act. (2001). 28/2001 Parliament of Victoria. Retrieved 10 February 2009, from http://www .legislation.vic.gov.au/Domino/Web_Notes/LDMS/PubStatbook .nsf/f932b66241ecf1b7ca256e92000e23be/469198B823669079C A256E5B00213EFE/$FILE/01-028a.pdf Wakefield, M., Trotter, L., Cameron, M., Woodward, A., Inglis, G., & Hill, D. (2003). Association between exposure to workplace secondhand smoke and reported respiratory and sensory symptoms: Cross-sectional study. Journal of Occupational and Environmental Medicine, 45, 622–627. Travers, M., Higbee, C., & Hyland, A. (2007). Vancouver Island outdoor tobacco smoke air monitoring study 2007. Retrieved 10 February 2009, from http://www.tobaccofreeair.org/documents /VancouverIslandOSAReport4-10-07.pdf World Health Organization. (2003). WHO framework convention on tobacco control. Retrieved 16 November 2009, from http: //whqlibdoc.who.int/publications/2003/9241591013.pdf U.S. Department of Health and Human Services. (2006). The health consequences of involuntary exposure to tobacco smoke: A report of the Surgeon General—executive summary. Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, Coordinating Center World Health Organization. (2009). WHO framework convention on tobacco control. Guidelines for implementation Article 5.3; Article 8; Article 11; Article 13, Retrieved 16 November 2009, from http://whqlibdoc.who.int/publications/2009 /9789241598224_eng.pdf 7