Tobacco Smoke Incursions in Multiunit Housing

Abstract

Objectives. We sought to describe the prevalence of secondhand tobacco smoke incursions reported by multiunit housing (MUH) residents, pinpoint factors associated with exposure, and determine whether smoke-free building policy was associated with prevalence of reported tobacco smoke incursions.

Methods. Data are from a 2011 nationally representative dual-frame survey (random-digit-dial and Internet panels) of US adults aged 18 years and older. Individuals who lived in MUH and who reported no smoking in their homes for the past 3 months, whether or not they reported being smokers themselves, were included in this study. Incursions were defined as smelling tobacco smoke in their building or unit.

Results. Of 562 respondents, 29.5% reported smoke incursions in their buildings. Of these, 16% reported incursions in their own unit, 36.2% of which occurred at least weekly. Government-subsidized housing and partial smoke-free policies were associated with a higher likelihood of reporting smoke incursions.

Conclusions. Many residents of multiunit housing are exposed to tobacco smoke in their units and buildings. Partial smoke-free policies do not appear to protect residents and might increase the likelihood of incursions in residents’ individual units.

The United States Surgeon General has extensively documented the harmful effects of smoking tobacco among adults¹ and the negative effects of secondhand tobacco smoke (SHS) on both adults and children.² Even brief exposures to SHS can result in sustained vascular injury³ and changes in endothelial function.⁴ In addition, studies have shown that very low levels of SHS exposure are associated with cognitive deficits⁵ and decreased antioxidant levels⁶ in children. Young children breathe faster than adults,⁷ have a tendency to mouth surfaces and objects,⁸ and may spend more of their time in the home,⁹ especially in places where outdoor activities might be perceived as unsafe.¹⁰ These all may increase their potential for tobacco smoke inhalation and ingestion.¹¹ In a recent US Centers for Disease Control and Prevention study, 54% of children aged 3 to 11 years had biological evidence of SHS exposure; however, only 18% had a household member who smoked in the home,¹² suggesting the significant contribution of other sources of tobacco smoke.

Approximately 25% of US residents live in multiunit housing,¹³ where air circulation patterns facilitate the spread of tobacco smoke from unit to unit.¹⁴ Tobacco smoke permeates housing complexes through air ducts, cracks in the floor and walls, stairwells, hallways, elevator shafts, electrical lines, and open windows.¹⁵ US residents spend about 69% of their time in private residences¹⁶; thus, drifting smoke from other residences may provide a significant source of exposure for those in nonsmoking homes. A study of real-time SHS transfer in multiunit housing (MUH) between smoke-free and smoke-permitting units demonstrated incursions into both smoke-free units and adjacent hallways¹⁵; another study of low-income MUH in Boston, Massachusetts, demonstrated SHS contamination in homes where residents reported that neither household members nor visitors smoked.¹⁷ In an analysis of national data, cotinine levels for children living in apartments were 45% higher than for those living in detached homes.¹⁸ These studies suggest a significant role for SHS exposure in MUH. Recently, studies have begun to examine the prevalence of incursions in MUH. The first major national study discovered that 29% of US multiunit housing residents lived in a smoke-free building, and of those who did not but had voluntary smoke-free home rules, 44% reported incursions in their unit.¹⁹ In New York State, 46.2% of residents of MUH reported experiencing an incursion in their home, and more than 9% experienced incursions daily.²⁰ Other local studies have shown similar rates of SHS incursions in MUH.^21,22

In 2012, the US Department of Housing and Urban Development (HUD) reissued recommendations that strongly encouraged all of their housing units go smoke-free²³; however, this is not yet a requirement, and policies regulating SHS exposure in personal living spaces remain limited.²² Estimates show that although most residents of MUH would prefer to live in a smoke-free environment,^20,22 most MUH developments do not have comprehensive smoke-free policies.²² We reported on a national survey of MUH residents to examine factors associated with tobacco smoke incursions. We hypothesized that type of smoke-free building policy (comprehensive, partial, or none) would be associated with prevalence of reported SHS incursions.

