Addictive Behaviors 40 (2015) 51–56
Contents lists available at ScienceDirect
Addictive Behaviors
Smoking policy change at a homeless shelter: Attitudes and effects Michael S. Businelle a,b,⁎, Insiya B. Poonawalla a, Darla E. Kendzor a,b, Debra M. Rios a, Erica L. Cuate a, Elaine J. Savoy c, Ping Ma a,b, Travis P. Baggett d, Jennifer Reingle a, Lorraine R. Reitzel e a
University of Texas School of Public Health, Department of Health Promotion and Behavioral Sciences, Dallas, TX, USA University of Texas Southwestern Medical Center, Dallas, TX, USA University of Houston, Department of Clinical Psychology, Houston, TX, USA d Massachusetts General Hospital, Division of General Internal Medicine, Boston, MA, USA e University of Houston, Department of Educational Psychology, Houston, TX, USA b c
H I G H L I G H T S • • • •
Sheltered homeless adults support the creation of large outdoor smoke free zones. Support for shelter-wide smoking bans is more limited. Smoking bans at shelter campuses may reduce carbon monoxide levels in smokers. Findings indicate potential consequences of partial and shelter-wide smoking bans.
a r t i c l e
i n f o
Available online 6 September 2014 Keywords: Homelessness Tobacco Smoking ban Policy
a b s t r a c t Homeless adults are exposed to more smokers and smoke in response to environmental tobacco cues more than other socioeconomically disadvantaged groups. Addressing the culture of smoking in homeless shelters through policy initiatives may support cessation and improve health in this vulnerable and understudied population. This study examined support for and expected/actual effects of a smoking ban at a homeless shelter. A 2-wave crosssectional study with an embedded cohort was conducted in the summer of 2013 two weeks before (wave 1) and two months after (wave 2) a partial outdoor smoking ban was implemented. A total of 394 homeless adults were surveyed (i.e., wave 1 [n = 155]; wave 2 [n = 150]; and 89 additional participants completed both waves). On average, participants were 43 years old, primarily African American (63%), male (72%), and had been homeless for the previous 12 months (median). Most participants were smokers (76%) smoking 12 cigarettes per day on average. Most participants supported the creation of a large smoke-free zone on the shelter campus, but there was less support for a shelter-wide smoking ban. Average cigarettes smoked per day did not differ between study waves. However, participants who completed both study waves experienced a reduction in expired carbon monoxide at wave 2 (W1 = 18.2 vs. W2 = 15.8 parts per million, p = .02). Expected effects of the partial ban were similar to actual effects. Partial outdoor smoking bans may be well supported by homeless shelter residents and may have a positive impact on shelter resident health. © 2014 Elsevier Ltd. All rights reserved.
1. Introduction Smoking prevalence has declined to 18.1% among US adults (Centers for Disease Control & Prevention, 2014), yet over 70% of homeless adults are current smokers (e.g., Baggett & Rigotti, 2010; Connor, Cook, Herbert, Neal, & Williams, 2002; Lee et al., 2005). Although homeless smokers and other socioeconomically disadvantaged domiciled smokers may have comparable numbers of quit attempts (Businelle, ⁎ Corresponding author at: University of Texas School of Public Health, 6011 Harry Hines Blvd., V8.112, Dallas, TX 75390-9128, USA. Tel.: +1 214 648 1070; fax: +1 214 648 1081. E-mail address: [email protected] (M.S. Businelle).
http://dx.doi.org/10.1016/j.addbeh.2014.08.013 0306-4603/© 2014 Elsevier Ltd. All rights reserved.
Cuate, Kesh, Poonawalla, & Kendzor, 2013), homeless smokers may have more difficulty maintaining smoking abstinence partially due to a culture of tobacco use (Baggett, Tobey, & Rigotti, 2013) and permissive shelter policies that result in frequent exposure to smoking. Recent findings indicate that homeless smokers are exposed to substantially more smokers and are more likely to report smoking in response to social goads and other cues than socioeconomically disadvantaged domiciled smokers (Businelle et al., 2013). Thus, outdoor smoking bans at shelters may reduce social cues to smoke and increase the likelihood of successful smoking cessation among homeless individuals (Baggett, Tobey, et al., 2013). Despite the alarmingly high negative health consequences of smoking in homeless individuals (Baggett, Hwang, et al., 2013;
52
M.S. Businelle et al. / Addictive Behaviors 40 (2015) 51–56
Baggett, Tobey, et al., 2013; Hwang, Wilkins, Tjepkema, O'Campo, & Dunn, 2009), few shelters have banned all smoking on their campuses (Apollonio & Malone, 2005; National Coalition for the Homeless, 2007). Furthermore, a survey of transitional shelters in the Los Angeles area indicated that only half of those with ≥200 beds had designated smoke free areas (Arangua, McCarthy, Moskowitz, Gelberg, & Kuo, 2007). Although many shelter administrators may recognize the harmful effects of smoking, their concerns about the potential negative consequences that may accompany shelter-wide smoking bans (e.g., violations of the smoking ban, greater presence of homeless individuals outside the shelter campus) may discourage this type of health promoting policy change. Notwithstanding these concerns, homeless individuals may actually support partial or full smoking bans on shelter grounds. For example, a survey of homeless adults at 26 shelters in Los Angeles found that 46% of respondents supported a “smoking ban in all outdoor common areas” (McCarthy & Dyrness, 2012). In addition, support for smoking bans has been shown to increase following implementation in workplace and medical settings (e.g., Borland, Owen, Hill, & Chapman, 1990; Unrod, Oliver, Heckman, Simmons, & Brandon, 2012), and this may also be the case for homeless shelter residents. However, no research has prospectively examined the impact of shelter smoking bans on the attitudes and behaviors of homeless smokers. The purpose of the current study was to examine shelter residents' attitudes about partial and full smoking bans at a homeless shelter in Dallas, Texas. We also examined the effects of a partial smoking ban on smoking behavior and expired carbon monoxide (CO), assessed two months after a partial ban was implemented. We tested the following hypotheses: 1) non-smokers would be more supportive of partial and shelter-wide smoking bans than smokers, 2) the number of supporters of the partial and full smoking bans would increase after the partial ban was implemented, and 3) the number of cigarettes smoked per day and CO levels among current smokers would be reduced after implementation of the partial smoking ban. 2. Methods 2.1. Participants Data collection took place at a large homeless shelter in Dallas, Texas. Homeless individuals were eligible to participate in this study if they spent the previous night at the shelter, earned a score of ≥ 4 on the Rapid Estimate of Adult Literacy in Medicine—Short Form (REALM-SF; Arozullah et al., 2007) indicating N6th grade English literacy level (i.e., required to complete tablet/laptop assessment items), and were at least 18 years old.
approved by the Institutional Review Board at the University of Texas School of Public Health. Flyers describing the study and indicating the dates and times of data collection were posted on the shelter campus. Interested shelter residents approached study staff on the scheduled data collection days. Informed consent was obtained from interested individuals and screening was completed on-site. Those who met the study inclusion criteria used tablet or laptop computers to complete study questionnaires. Questionnaire Development System software (version 2.6.1) was used whereby all questions were read aloud, through earphones. Participants responded to assessment items (using a mouse or touch screen) after the entire question was read. Participants who completed the assessments received a $20 gift card.
2.3. Measures All participants completed measures that assessed demographic characteristics including age, gender, race/ethnicity, and current period of homelessness. Participants provided a CO sample by breathing into a Vitalograph CO ecolyzer and were asked about their smoking status (never smoker, ex-smoker [smoked at least 100 cigarettes in lifetime, but no longer a smoker], and current smoker). Current smokers were also asked about current cigarettes smoked per day (CPD), number of lifetime quit attempts lasting ≥ 24 h, and readiness to quit smoking (Abrams et al., 2003). Finally, participants' attitudes about partial (“I support the creation of a smoke free zone in half of the shelter courtyard”) and shelter-wide (“I support a complete smoking ban at the shelter”) bans were assessed on 5 point scales ranging from strongly disagree to strongly agree (this scale was dichotomized so that those who strongly agreed or agreed were coded as “1” and those who were neutral, disagreed, or strongly disagreed were coded as “0”). To assess expected effects of the bans, all participants were asked “How would you react if smoking was no longer allowed in half the shelter courtyard?”, and “How would you react if the shelter no longer allowed smoking anywhere on the shelter campus?” To assess the actual effects of the partial ban, W2 participants were asked “What effect has the shelter ban on smoking in half the courtyard had on you?” (see Table 3 for response options). Finally, participants rated their agreement with the following statements “Banning all smoking at the shelter would improve the health of shelter guests” and “Banning all smoking at the shelter would reduce smoking in guests.” Response options ranged from strongly disagree to strongly agree (response options for these items were also dichotomized).
