University of Southern California Dissertations and Theses

Tobacco use change among formerly homeless supportive housing residents: socioecological barriers and facilitators to cessation

(USC Thesis Other)

Tobacco use change among formerly homeless supportive housing residents: socioecological barriers and facilitators to cessation

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

TOBACCO USE CHANGE AMONG FORMERLY HOMELESS SUPPORTIVE HOUSING
RESIDENTS:
SOCIOECOLOGICAL BARRIERS AND FACILITATORS TO CESSATION

by

Taylor Harris

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
SOCIAL WORK

August 2020

Copyright 2020 Taylor Harris

ii

Dedication
This work is dedicated to every person in Los Angeles who faces the terrifying and
traumatizing experience of being without a secure place to call home.
To the Transitions to Housing Participants, I am forever grateful for your willingness to
share your lives with me and the research team. Your vulnerability, openness, and resilience
have had a profound impact on my life in ways I cannot even begin to put into words. This is for
you.
In the wake of the murder of yet another unarmed Black person by a police officer, it
feels integral to acknowledge the deeply woven relationship between racism and homelessness.
Although these studies may not explicitly investigate the racial injustices incurred by Black
persons experiencing homelessness and supportive housing residents, the systemic racism and
consequential barriers to health and well-being are inherent in the issues examined in this body
of work. In combination with a promise to continue to dismantle the oppression and inequalities
faced by the Black community throughout my career, I dedicate this to every person who is born
with Black skin, including my husband. Your lives matter.

iii

Acknowledgements
To say “it takes a village” would be an understatement. This is an attempt to
acknowledge all the beings who have shaped my growth as a scholar and made this labor of love
possible.
To my mentor and dissertation chair. To Dr. Suzanne Wenzel for your commitment to
enhance the lives of persons experiencing homelessness, for imparting your wisdom, knowledge
and rigor upon me, for offering and instilling your resources in me, and for believing in me in
times when I did not believe in myself. I am utterly grateful for your grace and guidance.
To my dissertation and F31 committee members. To Dr. Ben Henwood for pushing me to
think deeper and shaping me as a writer. To Dr. Jennifer Unger for your inclusiveness, positivity
and pragmatism that have grounded me and propelled my scholarship. To Dr. Tom Valente for
your insight, directness and for making research really fun. To Dr. Joan Tucker for your support
and sensibility. To Dr. Chi Ping Chou for your mathematic brilliancy and mindful perspective.
To Suzanne Dworak-Peck School of Social Work Faculty and Staff. To Dr. Mike Hurlburt
for your care and consideration of student welfare that has been a source of comfort and set a
standard of excellence in mentorship. To Malinda Sampson for your patience, responsiveness
and care for my progression and achievement, and for always saving me a vegan lunch. To Dr.
Eric Rice for your candor, enthusiasm and understanding. To Dr. Carl Castro for including me in
your research, providing unique collaborative experiences and laying the foundation for my
future work with veterans. To Dr. Larry Palinkas for helping me shape this body of work and
reflecting my thoughts back in truly eloquent ways. To Dr. Monica Perez-Jolles, while our work
together came much later in my doctoral training, I am immensely grateful for your kindness and
inclusivity; I wish I had known your warm presence earlier.

iv

To my Transitions to Housing colleagues and friends. To Dr. Harmony Rhoades, your
intellect, constructive feedback and collegial nature has been so impactful on my development as
a researcher. I am a better social scientist because of you and I cannot thank you enough for your
teachings. To Hailey Winetrobe, you are one of the few friends who understands my professional
and personal worlds, which I cherish so much. Thank you for being a beacon of resilience and
leading by example from day one. To Monica Caraballo for always brightening my day, listening
reflectively and bringing light and love to this journey. To Adam Carranza and David Dent for
your shared dedication and commitment to the study that shaped my career, and for sharing your
experiences in the field that formulated empirical inquiries throughout my doctoral training and
beyond.
To my School of Social Work friends and colleagues. To Carrie Lucas for your heartfelt
conversations, scientific collaboration, and being the best conference roommate. To Carolina
Grest for being a confidant and willing ear to all personal and programmatic challenges and
triumphs faced in these past 6 years. To Cary Klemmer for being an insightful free spirit and an
understanding friend. To Sara Semborski for spending hours talking about statistics - your caring
nature and joyous personality that made these last few years so vibrant. To Robin Petering for
being a model of inclusivity and your commitment to community-based work with the homeless
populations of Los Angeles. To others who have been a source of warmth and support,
contributing to my development in this process: Judith Perrigo, Tasha Perdue, Joshua Ruscow,
Wichada La Motte-Kerr, Hadass Moore, Jaih Craddock, and Jessica Dodge. Thank you.
To my friends. I have known many of you for the greater part of my life and I am so
grateful for your lifelong friendships. We have grown up together. I am thankful you stuck by
me. You have seen me and supported me through all the lows and all the highs; the growing

v

pains, including the process of graduate school. Through all of it, you have been my constant.
Thank you for making me laugh and being there when I cry, for your directness and for your
truth. I attribute so much of my growth as a human being to our bonds. Thank you for being the
Miracle Grow to my soul: Lauren Hoffmann, Amber Nix, Vahan Petrossian, Victoria Le-
Shubert, Allyson Valley, Teal Berkowitz, Nina Polyne, Meredith Kalies, Yajaira Padilla, Sena
Cerna, Danielle Costello & Sasha Marfow.
To my family. To my mother and father who made endless sacrifices to position me to be
the first person in our family to graduate from college. You have always given me the resources I
needed and so much more. Thank you for never questioning my lengthy college stint and for
continuously encouraging me to pursue my dreams. You have shaped me into the woman I am
today. I promise to continue to make you proud. To my sister and best friend, you are the realest
person I know, and I feel truly lucky I got to grow up alongside you. Thank you for your tough
love approach, for making me laugh hysterically, for building a beautiful family that has brought
me so much happiness, and for being the person I can count on throughout my life. To my
brother, for willingly listening to my diatribes, for being so calm and sweet, and for accepting me
for who I am. To my grandmother, who is my archetype of determination, love and hope. You
have always believed in me, supported me, nurtured me and made me feel unstoppable. Thank
you for paving the way for the generations that follow you, for being an example of a nasty
woman (even if you did not always identify as one) and for your magical spirit. To my Naija
family, you have taught me so much about love and acceptance for which I am eternally grateful.
Not only have you expanded my family, you have given me a firsthand understanding of cultural
strength and kinship. You have gifted me with Jollof rice and my favorite human being. I am so
proud to be yours. Thank you.

vi

To my husband. Jay, you have been by my side through every paper, every revise and
resubmit, every rejection, every version of my grant, every computer crash, and every success.
You are a pillar of support and strength in my career and my everyday life. Thank you for being
my source of admiration and aspiration. Thank you for being my partner in this educational
process and in this life. There are simply not sufficient words that capture my gratitude for your
love, steadiness, and intellect. I could not have done this without you.

vii

Table of Contents
Dedication ....................................................................................................................................... ii
Acknowledgements ........................................................................................................................ iii
List of Tables ................................................................................................................................. xi
List of Figures ............................................................................................................................... xii
Abstract ........................................................................................................................................ xiii
Chapter 1: Introduction ................................................................................................................... 1
Homelessness and Tobacco Use ............................................................................................... 2
Permanent Supportive Housing and Cessation ......................................................................... 3
Guiding Conceptual Framework and Corresponding Aims ..................................................... 6
Physical and Social Environment Correlates of Tobacco Use............................................ 6
Intersection between Physical and Social Environments.................................................... 8
Individual Factors and Perceptions ..................................................................................... 9
References ............................................................................................................................... 12
Tables and Figures .................................................................................................................. 21
Chapter 2: Trajectories of Smoking Behavior and Change in the Transition from Homelessness
to Supportive Housing .................................................................................................................. 24
Introduction ............................................................................................................................. 24
Methods................................................................................................................................... 28
Measures ........................................................................................................................... 29
Analysis............................................................................................................................. 32
Results ..................................................................................................................................... 34
Participants ........................................................................................................................ 34

viii

RMLCA ............................................................................................................................ 35
Multinomial Logistic Regression: Univariable Models .................................................... 36
Multinomial Logistic Regression: Multivariable Models ................................................. 37
Discussion ............................................................................................................................... 37
Conclusion ........................................................................................................................ 42
References ............................................................................................................................... 43
Tables and Figures .................................................................................................................. 51
Chapter 3: Tobacco Risk Change in the Transition from Homelessness to Housing: The Role of
Housing and Support..................................................................................................................... 59
Introduction ............................................................................................................................. 59
Permanent Supportive Housing ........................................................................................ 59
Tobacco Risk .................................................................................................................... 63
Current Study .................................................................................................................... 64
Methods................................................................................................................................... 65
Measures ........................................................................................................................... 68
Analysis............................................................................................................................. 72
Results ..................................................................................................................................... 74
Descriptive Statistics ......................................................................................................... 74
Social Support Change ...................................................................................................... 75
Unconditional Models ....................................................................................................... 76
Conditional Main-Effect Models ...................................................................................... 76
Conditional Interactive Models......................................................................................... 77
Discussion ............................................................................................................................... 80

ix

Conclusion ........................................................................................................................ 85
References ............................................................................................................................... 88
Tables and Figures .................................................................................................................. 99
Chapter 4: Barriers and Facilitators to Cessation among Supportive Housing Residents: A
Qualitative Inquiry ...................................................................................................................... 119
Introduction ........................................................................................................................... 119
Method .................................................................................................................................. 124
Parent Study .................................................................................................................... 124
Measures ......................................................................................................................... 125
Current Study .................................................................................................................. 125
Analysis........................................................................................................................... 128
Results ................................................................................................................................... 129
Housing Environment: Facilitators ................................................................................. 130
Housing Environment: Barriers ...................................................................................... 133
Social Environment: Facilitators..................................................................................... 136
Social Environment: Barriers .......................................................................................... 142
Individual: Facilitator...................................................................................................... 145
Individual: Barrier ........................................................................................................... 149
Discussion ............................................................................................................................. 156
Limitations ...................................................................................................................... 161
Conclusion ...................................................................................................................... 162
References ............................................................................................................................. 165
Tables .................................................................................................................................... 174

x

Chapter 5: Conclusion................................................................................................................. 177
Introduction ........................................................................................................................... 177
Review of Major Findings and Integration with Existing Research ..................................... 179
Chapter 2: Trajectories of Smoking Behavior and Change in the Transition from
Homelessness to Supportive Housing ............................................................................. 179
Chapter 3: Tobacco Risk Change in the Transition from Homelessness to Housing: The
Role of Housing and Support .......................................................................................... 182
Chapter 4: Barriers and Facilitators to Cessation among Formerly Homeless Supportive
Housing Residents .......................................................................................................... 183
Integration, Triangulation, and Theoretical Application ...................................................... 186
Policy and Practice Implications, Recommendations, and Future Research ........................ 189
Physical Environment ..................................................................................................... 190
Social Environment ......................................................................................................... 191
Conclusion ............................................................................................................................ 196
References ............................................................................................................................. 197

xi

List of Tables
Table 1.1. Sample Statistics (N = 421) ......................................................................................... 21
Table 2.1. Descriptive Statistics for Total Sample of Tobacco Users and Each Latent Class ...... 51
Table 2.2. Model Fit Indexes for Models with Two to Four Classes ........................................... 54
Table 2.3. Univariable Multinomial Logistic Regression Models with Predictors of Latent Class
Membership .................................................................................................................................. 55
Table 2.4. Multivariable Multinomial Logistic Regression Models with Predictors of Latent
Class Membership ......................................................................................................................... 57
Table 3.1. Descriptive Statistics of Time-Invariant Measures Examined as Predictors and
Controls in Latent Growth Curve Models .................................................................................... 99
Table 3.2. Descriptive Statistics of Time-Variant Measures Examined as Predictors of Tobacco
Risk and Controls in Latent Growth Curve Models ................................................................... 100
Table 3.3. Results from Unconditional Growth Models (Models 1–3) ...................................... 102
Table 3.4. Conditional Main Effect Models (Models 4 and 5 [Including Covariates]) .............. 103
Table 3.5. Conditional Interactive Models (Models 4 and 5 [Including Covariates]) ................ 106
Table 4.1. Descriptive Statistics of Purposive Sample (N = 33) That Participated in
Semistructured Interviews .......................................................................................................... 174
Table 4.2. Summary of Thematic Findings at Multiple Socioecological Levels ........................ 175

xii

List of Figures
Figure 1.1. Socioecological Levels and Their Purported Effects on Smoking Behavior According
to Guiding Theoretical Frameworks ............................................................................................. 23
Figure 3.1. Linear and Quadratic Growth in Tobacco Risk as Adults Transition from
Homelessness to Supportive Housing ......................................................................................... 113
Figure 3.2. Difference in Slopes of Network Size over Time across Housing Models .............. 114
Figure 3.3. Difference in Slopes Representing Change in Neighbors in Networks over Time
across Housing Model................................................................................................................. 115
Figure 3.4. Difference in Slopes of Change in Nearby Network Members across Single-Site and
Scattered-Site Residents.............................................................................................................. 116
Figure 3.5. Differing Slopes Depicting Change in Conflicting Network Members across Skid
Row and DTLA PSH Residents .................................................................................................. 117
Figure 3.6. Slopes Representing Change in Neighbors across Skid Row and Other Neighborhood
PSH Residents ............................................................................................................................. 118

xiii

Abstract
Permanent supportive housing (PSH) offers permanent housing and supportive services
to formerly homeless tobacco users with high rates of mental health and substance use
conditions, enhancing opportunities to overcome the numerous barriers to cessation experienced
while homeless, including physical (e.g., housing instability, proximity to services) and social
(e.g., emotional support, provider support) environmental barriers, yet no empirical research has
explored whether these characteristics of PSH affect tobacco use behavior change. This body of
work is the first to examine associations between the physical, social and individual
characteristics of PSH and tobacco use behavior change, and the differential tobacco use change
pathways across chronic mental health conditions and risky substance use behaviors, among
homeless adults transitioning to housing, using multi-level social ecological data (i.e.,
quantitative, social network, qualitative). Results from this research enhance the identification of
new directions in efforts needed to improve cessation among supportive housing residents.
Specifically, identifying how housing, social support, mental health and substance use factors
can be altered or incorporated into efforts to improve cessation outcomes for PSH residents.

1

Chapter 1: Introduction
This dissertation is composed of three studies presented across three chapters. The
overarching goal of this dissertation is to identify socioecological hindrances to and facilitators
of smoking cessation among formerly homeless persons living in supportive housing, one of the
highest-risk groups of tobacco users. Although a great deal of research has explored tobacco use
in the general adult population, smoking is considered the “neglected addiction” for persons
experiencing homelessness (Baggett, Tobey, & Rigotti, 2013). Given the dearth of tobacco use
research in this population, this is the first body of work of its kind and thus, is largely
exploratory. The use of rich multilevel, multimethod data makes this a unique approach to
understand how social, physical, and individual factors come together to differentially affect
tobacco use change among supportive housing residents. The specific aims guiding this
dissertation are: (a) identify tobacco use trajectories in the transition from homelessness to
housing and determine whether mental health and substance use differentially affect patterns of
smoking change or maintenance; (b) identify the predictive roles of the physical and social
environment on smoking behavior change during the initial year in housing; and (c) explore
residents’ perceived facilitators of and barriers to tobacco use cessation as a function of tobacco
change behavior status.
To understand the phenomenon of tobacco use change from homelessness to housing, a
mixed-methods approach is necessary for a well-rounded understanding of the factors, and the
interrelation of these factors, that facilitate or hinder tobacco cessation. The initial study, Chapter
2, uses four waves of longitudinal, quantitative data to identify classes of tobacco use frequency
over time and determine mental health and substance use correlates of class membership.
Chapter 3 utilizes longitudinal quantitative and social network data to examine physical and

2

social environmental factors, changes in the social environment, and interactions between the
physical and social environments associated with changes in tobacco risk from homelessness to
housing. Chapter 4 uses primary qualitative data from semistructured interviews to explore
residents’ perceptions of the factors that facilitate or hinder their ability to cease using tobacco.
The use of multiple data sources and merging of quantitative and qualitative findings in Chapter
5 triangulates these findings, enhancing the credibility of evidence regarding tobacco change
patterns among supportive housing residents (Yin, 2003). By incorporating these data sources,
this dissertation provides researchers and service providers with an enhanced understanding of
tobacco change patterns and knowledge to target cessation across various contexts, improving
residents’ receipt of cessation treatments that meet their needs and ultimately improving
cessation rates for this population.
Homelessness and Tobacco Use
The 2015 National Health Interview Survey of households indicated 1 in 4 adults living
below the federal poverty line are tobacco users (Centers for Disease Control and Prevention,
2016). Omitted from these household data are homeless adults, who are among the most
vulnerable impoverished populations and report tobacco use rates up to 5 times that of housed
individuals living in poverty (Baggett et al., 2013; Connor, Cook, Herbert, Neal, & Williams,
2002; Okuyemi et al., 2006; Soar, Dawkins, Robson, & Cox, 2020). Although the physical harm
associated with tobacco use results in $170 billion spent annually on health care (Centers for
Disease Control and Prevention, 2016), efforts to address tobacco use among adults experiencing
homelessness often come second to other immediate mental and physical health needs, in
addition to basic needs such as housing. The use of tobacco products by homeless persons is
particularly problematic because it exacerbates already compromised health commonly

3

associated with harsh living conditions (Baggett et al., 2013). The adverse effects of tobacco use
and health costs are especially concerning for homeless adults, given their predisposition to poor
health and frequent use of high-cost medical services (Larimer et al., 2009; Martinez & Burt,
2006). Additionally, homeless persons engage in tobacco practices such as sniping (smoking
discarded cigarettes) that further risk their health (Tucker, Shadel, Golinelli, Mullins, & Ewing,
2015). These rates and risks have accumulated to create a major public health concern, with
tobacco use-related deaths occurring at twice the rate among homeless adults compared to
housed adults (Vijayaraghavan, Tieu, Ponath, Guzman, & Kushel, 2016).
Homeless persons’ increased risk of tobacco addiction is commonly attributed to poor
mental health and other substance use disorders in this population (Baggett et al., 2013; Harris,
Winetrobe, Rhoades, & Wenzel, 2019). In addition, social correlates impede cessation, including
but not limited to the normative acceptability of tobacco use (Businelle, Cuate, Kesh,
Poonawalla, & Kendzor, 2013; Okuyemi et al., 2006), diminished access to routine provider
support (Antoon, 2017), and lack of social support (Arnsten, Reid, Bierer, & Rigotti, 2004;
Gabrielian, Young, Greenberg, & Bromley, 2018). Moreover, although adults experiencing
homelessness have expressed motivation to cease tobacco use, the marginalization of this
population has led cessation to be a low-ranking priority among service providers (Businelle et
al., 2013; Okuyemi et al., 2006). Cessation is also markedly more difficult due to physical
environmental barriers, such as housing instability, proximity to health care settings, and means
of transportation for accessing cessation treatments (Antoon, 2017; Baggett et al., 2013).
Permanent Supportive Housing and Cessation
Permanent supportive housing (PSH) is a widely accepted solution to ending
homelessness through the provision of long-term, affordable housing combined with support

4

services (Tsemberis, Gulcur, & Nakae, 2004). PSH provides homeless adults with opportunities
to overcome physical and social environmental barriers via housing stability, increased access to
provider care, and improved social relationships (Larimer et al., 2009; Martinez & Burt, 2006;
Tsemberis, 2014; Tsemberis et al., 2004). However, the varying contexts of these
environments—such as whether PSH residents live in housing models that are clustered (single-
site) or dispersed (scattered) among other formerly homeless persons (Collins, Onwuegbuzie, &
Johnson, 2012), receive support from providers (Bao, Duan, & Fox, 2006), or are connected to
prosocial network members once housed (Rhoades et al., 2017)—may also lead to differential
tobacco use change.
The transition into PSH offers homeless adults opportunities for health behavior change
via enriching social environmental resources, such as increased access to health care providers
and other prosocial relationships (Tsemberis et al., 2004). Studies have empirically demonstrated
that support services in PSH increase access to appropriate health services for formerly homeless
persons (Larimer et al., 2009; Martinez & Burt, 2009; Tsemberis, 2014). Yet whether these
linkages result in higher cessation rates among residents remains largely unknown. Additionally,
the presence of other prosocial relationships in individual social networks are associated with
readiness to quit (Baggett & Rigotti, 2010) and successful cessation (Lee & Kahende, 2007).
Research has suggested social networks change as individuals transition to PSH, influencing
behavioral health outcomes and substance use (Henwood et al., 2017; Rhoades et al., 2017).
Thus, social network characteristics, including social roles (structural support) and the type of
support they provide (functional support), may contribute to tobacco use behavior change, but
whether social networks in PSH facilitate cessation or enhance risk propensity is absent in the
extant literature.

5

Physical environment characteristics are also correlates of health behavior change, such
that more positively perceived neighborhood and building characteristics are associated with
positive health practices

(Perry, Baranowski, & Parcel, 1990; Visser, 2007; Wright & Kloos,
2007); however, whether this is true for tobacco use has yet to be empirically examined.
Differential characteristics of residents’ neighborhood and building may positively or negatively
affect their tobacco use change.

