The Relationship Between Neighborhood Disorder and Barriers to Cessation in a Sample of Impoverished Inner-City Smokers in Baltimore, Maryland, United States

Abstract

Introduction

Economic disparities in rates of smoking have been well documented in many countries. These disparities exist on an individual and geographic or neighborhood level. This cross-sectional study examined the relationship between neighborhood physical and social disorder and barriers to smoking cessation among an impoverished urban sample.

Methods

A sample of current smokers were recruited through street outreach, posted advertisements, and word of mouth from impoverished neighborhoods in Baltimore, Maryland, USA for a study of psychosocial factors and smoking behaviors. Neighborhood disorder was assessed with a 10-item scale from the Block Environmental Inventory and barriers to cessation with a 9-item scale.

Results

In the multiple logistic regression model, perceived stress (aOR = 1.60, 95% CI = 1.32 to 1.95), neighborhood disorder (aOR= 1.34, 95% CI = 1.11 to 1.63), and level of nicotine dependence (aOR = 1.97), 95% CI = 1.62 to 2.40) were all strongly associated with barriers to cessation.

Conclusion

The results of this study suggest that neighborhood disorder may lead to barriers to cessation among low-income populations. The findings also indicate that tobacco control interventions should examine and address social and physical aspects of impoverished neighborhoods.

Implications

In many countries, tobacco control programs and policies have been less effective among low-income populations as compared to more affluent populations. Little is known about how neighborhood factors influence smoking cessation. This study examined the relationship between neighborhood disorder and barriers to cessation among a low-income population. We recruited a convenience sample of hard-to-reach cigarette smokers from low-income neighborhoods. Even after controlling for level of nicotine dependence and stress, neighborhood disorder was found to be associated with barriers to cessation. The findings suggest the important role of neighborhood disorder as a barrier to smoking cessation.

Introduction

Although tobacco control programs have been highly successful in more affluent communities in the United States and other countries, huge disparities exist in rates of smoking.¹ In some low-income communities in the United States, smoking remains the norm with smokers constituting more than half the adult population.²

There is a range of explanations as to why lower-income individuals and communities have higher rates of smoking and lower quit rates.³ These explanations for social class differences include social norms of greater acceptability and higher frequency of smoking^4–6; greater availability,^4,7 such as the buying and selling of single cigarettes⁸; coping with social and physical stressors^9,10 such as poor housing conditions, violence, and unemployment leading to psychological distress and depression,^11,12 which lowers motivation and ability to quit; tobacco advertisers targeting low-income neighborhoods^13,14; tobacco prevention programs not targeting or tailored to low-income communities^15,16; and lack of social support for cessation.¹⁷ Rates of violence,¹⁸ which are higher in some low-income urban neighborhoods, may also impede cessation^19,20; and, for individuals with limited access to many life opportunities, smoking can be a source of pleasure.^21,22

Low-income urban neighborhoods in the US often suffer from high rates of physical and social disorder.^23,24 Neighborhood disorder theory links physical and social neighborhood factors within a neighborhood to health consequences.^25,26 Those who live in neighborhoods with high levels of disorder and whose economic circumstances do not permit mobility may feel entrapped.²⁷ From chronic stressors and perceived entrapment, neighborhood disorder can contribute to psychological distress.^28–30

Neighborhood-level stress and perceived safety has been linked to smoking.^31,32 In analyses of a smoking cessation program, neighborhood social and physical disorder were found to be associated with a lower likelihood of achieving abstinence at 4 weeks,³³ and greater neighborhood-level poverty and unemployment were also correlated with reduced odds of abstinence.³⁴

Perceived barriers for smoking cessation refers to perceptions of factors that interfere with individuals’ ability to quit smoking. Measures of perceived barriers for smoking cessation have been found to successfully predict 1-month abstinence and 3-month smoking frequency.³⁵ Such measures have also been found to be associated with nicotine dependence, number of cigarettes smoked per day,^36,37 negative affect,³⁸ and smoking outcome expectancies.^38–40

