Original Investigation Indoor Air Pollution Levels Were Halved as a Result of a National Tobacco Ban in a New Zealand Prison

Transcription

1 Nicotine & Tobacco Research, Volume 15, Number 2 (February 201) 4 47 Original Investigation Indoor Air Pollution Levels Were Halved as a Result of a National Tobacco Ban in a New Zealand Prison Simon Thornley, B.H.B., M.B.Ch.B., M.P.H. (Hons), Ph.D.,1 Kim N. Dirks, B.Sc., M.Sc., Ph.D.,1 Richard Edwards, M.B., B.Chir., M.R.C.P., M.F.P.H. (UK), M.P.H., M.D., 2 Alistair Woodward, M.B., B.S., M.Med.Sci., Ph.D.,1 & Roger Marshall, B.Sc., M.Sc., Ph.D.1 1 Section of Epidemiology and Biostatistics, Faculty of Medical and Health Sciences, University of Auckland, Tamaki Campus, Auckland, New Zealand 2 Health Promotion and Policy Research Unit, Department of Public Health, University of Otago, Wellington, New Zealand Corresponding Author: Simon Thornley, B.H.B., M.B.Ch.B., M.P.H. (Hons), Ph.D., University of Auckland, Tamaki Campus, Auckland, New Zealand. Telephone: ext ; Fax: ; s.thornley@auckland.ac.nz Received October 25, 2011 ; accepted April 6, 2012 Abstract Introduction: Few studies have measured the effect of tobacco bans on secondhand smoke (SHS) exposure in prisons. From June 1, 2011, the sale of tobacco was prohibited in New Zealand prisons. One month later, the possession of tobacco was banned. We studied the indoor air quality before and after this policy was enforced. Methods: We measured indoor-fine-particulate (PM 2.5 ) concentrations using a TSI SidePak photometer. The instrument was placed in a staff base of a New Zealand maximum-security prison, adjacent to four 12-cell wings. Measurements were made before the sales restriction, during this period, and after the ban. Data were summarized using daily geometric means and generalized least squares regression. Results: A total of 7,107 observations were recorded at 5-min intervals, on 14 days before and 15 days after implementation, between 24 May and 5 August. Before the policy was implemented, the geometric mean was 6.58 μg/m (95% CI = ), which declined to 5.17 μg/m (95% CI = ) during the sales ban, and fell to 2.44 μg/m (95% CI = ) after the smoking ban. Regression analyses revealed an average 57% (95% CI = 42 68) decline in PM 2.5 concentrations, comparing the before and after periods. Conclusions: Our study showed a rapid and substantial improvement in indoor air quality after tobacco was banned at a prison. We conclude that prisoners have reduced their smoking in line with the ban, and that a significant health hazard has been reduced for staff and prisoners alike. Introduction Cigarette smoking is the single leading cause of preventable poor health and premature death in New Zealand ( Tobias & Turley, 2005 ). Beyond the well-recognized risks to smokers themselves, exposure to secondhand smoke (SHS) increases the risk of developing coronary heart disease and lung cancer by about 0% ( U.S. Department of Health and Human Services, 2006 ). According to the Department of Corrections, two- thirds of New Zealand prisoners are smokers ( Ministry of Health, 2006 ). As a result, Corrections Officers and other prison staff were, in the past, exposed frequently to SHS. To reduce this hazard and arson risks, the Department of Corrections implemented a national smoking ban in New Zealand prisons from Friday July 1, 2011 ( Department of Corrections, 2011 ). A 1- month ban on tobacco purchase (from June 1, 2011) preceded the policy. Both tobacco and cigarette lighters were considered contraband from July 1. Before this policy was enforced, prisoners could smoke freely within their cells and in outdoor areas, but not in indoor common areas. Staff were able to smoke in outdoor areas only. Before the rule was imposed, the Department boosted the provision of smoking cessation services, with effectively free nicotine replacement therapy made available to both staff and prisoners. Prisoners also had (and continue to have) access to telephone quit support. New Zealand was the first country to completely ban smoking in prisons nationwide, although several states in the U nited S tates have a similar policy, and Canada has enforced indoor bans in its institutions. Most prisons in Australia, in contrast, are yet to enact smokefree policies within their buildings. After the introduction of a similar ban in the state of North Carolina, pollutant levels (respirable particulates) fell by 77% ( Proescholdbell, Foley, Johnson, & Malek, 2008 ). We speculated that the ban would lead to a