2.4. Statistical analyses 2.2. Procedure The partial smoking ban was prompted by the results of another study at this shelter, which indicated that smokers who were trying to quit smoking were exposed to approximately 40 smokers each day, partially due to a lack of an outdoor shelter smoke free zone (Businelle et al., 2013, in press). Two separate waves of data collection occurred during the summer of 2013. Wave 1 (W1) data were collected over a one week period, two weeks prior to the implementation of the partial outdoor smoking ban that covered one-half (approximately 6750 ft2) of the shelter courtyard. At the time of W1 data collection, residents were unaware of plans for the ban implementation. Collection of wave 2 (W2) data was completed over a one week period that began two months after the partial ban was implemented. Participants who completed the W1 assessments were encouraged to complete the W2 assessments, but they were not required to do so. Therefore, some participants completed the W1 assessment or W2 assessment, and a subset of participants completed both W1 and W2 assessments. This study was
Participant characteristics, attitudes about smoking bans (i.e., partial and full bans), and the effects of a partial smoking ban were descriptively analyzed for the total sample and separately for each data collection wave. Independent sample t-tests and Chi-square tests were conducted to identify differences among participants who completed only one wave of data collection (i.e., either W1 or W2). Paired t-tests and McNemar's tests were conducted to test for within-subjects changes in study variables for those who completed both study waves. Changes in mean CPD and expired CO levels among smokers were analyzed using linear regression (i.e., for comparisons of participants who completed only W1 or W2) and linear multilevel mixed (LMM) modeling analyses (i.e., for comparisons of participants who completed both study waves). Adjusted models controlled for age, gender, race (White vs. non-White), and education. In addition, data were analyzed to determine if post-ban changes in CPD and CO differed by level of smoking (light smoker [≤10 CPD] vs. moderate to heavy smoker [N10 CPD]).
M.S. Businelle et al. / Addictive Behaviors 40 (2015) 51–56
53
p b .001). Analyses were also conducted with the cohort who completed both data collection waves, with similar results.
3. Results Overall, 394 unique shelter residents completed the study (i.e., W1 [n = 155]; W2 [n = 150]; and 89 additional participants completed both waves). Approximately 75% of all current shelter residents are enrolled in the study at each wave. Participants were 43.3 years old (SD = 11.8), predominantly male (71.8%), non-White (69.8%), and reported that their current period of homeless began 12.0 months ago (median). Most participants (75.9%) were current smokers who smoked an average of 12.0 CPD (SD = 7.2). Most current smokers (63%) reported that they planned to quit smoking within the next six months. Participants who completed only W1 were approximately five years younger than those who completed only W2 (p b .001) and four years younger than those who completed both data collection waves (p = .012). Cohorts were not significantly different on other demographic, homelessness, or smoking characteristics. Participant characteristics are presented in Table 1.
3.1. Support for partial and shelter-wide smoking bans At both study waves, the majority of participants supported (i.e., agreed or strongly agreed) the creation of a smoke free zone in half of the shelter courtyard. Additionally, there was a significant increase in support for the partial ban at W2 among residents who completed both study visits (W1 = 60.7% vs. W2 = 73.9%, p = .02) and a trend in the same direction for those who only completed one wave of data collection (W1 = 60.0% vs. W2 = 70.0%, p = .07; Table 2). Fewer participants supported a shelter-wide ban and support for the full ban did not differ by study wave or change over time within the embedded cohort (Table 2). A minority of smokers agreed that a shelterwide smoking ban would actually reduce smoking in shelter guests, but most smokers agreed that such a ban would improve the health of shelter residents (Table 2). Additional analyses compared non-smokers' and smokers' support for the partial and shelter-wide smoking bans among those who completed only one wave of data collection. Results indicated that nonsmokers and smokers were not significantly different in their support for the partial ban at W1 (71.1% vs. 56.4%, p = .111). However, nonsmokers were more supportive of the partial ban than smokers at W2 (84.2% vs. 65.2%, p = .027) and more supportive of the shelter-wide ban at W1 (63.2% vs. 17.9%, p b .001) and W2 (60.5% vs. 27.7%,
3.2. Comparison of the expected and actual effects of the partial smoking ban Many of the expected effects of a hypothetical partial smoking ban (assessed at W1) were similar to the actual effects of the partial ban (assessed at W2) among the cohort of smokers who completed both waves of data collection (see Table 3). Specifically, before and after the partial ban, similar numbers of participants reported that a partial ban would/did result in: 1) attempts to move to another shelter, 2) reductions in smoking, and 3) more time spent outside the shelter campus. However, fewer participants tried to quit smoking (see Table 3; p b .001) and fewer participants increased their smoking outside the shelter grounds than expected (see Table 3; p = .038) following the partial ban. Expected effects of the partial ban among those who completed W1 only and actual effects of the partial ban among those who completed W2 only are displayed in Table 3. Comparisons of those who completed only one wave of data collection should be interpreted with caution because other factors may have contributed to the observed differences across waves. 3.3. Expected effects of a shelter-wide smoking ban Smokers were asked to indicate their anticipated response(s) to a hypothetical shelter-wide smoking ban at each wave of assessment. No actual shelter-wide ban was implemented during the study period and there was very little difference across study waves in the expected effects of the hypothetical shelter-wide ban; therefore, all 299 smokers who participated in this study were included in the numbers below. Smokers reported that, in response to a shelter-wide smoking ban, 13.7% would try to move to a different shelter, 33.4% would spend more time away from the shelter, 42.8% would smoke more outside the shelter gates, 34.8% would try to quit smoking, and 25.8% would try to decrease the amount that they smoke. 3.4. Changes in CPD and CO across study waves After controlling for age, gender, race, and education, a betweensubjects comparison of smokers who completed only W1 or W2 (W1 = 12.3 vs. W2 = 12.0 CPD) and within-subjects analyses of
Table 1 Participant characteristics. Characteristic
Total unique N = 394
W1 n = 155
W2 n = 150
p value (W1 vs. W2)
W1 and W2 cohort n = 89
p value (W1 vs. cohort)
Age in years [M (SD)] Race (%) White African American All others Male (%) Unemployed (%) Current months of homelessness [M (SD)] Median (range: 0–366) Smoking status (%) Current smokers Ex-smokers Never smokers Cigarettes per day (CPD) [M (SD)]a Heaviness of smoking (%)a Light (≤10 CPD) Moderate/heavy (N10 CPD) Carbon monoxide level, ppm [M (SD)] Previous quit attempts lasting ≥24 h [M (SD)]a
43.3 (11.8)
40.6 (12.6)
45.4 (10.5)
b.001 .205
44.7 (11.7)
.012 .726
30.0 62.8 7.2 71.8 90.1 24.8 (39.5) 12
27.9 62.3 9.8 69.0 87.1 26.6 (42.9) 12.5
34.7 60.0 5.3 74.0 93.3 19.0 (23.6) 9.8
75.9 11.4 12.7 12.0 (7.2)
75.5 11.0 13.5 12.3 (8.1)
74.7 12.7 12.7 12.0 (6.8)
53.2 46.8 13.7 (12.5) 5.1 (3.1)
56.4 43.6 14.2 (13.5) 5.3 (3.0)
49.1 50.9 12.4 (10.5) 4.8 (3.1)
.337 .067 .081
25.8 68.5 5.7 73.0 89.9 31.0 (51.2) 14
.888
.798 .289
.195 .199
W1 = completed wave 1 only; W2 = completed wave 2 only; W1 and W2 cohort = completed both wave 1 and wave 2. a Only current smokers were asked these questions.