PSH locations vary, with a tendency for single-site housing in
Los Angeles to be located on Skid Row, one of the most densely concentrated neighborhoods of
homeless persons and services in the United States (Hsu, Simon, Henwood, Wenzel, & Couture,
2016; Solari, Cortes, Henry, Matthews, & Morris, 2014), whereas scattered-site units are
typically situated in neighborhoods without such concentrations (Hwang, Stergiopoulos,
O’Campo, & Gozdzik, 2012). Individuals living on Skid Row, where tobacco product access is
abundant through street vendors, may be more likely to use tobacco compared to those in
neighborhoods where tobacco access is less prevalent (Bunnell et al., 2012). Building and
neighborhood characteristics are associated with changing health behaviors like exercise
(Stronegger, Titze, & Oja, 2010) and social integration (Yanos, Barrow, & Tsemberis, 2004), yet
how these factors may affect tobacco use change for supportive housing residents is unclear.
Furthermore, the physical and social environments of PSH are often interrelated, such
that those in single-site housing models may have on-site health care clinics, which could
increase the likelihood of receiving cessation advice from a provider. Research suggests
availability of health care (Gibson, Ghosh, Morano, & Altice, 2014; Wang & Luo, 2005) and
geographic proximity

to such services

(Arcury et al., 2005; Guagliardo, 2004; Wang & Luo,
2005) contribute to accessing providers. Likewise, some neighborhoods, such as Skid Row, host
a large percentage of homeless and formerly homeless persons, whereas other neighborhoods

6

have more diverse communities (Hsu et al., 2016; Solari et al., 2014). Living in neighborhoods
like Skid Row may also decrease the likelihood of cessation due to the normative acceptability
when living in large groups of persons that use tobacco at high rates (Businelle et al., 2013;
Okuyemi et al., 2006). These social and physical environmental factors likely come together with
individual-level factors, such as mental health and other substance use, to form varying tobacco
change patterns in the transition from homelessness to housing. However, these relationships
have yet to be examined empirically.
Guiding Conceptual Framework and Corresponding Aims
Physical and Social Environment Correlates of Tobacco Use
Social learning theory (SLT) provides a theoretical basis critical to understanding how
tobacco use behavior may change in PSH. SLT can explain how individuals acquire and
maintain behavioral patterns and creates the foundation for intervention strategies (Bandura,
1985). SLT posits that the social environment is a mode by which behavior change occurs as
learning is acquired through social interactions (Bandura, 1985; Bandura & Walters, 1977). The
social environment changes in the transition to housing as social network members disappear or
are replaced; some residents restore severed familial relationships, whereas others may distance
themselves from risky peers (Henwood et al., 2017; Rhoades et al., 2017).
Changes in structural support, described as the existence of family or friends in an
individual’s network, may affect tobacco use change (Wills & Ainette, 2012). Structural support
has been found to prevent tobacco use and increase successful cessation (Brown & Rinelli, 2010;
Caponnetto & Polosa, 2008; Lee & Kahende, 2007; May & West, 2000). For example,
adolescents who reside in two-parent households are less likely to try tobacco (Brown & Rinelli,
2010). Married individuals also have higher rates of cessation (Caponnetto & Polosa, 2008; Lee

7

& Kahende, 2007). Functional support, referred to as the quality of these relationships and the
forms of support they provide, such as emotional support, may also change in the transition to
PSH (Will & Ainette, 2012). Many effective behavioral treatments for tobacco use incorporate
functional support through methods such as supportive peers (Michie, Churchill, & West, 2011;
Stead, Hartmann-Boyce, Perera, & Lancaster, 2013; Stead & Lancaster, 2012). Additionally,
naturally occurring forms of these supports facilitate successful cessation outcomes (Hennrikus
et al., 2010; May, West, Hajek, McEwen, & McRobbie, 2006; Westmaas, Bontemps-Jones, &
Bauer, 2010). Once individuals are housed, increased provider support may also affect their
tobacco use change. In PSH, residents are connected to case managers and more likely to use
primary care services compared to homeless groups (Larimer et al., 2009; Martinez & Burt,
2006; Tsemberis, 2014), which should increase rates of ceasing tobacco use (Stead, Buitrago, et
al., 2013). However, research has also suggested that health care providers often focus on other
chronic health and mental health conditions in this population (Porter et al., 2017; Sherman et al.,
2016). Case managers, who are often tobacco users themselves (Baggett et al., 2013; Okuyemi et
al., 2006), experience challenges addressing other behavioral health concerns such as substance
use and community integration (Stergiopoulos et al., 2014; Tiderington, Stanhope, & Henwood,
2013), which may be the case for tobacco use as well.
The physical environment may also contribute to tobacco use change because it provides
the context for behavior to be shaped and transformed. According to SLT, the external
environment, or physical environment, sets the stage for behavior development and change via
individuals’ responses to the stimuli in these settings, which subsequently form behavioral
patterns (Bandura, 1985). When applied to PSH residents’ tobacco use change, the physical
environment, which is composed of varying housing models, may shape differential access to

8

providers and subsequent opportunities for tobacco behavior change. For example, scattered-site
housing agencies, where residents are dispersed in standard apartment settings, may use assertive
community treatment teams to deliver health care services in residents’ homes, whereas others
may employ a more autonomous approach of referring residents to services outside their home
(Henwood, Cabassa, Craig, & Padgett, 2013; Tsemberis et al., 2004). Some single-site settings,
where homeless persons are clustered together, host on-site health care clinics, which minimizes
structural barriers, such as distance and transportation, and may increase access to provider
support (Andersen, Davidson, & Baumeister, 2013; Hastings, 2009). Increased use of appropriate
health services improves the likelihood of speaking to a provider about tobacco use. According
to U.S. clinical practice guidelines, every tobacco user seen in a health care setting should
receive strong advice to quit (Stead, Buitrago, et al., 2013). However, characteristics of the
physical environment may reduce or enhance receipt of provider support, resulting in differential
tobacco use behavior change.
Intersection between Physical and Social Environments
The social ecological model provides further conceptual guidance for understanding how
the social and physical environments are associated with tobacco use change. The social
ecological model suggests public health problems can be understood as occurring at a given level
(e.g., intrapersonal, interpersonal, and institutional) and individual behavior can be better
understood if the context in which it occurs is understood (Visser, 2007). These levels
correspond with the individual, social environment, and physical environment, respectively, and
are nested in one another such that the intrapersonal level is nested in the interpersonal level, and
these two levels are nested in the institutional level. These levels can be examined independently,
such as how individual cognition affects individual behavior. However, in accordance with the

9

model, health behaviors such as tobacco use are best understood and changed when numerous
levels and their nested nature are considered simultaneously (Macintyre & Ellaway, 2000).
Research that contains socioecological data at numerous levels is often limited due to the
resources required to obtain such diverse sources of data, making it difficult to examine
relationships such as those among the social environment, physical environment, and tobacco use
(Hwang et al., 2012). However, data from these dissertation studies contain unique
socioecological information that provides a rare opportunity to examine these relationships.
Together, SLT and the social ecological model support examining the interactive effects
of the social environment in the relationship between the physical environment and tobacco use
behavior. SLT emphasizes the role of the social environment in individual behavior because
behaviors are acquired and changed through social interaction (Bandura & Walters, 1977; Perry
et al., 1990). According to SLT, the physical environment serves as the context for the formation
of the social environment (Bandura, 1985). The social ecological model adds to this
understanding by purporting that various levels exist in nested hierarchies, with the intrapersonal
being the lowest level and each level being contained by another larger system (Visser, 2007).
These levels comprise a chain of linkages, with health behavior processes moving through higher
levels to lower levels, ultimately affecting individual health behavior (Macintyre & Ellaway,
2000). Thus, according to these theoretical models, tobacco use behavior processes will move
from the institutional level, or physical environment, through the interpersonal level, or social
environment, to influence the intrapersonal level, or individual.
Individual Factors and Perceptions
Both SLT and the social ecological model highlight the importance of the intrapersonal
level, or individual, in health behavior change. In SLT, the perceptions and expectations of

10

individuals are integral components of behavioral learning and behavior change (Bandura &
Walters, 1977; Perry et al., 1990). In the social ecological model, the intrapersonal level is the
epicenter of the ecosystem, with individual attitudes and beliefs serving as critical components of
this level (Visser, 2007). Incorporating the individual perspective allows for novel information to
emerge (Creswell, Shope, Plano Clark, & Green, 2006), such as facilitators of and hindrances to
tobacco use change that were not previously considered or included in the parent study
questionnaire. Identifying subgroups that increased, decreased, or maintained tobacco use
enhances the ability to understand these factors. Via qualitative methods that gather perceptions
and expectations of tobacco use behavior across subgroups, value-laden questions can inform
intervention efforts regarding facilitators of and barriers to cessation.
Guided by the social ecological model (Bronfenbrenner, 1977) and SLT (Bandura &
Walters, 1977), this mixed-methods study is the first to examine how the social and physical
environments of PSH contribute to tobacco use behavior. Drawing from the parent HIV Risk,
Drug Use, Social Networks: Homeless Persons Transitioned to Housing Study (NIDA Grant No.
R01DA036345), the current study (NIDA Grant No. F31DA045429-02) used four waves of
quantitative and social network data from a sample of 421 formerly homeless adult PSH
residents (see Table 1.1) to determine whether established mental health and substance use
correlates of tobacco use among homeless adults also affect pathways of tobacco use change as
adults experiencing homelessness transition to supportive housing. Additionally, it investigated
social and physical environmental factors associated with longitudinal tobacco change and
whether the social and physical environments interact to differentially affect tobacco risk
outcomes. To further understand and contextualize the impact and interconnectedness of these
socioecological levels and their effect on individual smoking behavior, the researcher gathered

11

residents’ perceptions of the contributors and hindrances to cessation through qualitative
semistructured interviews with a purposive sample of participants from the parent study who
increased, decreased, or maintained their tobacco use over time.

12

References
Andersen, R. M., Davidson, P. L., & Baumeister, S. E. (2013). Improving access to care. In G. F.
Kominski (Ed.), Changing the U.S. health care system: Key issues in health services
policy and management (pp. 33–69). San Francisco, CA: Jossey-Bass.
Antoon, B. G. (2017). Reducing tobacco-related health disparities: Exploring the barriers and
facilitators to smoking cessation among individuals experiencing homelessness (Doctoral
dissertation). University of Louisville, Louisville, KY.
Arcury, T. A., Gesler, W. M., Preisser, J. S., Sherman, J., Spencer, J., & Perin, J. (2005). The
effects of geography and spatial behavior on health care utilization among the residents
of a rural region. Health Services Research, 40, 135–156. doi:10.1111/j.1475-
6773.2005.00346.x
Arnsten, J. H., Reid, K., Bierer, M., & Rigotti, N. (2004). Smoking behavior and interest in
quitting among homeless smokers. Addictive Behaviors, 29, 1155–1161.
doi:10.1016/j.addbeh.2004.03.010
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.
doi:10.1016/j.amepre.2010.03.024
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.
doi:10.1056/NEJMp1301935
Bandura, A. (1985). Model of causality in social learning theory. In M. J. Mahoney (Ed.),
Cognition and psychotherapy (pp. 81-99). New York, NY: Plenum Press.

13

Bandura, A., & Walters, R. H. (1977). Social learning and personality development. London,
United Kingdom: Holt, Rinehart, and Winston.
Bao, Y., Duan, N., & Fox, S. A. (2006). Is some provider advice on smoking cessation better
than no advice? An instrumental variable analysis of the 2001 National Health Interview
Survey. Health Services Research, 41, 2114–2135. doi:10.1111/j.1475-
6773.2006.00592.x
Bronfenbrenner, U. (1977). Toward an experimental ecology of human development. American
Psychologist, 32, 513–531. doi:10.1037/0003-066X.32.7.513
Brown, S. L., & Rinelli, L. N. (2010). Family structure, family processes, and adolescent
smoking and drinking. Journal of Research on Adolescence, 20, 259–273.
doi:10.1111/j.1532-7795.2010.00636.x
Bunnell, R., O’Neil, D., Soler, R., Payne, R., Giles, W. H., Collins, J., … Communities Putting
Prevention to Work Program Group. (2012). Fifty communities putting prevention to
work: Accelerating chronic disease prevention through policy, systems and
environmental change. Journal of Community Health, 37, 1081–1090.
doi:10.1007/s10900-012-9542-3
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. doi:10.2105/AJPH.2013.301336
Caponnetto, P., & Polosa, R. (2008). Common predictors of smoking cessation in clinical
practice. Respiratory Medicine, 102, 1182–1192. doi:10.1016/j.rmed.2008.02.017
Centers for Disease Control and Prevention. (2016). Current cigarette smoking among adults—
United States, 2005–2015. Morbidity and Mortality Weekly Report, 65, 1205–1211.

14

Collins, K. M. T., Onwuegbuzie, A. J., & Johnson, R. B. (2012). Securing a place at the table: A
review and extension of legitimation criteria for the conduct of mixed research. American
Behavioral Scientist, 56, 849–865. doi:10.1177/0002764211433799
Connor, S. E., Cook, R. L., Herbert, M. I., Neal, S. M., & Williams, J. T. (2002). Smoking
cessation in a homeless population. Journal of General Internal Medicine, 17, 369–372.
doi:10.1046/j.1525-1497.2002.10630.x
Creswell, J. W., Shope, R., Plano Clark, V. L., & Green, D. O. (2006). How interpretive
qualitative research extends mixed methods research. Research in the Schools, 13, 1–11.
Gabrielian, S., Young, A. S., Greenberg, J. M., & Bromley, E. (2018). Social support and
housing transitions among homeless adults with serious mental illness and substance use
disorders. Psychiatric Rehabilitation Journal, 41, 208–215. doi:10.1037/prj0000213
Gibson, B. A., Ghosh, D., Morano, J. P., & Altice, F. L. (2014). Accessibility and utilization
patterns of a mobile medical clinic among vulnerable populations. Health & Place, 28,
153–166. doi:10.1016/j.healthplace.2014.04.008
Guagliardo, M. F. (2004). Spatial accessibility of primary care: Concept, methods and
challenges. International Journal of Health Geographics, 3, 3. doi:10.1186/1476-072X-
3-3
Harris, T. E., Winetrobe, H., Rhoades, H., & Wenzel, S. L. (2019). Mental health, substance use,
and demographic correlates of tobacco use and cessation attempts among homeless
adults. Journal of Dual Diagnosis, 15, 76–87. doi:10.1080/15504263.2019.1579947
Hastings, A. (2009). Poor neighbourhoods and poor services: Evidence on the ‘rationing’ of
environmental service provision to deprived neighbourhoods. Urban Studies, 46, 2907–
2927. doi:10.1177/0042098009344995

15

Hennrikus, D., Pirie, P., Hellerstedt, W., Lando, H. A., Steele, J., & Dunn, C. (2010). Increasing
support for smoking cessation during pregnancy and postpartum: Results of a randomized
controlled pilot study. Preventive Medicine, 50, 134–137.
doi:10.1016/j.ypmed.2010.01.003
Henwood, B. F., Cabassa, L. J., Craig, C. M., & Padgett, D. K. (2013). Permanent supportive
housing: Addressing homelessness and health disparities? American Journal of Public
Health, 103, S188–S192. doi:10.2105/AJPH.2013.301490
Henwood, B. F., Rhoades, H., Hsu, H.-T., Couture, J., Rice, E., & Wenzel, S. L. (2017). Changes
in social networks and HIV risk behaviors among homeless adults transitioning into
permanent supportive housing: A mixed methods pilot study. Journal of Mixed Methods
Research, 11, 124–137. doi:10.1177/1558689815607686
Hsu, H.-T., Simon, J. D., Henwood, B. F., Wenzel, S. L., & Couture, J. (2016). Location,
location, location: Perceptions of safety and security among formerly homeless persons
transitioned to permanent supportive housing. Journal of the Society for Social Work and
Research, 7, 65–88. doi:10.1086/685034
Hwang, S. W., Stergiopoulos, V., O’Campo, P., & Gozdzik, A. (2012). Ending homelessness
among people with mental illness: The At Home/Chez Soi randomized trial of a Housing
First intervention in Toronto. BMC Public Health, 12, 787. doi:10.1186/1471-2458-12-
787
Larimer, M. E., Malone, D. K., Garner, M. D., Atkins, D. C., Burlingham, B., Lonczak, H. S., …
Marlatt, G. A. (2009). Health care and public service use and costs before and after
provision of housing for chronically homeless persons with severe alcohol problems.

16

Journal of the American Medical Association, 301, 1349–1357.
doi:10.1001/jama.2009.414
Lee, C.-W., & Kahende, J. (2007). Factors associated with successful smoking cessation in the
United States, 2000. American Journal of Public Health, 97, 1503–1509.
doi:10.2105/AJPH.2005.083527
Macintyre, S., & Ellaway, A. (2000). Ecological approaches: Rediscovering the role of the
physical and social environment. Social Epidemiology, 9, 332–348.
Martinez, T. E., & Burt, M. R. (2006). Impact of permanent supportive housing on the use of
acute care health services by homeless adults. Psychiatric Services, 57, 992–999.
doi:10.1176/ps.2006.57.7.992
May, S., & West, R. (2000). Do social support interventions (“buddy systems”) aid smoking
cessation? A review. Tobacco Control, 9, 415–422. doi:10.1136/tc.9.4.415
May, S., West, R., Hajek, P., McEwen, A., & McRobbie, H. (2006). Randomized controlled trial
of a social support (‘buddy’) intervention f)or smoking cessation. Patient Education and
Counseling, 64, 235–241. doi:10.1016/j.pec.2006.02.008
Michie, S., Churchill, S., & West, R. (2011). Identifying evidence-based competences required
to deliver behavioural support for smoking cessation. Annals of Behavioral Medicine, 41,
59–70. doi:10.1007/s12160-010-9235-z
Okuyemi, K. S., Caldwell, A. R., Thomas, J. L., Born, W., Richter, K. P., Nollen, N., …
Ahluwalia, J. S. (2006). Homelessness and smoking cessation: Insights from focus
groups. Nicotine & Tobacco Research, 8, 287–296. doi:10.1080/14622200500494971
Perry, C. L., Baranowski, T., & Parcel, G. S. (1990). How individuals, environments, and health
behavior interact: Social learning theory. In K. Glanz, F. M. Lewis, & B. K. Rimer (Eds),

17

Health behavior and health education: Theory, research, and practice (pp. 161–186).
San Francisco, CA: Jossey-Bass.
Porter, M., Harvey, J., Gavin, J. K., Carpenter, M. J., Cummings, K. M., Pope, C., & Diaz, V. A.
(2017). A qualitative study to assess factors supporting tobacco use in a homeless
population. AIMS Medical Science, 4, 83–98. doi:10.3934/medsci.2017.1.83
Rhoades, H., Hsu, H., Harris, T., Rice, E., Wenzel, S. L., & Winetrobe, H. (2017). Social
network change after moving into permanent supportive housing: Who stays and who
goes? Manuscript under review.
Sherman, S. E., Link, A. R., Rogers, E. S., Krebs, P., Ladapo, J. A., Shelley, D. R., … Grossman,
E. (2016). Smoking-cessation interventions for urban hospital patients: A randomized
comparative effectiveness trial. American Journal of Preventive Medicine, 51, 566–577.
doi:10.1016/j.amepre.2016.06.023
Soar, K., Dawkins, L., Robson, D., & Cox, S. (2020). Smoking amongst adults experiencing
homelessness: A systematic review of prevalence rates, interventions and the barriers and
facilitators to quitting and staying quit. Journal of Smoking Cessation, 15, 94–108.
Solari, C. D., Cortes, A., Henry, M., Matthews, N., & Morris, S. (2014). The 2013 Annual
Homeless Assessment Report (AHAR) to Congress: Part 2, estimates of homelessness in
the United States. Washington, DC: U.S. Department of Housing and Urban
Development, Office of Community Planning and Development.
Stead, L. F., Buitrago, D., Preciado, N., Sanchez, G., Hartmann-Boyce, J., & Lancaster, T.
(2013). Physician advice for tobacco use cessation. Cochrane Database of Systematic
Reviews, CD000165. doi:10.1002/14651858.CD000165.pub4

18

Stead, L. F., Hartmann-Boyce, J., Perera, R., & Lancaster, T. (2013). Telephone counselling for
smoking cessation. Cochrane Database of Systematic Reviews, CD002850.
doi:10.1002/14651858.CD002850.pub3
Stead, L. F., & Lancaster, T. (2012). Behavioural interventions as adjuncts to pharmacotherapy
for smoking cessation. Cochrane Database of Systematic Reviews, CD009670.
doi:10.1002/14651858.CD009670.pub2
Stergiopoulos, V., Gozdzik, A., O’Campo, P., Holtby, A. R., Jeyaratnam, J., & Tsemberis, S.
(2014). Housing First: Exploring participants’ early support needs. BMC Health Services
Research, 14, 167. doi:10.1186/1472-6963-14-16
Stronegger, W. J., Titze, S., & Oja, P. (2010). Perceived characteristics of the neighborhood and
its association with physical activity behavior and self-rated health. Health & Place, 16,
736–743. doi:10.1016/j.healthplace.2010.03.005
Tiderington, E., Stanhope, V., & Henwood, B. F. (2013). A qualitative analysis of case
managers’ use of harm reduction in practice. Journal of Substance Abuse Treatment,
44(1), 71-77. doi:10.1016/j.jsat.2012.03.007
Tsemberis, S. (2014). Evidence-based values in housing and treatment. Psychiatric Services, 65,
277. doi:10.1176/appi.ps.650304
Tsemberis, S., Gulcur, L., & Nakae, M. (2004). Housing First, consumer choice, and harm
reduction for homeless individuals with a dual diagnosis. American Journal of Public
Health, 94, 651–656. doi:10.2105/AJPH.94.4.651
Tucker, J. S., Shadel, W. G., Golinelli, D., Mullins, L., & Ewing, B. (2015). Sniping and other
high-risk smoking practices among homeless youth. Drug and Alcohol Dependence, 154,
105–110. doi:10.1016/j.drugalcdep.2015.06.036

19

Vijayaraghavan, M., Tieu, L., Ponath, C., Guzman, D., & Kushel, M. (2016). Tobacco cessation
behaviors among older homeless adults: Results from the HOPE HOME Study. Nicotine
& Tobacco Research, 18, 1733–1739. doi:10.1093/ntr/ntw040
Visser, M. (2007). The social ecological model as theoretical framework in community
psychology. In N. Duncan, B. Bowman, A. Naidoo, J. Pillay, & V. Roos (Eds.),
Community psychology: Analysis, context and action (pp. 102–116). Cape Town, South
Africa: UCT Press.
Wang, F., & Luo, W. (2005). Assessing spatial and nonspatial factors for healthcare access:
Towards an integrated approach to defining health professional shortage areas. Health &
Place, 11, 131–146. doi:10.1016/j.healthplace.2004.02.003
Westmaas, J. L., Bontemps-Jones, J., & Bauer, J. E. (2010). Social support in smoking cessation:
Reconciling theory and evidence. Nicotine & Tobacco Research, 12, 695–707.
doi:10.1093/ntr/ntq077
Wills, T. A., & Ainette, M. G. (2012). Social networks and social support. In A. Baum, T. A.
Revenson, & J. Singer (Eds.), Handbook of health psychology (2nd ed., pp. 465–492).
New York, NY: Psychology Press.
Wright, P. A., & Kloos, B. (2007). Housing environment and mental health outcomes: A levels
of analysis perspective. Journal of Environmental Psychology, 27, 79–89.
doi:10.1016/j.jenvp.2006.12.001
Yanos, P. T., Barrow, S. M., & Tsemberis, S. (2004). Community integration in the early phase
of housing among homeless persons diagnosed with severe mental illness: Successes and
challenges. Community Mental Health Journal, 40, 133–150.
doi:10.1023/B:COMH.0000022733.12858.cb

20

Yin, R. K. (2003). Case study research: Design and methods (3rd ed.). Thousand Oaks, CA:
Sage.

21

Tables and Figures
Table 1.1. Sample Statistics (N = 421)
Variable n (%)
Demographics
Age
a
54.44 (7.53)
Male 301 (71.50)
Black 234 (55.58)
Latino 67 (15.91)
Other race 37 (8.79)
White 83 (19.71)
≤ High school education 191 (45.37)
Not heterosexual 47 (11.16)
Veteran 125 (30.00)
Homelessness
Lifetime literal (years)
a
6.02 (6.87)
Lifetime mental health conditions
Schizophrenia 117 (27.79)
PTSD 121 (28.74)
Anxiety 235 (55.82)
Depression 226 (53.68)
Bipolar disorder 129 (30.64)
Substance use (past 3 months)
Alcohol 213 (50.59)

22

Marijuana 112 (26.60)
Illicit drugs 78 (18.53)
Any tobacco 254 (60.33)
Daily tobacco 209 (49.64)
a
M (SD).

23

Figure 1.1. Guiding Conceptual Frameworks and their Relationship to Smoking Behavior

Note. The large arrows demonstrate the effect of each sociological level on the other reflecting the Socio-ecological
Model of Health Behavior, while the triangle represents the proposed associations according to Social Learning
Theory.