sustained drop in indoor-fine-particulate (PM2.5 ) concentrations. To our knowledge, few studies have assessed the effect of comprehensive prison smoking bans on indoor air quality. We aimed to determine whether a tobacco ban would improve the indoor air quality of a maximum-security prison. Methods The study was carried out at Auckland Prison, a maximumsecurity, long stay, prison in Albany, in rural surrounds, 0 km from the center of the city. doi:10.109/ntr/nts127 Advance Access publication May 15, 2012 The Author Published by Oxford University Press on behalf of the Society for Research on Nicotine and Tobacco. All rights reserved. For permissions, please journals.permissions@oup.com 4

2 Air pollution levels halved in a New Zealand prison We placed a fine particle ( 2.5 μ m in diameter) monitor (TSI SidePak AM510 Personal Aerosol Monitor, TSI Inc, St. Paul) in the West Division, Unit 2 staff base on May 24, 2011 ( 1 week before the sales ban), and ceased measurements on August 10, Although we had planned to place the monitor closer to the source of pollution (prisoner s cells), fear of tampering with the device lead us to place it in the staff base. The particular base chosen had a door that was often open and in direct communication with the area containing the prisoners cells. The monitor was set on a ledge that had free airflow from the common, recreation area, and four wings of individual cells. The wings spread in a V shape, from the common area, with two upstairs and two downstairs, each containing 12 prisoners. At the end of each wing were communal bathrooms. Prison staff reported that all rooms were occupied during the period of the study. Staff estimated that about 80% of prisoners smoked in this area before the ban (compared with the average of about two- thirds of inmates being smokers in prisons nationwide). Smoking prevalence is much lower among the general adult population in New Zealand ( ~20% ; Ministry of Health, 2008 ). The meter measured the concentration of PM 2.5 by estimating the dispersion of a laser beam that passed through continuously pumped air, sampled from the immediate environment. The monitor was zero-calibrated at each visit (weekly to monthly) and was fitted with a 2.5 μ m impactor with an air flow rate of 1.7 L/ min. When visits were made, the monitor was reset to log data and the impactor regreased. The air flow rate was validated in the New Zealand setting using a pneumotachograph (Hans Rudolph 481 pneumotachograph, vacuumed differential pressure transducer 4500, Vacumetrics, CA ), and was within 10% of the stated flow rate. A calibration factor of 0.2 was applied based on calibration work with a ThermoMIE personaldataram model pdr-1200 real-time aerosol monitor (ThermoAndersen, Inc., Smyrna, GA). Additional calibration work subsequently has calculated a very similar calibration factor for secondhand smoke ( ; Jiang et al., 2011 ). This type of monitor had been used extensively in previous similar studies assessing particulate levels from SHS, including in a study from Delaware where the SidePak AM510 was calibrated against standard pump-andfilter gravimetric methods ( Repace, 2004 ). The monitor was fixed in place for months in total. Air pollution levels were recorded before any restrictions were introduced ( baseline ), during the first month in which purchase was restricted ( sales ban ), then for 2 months after the policy was enacted ( smoking ban ). During our visits to prison, we informally asked four staff about their perceptions of compliance with the new smokefree policy; however, this was a minor adjunct to the study, and no prisoners were interviewed. Statistical Analysis Data were recorded at 5-min intervals and logged within the device. The monitors were checked and data downloaded at weekly to monthly intervals throughout the observation period. The data were then converted to daily averages, and comparisons were made between the baseline, sales ban, and smoking ban phases of the monitoring, using regression methods. Serial correlation between measurements was investigated and models that incorporate an autocorrelated error term were used to model the effect of time on the PM 2.5 concentration. We used the gls (generalized least squares) function, from the nlme library that accompanies R software, to carry out these analyses ( R Development Core Team, 2007 ). Time was considered, in separate analyses, as both a continuous and a categorical variable. The categories included baseline (before the sales ban), the