.509 .517 .494 .839
78.7 10.1 11.2 11.6 (6.1) 54.3 45.7 15.0 (13.6) 5.1 (3.2)
.581 .968
.640 .618
54
M.S. Businelle et al. / Addictive Behaviors 40 (2015) 51–56
Table 2 Attitudes and opinions about partial and shelter-wide smoking bans. Attitude
I support the creation of a smoke free zone in half the shelter courtyard Strongly agree or agree (%) Strongly disagree or disagree (%) I support a complete smoking ban at the shelter Strongly agree or agree (%) Strongly disagree or disagree (%) Smoker's opinions about shelter-wide smoking ban Banning all smoking at the shelter would reduce smoking in guests Strongly agree or agree (%) Strongly disagree or disagree (%) Banning all smoking at the shelter would improve guest health Strongly agree or agree (%) Strongly disagree or disagree (%)
W1 n = 155
W2 n = 150
W1 vs. W2
W1 and W2 cohort n = 89
Pre-ban
Post-ban
p
Pre-ban
Post-ban
p
60.0 16.1
70.0 11.3
.067
60.7 11.2
73.9 12.5
.019
29.0 47.1
36.0 40.7
.194
30.3 46.1
33.0 52.3
n = 117
n = 112
p
n = 70
27.4 45.3
42.9 38.4
.043
38.6 48.6
33.3 49.3
.523
53.8 18.8
55.4 25.0
.820
58.6 18.6
49.3 29.0
.230
.832 p
W1 = completed wave 1 only; W2 = completed wave 2 only; W1 and W2 cohort = completed both wave 1 and wave 2.
smokers who completed both study waves (W1 = 11.6 vs. W2 = 12.0 CPD) indicated no differences in CPD before and after the partial ban (all ps N .68). However, within-subjects analyses indicated that CO was significantly lower after implementation of the partial smoking ban among participants who completed both study waves (W1 = 15.0 vs. W2 = 12.7 parts per million [ppm], p = .02). Follow-up analyses indicated that significant reductions in CO were found among smokers (W1 = 18.8 vs. W2 = 15.8 ppm, p = .02), but CO was not significantly different among non-smokers who completed both study waves (W1 = 1.1 vs. W2 = 0.9 ppm). Results of a between-subjects CO analysis that compared those who completed W1 to those who completed W2 were in a similar direction, though non-significant (W1 = 14.2 vs. W2 = 12.4 ppm, p = .20). We conducted additional analyses to determine if the partial ban had a differential impact on CPD and CO among smokers by heaviness of smoking (i.e., ≤10 vs. N 10 CPD). We found no change in CPD following the partial smoking ban for light or moderate/heavy smokers in within- and between-subjects analyses. However, CO levels were lower at W2 among moderate to heavy smokers (N10 CPD), for both between-subjects (W1 = 23.9 vs. W2 = 18.6 ppm, p b .01) and within-subjects (W1 = 25.5 vs. W2 = 20.7 ppm, p b .05) analyses. No significant change in CO was detected among light smokers. 4. Discussion The current study is the first to examine pre-ban attitudes and postban effects of a partial smoking ban in a homeless shelter setting. Results yielded three key findings with potential implications for shelter policies concerning smoking. First, most shelter residents (smokers and non-smokers) supported a partial outdoor smoking ban on the shelter grounds, both before and after the partial ban was implemented.
However, residents were far less likely to support a shelter-wide smoking ban, even though most smokers in this study believed that a complete ban would improve the health of shelter residents. Compared to smokers, non-smokers were 15% to 45% more likely to support the smoking bans. Second, the partial ban was related to reductions in carbon monoxide levels, potentially due to changes in active smoking or reduced environmental tobacco smoke exposure among some smokers. Finally, expected and actual effects of the partial smoking ban indicated that this type of shelter policy change may be well tolerated with few negative consequences. Thus, smoking bans at homeless shelters may be an important component of comprehensive shelter policies that aim to improve the health of shelter residents. An overwhelming majority of shelter residents supported the creation of a smoke free zone encompassing one half of the outdoor shelter courtyard. Consistent with previous studies, the partial smoking ban became more popular with smokers following implementation (Borland et al., 1990; Unrod et al., 2012). This level of support for the partial smoking ban is especially important considering that 76% of shelter residents were smokers. It is notable that shelter security has reported few violations (i.e., smoking in the smoke-free zone) since the ban was implemented. Additionally, shelter security recently reported that the smoke free courtyard area is now used more than before the ban was implemented. Prior to learning about the implementation of the partial smoking ban, many W1 participants reported that they would try to quit (24% to 31%) and/or reduce the amount that they smoke (27%–28%) if a partial outdoor smoking ban was implemented. Findings indicated 8.7% of all smokers who completed both study waves reported that they initiated a quit attempt and 26.1% reported that they reduced the amount they smoke in the two months following the partial smoking ban. Thus, partial smoking bans on shelter campuses may not motivate substantial
Table 3 Expected and actual effects of partial smoking ban among smokers, n = 299. W1 only n = 117
W2 only n = 112
W1 vs. W2
W1 and W2 cohort n = 70
If smoking was no longer allowed in half the shelter courtyard
Expected effects partial ban (%)
Actual effects partial ban (%)
p
W1 expected effects partial ban (%)
W2 actual effects partial ban (%)
p
I would try (have tried) to move to another shelter I would try (have tried) to quit smoking I would try to (have) decrease(d) the amount I smoke I would spend (now spend) more time outside the shelter I would smoke (now smoke) more outside the shelter gates I would just smoke on the smoking side of the courtyard
8.5 23.9 27.4 17.9 23.9 56.4
3.6 8.0 25.9 7.1 15.2 –
.117 .001 .804 .014 .096
4.3 31.4 27.5 15.7 24.3 58.6
2.9 8.7 26.1 14.5 10.1 –
.658 b.001 .843 .999 .038
W1 = completed wave 1 only; W2 = completed wave 2 only; W1 and W2 cohort = completed both wave 1 and wave 2.
M.S. Businelle et al. / Addictive Behaviors 40 (2015) 51–56
numbers of homeless adults to initiate a quit attempt, but may be more likely to prompt efforts to reduce the number of cigarettes smoked. Importantly, self-reported “reductions in smoking” did not correspond to significant reductions in self-reported CPD in pre–post ban analyses. Future research using more sensitive methods (e.g., ecological momentary assessment) to follow residents over a longer period of time are needed to further examine the effects of smoking bans (partial and full) and whether smoking bans have an impact on the culture of tobacco use among sheltered homeless adults. Prior to the partial smoking ban, 16% to 18% of current smokers reported that they would likely spend more time outside the shelter campus if a partial smoking ban was implemented. W2 findings indicated that 15% of smokers who completed both study visits actually reported spending more time outside the shelter campus after the partial ban was implemented. These findings coupled with increased support of the partial ban at the W2 assessment may indicate that partial smoking bans at shelters may be well-tolerated by the shelter population. Study findings indicated that the partial smoking ban was related to a reduction in CO among moderate/heavy smokers who completed both W1 and W2. The decrease in CO coupled with no self-reported change in CPD raises the possibility that these smokers may not have reduced the amount that they smoke, but rather the partial ban may have resulted in reduced exposure to secondhand smoke or changes in the way that some smokers smoke (e.g., smoking less of each cigarette). The partial ban may have led residents to smoke less in groups, smoke more outside the shelter grounds, and/or spend significant amounts of time in the smoke-free portion of the shelter courtyard, thereby reducing inhalation of secondhand smoke. It is interesting that only 10% of smokers who completed both data collection waves stated that they smoked more outside the shelter campus following the partial ban and analysis indicated no difference in the amount of time spent off campus pre- and post-ban (i.e., within subjects: W1 = 11.0 h vs. W2 = 11.5 h; between subjects: W1 only = 10.7 h vs. W2 only = 10.5 h). Thus, it is possible that expired CO was reduced directly due to the creation of the smoke-free zone. Importantly, support for a shelter-wide outdoor smoking ban was much lower than support for the partial ban. Overall, 46%–52% of all shelter residents disagreed or strongly disagreed with the potential policy of banning all smoking on shelter grounds. However, only 14% of all smokers reported that they would try to move to another shelter and 33% of shelter guests reported that they would likely spend more time outside the shelter campus if a complete smoking ban was implemented. These results are in line with previous studies of smoking bans on medical campuses which have indicated that patient volume is not significantly affected by shelter-wide smoking bans (Gadomski, Stayton, Krupa, & Jenkins, 2010; Wheeler et al., 2007). On the other hand, 35% of smokers reported that they would try to quit smoking and 26% of smokers reported that they would try to reduce the amount they smoke if a shelter-wide smoking ban was implemented. These findings highlight some of the potential positive and negative consequences that should be considered by shelter administration when contemplating this type of shelter-wide policy change. Multiple limitations of this research warrant mentioning. First, data were collected at one shelter with a large courtyard area in a warm weather climate. Thus, findings may not generalize to other types of shelters in other areas of the country. However, the current study sample is very similar to the overall homeless population in Dallas (see Metro Dallas Homeless Alliance, 2014), but is more male and African American than the overall US homeless adult population (see Substance Abuse & Mental Health Services Administration, 2011). Notably, smoking prevalence for this sample is very similar to what has been found in other samples of homeless adults (Baggett & Rigotti, 2010; Connor et al., 2002; Lee et al., 2005). Second, although the partial smoking ban could have played a role in attenuation of CO among participants who completed both study waves, there may be other variables (e.g., time since last cigarette, reductions in the portion of