24

Chapter 2: Trajectories of Smoking Behavior and Change in the Transition from
Homelessness to Supportive Housing
Introduction
Adults experiencing homelessness smoke cigarettes at disproportionately higher rates
than housed persons (Baggett, Tobey, & Rigotti, 2013). In the general population of adults and
among adults experiencing homelessness, demographic and clinical characteristics are associated
with increased rates of tobacco use. Yet for the homeless population, cessation outcomes have
not been promising following intervention—specifically, randomized controlled trials using
nicotine replacement therapy and motivational interviewing or counseling and pharmacotherapy
have seen low quit rates (Carpenter et al., 2015; Okuyemi et al., 2006; Segan, Maddox, &
Borland, 2015; Shelley, Cantrell, Wong, & Warn, 2010). Individuals in this population face a
multitude of hindrances to cessation while homeless that likely affect their ability to change their
smoking. Permanent supportive housing (PSH) provides affordable housing and supportive
services to adults experiencing homelessness. Whether tobacco use changes once homeless
adults are housed, and whether these demographic and clinical characteristics affect change
trajectories, is largely unknown.
Estimates of smoking are upward of 80% for persons experiencing homelessness, nearly
4 times that of the housed population of adults (Baggett, Lebrun-Harris, & Rigotti, 2013).
Although smoking rates in the United States have declined during the last 50 years, rates of
smoking among persons experiencing homelessness have remained alarmingly high (Centers for
Disease Control and Prevention, 2017). Persons experiencing homelessness desire to quit
smoking (Baggett, Lebrun-Harris, et al., 2013), and research has indicated they attempt cessation
at rates comparable to the general population of tobacco users (Businelle, Cuate, Kesh,

25

Poonawalla, & Kendzor, 2013), yet cessation outcomes are an ongoing challenge for this
population and are distinctly lower than that of domiciled persons (Businelle et al., 2014;
Okuyemi et al., 2013; Rash, Petry, & Alessi, 2018; Segan et al., 2015).
Homelessness and its associated stressors have been identified as contributors to high
rates of smoking and barriers to cessation (Chen, Nguyen, Malesker, & Morrow, 2016; Porter et
al., 2017). The prevalence of tobacco use among adults experiencing homelessness is often
attributed to high rates of substance use and poor mental health. Several studies have found that
mental health conditions and substance use behavior are associated with smoking among
homeless populations (Baggett, Tobey, et al., 2013; Harris, Winetrobe, Rhoades, & Wenzel,
2019; Tsai & Rosenheck, 2012; Vijayaraghavan, Tieu, Ponath, Guzman, & Kushel, 2016).
Although mental health conditions and substance use are associated with increased odds of
smoking for homeless adults, several studies have also found that the presence of some mental
health conditions facilitates cessation attempts for this population (Harris, Winetrobe, et al.,
2019; Vijayaraghavan et al., 2016).
Despite the extant vulnerability of adults experiencing homelessness, not much attention
has been paid to demographic variation in risk of tobacco use in this population. There is,
however, substantial support in the literature on the general population of tobacco users to
suggest that rates of tobacco use and cessation outcomes vary across demographic characteristics
(Creamer et al., 2019). For example, in nationally representative data, smoking rates are highest
among adults aged 25–44 years and 45–64 years (Creamer et al., 2019), and older adults are less
likely to cease smoking than younger adults (Messer, Trinidad, Al-Delaimy, & Pierce, 2008). In
a review of 17 papers examining differences in quit attempts across race, more than half
indicated that African American or Black identifying adults were more likely than Caucasian or

26

White identifying adults to attempt to quit smoking, whereas Caucasian or White identifying
adults demonstrated better longitudinal abstinence rates than African American or Black
identifying adults (Kulak, Cornelius, Fong, & Giovino, 2016). Although these demographic
differences may be less prominent or understudied in the homeless adult population, some recent
research has suggested that demographic characteristics including race (Pinsker et al., 2018) and
education (Harris, Winetrobe, et al., 2019) may be associated with tobacco use and cessation
behavior in this population.
Housing is a potential facilitator of cessation. Housing is often a catalyst for health
recovery and improved quality of life for persons experiencing homelessness (Henwood,
Matejkowski, Stefancic, & Lukens, 2014; Leff et al., 2009). Positive improvements in physical
and mental health have been found following adults’ entry into PSH (Fitzpatrick-Lewis et al.,
2011; Harris, Rhoades, Duan, & Wenzel, 2019), which may affect tobacco use and cessation
outcomes; however, little is known about how housing affects tobacco use behavior. To the
author’s knowledge, only one study has examined changes in tobacco use in PSH, finding no
significant changes after 12 months (Tsai & Rosenheck, 2012). Ultimately, more longitudinal
research is needed to understand the impact of housing on tobacco use behavior over time,
whether there are differential patterns of smoking as adults experiencing homelessness transition
to PSH, and what factors are associated with varying tobacco use trajectories, making this a rich
and rare opportunity to examine these relationships for this highly vulnerable population.
The current study is the first to use repeated measures of tobacco use frequency to
identify patterns of tobacco use behavior in the transition from homelessness to housing.
Specifically, this study draws on four waves of longitudinal tobacco use data from a cohort of
adults who were homeless at baseline, across their initial year in PSH in Los Angeles, using a

27

repeated-measures latent class analysis approach (RMLCA). A RMLCA facilitates the ability to
group tobacco-using participants based on their shared smoking behavior trajectories and
identify demographic and clinical characteristics associated with membership of a given
subgroup, referred to as a latent class (Collins & Lanza, 2010). Therefore, this study also
examined demographic (age, gender, veteran status, education, race and ethnicity, and lifetime
history of homelessness); clinical (psychological functioning and lifetime mental health
diagnoses [schizophrenia, PTSD, bipolar disorder, depression, and anxiety]); substance use
(cannabis, binge drinking, and illicit substance use); and cessation as correlates of class
membership.
Although this is the initial study to examine these relationships among this population,
several hypotheses were developed based on prior literature: (1) Given the findings of Tsai and
Rosenheck (2012), the author hypothesized that there will be a latent class of individuals who did
not reduce their tobacco use over time; (2) based on studies with adults experiencing
homelessness that have found specific mental health conditions (i.e., schizophrenia, bipolar
disorder, and PTSD) and illicit substance are associated with high rates of smoking and are
barriers to cessation, the author hypothesized that these conditions will be associated with a
latent class that did not reduce smoking; and (3) given two studies of adults experiencing
homelessness that have found a major depression diagnosis is associated with increased cessation
attempts (Harris, Winetrobe, et al., 2019; Vijayaraghavan et al., 2016), the author hypothesized
that depression will be associated with a latent class that decreases tobacco use over time, if such
a subgroup is identified. Understanding differential tobacco use trajectories that occur as adults
experiencing homelessness transition to supportive housing and identifying characteristics that

28

drive differential patterns of change may be critical for targeting subpopulations and tailoring
efforts to overcome barriers inhibiting cessation and improve cessation rates for this population.
Methods
Data for the current study were drawn from the Transitions to Housing Study (NIDA
Grant No. R01DA036345), a longitudinal prospective cohort study designed to understand
clinical, behavioral, and social network outcomes in the transition from homelessness to
supportive housing in Los Angeles. Participants were recruited between 2014 and 2016 via
collaborations with 26 agencies that constitute the majority of supportive housing providers in
Los Angeles County. Participants were eligible if they were 39 years old or older, spoke English
or Spanish, and were currently homeless unaccompanied adults (without minor children).
Because the primary focus of the Transitions to Housing Study was HIV risk and prevention, the
age and nonparenting requirements were implemented to maximize the ability to detect changes
in HIV risk outcomes by minimizing variability due to developmental life stage or current
parenting status. Given the emphasis of PSH on addressing chronic homelessness and
accompanying health conditions, individuals in PSH increasingly tend to be older individuals not
living with children (U.S. Department of Housing and Urban Development, 2015). Persons
experiencing homelessness are typically placed into supportive housing in L.A. County through a
housing-, social services-, or hospital-appointed staff member. The Vulnerability Index Service
Prioritization Decision Assistance Tool is used to assess health and risk status and prioritizes
placement of the most vulnerable persons into supportive housing (United Way of Greater Los
Angeles, 2016).
At baseline, while experiencing homelessness, 421 adults were recruited to participate in
quantitative and social network surveys that took 1–1.5 hours to complete and were led by

29

trained interviewers using iPads. Follow-up surveys were conducted at 3 (n = 405; 96%
retention), 6 (n = 400; 95% retention), and 12 (n = 383; 91% retention) months posthousing.
Interviewers maintained contact with participants between interviews via monthly phone calls.
Although participants who were not interviewed at a midpoint wave were eligible for later
interviews, there was attrition due to death, incarceration, withdrawal, and loss of contact. This
study was restricted to a slightly smaller sample of participants who used tobacco (n = 321, 76%)
at one or more waves of data collection. Participants completed written informed consent prior to
their participation. All study procedures were approved by the University of Southern
California’s Institutional Review Board. Additionally, the study received a certificate of
confidentiality from the U.S. Department of Health and Human Services to protect participant
data from subpoena.
Measures
Tobacco frequency. Past-3-month smoking frequency was used as the repeated indicator
of latent class in the RMLCA. Participants’ smoking frequency was assessed at each wave using
the National Institute on Drug Abuse-Modified Alcohol, Smoking and Substance Involvement
Screening Test (NIDA-Modified ASSIST; National Institute on Drug Abuse, 2012). Respondents
reported whether they used each substance never, once or twice, monthly, weekly, or daily or
almost daily in the past 3 months. The NIDA-Modified ASSIST has been well regarded for its
accuracy of self-reported substance use (Bailey, 2018), good to excellent sensitivity and
specificity (Zgierska, Amaza, Brown, Mundt, & Fleming, 2014), and excellent test–retest
reliability (McNeely et al., 2014).
Tobacco cessation. A separate question on the NIDA-Modified ASSIST asked
participants who endorsed using tobacco whether they had “tried and failed to control, cut down,

30

or stop using tobacco.” Response options included no; yes, but not in the past 3 months; and yes,
in the past 3 months. An indicator of past-3-month cessation attempt versus no attempt was
created for each wave.
Age. Age was captured at baseline.
Race and ethnicity. A combined measure of race and ethnicity was created for analyses,
leading to four subgroups: Black or African American; White or Caucasian; other (Native
American, Alaska Native, Asian, Native Hawaiian, Pacific Islander, or multiracial); and Hispanic
or Latino (Wenzel, 2005).
Gender. Response options for gender on the quantitative survey included male, female,
and male-to-female transgender. Four participants identified as male-to-female transgender and
were subsequently combined with the female subgroup due to size.
Education. A variable captured participants’ highest level of education (California
Health Interview Survey, 2011). Each “grade” after high school was equivalent to 1 year of
education following high school (e.g., vocational, trade school, any college).
Veteran status. A dichotomous measure of veteran status indicated whether participants
had served in any branch of the U.S. military.
Literal homelessness. Participants’ lifetime history of unsheltered homelessness (i.e.,
staying outside on the streets, park, or beach; in a shelter; in an abandoned building, garage, or
shed; in an indoor public place; in a car, truck, van, or recreational vehicle; or on a bus, subway,
or train) prior to entering supportive housing was included (National Center for Health Statistics,
2016). The number of years of unsheltered homelessness was used as the indicator in analyses.

31

Income. Past-30-day income from all sources (jobs, tips, welfare, general relief,
governmental assistance, and any illegal activities or under-the-table payments) was assessed
across each wave.
Diagnosed mental health conditions. At baseline, participants were asked about their
lifetime histories of diagnosed chronic mental health conditions (i.e., schizophrenia, major
depressive disorder, bipolar disorder, anxiety, PTSD). This measure drew from an adapted item
in the National Health Interview Survey (National Center for Health Statistics, 2016) and
response options as informed by prior literature (Bassuk, Buckner, Perloff, & Bassuk, 1998;
Hwang, 2001; Wenzel, 2009).
Current psychiatric disability. At baseline and all follow-ups, participants’ past-month
overall psychological functioning was assessed using the Modified Colorado Symptom Index
(Conrad et al., 2001). A cutoff score of 16 indicated the presence of a psychiatric disability
(Boothroyd & Chen, 2008; Conrad et al., 2001).
Illicit substance use. The NIDA-Modified ASSIST was used to determine participants’
use of cannabis and illicit substances (cocaine, methamphetamine, hallucinogens, and street
opioids [i.e., heroin, opium]) at each wave (National Institute on Drug Abuse, 2012). Misuse of
prescription stimulants, sedatives, or opioids was determined by items adapted from prior
research (Al-Tayyib, Rice, Rhoades, & Riggs, 2014; Rice, 2011). Participants who endorsed use
of these prescriptions were asked a follow-up question about misuse of each type of prescription.
For analyses, illicit substances from the NIDA-Modified ASSIST were combined with the
prescription misuse measures to create an indicator of any illicit substance use.
Cannabis use. Determination of residents’ cannabis use also drew from the NIDA-
Modified ASSIST and was dichotomized to any past-3-month use and no use.

32

Binge drinking. An item adapted from the National Institute on Alcohol Abuse and
Alcoholism (2003) was used to assess past-3-month binge drinking at each survey. Binge
drinking was defined as consuming at least four (for women) or five (for men) alcoholic drinks
during a 2-hour period (National Institute on Alcohol Abuse and Alcoholism, 2003).
Respondents who reported any past-3-month binge drinking were considered positive.
Analysis
An RMLCA was used to determine subgroups that reflected distinct patterns of smoking
change in the transition from homelessness to housing. Tobacco use frequency (never, once or
twice, monthly, weekly, or daily or almost daily) at each wave (baseline and 3, 6, and 12
months), totaling four variables, served as the item indicators for the RMLCA. The analysis was
restricted to participants who reported using tobacco at least once across the four surveys, due to
interest in identifying trajectories for tobacco users. This resulted in 321 participants (76.24%)
who served as the analytic subgroup. To determine the number of subgroups, or classes, that
reflect variations in smoking behavior over time, estimations for one to four classes were
conducted. Determination of the best model fit was based on several fit indexes including the
Akaike information criterion (AIC), Bayesian information criteria (BIC), Lo-Mendell-Rubin
likelihood ratio test, and entropy and interpretability of latent classes. All models and model
comparisons were conducted using Mplus version 8.0.
An RMLCA is ideal for these data and the aims of this study for several reasons. An
RMLCA can garner new information about how tobacco use changes, by grouping participants
who share similar patterns, and the proportion of persons in each pattern (Collins & Lanza,
2010). An RMLCA allows for identification of latent patterns of categorical items—for studies
of behavior change over time, this approach determines behavioral trends and an opportunity to

33

understand what factors contribute to each pattern of change. Latent class modeling is considered
a person-centered approach, which can identify differential patterns of behavior and is
particularly valuable compared with variable-centered approaches such as mixed-effects
modeling that group persons based on a grouped set of independent and dependent variables
(Collins & Lanza, 2010). This study examined changes in tobacco use over time using a mixed-
effects model with a logit function to determine whether time had a significant effect on changes
in tobacco frequency and to demonstrate the benefit of using an RMLCA. Using the tobacco
frequency item measured repeatedly across four interviews allowed for understanding of the
complexity of responses and clustered them for ease of interpretability (Collins & Lanza, 2010).
The RMLCA provided the proportions expected for each class, resting on the assumption of
conditional independence; that is, the latent variable explains all variation between the repeated
measures of tobacco frequency used to identify each class. Results from model comparisons for
this study indicated the model with three latent classes was the best fit for these data.
Following the identification of class membership, the relationship between
demographics, clinical characteristics, and cessation attempts with class membership were
examined. Baseline demographic variables (age, race and ethnicity, gender, education, literal
homelessness, and veteran status) and lifetime mental health diagnoses (schizophrenia, PTSD,
bipolar disorder, depression, and anxiety) were examined using multinomial logistic regression.
To account for time-variant predictors (income, psychological functioning, substance use, and
cessation) of class membership, distributed lag models were used. A distributed lag model is
ideal for time-series data, wherein time-varying measures are regressed to predict a multinomial
logistic outcome, because it accounts for time and change in explanatory variables to be captured
and fits the measure to the person level, allowing it to predict the person-level latent class

34

outcomes (Gasparrini & Leone, 2014). Each time-variant measure at baseline (T0) was modeled
simultaneously with a change score that captured differences between baseline and scores of that
variable at the final follow-up (T1; 12 months postbaseline). Each independent variable was
assessed individually in univariable models, with class membership levels serving as the
multinomial categorical dependent variable. Comparisons were made across each potential
pairing of latent classes (one vs. two, one vs. three, and two vs. three). All univariable models,
including distributed lag models with associations significant at p < .10 were subsequently
analyzed in multivariable models, which included all demographic control variables (regardless
of statistical significance in univariable models). In controlled models, independent variables
were considered statistically significant at p < .05. Clinical and substance use measures were
examined in separate multivariable models due to comorbidity and correlation. All multinomial
logistic regression models were conducted using Stata version 16.
Results
Participants
Across all participants who reported smoking during the study (N = 321), the average age
was 54.54 years (SD = 7.21; range: 39–79). The sample was predominantly male (74.14%), and
more than half of the sample identified as African American or Black (57.01%). Approximately
a third of the sample (31.15%) had formerly military service experience and on average,
residents had received 1 year of education after high school (SD = 3.68). Participants had
experienced an average of 6.38 years of literal homelessness across their lifetime. The most
prevalent mental health diagnosis reported was depression (57.01%), with more than half of the
sample endorsing receipt of this diagnosis by a doctor. Regarding time-varying measures, at
baseline, the total sample reported an average of $565.41 (SD = $352.92) in monthly income and

35

an average of $627.07 (SD = $367.28) at 12 months. At baseline, 222 (69.16%) participants were
positive for a psychiatric disability, whereas 150 (52.26%) were positive at the 12-month follow-
up. Baseline rates of cannabis use (31.97%), hard drug use (22.12%), and binge drinking
(16.82%) were relatively consistent across the study period, with the exception of an uptick in
cannabis (38.46%) and binge drinking (20%) 6 months posthousing, whereas hard drug use
decreased at this follow-up (18%). Cessation attempts were reported by 39.09% of the sample at
baseline, and that rate remained relatively consistent over time. See Table 2.1 for descriptive
statistics across the sample.
RMLCA
An assessment of fit indexes, model parameters of probability of class membership, and
gamma parameters of latent class prevalence revealed that a three-class model best fit the data.
Classes were labeled and described as (1) mixed-frequency decreasers, for those who used
tobacco at differing frequencies over time while demonstrating an overall trend of decreasing
tobacco use over time (n = 69); (2) high-frequency maintainers, for those who were
predominantly daily or almost daily smokers and remained so over time; and (3) low-frequency
increasers, for the subset of tobacco users who were generally nonsmokers at baseline, but
significantly increased their smoking to monthly or weekly consumption over time. A summary
of model indexes for the three-class model, including model fit and selection criteria, is reported
in Table 2.2, along with a comparison of models. Figure 2.1 depicts the linear trajectories of
smoking across each latent class. Rates of tobacco use, along with all descriptive statistics of
independent variables, and cessation across the three latent classes can be found in Table 2.1.

36

Multinomial Logistic Regression: Univariable Models
Comparisons across latent classes revealed several differences across demographic
characteristics (See Table 2.3). Compared to high-frequency maintainers, mixed-frequency
decreasers were more than 3 times as likely to identify as Latino (p < .01). Compared to low-
frequency increasers, high-frequency maintainers were 46% less likely to be White or Caucasian
(p = .08), 59% less likely to identify as a race that fell in the other racial category (p = .09), and
61% less likely to identify as Latino (p = .04). Regarding lifetime mental health diagnoses,
mixed-frequency decreasers had 36% lower odds of having a diagnosis of PTSD (p = .09) and
53% lower odds of having a diagnosis of bipolar disorder (p = .01) compared to high-frequency
maintainers. The odds of having a diagnosis of schizophrenia were 2 times greater (p = .03) for
high-frequency maintainers compared to low-frequency increasers. High-frequency maintainers
also had 2 times greater odds (p = .01) of having a diagnosis of bipolar disorder than low-
frequency increasers. There were no statistically significant differences in demographics or
mental health diagnoses between mixed-frequency decreasers and low-frequency increasers.
Several time-variant measures significantly changed over time. Compared to baseline,
there was an approximate increase of $62 in past-month income at the 12-month interview (p <
.01). Participants had 62% lower odds of a positive screening for a psychiatric disability 3
months after receiving housing (p < .01), 67% lower odds at 6 months (p < .01), and 81% lower
odds at 12 months (p < .01), compared to baseline. The odds of using cannabis were 2.03 times
greater at the 6-month posthousing follow-up (p = .02) and nearly 1.83 times greater at 12
months (p = .04), compared to baseline, whereas the odds of using hard drugs were 45% lower at
6 months (p = .04). At 12-months posthousing, the odds of attempting cessation were 33% lower
than at baseline (p = .05). In the distributed lagged models, past-3-month cannabis use at baseline

37

was associated with 2.44 greater odds (p = .07) of being a mixed-frequency decreaser and 2.38
greater odds (p = .05) of being a high-frequency maintainer when compared to low-frequency
increasers. A change in cannabis use over time was associated with 2.97 greater odds (p = .03) of
being a member of the mixed-frequency decreaser class and 3.10 greater odds (p < .01) of being
a high-frequency maintainer when compared to low-frequency increasers. Binge drinking at
baseline was associated with 3.80 greater odds (p = .05) of being a mixed-frequency decreaser
compared to low-frequency increasers. Due to small cell size, low-frequency increasers were
omitted from the model examining cessation as a predictor of class membership. However, there
were no significant differences between high-frequency maintainers and mixed-frequency
decreasers.
Multinomial Logistic Regression: Multivariable Models
Most measures that were significant in univariable models were significant in
multivariable models with the expectation of schizophrenia and PTSD. Race remained
significant, however only Hispanic or Latino participants significantly differed from the Black or
African American subgroup. That is, compared to high-frequency maintainers, the odds of being
Hispanic or Latino were 7.90 times greater for the mixed-frequency decreasers (p < .01).
Compared to low-frequency increasers, the odds of being Hispanic or Latino were 15% lower for
the high-frequency maintainers (p = .02). See Table 2.4 for results from multivariable models.
Discussion
This study identified trajectories of tobacco use frequency among adults experiencing
homelessness during their initial year transitioning to PSH in Los Angeles. Of three identified
classes, two classes were characterized by behavior change, with one class increasing its
smoking and one class decreasing its smoking over time, whereas the third class, composed of

38

daily smokers, maintained their smoking behavior over time. When aggregated, the sample
demonstrated significant decreases in tobacco use over time. These findings differ from the study
by Tsai and Rosenheck (2012), which did not find significant change in smoking in a cohort of
veterans in a supportive housing program. Although the overall findings of smoking change over
time may be distinct from that of Tsai and Rosenheck (2012), they nonetheless highlight the
importance of using an RMLCA, because it allowed these trends in smoking change and stability
to be parsed, including the identification of two integral target groups of residents: one that may
require cessation interventions in the nascent months of housing to reduce further exacerbation
of smoking, and a particularly high-risk group of daily smokers who maintained their tobacco
use over time.
An RMLCA also allowed for examination of factors associated with class membership,
which bolstered the ability to identify characteristics that define each class. The only
demographic characteristic that emerged as a correlate of class membership was race and
ethnicity. The low-frequency increasers and mixed-frequency decreasers had more representation
of Latino-identifying participants when compared to Black-identifying participants. In the
general population of adult smokers, 15% of non-Hispanic Black-identifying adults are smokers,
whereas 10% of Hispanic or Latino persons are smokers (Creamer et al., 2019). National and
state data suggest Hispanic or Latino adults are more likely to be light smokers than other racial
groups (Bandiera, Assari, Livaudais-Toman, & Pérez-Stable, 2016; Pulvers et al., 2015).
Although the mechanisms underlying these differences are less known, collectivism and cultural
strength may be protective in Hispanics and Latinos’ smoking behavior, as it has been for other
health behaviors (Ai, Aisenberg, Weiss, & Salazar, 2014; Elder, Ayala, Parra-Medina, &
Talavera, 2009). This sense of community among Latino communities has also been protective

39

against experiences of discrimination, which has been tied to increased smoking frequency
among racial minorities (Brondolo et al., 2015). However, the greater proportions of Latino
residents in classes that increased or decreased smoking may suggest that other factors not
examined in this study may be contributing to these distinct trends. Future research would
benefit from a deeper qualitative exploration of the role of race and ethnicity in homeless adults’
smoking behavior and cessation.
Poor mental health is believed to be one of the primary reasons for the high rates of
tobacco use among persons experiencing homelessness. Several studies have found that some
mental health conditions, specifically those that may be more severe and characterized by
psychotic features or chronic mood dysregulation, were associated with increased rates of
tobacco use (Baggett, Tobey, et al., 2013; Harris, Winetrobe, et al., 2019; Tsai & Rosenheck,
2012; Vijayaraghavan et al., 2016). In this study, bipolar disorder emerged as a correlate of class
membership, indicating that the high-frequency maintainer class had significantly more members
with bipolar disorder than the two lower-risk classes. Smokers with bipolar disorder have higher
rates of suicidal ideation and substance abuse, more severe mental health conditions, and
lengthier hospital stays (Dodd et al., 2010; Ostacher et al., 2006), making this a particularly
concerning subgroup of tobacco users. These mental health and risk comorbidities can and
should be incorporated into treatment. In fact, studies that have combined mental health
treatment components into cessation interventions have proven to be much more effective in
promoting long-term abstinence (Prochaska, 2011). It is important to note that although some
aspects of mental health can make cessation more challenging, the presence of a mental health
condition does not diminish this population’s readiness to quit or attempts at cessation
(Prochaska, 2011).

40

Across the total sample, residents increased their cannabis use over time and this change
was associated with a greater likelihood of being a mixed-frequency decreaser or high-frequency
maintainer. Given both classes demonstrated this trend, this is likely indicative of the tendency
for more frequent tobacco users to also use cannabis more frequently (Agrawal, Budney, &
Lynskey, 2012; Hindocha, Freeman, Ferris, Lynskey, & Winstock, 2016), rather than the mixed-
frequency decreasers substituting cannabis for tobacco. Nonetheless, this is worth exploring in
future research, because decreased hard drug use may be at play in this relationship and may
suggest that some level of harm reduction is involved in these changes. The relationship between
cannabis and tobacco is a complex one in that it can vary across environments, with a systematic
review finding that routes of administration vary substantially across geographic regions
(Agrawal et al., 2012). Compared to European countries, Americans are significantly less likely
to use cannabis joints with tobacco as a route of administration, which is important because use
of a nontobacco route of administration of cannabis is associated with a substantial increase in
the likelihood to desire to and intend to quit using tobacco (Hindocha et al., 2016). Additionally,
cost and availability of one of these substances can affect use of the other (Agrawal et al., 2012).
Data for the present study were collected prior to the increase in cigarette taxation that went into
effect in 2017 in Los Angeles; however, cannabis dispensaries had become more commonplace,
even prior to the legalization of cannabis in 2018 (Goldstein & Sumner, 2019). Moreover, e-
cigarette use and noncombustible methods of cannabis use were not widely popular during the
study observation period and may affect current tobacco use trajectories for this population. In
other states where legalization and recreational cannabis use has been established, increased rates
of tobacco and cannabis use have been observed (Wang & Cataldo, 2016). For a high-risk
population like PSH residents, cannabis use can often seem like a secondary concern to the high

41

rates of illicit substance use. However, in the context of improving cessation outcomes, more
attention to this relationship may be warranted.
This study also examined associations between cessation attempts and class membership.
Among the classes that were compared, high-frequency maintainers and mixed-frequency
decreasers, no statistically significant differences existed in cessation attempts. Although this
may seem contradictory given that mixed-frequency decreasers were characterized by a decrease
in tobacco use, this is likely an indication that the two classes were attempting cessation at equal
rates. Cessation attempts, although a potential indicator of motivation to cease smoking, are often
an indicator of future cessation attempts (Zhou et al., 2009). In contrast, reductions in smoking
have demonstrated associations with future cessation (Hyland et al., 2005). Only two indicators
significantly differentiated these groups, which may suggest future research should explore
factors beyond individual characteristics, including environmental factors, associated with class
membership.
Limitations
This study comes with some limitations. Although the proportion of smokers in this
sample was rather large the smallest latent class was unable to be compared to the two larger
latent classes due to small cell size. Tobacco use frequency was reported using a Likert scale,
and the use of a different measure of tobacco use frequency such as number of cigarettes per day,
may result in different latent classes. All mental health and substance use measures were self-
reported which may increase bias, although this population’s self-reported health information,
including mental health symptomatology, has demonstrated high internal consistency and
healthcare utilization has demonstrated good concurrent validity, even among chronically
homeless persons with risky substance use (Klinkenberg et al., 2002; Conrad et al., 2001).