There are several pathways that may link neighborhood disorder to barriers to cessation. High rates of public smoking may make it appear normative and hence going against the prevailing norms may be difficult, and if more people publically smoke in these neighborhoods there may be less encouragement to quit.^4,5,7 There may be few options for pleasurable activities in the neighborhood, hence smoking as seen as one of the few pleasurable activities.⁴¹ These neighborhoods may have more advertisements in stores for cigarettes, which may cue smoking behaviors.^42,43 It is also possible that smoking provides a sense of control and means of coping in a chaotic environment and cessation would remove this coping mechanism.^9,10,44 Few empirical studies have examined how neighborhood factors may lead to barriers to cessation. In a qualitative study of residents of disadvantaged areas in Adelaide, Australia, stress due to the social environment was perceived as a significant barrier to cessation.⁴⁵

In the present study, we recruited current tobacco smokers from low-income neighborhoods in Baltimore, Maryland, United States for a study on psychosocial factors and tobacco use. We anticipated that neighborhood disorder would be associated with perceived barriers to smoking cessation. We also examined the potential confounders of perceived stress, which has been found to be associated with quit attempts,^21,46 as well as level of nicotine dependence,^47–49 which has been found to be associated with both quit behaviors and social economic status.

Methods

Data for the analyses were from the Tobacco Use in Drug Environment (TIDE) study in Baltimore, MD from September 2013 to May 2015. Study participants were recruited through street outreach, posted advertisements, and by word of mouth. Participants, who were first verbally screened by telephone or face-to-face, were eligible to participate if they were (1) 18 years or older and (2) a current smoker, which was defined as smoking more than 100 cigarettes in their lifetime and smoking in the last week. Eligible participants provided written informed consent before the start of a face-to-face interview. A trained staff member administered the survey, and sensitive information was collected via audio computer-assisted self-interviewing (ACASI). NicCheck® I Test Strips were used to test for nicotine and metabolites.⁵⁰ The study was approved by the Johns Hopkins Bloomberg School of Public Health Institutional Review Board.

Measures

The Outcome of Barriers to Cessation

Based on the 19 items from the Barriers to Cessation Scale.⁵¹ The barriers were rated on a 4-point scale of “not a barrier,” “a small barrier,” “a medium barrier,” and “a large barrier.” We removed nine items that did not apply to the whole sample, such as items that referred to a partner, or were not appropriate to a highly impoverished sample, such as work. Some of the items were also modified for clarity (Appendix A). A factor analyses revealed that 9 of the 10 items loaded highly on one factor accounting for 44% of the variance. The 10th item, on gaining weight as a barrier, was removed. The scale’s Cronbach’s alpha was .853. Two items had significant missing data (18% and 5%), and linear interpolation was used to replace missing values. The scale range was 0–27 and the median was 11.

Nicotine Dependence

Level of nicotine dependence was computed from the 9-item measure in the PATH study.⁵² The items, such as “I usually want to smoke a cigarette right after I wake up” had a 5-point Likert scale that ranged from “Not true of me at all” to “Extremely true of me.” One item had substantial missing data (3%), and linear interpolation was used to replace missing values. We conducted a factor analysis on the tobacco dependence scale, and it revealed one strong factor which accounted for 46.6% of the variance. The Cronbach’s alpha was .856. The range was 9–45 and the median was 33. Two items from Fagerstrom Test for Nicotine Dependence were also included in the survey.

Neighborhood Disorder

Neighborhood disorder was assessed with a 10-item, three point scale, based on Perkins and Taylor’s validated Block Environmental Inventory.⁵³ Participants were asked whether issues such as vacant housing, trash on the streets, and groups of teenagers hanging out on the street, were “not a problem,” “somewhat of a problem,” or “a big problem” on their block. The Cronbach’s alpha was .907. The range was 0–16 and the median was 3.

Perceived Stress

Stress was assessed with the 4-item Perceived Stress Scale that included items such as “During the past 30 days, how often have you felt that you were unable to control the important things in your life?”⁵⁴ The response categories were “never,” “almost never,” “sometimes,” “fairly often,” and “very often.” The Cronbach’s alpha was .656. The range was 0–16 and the median was 5.

Demographic and Substance Use Measures

Demographic and substance use measures included age, sex at birth, race/ethnicity, level of education (grade 11 or less versus grade 12 or higher), having a main partner, homelessness in the prior 6 months, cocaine or heroin use in the prior 6 months, and current employment status (unemployed vs. employed).