METHODS

Our data were from the 2011 Social Climate Survey of Tobacco Control, an annual nationally representative, cross-sectional survey developed by tobacco-control researchers to study behaviors and trends related to tobacco use and exposure in the United States. The final sample comprised 2 distinct sampling frames of US adults aged 18 years and older: a random-digit-dial (RDD) survey of landline telephone numbers, and the KnowledgeNetworks (KN) probability-based Internet panel.^24,25 The KN panel is a nationally representative Internet panel based on a sampling frame that includes both listed and unlisted landline numbers, as well as panelists without a landline telephone. The panel does not accept self-selected volunteers²⁴ and provides sample coverage for 97% of US households.²⁶

The Social Climate Survey of Tobacco Control has used a dual-frame methodology since 2010 to account for the high rate of cell phone substitution among smokers and young men,²⁷ both being important groups to sample adequately in tobacco research. Details of the dual-frame sampling strategy, and comparisons to other national surveys, have been published elsewhere.²⁸ The sample was weighted to adjust for age, race, gender, and US Census region, based on current US estimates, as well as for the frame overlap among Internet panel respondents who also had a landline telephone and were therefore also eligible for the RDD frame. Weights were capped so as not to exceed 5 to prevent an underrepresented group with a high weight from being overly influential. Both the RDD and KN panel data were collected from October to November 2011.

Measures

All measures in this study except for US Census region were based on self-report. The study sample was made up of MUH residents who reported no smoking in their home for the past 3 months. Multiunit housing was defined by answering,“A one-family house attached to one or more houses” or “An apartment or condominium building” to the question “Which of the following best describes where you live?” No smoking in the home in at least the past 3 months was defined by a negative response to the question “Over the past 3 months, has anyone smoked anywhere in your home?” Multiunit housing building and unit smoke incursions represented the 2 outcome variables with “Yes,” “No,” or “Don’t know/Refused” responses. Participants who reported living in MUH were asked “Do you ever smell tobacco smoke in your multiunit building?” Those who responded “Yes” to smoke incursions in their MUH building were then asked, “Do you ever smell tobacco smoke inside your unit?” Respondents reporting “Yes” to the building or unit smoke incursion questions were also asked about the frequency of smelling smoke with “Daily,” “Weekly,” “Monthly,” “Rarely,” “Never,” “Don’t know,” and “Refused” response options.

To assess MUH smoking rules, respondents were asked “Which of the following best describes your apartment or attached housing?” with the following response options: “Smoking is allowed on the property in shared areas and in the apartments or units,” “Smoking is allowed only in apartments or unit,” “Smoking is not allowed in apartments or units, but is allowed on the property,” and “Smoking is not allowed at all, not even in apartments or units”; “Do not know” was also permitted as a response. Demographic variables including categorized race, highest educational level attained, 2010 before-tax household income, age, and US Census region. US Census region was derived from the landline telephone number in the RDD sample frame and by respondent state in the KN frame. Participants were also asked about subsidized housing with the following question: “Do you receive any government assistance to pay for your rent or housing?” They were also asked “How many children under 18 years of age currently live in your household?” Respondents who reported having smoked at least 100 cigarettes in their entire life and currently smoke some days or everyday were categorized as current cigarette smokers. The sampling frame indicates whether the respondent participated through the RDD survey of landline telephone numbers or the KN probability-based Internet panel.

Data Analysis

We performed univariable analyses to describe the frequency and proportion of select characteristics (race, gender, age, income, education, geographic region, subsidized housing, children living at home, smoking status, MUH smoking rules, type of MUH housing, and tobacco smoke incursions). The remaining analyses were performed on MUH residents who reported no smoking in their home for at least 3 months. The Rao–Scott χ² test was performed and P values, unadjusted odds ratios (ORs), and 95% confidence intervals (CIs) were reported to assess the bivariable association between reporting smoke incursions versus not reporting smoke incursions and the respondent characteristics. We used 2 multivariable logistic regression models to estimate adjusted ORs and 95% CIs for characteristics associated with (1) reporting building smoke incursions versus not reporting building smoke incursions and (2) reporting unit smoke incursions versus not reporting unit smoke incursions. Characteristics with a P level of less than .2 in the bivariable analyses were entered into the multivariable logistic regression model. For the building incursion model these variables were gender, income, education, government-subsidized housing, children living in the home, smoking status, MUH smoking rules, and type of MUH housing. Because subsidized housing and income were collinear, we excluded income from the model; this did not appreciably change the model estimates. We also included race, age, and US Census region in this model for substantive purposes; this addition did not substantially affect the magnitude of the ORs or the precision of the 95% CIs. For the unit incursion model the following variables were included: educational level, children living in the home, smoking status, MUH smoking rules, and type of MUH. Variables with categories that had similar bivariable associations with the outcome were collapsed in the multivariable analyses (race, income, educational level, and MUH rules) to reduce the number of degrees of freedom. We used the Hosmer–Lemeshow test and c statistic to evaluate model fit. Bivariable and multivariable analyses excluded persons with responses that were missing or recorded as “Don’t know.” P values were 2-sided, and a P level of less than .05 was considered to indicate statistical significance. Except where noted, weighted data are reported.