42

Additionally, these data draw from residents’ living in Los Angeles, and environmental factors
not explored in this study may affect residents’ tobacco use differentially across other regions.
Conclusion
Our findings suggest that housing may facilitate reductions in tobacco use among some
subgroups of PSH residents and that racial, clinical, and substance use factors are associated with
differential smoking frequency trajectories, demonstrating the value of this person-centered
analysis. Although this was the first study to find tobacco use declined from homelessness to
housing, the use of an RMLCA demonstrated that this approach was applicable to a portion of
PSH residents and that the largest portion of tobacco users were daily smokers and maintained
their use over time. Clinical trials and comparative treatment studies have faced challenges to
improve cessation outcomes among adults experiencing homelessness; therefore, this study may
offer hope in achieving this population’s desire to quit smoking. Findings from this study
highlight the great need for cessation interventions to be implemented in supportive housing
programs. Efforts that address mental health symptoms, tobacco use, and co-occurring substance
use are warranted. This includes navigating ways to manage mental health symptomatology that
serves as a barrier and incorporating tobacco use into harm reduction practices. Additionally,
tailoring cessation interventions to incorporate racial and cultural factors that relate to smoking
and affect cessation may be beneficial to advance much-needed improvement in cessation
outcomes for this population.

43

References
Agrawal, A., Budney, A. J., & Lynskey, M. T. (2012). The co‐occurring use and misuse of
cannabis and tobacco: A review. Addiction, 107, 1221–1233. doi:10.1111/j.1360-
0443.2012.03837.x
Ai, A. L., Aisenberg, E., Weiss, S. I., & Salazar, D. (2014). Racial/ethnic identity and subjective
physical and mental health of Latino Americans: An asset within? American Journal of
Community Psychology, 53, 173–184. doi:10.1007/s10464-014-9635-5
Al-Tayyib, A. A., Rice, E., Rhoades, H., & Riggs, P. (2014). Association between prescription
drug misuse and injection among runaway and homeless youth. Drug and Alcohol
Dependence, 134, 406–409. doi:10.1016/ j.drugalcdep.2013.10.027
Baggett, T. P., Lebrun‐Harris, L. A., & Rigotti, N. A. (2013). Homelessness, cigarette smoking
and desire to quit: Results from a US national study. Addiction, 108, 2009–2018.
doi:10.1111/add.12292
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.
doi:10.1056/NEJMp1301935
Bailey, B. (2018). By when do I have to quit? Could I just cut down some? Evidence to help
inform difficult questions in pregnancy smoking interventions. American Journal of
Obstetrics & Gynecology, 218, S137–S138. doi:10.1016/j.ajog.2017.10.085
Bandiera, F. C., Assari, S., Livaudais-Toman, J., & Pérez-Stable, E. J. (2016). Latino and Black
smokers in the Health and Retirement Study are more likely to quit: The role of light
smoking. Tobacco Induced Diseases, 14, 23. doi:10.1186/s12971-016-0090-y

44

Bassuk, E. L., Buckner, J. C., Perloff, J. N., & Bassuk, S. S. (1998). Prevalence of mental health
and substance use disorders among homeless and low-income housed mothers. American
Journal of Psychiatry, 155, 1561–1564. doi:10.1176/ajp.155.11.1561
Boothroyd, R. A., & Chen, H. J. (2008). The psychometric properties of the Colorado Symptom
Index. Administration and Policy in Mental Health and Mental Health Services
Research, 35, 370–378. doi:10.1007/ s10488-008-0179-6
Brondolo, E., Monge, A., Agosta, J., Tobin, J. N., Cassells, A., Stanton, C., & Schwartz, J.
(2015). Perceived ethnic discrimination and cigarette smoking: Examining the
moderating effects of race/ethnicity and gender in a sample of Black and Latino urban
adults. Journal of Behavioral Medicine, 38, 689–700. doi:10.1007/s10865-015-9645-2
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. doi:10.2105/AJPH.2013.301336
Businelle, M. S., Kendzor, D. E., Kesh, A., Cuate, E. L., Poonawalla, I. B., Reitzel, L. R., …
Wetter, D. W. (2014). Small financial incentives increase smoking cessation in homeless
smokers: A pilot study. Addictive Behaviors, 39, 717–720.
doi:10.1016/j.addbeh.2013.11.017
California Health Interview Survey. (2011). Adult questionnaire, version 10.3. Los Angeles, CA:
UCLA Center for Health and Policy Research.
Carpenter, V. L., Hertzberg, J. S., Kirby, A. C., Calhoun, P. S., Moore, S. D., Dennis, M. F., …
Beckham, J. C. (2015). Multi-component smoking cessation treatment including mobile
contingency management for smoking cessation in homeless veteran smokers. Journal of
Clinical Psychiatry, 76, 959–964. doi:10.4088/JCP.14m09053

45

Centers for Disease Control and Prevention. (2017). State Tobacco Activities Tracking &
Evaluation (STATE) system: Map of current cigarette use among adults. Retrieved from
https://www.cdc.gov/statesystem/cigaretteuseadult.html
Chen, J. S., Nguyen, A. H., Malesker, M. A., & Morrow, L. E. (2016). High-risk smoking
behaviors and barriers to smoking cessation among homeless individuals. Respiratory
Care, 61, 640–645. doi:10.4187/respcare.04439
Collins, L. M., & Lanza, S. T. (2010). Latent class and latent transition analysis: With
applications in the social, behavioral, and health sciences. New York, NY: John Wiley
& Sons.
Conrad, K. J., Yagelka, J. R., Matters, M. D., Rich, A. R., Williams, V., & Buchanan, M. (2001).
Reliability and validity of a modified Colorado Symptom Index in a national homeless
sample. Mental Health Services Research, 3, 141–153. doi:10.1023/A:1011571531303
Creamer, M. R., Wang, T. W., Babb, S., Cullen, K. A., Day, H., Willis, G., … Neff, L. (2019).
Tobacco product use and cessation indicators among adults—United States, 2018.
Morbidity and Mortality Weekly Report, 68(45), 1013–1019.
Dodd, S., Brnabic, A. J., Berk, L., Fitzgerald, P. B., de Castella, A. R., Filia, S., … Kulkarni, J.
(2010). A prospective study of the impact of smoking on outcomes in bipolar and
schizoaffective disorder. Comprehensive Psychiatry, 51, 504–509.
doi:10.1016/j.comppsych.2009.12.001
Elder, J. P., Ayala, G. X., Parra-Medina, D., & Talavera, G. A. (2009). Health communication in
the Latino community: Issues and approaches. Annual Review of Public Health, 30, 227–
251. doi:10.1146/annurev.publhealth.031308.100300

46

Fitzpatrick-Lewis, D., Ganann, R., Krishnaratne, S., Ciliska, D., Kouyoumdjian, F., & Hwang, S.
W. (2011). Effectiveness of interventions to improve the health and housing status of
homeless people: a rapid systematic review. BMC public health, 11(1), 638.
Gasparrini, A., & Leone, M. (2014). Attributable risk from distributed lag models. BMC Medical
Research Methodology, 14, 55. doi:10.1186/1471-2288-14-55
Goldstein, R., & Sumner, D. (2019). California cannabis regulation: An overview. California
Agriculture, 73, 101–102. doi:10.3733/ca.2019a0021
Harris, T., Rhoades, H., Duan, L., & Wenzel, S. L. (2019). Mental health change in the transition
to permanent supportive housing: The role of housing and social networks. Journal of
Community Psychology, 47, 1834–1849. doi:10.1002/jcop.22230
Harris, T., Winetrobe, H., Rhoades, H., & Wenzel, S. (2019). The role of mental health and
substance use in homeless adults’ tobacco use and cessation attempts. Journal of Dual
Diagnosis, 15, 76–87. doi:10.1080/15504263.2019.1579947
Henwood, B. F., Matejkowski, J., Stefancic, A., & Lukens, J. M. (2014). Quality of life after
housing first for adults with serious mental illness who have experienced chronic
homelessness. Psychiatry Research, 220, 549–555. doi:10.1016/j.psychres.2014.07.072
Hindocha, C., Freeman, T. P., Ferris, J. A., Lynskey, M. T., & Winstock, A. R. (2016). No
smoke without tobacco: A global overview of cannabis and tobacco routes of
administration and their association with intention to quit. Frontiers in Psychiatry, 7, 104.
doi:10.3389/fpsyt.2016.00104
Hwang, S. W. (2001). Homelessness and health. Canadian Medical Association Journal, 164,
229–233.

47

Hyland, A., Levy, D. T., Rezaishiraz, H., Hughes, J. R., Bauer, J. E., Giovino, G. A., &
Cummings, K. M. (2005). Reduction in amount smoked predicts future cessation.
Psychology of Addictive Behaviors, 19, 221–225. doi:10.1037/0893-164X.19.2.221
Klinkenberg, W. D., Calsyn, R. J., Morse, G. A., McCudden, S., Richmond, T. L., & Burger, G.
K. (2002). Consistency of recall of sexual and drug-using behaviors for homeless persons
with dual diagnosis. AIDS and Behavior, 6(4), 295-307.
Kulak, J. A., Cornelius, M. E., Fong, G. T., & Giovino, G. A. (2016). Differences in quit
attempts and cigarette smoking abstinence between whites and African Americans in the
United States: Literature review and results from the International Tobacco Control US
Survey. Nicotine & Tobacco Research, 18, S79–S87. doi:10.1093/ntr/ntv228
Leff, H. S., Chow, C. M., Pepin, R., Conley, J., Allen, I. E., & Seaman, C. A. (2009). Does one
size fit all? What we can and can’t learn from a meta‐analysis of housing models for
persons with mental illness. Psychiatric Services, 60, 473–482.
doi:10.1176/appi.ps.60.4.473
McNeely, J., Strauss, S. M., Wright, S., Rotrosen, J., Khan, R., Lee, J. D., & Gourevitch, M. N.
(2014). Test–retest reliability of a self-administered Alcohol, Smoking and Substance
Involvement Screening Test (ASSIST) in primary care patients. Journal of Substance
Abuse Treatment, 47, 93–101. doi:10.1016/j.jsat.2014.01.007
Messer, K., Trinidad, D. R., Al-Delaimy, W. K., & Pierce, J. P. (2008). Smoking cessation rates
in the United States: A comparison of young adult and older smokers. American Journal
of Public Health, 98, 317–322. doi:10.2105/AJPH.2007.112060
National Center for Health Statistics. (2016). 2014 National Health Interview Survey. Retrieved
from https://www.cdc. gov/nchs/nhis/data-questionnaires-documentation.htm

48

National Institute on Alcohol Abuse and Alcoholism. (2003). Recommended alcohol questions.
Retrieved from https://www.niaaa.nih.gov/research/guidelines-and
resources/recommended-alcohol-questions
National Institute on Drug Abuse. (2012). Resource guide: Screening for drug use in general
medical settings: The NIDA Quick Screen. Retrieved from https://www.
drugabuse.gov/publications/resource-guide-screeningdrug-use-in-generalmedical-
settings/nida-quick-screen
Okuyemi, K. S., Goldade, K., Whembolua, G. L., Thomas, J. L., Eischen, S., Sewali, B., …
Resnicow, K. (2013). Motivational interviewing to enhance nicotine patch treatment for
smoking cessation among homeless smokers: A randomized controlled trial. Addiction,
108, 1136–1144. doi:10.1111/add.12140
Okuyemi, K. S., Thomas, J. L., Hall, S., Nollen, N. L., Richter, K. P., Jeffries, S. K., …
Ahluwalia, J. S. (2006). Smoking cessation in homeless populations: A pilot clinical trial.
Nicotine & Tobacco Research, 8, 689–699. doi:10.1080/14622200600789841
Ostacher, M. J., Nierenberg, A. A., Perlis, R. H., Eidelman, P., Borrelli, D. J., Tran, T. B., …
Sachs, G. S. (2006). The relationship between smoking and suicidal behavior,
comorbidity, and course of illness in bipolar disorder. Journal of Clinical Psychiatry, 67,
1907–1911. doi:10.4088/jcp.v67n1210
Porter, M., Harvey, J., Gavin, J. K., Carpenter, M. J., Cummings, K. M., Pope, C., & Diaz, V. A.
(2017). A qualitative study to assess factors supporting tobacco use in a homeless
population. AIMS Medical Science, 4, 83–98. doi:10.3934/medsci.2017.1.83
Pinsker, E. A., Hennrikus, D. J., Erickson, D. J., Call, K. T., Forster, J. L., & Okuyemi, K. S.
(2018). Trends in self-efficacy to quit and smoking urges among homeless smokers

49

participating in a smoking cessation RCT. Addictive Behaviors, 78, 43–50.
doi:10.1016/j.addbeh.2017.10.025
Prochaska, J. J. (2011). Smoking and mental illness—breaking the link. New England Journal of
Medicine, 365, 196–198. doi:10.1056/NEJMp1105248
Pulvers, K., Romero, D. R., Blanco, L., Sakuma, K. L. K., Ahluwalia, J. S., & Trinidad, D. R.
(2015). Light and intermittent smoking among California Black, Hispanic/Latino, and
non-Hispanic White men and women. Nicotine & Tobacco Research, 17, 755–759.
doi:10.1093/ntr/ntu221
Rash, C. J., Petry, N. M., & Alessi, S. M. (2018). A randomized trial of contingency
management for smoking cessation in the homeless. Psychology of Addictive Behaviors,
32, 141–148. doi:10.1037/adb0000350
Rice, E. (2011). HIV risk in large social networks of homeless youth (Grant No.
R01MH093336). Bethesda, MD: National Institute of Mental Health.
Segan, C. J., Maddox, S., & Borland, R. (2015). Homeless clients benefit from smoking
cessation treatment delivered by a homeless persons’ program. Nicotine and Tobacco
Research, 17, 996–1001. doi:10.1093/ntr/ntv062
Shelley, D., Cantrell, J., Wong, S., & Warn, D. (2010). Smoking cessation among sheltered
homeless: A pilot. American Journal of Health Behavior, 34, 544–552.
doi:10.5993/ajhb.34.5.4
Tsai, J., & Rosenheck, R. A. (2012). Smoking among chronically homeless adults: Prevalence
and correlates. Psychiatric Services, 63, 569–576. doi:10.1176/appi.ps.201100398
U.S. Department of Housing and Urban Development. (2015). 2015 Annual homeless
Assessment report to Congress Part 2 – people living in permanent supportive housing in

50

the US. Washington, DC: Author. Retrieved from https://www.hudexchange.info/onecpd/
assets/File/2015-AHAR-Part-2-Section-7.pdf
United Way of Greater Los Angeles. (2016). Greater Los Angeles Coordinated Entry System
survey packet version 2.2. Los Angeles, CA: Author.
Vijayaraghavan, M., Tieu, L., Ponath, C., Guzman, D., & Kushel, M. (2016). Tobacco cessation
behaviors among older homeless adults: Results from the HOPE HOME study. Nicotine
& Tobacco Research, 18, 1733–1739. doi:10.1093/ntr/ntw040
Wang, J. B., & Cataldo, J. K. (2016). Medical marijuana legalization and co-use in adult
cigarette smokers. American Journal of Health Behavior, 40, 205–214.
doi:10.5993/AJHB.40.2.5
Wenzel, S. (2005). Alcohol use and HIV risk among impoverished women (Grant No.
5R01AA015301). Bethesda, MD: National Institute on Alcohol Abuse and Alcoholism.
Wenzel, S. (2009). Heterosexual HIV risk behavior in homeless men (Grant No.
1R01HD059307). Rockville, MD: Eunice Kennedy Shriver National Institute for Child
Health and Human Development.
Zgierska, A., Amaza, I. P., Brown, R. L., Mundt, M., & Fleming, M. F. (2014). Unhealthy drug
use: How to screen, when to intervene. Journal of Family Practice, 63, 524–530.
Zhou, X., Nonnemaker, J., Sherrill, B., Gilsenan, A. W., Coste, F., & West, R. (2009). Attempts
to quit smoking and relapse: Factors associated with success or failure from the
ATTEMPT cohort study. Addictive Behaviors, 34, 365–373.
doi:10.1016/j.addbeh.2008.11.013

51

Tables and Figures
Table 2.1. Descriptive Statistics for Total Sample of Tobacco Users and Each Latent Class

Total Mixed-Frequency
Decreasers
High-Frequency
Maintainers
Low-frequency
increasers
N = 321 n = 71 n = 194 n = 56
n (%) or M (SD) n (%) or M (SD) n (%) or M (SD) n (%) or M (SD)
Demographics
Age 54.54 (7.21) 54.36 (8.10) 55.01 (6.60) 55.49 (7.76)
Gender (female) 83 (25.86) 19 (29.01) 50 (25.52) 14 (26.79)
Race and ethnicity

Black/AA (ref) 183 (57.01) 38 (53.03) 120 (62.24) 25 (45.64)
White/Caucasian 78 (24.30) 14 (20.29) 47 (24.23) 17 (30.36)
Other 21 (6.54) 4 (5.68) 11 (5.67) 6 (10.71)
Latino 39 (12.15) 15 (21.59) 16 (8.25) 8 (14.29)
Veteran 100 (31.15) 17 (23.94) 66 (34.02 17 (30.36)
Education 14.3 (3.68) 14.13 (3.95) 14.34 (3.53) 14.93 (3.69)
Homelessness 6.38 (7.22) 5.89 (6.09) 6.86 (7.53) 5.33 (7.36)
Income

Baseline 565.41 (352.92) 530.03 (347.96) 569.72 (355.88) 593.94 (351.43)
3 months 556.53 (353.38) 510.64 (343.21) 563.11 (357.48) 590.56 (352.36)
6 months 564.81 (355.32) 523.32 (350.99) 569.91 (356.87) 598.27 (357.25)
12 months 627.07
(367.28)**
612.69 (373.24) 620.71 (364.07) 669.59 (375.48)

52

Clinical characteristics
Schizophrenia 101 (31.46) 19 (28.79) 70 (32.43) 12 (18.84)
PTSD 109 (33.96) 20 (28.17) 73 (37.63) 16 (28.57)
Bipolar disorder 109 (33.96) 18 (25.35) 79 (40.72) 12 (21.43)
Depression 183 (57.01) 37 (52.11) 118 (60.82) 28 (50.00)
Anxiety 160 (49.84) 36 (50.70) 98 (50.52) 26 (46.43
Psychological functioning
Baseline 222 (69.16) 49 (69.01) 137 (70.62) 36 (64.29)
3 months 183 (60.00)** 40 (56.34) 110 (56.70) 33 (58.93)
6 months 176 (58.47)** 33 (46.48) 116 (59.79) 27 (48.21)
12 months 150 (52.26)** 34 (47.89) 94 (48.45) 22 (39.29)
Substance Use

Cannabis Use
Baseline 102 (31.97) 24 (33.80) 65 (33.51) 13(23.21)
3 months 103 (33.88) 29 (40.85) 62 (31.96) 12 (21.43)
6 months 115 (38.46)* 28 (39.44) 74 (38.14) 13 (23.21)
12 months 107 (37.28)* 26 (36.62) 71 (36.60) 10 (17.86)
Hard drug use

Baseline 71 (22.12) 17 (23.94) 44 (22.68) 10 (17.86)
3 months 65 (21.31)* 17 (23.94) 40 (20.62) 8 (14.29)
6 months 54 (18.00) 15 (21.23) 34 (17.53) 5 (8.93)
12 months 68 (23.69) 16 (22.53) 44 (22.68) 8 (12.29)
Binge drinking

53

Baseline 54 (16.82) 15 (21.23) 34 (17.53) 5 (8.93)
3 months 51 (16.72) 16 (22.53) 28 (14.43) 7 (12.50)
6 months 60 (20.00) 16 (22.53) 36 (18.56) 8 (14.29)
12 months 51 (17.77) 14 (19.72) 30 (15.46) 7 (12.50)
Cessation attempts

Baseline 120 (39.09) 26 (36.62) 93 (47.93) 1 (1.79)
3 months 106 (43.09) 27 (38.03) 79 (40.72) 0 (0.0)
6 months 96 (40.68) 20 (28.17) 75 (38.66) 1 (1.79)
12 months 85 (37.44)* 18 (25.35) 66 (34.02) 1 (1.79)
*p < .05 relative to baseline. **p < .01 relative to baseline.

54

Table 2.2. Model Fit Indexes for Models with One to Four Classes

Model 1 Model 2 Model 3 Model 4
Log likelihood 659.54 331.59 269.25 231.08
χ
2
df 593 584 573 556
Parameters 16 33 50 67
AIC 2,661.76 2,161.16 2,080.59 2,077.56
BIC 2,722.10 2,285.62 2,269.16 2,330.25
Lo-Mendell-Rubin -- 529.2 112.62 36.50
Entropy -- .92 .89 .90
LMR probability -- p < .001 p < .001 p = .88

55

Table 2.3. Univariable Multinomial Logistic Regression Models with Predictors of Latent Class
Membership
Mixed-Frequency
Decreasers vs.
High-Frequency
Maintainers (ref)
Mixed-Frequency
Decreasers vs.
Low-frequency
increasers (ref)
High-Frequency
Maintainers vs.
Low-frequency
increasers (ref)

OR p OR p OR p
Age 0.99 .56 0.98 .43 0.99 .66
Gender (female) 1.16 .64 1.12 .78 0.97 .93
Race and ethnicity

Black or African American (ref)
White or Caucasian 0.97 .93 0.52 .14 0.54 .08
Other 1.18 .75 0.49 .25 0.41 .09
Hispanic or Latino 3.08 < .01 1.20 .59 0.39 .04
Veteran 0.80 .48 0.80 .25 0.80 .46
Education 0.98 .69 0.94 .23 0.96 .28
Homelessness (years) 0.98 .31 1.01 .66 1.03 .25
Schizophrenia 0.84 .54 1.74 .16 2.07 .03
PTSD 0.64 .09 0.85 .65 1.34 .32
Bipolar disorder 0.47 .01 1.02 .97 2.18 .01
Depression 0.74 .20 1.05 .88 1.41 .20
Anxiety 1.09 .72 1.30 .42 1.19 .64
Income (baseline) 0.99 .64 0.99 .37 0.99 .53
Income (change) 1.00 .76 0.99 .71 0.99 .52

56

Psychiatric disability
(baseline)
1.07 .84 1.54 .36 1.43 .37
Psychological disability
(change)
0.96 .90 1.24 .59 1.30 .47
Cannabis use (baseline) 1.02 .93 2.44 .07 2.38 .05
Cannabis use (change) 0.95 .91 2.97 .03 3.10 < .01
Hard drug use (baseline) 1.06 .88 1.73 .33 1.64 .33
Hard drug use (change) 0.98 .96 1.66 .32 1.69 .25
Binge drinking (baseline) 1.62 .29 3.80 .05 2.35 .17
Binge drinking (change) 1.34 .41 1.30 .62 0.96 .94
Note. Values in bold are significant at p < .10.

