Smoking Behaviors and Cessation Interests Among Multiunit Subsidized Housing Tenants, Columbus, Ohio, 2011

Introduction Cessation services have been recommended to complement smoke-free policies in subsidized multiunit housing, but little is known about smoking- and cessation-related characteristics among subsidized housing tenants. This study examined smoking behaviors and cessation-related interests in a population of subsidized housing tenants. Methods A face-to-face survey was conducted in August to October 2011 with a probability sample of private subsidized housing lease holders in Columbus, Ohio (N = 301, 64% response rate). Results Almost half (47.5%) of respondents were current smokers, including smokers of cigarettes or small cigars. Smokers were less likely than nonsmokers to have health insurance and more likely to be at risk for food insecurity. Among smokers, 20.3% did not smoke daily and 35.0% smoked 5 or fewer cigarettes per day. More than half (61.3%) purchased single cigarettes in the past month, with higher rates among nondaily smokers. Most smokers intended to quit within 6 months or less (60.1%) and were interested in using nicotine replacement therapy (NRT) (65.0%). Most respondents had Medicaid but only 30.4% knew Medicaid covered cessation medications. Conclusions This population of subsidized housing tenants had high rates of smoking, including light smoking. Interest in NRT was high and access can be improved by increasing awareness of Medicaid coverage among clients and health care providers. However, more research is needed about scalable, evidence-based cessation strategies for low-socioeconomic status and light smokers. Strategies to address environmental factors such as availability of single cigarettes should also be considered in parallel with smoke-free policies.


Introduction
Smoke-free policies in publicly subsidized multiunit housing (MUH) are increasingly being promoted to protect nonsmokers of low socioeconomic status (SES) from involuntary secondhand smoke exposure in the home (1). Although these policies do not require smokers to quit smoking, it is recommended that cessation services accompany such policies (1). In the only outcome evaluation of a smoke-free policy in subsidized MUH, 15% of smokers reported quitting within about 18 months after policy implementation (2). Although self-reported data were not biochemically validated, this finding at least suggests smokers may be interested in quitting in response to smoke-free policies.
Smoking behaviors and cessation interests among smokers in subsidized housing are not well understood, and such information could be used to inform cessation programs. Estimates of smoking rates among subsidized housing tenants suggest that they exceed national estimates for the lowest SES groups (3,4). In general, low-SES smokers have higher levels of nicotine dependence and are less likely to intend to quit or to successfully quit smoking compared with their higher-SES counterparts (5,6). However, mean annual household income for subsidized housing tenants is approximately $12,000 (7), whereas most definitions of low SES in previous studies include notably higher incomes. Cigarette consumption could be lower among subsidized housing tenants because of limited financial resources or higher because of financial and other stressors. Tenants may be less interested in quitting because of the high prevalence of smoking or more interested because of high financial stress (8).
Evidence-based treatments such as behavioral counseling and nicotine replacement therapy (NRT) double a smoker's chance of quitting successfully and are recommended for all smokers, including those who are low SES (9). However, low-SES smokers are less likely to use NRT or other cessation aids compared with their higher-SES counterparts (10). Barriers to use include lack of knowledge about the effectiveness of such aids and limited availability (11). Although low -SES smokers who visit health care providers are as likely as other smokers to receive advice about quitting from health care providers (10), they are less likely to receive assistance with quitting (12).
Our study examined smoking behaviors and cessation-related interests among smokers in subsidized MUH collected as part of a larger survey about voluntary home smoking restrictions and support for smoke-free housing policies. There were 3 primary aims: 1) compare sociodemographic and health characteristics between smokers and nonsmokers; 2) describe smoking behaviors and cessation-related interests among smokers; and 3) identify differences in sociodemographic, smoking, and cessation-related characteristics between smokers with different levels of cigarette consumption. The third aim was identified based on preliminary analyses that showed that the survey population had a large proportion of nondaily and light smokers.

Sample
The study population was tenants in approximately 1,000 subsidized MUH units in 184 buildings across 5 urban neighborhoods and managed by a private company in Columbus, Ohio. No buildings or units were covered by a smokefree housing policy. By using units that were occupied as of July 2011 (n = 914), a stratified random sample (n = 475) was selected, and the primary lease holder in each unit was eligible to participate in the study. Administrative data provided by the property management company were used to stratify units by the age of the youngest child (<5 years; 5 -17 years; or no children <18 years) because this factor was expected to be associated with voluntary home smoking restrictions and support for smoke-free policies, which were the primary outcomes in the larger study. Units confirmed as vacant by the property management company at the time they were fielded were ineligible (n = 3).