Individual totals may not add up to category totals because of rounding weighted values. SAS version 9.3 (SAS Institute, Cary, NC) survey procedures were used to account for the complex sampling design (the use of sampling frames and weighted data); these included the Rao–Scott χ² test and design-based variances using a Taylor series expansion approximation for the estimated P values and ORs. SAS version 9.3 was used for all analyses.

RESULTS

In the RDD frame, of 1735 eligible respondents contacted, 1500 (86.46%) completed surveys. For the Internet panel frame, 2476 panelists were randomly drawn from the probability panel; 1597 responded to the invitation, yielding a final stage completion rate of 64.5%.

Univariable Results

The weighted sample size was 3059. Almost one quarter of respondents reported living in MUH (23.9%, n = 731). Of these, more than three quarters reported no smoking in their home for the past 3 months (76.8%, n = 562). The remainder of our analysis is on the subgroup of MUH residents who reported no smoking in their home for the past 3 months (n = 562). Of this group 12.1% were current smokers. The cohort was about half men and half White, with the majority of respondents younger than 45 years; demographic characteristics of the sample are detailed in Table 1. About 30% of the respondents had a high school education or less, and 22% had incomes below $20 000; 7.6% reporting receiving government subsidies for their housing, and 29.1% had a child living at home.

TABLE 1—

Frequency of Select Characteristics: Social Climate Survey of Tobacco Control, United States, 2011

	MUH Residents (n = 731), % (No.)	MUH Residents With No Home Smoking for ≥ 3 Months (n = 562), % (No.)
Race/ethnicity
White	46.5 (340)	47.6 (268)
African American	24.3 (178)	22.8 (128)
Other	29.2 (213)	29.6 (167)
Gender
Male	47.9 (350)	48.4 (272)
Female	52.1 (380)	51.6 (290)
2010 household income (n = 689, 533), $
≤ 20 000	24.0 (165)	22.0 (117)
20 000–75 000	54.3 (374)	54.5 (290)
≥ 75 000	21.8 (150)	23.5 (125)
Age (n = 724, 556), y
18–29	29.1 (211)	31.1 (173)
30–44	33.3 (241)	33.4 (186)
45–59	19.4 (140)	17.9 (100)
≥ 60	18.3 (132)	17.6 (98)
Education
< high school	11.6 (85)	5.6 (32)
High school	25.7 (187)	23.9 (134)
Some college	27.3 (199)	28.4 (159)
≥ bachelor’s degree	35.5 (259)	42.1 (236)
US Census region
Northeast	20.5 (150)	21.1 (119)
Midwest	16.8 (123)	12.6 (71)
South	26.9 (196)	28.8 (162)
West	35.8 (262)	37.6 (211)
Type of MUH
1-family house attached to ≥ 1 house	31.7 (232)	33.1 (186)
Apartment or condominium building	68.3 (499)	66.9 (376)
Government-subsidized housing (n = 674, 519)
Yes	10.2 (69)	7.6 (40)
No	89.8 (605)	92.4 (480)
Children living in home
Yes	29.1 (213)	29.1 (164)
No	70.9 (518)	70.9 (398)
Smoking status (n = 719, 559)
Current smoker	22.3 (160)	12.1 (68)
Not a current smoker	77.7 (558)	87.9 (491)
Sampling frame
Internet	79.9 (584)	79.4 (447)
Random digit dial	20.1 (147)	20.6 (116)

Note. MUH = multiunit housing. The sample size was n = 3059. For each characteristic, n = (number of building incursion responses, number of unit incursion responses).