57

Table 2.4. Multivariable Multinomial Logistic Regression Models with Predictors of Latent
Class Membership
Mixed Frequency
Decreasers vs. High
Frequency
Maintainers (ref)
Mixed Frequency
Decreasers vs. Low
frequency
increasers (ref)
High Frequency
Maintainers vs. Low
frequency increasers
(ref)
Race and ethnicity OR p OR p OR p
Black or African American
(ref)

White 2.00 0.97 1.36 0.20 1.18 0.13
Other 4.05 0.77 0.45 0.27 1.19 0.09
Latino 7.90 < .01 3.22 0.88 0.85 0.02
Schizophrenia 0.61 0.13 1.14 0.77 1.87 0.10
PTSD 0.68 0.24 1.05 0.89 1.55 0.21
Bipolar disorder 0.47 0.02 1.17 0.71 2.45 0.02
Cannabis use (baseline) 1.27 0.50 2.56 0.07 2.02 0.14
Cannabis use (change) 1.21 0.65 2.85 0.03 2.37 0.03
Binge drinking (baseline) 1.70 0.25 4.13 0.04 2.42 0.19
Binge drinking (change) 1.36 0.41 1.36 0.57 1.01 0.99
Note. Values in bold are significant at p < .10.

58

Figure 2.1. Longitudinal Smoking Frequency Change and Stability across Latent Classes
Note. Multilevel linear regression models confirmed Mixed Frequency Decreasers’ smoking frequency significantly
decreased over time (β=-.28, p=.01, 95% CI=-.49, -.07), while Low Frequency increasers significantly increased
their smoking over time (β=.13, p=.03, 95% CI=.01, .24). There was no significant change in smoking frequency
among the High Frequency Maintainers (β=-.04, p=.18, 95% CI=-.10, .02).

59

Chapter 3: Tobacco Risk Change in the Transition from Homelessness to Housing: The
Role of Housing and Support
Introduction
In 2019, the Centers for Disease Control and Prevention announced that smoking among
adults in the United States had reached an all-time low of 14%. Although this may be true for
housed adults, for adults experiencing homelessness, rates of smoking have remained consistent
at approximately 75%, five times above the national average (Soar, Dawkins, Robson, & Cox,
2020). Efforts to address smoking may be catching up to the level of public health attention it
requires for the general population, but for adults experiencing homelessness, these efforts are
often a low priority (Baggett, Tobey, & Rigotti, 2013; Okuyemi et al., 2006). Additionally,
adults experiencing homelessness face numerous barriers to making changes in their smoking
behavior, and these barriers are often multimorbid and multilayered (Baggett et al., 2013). Many
of these barriers are systemic, including the experience of homelessness itself, wherein this
population is faced with numerous physical and social environmental impediments including
diminished access to routine health care, normative acceptability of smoking, and daily stressors
of achieving basic survival needs (Baggett et al., 2013).
Permanent Supportive Housing
Through the provision of subsidized housing and supportive services, permanent
supportive housing (PSH) may address these barriers and improve smoking behavior outcomes.
PSH rests on two pillars—housing and support—that are often interconnected. Housing is the
structural factor, or physical environment, that serves as the setting and conduit for the social
environment, which is where the social, cultural, economic, and occupational sustenance of daily
life comes together to affect individuals’ health and well-being (Bandura & Walters, 1977;

60

Lounsbury & Mitchell, 2009). The guiding principles of recovery in PSH, in addition to theories
of health behavior and behavior change, rest on the belief that as physical and social
environmental barriers are addressed in PSH, individual improvement is achieved (Bandura &
Walters, 1977; Lounsbury & Mitchell, 2009; Tsemberis, Gulcur, & Nakae, 2004). Although
formerly homeless PSH residents may share a common building or neighborhood, they all have
individual apartments and subsequently, individual social worlds in which they exist. Given this,
understanding residents’ smoking behavior requires an examination of the physical (housing)
and social (support) characteristics that commonly affect individual smoking behavior, while
recognizing the individuality of residents’ social environments and changes in these
environments.
Housing. Housing and support are fundamental to all PSH, yet what they look like can be
quite varied. Two broad and dominant factors unique to PSH are the housing model and
neighborhood. In scattered-site housing models, PSH residents live in housing units rented from
private landlords in buildings that house the general population, whereas single-site models
house PSH residents together in a building dedicated to formerly homeless tenants (Harris,
dunton et al., 2019; Henwood, Harris, et al., 2018). Although no studies have compared
residents’ smoking behavior across housing model type, a study that examined differences in
illicit substance use found single-site residents were more likely to increase injection drug use
over time (Whittaker, Dobbins, Swift, Flatau, & Burns, 2017). Given the tendency to house
adults experiencing homelessness in spurned, peripheral neighborhoods regardless of the housing
model type, the impact of residents’ neighborhoods on their smoking behavior may be particular
to some geographic regions. In Los Angeles, for example, the Skid Row neighborhood has
among the highest number of unsheltered homeless persons and homelessness services per capita

61

(Henry, Cortes, & Shivij, 2018), and it sits in a thriving downtown neighborhood. Similar to
other metropolises, Los Angeles’ downtown serves as the hub for public transportation; however,
unlike other metropolises, it is not the epicenter of the city. Moreover, Los Angeles has high
density, atypical sprawl, and severe public transportation deficits, making life outside of these
areas absent of the vibrant street life and close-knit communities reflective in other large
metropolitan areas (Eidlin, 2005; Reese, Deverteuil, & Thach, 2010; Wei & Ewing, 2018). These
neighborhoods (i.e., Skid Row, downtown Los Angeles [DTLA], and others) differentially affect
other behavioral health outcomes for this population, including service utilization, social
integration, perceived safety, substance use, and sexual risk (Harris, Dunton et al., 2019;
Henwood, Lahey, Harris, Rhoades, & Wenzel, 2013; Hsu, Simon, Henwood, Wenzel, &
Couture, 2016).
Support. Support in PSH is most commonly thought of in relation to the supportive
services residents receive. Indeed, much research has suggested that this aspect of provider
support improves in PSH, such that residents’ utilization of high-cost services decreases and
some aspects of health progress (Harris, Rhoades, Duan, & Wenzel, 2019; National Academies
of Sciences, Engineering, and Medicine, 2018; Wenzel et al., 2019). Case managers and other
providers, although integral to recovery, do not comprise the entirety of residents’ support
system. Aligning with the goal of recovery in PSH, support must consider connections and
socialization beyond provider support (Tsemberis, 2010). Structural support, including the size,
relationship (e.g., family, neighbor), and proximity of persons in a social environment, may
affect smoking behavior (Cohen & Syme, 1985; Uchino, Cacioppo, & Kiecolt-Glaser, 1996). For
example, married persons have higher cessation rates (Caponnetto & Polosa, 2008), whereas
having more risk-engaging peers is associated with increased risk engagement (Blok, de Vlas,

62

van Empelen, & van Lenthe, 2017). The functional component of support is concerned with the
delivery of various types of support (emotional, tangible [material goods], and informational)
from people in a social environment (Cohen & Syme, 1985). Many effective behavioral
treatments for tobacco use incorporate functional support such as training family members to
facilitate supportive cessation interventions (Chan et al., 2017; Rohrbaugh, Shoham, Skoyen,
Jensen, & Mehl, 2012). Additionally, naturally occurring forms of these supports facilitate
successful cessation outcomes (Westmaas, Bontemps-Jones, & Bauer, 2010). An ideal way of
examining residents’ social environments and changes in these environments is through social
network analysis. Specifically, an egocentric network approach focuses on the individual (e.g.,
resident) and the network of relations surrounding the individual, allowing inferences to be made
about the features of personal networks, including these varying aspects of support and their
effect on other outcomes (Chung, Hossain, & Davis, 2005).
Interconnectedness of housing and support. Because the housing environment sets the
stage for the social environment and variation in housing exists, social environments also might
vary across these factors, subsequently affecting smoking behavior differentially (Bandura &
Walters, 1977; Lounsbury & Mitchell, 2009). The housing model often dictates the mode by
which persons receive provider support, because single-site residences typically have case
managers located on site and some buildings host on-site health clinics, whereas scattered-site
residents may receive services via mobile community treatment teams or in regions like Los
Angeles, tend to use services off site (Henwood, Harris, et al., 2018). Studies have found single-
site residents utilize services at higher rates (Chinchilla et al., 2019; Somers et al., 2017;
Whittaker et al., 2017). The housing model also often dictates residents’ immediate social
environment; single-site residents are more likely to know their neighbors with shared lived

63

experiences, whereas residents in scattered-site housing report higher rates of feeling isolated
(Yanos, Barrow, & Tsemberis, 2004). Yet qualitative studies have identified increased tension
among single-site residents as they attempt to move away from their “homeless identities”
(Henwood, Lahey, et al., 2018), and that scattered-site residents may benefit from distance from
risky peers (Yanos, Felton, Tsemberis, & Frye, 2007). Residents’ neighborhoods may interact
with support to affect smoking behavior. In Los Angeles, Skid Row residents are more proximal
to provider support, and transportation deficits in other neighborhoods may make getting to
providers a challenge, particularly for a population characterized by poor health (Baggett et al.,
2013). Research has demonstrated that other aspects of support vary across neighborhoods,
including findings that Skid Row has higher rates of functional support, despite being a high-risk
environment (Harris, Dunton et al., 2019). Moreover, the abundance of street selling of cigarettes
in Skid Row may negate any added support benefits that would theoretically improve smoking
outcomes.
Tobacco Risk
Smoking resembles other high-risk behaviors, and for adults experiencing homelessness
and formerly homeless PSH residents, treating it as such is of particular importance. Historically,
smoking has often been considered an act of physical preference, a framing that diminishes the
physical addiction and its effects on mood, the habitual nature and psychological dependence of
nicotine, its use as coping mechanism, its deeply woven relationship with daily activities and
social interactions, and its effect on social environments and physical spaces (Alberg, Shopland,
& Cummings, 2014; Alemanno, 2012; Garrett, Dube, Babb, & McAfee, 2014). Treating smoking
as the detriment to health that it undoubtedly is has been commonplace; however, it may
overlook the complexity of tobacco use and the associated risk (Hughes et al., 2014; McCarthy,

64

Ebssa, Witkiewitz, & Shiffman, 2015). Risk behaviors are defined as conscious or unconscious
actions whose outcomes and the level of harm or possible loss involved are uncertain (Trimpop,
1994). Although the magnitude of the effect is unknown, risk behaviors affect the physical,
economic, or psychosocial well-being of the individual engaging in risk and others (Schonberg,
Fox, & Poldrack, 2011). Treating tobacco use as a risk behavior, particularly among adults
experiencing homelessness and supportive housing residents, may facilitate holistic efforts to
address the socioecological nature of this issue.
Supportive housing programs are typically guided by the Housing First (HF) model, an
approach that addresses the high rates of substance use and risk among formerly homeless
residents using harm reduction (Tsemberis et al., 2004). By working with clients at their own
pace, harm reduction reduces the risk and harmful effects of substance use that affect the
individual, community, and society (Watson, Shuman, Kowalsky, Golembiewski, & Brown,
2017). Harm reduction is considered fundamental to the recovery-oriented practices of HF by
moving away from a focus on reducing individual health risk behaviors toward attention to the
resident’s overall well-being (Watson et al., 2017). Although variations exist regarding the
implementation and fidelity of the HF model and harm reduction practices, tobacco use overall
has been omitted from harm reduction efforts despite its life-threatening effects (Hawk et al.,
2017; Prochaska, 2010). To prioritize smoking and align with the principles and practices of the
HF model, this study examined smoking behavior as a construct of tobacco risk.
Current Study
The social environments of adults experiencing homelessness may change as their
physical environments change—that is, as they transition to PSH—and these changes may affect
their smoking behavior. This study is the first to examine social and physical environmental

65

factors associated with formerly homeless PSH residents’ tobacco risk, a measure that
encompasses the complexity of smoking behavior—an important distinction given the
vulnerability of this population and cessation hurdles it faces (Baggett et al., 2013). Specifically,
this study examined: (a) changes in tobacco risk and support in the transition from homelessness
across residents’ initial year in PSH; (b) the cross-sectional and longitudinal changes in support
associated with tobacco risk; and (c) the main and interactive effects of housing (i.e., housing
model, neighborhood) and support (i.e., structural, functional, provider), and housing and
changes in support on tobacco risk. To achieve this goal, this study used longitudinal quantitative
and social network data from a prospective cohort study of adults who transitioned from
homelessness to housing in Los Angeles, using egocentric network analysis and latent growth
curve modeling. Examining these relationships offered the opportunity to assess whether
residents’ locales and the structural nature of their housing affected their smoking behavior,
which may inform policies and practices in PSH to improve housing as a social determinant of
health. Results from this study may also aid in identifying the social deficits and strengths of
PSH residents, a population that struggles with social isolation. Last, understanding how
residents’ social and physical environments come together to affect their tobacco risk may
facilitate targeted and tailored cessation interventions that address multilevel barriers.
Methods
These longitudinal data were drawn from a sample of adults experiencing homelessness
(National Institute on Drug Abuse Grant No. R01DA036345) who entered supportive housing in
Los Angeles County (LAC) between August 2014 and January 2016. Researchers collaborated
with 26 housing providers in LAC to recruit participants. These agencies make up most housing
providers in LAC, including some of the largest supportive housing agencies in the county.

66

Participants were screened for study eligibility over the phone or in person and were deemed
eligible if they were 39 years old or older, spoke English or Spanish, were unaccompanied
(living without minor children), and were entering housing within 20 miles of downtown Los
Angeles. Given the prevalence of persons experiencing homelessness moving into PSH in Long
Beach, a city in LAC, an exception was made to include participants moving to this area if it
exceeded the 20‐mile radius. Confirmation by agency personnel that the participant moved into
PSH was also required for study eligibility.
Eligible participants completed informed consent, and those who enrolled participated in
four interviews: baseline (before or within 5 days of PSH move in) and 3, 6, and 12 months after
housing. Interviews across all points consisted of a questionnaire and social network interview
(SNI) that together took approximately 1.5 hours to complete. Interviews were conducted one ‐
on‐one by trained study personnel who collected respondents’ answers with iPads using
Qualtrics survey software for the questionnaire and a social network data collection app designed
by the study team (Rice, 2011). Interviews were conducted in private locations such as agency
conference rooms, building lobbies, cafes, a local outdoor space with sufficient privacy, and
participants’ apartments (follow-up interviews only). Participants received $20 at baseline, with
increases in $5 increments at each follow ‐up (3-month = $25, 6-month = $30, and 12-month =
$35). Persons not interviewed at any survey midpoint were eligible to participate in later
interviews, excluding those who passed away or withdrew from the study.
Calendars were used to aid participants’ recall of information collected in the
questionnaire and SNI, and at follow-up interviews, participants were reminded about when their
previous interview had taken place. The questionnaire collected participants’ demographic
characteristics at baseline, along with service utilization, tobacco and other substance use, and

67

housing information, which were collected at each wave. The SNI was also conducted at each
wave and gathered characteristic information regarding the nature of participants’ relationships
with network members, their proximity to network members, and the social support that network
members provided to respondents for all nominated individuals. These networks were considered
egocentric, such that each network constituted the direct ties of an index person—in this case,
each participant (Neaigus, 1998).
The SNI began by asking participants to identify any individuals they interacted with
during the previous 3 months via phone, in person, online, or through letters. Participants were
asked to first nominate social network members whom they considered most important to them.
Following this, interviewers went through each form of communication (phone, in person,
online, or through letters) separately to allow participants time to reflect on persons they had
communicated with during the past 3 months. Basic prompts were used as needed to aid in
participants’ recall of past-3-month contact with network members, such as “Was there anyone
else you talked to online?” No minimum or maximum number of network members was required
for the SNI. A “by questions” approach was used, which asks a given question for all nominated
individuals (Kogovšek & Ferligoj, 2005). The SNI application generated an egocentric network
that allowed participants to easily view the network member they had nominated as each
question was asked, if they wished.
All study procedures were approved by the University of Southern California’s
Institutional Review Board. In addition, the study received a Certificate of Confidentiality from
the U.S. Department of Health and Human Services to protect participant data from subpoena.

68

Measures
Tobacco risk. The National Institute on Drug Abuse-Modified Alcohol, Smoking and
Substance Involvement Screening Test (NIDA-Modified ASSIST; National Institute on Drug
Abuse, 2012) was used to assess tobacco risk. The NIDA-Modified ASSIST features six
questions gauging tobacco use frequency, smoking urges, interference in responsibilities or
motivation, concern from loved ones, and failed quit attempts that are summed to create an
indicator of tobacco risk. These items are recorded on a Likert scale of never, once or twice,
monthly, weekly, and daily or almost daily. The NIDA-Modified ASSIST is considered an
accurate measure of self-reported substance use and has been used with homeless and other
vulnerable populations (Bailey, 2018; Oga, Mark, Peters, & Coleman-Cowger, 2020; Weber,
Thompson, Schmiege, Peifer, & Farrell, 2013). It has demonstrated good to excellent sensitivity
and specificity (Zgierska, Amaza, Brown, Mundt, & Fleming, 2014) and excellent test–retest
reliability (McNeely et al., 2014).
Housing. Participants’ self-reported addresses were gathered at each follow-up interview
and used to create a housing model measure denoting whether participants were housed in
scattered-site or single-site models. This measure was created using study interviewers’ existing
knowledge of their participants’ building type, available information through agencies’ websites,
and confirmation by agency personnel in the rare case of uncertainty about whether the building
was a single-site or scattered-site setting. Residents’ buildings were considered time invariant
because there was little movement and 95% were stably housed at the 12-month interview.
Seven participants moved between their third and fourth month in housing from a single-site
building to separate scattered-site buildings. There were no statistically significant differences
between these seven individuals and other tobacco-using participants in the sample.

69

Neighborhood. A measure of location was created, consisting of three levels: Skid Row,
DTLA, and other. Levels were based on city-established neighborhood borders for DTLA and
prior identification of Skid Row borders (Hsu et al., 2016; Solari et al., 2014). These levels were
considered importantly distinct for several reasons: (a) Skid Row is the largest area of
unsheltered homeless persons in the nation (Henry et al., 2018), serving as the epicenter for
homeless persons and homelessness services in Los Angeles; (b) although Skid Row sits in
DTLA, the remaining downtown region is in stark contrast to Skid Row, characterized by
expansion, gentrification, and an upper-middle-class population (Reese et al., 2010); and (c)
neighborhoods outside of Skid Row and DTLA fall into the “other” category, considered an
important distinction given the deficits in public transportation and immense sprawl of Los
Angeles, which can be particularly challenging for vulnerable populations (Frumkin, 2016), and
for purposes of analysis. These neighborhoods have been identified as correlates of other
behavioral health and social outcomes for persons experiencing homelessness and PSH residents
(Harris, Dunton et al., 2019; Henwood, Lahey, et al., 2013; Hsu et al., 2016). Participants’
addresses were also used to determine their neighborhoods via Google Maps. Two research team
members coded each neighborhood separately, and simultaneously co-coded each participant’s
neighborhood for confirmation.
Support. Social network measures used in this study were drawn from previous studies
with homeless populations (Green, Tucker, Golinelli, & Wenzel, 2013; Rhoades et al., 2011;
Rice, 2011; Tucker et al., 2009). All network measures were gathered at each wave of data
collection and captured past-3-month networks. Summed totals of each network measure at each
wave were included in analyses. Except for network providers, the remaining provider support
measures drew from the quantitative survey and assessed past-3-month behavior at each wave.

70

Structural support.
Size. Network size was the total number of network members a given participant
nominated in each SNI at each wave.
Member role. After the nomination of a participant’s network members, subsequent
questions assessed network member roles or relationships with the participant across the
following categories: relative, romantic partner, veteran, neighbor (those who live in their
building), or someone they had met while homeless.
Network member distance. To determine the role of distant network members, a question
related to physical proximity was asked: “Who lives so far away that you have trouble seeing
them when you want to?” For those who lived nearby, a question asked, “Who lives so close that
you can easily see them when you want to?”
Functional support. Each type of functional support drew from distinct questions in the
SNI. Emotional support was derived from a question that asked, “Who do you feel emotionally
close to most of the time?” Tangible support drew from a measure that asked participants who in
their networks “had provided them with or assisted with getting money, clothes, food, or a place
to stay.” Informational support drew from an item that asked participants whom in their networks
they “could go to for advice about a problem if they needed it.” Given research identifying
conflict as a factor that may interact with housing model, neighborhood, and smoking behavior, a
measure was included. Conflict was assessed by a question in the SNI that asked participants
whom in their social networks they “get into arguments or disagreements with.”
Provider support.
Providers in networks. The SNI gathered nominated case managers, doctors, and
emotional health counselors (i.e., therapists and psychiatrists) at each wave. Although these

71

network member types could be considered components of structural support, it is important to
consider them as provider support measures because participants were not required to nominate
providers. Therefore providers in networks may be indicative of particularly salient provider–
client relationships.
Case management visits. The number of case management visits in the past 3 months was
collected at each wave. For residents who reported not having a case manager, their case
management visits were considered zero.
Health care visits. Participants were asked, “How many outpatient medical care visits did
you have in the past 3 months?” at each wave. It was clarified that health care visits referred to
outpatient procedures.
Mental health visits. Participants were asked, “How many outpatient therapy visits for
emotional or psychiatric problems did you have?” The total number of mental health visits in the
past 3 months was recorded at each wave.
Control measures. Several demographic measures were included as covariates in these
analyses. Demographic measures used as controls included participants’ age, race and ethnicity,
gender, veteran status, education, and income. Race and ethnicity categories included Black or
African American, White, Hispanic or Latino, and other (Native American or Alaska Native,
Asian, Native Hawaiian or Pacific Islander, and multiracial; Wenzel, 2005). Participants’ gender
identities included male, female, and male-to-female transgender. Three participants who
identified as male-to-female transgender were included in the female gender category for
analytic purposes. The education measure, adapted from the California Health Interview Survey
(UCLA Center for Health and Policy Research, 2014), assessed participants’ highest level of

72

education. Income was captured across wave, and the cross-sectional and longitudinal effects of
this measure were included in the controlled models.
Analysis
This study examined the main and interactive effects of housing, social network, and
service factors, including cross-sectional and longitudinal changes in social network and service
factors, as predictors of longitudinal tobacco risk in the transition from homelessness to housing.
Data for this study are hierarchical in nature such that they contain two levels: repeated measures
across time (baseline [prior to entering housing] and 3, 6, and 12 months), including time-
invariant and time-varying social network and service measures at level 1. These repeated
observations are nested in participants, along with the change in time-varying measures that
serve as level 2. To achieve these analytic goals, linear growth curve modeling (LGCM) was
used, because this approach captures longitudinal change at the population (sample) level and the
individual level. Specifically, LGCM is ideal for this inquiry because individual differences can
be examined over time by allowing the coefficients of time-varying measures to vary across
persons (Curran & Bauer, 2011).
Although housing factors (housing model and neighborhood) were time invariant, social
network measures (sum of network member types and sum of supportive network members) and
service factors (case management visits, mental health services, and medical services) were time
variant and thus, these measures had to be centered to understand whether changes in network
measures predicted changes in tobacco risk. The observed repeated measure for each individual
and time-and-individual-specific residuals represent the within-person effect. Intercept and
slopes can also vary randomly across persons, and this variation was captured by determining
individual deviations from the overall mean intercept and slope of the sample, which is

73

represented by the between-person effect (Curran & Bauer, 2011). All time-variant measures
were subsequently centered as two parts: the within ‐person mean and the deviation from the
mean at each time point (between-effect). The estimated coefficients for the deviation yield an
estimate of the within ‐subject effect (a longitudinal change in the time-varying predictor
associated with the outcome), whereas the parameter estimates associated with the overall mean
were interpreted as the between ‐subject effect (a cross ‐sectional association between the measure
and tobacco risk; Fitzmaurice, Laird, & Ware, 2011). Of note, egocentric network analysis treats
each individual, or ego, as a separate case; thus, characteristics of the ego’s network are
summarized to the participant level, allowing them to be treated traditionally in standard
quantitative analyses (DeJordy & Halgin, 2008). All centering of measures was conducted using
the XTCENTER function in Stata version 16 (Dzubur, 2015).
A five-step model process was used to examine changes in tobacco risk and housing,
support factors, and changes in support predictive of tobacco risk. The initial full unconditional
model (Model 1) examined the significance of the intercept to determine the appropriateness of
this hierarchical approach (Albright & Marinova, 2010). The unconditional LGCM (Model 2)
included the examination of a random intercept and a linear change (slope) in tobacco risk. This
allowed for determination of a straight-line, or linear, trend in tobacco risk over time. Following
that, a quadratic growth model (Model 3) determined whether the growth in tobacco risk was
better fit to a curved line. Model 3 included the intercept of tobacco risk and the slopes of both
the linear and quadratic measures of time. Model 3 introduced level 1 and 2 predictors to predict
cross-sectional associations between measures of support and tobacco risk, change (slopes) in
time-varying measures, and the interaction between housing and social support measures as
predictive of tobacco risk. Model 4 involved two steps. The first provided the main effect of each