sales ban period (June 2011), and the tobacco ban period (July 1 onward). Plots of residuals by fitted values were used to investigate the model fit. Approval for the study was sought through the Department of Corrections external research committee, which facilitated the research process. Results A broken time series of PM 2.5 measurements was obtained between May 17 and August 5, Daily summaries of the raw data are recorded in the Table 1. Between 2 and June 27 and June 0 to July 4, no data were recorded. The periods of missing data were due to the monitor being inadvertently unplugged by prison staff. We observed a gradual decline in maxima and geometric means with a modest decline in minima. Before the ban, the geometric mean 5-min PM 2.5 ranged from to 1 μ g/m, Table 1. Daily Summaries of Indoor-Fine- Particulate Concentration ( μ g/m ) During the S tudy P eriod Period Date Geometric mean Maximum Minimum Baseline 24/ / / / / / / / ,700 1 Sales ban 1/ / / / / / Smoking ban 4/ / / / / / / / / / / / / / /

3 Nicotine & Tobacco Research, Volume 15, Number 2 (February 201) while after the ban, a single peak occurred on July 18 (10.6 μg/m ), with subsequent means ranging from 2 to 5 μg/m. There were a total of 7,107 five-minute PM recordings. The PM 2.5 values were approximately log-normally distributed and were, therefore, log-transformed to stabilize their variance in regression analyses, and associated geometric means were reported. Of the 7,107 observations, 2,17 were recorded before the sales ban, 1,409 during the sales ban, and,561 after the tobacco ban was enforced. Before the sales ban, the geometric mean daily average was 6.58 μg/m (95% CI = ), which declined to 5.17 μ g/m (95% CI = ) during the sales ban, and fell to 2.44 μ g/m (95% CI = ) after the smoking ban. The distribution of 5-minute PM 2.5 during these periods is depicted in the Figure 1. The plot shows a reduction in the median values, with fewer high, extreme measurements after the sales and smoking ban was enforced. Regression analysis was undertaken on log-transformed values to determine a time trend. We added 0.25 μg/m to zero values to allow the log transformation. The autocorrelation between measurements was about 0.9 with serial decay of the coefficient with increasing lag. The partial autocorrelation coefficient indicated little ( ~ 0.07) residual correlation after the correlation between immediately adjacent terms had been adjusted for. A locally weighted regression (lowess) plot of the logtransformed PM 2.5 concentration by time, revealed a lineardownward trend in fine particle levels throughout the period of the study. After regression analysis, using a first order autocorrelation term, we derived an exponentiated beta-coefficient of (95% CI = ), which indicated the average daily percentage decline in particulate during the study was 1.%. This was equivalent to a halving of concentration during any 52- day period (95% CI = days), throughout the course of the study. During the 71 days of monitoring, therefore, the magnitude of the decline was 61% (95% CI = 48% 71%). When Figure 1. Boxplot of 5-min PM 2.5 concentration, by prison tobacco policy period ( 0.25 added to zero-value measurements so that they may be depicted on a log scale). we divided time into three categorical periods baseline, sales ban, and smoking ban a model that incorporated a first order autoregression term indicated the sales ban resulted in a 4% (exponentiated beta-coefficient 0.66; 95% CI = ) and the tobacco ban was associated with a 57% drop (exponentiated beta-coefficient 0.4; 95% CI = ) in fine particle levels, compared with the preban period. Plots of residuals versus fitted values indicated acceptable model fit. We repeated the regression analysis by adding 0.5 μg/m (rather than 0.25) to 0 PM 2.5 values (to allow the log transformation), but effect estimates were effectively unchanged (<1% change in point estimates). We then limited the analysis to data collected during waking hours (07:00 22:59), when most smoking was likely to have occurred. The magnitude of the drop associated with the ban was higher for measures restricted to these intervals (baseline: geometric mean 9.64 μg/m, 95% CI = ; during sales ban: 6.57, 95% CI = ; during smoking ban: 2.6, 95% CI = ). Staff reported that prisoners had, by and large, complied with the prison policy, and no tobacco had been seized after the ban date. Occasional cigarette lighters had been found and confiscated. Although an increase in violence had been feared, no