Almost half of the respondents lived in MUH that allowed smoking, either only in units, or in units and shared areas. Almost one third (28.9%) reported smelling tobacco smoke in their MUH building, with 53.8% of those reporting ever smelling tobacco smoke in their unit. Of those, 39.6% reported smelling tobacco smoke weekly, and 21.0% daily (Table 2). Most (60.3%) of the respondents who smelled SHS in their buildings either smelled it in the hallway or lobby, with almost 20% smelling it daily or weekly. For the building SHS incursion outcome, 2.5% (n = 14) of respondents had missing data, refused to answer, or answered, “Don’t know.” All respondents answered “Yes” or “No” for the unit smoke incursion question.

TABLE 2—

Self Reported Smoking Policies and Incursions Among Multiunit Housing Residents With No Home Smoking for ≥ 3 Months: Social Climate Survey of Tobacco Control, United States, 2011

	Overall, % (No.)
Which best describes your apartment or attached housing?
Smoking is allowed on the property in shared areas and in the apartments or units	34.9 (196)
Smoking is allowed only in apartments or units	13.7 (77)
Smoking is not allowed in apartments or unit, but is allowed on the property	19.0 (107)
Smoking is not allowed at all—even in apartments or units	27.7 (156)
Don’t know smoking rules/refused	4.7 (27)
Do you ever smell tobacco smoke in your multiunit building? (n = 561)
Yes	28.9 (162)
No	68.9 (386)
Don’t know/refused	2.2 (12)
Do you ever smell tobacco smoke in your unit? (n = 162a)
Yes	53.8 (87)
No	46.2 (75)
Don’t know/refused	0
How often do you smell it in your unit? (n = 87b)
Rarely/never	27.6 (24)
Monthly	11.8 (10)
Weekly	39.6 (34)
Daily	21.0 (18)
Do you smell tobacco smoke in the hallway or lobby? (n = 101a)
Yes	60.3 (61)
No	39.7 (40)
Don’t know/refused	0
How often do you smell it in the hallway or lobby? (n = 61c)
Rarely/never	59.9 (37)
Monthly	20.3 (12)
Daily/weekly	19.7 (12)

Note. The total sample size was n = 562.

^aStatistic represents those responding “yes” to smelling smoke in their building.

^bStatistic represents those responding “yes” to smelling smoke in their unit.

^cStatistic represents those responding “yes” to smelling smoke in the hallway or lobby.

Bivariable Results

Building incursions.

We associated household income, living in subsidized housing, having children living in the home, smoking status, type of MUH, and type of MUH smoking rule with building incursions (Table 3). Of the respondents with children living in the home, 40.2% reported building smoke incursions compared with 25.1% of respondents without children in the home (P = .02). Similarly, 55.2% of respondents residing in government-subsidized housing reported smelling smoke in their MUH building compared with 26.1% of respondents who did not live in subsidized housing (P = .01), with building incursions reported most frequently in the lowest household income group. Building SHS incursions varied by type of MUH. Respondents from apartments or condominiums were more likely to report building SHS incursions (38.3%) compared with residents of 1-family houses attached to 1 or more houses (11.4%; P < .001). Compared with MUH where no smoking is allowed anywhere, building SHS incursions were reported more frequently in MUH where smoking is allowed on the property in shared areas and in the apartments or units (36.2%) and where smoking is allowed only in apartments or units (41.6%). The proportion of respondents reporting building incursions who lived in MUH where smoking is not allowed in apartments or units but is allowed on the property (22.7%) is similar to those who live where smoking is not allowed whatsoever (22.5%; P = .1). Building tobacco smoke incursions are not appreciably different when comparing respondents living where smoking is allowed only in apartments or units to MUH residents living where smoking is allowed on the property in shared areas and in the apartments or units (OR = 1.3; 95% CI = 0.6, 2.9).