74

measure in a given interaction, along with the interactive effect. In the second step of Model 4,
significant predictors were examined simultaneously in a single model, and a final conditional
model (Model 5) examined housing and social factors along with level 2 covariates, which
included demographic and clinical characteristics, to determine whether the predictors of interest
had significant effects on growth.
Although this study was concerned with fixed effects, random effects were included in
the models to account for variation across individual observed effects. Level 1 and 2 predictors
are assumed to be fixed in multilevel models, yet the variability of observed effects across
individuals (level 1) can vary across groups and thus, may be treated randomly. Intercepts and
slopes were considered random to account for intraindividual variation in tobacco risk and
changes in support measures. Unstructured covariance matrixes were employed to account for
distinct estimations of variances and covariances without an expected pattern (Raudenbush &
Bryk, 2002). The estimation of unknown parameters in the models was based on restricted
maximum-likelihood estimation. Multilevel modeling was used to conduct all LGCM models—
specifically, the XTMIXED procedure in Stata version 16 was used.
Results
Descriptive Statistics
At baseline, 421 persons were enrolled in the study; 405 completed 3‐month interviews
(96% retention), 400 completed 6‐month interviews (95% retention), and 383 completed 12‐
month interviews (91% retention). Reasons for loss at each point included death, incarceration,
withdrawal, and loss of contact. Of the 421 participants in the total sample, 321 (76.25%) were
identified as tobacco users at one or more waves, and subsequently served as the analytic sample.
Among the tobacco users, the average age was 54.54 (SD = 7.21), 25.86% identified as female,

75

more than half (57.01%) identified as Black or African American, 24.3% identified as White or
Caucasian, 6.54% were considered part of the “other” racial category, and 12.15% identified as
Hispanic or Latino. The analytic sample featured 100 (31.15%) veterans. Regarding housing
descriptives, 66.04% lived in single-site housing models, 22.74% lived in Skid Row, 20.25%
lived in DTLA, and the remaining participants lived in “other” neighborhoods. Regarding control
measures significantly associated with tobacco risk, women had higher tobacco risk scores by an
average of 2.04 units (p = .02, 95% CI = 0.38, 3.70). A cross-sectional relationship existed
between income and tobacco risk—for each additional dollar in income reported, tobacco risk
scores increased by 0.04 (p < .01, 95% CI = 0.02, 0.06). All descriptive statistics of the sample
can be found in Table 3.1.
Social Support Change
Several time-variant measures (Table 3.2) significantly changed. For every 1-unit
increase in time, or each additional wave, there was a 0.42-unit decrease in social network size (p
< .01, 95% CI = -0.57, -0.27), a 0.16-unit increase in relatives in social networks (p < .01, 95%
CI = 0.07, 0.24), a 0.19-unit increase in neighbors in social networks (p < .01, 95% CI = 0.13,
0.24), a 0.32-unit decrease in homeless social network members (p < .01, 95% CI = -0.38, -0.26),
and a 0.19-unit decrease in network members who lived nearby (p < .01, 95% CI = -0.29, -0.08).
Regarding support, for each additional wave, there was a 0.12-unit decrease in the number of
network members to whom participants felt emotionally close (p = .02, 95% CI = -0.24, -0.02), a
0.17-unit decrease in the number of network members providing tangible support (p < .02, 95%
CI = -0.24, -0.11), and a 0.36-unit decrease in network members providing instrumental support
(p < .01, 95% CI = -0.46, -0.26). Regarding provider support, there was a 0.21-unit decrease in
case managers (p < .01, 95% CI = -0.25, -0.17) and a 0.06-unit decrease in emotional health

76

counselors (p < .01, 95% CI = -0.08, -0.03) in social networks at each wave. The number of
sessions with case managers also declined, with the aggregated average declining by
approximately 4.02 sessions across each wave (p < .01, 95% CI = -4.78, -3.27). Likewise, mental
health visits declined by 0.49 units across each wave (p < .01, 95% CI = -0.78, -0.21), as did
health care visits by 0.44 units (p < .01, 95% CI = -0.67, -0.22).
Unconditional Models
Table 3.3 presents the fixed effects of predictors for Models 1–3. Model 2 identified a
significant linear decrease in tobacco risk over time (p < .01, 95% CI = -2.78, -0.82). Model 3
demonstrated a negative linear relationship (p < .01, 95% CI = -3.01, -0.86) with a positive
quadratic term (p < .01, 95% CI = 0.22, 0.92), indicating the growth rate declined at the 3- and 6-
month marks and began to demonstrate an upward trend at the 12-month mark (see Figure 3.1).
Conditional Main-Effect Models
In the main-effect model without controls, on average, estimated tobacco risk scores on
the NIDA-Modified ASSIST were 2.15 points higher for participants living in DTLA across
waves (p < .04, 95% CI = 0.03, 4.27). Several cross-sectional associations were identified—
network size, romantic partners, relatives, veterans, distant network members, instrumental
support, and mental health visits were positively associated with tobacco risk, indicating that
higher rates of these measures predicted higher scores of tobacco risk (see Table 3.4 for all
parameter estimates). Results identified a positive relationship between change in network size
and tobacco risk, such that with every 1-unit increase in network size over time, participants’
tobacco risk scores increased by 0.15 (p = .03, 95% CI = 0.01, 0.28). However, it should be
noted that the slopes of these lines decreased over time. This was true for changes in case
managers in networks as well. Across the initial year in housing, gaining a case manager in one’s

77

social network was associated with a 0.08-point increase in tobacco risk (p = .08, 95% CI = -
0.24, 0.41); however, case managers in networks declined over time. Although case management
visits and change in counselors were not significant at the p < .05 level, they were retained in
models with controls because they were significant at the p < .10 level.
The main effects largely remained significant in controlled models (see Model 5 in Table
3.4), except for the cross-sectional relationship between romantic partners and tobacco risk, and
the change in case managers and network size, which were no longer significant predictors in
controlled models.
Conditional Interactive Models
Housing model significantly moderated the relationship between several network
measures and tobacco risk including network size, neighbors and nearby network members. The
main effects for these support measures were not statistically significant (although size
approached statistical significance), suggesting these are “spreading” interactions, which
indicates that one group has a stronger effect than the other, or “crossover” interactions, which
suggests that the two groups have opposing effects on the outcome (Loftus, 1978). To interpret
these relationships, the slopes of each interaction were plotted (see Figures 3.2–3.6). At baseline,
for each additional network member, residents in scattered-site housing tobacco risk scores were
reduced by .07 (p = .04, 95% CI = -0.13, -0.01). Figure 3.2 demonstrates a clear crossover effect,
which suggests that the effect of change network size on tobacco risk is distinctly opposite across
housing models. That is, for single-site residents, an increase in network size over time was
associated with increased tobacco risk, whereas for scattered-site residents, an increase in
network size was associated with decreased tobacco risk. At the 3-month interview (the first
wave in which neighbors in networks were assessed), for every additional neighbor in their

78

networks, residents in scattered-site housing had tobacco risk scores that decreased by 1.39
points (p < .01, 95% CI = -2.31, -0.47). Figure 3.3 demonstrates the spreading effect—over time,
the gap in neighbors widens, with a stronger relationship for single-site residents. For scattered-
site residents, there was a negative relationship between nearby network members and tobacco
risk at baseline; for each additional nearby network member, there was a 0.62-point decrease in
tobacco risk scores (p = .03, 95% CI = -1.16, -0.08). Although residents in both housing models
had relatively similar numbers of nearby network members at baseline (while homeless), the
spread in nearby network members widens over time (see Figure 3.4), and the relationship
between neighbors and tobacco risk was stronger for single-site residents.
Neighborhood was also a significant moderator. At any given wave, for each additional
distant network member nominated, DTLA residents’ tobacco risk scores decreased by 0.86 (p =
.03, 95% CI = -1.66, -0.07). The negative interaction effect indicated the gap in tobacco risk
widens between those in Skid Row and those in DTLA, cross-sectionally. Neighborhood
moderated the relationship between change in conflicting network members and tobacco risk. At
baseline for each additional conflicting network member, DTLA residents’ tobacco risk scores
were reduced by 1.26 (p = .03, 95% CI = -2.42, -0.10). The main effect of change in conflict was
not significantly associated with tobacco risk, suggesting this is also a crossover interaction (see
Figure 3.5). Specifically, the relationship between network member conflict and tobacco risk was
positive for Skid Row residents; gaining a conflicting network member over time was associated
with a higher tobacco risk score, whereas for DTLA residents, gaining a conflicting social
network member was associated with a decrease in tobacco risk. When comparing Skid Row and
DTLA residents, as the number of neighbors in networks increased, the gap in tobacco risk
narrowed between these two neighborhoods (p = .03, 95% CI = 0.55, 5.03), cross-sectionally.

79

Change in neighbors in networks was identified as a significant interaction effect when
comparing Skid Row residents and those in “other” neighborhoods (p = .03, 95% CI = 0.13,
2.09). Figure 3.6 demonstrates that this was a spreading effect, with a stronger relationship
between changes in neighbors and tobacco risk for Skid Row residents. Neighborhood
moderated the relationship between nearby network members and tobacco risk as well. As
nearby network members increased, the gap in tobacco risk widens cross-sectionally between
residents in Skid Row and those in other neighborhoods (p = .04, 95% CI = -2.18, -0.03). At any
given wave, for each additional case manager nominated, residents in other neighborhoods had
tobacco risk scores decrease by 3.17. The significant negative estimate for the interaction
between neighborhood and case managers indicated that as case managers increased, the gap in
tobacco risk widens between residents in Skid Row and other neighborhoods, cross-sectionally
(p = .01, 95% CI = -5.70, -0.63).
In the controlled models (see Table 3.5), all interaction effects with housing model
remained significant. The interaction effects of neighborhood, specifically DTLA and Skid Row
comparisons, also remained significant. The interaction between neighborhood and counselors in
social networks became significant in the controlled model, suggesting that as counselors in
networks increased, the differences in tobacco risk between residents in Skid Row and residents
in DTLA widens cross-sectionally (p = .03, 95% CI = -9.94, -0.64). When comparing differences
between other neighborhood residents and Skid Row residents, the interactive effects of
neighborhood and neighbors in networks and neighborhood and nearby network members were
no longer significant; however, the interactive effect of neighborhood and case managers did
remain significant.

80

Discussion
This study examined measures of support, their change over time, and their interaction
with housing factors as predictors of longitudinal tobacco risk as adults transitioned from
homelessness to supportive housing in Los Angeles. Findings indicate that tobacco risk
significantly declined over time; however, it featured a concave slope, indicating that this may be
an initial trend and tobacco risk may rise as residents remain in housing longer. The significant
quadratic slope could be an indication that upon entry into PSH, residents are initially eager and
proactive about addressing behavioral health issues, but with time, this motivation dissipates.
Interestingly, research with this sample (including nontobacco users) found residents’ self-
reported health goals were much higher at the 3- and 6-month marks compared to baseline (while
homeless) and peaked at the 6-month mark, with goals shifting to a focus on relocating to a new
apartment at the 12-month mark (Wenzel et al., 2018). Although housing is often considered a
launching pad for recovery and improvements in health and well-being (Henwood et al., 2013),
this combination of findings may suggest that the stressors of relocation and their impact on
tobacco risk reflect similar stressors of securing housing. It is important to understand in future
research how housing security and concerns about safety may affect these relationships and their
change over time.
It was presumed in part that support would improve (increase) post housing and
subsequently tobacco risk would decrease, however most support measures significantly
decreased over time. Residents had smaller social networks; fewer nearby network members;
fewer case managers and mental health counselors; fewer case management, health care, and
mental health visits; and less functional support across all three subtypes assessed (informational,
emotional, and tangible). These findings correspond with a large body of research that has found

81

PSH residents struggle with isolation and community integration (Harris et al., 2018; Hawkins &
Abrams, 2007; Padgett, Tiderington, Smith, Derejko, & Henwood, 2016; Tsai, Mares, &
Rosenheck, 2012). Residents may be distancing themselves as a means of improving their social
relationships and healing from the chaos and consequential traumas of homelessness and their
past (Henwood et al., 2013; Padgett, Henwood, Abrams, & Davis, 2008). However, findings
from this study indicate that even after a year, many of these support components had declined,
which could suggest the marginalization of this population is being maintained or even enhanced
(Henwood et al., 2013; Hopper, 2012). Although some aspects of structural support increased,
including relatives and neighbors, the decline of nearly all measures of provider support is
concerning because residents may be losing key players to support the process of swapping
negative social influences for prosocial ones.
Components of structural, functional, and provider support emerged as significant main
effects in these analyses. Having larger networks and more romantic partners, relatives, and
veterans in networks was positively associated with increased tobacco risk. Because the tobacco
risk construct includes a question that inquires about concern from loved ones, this may
demonstrate the closeness and care expressed by these network members intended to improve
residents’ smoking outcomes. Larger networks may simply mean that residents have a higher
probability of having network members who show concern. A more dismal interpretation of the
direction of these relationships would be that the tension and strain in these relationships may
facilitate increased tobacco risk, and that increased risk among these network members and
larger risk networks are driving increased tobacco risk (Blok et al., 2017). The positive main
effect of informational support, counselors, mental health sessions, and case management
sessions may capture the exacerbated risk or high rates of mental conditions common among

82

formerly homeless tobacco users (Harris, Winetrobe, Rhoades, & Wenzel, 2019), suggesting
these individuals are appropriately connected to services and have enhanced exposure to sources
of advice to navigate problems and develop alternative coping strategies to smoking. Given this,
the absence of associations of health care providers and health care visits with tobacco risk are of
substantial concern, because the U.S. Public Health Service guidelines that recommend all health
care providers deliver brief cessation interventions during all encounters with tobacco-using
patients (Fiore et al., 2008). Health care providers working with adults experiencing
homelessness have indicated that competing medical, psychiatric, and social issues hinder their
delivery of cessation interventions (Baggett et al., 2013). The strong desire to cease smoking
reported by this population (Baggett et al., 2013; Okuyemi et al., 2006) and increased health
goals once housed (Wenzel et al., 2018) warrant enhancing provider resources to ensure their
prioritization of cessation.
Changes in residents’ network size, number of neighbors in their network, and number of
nearby networks were associated with tobacco risk, and each of these associations demonstrated
higher risk for single-site residents. Although these findings may suggest some aspects of
scattered-site residents’ networks protect against tobacco risk, these associations also point to the
potential increased isolation scattered-site residents face. This is contradictory to prior research
from cross-sectional and qualitative studies that found scattered-site residents are no more
isolated than residents in other housing settings (Gulcur, Tsemberis, Stefancic, & Greenwood,
2007; Henwood et al., 2015; Yanos et al., 2007), a potential insight from these longitudinal data.
Given that all scattered-site housing in Los Angeles is in “other” neighborhoods, that is, no
scattered-site units are in Skid Row or DTLA, the decreases in network size and nearby network
members may highlight the transportation and proximity challenges these residents face in

83

accessing network members. Regarding changes in neighbors, a study by Patterson,
Moniruzzaman, and Somers (2014) found that residents in single-site housing were more likely
to know their neighbors, but this did not predict increased interaction with them or an increased
sense of community. The current findings indicate there may need to be increased attention to
tailored approaches aimed at improving Los Angeles-based scattered-site residents’ socialization
and building community in their immediate surroundings, in addition to aiding them in
connecting with networks members they may have known before housing and are struggling to
connect with.
Neighborhood moderated the relationship between several support measures and tobacco
risk, including number of neighbors, distant network members, counselors, and change in
conflict over time. Overall, these findings suggest that DTLA residents may have lower rates of
tobacco risk as a result of reduced or declining aspects of support. This supports the notion that
residents in DTLA, all of whom live in single-site housing only a few blocks from Skid Row,
may be distancing themselves from the homeless communities they were entrenched in only a
short time before, including those who live in their buildings. A qualitative study found residents
living in this area reported isolating to avoid the temptation to engage in risk behaviors, namely
illicit substance use (Henwood, Lahey, et al., 2018). DTLA residents’ physical and social
environments may seemingly be idyllic because they are housed in an increasingly affluent
neighborhood, have easy access to public transportation, and are surrounded by a multitude of
establishments and activities utilized by the general population. However, social isolation may
not be a product of the built environment, but rather unintended inequities that result from the
built environment (Gillespie, 2018). Thwarting the temptations of Skid Row may also lead
DTLA residents to avoid the abundance of services offered there, and their high rates of poverty

84

may limit their ability to take part in neighborhood offerings, making these residents particularly
susceptible to isolation. Interestingly, DTLA residents had more distant members in their
networks, which was positively associated with increased tobacco risk, and they had more
counselors consistently across all waves. Mental health counselors may be working to manage
DTLA residents’ risk via their immediate social environments, but they may not be considering
the risk associated with distant network members, although this again could be a function of
network members expressing concern from a distance and should be explored in future research.
When comparing other neighborhoods and Skid Row, neighborhood moderated the
relationship between case managers in networks and tobacco risk, the sole significant interaction
in controlled models. For Skid Row residents, case managers were positively associated with
tobacco risk, whereas for other neighborhood residents, more case managers lowered tobacco
risk. Because residents do not have to nominate providers, the inclusion of a case manager may
speak to the closeness and connection of this relationship for those who do nominate them. This
interaction may reflect several aspects of case management delivery in PSH in Los Angeles.
Homeless adults entering housing on Skid Row have indicated they anticipate utilizing the same
service providers after being housed (Henwood et al., 2013); however, at the time this study took
place, those living in or around Skid Row were commonly working with a case manager at a
local homelessness services agency and once housed, were assigned to a new case manager
(Westat & USC Suzanne Dworak Peck School of Social Work, 2020). Because homelessness
services agencies outside of Skid Row are much sparser and more dispersed, those in other
neighborhoods may have had more continuity of care. This finding may also reflect, in part, the
strategic placement of residents perceived as more service dependent in areas where services are
more readily available (Chinchilla et al., 2019). Residents in other neighborhoods may be

85

perceived as higher functioning and less demanding; subsequently, case manager–resident
relationships might be more copacetic. It may also simply speak to the levels of crime, poverty,
and risk that Skid Row residents face daily. Case managers across regions may be making
similar attempts to address residents’ health and well-being, but the environmental barriers of
Skid Row might hinder the positive impact of case managers observed in other neighborhoods.
Limitations
While the rare and rich data are certainly strengths to this study, they come with some
limitations. These data draw from persons residing in Los Angeles and given the unique
topography and the concentration of homeless persons and services to specific regions, other
studies that examine environmental correlates of smoking behavior may differ in findings. The
support measures examined in this study were assumed to be prosocial, however the assumption
that more social network members is positive may not be the case, particularly for this
population. The outcome measured in this study is a sum score of tobacco risk, which the authors
consider a strength, however this is a relatively novel way of examining smoking behavior and
more research is needed to understand this construct. The measure of tobacco risk also includes a
question about family members and friends who have expressed concern participants’ smoking,
making it difficult to ascertain whether some of the associations identified are a function of risk
expressed concern and care. Last, this study did not examine the role of mental health and/or
other substance as control measures, and these may be important issues to consider in future
research that examined these relationships.
Conclusion
This study examined tobacco use as a risk factor, rather than as a more commonly used
outcome of individual smoking behavior such as smoking frequency or days of abstinence.

86

Given this population’s disproportionately high rates of tobacco use, struggles with poor health,
and other high risk behaviors; low prioritization of cessation by providers; and goals of recovery
and harm reduction services, it felt important to highlight smoking behavior as a function of the
environment. This study highlights the impact of supportive housing residents’ physical and
social environments on their tobacco risk and the differential effects of these environments, even
among bordering neighborhoods. Although PSH may address some of the barriers this
population faces in improving smoking behavior outcomes, this study also demonstrates the
ongoing challenges residents’ face once housed, particularly regarding their social environments.
Skid Row residents, all of whom are living in single-site housing, appear to be at substantially
higher risk. Despite their higher rates of functional support, the concentration of supportive
services in the area, and increased access to public transportation, the risk and marginalization of
this community likely outweigh any potential benefits to their recovery and well-being in PSH.
Across all residents, these results point to deficits in support and the likelihood that many
residents may be replacing risk behaviors with isolative behaviors, which can also be detrimental
to their health (Antonucci, Ajrouch, Webster, & Zahodne, 2019).
These findings warrant several recommendations for practice and future research. First,
they highlight the integral need for providers to prioritize this population’s smoking behavior and
to address environmental hindrances in their targeted efforts. For health care providers, this
entails increased assessment of smoking behavior and delivery of brief interventions. For PSH
providers on the frontlines, this involves the incorporation of tobacco use into community
integration and harm reduction efforts that align with the goals of HF. Specifically, for case
managers in PSH, this involves assessing tobacco risk and other risk behaviors frequently,
working with residents to understand their readiness to change, responding to risk in a

87

compassionate manner, and linking them with cessation resources that meet their needs. It also
requires strategic efforts to assist residents in navigating their risk environments while
simultaneously supporting their socialization within prosocial environments. Regarding housing
policies, improving residents’ housing choice, another principle of the HF model, and reducing
the practice of housing this population in marginalized neighborhoods may facilitate residents’
connection to their social environments and minimize isolation. Moreover, public policies aimed
at improving accessible transportation, including specialized public transportation services for
disabled persons, is fundamental to helping residents address the systemic barriers they face to
improving their tobacco risk outcomes, even once housed.

88

References
Alberg, A. J., Shopland, D. R., & Cummings, K. M. (2014). The 2014 surgeon general’s report:
Commemorating the 50th anniversary of the 1964 report of the advisory committee to the
US surgeon general and updating the evidence on the health consequences of cigarette
smoking. American Journal of Epidemiology, 179, 403–412. doi:10.1093/aje/kwt335
Albright, J. J., & Marinova, D. M. (2010). Estimating multilevel models using SPSS, Stata, SAS,
and R. Retrieved from http://hdl.handle.net/2022/19737
Alemanno, A. (2012). Nudging smokers: The behavioural turn of tobacco risk regulation.
European Journal of Risk Regulation, 3, 32–42. doi:10.1017/S1867299X00001781
Antonucci, T. C., Ajrouch, K. J., Webster, N. J., & Zahodne, L. B. (2019). Social relations across
the life span: Scientific advances, emerging issues, and future challenges. Annual Review
of Developmental Psychology, 1, 313–336. doi:10.1146/annurev-devpsych-121318-
085212
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.
doi:10.1056/NEJMp1301935
Bailey, B. (2018). By when do I have to quit? Could I just cut down some? Evidence to help
inform difficult questions in pregnancy smoking interventions. American Journal of
Obstetrics & Gynecology, 218, S137–S138. doi:10.1016/j.ajog.2017.10.085
Bandura, A., & Walters, R. H. (1977). Social learning theory (Vol. 1). Englewood Cliffs, NJ:
Prentice-Hall.

89

Blok, D. J., de Vlas, S. J., van Empelen, P., & van Lenthe, F. J. (2017). The role of smoking in
social networks on smoking cessation and relapse among adults: A longitudinal study.
Preventive Medicine, 99, 105–110. doi:10.1016/j.ypmed.2017.02.012
Caponnetto, P., & Polosa, R. (2008). Common predictors of smoking cessation in clinical
practice. Respiratory Medicine, 102, 1182–1192. doi:10.1016/j.rmed.2008.02.017
Centers for Disease Control and Prevention. (2019, November 14). Cigarette smoking among
U.S. adults hits all-time low [Press release]. Retrieved from
https://www.cdc.gov/media/releases/2019/p1114-smoking low.html
Chan, S. S. C., Cheung, Y. T. D., Fong, D. Y. T., Emmons, K., Leung, A. Y. M., Leung, D. Y.
P., & Lam, T. H. (2017). Family-based smoking cessation intervention for smoking
fathers and nonsmoking mothers with a child: A randomized controlled trial. Journal of
Pediatrics, 182, 260–266. doi:10.1016/j.jpeds.2016.11.021
Chinchilla, M., Gabrielian, S., Horan, W. P., Glasmeier, A., Hellemann, G., Wynn, J. K., …
Green, M. F. (2019). Comparing tenant and neighborhood characteristics of the VA’s
project- vs. tenant-based supportive housing program in Los Angeles County. Journal of
Health Care for the Poor and Underserved, 30, 1373–1393. doi:10.1353/hpu.2019.0098
Chung, K. K., Hossain, L., & Davis, J. (2005, November). Exploring sociocentric and egocentric
approaches for social network analysis. In Proceedings of the 2nd international
conference on knowledge management in Asia Pacific (pp. 1-8). Wellington, New
Zealand: Victoria University.
Cohen, S., & Syme, S. L. (1985). Issues in the study and application of social support. In Social
support and health (pp. 3–22). San Diego, CA: Academic Press.