Analysis

Barriers to Cessation were dichotomized based on a median split, and chi-square and t-tests were used to examine associations between high and low levels of barriers to cessation. The three variables that were scales (perceived stress, level of nicotine dependence, and neighborhood disorder) were transformed into z-scores for ease of interpretation. Pearson correlations were used to examine the correlations between the independent variables. The data were modeled using OLS multiple regression, treating barriers to cessation as a continuous variable. Logistic regression was used to assess the independent association of level of neighborhood disorder with the outcome of barriers to cession, adjusting for potential confounders, including perceived stress and nicotine dependence, using a median split to construct a dichotomous variable. The multiple logistic regression analyses included demographic variables as well as those variables with p values <.15 in the bivariate analyses. Out of the 593 respondents, the analyses were restricted to 580 cases. The perceived stress scale and neighborhood disorder measure had one missing value each and barriers to cessation had 11 missing values. SPSS 24 and Stata 14 were used for the analyses.

Results

The sample was predominantly male (62.8%) with a mean age of 47 years and the median age of 49 years. More than half (56.9%) had graduated from high school, and most (83.3%) were unemployed, African-American (89.7%), and single (53.8%) and 31.4% were either married or in a committed relationship. In examining the Barriers to Cessation scale, the individual items that were endorsed most often as “a large barrier” were withdrawal symptoms (33.4%), seeing other people smoke (33.1%), and thinking that you will start smoking again (27.2%). Family members’ encouragement to smoke was reported as “not a barrier” by almost half (45.5%) the sample, as was friends’ encouragement to smoke (55.5%). As seen in Tables 1 and 2, the level of perceived barriers to cessation was not associated with level of education, employment status, sex at birth, or age. Neighborhood disorder, level of nicotine dependence and perceived stress were all strongly and positively associated (p < .001) with levels of barriers to cessation. Neighborhood disorder was correlated with level of dependence (.131, p < .01) and stress (.092, p < .05)

Table 1.

Chi-Square Analyses of the Association Between Participant Characteristics and Barriers to Smoking Cessation

	Barriers to cessation < 11		Barriers to cessation ≥ 11		Total		χ2	p
	n	%	n	%	n	%
Sex at birth
Male	188	66.0	176	59.7	364	62.8	2.465	.116
Female	97	34.0	119	40.3	216	37.2
Education
Grade 11 or less	126	44.2	124	42.0	250	43.1	0.280	.597
Grade 12 or higher	159	55.8	171	58.0	330	56.9
Employment status
Employed	52	18.2	45	15.3	97	16.7	0.931	.335
Unemployed	233	81.8	250	84.7	483	83.3
Main sex partner
Yes	150	52.6	146	49.5	299	51.6	0.262	.609
No	135	47.4	149	50.5	281	48.4
Homeless
Yes	38	13.3	48	16.3	86	14.8	0.991	.320
No	247	86.7	247	83.7	494	85.2
Heroin or cocaine use (past 6 months)
Yes	65	22.9	80	27.1	145	25.0	1.380	.240
No	219	77.1	215	72.9	434	75.0

Table 2.

t-Test Analyses of the Association Between Participant Characteristics and Barriers to Smoking Cessation

	n	Barriers to cessation < 11		Barriers to cessation ≥ 11		t	p
		Mean	SD	Mean	SD
Age	580	47.74	10.39	46.52	9.42	1.476	.141
Perceived stress	580	−0.28	0.93	0.27	0.99	−6.881	<.001
Neighborhood disorder	580	−0.23	0.91	0.19	1.02	−5.195	<.001
Level of nicotine dependence	580	−0.33	1.04	0.33	0.85	−8.408	<.001

In the OLS regression model, perceived stress, neighborhood disorder, and level of nicotine dependence were all strongly associated with barriers to cessation (Table 3). Level of education was marginally associated with barriers to cessation. Similar to the OLS regression model, in the multiple logistic regression model, perceived stress, neighborhood disorder, and level of nicotine dependence were all strongly associated with barriers to cessation (Table 4). Level of education was positively and significantly associated with barriers to cessation in the logistic model. In the logistic regression analyses when perceived stress and level of nicotine dependence were not included in the first block of variables but added to the model in the second block as potential confounders, the association between neighborhood disorder and perceived barriers to cessation changed by only a small amount (from OR = 1.48 to OR = 1.34). We also examined a logistic regression model that included two items from Fagerstrom Test for Nicotine⁵⁵ as a measure of nicotine dependence, rather than the PATH measure of nicotine dependence, and found highly similar associations of statistical significance as compared to when we included the PATH measure of level of nicotine dependence.

Table 3.