increase in incidents was reported by the unit manager, who accompanied us, while obtaining measurements from the monitor. Discussion We observed a reduction in indoor air concentrations of fine particles after a comprehensive tobacco ban in a maximumsecurity prison in Auckland, New Zealand. Overall, levels of particulates were very low after the ban was enacted and this, together with reports from staff, suggests that short-term breaches of the ban were low. The study was limited by its scope, missing data, and inability to monitor close to the source of pollution. Due to limited resources, only one particulate monitor was available, so that our results only describe the effect on indoor air quality in one section of the prison. The site was chosen due to the continuity of airflow with the prisoner s cells and also to provide an assessment of real, workplace exposures for Corrections officers. Also, the prison site, located in a semi-rural setting, was not close to any other sources of pollution, such as traffic or industrial sites. Missing information resulted in an interrupted time series, so that we cannot describe the full trajectory of particulate measurements during the sales ban period. However, we can infer that concentrations were lower after the ban was implemented than before. Ideally, we would have liked to monitor in the common area used by prisoners outside their cells. However, due to the nature of the equipment and the likelihood of tampering, a compromise had to be made. In addition, the monitor was close to a microwave and toaster. Both of these have the potential to release particulate matter into air (such as burnt toast). We speculated that some of the outlying measurements were due to this source. The extreme values were all recorded before the 45

4 Air pollution levels halved in a New Zealand prison tobacco ban, so may have been a result of smoking in close proximity to the monitor. If extreme values, greater than 1.5 times the interquartile range of the log-transformed measurements, (>100 μ g/m ) were excluded in the regression analysis, the before after effect of the total ban was reduced by three percentage-points (55%). Many other studies have documented reductions in indoor levels of particulates as a result of smoking bans in public places, such as pubs and bars ( Brennan et al., 2010 ; Connolly et al., 2009 ; Valente et al., 2007 ). Studies in prisons are rarer. Three other reports, to our knowledge, have similarly analyzed the impact of a smoking ban in prisons: two in the United States and another in Switzerland. One study was undertaken in the state of North Carolina, United States, before and after a complete smoking ban ( Proescholdbell et al., 2008 ). While this study monitored a variety of sites in six prisons before and after the smoking ban, the monitoring only occurred for short periods (mean: 96 min). Also, the time immediately after the ban was not monitored: only 2 month s intervals before and 12 months after implementing the ban. Despite the methodological differences, the U.S. study showed a consistent decline of fine particulate concentration, with an average decline of % in the 24 areas of the six prisons monitored. Our analysis was consistent with this result, showing a 57% overall decline when the ban periods were divided into categories in the regression analysis. A second study of indoor air quality in two prisons in Vermont and one in Massachusetts, before and after a tobacco ban, showed a decline in the indoor concentration of aerosolized nicotine, from 11 to μ g/m in weekly means ( Hammond & Emmons, 2005 ). Measurements were taken from a range of living quarters and common areas in the prisons. This study detected a larger percentage decline in nicotine levels than we observed with PM 2.5. Nicotine is more specific than particulate to emissions from cigarettes, and so such a finding is expected. In Switzerland, a partial ban of smoking (limiting use of cigarettes to cells and outside areas only) resulted in a drop in both nicotine (70%) and particulate matter (PM 10 ; 45%), when the air quality was assessed 2 months before and 11 months after the policy enactment in one prison ( Ritter, Huynh, Etter, & Elger, 2011 ). The authors described the difference in nicotine as statistically significant, but not the change in particulate levels. Their findings, of a 45% decrease, is similar to ours of a 57% overall reduction. Inmate views were sought in response to an indoor ban, effected in prisons in