90

Curran, P. J., & Bauer, D. J. (2011). The disaggregation of within-person and between-person
effects in longitudinal models of change. Annual Review of Psychology, 62, 583–619.
doi:10.1146/annurev.psych.093008.100356
DeJordy, R., & Halgin, D. (2008). Introduction to ego network analysis. Retrieved from
http://www.analytictech.com/e-net/pdwhandout.pdf
Dzubur, E. (2015). XTCENTER: Stata module to disaggregate within and between-person effects
by centering variables for mixed and melogit models. Retrieved from
https://econpapers.repec.org/software/bocbocode/s458076.htm
Eidlin, E. (2005). The worst of all worlds: Los Angeles, California, and the emerging reality of
dense sprawl. Transportation Research Record, 1902, 1–9.
doi:10.1177/0361198105190200101
Fiore, M. C., Jaén, C. R., Baker, T. B., Bailey, W. C., Benowitz, N. L., Curry, S. J., … Wewers,
M. E. (2008). Treating tobacco use and dependence: 2008 update. Rockville, MD: U.S.
Department of Health and Human Services, U.S. Public Health Service.
Fitzmaurice, G. M., Laird, N. M., & Ware, J. H. (2011). Applied longitudinal analysis (2nd ed.).
Hoboken, NJ: Wiley.
Frumkin, H. (2016). Urban sprawl and public health. Public Health Reports, 117, 201–217.
doi:10.1093/phr/117.3.201
Garrett, B. E., Dube, S. R., Babb, S., & McAfee, T. (2014). Addressing the social determinants
of health to reduce tobacco-related disparities. Nicotine & Tobacco Research, 17, 892–
897. doi:10.1093/ntr/ntu266
Gillespie, A. (2018). A social ecological approach to exploring the relationship between the built
environment and social isolation and loneliness: Implications for public policy. Retrieved

91

from https://ncceh.ca/documents/internal/exploring-relationship-between-built-
environment-and-social-isolation-and
Green, H. D., Tucker, J. S., Golinelli, D., & Wenzel, S. L. (2013). Social networks, time
homeless, and social support: A study of men on Skid Row. Network Science, 1, 305–
320. doi:10.1017/nws.2013.18
Gulcur, L., Tsemberis, S., Stefancic, A., & Greenwood, R. M. (2007). Community integration of
adults with psychiatric disabilities and histories of homelessness. Community Mental
Health Journal, 43, 211–228. doi:10.1007/s10597-006-9073-4
Harris, T., Dunton, G., Henwood, B., Rhoades, H., Rice, E., & Wenzel, S. (2019). Los Angeles
housing models and neighbourhoods’ role in supportive housing residents’ social
integration. Housing Studies, 34, 609–635. doi:10.1080/02673037.2018.1462308
Harris, T., Rhoades, H., Duan, L., & Wenzel, S. L. (2019). Mental health change in the transition
to permanent supportive housing: The role of housing and social networks. Journal of
Community Psychology, 47, 1834–1849. doi:10.1002/jcop.22230
Harris, T., Winetrobe, H., Rhoades, H., & Wenzel, S. (2019). The role of mental health and
substance use in homeless adults’ tobacco use and cessation attempts. Journal of Dual
Diagnosis, 15, 76–87. doi:10.1080/15504263.2019.1579947
Hawk, M., Coulter, R. W., Egan, J. E., Fisk, S., Friedman, M. R., Tula, M., & Kinsky, S. (2017).
Harm reduction principles for healthcare settings. Harm Reduction Journal, 14, 70.
doi:10.1186/s12954-017-0196-4
Hawkins, R. L., & Abrams, C. (2007). Disappearing acts: The social networks of formerly
homeless individuals with co-occurring disorders. Social Science and Medicine, 65,
2031–2042. doi:10.1016/j.socscimed.2007.06.019

92

Henry, M., Mahathey, A., Morrill, T., Robinson, A., Shivji, A., & Watt, R. (2018). The 2018
Annual Homeless Assessment Report (AHAR) to Congress: Part 1: Point-in-time
estimates of homelessness. Washington, DC: U.S. Department of Housing and Urban
Development.
Henwood, B. F., Harris, T., Woo, D., Winetrobe, H., Rhoades, H., & Wenzel, S. L. (2018).
Availability of comprehensive services in permanent supportive housing in Los Angeles.
Health & Social Care in the Community, 26, 207–213. doi:10.1111/hsc.12510
Henwood, B. F., Hsu, H. T., Dent, D., Winetrobe, H., Carranza, A., & Wenzel, S. (2013).
Transitioning from homelessness: A “fresh-start” event. Journal of the Society for Social
Work and Research, 4, 47–57. doi:10.5243/jsswr.2013.4
Henwood, B. F., Lahey, J., Harris, T., Rhoades, H., & Wenzel, S. L. (2018). Understanding risk
environments in permanent supportive housing for formerly homeless adults. Qualitative
Health Research, 28, 2011–2019. doi:10.1177/1049732318785355
Henwood, B. F., Stefancic, A., Petering, R., Schreiber, S., Abrams, C., & Padgett, D. K. (2015).
Social relationships of dually diagnosed homeless adults following enrollment in housing
first or traditional treatment services. Journal of the Society for Social Work and
Research, 6, 385–406. doi:10.1086/682583
Hopper, K. (2012). The counter-reformation that failed? A commentary on the mixed legacy of
supported housing. Psychiatric Services, 63, 461–463. doi:10.1176/appi.ps.201100379
Hsu, H. T., Simon, J. D., Henwood, B. F., Wenzel, S. L., & Couture, J. (2016). Location,
location, location: Perceptions of safety and security among formerly homeless persons
transitioned to permanent supportive housing. Journal of the Society for Social Work and
Research, 7, 65–88. doi:10.1086/685034

93

Hughes, J. R., Solomon, L. J., Naud, S., Fingar, J. R., Helzer, J. E., & Callas, P. W. (2014).
Natural history of attempts to stop smoking. Nicotine & Tobacco Research, 16, 1190–
1198. doi:10.1093/ntr/ntu052
Kogovšek, T., & Ferligoj, A. (2005). Effects on reliability and validity of egocentered network
measurements. Social Networks, 27, 205–229. doi:10.1016/j.socnet.2005.01.001
Loftus, G. R. (1978). On interpretation of interactions. Memory & Cognition, 6, 312–319.
doi:10.3758/BF03197461
Lounsbury, D. W., & Mitchell, S. G. (2009). Introduction to special issue on social ecological
approaches to community health research and action. American Journal of Community
Psychology, 44, 213–220. doi:10.1007/s10464-009-9266-4
McCarthy, D. E., Ebssa, L., Witkiewitz, K., & Shiffman, S. (2015). Paths to tobacco abstinence:
A repeated-measures latent class analysis. Journal of Consulting and Clinical
Psychology, 83, 696–708. doi:10.1037/ccp0000017
McNeely, J., Strauss, S. M., Wright, S., Rotrosen, J., Khan, R., Lee, J. D., & Gourevitch, M. N.
(2014). Test–retest reliability of a self-administered Alcohol, Smoking and Substance
Involvement Screening Test (ASSIST) in primary care patients. Journal of Substance
Abuse Treatment, 47, 93–101. doi:10.1016/j.jsat.2014.01.007
National Academies of Sciences, Engineering, and Medicine. (2018). Permanent supportive
housing: Evaluating the evidence for improving health outcomes among people
experiencing chronic homelessness. Washington, DC: National Academies Press.
National Institute on Drug Abuse. (2012). Resource guide: Screening for drug use in general
medical settings: The NIDA Quick Screen. Retrieved from

94

https://www.drugabuse.gov/publications/resource-guide-screeningdrug-use-in-
generalmedical-settings/nida-quick-screen
Neaigus, A. (1998). The network approach and interventions to prevent HIV among injection
drug users. Public Health Reports, 113(Suppl. 1), 140–150.
Oga, E. A., Mark, K., Peters, E. N., & Coleman-Cowger, V. H. (2020). Validation of the NIDA-
Modified ASSIST as a screening tool for prenatal drug use in an urban setting in the
United States. Journal of Addiction Medicine. Advance online publication.
doi:10.1097/ADM.0000000000000614
Okuyemi, K. S., Thomas, J. L., Hall, S., Nollen, N. L., Richter, K. P., Jeffries, S. K., …
Ahluwalia, J. S. (2006). Smoking cessation in homeless populations: A pilot clinical trial.
Nicotine & Tobacco Research, 8, 689–699. doi:10.1080/14622200600789841
Padgett, D. K., Henwood, B., Abrams, C., & Davis, A. (2008). Engagement and retention in
services among formerly homeless adults with co-occurring mental illness and substance
abuse: Voices from the margins. Psychiatric Rehabilitation Journal, 31, 226–233.
doi:10.2975/31.3.2008.226.233
Padgett, D. K., Tiderington, E., Smith, B. T., Derejko, K. S., & Henwood, B. F. (2016). Complex
recovery: Understanding the lives of formerly homeless adults with complex needs.
Journal of Social Distress and the Homeless, 25, 60–70.
doi:10.1080/10530789.2016.1173817
Patterson, M. L., Moniruzzaman, A., & Somers, J. M. (2014). Community participation and
belonging among formerly homeless adults with mental illness after 12 months of
Housing First in Vancouver, British Columbia: A randomized controlled trial.
Community Mental Health Journal, 50, 604–611. doi:10.1007/s10597-013-9672-9

95

Prochaska, J. J. (2010). Failure to treat tobacco use in mental health and addiction treatment
settings: A form of harm reduction? Drug and Alcohol Dependence, 110, 177–182.
doi:10.1016/j.drugalcdep.2010.03.002
Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models: Applications and data
analysis methods. Thousand Oaks, CA: Sage.
Reese, E., Deverteuil, G., & Thach, L. (2010). ‘Weak‐center’gentrification and the contradictions
of containment: Deconcentrating poverty in downtown Los Angeles. International
Journal of Urban and Regional Research, 34, 310–327. doi:10.1111/j.1468-
2427.2010.00900.x
Rhoades, H., Wenzel, S. L., Golinelli, D., Tucker, J. S., Kennedy, D. P., Green, H. D., & Zhou,
A. (2011). The social context of homeless men’s substance use. Drug and Alcohol
Dependence, 118, 320–325. doi:10.1016/j.drugalcdep.2011.04.011
Rice, E. (2011). HIV risk in large social networks of homeless youth (Grant No.
R01MH093336). Bethesda, MD: National Institute of Mental Health.
Rohrbaugh, M. J., Shoham, V., Skoyen, J. A., Jensen, M., & Mehl, M. R. (2012). We ‐talk,
communal coping, and cessation success in a couple ‐focused intervention for health ‐
compromised smokers. Family Process, 51, 107–121. doi:10.1111/j.1545-
5300.2012.01388.x
Schonberg, T., Fox, C. R., & Poldrack, R. A. (2011). Mind the gap: Bridging economic and
naturalistic risk-taking with cognitive neuroscience. Trends in Cognitive Sciences, 15,
11–19. doi:10.1016/j.tics.2010.10.002
Soar, K., Dawkins, L., Robson, D., & Cox, S. (2020). Smoking amongst adults experiencing
homelessness: A systematic review of prevalence rates, interventions and the barriers and

96

facilitators to quitting and staying quit. Journal of Smoking Cessation, 15, 94–108.
doi:10.1017/jsc.2020.11
Solari, C. D., Cortes, A., Henry, M., Matthews, N., Morris, S., Khadduri, J., & Culhane, D. P.
(2014). The 2013 annual homeless assessment report (AHAR) to Congress, part 2.
Retrieved from https://www.hudexchange.info/onecpd/assets/File/2013-AHAR-Part-
2.pdf
Somers, J. M., Moniruzzaman, A., Patterson, M., Currie, L., Rezansoff, S. N., Palepu, A., &
Fryer, K. (2017). A randomized trial examining Housing First in congregate and scattered
site formats. PLOS One, 12, e0168745. doi:10.1371/journal.pone.0168745
Trimpop, R. (1994). The psychology of risk taking behavior. Amsterdam, Netherlands: Elsevier
Science.
Tsai, J., Mares, A. S., & Rosenheck, R. A. (2012). Does housing chronically homeless adults
lead to social integration? Psychiatric Services, 63, 427–434.
doi:10.1176/appi.ps.201100047
Tsemberis, S. (2010). Housing First: Ending homelessness, promoting recovery and reducing
costs. In I. G. Ellen & B. O’Flaherty (Eds.), How to house the homeless (pp. 37–56). New
York, NY: Russell Sage Foundation.
Tsemberis, S., Gulcur, L., & Nakae, M. (2004). Housing First, consumer choice, and harm
reduction for homeless individuals with a dual diagnosis. American Journal of Public
Health, 94, 651–656. doi:10.2105/ajph.94.4.651
Tucker, J. S., Kennedy, D., Ryan, G., Wenzel, S. L., Golinelli, D., & Zazzali, J. (2009).
Homeless women’s personal networks: Implications for understanding risk behavior.
Human Organization, 68, 129–140. doi:10.17730/humo.68.2.m23375u1kn033518

97

Uchino, B, N., Cacioppo, J. T., & Kiecolt-Glaser, J. K. (1996). The relationship between social
support and physiological processes: A review with emphasis on underlying mechanisms
and implications for health. Psychological Bulletin, 119, 488–531. doi:10.1037/0033-
2909.119.3.488
UCLA Center for Health and Policy Research. (2014). Adult California Health Interview Survey
(2011 ‐2012). Los Angeles, CA: Author.
Watson, D. P., Shuman, V., Kowalsky, J., Golembiewski, E., & Brown, M. (2017). Housing First
and harm reduction: A rapid review and document analysis of the US and Canadian open-
access literature. Harm Reduction Journal, 14, 30. doi:10.1186/s12954-017-0158-x
Weber, M., Thompson, L., Schmiege, S. J., Peifer, K., & Farrell, E. (2013). Perception of access
to health care by homeless individuals seeking services at a day shelter. Archives of
Psychiatric Nursing, 27, 179–184. doi:10.1016/j.apnu.2013.05.001
Wei, Y. D., & Ewing, R. (2018). Urban expansion, sprawl and inequality. Landscape and Urban
Planning, 177, 259–265. doi:10.1016/j.landurbplan.2018.05.021
Wenzel, S. (2005). Alcohol use and HIV risk among impoverished women (Grant No.
5R01AA015301). Bethesda, MD: National Institute on Alcohol Abuse and Alcoholism.
Wenzel, S. L., Rhoades, H., La Motte-Kerr, W., Duan, L., Harris, T., Rice, E., & Henwood, B. F.
(2019). Do HIV risk and prevention behaviors change over time among adults in
permanent supportive housing? AIDS Care, 31, 1172–1177.
doi:10.1080/09540121.2019.1576849
Wenzel, S. L., Rhoades, H., Moore, H., Lahey, J., Henwood, B., La Motte ‐Kerr, W., & Bird, M.
(2018). Life goals over time among homeless adults in permanent supportive housing.
American Journal of Community Psychology, 61, 421–432. doi:10.1002/ajcp.12237

98

Westat & University of Southern California. (2020). Evaluation of strategy D7 to expand and
enhance services provided through permanent supportive housing: Presentation to the
Homelessness Policy Research Institute. Los Angeles, CA: Author.
Westmaas, J. L., Bontemps-Jones, J., & Bauer, J. E. (2010). Social support in smoking cessation:
Reconciling theory and evidence. Nicotine & Tobacco Research, 12, 695–707.
doi:10.1093/ntr/ntq077
Whittaker, E., Dobbins, T., Swift, W., Flatau, P., & Burns, L. (2017). First examination of
varying health outcomes of the chronically homeless according to Housing First
configuration. Australian and New Zealand Journal of Public Health, 41, 306–308.
doi:10.1111/1753-6405.12631
Yanos, P. T., Barrow, S. M., & Tsemberis, S. (2004). Community integration in the early phase
of housing among homeless persons diagnosed with severe mental illness: Successes and
challenges. Community Mental Health Journal, 40, 133–150.
doi:10.1023/B:COMH.0000022733.12858.cb
Yanos, P. T., Felton, B. J., Tsemberis, S., & Frye, V. A. (2007). Exploring the role of housing
type, neighborhood characteristics, and lifestyle factors in the community integration of
formerly homeless persons diagnosed with mental illness. Journal of Mental Health, 16,
703–717. doi:10.1080/09638230701496378
Zgierska, A., Amaza, I. P., Brown, R. L., Mundt, M., & Fleming, M. F. (2014). Unhealthy drug
use: How to screen, when to intervene. Journal of Family Practice, 63, 524–530.

99

Tables and Figures
Table 3.1. Descriptive Statistics of Time-Invariant Measures Examined as Predictors and
Controls in Latent Growth Curve Models

Baseline
(n = 321)
M (SD) or n (%)
Demographic characteristics
Age 54.54 (7.21)
Gender (female) 83 (25.86)
Race

Black or African American) 183 (57.01)
White 78 (24.30)
Other 21 (6.54)
Latino 39 (12.15)
Veteran 100 (31.15)
Education 14.3 (3.68)
Housing model

Single-site 212 (66.04)
Scattered-site 106 (33.02)
Neighborhood

Skid Row 73 (22.74)
Downtown Los Angeles 65 (20.25)
Other 182 (56.70)

100

Table 3.2. Descriptive Statistics of Time-Variant Measures Examined as Predictors of Tobacco
Risk and Controls in Latent Growth Curve Models

Baseline 3 Months 6 Months 12 Months
(n = 321) (n = 304) (n = 299) (n = 287)
M (SD) M (SD) M (SD) M (SD)
Support measures
Income**
b
565.41
(352.92)
556.53
(353.38)
564.81
(355.32)
627.07
(367.28)**
Structural support

Size** 7.71 (4.57) 7.15 (3.90) 6.46 (3.97) 6.56 (3.86)
Relatives** 2.01 (2.34) 2.26 (2.45) 2.36 (2.55) 2.51 (2.88)
Romantic partners 0.42 (0.60) 0.49 (0.76) 0.47 (0.67) 0.43 (0.56)
Neighbors**
a

0.97 (1.46) 0.99 (1.59) 1.06 (1.39)
Veterans* 0.59 (1.11) 0.56 (0.97) 0.46 (0.84) 0.50 (0.91)
Currently homeless** 1.41 (2.07) 0.76 (1.33) 0.41 (0.89) 0.46 (0.88)
Nearby** 2.64 (2.89) 2.46 (2.58) 2.16 ( 2.49) 2.13 (2.20)
Distant 3.63 (3.72) 3.90 (3.55) 3.37 (3.17) 3.43 (3.21)
Functional support

Emotional* 2.92 (2.82) 2.52 (2.58) 2.38 (2.33) 2.55 (3.12)
Tangible** 1.30 (1.78) 1.06 (1.58) 0.73 (1.10) 0.82 (1.33)
Informational** 2.59 (2.62) 1.98 (2.46) 1.62 (2.21) 1.53 (2.06)
Conflict 1.07 (1.59) 0.93 (1.19) 1.04 (1.30) 1.11 (1.52)
Provider support

101

Case manager** 1.16 (1.30) 0.74 (1.13) 0.61 (0.90) 0.53 (0.73)
Doctor 0.32 (0.56) 0.35 (0.61) 0.29 (0.52) 0.32 (0.57)
Counselors** 0.47 (0.74) 0.33 (0.57) 0.33 (0.67) 0.29 (0.57)
Case management visits** 19.89 (22.71) 10.93 (14.88) 7.93 (12.63) 7.73 (12.68)
Health care visits** 4.21 (6.24) 3.59 (5.71) 3.00 ( 3.74) 2.96 (4.53)
Mental health visits** 3.99 (7.87) 3.28 (6.45) 2.93 (6.95) 2.51 (5.64)
*Significant change over time at p < .05.
**Significant change over time at p < .01.
a
Assessed only at follow-up waves (posthousing).
b
Examined as a time-varying control in final conditional growth curve models.

102

Table 3.3. Results from Unconditional Growth Models (Models 1–3)
Model 1 (Fully Unconditional Model)
Effect Estimate SE Pr>|t| 95% CI
Intercept 18.07 0.363 < .01 17.36, 18.78
Model 2 (Unconditional Linear Growth Model)
Intercept 18.43 0.43 < .01 17.98, 19.78
Time -1.80 0.5 < .01 -2.78, -0.82
Model 3 (Unconditional Quadratic Growth Model)
Intercept 18.96 0.47 < .01 18.05, 19.87
Time -1.93 0.55 < .01 -3.01, -0.86
Time 2 0.57 0.18 < .01 0.22, 0.92

103

Table 3.4. Conditional Main Effect Models (Models 4 and 5 [Including Covariates])
Model 4 (Conditional Main Effect Model)
Effect Estimate SE Pr>|t| 95% CI
Scattered-site 0.03 0.78 .97 -1.50, 1.56
DTLA 2.15 1.08 .04 0.03, 4.27
Other 0.96 0.89 .29 -0.80, 2.71
Size 0.38 0.11 < .01 0.17, 0.59
Size Δ 0.15 0.07 .03 0.01, 0.28
Romantic partner 1.63 0.77 .04 0.12, 3.14
Romantic partner Δ 0.21 0.41 .60 -0.58, 1.01
Relative 0.63 0.15 < .01 0.33, 0.93
Relative Δ -0.01 0.13 .92 -0.26, 0.23
Neighbors 0.50 0.39 .20 -0.25, 1.35
Neighbors Δ 0.12 0.18 0.48 -0.22, 0.46
Veterans 1.02 0.44 .02 0.16, 1.89
Veterans Δ 0.32 0.29 .27 -0.25, 0.88
Currently homeless -0.06 0.36 .87 -0.77, 0.65
Currently homeless Δ 0.23 0.15 .13 -0.07, 0.53
Nearby 0.05 0.19 .79 -0.32, 0.43
Nearby Δ 0.12 0.10 .21 -0.07, 0.31
Distant 0.36 0.13 < .01 0.11, 0.61
Distant Δ 0.07 0.09 .40 -0.10, 0.25
Emotional 0.29 0.18 .12 -0.07, 0.65

104

Emotional Δ 0.02 0.09 .84 -0.16, 0.20
Tangible 0.54 0.36 .12 -0.14, 1.22
Tangible Δ 0.05 0.16 .75 -0.25, 0.35
Instrumental 0.56 0.22 < .01 0.14, 0.99
Instrumental Δ 0.03 0.12 .78 -0.19, 0.26
Conflict 0.51 0.37 .17 -0.21, 1.24
Conflict Δ 0.08 0.17 .62 -0.24, 0.41
Case manager 0.35 0.42 .40 -0.47, 1.17
Case manager Δ 0.64 0.24 < .01 0.17, 1.10
Doctor 0.02 0.74 .96 -1.44, 1.48
Doctor Δ -0.22 0.52 .68 -1.24, 0.80
Counselor 1.03 0.78 .19 -0.50, 2.55
Counselor Δ 0.71 0.41 .08 -0.09, 1.50
Case management visits 0.05 0.03 .09 -0.01, 0.12
Case management visits Δ 0.02 0.01 .12 -0.01, 0.04
Health care visits 0.10 0.11 .38 -0.12, 0.32
Health care visits Δ 0.02 0.05 .73 -0.08, 0.12
Mental health visits 0.21 0.06 < .01 0.09, 0.33
Mental health visits Δ 0.05 0.03 .18 -0.22, 0.11
Model 5 (Conditional Main Effect Model with Covariates)
DTLA 2.51 1.07 .02 0.41, 4.60
Other 1.18 0.89 .19 -0.56, 2.92
Size 0.39 0.11 < .01 0.18, 0.61

105

Size Δ 0.08 0.07 .23 -0.05, 0.22
Romantic partner 1.20 0.74 .11 -0.25, 2.66
Romantic partner Δ 0.15 0.39 .71 -0.62, 0.91
Relative 0.55 0.17 < .01 0.20, 0.89
Relative Δ 0.00 0.12 1.00 -0.24, 0.24
Veterans 1.30 0.48 .01 0.36, 2.23
Veterans Δ 0.15 0.28 .61 -0.41, 0.70
Distant 0.39 0.13 < .01 0.14, 0.65
Distant Δ 0.04 0.08 .62 -0.12, 0.20
Instrumental 0.56 0.21 .01 0.14, 0.97
Instrumental Δ -0.04 0.11 .71 -0.25, 0.17
Counselor 0.59 0.28 .03 -0.50, 2.55
Counselor Δ 0.14 0.08 .09 -0.02, 0.30
Case manager 0.44 0.49 .37 -0.52, 1.41
Case manager Δ 0.33 0.24 .17 -0.14, 0.80
Case management visits 0.24 0.07 < .01 0.10, 0.38
Case management visits Δ 0.03 0.03 .36 -0.03, 0.10
Mental health visits 0.24 0.07 < .01 0.10, 0.38
Mental health visits Δ 0.03 0.03 .36 -0.03, 0.10
Note. All effects significant at p < .10 were retained in controlled models and indicated in bold.