55
each cigarette smoked) that could also account for this observed change. Likewise, differences in participants' attitudes, opinions, and effects of the partial smoking ban among those who completed only one wave of data collection (i.e., W1 vs. W2 analyses) should be interpreted with caution since unmeasured group differences may have had an impact on these results. Third, the observed attenuation of CO at the W2 assessment may be statistically significant, but may not be clinically significant. Importantly, the observed reduction in CO was not accompanied by a reported reduction in CPD, but 26% of smokers who completed both the W1 and W2 assessments reported that they were smoking fewer cigarettes at the W2 assessment. Smokers commonly round CPD estimates (see Shiffman, 2009) so small changes in CPD may not have been reported by many participants. Future research should examine whether partial smoking bans at shelters have accumulating benefits beyond the period examined in the current study. Although smoking is a major contributor to premature death in homeless adults, many shelters allocate few resources toward helping homeless smokers quit (Okuyemi et al., 2006). Allocation of smoking cessation resources (e.g., offering smoking cessation counseling) and banning smoking on shelter grounds may help to attenuate smoking related health disparities among homeless adults (Porter, Houston, Anderson, & Maryman, 2011). We found that most homeless smokers want to quit and most homeless adults were supportive of outdoor smoke free zones on the shelter campus. Findings identify and quantify potential barriers to partial and shelter-wide smoking bans and thus may inform tobacco control policies at homeless shelters. Partial smoking bans may be a low cost way to reduce environmental tobacco smoke exposure and may ultimately have an effect on shelter culture as it relates to smoking. In addition, smoking bans may prompt homeless smokers to think about reducing or quitting smoking. Role of funding sources This work was supported by institutional funding provided by the University of Texas Health Science Center, School of Public Health (to MSB) and the University of Texas MD Anderson Cancer Center (to LRR). Data analysis and manuscript preparation were additionally supported through grant MRSGT-10-104-01-CPHPS (to Kendzor) and MRSGT-12-11401-CPPB (to Businelle) awarded by the American Cancer Society. The funding sources had no further role in the study design; in the collection, analysis and interpretation of data; in the writing of the report; or in the decision to submit the paper for publication. Contributors Drs. Businelle, Kendzor and Reitzel were the senior researchers who conceptualized the study, oversaw data collection, and wrote the first draft of the paper. Mrs. Poonawalla, Dr. Ma, and Dr. Reingle conducted the statistical analyses and wrote the first and revised drafts of the Results section. Dr. Baggett and Dr. Reingle critically reviewed the manuscript and made substantial contributions to the manuscript focus. Dr. Businelle, Mrs. Poonawalla, Dr. Kendzor, Ms. Rios, Ms. Cuate, Ms. Savoy, and Dr. Reitzel collected study data. All authors reviewed and approved the manuscript prior to initial submission. Conflicts of interests The authors declare no conflicts of interest pertaining to this work. Acknowledgments We thank the staff at the Bridge Homeless Assistance Center for their work and support throughout the data collection portion of this project. In addition, we thank Jay Dunn (Bridge CEO) for his notable efforts that enabled this research.
References Abrams, D. B., Niaura, R., Brown, R. A., Emmons, K. M., Goldstein, M. G., & Monti, P. M. (2003). Assessment to inform smoking cessation treatment. In D. B. Abrams, & R. Niaura (Eds.), The tobacco dependence treatment handbook: A guide to best practices (pp. 27–72). New York: Guilford Press. Apollonio, D. E., & Malone, R. E. (2005). Marketing to the marginalised: Tobacco industry targeting of the homeless and mentally ill. Tobacco Control, 14, 409–415. Arangua, L., McCarthy, W. J., Moskowitz, R., Gelberg, L., & Kuo, T. (2007). Are homeless transitional shelters receptive to environmental tobacco control interventions? Tobacco Control, 16, 143–144. Arozullah, A. M., Yarnold, P. R., Bennett, C. L., Soltysik, R. C., Wolf, M. S., Ferreira, R. M., et al. (2007). Development and validation of a short-form, rapid estimate of adult literacy in medicine. Medical Care, 45, 1026–1033.
56
M.S. Businelle et al. / Addictive Behaviors 40 (2015) 51–56
Baggett, T. P., Hwang, S. W., O'Connell, J. J., Porneala, B. C., Stringfellow, E. J., Orav, E. J., et al. (2013). Mortality among homeless adults in Boston: Shifts in causes of death over a 15-year period. Journal of the American Medical Association Internal Medicine, 173, 189–195. Baggett, T. P., & Rigotti, N. A. (2010). Cigarette smoking and advice to quit in a national sample of homeless adults. American Journal of Preventive Medicine, 39, 164–172. Baggett, T. P., Tobey, M. L., & Rigotti, N. A. (2013). Tobacco use among homeless people — Addressing the neglected addiction. New England Journal of Medicine, 369, 201–204. Borland, R., Owen, N., Hill, D., & Chapman, S. (1990). Changes in acceptance of workplace smoking bans following their implementation: A prospective study. Preventive Medicine, 19, 314–322. Businelle, M. S., Cuate, E. L., Kesh, A., Poonawalla, I. B., & Kendzor, D. E. (2013). Comparing homeless smokers to economically disadvantaged domiciled smokers. American Journal of Public Health, 103, S218–S220. Businelle, M. S., Ma, P., Kendzor, D. E., Reitzel, L. R., Chen, M., Lam, C. Y., et al. (2014). Predicting quit attempts among homeless smokers seeking cessation treatment: An ecological momentary assessment study. Nicotine & Tobacco Research (in press). Centers for Disease Control and Prevention (2014). Current cigarette smoking among adults — United States, 2005–2012. Morbidity and Mortality Weekly Report, 63, 29–34. Connor, S. E., Cook, R. L., Herbert, M. I., Neal, S. M., & Williams, J. T. (2002). Smoking cessation in a homeless population: There is a will, but is there a way? Journal of General Internal Medicine, 17, 369–372. Gadomski, A. M., Stayton, M., Krupa, N., & Jenkins, P. (2010). Implementing a smoke-free medical campus: Impact on inpatient and employee outcomes. Journal of Hospital Medicine, 5, 51–54. Hwang, S. W., Wilkins, R., Tjepkema, M., O'Campo, P. J., & Dunn, J. R. (2009). Mortality among residents of shelters, rooming houses, and hotels in Canada: 11 year follow-up study. British Medical Journal, 339, b4036. Lee, T. C., Hanlon, J. G., Ben-David, J., Booth, G. L., Cantor, W. J., Connelly, P. W., et al. (2005). Risk factors for cardiovascular disease in homeless adults. Circulation, 111, 2629–2635.
McCarthy, W. J., & Dyrness, G. (2012). CPPW-L.A. Smoking reduction program for transitional shelters for the homeless. Retrieved May 2, 2013, from. http://publichealth.lacounty. gov/tob/pdf/bh/Bill_McCarthy_Homeles_%20Military_Communities.pdf (May 2, 2013) Metro Dallas Homeless Alliance (2014). 2013 point-in-time homeless count and census. Retrieved July 21, 2014, from http://www.mdhadallas.org/Libraries/MDHA_ Documents/2013_Point-in-Time_Homeless_Count_Census_Report.sflb.ashx National Coalition for the Homeless (2007). The National Survey on Tobacco Prevention at Homeless Service Providers. from. http://www.nationalhomeless.org/factsheets/ tobacco.pdf Okuyemi, K. S., Caldwell, A.R., Thomas, J. L., Born, W., Richter, K. P., Nollen, N., et al. (2006). Homelessness and smoking cessation: Insights from focus groups. Nicotine & Tobacco Research, 8, 287–296. Porter, J., Houston, L., Anderson, R. H., & Maryman, K. (2011). Addressing tobacco use in homeless populations: Recommendations of an expert panel. Health Promotion Practice, 12, 144S–151S. Shiffman, S. (2009). How many cigarettes did you smoke? Assessing cigarette consumption by global report, time-line follow-back, and ecological momentary assessment. Health Psychology, 28, 519–526. Substance Abuse and Mental Health Services Administration (2011). Current statistics on the prevalence and characteristics of people experiencing homelessness in the United States. Retrieved July 21, 2014, from. http://homeless.samhsa.gov/ResourceFiles/hrc_ factsheet.pdf Unrod, M., Oliver, J. A., Heckman, B. W., Simmons, V. N., & Brandon, T. H. (2012). Outdoor smoking ban at a cancer center: Attitudes and smoking behavior among employees and patients. Journal of Public Health Management and Practice, 18. Wheeler, J. G., Pulley, L., Felix, H. C., Bursac, Z., Siddiqui, N. J., Stewart, M. K., et al. (2007). Impact of a smoke-free hospital campus policy on employee and consumer behavior. Public Health Reports, 122, 744–752.