Quebec, Canada, in Interviews with smoking prisoners revealed that only 40% of smokers thought that the policy had reduced the amount they smoked. About half of inmates interviewed thought their exposure to SHS had reduced ( Lasnier et al., 2011 ). The absolute levels of fine particles that we measured throughout the study were low. The recorded low levels, however, were likely to be due to the proximity of the monitor (placed in the staff base) rather than at the source of pollution (prisoner s cells). For comparison, the Ministry for the Environment in New Zealand suggests that a daily average of 25 μg/m or greater for ambient air is cause for concern ( Ministry for the Environment, 2002 ). The World Health Organization, similarly, recommends that levels of PM 2.5 should be below an annual mean of 10 μg/m and 25 μg/m as a daily mean ( World Health Organization, 2005 ). The rationale underlying the WHO guideline is based on an American Cancer Society Study of the effects of long-term exposure to particulates, recorded outdoors, on mortality. A threshold was identified at the lower range of particles at which adverse effects on survival were detected ( Krewski et al., 2005 ). The daily means were well under that threshold, even before the ban was enforced. While the evidence for adverse health effects is strongest for chronic exposure, the daily standard is also recommended to prevent temporary peaks that may lead to excess mortality and morbidity. Relative to these standard levels, enforcement of the smoking ban has brought about an improvement in the concentration of fine particles, such that daily indoor geometric means were well below the World Health Organization s annual threshold of 10 μg/m on all but 1 of 15 days after the smoking ban, whereas before the ban, the W orld H ealth O rganization annual average air quality standard of 10 μg/m was exceeded on 2 out of 14 days. Although we monitored for a short time, the annual average was unlikely to exceed this threshold, and so the levels of fine particles we observed were low overall. Despite the low absolute levels of pollution, which were likely to be much higher within cells when cigarettes were being smoked, we believe the more important finding is the relative change in pollutant level. Such relative changes indicate that exposure to particulates, within cells, is also likely to be dropping by similar ratios for staff and prisoners alike. Very high isolated levels, which occurred before the policy was enacted, were less commonly observed after the ban. The source of the PM 2.5 that caused these peaks is unknown. Comparisons of the levels found in our study, however, with W orld H ealth O rganization guideline levels should be carried out with caution. First, guidelines apply to ambient air pollution. The World Health Organization observes that no threshold for PM 2.5 has been identified below which no damage to health is observed so that any involuntary exposure of nonsmokers to SHS should be avoided, if possible ( World Health Organization, 2005 ). The high levels of carcinogens in SHS mean that the risk of cancer is likely to be higher when people are exposed to this source, compared with ambient air pollution, when levels of fine particles are otherwise comparable ( International Agency for Research on Cancer, 2004 ). To balance this concern, however, the exposure of officers and inmates to particulates may be lower than the measured levels we report, due to time spent in other areas of the prison compound than the cell areas (such as working or training), and for prison officers, time away from the prison compound. About 7,000 prisoners inhabit 19 prisons in New Zealand and the Department of Corrections employs about 6,000 staff to service them. We observed that indoor air quality improved as a result of a comprehensive smoking ban, which suggested that the ban has successfully restricted the smoking of cigarettes and so resulted in a safer environment for prison employees, prisoners, and visitors. The ban was accompanied by the provision of nicotine replacement therapy to both staff and prisoners to assist quit attempts and defuse resistance to the change in policy. The improvements in air quality support the prison staff s observations that prisoner compliance with the smoking ban was high. Future studies could be undertaken to investigate the 46