Data collection
An interviewer-administered, face-to-face survey was conducted at tenants' homes from August to October 2011. A personalized letter was sent to lease holders in selected units 1 week before the first in-person visit. Teams of 2 interviewers (1 community resident and 1 graduate student) made at least 5 in-person attempts to contact each lease holder at different days and times. Visits took an average of 27.0 minutes to complete and participants were given a $5 grocery store gift card. The study was approved by the university's institutional review board, and participants provided informed consent.

Measures
All measures were self-reported. Sociodemographic characteristics collected were age, race/ethnicity, sex, age of youngest child under 18 years living in household, presence of a child with asthma diagnosis, educational attainment, and employment status. Health-related characteristics collected were health insurance status, type of insurance (Medicaid, Medicare, or private), general health status, and physical limitations during the past month (not at all or very little vs somewhat or quite a lot). Food insecurity was measured by using a previously validated 2-item scale (13); those who reported at least 1 of the following statements was true often or sometimes in the past 12 months were classified as at risk for food insecurity: "We worried whether our food would run out before we got money to buy more," and "The food we bought just didn't last and we didn't have money to get more." All participants were asked if they had smoked at least 100 cigarettes in their lifetime and how many days per week they smoke cigarettes now. Interviewers instructed respondents to include small cigars in response to these questions because there was not a separate question about small cigar use. Never smokers had never smoked 100 cigarettes and smoked zero days per week now. Former smokers had smoked 100 cigarettes in their lifetime but smoked zero days now. Current smokers smoked 1 to 7 days per week now regardless of whether they had smoked 100 cigarettes in their lifetime; this definition is appropriate for measuring lower levels of cigarette consumption that are common among young adult and nonwhite smokers (14).
Current smokers also reported how many cigarettes they usually smoke on days they smoke. Cigarette consumption levels were classified as nondaily (1-6 days/week), light daily (1-5 cigarettes per day [CPD]), 7 d/wk), and nonlight daily (≥6 CPD, 7 d/wk). Categorical responses for time to first cigarette after waking (TTF) were dichotomized into 30 minutes or less and more than 30 minutes. Strength of urges to smoke in the past 24 hours (15) was dichotomized as none, slight, or moderate versus strong or very strong. Current smokers were asked, "Is your usual cigarette brand menthol or nonmenthol?" and "How often in the past month have you bought cigarettes loose out of the pack?" (16).
Based on the 2-question stage of change measure (17), smokers who intended to quit within 6 months or 30 days were grouped versus those who did not intend to quit. Although intending to quit within 30 days is a better predictor of short-term cessation than within 6 months, both predict future quit attempts (17), and there were too few respondents who intended to quit within 30 days to conduct separate analyses for this group. Current smokers were also asked, "If you decided to quit smoking, how interested would you be in using each of the following to help you quit if they were free?" Methods were nicotine patches, gum, or lozenges; other medication like Chantix; face-to-face counseling program; telephone counseling program (also called a quitline); and Internet counseling program. Response options included not at all, very, or somewhat interested. Respondents with health insurance were asked whether their insurance would pay for medications, face-to-face programs, or telephone programs to help people quit smoking.

Data analysis
Because our study focused on current smokers, we grouped never and former smokers as nonsmokers. We compared sociodemographic and health characteristics between current and nonsmokers and across cigarette consumption levels. Chi-square tests of association with Rao-Scott adjustment for the stratified survey design were used for categorical variables and Kruskal-Wallis tests for continuous variables. Survey logistic models with adjustment for stratification were used for most multivariable analyses. Exact logistic regression models that also controlled for survey design strata were used for analyses comparing smoking-related behaviors and characteristics across consumption levels due to zero or small sample sizes. Analyses related to awareness of insurance benefits were limited to respondents who reported Medicaid coverage because cessation-related benefits are not consistent across other insurance types. Analyses were conducted with SAS 9.2 (SAS Institute Inc, Cary, North Carolina).