TABLE 3—

Bivariable Associations Between Self-Reported Smoke Incursions Among Multiunit Housing Residents With No Home Smoking for ≥ 3 Months and Select Characteristics: Social Climate Survey of Tobacco Control, United States, 2011

	Reported Building Incursions (n = 162)			Reported Unit Incursions (n = 87)a
Select Characteristics	% (No.)	OR (95% CI)b	Pc	% (No.)	OR (95% CI)b	Pc
Race			.61			.31
White	27.7 (72)	1.0 (Ref)		45.4 (33)	1.0 (Ref)
African American	27.5 (34)	1.0 (0.5, 2.0)		68.9 (24)	2.7 (0.7, 9.6)
Other	34.1 (56)	1.4 (0.7, 2.7)		58.4 (31)	1.5 (0.5, 4.4)
Gender			.09			.64
Female	24.7 (69)	1.0 (Ref)		50.7 (35)	1.0 (Ref)
Male	34.6 (93)	1.6 (0.9, 2.8)		56.1 (52)	1.3 (0.5, 3.1)
Age, (n = 542, 161), y			.62			.28
18–29	28.6 (49)	1.1 (0.5, 2.3)		48.7 (24)	1.0 (0.3, 3.7)
30–44	27.3 (50)	1.0 (Ref)		48.3 (24)	1.0 (Ref)
45–59	38.0 (37)	1.6 (0.7, 3.6)		73.5 (27)	3.0 (0.8, 11.2)
≥ 60	27.9 (25)	1.0 (0.5, 2.3)		44.0 (11)	0.8 (0.2, 3.2)
2010 household income (n=521, 154), $			.04			.3
≤ 20 000	40.1 (46)	1.5 (0.8, 3.1)		68.4 (32)	2.3 (0.7, 7.2)
20 000–75 000	30.3 (85)	1.0 (Ref)		48.8 (42)	1.0 (Ref)
≥ 75 000	18.4 (23)	0.5 (0.2, 1.1)		50.2 (11)	1.1 (0.3, 3.8)
Education			.05			.19
< high school	62.0 (19)	4.0 (1.3, 11.8)		80.4 (16)	3.1 (0.7, 13.8)
High school	26.6 (34)	0.9 (0.4, 1.9)		45.8 (16)	0.6 (0.2, 2.1)
Some college	26.1 (41)	0.9 (0.4, 1.8)		42.2 (17)	0.6 (0.2, 1.9)
≥ bachelor’s degree	29.2 (68)	1.0 (Ref)		57.2 (39)	1.0 (Ref)
US Census region			.73			.88
West	31.4 (65)	1.0 (Ref)		58.4 (38)	1.0 (Ref)
Northeast	31.6 (36)	1.0 (0.5, 2.2)		55.0 (20)	0.9 (0.2, 3.2)
Midwest	32.4 (23)	1.1 (0.5, 2.3)		48.6 (11)	0.7 (0.2, 2.4)
South	24.3 (38)	0.7 (0.3, 1.5)		48.0 (18)	0.7 (0.2, 2.2)
Government-subsidized housing (n = 513, 145)			.01			.94
No	26.1 (124)	1.0 (Ref)		54.5 (67)	1.0 (Ref)
Yes	55.2 (22)	3.5 (1.4, 8.9)		53.1 (12)	1.0 (0.3, 3.6)
Children living in home			.02			.02
No	25.1 (98)	1.0 (Ref)		64.5 (63)	1.0 (Ref)
Yes	40.2 (64)	2.0 (1.1, 3.7)		37.7 (24)	0.3 (0.1, 0.9)
Smoking status			.03			< .001
Current smoker	11.5 (8)	1.0 (Ref)		98.5 (8)	1.0 (Ref)
Not a current smoker	32.1 (154)	3.7 (1.1, 12.4)		51.6 (80)	0.02 (0.002, 0.2)
Which best describes your apartment or attached housing			.1			.07
Smoking is allowed on the property in shared areas and in the apartments or units	36.2 (71)	2.0 (0.9, 4.3)		47.2 (34)	1.0 (0.3, 3.6)
Smoking is allowed only in apartments or units	41.6 (32)	2.5 (1.0, 6.4)		84.3 (27)	5.7 (1.0, 32.1)
Smoking is not allowed in apartments or unit, but is allowed on the property	22.7 (24)	1.0 (0.4, 2.6)		40.1 (10)	0.7 (0.1, 3.7)
Smoking is not allowed at all—even in apartments or units	22.5 (35)	1.0 (Ref)		48.6 (17)	1.0 (Ref)
Which of the following best describes where you live			< .001			.01
1-family house attached to ≥ 1 house	11.4 (20)	1.0 (Ref)		23.1 (5)	1.0 (Ref)
Apartment or condominium building	38.3 (142)	4.8 (2.4, 9.9)		58.2 (82)	4.7 (1.2, 17.5)

Note. CI = confidence interval; OR = odds ratio. For each characteristic, n = (number of building incursion responses, number of unit incursion responses).