Contents lists available at ScienceDirect
Addictive Behaviors
Smoking policy change at a homeless shelter: Attitudes and effects Michael S. Businelle a,b,⁎, Insiya B. Poonawalla a, Darla E. Kendzor a,b, Debra M. Rios a, Erica L. Cuate a, Elaine J. Savoy c, Ping Ma a,b, Travis P. Baggett d, Jennifer Reingle a, Lorraine R. Reitzel e a
University of Texas School of Public Health, Department of Health Promotion and Behavioral Sciences, Dallas, TX, USA University of Texas Southwestern Medical Center, Dallas, TX, USA University of Houston, Department of Clinical Psychology, Houston, TX, USA d Massachusetts General Hospital, Division of General Internal Medicine, Boston, MA, USA e University of Houston, Department of Educational Psychology, Houston, TX, USA b c
H I G H L I G H T S • • • •
Sheltered homeless adults support the creation of large outdoor smoke free zones. Support for shelter-wide smoking bans is more limited. Smoking bans at shelter campuses may reduce carbon monoxide levels in smokers. Findings indicate potential consequences of partial and shelter-wide smoking bans.
a r t i c l e
i n f o
Available online 6 September 2014 Keywords: Homelessness Tobacco Smoking ban Policy
a b s t r a c t Homeless adults are exposed to more smokers and smoke in response to environmental tobacco cues more than other socioeconomically disadvantaged groups. Addressing the culture of smoking in homeless shelters through policy initiatives may support cessation and improve health in this vulnerable and understudied population. This study examined support for and expected/actual effects of a smoking ban at a homeless shelter. A 2-wave crosssectional study with an embedded cohort was conducted in the summer of 2013 two weeks before (wave 1) and two months after (wave 2) a partial outdoor smoking ban was implemented. A total of 394 homeless adults were surveyed (i.e., wave 1 [n = 155]; wave 2 [n = 150]; and 89 additional participants completed both waves). On average, participants were 43 years old, primarily African American (63%), male (72%), and had been homeless for the previous 12 months (median). Most participants were smokers (76%) smoking 12 cigarettes per day on average. Most participants supported the creation of a large smoke-free zone on the shelter campus, but there was less support for a shelter-wide smoking ban. Average cigarettes smoked per day did not differ between study waves. However, participants who completed both study waves experienced a reduction in expired carbon monoxide at wave 2 (W1 = 18.2 vs. W2 = 15.8 parts per million, p = .02). Expected effects of the partial ban were similar to actual effects. Partial outdoor smoking bans may be well supported by homeless shelter residents and may have a positive impact on shelter resident health. © 2014 Elsevier Ltd. All rights reserved.
1. Introduction Smoking prevalence has declined to 18.1% among US adults (Centers for Disease Control & Prevention, 2014), yet over 70% of homeless adults are current smokers (e.g., Baggett & Rigotti, 2010; Connor, Cook, Herbert, Neal, & Williams, 2002; Lee et al., 2005). Although homeless smokers and other socioeconomically disadvantaged domiciled smokers may have comparable numbers of quit attempts (Businelle, ⁎ Corresponding author at: University of Texas School of Public Health, 6011 Harry Hines Blvd., V8.112, Dallas, TX 75390-9128, USA. Tel.: +1 214 648 1070; fax: +1 214 648 1081. E-mail address: [email protected] (M.S. Businelle).
http://dx.doi.org/10.1016/j.addbeh.2014.08.013 0306-4603/© 2014 Elsevier Ltd. All rights reserved.
Cuate, Kesh, Poonawalla, & Kendzor, 2013), homeless smokers may have more difficulty maintaining smoking abstinence partially due to a culture of tobacco use (Baggett, Tobey, & Rigotti, 2013) and permissive shelter policies that result in frequent exposure to smoking. Recent findings indicate that homeless smokers are exposed to substantially more smokers and are more likely to report smoking in response to social goads and other cues than socioeconomically disadvantaged domiciled smokers (Businelle et al., 2013). Thus, outdoor smoking bans at shelters may reduce social cues to smoke and increase the likelihood of successful smoking cessation among homeless individuals (Baggett, Tobey, et al., 2013). Despite the alarmingly high negative health consequences of smoking in homeless individuals (Baggett, Hwang, et al., 2013;
52
M.S. Businelle et al. / Addictive Behaviors 40 (2015) 51–56
Baggett, Tobey, et al., 2013; Hwang, Wilkins, Tjepkema, O'Campo, & Dunn, 2009), few shelters have banned all smoking on their campuses (Apollonio & Malone, 2005; National Coalition for the Homeless, 2007). Furthermore, a survey of transitional shelters in the Los Angeles area indicated that only half of those with ≥200 beds had designated smoke free areas (Arangua, McCarthy, Moskowitz, Gelberg, & Kuo, 2007). Although many shelter administrators may recognize the harmful effects of smoking, their concerns about the potential negative consequences that may accompany shelter-wide smoking bans (e.g., violations of the smoking ban, greater presence of homeless individuals outside the shelter campus) may discourage this type of health promoting policy change. Notwithstanding these concerns, homeless individuals may actually support partial or full smoking bans on shelter grounds. For example, a survey of homeless adults at 26 shelters in Los Angeles found that 46% of respondents supported a “smoking ban in all outdoor common areas” (McCarthy & Dyrness, 2012). In addition, support for smoking bans has been shown to increase following implementation in workplace and medical settings (e.g., Borland, Owen, Hill, & Chapman, 1990; Unrod, Oliver, Heckman, Simmons, & Brandon, 2012), and this may also be the case for homeless shelter residents. However, no research has prospectively examined the impact of shelter smoking bans on the attitudes and behaviors of homeless smokers. The purpose of the current study was to examine shelter residents' attitudes about partial and full smoking bans at a homeless shelter in Dallas, Texas. We also examined the effects of a partial smoking ban on smoking behavior and expired carbon monoxide (CO), assessed two months after a partial ban was implemented. We tested the following hypotheses: 1) non-smokers would be more supportive of partial and shelter-wide smoking bans than smokers, 2) the number of supporters of the partial and full smoking bans would increase after the partial ban was implemented, and 3) the number of cigarettes smoked per day and CO levels among current smokers would be reduced after implementation of the partial smoking ban. 2. Methods 2.1. Participants Data collection took place at a large homeless shelter in Dallas, Texas. Homeless individuals were eligible to participate in this study if they spent the previous night at the shelter, earned a score of ≥ 4 on the Rapid Estimate of Adult Literacy in Medicine—Short Form (REALM-SF; Arozullah et al., 2007) indicating N6th grade English literacy level (i.e., required to complete tablet/laptop assessment items), and were at least 18 years old.
approved by the Institutional Review Board at the University of Texas School of Public Health. Flyers describing the study and indicating the dates and times of data collection were posted on the shelter campus. Interested shelter residents approached study staff on the scheduled data collection days. Informed consent was obtained from interested individuals and screening was completed on-site. Those who met the study inclusion criteria used tablet or laptop computers to complete study questionnaires. Questionnaire Development System software (version 2.6.1) was used whereby all questions were read aloud, through earphones. Participants responded to assessment items (using a mouse or touch screen) after the entire question was read. Participants who completed the assessments received a $20 gift card.
2.3. Measures All participants completed measures that assessed demographic characteristics including age, gender, race/ethnicity, and current period of homelessness. Participants provided a CO sample by breathing into a Vitalograph CO ecolyzer and were asked about their smoking status (never smoker, ex-smoker [smoked at least 100 cigarettes in lifetime, but no longer a smoker], and current smoker). Current smokers were also asked about current cigarettes smoked per day (CPD), number of lifetime quit attempts lasting ≥ 24 h, and readiness to quit smoking (Abrams et al., 2003). Finally, participants' attitudes about partial (“I support the creation of a smoke free zone in half of the shelter courtyard”) and shelter-wide (“I support a complete smoking ban at the shelter”) bans were assessed on 5 point scales ranging from strongly disagree to strongly agree (this scale was dichotomized so that those who strongly agreed or agreed were coded as “1” and those who were neutral, disagreed, or strongly disagreed were coded as “0”). To assess expected effects of the bans, all participants were asked “How would you react if smoking was no longer allowed in half the shelter courtyard?”, and “How would you react if the shelter no longer allowed smoking anywhere on the shelter campus?” To assess the actual effects of the partial ban, W2 participants were asked “What effect has the shelter ban on smoking in half the courtyard had on you?” (see Table 3 for response options). Finally, participants rated their agreement with the following statements “Banning all smoking at the shelter would improve the health of shelter guests” and “Banning all smoking at the shelter would reduce smoking in guests.” Response options ranged from strongly disagree to strongly agree (response options for these items were also dichotomized).