Results
Completed surveys were obtained from lease holders in 301 units (63.8% response rate). Analysis of administrative data from the property management company showed that nonrespondents did not differ from respondents in lease holder age, age of youngest child, or neighborhood of residence.

Characteristics of smokers and nonsmokers
Overall, 47.5% (n = 143) of respondents were current smokers, 12.3% (n = 37) were former smokers, and 40.2% were never smokers (n = 121); therefore, 52.5% were nonsmokers. Fifteen current smokers had not smoked 100 cigarettes in their lifetime. Smokers were less likely than nonsmokers to have at least a high school education, be employed, or have health insurance, and were more likely to have physical limitations or be at risk of food insecurity (Table 1). After controlling for these differences, smokers were less likely to have health insurance (adjusted odds ratio [AOR], 0.45; 95% confidence interval [CI], 0.21-1.00) and more likely to be at risk of food insecurity (AOR, 1.73; 95% CI, 1.07-2.81).

Smoking-related behaviors
Among smokers, 20.3% were nondaily, 35.0% were light daily, and the remaining 44.8% were nonlight daily. Only 14.7% of all smokers consumed more than 10 CPD. Median age was the only sociodemographic or health characteristic that differed significantly by consumption level (P = .002); age was lower for nondaily (23.2 years) and light daily (25.0 years) than for nonlight daily smokers (31.5 years). After controlling for age -the only demographic variable that differed by consumption level -the odds of smoking within 30 minutes of waking and having strong urges to smoke remained significantly higher for nonlight daily versus light daily smokers (AOR, 4.5; 95% CI, 1.8-11.7, and AOR, 4.2; 95% CI, 1.6-12.3, respectively). The odds of smoking within 30 minutes of waking were lower for nondaily smokers compared with light daily smokers (AOR, 0.1; 95% CI, 0.0-0.5). Odds of strong urges are not reported for nondaily smokers because of zero cell size. More than half of smokers had ever purchased single cigarettes in the past month (Table 2), including 27.5% who purchased singles frequently (ie, many times or just about all the time). Nondaily smokers were more likely to have purchased singles in the past month compared with nonlight daily smokers after controlling for age (AOR, 5.6; 95% CI, 1.6-25.6).

Cessation intentions and interests
More than half of smokers intended to quit within the next 6 months or less (Table 2), including 25.9% within the next month and an additional 34.3% within 6 months. Intentions did not differ significantly by consumption level. Smokers were most interested in using NRT as a cessation aid if they decided to quit smoking ( Table 2). Interest in using cessation aids increased with consumption levels but differences were not significant ( Table 2).

Awareness of Medicaid coverage
More than three-fourths (78.7%) of respondents reported having Medicaid insurance, which covers all cessation medications but no counseling (face-to-face or telephone) in Ohio. At the time of the survey, low-income smokers could receive some telephone counseling through the Ohio Tobacco Quitline supported by the state. Of respondents with Medicaid, less than one-third (30.4%) stated that Medicaid would pay for cessation medications. Lower percentages reported that Medicaid would pay for face-to-face (19.4%) or telephone (17.3%) counseling. There were no differences in knowledge of medication coverage between smokers and nonsmokers or by quit intentions, consumption level, or interest in using NRT or other medications among smokers.