^aThis question was asked of all respondents who reported “yes” to the question “Do you ever smell smoke in your multiunit building?”(n = 162).

^bThis represents unadjusted odds ratios and 95% CI comparing respondents who reported building/unit incursions to those who did not report incursions. Models for building incursions are based on 548 respondents; for unit incursions, 162 respondents.

^cP values were determined by the Rao–Scott χ² test.

Unit incursions.

Individual unit incursions were reported more often among respondents without children in the home, compared with those with children in the home (64.5% vs 37.7%; P = .02). Respondents who reported being smokers were also more likely than nonsmokers to report unit incursion (98.5% vs 51.6%; P < .001) as were residents of apartment or condominiums (58.2%) compared with those living in 1-family attached houses (23.1%; P = .01). Respondents who lived in MUH where smoking is allowed only in apartments or units reported unit SHS incursions more frequently (84%) compared with those who lived in MUH with any other type of smoking rule (40%–49%; P = .07). When compared with MUH residents living where smoking is allowed on the property in shared areas and in the apartments or units, unit SHS incursions were higher among respondents living where smoking is allowed only in apartments or units (OR = 6.0; 95% CI = 1.4, 26.1).

Multivariable Results

Building incursions.

Respondents who reported building incursions compared with those who did not report building incursions were more likely to be men than women (OR = 2.0; 95% CI = 1.0, 4.0), to have children living in the home compared with no children at home (OR = 2.5; 95% CI = 1.2, 5.4), to live where smoking is allowed on the property in shared areas and in the apartments or units compared with not allowed at all (OR = 3.3; 95% CI = 1.4, 8.0), and to live in an apartment or condominium compared with a 1-family attached house (OR = 6.2; 95% CI = 2.4, 15.9; Table 4). Respondents whose highest education level was high school were less likely to report building incursions compared with those whose highest level of education was at least a bachelor’s degree (OR = 0.4; 95% CI = 0.2, 1.1). Nine percent (n = 52) of respondents in the cohort had missing data for 1 or more of the model variables and were excluded from the analyses. Model fit for the building incursion model was statistically significant (Hosmer–Lemeshow test P < .05) and the c statistic was 0.74.

TABLE 4—

Multivariable Associations Between Self-Reported Smoke Incursions Among Multiunit Housing Residents With No Home Smoking for ≥ 3 Months and Select Characteristics: Social Climate Survey of Tobacco Control, United States, 2011

Characteristics	OR (95% CI)a
Building incursions (n = 496)
Gender
Male	2.0 (1.0, 4.0)
Female (Ref)	1.0
Education
High school	0.4 (0.2, 1.1)
< high school	2.4 (0.7, 7.8)
Some college	0.8 (0.3, 1.9)
≥ bachelor’s degree (Ref)	1.0
Government-subsidized housing
Yes	3.5 (1.3, 9.4)
No (Ref)	1.0
Children living in home
Yes	2.5 (1.2, 5.4)
No (Ref)	1.0
Smoking status
Not a current smoker	4.7 (0.9, 24.6)
Currently smokes some days or every day (Ref)	1.0
MUH smoking rules
Smoking is not allowed at all—even in apartments or units (Ref)	1.0
Smoking is allowed on the property in shared areas and in the apartments or units	3.3 (1.4, 8.0)
Smoking is allowed only in apartments or units	2.1 (0.7, 6.0)
Smoking is not allowed in apartments or unit, but is allowed on the property	1.4 (0.4, 4.5)
Type of MUH
1-family house attached to ≥ 1 house (Ref)	1.0
Apartment or condominium building	6.2 (2.4, 15.9)
Race/ethnicity
White (Ref)	1.0
African American	1.1 (0.5, 2.4)
Other	1.3 (0.6, 2.9)
Age, y
18–29	0.7 (0.3, 1.9)
30–44 (Ref)	1.0
45–59	2.1 (0.9, 5.2)
≥ 60	1.1 (0.4, 2.8)
US Census region
West (Ref)	1.0
Northeast	1.4 (0.6, 3.4)
Midwest	1.2 (0.4, 3.1)
South	0.8 (0.3, 1.9)
Unit incursions (n = 162)
Education
< high school	10.2 (1.8, 58.8)
≥ high school graduate	1.0
Children living in home
Yes	0.2 (0.06, 0.6)
No (Ref)	1.0
Smoking status
Not a current smoker	0.001 (< 0.001, 0.03)
Currently smokes some days or every day (Ref)	1.0
MUH smoking rules
Smoking is not allowed at all—even in apartments or units (Ref)	1.0
Smoking is allowed in the apartments or units or on the property	1.1 (0.3, 4.5)
Smoking is allowed only in apartments or units	8.9 (1.5, 52.6)
Type of MUH
1-family house attached to > 1 house (Ref)	1.0
Apartment or condominium building	13.6 (2.6, 69.9)