2.4. Statistical analyses 2.2. Procedure The partial smoking ban was prompted by the results of another study at this shelter, which indicated that smokers who were trying to quit smoking were exposed to approximately 40 smokers each day, partially due to a lack of an outdoor shelter smoke free zone (Businelle et al., 2013, in press). Two separate waves of data collection occurred during the summer of 2013. Wave 1 (W1) data were collected over a one week period, two weeks prior to the implementation of the partial outdoor smoking ban that covered one-half (approximately 6750 ft2) of the shelter courtyard. At the time of W1 data collection, residents were unaware of plans for the ban implementation. Collection of wave 2 (W2) data was completed over a one week period that began two months after the partial ban was implemented. Participants who completed the W1 assessments were encouraged to complete the W2 assessments, but they were not required to do so. Therefore, some participants completed the W1 assessment or W2 assessment, and a subset of participants completed both W1 and W2 assessments. This study was
Participant characteristics, attitudes about smoking bans (i.e., partial and full bans), and the effects of a partial smoking ban were descriptively analyzed for the total sample and separately for each data collection wave. Independent sample t-tests and Chi-square tests were conducted to identify differences among participants who completed only one wave of data collection (i.e., either W1 or W2). Paired t-tests and McNemar's tests were conducted to test for within-subjects changes in study variables for those who completed both study waves. Changes in mean CPD and expired CO levels among smokers were analyzed using linear regression (i.e., for comparisons of participants who completed only W1 or W2) and linear multilevel mixed (LMM) modeling analyses (i.e., for comparisons of participants who completed both study waves). Adjusted models controlled for age, gender, race (White vs. non-White), and education. In addition, data were analyzed to determine if post-ban changes in CPD and CO differed by level of smoking (light smoker [≤10 CPD] vs. moderate to heavy smoker [N10 CPD]).
M.S. Businelle et al. / Addictive Behaviors 40 (2015) 51–56
53
p b .001). Analyses were also conducted with the cohort who completed both data collection waves, with similar results.
3. Results Overall, 394 unique shelter residents completed the study (i.e., W1 [n = 155]; W2 [n = 150]; and 89 additional participants completed both waves). Approximately 75% of all current shelter residents are enrolled in the study at each wave. Participants were 43.3 years old (SD = 11.8), predominantly male (71.8%), non-White (69.8%), and reported that their current period of homeless began 12.0 months ago (median). Most participants (75.9%) were current smokers who smoked an average of 12.0 CPD (SD = 7.2). Most current smokers (63%) reported that they planned to quit smoking within the next six months. Participants who completed only W1 were approximately five years younger than those who completed only W2 (p b .001) and four years younger than those who completed both data collection waves (p = .012). Cohorts were not significantly different on other demographic, homelessness, or smoking characteristics. Participant characteristics are presented in Table 1.
3.1. Support for partial and shelter-wide smoking bans At both study waves, the majority of participants supported (i.e., agreed or strongly agreed) the creation of a smoke free zone in half of the shelter courtyard. Additionally, there was a significant increase in support for the partial ban at W2 among residents who completed both study visits (W1 = 60.7% vs. W2 = 73.9%, p = .02) and a trend in the same direction for those who only completed one wave of data collection (W1 = 60.0% vs. W2 = 70.0%, p = .07; Table 2). Fewer participants supported a shelter-wide ban and support for the full ban did not differ by study wave or change over time within the embedded cohort (Table 2). A minority of smokers agreed that a shelterwide smoking ban would actually reduce smoking in shelter guests, but most smokers agreed that such a ban would improve the health of shelter residents (Table 2). Additional analyses compared non-smokers' and smokers' support for the partial and shelter-wide smoking bans among those who completed only one wave of data collection. Results indicated that nonsmokers and smokers were not significantly different in their support for the partial ban at W1 (71.1% vs. 56.4%, p = .111). However, nonsmokers were more supportive of the partial ban than smokers at W2 (84.2% vs. 65.2%, p = .027) and more supportive of the shelter-wide ban at W1 (63.2% vs. 17.9%, p b .001) and W2 (60.5% vs. 27.7%,
3.2. Comparison of the expected and actual effects of the partial smoking ban Many of the expected effects of a hypothetical partial smoking ban (assessed at W1) were similar to the actual effects of the partial ban (assessed at W2) among the cohort of smokers who completed both waves of data collection (see Table 3). Specifically, before and after the partial ban, similar numbers of participants reported that a partial ban would/did result in: 1) attempts to move to another shelter, 2) reductions in smoking, and 3) more time spent outside the shelter campus. However, fewer participants tried to quit smoking (see Table 3; p b .001) and fewer participants increased their smoking outside the shelter grounds than expected (see Table 3; p = .038) following the partial ban. Expected effects of the partial ban among those who completed W1 only and actual effects of the partial ban among those who completed W2 only are displayed in Table 3. Comparisons of those who completed only one wave of data collection should be interpreted with caution because other factors may have contributed to the observed differences across waves. 3.3. Expected effects of a shelter-wide smoking ban Smokers were asked to indicate their anticipated response(s) to a hypothetical shelter-wide smoking ban at each wave of assessment. No actual shelter-wide ban was implemented during the study period and there was very little difference across study waves in the expected effects of the hypothetical shelter-wide ban; therefore, all 299 smokers who participated in this study were included in the numbers below. Smokers reported that, in response to a shelter-wide smoking ban, 13.7% would try to move to a different shelter, 33.4% would spend more time away from the shelter, 42.8% would smoke more outside the shelter gates, 34.8% would try to quit smoking, and 25.8% would try to decrease the amount that they smoke. 3.4. Changes in CPD and CO across study waves After controlling for age, gender, race, and education, a betweensubjects comparison of smokers who completed only W1 or W2 (W1 = 12.3 vs. W2 = 12.0 CPD) and within-subjects analyses of
Table 1 Participant characteristics. Characteristic
Total unique N = 394
W1 n = 155
W2 n = 150
p value (W1 vs. W2)
W1 and W2 cohort n = 89
p value (W1 vs. cohort)
Age in years [M (SD)] Race (%) White African American All others Male (%) Unemployed (%) Current months of homelessness [M (SD)] Median (range: 0–366) Smoking status (%) Current smokers Ex-smokers Never smokers Cigarettes per day (CPD) [M (SD)]a Heaviness of smoking (%)a Light (≤10 CPD) Moderate/heavy (N10 CPD) Carbon monoxide level, ppm [M (SD)] Previous quit attempts lasting ≥24 h [M (SD)]a
43.3 (11.8)
40.6 (12.6)
45.4 (10.5)
b.001 .205
44.7 (11.7)
.012 .726
30.0 62.8 7.2 71.8 90.1 24.8 (39.5) 12
27.9 62.3 9.8 69.0 87.1 26.6 (42.9) 12.5
34.7 60.0 5.3 74.0 93.3 19.0 (23.6) 9.8
75.9 11.4 12.7 12.0 (7.2)
75.5 11.0 13.5 12.3 (8.1)
74.7 12.7 12.7 12.0 (6.8)
53.2 46.8 13.7 (12.5) 5.1 (3.1)
56.4 43.6 14.2 (13.5) 5.3 (3.0)
49.1 50.9 12.4 (10.5) 4.8 (3.1)
.337 .067 .081
25.8 68.5 5.7 73.0 89.9 31.0 (51.2) 14
.888
.798 .289
.195 .199
W1 = completed wave 1 only; W2 = completed wave 2 only; W1 and W2 cohort = completed both wave 1 and wave 2. a Only current smokers were asked these questions.