Discussion
A better understanding of smoking behaviors and cessation interests among low-SES smokers in subsidized MUH will help inform cessation and other interventions to accompany smoke-free housing policies. Strengths of this study included a probability sample and high response rate. Because they qualified for subsidized housing, all respondents had very low incomes; additionally, they were predominantly African American and female. Our findings were similar to national estimates for African Americans (18) in that almost one-fourth of smokers did not smoke every day. However, light daily smoking was 3 times more common than among African American smokers in general (18). Light smoking is a concern because it causes similar risks of cardiovascular disease as heavier smoking and higher risks of lung and other cancers, respiratory disease, and all-cause mortality than nonsmoking (19).
Smokers reported notably higher rates of quit intentions than other low-SES populations (5), possibly because of high rates of financial stress (almost 60% reported food insecurity in the past year) (8) or nondaily and light daily smoking (20). Although intentions to quit have not been associated with long-term cessation, they are associated with shortterm cessation among smokers with previous quit attempts and with increased quit attempts (17). Intentions to quit may also reflect receptivity to smoking cessation supports (17). However, because both low-SES and African American smokers have lower cessation rates than other smokers (5,10), interventions are needed to transform high intentions into successful cessation.
Few studies have examined the effectiveness of evidence-based cessation treatment among very low-SES, light, or nondaily smokers. Guidelines recommend evidence-based cessation aids for low-SES smokers (9) but some cessation aids have been less effective among low-SES (21,22) and light (23) smokers. A community-based intervention with women living in public housing that included nicotine patches and counseling was effective at 6 months (24). However, only daily smokers were eligible for the study and the intervention would likely be cost-prohibitive to implement across a large number of subsidized housing settings. More research is needed to design scalable, culturally appropriate strategies to address a range of consumption levels and contextual realities (eg, food insecurity, availability of single cigarettes) of very low-SES smokers.
The sale of single cigarettes is illegal, and observational data from tobacco retail outlets do not typically reveal single cigarette sales (25). However, the high rate of self-reported single cigarette purchasing was consistent with a previous study of young urban African American adults in Baltimore (16). Smokers purchase single cigarettes for different reasons, including convenience, low immediate costs, attempts to quit or cut back, and urges to smoke when they see singles for sale (16,26). More research is needed to determine whether single cigarettes serve to maintain or treat dependence among very low-SES populations, and whether singles are primarily purchased from commercial venues or other individuals.
For subsidized housing tenants who have high levels of nicotine dependence, Medicaid ensures access to cessation medications. Although awareness rates were similar to other surveys of smokers who have Medicaid insurance (11,27,28), 70% of respondents were not aware of this coverage; we found no differences by smoking status, intentions to quit, or intentions to use cessation medications. Several previous studies have evaluated interventions to increase awareness and use of insurance coverage for treatment of tobacco dependence. The most promising was a health maintenance organization-based communication campaign that increased use of cessation medications and the telephone quitline among Medicaid smokers in Wisconsin compared with fee-for-service Medicaid enrollees without a campaign (29). The success of this intervention highlights the importance of involving health care providers, many of whom are also unaware of Medicaid cessation benefits (28), and using integrated, multicomponent strategies. Campaigns to increase awareness and use of Medicaid cessation coverage among MUH tenants could focus on settings where most tenants receive health care services as well as housing offices, tenant newsletters or mailings, and resident support services. Additional barriers, such as perceived effectiveness of medications, must be addressed in parallel with awareness of benefits (11).

Limitations
Smokers may have been less likely to respond than nonsmokers, and self-reported smoking status was not biochemically verified. However, this means that the reported smoking rate, which was much higher than in the general population, is conservative. Sample sizes were small for nondaily smokers, which limited power for comparisons across consumption levels. Generalizability to all subsidized housing populations may be limited. In particular, age was associated with consumption levels, and older tenants were underrepresented in the study population compared with national estimates for subsidized housing tenants. However, a large minority of all subsidized housing tenants nationally are African American and have children, and a majority are female (7), making them comparable to our study population.

Conclusions
Smoking in general, and light smoking specifically, was much more common in this subsidized housing population compared with the general population. Medicaid coverage for cessation medications should be promoted to tenants and health care providers because most daily smokers were interested in using NRT but most did not know Medicaid covers these medications. Policies and interventions to address environmental factors such as availability of single cigarettes should be considered in parallel with smoke-free housing policies, especially because few tobacco dependence treatments have been tested or shown to be effective with very low-SES or light smokers. Denominators for percentages are not always the total n for the group because of missing data. Respondents who reported that at least 1of 2 statements about food insecurity were true often or sometimes in the past 12 months (13) were classified as at risk for food insecurity. Abbreviation: TTF, time to first cigarette after waking. Nondaily = smoke on 1 to 6 d/wk; light daily = 1 to 5 cigarettes per day, 7 days per week; nonlight daily = 6 or more cigarettes per day, 7 days per week.

Tables
Statistical tests were χ tests of association (with Rao-Scott adjustment for survey design). Denominators for percentages are not always the total n for the group because of missing data. Nondaily smokers were excluded from comparisons across consumption levels because of small or zero cell size. "Don't know" (n = 19) and refused (n = 1) responses were excluded; "don't know" responses were primarily from respondents who only smoked small cigars.
Defined as "ever" (ie, a few times, many times, or just about all of the time) bought single cigarettes loose out of the pack in the past month (versus "never").
"Interested" was defined as "very" or "somewhat" versus "not at all."