Note. CI = confidence interval; MUH = multiunit housing; OR = odds ratio.

^aAdjusted odds ratios and 95% CI comparing respondents who reported building/unit incursions to those who did not report incursions. Models for building incursions are based on 548 respondents; for unit incursions, 162 respondents.

Unit incursions.

Reporting unit SHS incursions was positively associated with having less than a high school education compared with greater than a high school education (OR = 10.2; 95% CI = 1.8, 58.8), living in MUH where smoking is allowed only in apartments or units compared with not allowed at all (OR = 8.9; 95% CI = 1.5, 52.6), and living in an apartment or condominium compared with a 1-family attached house (OR = 13.6; 95% CI = 2.6, 69.9). Unit SHS incursion was negatively associated with having children living at home (OR = 0.2; 95% CI = 0.06, 0.60) and not being a current smoker (OR = 0.001; 95% CI = 0.001, 0.03). The model fit for the building incursion model was not statistically significant (Hosmer–Lemeshow test P = .05) and the c statistic was 0.74.

DISCUSSION

Nearly one third of multiunit housing residents who did not smoke in their own homes reported SHS incursions in their buildings. A significant proportion of these residents reported SHS incursions in their own units with almost 8% reporting these incursions daily. Previous studies have shown a strong preference among residents for having the option to live in smoke-free MUH,²⁰ and this study provides additional support for efforts to create completely smoke-free living options for residents.

The most recent national data found an incursion rate of 46.2% using the question “In the past 12 months, how often has tobacco smoke entered your unit from somewhere else in or around your building,” and incursion defined as answering “often,” “sometimes,” or “rarely.”¹⁹ We framed our questions somewhat differently, asking initially about building incursions, which were reported by 28.9%, and then asking only those residents who reported building incursions about their unit incursions. Of these noting a higher percentage than in the previous study, 53.8%, reported unit incursions. It is possible that we missed people with unit incursions who didn’t fully understand that the building incursion question would include units, and we have changed our survey question for future collection. In addition, Licht et al. asked about SHS incursions for MUH residents with smoke-free home rules, whereas we asked the question of residents who had not had anyone smoke in their home in the past 3 months; residents who endorse a smoke-free home rule might have different experiences than those whose homes might be incidentally smoke-free. These differences highlight the subtlety of asking about SHS incursions, as well as the need to find an objective measure to document exposure, such as an air nicotine monitor, and to assess compliance with smoke-free MUH regulations as they are implemented.