.509 .517 .494 .839
78.7 10.1 11.2 11.6 (6.1) 54.3 45.7 15.0 (13.6) 5.1 (3.2)
.581 .968
.640 .618
54
M.S. Businelle et al. / Addictive Behaviors 40 (2015) 51–56
Table 2 Attitudes and opinions about partial and shelter-wide smoking bans. Attitude
I support the creation of a smoke free zone in half the shelter courtyard Strongly agree or agree (%) Strongly disagree or disagree (%) I support a complete smoking ban at the shelter Strongly agree or agree (%) Strongly disagree or disagree (%) Smoker's opinions about shelter-wide smoking ban Banning all smoking at the shelter would reduce smoking in guests Strongly agree or agree (%) Strongly disagree or disagree (%) Banning all smoking at the shelter would improve guest health Strongly agree or agree (%) Strongly disagree or disagree (%)
W1 n = 155
W2 n = 150
W1 vs. W2
W1 and W2 cohort n = 89
Pre-ban
Post-ban
p
Pre-ban
Post-ban
p
60.0 16.1
70.0 11.3
.067
60.7 11.2
73.9 12.5
.019
29.0 47.1
36.0 40.7
.194
30.3 46.1
33.0 52.3
n = 117
n = 112
p
n = 70
27.4 45.3
42.9 38.4
.043
38.6 48.6
33.3 49.3
.523
53.8 18.8
55.4 25.0
.820
58.6 18.6
49.3 29.0
.230
.832 p
W1 = completed wave 1 only; W2 = completed wave 2 only; W1 and W2 cohort = completed both wave 1 and wave 2.
smokers who completed both study waves (W1 = 11.6 vs. W2 = 12.0 CPD) indicated no differences in CPD before and after the partial ban (all ps N .68). However, within-subjects analyses indicated that CO was significantly lower after implementation of the partial smoking ban among participants who completed both study waves (W1 = 15.0 vs. W2 = 12.7 parts per million [ppm], p = .02). Follow-up analyses indicated that significant reductions in CO were found among smokers (W1 = 18.8 vs. W2 = 15.8 ppm, p = .02), but CO was not significantly different among non-smokers who completed both study waves (W1 = 1.1 vs. W2 = 0.9 ppm). Results of a between-subjects CO analysis that compared those who completed W1 to those who completed W2 were in a similar direction, though non-significant (W1 = 14.2 vs. W2 = 12.4 ppm, p = .20). We conducted additional analyses to determine if the partial ban had a differential impact on CPD and CO among smokers by heaviness of smoking (i.e., ≤10 vs. N 10 CPD). We found no change in CPD following the partial smoking ban for light or moderate/heavy smokers in within- and between-subjects analyses. However, CO levels were lower at W2 among moderate to heavy smokers (N10 CPD), for both between-subjects (W1 = 23.9 vs. W2 = 18.6 ppm, p b .01) and within-subjects (W1 = 25.5 vs. W2 = 20.7 ppm, p b .05) analyses. No significant change in CO was detected among light smokers. 4. Discussion The current study is the first to examine pre-ban attitudes and postban effects of a partial smoking ban in a homeless shelter setting. Results yielded three key findings with potential implications for shelter policies concerning smoking. First, most shelter residents (smokers and non-smokers) supported a partial outdoor smoking ban on the shelter grounds, both before and after the partial ban was implemented.
However, residents were far less likely to support a shelter-wide smoking ban, even though most smokers in this study believed that a complete ban would improve the health of shelter residents. Compared to smokers, non-smokers were 15% to 45% more likely to support the smoking bans. Second, the partial ban was related to reductions in carbon monoxide levels, potentially due to changes in active smoking or reduced environmental tobacco smoke exposure among some smokers. Finally, expected and actual effects of the partial smoking ban indicated that this type of shelter policy change may be well tolerated with few negative consequences. Thus, smoking bans at homeless shelters may be an important component of comprehensive shelter policies that aim to improve the health of shelter residents. An overwhelming majority of shelter residents supported the creation of a smoke free zone encompassing one half of the outdoor shelter courtyard. Consistent with previous studies, the partial smoking ban became more popular with smokers following implementation (Borland et al., 1990; Unrod et al., 2012). This level of support for the partial smoking ban is especially important considering that 76% of shelter residents were smokers. It is notable that shelter security has reported few violations (i.e., smoking in the smoke-free zone) since the ban was implemented. Additionally, shelter security recently reported that the smoke free courtyard area is now used more than before the ban was implemented. Prior to learning about the implementation of the partial smoking ban, many W1 participants reported that they would try to quit (24% to 31%) and/or reduce the amount that they smoke (27%–28%) if a partial outdoor smoking ban was implemented. Findings indicated 8.7% of all smokers who completed both study waves reported that they initiated a quit attempt and 26.1% reported that they reduced the amount they smoke in the two months following the partial smoking ban. Thus, partial smoking bans on shelter campuses may not motivate substantial
Table 3 Expected and actual effects of partial smoking ban among smokers, n = 299. W1 only n = 117
W2 only n = 112
W1 vs. W2
W1 and W2 cohort n = 70
If smoking was no longer allowed in half the shelter courtyard
Expected effects partial ban (%)
Actual effects partial ban (%)
p
W1 expected effects partial ban (%)
W2 actual effects partial ban (%)
p
I would try (have tried) to move to another shelter I would try (have tried) to quit smoking I would try to (have) decrease(d) the amount I smoke I would spend (now spend) more time outside the shelter I would smoke (now smoke) more outside the shelter gates I would just smoke on the smoking side of the courtyard
8.5 23.9 27.4 17.9 23.9 56.4
3.6 8.0 25.9 7.1 15.2 –
.117 .001 .804 .014 .096
4.3 31.4 27.5 15.7 24.3 58.6
2.9 8.7 26.1 14.5 10.1 –
.658 b.001 .843 .999 .038
W1 = completed wave 1 only; W2 = completed wave 2 only; W1 and W2 cohort = completed both wave 1 and wave 2.
M.S. Businelle et al. / Addictive Behaviors 40 (2015) 51–56
numbers of homeless adults to initiate a quit attempt, but may be more likely to prompt efforts to reduce the number of cigarettes smoked. Importantly, self-reported “reductions in smoking” did not correspond to significant reductions in self-reported CPD in pre–post ban analyses. Future research using more sensitive methods (e.g., ecological momentary assessment) to follow residents over a longer period of time are needed to further examine the effects of smoking bans (partial and full) and whether smoking bans have an impact on the culture of tobacco use among sheltered homeless adults. Prior to the partial smoking ban, 16% to 18% of current smokers reported that they would likely spend more time outside the shelter campus if a partial smoking ban was implemented. W2 findings indicated that 15% of smokers who completed both study visits actually reported spending more time outside the shelter campus after the partial ban was implemented. These findings coupled with increased support of the partial ban at the W2 assessment may indicate that partial smoking bans at shelters may be well-tolerated by the shelter population. Study findings indicated that the partial smoking ban was related to a reduction in CO among moderate/heavy smokers who completed both W1 and W2. The decrease in CO coupled with no self-reported change in CPD raises the possibility that these smokers may not have reduced the amount that they smoke, but rather the partial ban may have resulted in reduced exposure to secondhand smoke or changes in the way that some smokers smoke (e.g., smoking less of each cigarette). The partial ban may have led residents to smoke less in groups, smoke more outside the shelter grounds, and/or spend significant amounts of time in the smoke-free portion of the shelter courtyard, thereby reducing inhalation of secondhand smoke. It is interesting that only 10% of smokers who completed both data collection waves stated that they smoked more outside the shelter campus following the partial ban and analysis indicated no difference in the amount of time spent off campus pre- and post-ban (i.e., within subjects: W1 = 11.0 h vs. W2 = 11.5 h; between subjects: W1 only = 10.7 h vs. W2 only = 10.5 h). Thus, it is possible that expired CO was reduced directly due to the creation of the smoke-free zone. Importantly, support for a shelter-wide outdoor smoking ban was much lower than support for the partial ban. Overall, 46%–52% of all shelter residents disagreed or strongly disagreed with the potential policy of banning all smoking on shelter grounds. However, only 14% of all smokers reported that they would try to move to another shelter and 33% of shelter guests reported that they would likely spend more time outside the shelter campus if a complete smoking ban was implemented. These results are in line with previous studies of smoking bans on medical campuses which have indicated that patient volume is not significantly affected by shelter-wide smoking bans (Gadomski, Stayton, Krupa, & Jenkins, 2010; Wheeler et al., 2007). On the other hand, 35% of smokers reported that they would try to quit smoking and 26% of smokers reported that they would try to reduce the amount they smoke if a shelter-wide smoking ban was implemented. These findings highlight some of the potential positive and negative consequences that should be considered by shelter administration when contemplating this type of shelter-wide policy change. Multiple limitations of this research warrant mentioning. First, data were collected at one shelter with a large courtyard area in a warm weather climate. Thus, findings may not generalize to other types of shelters in other areas of the country. However, the current study sample is very similar to the overall homeless population in Dallas (see Metro Dallas Homeless Alliance, 2014), but is more male and African American than the overall US homeless adult population (see Substance Abuse & Mental Health Services Administration, 2011). Notably, smoking prevalence for this sample is very similar to what has been found in other samples of homeless adults (Baggett & Rigotti, 2010; Connor et al., 2002; Lee et al., 2005). Second, although the partial smoking ban could have played a role in attenuation of CO among participants who completed both study waves, there may be other variables (e.g., time since last cigarette, reductions in the portion of