Ordinary Least Squares (OLS) Multiple Regression Models of Association Between Participant Characteristics and Barriers to Smoking Cessation

	β coefficient	SE	p value
Age	−0.018	0.026	.492
Sex at birth
Male	Ref.	—	—
Female	0.586	0.530	.270
Education
Grade 11 or less	Ref.	—	—
Grade 12 or higher	1.027	0.523	.050
Employment
Employed	Ref.	—	—
Unemployed	0.537	0.684	.433
Perceived stress	1.547	0.266	<.001
Neighborhood disorder	0.870	0.267	.001
Level of nicotine dependence	2.223	0.260	<.001

Table 4.

Multiple Logistic Regression Model of the Association Between Participant Characteristics and Barriers to Smoking Cessation

	Odds ratio	95% CI	p value
Age	0.994	(0.976, 1.103)	.550
Sex at birth
Male	Ref.	—	—
Female	1.339	(0.915, 1.958)	.133
Education
Grade 11 or less	Ref.	—	—
Grade 12 or higher	1.540	(1.055, 2.246)	.025
Employment
Employed	Ref.	—	—
Unemployed	1.072	(0.658, 1.745)	.780
Perceived stress	1.603	(1.320, 1.947)	<.001
Neighborhood disorder	1.343	(1.109, 1.626)	.003
Level of nicotine dependence	1.973	(1.621, 2.401)	<.001

Discussion

The study findings suggest that neighborhood disorder was associated with perceived barriers to cessation among urban inner-city residents. This association continued to be statistically significant even after adjusting for perceived stress and level of nicotine dependence. As smoking is normative in many inner-city neighborhoods, there may be less support for cessation, especially in neighborhoods that have high levels of disorder. In such areas, smoking cessation may not be seen as a high priority, as compared to more pressing needs such as avoiding violence and acquiring resources to survive.³ For individuals who may also perceive that they have little control over their social and physical environments, smoking may provide a small amount of perceived control or pleasure.⁵⁶

Several study limitations should be noted. The study was a convenience sample and the majority of data was self-reported, though tobacco use was verified with nicotine assays, and did not include objective measures of neighborhood disorder. Moreover, the study recruited in highly impoverished urban areas; hence, the findings may not generalize to more affluent and less urban areas. This was not a random sample and there may have been unaccounted clustering by neighborhoods. It is also possible that stress plays a larger role in the relationship between neighborhood disorder and barriers to cessation, but that our measure of perceived stress did not measure these factors. We also did not measure other neighborhood factors, such as social cohesion or collective efficacy, or have measure of objective neighborhood factors such as poverty, income, crime, and abandoned buildings that may also influence smoking behaviors. It is also likely that neighborhood factors have a cumulative impact on smoking behaviors and this was a cross-sectional study. The barriers to cessation scale only included nine items from a single factor. It is highly likely that there are other important barriers that we did not measure. More information is also needed on the relative importance of perceived barriers to cessation as compared to other social and environmental factors that promote cessation. In piloting the barriers to cessation items, some individuals reported more than one partner. For future surveys of barriers to cessation it may be prudent to phrase items regarding main partners as a “spouse or main partner.” The Barriers to Cessation scale also asks about encouragement to smoke from family and friends. For this population, it also may be useful to ask questions about smokers and nonsmokers in the same household who may not be considered family or friends.

We also found that higher levels of nicotine dependence were also associated with perceived barriers to cessation, which highlights that drug dependence is linked to perceived social-structural factors as well as an individual-level variables. However, even when nicotine dependence was in the model, neighborhood disorder was still associated with perceived barriers to cessation. These findings suggest that strictly individual-level interventions for cessation may not be as effective in neighborhoods with high levels of disorder. To develop effective interventions for neighborhoods with high levels of disorder, tobacco control researchers and practitioners need to address neighborhood factors. Household and community-based interventions, as well as interventions that address poverty as the fundamental cause of poor health, are needed. Moreover, interventions that do address poverty should assess how such programs may reduce the barriers to cessation and increase both quit attempts and actual cessation. In conclusion, the study results suggest that neighborhood disorder is associated with perceived barriers to cessation in an impoverished community. Future research should both examine the mechanisms as well as test smoking cessation programs that take into consideration neighborhood social disorder.

Funding

This research was supported by FDA and NIH grant support, Award # 5R01DA032217-04S1 and Research reported in this publication was supported by NIDA/NIH and FDA Center for Tobacco Products (CTP).

Declaration of Interests

None declared.