The most surprising finding in our study is that having partial smoke-free policies (i.e., allowing smoking only in units, but restricting it in common areas) was associated with a higher likelihood of SHS unit incursion for nonsmoking residents with nonsmoking homes. It is likely that restricting smoking to units only increases smoking behavior inside these units and therefore increases the likelihood of smoke drifting from one unit to another. Alternatively, residents who have chosen housing with smoking restrictions might be more aware of incursions and more likely to report. This finding is consistent with studies that have examined the transfer of smoke between units.¹⁵ As residents, building managers, and policymakers craft restrictions, they need to be aware that partial smoke-free policies might increase the risk of exposure for nonsmoking residents in their own homes. The results of our study suggest that the most effective solution for minimizing or eliminating SHS incursions in MUH is a comprehensive smoke-free policy that prohibits smoking in all areas of multiunit housing, public and private. Although some may be concerned about the rights of smoking residents, previous work has provided a strong legal rationale for the restriction of smoking in multiunit housing.²⁹ Smoke-free policies could proceed as leases are renewed, and tobacco cessation programs and access to cessation medications for the smoking residents could be made available to those affected by new regulations.²⁹

The fact that residents in government-subsidized housing were more likely to report building incursion raises concerns because they may not have the option for changing their housing situation, even if they or their families are experiencing negative health effects from the smoke. A recent study of those in subsidized housing found high rates of exposure (88%) as measured by cotinine assessments.³⁰ Tobacco smoke incursions can have a significant negative effect on quality of life; in a previous study, 77% of residents who experienced a unit incursion indicated that they were bothered by it.²⁰ Residents receiving government subsidies are also more likely to be elderly, disabled, or have children in the home; these are all populations that spend more time in the home, and may have medical conditions aggravated by smoke exposure.²³ In addition, residents receiving government subsidies tend to support completely smoke-free building policies.³¹ The growing body of evidence supports efforts by HUD to ban smoking completely in all areas of government-subsidized housing.²³

We had several other groups with lower likelihood of reporting incursions. Those with at least a high school education were less likely to report incursion, even controlling for socioeconomic status and type of MUH; although we can’t tell from the data, those with more education may be more likely to be in the workforce and, thus, less likely to be at home to notice incursions. Smokers may have been less likely to report building incursions because they are habituated to the smell and less likely to be irritated by it. The inverse of this association was found for unit incursions (nonsmokers were less likely to report unit incursions). However, this could be attributed to the small sample size in this group.

Homes with children were more likely to report building incursions, and nearly 20% of MUH residents are children younger than 18 years.¹³ Of special concern are the approximately 800 000 children who live in government-subsidized MUH who may be at increased risk for tobacco smoke exposure in their homes.³² Living in poverty is associated with decreased cognitive functioning,³³ even when controlling for genetic contributions.³⁴ Because even small amounts of tobacco smoke exposure have been associated with childhood cognitive deficits⁵ and school absenteeism,³⁵ chronic low-level exposures from adjacent units or common areas might place children already at high risk for low achievement at an even greater disadvantage. It is unclear why homes with children were less likely to report unit incursions; however, our unit incursion sample was small and future research with larger populations of homes with children may help us understand this relationship better.

Limitations

This study has several limitations. Although our dual-frame methodology is designed to reduce the potential for sample bias associated with either RDD or Internet panel samples alone, we still can’t eliminate the potential for sample ascertainment bias. Our incursion data are based on self-report of smelling tobacco smoke; we don’t have biological confirmation of incursions. However, our previous work has confirmed higher cotinine in children living in nonsmoking units of multiunit housing compared with detached homes.¹⁸ Our sample of families with children was small, which limits our ability to determine incursion rates in homes with children. Our overall unit incursion sample was also small, as indicated by the wide 95% CIs, and should be interpreted with caution. The data are cross-sectional, and we can’t say whether the different regulations were the cause of the increased incursions, or just a marker for an unmeasured confounder; the suboptimal model fit for building incursions suggests that factors not assessed in this study contribute to the relationship. However we believe that our data support previous studies that have found a significant burden of SHS incursions in MUH, and provide further information on the factors associated with experiencing incursions, as well as the complex role played by different degrees of smoking restrictions.

Conclusions

Many residents of multiunit housing are exposed to tobacco smoke by their neighbors, increasing the risk that they or their children will suffer health consequences and negatively affecting quality of life. Partial smoke-free policies do not appear to reduce self-reported SHS exposure residents, and in fact, might increase the risk of incursions in residents’ individual units. Biological confirmation of incursions would further elucidate the impact of partial smoke-free policies on MUH residents. Eliminating all smoking in government-supported MUH would protect those residents most likely to report incursions and the least able to change their residence to avoid the exposure. Everyone should have the right to live in smoke-free housing.