55
each cigarette smoked) that could also account for this observed change. Likewise, differences in participants' attitudes, opinions, and effects of the partial smoking ban among those who completed only one wave of data collection (i.e., W1 vs. W2 analyses) should be interpreted with caution since unmeasured group differences may have had an impact on these results. Third, the observed attenuation of CO at the W2 assessment may be statistically significant, but may not be clinically significant. Importantly, the observed reduction in CO was not accompanied by a reported reduction in CPD, but 26% of smokers who completed both the W1 and W2 assessments reported that they were smoking fewer cigarettes at the W2 assessment. Smokers commonly round CPD estimates (see Shiffman, 2009) so small changes in CPD may not have been reported by many participants. Future research should examine whether partial smoking bans at shelters have accumulating benefits beyond the period examined in the current study. Although smoking is a major contributor to premature death in homeless adults, many shelters allocate few resources toward helping homeless smokers quit (Okuyemi et al., 2006). Allocation of smoking cessation resources (e.g., offering smoking cessation counseling) and banning smoking on shelter grounds may help to attenuate smoking related health disparities among homeless adults (Porter, Houston, Anderson, & Maryman, 2011). We found that most homeless smokers want to quit and most homeless adults were supportive of outdoor smoke free zones on the shelter campus. Findings identify and quantify potential barriers to partial and shelter-wide smoking bans and thus may inform tobacco control policies at homeless shelters. Partial smoking bans may be a low cost way to reduce environmental tobacco smoke exposure and may ultimately have an effect on shelter culture as it relates to smoking. In addition, smoking bans may prompt homeless smokers to think about reducing or quitting smoking. Role of funding sources This work was supported by institutional funding provided by the University of Texas Health Science Center, School of Public Health (to MSB) and the University of Texas MD Anderson Cancer Center (to LRR). Data analysis and manuscript preparation were additionally supported through grant MRSGT-10-104-01-CPHPS (to Kendzor) and MRSGT-12-11401-CPPB (to Businelle) awarded by the American Cancer Society. The funding sources had no further role in the study design; in the collection, analysis and interpretation of data; in the writing of the report; or in the decision to submit the paper for publication. Contributors Drs. Businelle, Kendzor and Reitzel were the senior researchers who conceptualized the study, oversaw data collection, and wrote the first draft of the paper. Mrs. Poonawalla, Dr. Ma, and Dr. Reingle conducted the statistical analyses and wrote the first and revised drafts of the Results section. Dr. Baggett and Dr. Reingle critically reviewed the manuscript and made substantial contributions to the manuscript focus. Dr. Businelle, Mrs. Poonawalla, Dr. Kendzor, Ms. Rios, Ms. Cuate, Ms. Savoy, and Dr. Reitzel collected study data. All authors reviewed and approved the manuscript prior to initial submission. Conflicts of interests The authors declare no conflicts of interest pertaining to this work. Acknowledgments We thank the staff at the Bridge Homeless Assistance Center for their work and support throughout the data collection portion of this project. In addition, we thank Jay Dunn (Bridge CEO) for his notable efforts that enabled this research.
References Abrams, D. B., Niaura, R., Brown, R. A., Emmons, K. M., Goldstein, M. G., & Monti, P. M. (2003). Assessment to inform smoking cessation treatment. In D. B. Abrams, & R. Niaura (Eds.), The tobacco dependence treatment handbook: A guide to best practices (pp. 27–72). New York: Guilford Press. Apollonio, D. E., & Malone, R. E. (2005). Marketing to the marginalised: Tobacco industry targeting of the homeless and mentally ill. Tobacco Control, 14, 409–415. Arangua, L., McCarthy, W. J., Moskowitz, R., Gelberg, L., & Kuo, T. (2007). Are homeless transitional shelters receptive to environmental tobacco control interventions? Tobacco Control, 16, 143–144. Arozullah, A. M., Yarnold, P. R., Bennett, C. L., Soltysik, R. C., Wolf, M. S., Ferreira, R. M., et al. (2007). Development and validation of a short-form, rapid estimate of adult literacy in medicine. Medical Care, 45, 1026–1033.
56
M.S. Businelle et al. / Addictive Behaviors 40 (2015) 51–56
Baggett, T. P., Hwang, S. W., O'Connell, J. J., Porneala, B. C., Stringfellow, E. J., Orav, E. J., et al. (2013). Mortality among homeless adults in Boston: Shifts in causes of death over a 15-year period. Journal of the American Medical Association Internal Medicine, 173, 189–195. Baggett, T. P., & Rigotti, N. A. (2010). Cigarette smoking and advice to quit in a national sample of homeless adults. American Journal of Preventive Medicine, 39, 164–172. Baggett, T. P., Tobey, M. L., & Rigotti, N. A. (2013). Tobacco use among homeless people — Addressing the neglected addiction. New England Journal of Medicine, 369, 201–204. Borland, R., Owen, N., Hill, D., & Chapman, S. (1990). Changes in acceptance of workplace smoking bans following their implementation: A prospective study. Preventive Medicine, 19, 314–322. Businelle, M. S., Cuate, E. L., Kesh, A., Poonawalla, I. B., & Kendzor, D. E. (2013). Comparing homeless smokers to economically disadvantaged domiciled smokers. American Journal of Public Health, 103, S218–S220. Businelle, M. S., Ma, P., Kendzor, D. E., Reitzel, L. R., Chen, M., Lam, C. Y., et al. (2014). Predicting quit attempts among homeless smokers seeking cessation treatment: An ecological momentary assessment study. Nicotine & Tobacco Research (in press). Centers for Disease Control and Prevention (2014). Current cigarette smoking among adults — United States, 2005–2012. Morbidity and Mortality Weekly Report, 63, 29–34. Connor, S. E., Cook, R. L., Herbert, M. I., Neal, S. M., & Williams, J. T. (2002). Smoking cessation in a homeless population: There is a will, but is there a way? Journal of General Internal Medicine, 17, 369–372. Gadomski, A. M., Stayton, M., Krupa, N., & Jenkins, P. (2010). Implementing a smoke-free medical campus: Impact on inpatient and employee outcomes. Journal of Hospital Medicine, 5, 51–54. Hwang, S. W., Wilkins, R., Tjepkema, M., O'Campo, P. J., & Dunn, J. R. (2009). Mortality among residents of shelters, rooming houses, and hotels in Canada: 11 year follow-up study. British Medical Journal, 339, b4036. Lee, T. C., Hanlon, J. G., Ben-David, J., Booth, G. L., Cantor, W. J., Connelly, P. W., et al. (2005). Risk factors for cardiovascular disease in homeless adults. Circulation, 111, 2629–2635.
McCarthy, W. J., & Dyrness, G. (2012). CPPW-L.A. Smoking reduction program for transitional shelters for the homeless. Retrieved May 2, 2013, from. http://publichealth.lacounty. gov/tob/pdf/bh/Bill_McCarthy_Homeles_%20Military_Communities.pdf (May 2, 2013) Metro Dallas Homeless Alliance (2014). 2013 point-in-time homeless count and census. Retrieved July 21, 2014, from http://www.mdhadallas.org/Libraries/MDHA_ Documents/2013_Point-in-Time_Homeless_Count_Census_Report.sflb.ashx National Coalition for the Homeless (2007). The National Survey on Tobacco Prevention at Homeless Service Providers. from. http://www.nationalhomeless.org/factsheets/ tobacco.pdf Okuyemi, K. S., Caldwell, A.R., Thomas, J. L., Born, W., Richter, K. P., Nollen, N., et al. (2006). Homelessness and smoking cessation: Insights from focus groups. Nicotine & Tobacco Research, 8, 287–296. Porter, J., Houston, L., Anderson, R. H., & Maryman, K. (2011). Addressing tobacco use in homeless populations: Recommendations of an expert panel. Health Promotion Practice, 12, 144S–151S. Shiffman, S. (2009). How many cigarettes did you smoke? Assessing cigarette consumption by global report, time-line follow-back, and ecological momentary assessment. Health Psychology, 28, 519–526. Substance Abuse and Mental Health Services Administration (2011). Current statistics on the prevalence and characteristics of people experiencing homelessness in the United States. Retrieved July 21, 2014, from. http://homeless.samhsa.gov/ResourceFiles/hrc_ factsheet.pdf Unrod, M., Oliver, J. A., Heckman, B. W., Simmons, V. N., & Brandon, T. H. (2012). Outdoor smoking ban at a cancer center: Attitudes and smoking behavior among employees and patients. Journal of Public Health Management and Practice, 18. Wheeler, J. G., Pulley, L., Felix, H. C., Bursac, Z., Siddiqui, N. J., Stewart, M. K., et al. (2007). Impact of a smoke-free hospital campus policy on employee and consumer behavior. Public Health Reports, 122, 744–752.