5 Nicotine & Tobacco Research, Volume 15, Number 2 (February 201) extent of compliance with the policy. This is the first time in the world that tobacco has been banned completely from a nation s prisons, and the improvement in indoor air quality, as a result, sets a precedent for other nations to introduce a similar comprehensive policy. Funding This study was funded by the University of Auckland. Declaration of Interests None. Acknowledgments We gratefully acknowledge the Department of Corrections for supporting this project, and Mr. Pete Phelan for his help in gaining access to the study site and information about the implementation of the policy. References Brennan, E., Cameron, M., Warne, C., Durkin, S., Borland, R., Travers, M. J., et al. (2010 ). Secondhand smoke drift: Examining the influence of indoor smoking bans on indoor and outdoor air quality at pubs and bars. Nicotine & Tobacco Research, 12, doi:10.109/ntr/ntp204 Connolly, G. N., Carpenter, C. M., Travers, M. J., Cummings, K. M., Hyland, A., Mulcahy, M., et al. (2009 ). How smoke-free laws improve air quality: A global study of Irish pubs. Nicotine & Tobacco Research, 11, doi:10.109/ntr/ntp08 Department of Corrections. (2011 ). Corrections news: May to June Retrieved from http :// /newsand - publications / magazines - and - newsletters / corrections - news /2012 /corrections_news_may-june_2011 /last_puff.html Hammond, S. K., & Emmons, K. M. (2005 ). Inmate exposure to secondhand smoke in correctional facilities and the impact of smoking restrictions. Journal of Exposure Analysis and Environmental Epidemiology, 15, doi:10.108/sj.jea [pii] International Agency for Research on Cancer. (2004 ). Tobacco smoke and involuntary smoking (vol. 8 ). Lyon, France : Author. Jiang, R. T., Acevedo-Bolton, V., Cheng, K. C., Klepeis, N. E., Ott, W. R., & Hildemann, L. M. ( 2011 ). Determination of response of real-time SidePak AM510 monitor to secondhand smoke, other common indoor aerosols, and outdoor aerosol. Journal of Environmental Monitoring, 1, doi:10.109/ C0em0072c Krewski, D., Burnett, R., Jerrett, M., Pope, C. A., Rainham, D., Calle, E., et al. ( 2005 ). Mortality and long-term exposure to ambient air pollution: Ongoing analyses based on the American Cancer Society cohort. Journal of Toxicology and Environmental Health, Part A, 68, doi:u1guu27wn00558x [pii] 1080/ Lasnier, B., Cantinotti, M., Guyon, L., Royer, A., Brochu, S., & Chayer, L. (2011 ). Implementing an Indoor Smoking Ban in Prison: Enforcement Issues and Effects on Tobacco Use, Exposure to Second-hand Smoke and Health of Inmates. Canadian Journal of Public Health-Revue Canadienne De Sante Publique, 102, Ministry for the Environment. ( 2002 ). Ambient air quality guidelines. Wellington, New Zealand : Author. Ministry of Health. (2006 ). Results from the prisoner health survey. Wellington, New Zealand : Author. Retrieved from http :// www. moh. govt. nz / moh. nsf / pagesmh / 5650 /$ File / prisoner - health-survey-2005.pdf Ministry of Health. (2008 ). A Portrait of Health: Key Results of the 2006/07 New Zealand Health Survey. Wellington, New Zealand : Author. Proescholdbell, S. K., Foley, K. L., Johnson, J., & Malek, S. H. ( 2008 ). Indoor air quality in prisons before and after implementation of a smoking ban law. Tobacco Control, 17, doi:10.116/tc R Development Core Team. (2007 ). R: A language and environment for statistical computing. Vienna, Austria : Author. Repace, J. ( 2004 ). Respirable particles and carcinogens in the air of delaware hospitality venues before and after a smoking ban. Journal of Occupational and Environmental Medicine, 46, doi: /01.jom Ritter, C., Huynh, C. K., Etter, J.-F., & Elger, B. S. (2011 ). Exposure to tobacco smoke before and after a partial smoking ban in prison: Indoors air quality measures. Tobacco Control. doi:10.116/ tc Tobias, M., & Turley, M. (2005 ). Causes of death classified by risk and condition, New Zealand Australian and New Zealand Journal of Public Health, 29, doi: /j x tb00740.x U.S. Department of Health and Human Services. (2006 ). The health consequences of involuntary exposure to tobacco smoke: A report of the Surgeon General. Atlanta, GA : US Department of Health and Human Services, Centers for Disease Control and Prevention. National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. Valente, P., Forastiere, F., Bacosi, A., Cattani, G., Di Carlo, S., Ferri, M., et al. (2007 ). Exposure to fine and ultrafine particles from secondhand smoke in public places before and after the smoking ban, Italy Tobacco Control, 16, doi:10.116/tc World Health Organization. (2005 ). WHO air quality guidelines global update Copenhagen, Denmark : Author. 47