106

Table 3.5. Conditional Interactive Models (Models 4 and 5 [Including Covariates])
Interaction with Housing Model
Model 4 (Conditional Interactive Model)
Effect Estimate SE Pr>|t| 95% CI
Size -0.11 0.24 0.64 (-.57, .36)
Size Δ -0.07 0.03 0.04 (-.13, -.01)
Romantic partner 0.69 1.53 0.65 (-2.32, 3.70)
Romantic partner Δ -0.91 0.80 0.26 (-2.47, .66)
Relative 0.25 0.94 0.79 (-1.60, 2.10)
Relative Δ -0.29 0.41 0.48 (-1.09, 0.51)
Neighbors 1.44 1.03 0.16 (-0.59, 3.47)
Neighbors Δ -1.39 0.47 0.00 (-2.31, -0.47)
Veterans 0.83 1.34 0.54 (-1.79, 3.45)
Veterans Δ -0.89 0.81 0.27 (-2.49, 0.71)
Currently homeless 1.26 1.32 0.34 (-1.32, 3.84)
Currently homeless Δ 0.06 0.49 0.90 (-0.89, 1.01)
Nearby 0.89 0.60 0.14 (-0.28, 2.07)
Nearby Δ -0.62 0.28 0.03 (-1.16, -0.08)
Distant -0.25 0.35 0.47 (-0.93, 0.43)
Distant Δ 0.25 0.23 0.27 (-0.19, 0.69)
Emotional -0.21 0.50 0.68 (-1.18, 0.77)
Emotional Δ 0.10 0.25 0.67 (-0.38, 0.59)
Tangible -1.84 1.14 0.11 (-4.07, 0.39)
Tangible Δ -0.58 0.44 0.19 (-1.45, 0.28)
Informational 0.60 0.60 0.32 (-0.57, 1.77)
Informational Δ 0.16 0.31 0.60 (-0.44, 0.76)
Conflict 1.31 1.02 0.20 (-0.68, 3.31)
Conflict Δ -0.80 0.42 0.06 (-1.63, 0.03)
Case manager -0.18 1.09 0.87 (-2.32, 1.95)
Case manager Δ 0.56 0.54 0.30 (-0.51, 1.62)

107

Doctor -0.71 2.09 0.74 (-4.81, 3.39)
Doctor Δ 1.31 1.20 0.27 (-1.03, 3.66)
Counselor 1.56 2.09 0.46 (-2.54, 5.65)
Counselor Δ 1.21 1.05 0.25 (-0.84, 3.26)
Case management visits 0.04 0.09 0.68 (-0.13, 0.21)
Case management visits Δ 0.01 0.03 0.73 (-0.05, 0.07)
Healthcare visits 0.02 0.26 0.93 (-0.49, 0.54)
Healthcare visits Δ 0.13 0.13 0.30 (-0.12, 0.39)
Mental health visits -0.15 0.21 0.47 (-0.56, 0.26)
Mental health visits Δ 0.01 0.11 0.94 (-0.20, 0.22)
Model 5 (Conditional Interactive Model with Controls)
Effect Estimate SE Pr>|t| 95% CI
Size -0.11 0.24 0.66 (-0.57, 0.36)
Size Δ -0.07 0.03 0.04 (-0.13, -.99)
Relative 0.01 0.39 0.99 (-0.75, 0.76)
Relative Δ 0.05 0.15 0.76 (-0.25, .34)
Neighbors 0.78 0.92 0.40 (-1.03, 2.59)
Neighbors Δ -0.87 0.42 0.04 (-1.69, -0.05)
Currently homeless 0.02 0.93 0.99 (-1.81, 1.85)
Currently homeless Δ -0.03 0.33 0.93 (-0.68, 0.62)
Nearby -0.13 0.53 0.80 (-1.17, 0.90)
Nearby Δ -0.45 0.21 0.03 (-0.86, -0.04)
Distant -0.06 0.27 0.82 (-0.60, 0.47)
Distant Δ 0.30 0.17 0.08 (-0.03, 0.64)
Informational 0.15 0.53 0.78 (-0.89, 1.18)
Informational Δ 0.31 0.26 0.23 (-0.20, 0.81)
Conflict 0.64 0.77 0.41 (-0.87, 2.16)
Conflict Δ -0.14 0.33 0.66 (-0.78, 0.49)
Case manager -0.22 1.09 0.84 (-2.36, 1.90)
Case manager Δ 0.32 0.27 0.23 (-0.20, 0.86)
Counselor 0.64 0.77 0.41 (-0.87, 2.16)

108

Counselor Δ -0.14 0.33 0.66 (-0.78, 0.49)

Interaction with Neighborhood - DTLA
Model 4 (Conditional Interactive Model)
Effect Estimate SE Pr>|t| 95% CI
Size -0.52 0.33 0.11 (-1.16, .13)
Size Δ 0.16 0.21 0.45 (-.27, .45)
Romantic partner -2.49 2.39 0.30 (-7.18, 2.20)
Romantic partner Δ -0.46 1.97 0.81 (-4.30, 3.39)
Relative -0.91 0.51 0.07 (-1.91, 0.08)
Relative Δ -0.01 0.48 0.98 (-0.95, 0.92)
Neighbors 2.79 1.14 0.02 (0.55, 5.03)
Neighbors Δ 0.49 0.53 0.36 (-0.55, 1.54)
Veterans 0.56 1.63 0.73 (-2.63, 3.75)
Veterans Δ 0.43 0.90 0.64 (-1.33, 2.18)
Currently homeless -1.21 0.98 0.22 (-3.14, 0.72)
Currently homeless Δ 0.04 0.42 0.92 (-0.78, 0.86)
Nearby 0.12 0.55 0.82 (-0.95, 1.20)
Nearby Δ 0.08 0.28 0.77 (-0.46, 0.62)
Distant -0.86 0.41 0.03 (-1.66, -0.07)
Distant Δ 0.01 0.26 0.96 (-0.50, 0.53)
Emotional -0.09 0.59 0.88 (-1.24, 1.07)
Emotional Δ 0.17 0.49 0.73 (-0.79, 1.14)
Tangible -0.13 1.04 0.90 (-2.17, 1.90)
Tangible Δ -0.10 0.43 0.82 (-0.95, 0.75)
Informational -1.26 0.68 0.06 (-2.58, 0.07)
Informational Δ -0.02 0.30 0.96 (-0.60, 0.56)
Conflict -1.09 1.25 0.38 (-3.54, 1.36)
Conflict Δ -1.26 0.59 0.03 (-2.42, -0.10)
Case manager -1.44 1.39 0.30 (-4.16, 1.27)
Case manager Δ 0.12 0.62 0.84 (-1.10, 1.34)

109

Doctor -2.61 2.67 0.33 (-7.84, 2.61)
Doctor Δ -2.49 1.77 0.16 (-5.96, 0.98)
Counselor -4.44 2.41 0.07 (-9.16, 0.28)
Counselor Δ -0.79 1.38 0.57 (-3.50, 1.92)
Case management visits -0.03 0.12 0.79 (-0.27, 0.20)
Case management visits Δ 0.01 0.04 0.85 (-0.08, 0.09)
Healthcare visits -0.58 0.45 0.20 (-1.47, 0.31)
Healthcare visits Δ -0.08 0.21 0.71 (-0.50, 0.34)
Mental health visits -0.02 0.31 0.96 (-0.63, 0.60)
Mental health visits Δ 0.08 0.12 0.53 (-0.16, 0.32)
Model 5 (Conditional Interactive Model with Controls)
Effect Estimate SE Pr>|t| 95% CI
Size -0.63 0.32 0.05 (-1.27, 0.00)
Size Δ 0.16 0.21 0.44 (-0.25, .58)
Relative -1.00 0.53 0.06 (-2.04, 0.03)
Relative Δ 0.03 0.48 0.94 (-0.90, .97)
Neighbors 2.56 1.15 0.03 (0.31, 4.81)
Neighbors Δ 0.44 0.53 0.41 (-0.61, 1.49)
Currently homeless -1.16 0.98 0.24 (-3.08, 0.76)
Currently homeless Δ 0.00 0.42 1.00 (-0.82, 0.82)
Nearby 0.03 0.54 0.96 (-1.04, 1.09)
Nearby Δ 0.07 0.28 0.80 (-0.47, 0.62)
Distant -0.98 0.40 0.01 (-1.76, -.20)
Distant Δ 0.02 0.26 0.95 (-0.50, 0.53)
Informational -1.03 0.68 0.13 (-2.36, 0.30)
Informational Δ 0.00 0.30 1.00 (-0.58, 0.58)
Conflict -0.99 1.25 0.43 (-3.44, 1.46)
Conflict Δ -1.29 0.59 0.03 (-2.45, -0.13)
Case manager -1.57 1.39 0.26 (-4.29, 1.15)
Case manager Δ 0.17 0.62 0.77 (-1.04, 1.40
Counselor -5.29 2.37 0.03 (-9.94, -0.64)

110

Counselor Δ -0.76 1.38 0.58 (-3.47, 1.94)

Interaction with Neighborhood - Other
Model 4 (Conditional Interactive Model)
Effect Estimate SE Pr>|t| 95% CI
Size -0.33 0.32 0.29 (-0.95, .13)
Size Δ 0.09 0.18 0.63 (-0.27, 0.45)
Romantic partner -0.58 2.40 0.81 (-5.28, 4.11)
Romantic partner Δ -1.51 1.17 0.20 (-3.82, 0.79)
Relative 0.07 0.38 0.85 (-0.68, 0.82)
Relative Δ -0.53 0.32 0.10 (-1.16, 0.10)
Neighbors 0.30 1.03 0.77 (-1.73, 2.33)
Neighbors Δ 1.11 0.50 0.03 (0.13, 2.09)
Veterans 0.20 1.66 0.90 (-3.05, 3.46)
Veterans Δ 0.35 0.84 0.68 (-1.31, 2.00)
Currently homeless -2.21 1.29 0.09 (-4.73, 0.32)
Currently homeless Δ -0.11 0.48 0.82 (-1.06, 0.84)
Nearby -1.10 0.55 0.04 (-2.18, -0.03)
Nearby Δ 0.31 0.27 0.24 (-0.21, 0.84)
Distant -0.22 0.39 0.58 (-0.99, 0.55)
Distant Δ 0.05 0.24 0.83 (-0.42, 0.52)
Emotional -0.13 0.31 0.68 (-0.73, 0.47)
Emotional Δ 0.01 0.24 0.98 (-0.46, 0.47)
Tangible 1.20 1.14 0.29 (-1.02, 3.43)
Tangible Δ 0.29 0.44 0.51 (-0.57, 1.15)
Informational -1.15 0.63 0.07 (-2.38, 0.08)
Informational Δ 0.25 0.28 0.39 (-0.31, 0.80)
Conflict -1.55 1.26 0.22 (-4.02, 0.91)
Conflict Δ 0.28 0.58 0.63 (-0.86, 1.42)
Case manager -3.17 1.29 0.01 (-5.70, -0.63)
Case manager Δ 0.51 0.64 0.42 (-0.74, 1.77)

111

Doctor -3.24 2.59 0.21 (-8.31, 1.83)
Doctor Δ 0.32 1.53 0.84 (-2.68, 3.32)
Counselor -4.19 2.49 0.09 (-9.08, 0.70)
Counselor Δ -0.70 1.20 0.56 (-3.05, 1.64)
Case management visits -0.05 0.09 0.61 (-0.23, 0.13)
Case management visits Δ 0.05 0.04 0.18 (-0.02, 0.13)
Healthcare visits -0.15 0.39 0.70 (-0.91, 0.61)
Healthcare visits Δ -0.17 0.18 0.36 (-0.52, 0.19)
Mental health visits -0.05 0.25 0.83 (-0.54, 0.43)
Mental health visits Δ 0.09 0.09 0.31 (-0.09, 0.27)
Model 5 (Conditional Interactive Model with Controls)
Effect Estimate SE Pr>|t| 95% CI
Size -0.40 0.26 0.13 (-0.92, 0.12)
Size Δ 0.03 0.16 0.83 (-0.27, 0.34)
Relative -0.18 0.38 0.63 (-0.92, 0.56)
Relative Δ -0.38 0.28 0.17 (-0.93, 0.17)
Neighbors 0.67 0.96 0.49 (-1.21, 2.55)
Neighbors Δ 0.57 0.47 0.22 (-0.35, 1.49)
Currently homeless -1.48 0.99 0.14 (-3.43, 0.47)
Currently homeless Δ -0.07 0.35 0.83 (-0.77, ,0.62)
Nearby -0.88 0.49 0.07 (-1.83, 0.08)
Nearby Δ -0.07 0.21 0.74 (-0.49, 0.34)
Distant -0.32 0.33 0.33 (-0.96, 0.32)
Distant Δ 0.20 0.19 0.29 (-0.17, 0.57)
Informational -1.07 0.56 0.06 (-2.16, 0.02)
Informational Δ 0.32 0.25 0.19 (-0.16, 0.81)
Conflict -0.68 1.04 0.52 (-2.72, 1.37)
Conflict Δ -0.26 0.50 0.60 (-1.24, 0.71)
Case manager -2.44 1.07 0.02 (-4.54, -.34)
Case manager Δ 0.58 0.50 0.25 (-0.40, 1.57)
Counselor -3.74 2.05 0.07 (-7.77, 0.29)

112

Counselor Δ 0.04 1.00 0.97 (-1.93, 2.00)
Note. All effects significant at p < .10 were retained in controlled models and indicated in bold.

113

Figure 3.1. Linear and Quadratic Growth in Tobacco Risk as Adults Transition from
Homelessness to Supportive Housing

114

Figure 3.2. Difference in Slopes of Network Size over Time across Housing Models

115

Figure 3.3. Difference in Slopes Representing Change in Neighbors in Networks over Time
across Housing Model

116

Figure 3.4. Difference in Slopes of Change in Nearby Network Members across Single-Site and
Scattered-Site Residents

117

Figure 3.5. Differing Slopes Depicting Change in Conflicting Network Members across Skid
Row and DTLA PSH Residents

118

Figure 3.6. Slopes Representing Change in Neighbors across Skid Row and Other Neighborhood
PSH Residents

119

Chapter 4: Barriers and Facilitators to Cessation among Supportive Housing Residents: A
Qualitative Inquiry
Introduction
National estimates suggest that 1 in 5 adults smoke cigarettes (Centers for Disease
Control and Prevention, 2020), whereas smoking among persons experiencing homelessness is
estimated to be between 60% and 80% (Vijayaraghavan, Elser, & Apollonio, 2019). Smoking
exacerbates the risk and rates of chronic health conditions and death (U.S. Department of Health
and Human Services, 2014). Just as persons experiencing homelessness smoke at higher rates
than the housed adult population, so too are they affected by tobacco use at greater rates. Due to
risky tobacco practices (e.g., sniping [smoking discarded cigarettes], sharing cigarettes, remaking
cigarettes; Tucker, Shadel, Golinelli, Mullins, & Ewing, 2015; Vijayaraghavan et al., 2018) and a
high rate of already compromised health, tobacco-related conditions are the leading cause of
death for persons experiencing homelessness (Vijayaraghavan et al., 2018). Moreover, rates of
tobacco-related mortality among persons experiencing homelessness are double that of the
housed population (Vijayaraghavan et al., 2016).
Several studies have found that persons experiencing homelessness express a desire to
quit and are motivated to cease smoking (Businelle et al., 2014; Okuyemi et al., 2006); however,
their rate of cessation is markedly lower than that of the housed adult population (Businelle,
Cuate, Kesh, Poonawalla, & Kendzor, 2013; Vijayaraghavan et al., 2018). This population faces
numerous socioecological barriers to cessation while homeless, including the experience of
homelessness itself (Chen, Nguyen, Malesker, & Morrow, 2016). Homeless persons’ increased
rates of smoking and reduced success in ceasing smoking is partly associated with poor mental
health and other substance use disorders in this population (Baggett, Tobey, & Rigotti, 2013;

120

Harris, Winetrobe, Rhoades, & Wenzel, 2019). Additionally, social impediments hinder
cessation, including but not limited to the normative acceptability of tobacco use (Businelle et
al., 2013; Okuyemi et al., 2006), pressure to smoke in social situations (Pratt et al., 2019),
diminished access to routine provider support, and low prioritization of cessation among
providers (Antoon, 2017). Cessation is also markedly more difficult due to physical
environmental barriers, such as housing instability, limited transportation, excess distance to
health care and cessation treatments, and living in areas where access to cigarettes is abundant
(Antoon, 2017; Baggett et al., 2013).
Permanent supportive housing (PSH) provides adults experiencing homelessness with the
opportunity to overcome many physical, social, and individual barriers faced while homeless.
PSH is considered the superior approach to ending homelessness by providing homeless adults
with the combination of affordable, long-term housing and supportive services (Tsemberis,
Gulcur, & Nakae, 2004). As homeless adults transition to PSH, they potentially experience
increased housing stability and security, enriched social support by way of increased access to
health care providers and other prosocial relationships (Tsemberis et al., 2004), and improved
health and mental health (Harris, Rhoades, Duan, & Wenzel, 2019). Yet research suggests that
environmental barriers may remain, including concerns about neighborhood safety and security
in PSH (Henwood, Lahey, Harris, Rhoades, & Wenzel, 2018) and decreased social networks and
social support in housing (Harris, Rhoades et al., 2019), which may perpetuate barriers to
cessation in housing.
The social ecological model provides a conceptual guide for understanding factors that
affect individual health behavior and changes in these behaviors (Bronfenbrenner, 1977; Visser,
2007). According to the social ecological model, the ecosystem has multiple levels that, for the

121

purposes of this study, include three integral levels: institutional, interpersonal, and
intrapersonal, corresponding to the housing environment, social environment, and individual,
respectively. The socioecological levels are nested in one another such that an individual exists
in a social environment, which exists in the housing environment. These levels could be
considered independently; however, this would overlook the interconnectivity of these levels that
make up a system (Macintyre & Ellaway, 2000). Using a socioecological approach allows for
individual health behaviors to be understood in the context of their environments, which allows
for modifying individual behavior via targeted efforts at multiple levels rather than solely
targeting the individual level (Stokols, 1996).
This conceptual model is particularly pertinent for this study because the environmental
contexts of PSH can vary. The housing environment can vary across housing models, with two
predominant models of PSH seen across the United States: scattered-site and single-site housing.
Residents in scattered-site housing live in existing market-rental accommodations, where they
are scattered among buildings with members of the general population, whereas single-site
housing residents are clustered in a building designated for persons experiencing homelessness
(Somers et al., 2017). Neighborhoods can vary in characteristics, and the literature on tobacco
use has highlighted the associations between smoking and number of tobacco retailers (Pearce,
Hiscock, Moon, & Barnett, 2009; Pulakka et al., 2016), marketing (Lee, Henriksen, Rose,
Moreland-Russell, & Ribisl, 2015), social norms and practices (Karasek, Ahern, & Galea, 2012),
and concerns about neighborhood safety (Patterson, Seravalli, Hanlon, & Nelson, 2012).
As the social ecological model suggests, these housing environments are connected to
and affect residents’ social environments. Regarding PSH, this is evident in how housing models
and neighborhoods affect delivery and receipt of services. Although a pillar of all PSH,

122

supportive services can also vary such that services exist on site for some and off site for others.
Although Housing First guidelines indicate scattered-site residents should receive wraparound
care on site (Tsemberis et al., 2004), this practice is more prevalent in some regions than others
(Henwood, Harris, et al., 2018). For single-site residents, access to on-site case management
services is generally standard, and some have mental health and medical services delivered on
site by visiting providers or in the case of some newer buildings, in an on-site clinic (Henwood
Harris et al., 2018). For residents in neighborhoods where access to services is limited or
proximally challenging, access to routine care and cessation treatments may be a barrier (Billi,
Pai, & Spahlinger, 2007; Tailor et al., 2019). In addition to the potential variation in support
from providers, residents likely have differences in their relationships with friends, families, and
neighbors and may be affected by their physical environments. For example, research has shown
that residents living in Skid Row, the neighborhood with the highest concentration of homeless
persons and services in the nation, report higher rates of social support compared to those living
in other Los Angeles neighborhoods, which is likely a function of the number of social services
and the street community concentrated in Skid Row (Harris Dunton et al., 2019). Unfortunately,
PSH residents across contexts report a great deal of conflict in their relationships, and the
struggles of distancing themselves from negative social influences while generating prosocial
ties and thwarting isolation in housing has been an ongoing challenge (Harris Rhoades et al.,
2019; Henwood, Matejkowski, Stefancic, & Lukens, 2014; Padgett, Henwood, Abrams, &
Drake, 2008; Yanos, Felton, Tsemberis, & Frye, 2007).
The varying contexts of PSH may affect motivation and confidence to quit smoking.
Although the social ecological model purports that housing and social environments and their
interrelation ultimately affect individual health behavior, the individual level is nonetheless an

123

integral consideration (Bronfenbrenner, 1977). The intrapersonal level is the epicenter of the
ecosystem, with individual attitudes and beliefs serving as critical components of this level
(Visser, 2007). Whereas an outsider could hypothesize what barriers might exist for residents in
PSH, this is an opportunity to give a voice to a marginalized, underserved population. Very little
research exists on tobacco use behavior among PSH residents, making this an opportune time to
examine this issue qualitatively. Although there has been some qualitative exploration of PSH
residents’ perceptions about smoke-free policies in housing (Vijayaraghavan, Hurst, & Pierce,
2017), this was the first study to explore numerous barriers to and facilitators of cessation at
multiple levels of the socioecological system.
Drawing on data from 33 semistructured interviews with residents in varying supportive
housing models and neighborhoods across Los Angeles County, this study explored housing,
social, and individual barriers to and facilitators of cessation among tobacco users. Purposively
sampled from a longitudinal cohort study of residents transitioning from homelessness to
housing, these residents also presented with differential tobacco use trajectories across their
initial year in supportive housing. Incorporating individual perspectives using a qualitative
approach allows for novel information to emerge (Creswell, Shope, Plano Clark, & Green,
2006), such as facilitators and hindrances of tobacco use change, and identifying subgroups that
increased, decreased, and maintained tobacco use enhances the ability to understand these
factors. Using qualitative methods to gather perceptions and expectations of tobacco use
behavior across subgroups, value-laden questions can inform intervention efforts of facilitators
of and barriers to cessation.

124

Method
This study drew on existing longitudinal quantitative data from a parent study (NIDA
Grant No. R01DA036345). Quantitative data from the parent study were used to purposively
sample participants for the current study. All procedures for the parent and current study were
approved by the University of Southern California’s Institutional Review Board.
Parent Study
The Transitions study is a prospective study with four repeated measurements: before
entering PSH (baseline, N = 421) and 3, 6, and 12 months postbaseline. Participants were
unaccompanied adults placed in PSH between August 2014 and January 2016. Eligibility criteria
were that participants were confirmed as entering PSH in Los Angeles County, were 39 years of
age or older, and could complete an interview in English or Spanish. To recruit this cohort, the
study team collaborated with 26 organizations that provide both scattered-site and single-site
models of PSH throughout the county and constitute the majority of PSH providers in the region.
The primary focus of the parent study was HIV risk and prevention; therefore, the age and
nonparenting requirements were implemented to maximize the ability to detect changes in HIV
risk outcomes by minimizing variability due to developmental life stage or current parenting
status.
Interviews across all timepoints consisted of a questionnaire and social network interview
that together took approximately 1.5 hours to complete. Interviews were conducted one-on-one
by trained study personnel who collected respondents’ answers with iPads using Qualtrics survey
software for the questionnaire and a social network data collection app. Participants received $20
at baseline with increased incentives in $5 increments at each follow-up (3 months = $25, 6

125

months = $30, 12 months = $35). Surveys covered topics including housing information, risk
exposure, social support, mental health, substance use, and service needs and utilization.
Measures
Tobacco use was self-reported across all points of the quantitative survey using the
NIDA-modified Alcohol, Smoking and Substance Involvement Screening Test (National
Institute on Drug Abuse, 2012). Respondents reported whether they used tobacco never, once or
twice, monthly, weekly, or daily or almost daily in the past 3 months. These data were used to
purposively sample residents who increased, decreased, or maintained their tobacco use across
their initial year in housing.
Participants’ gender, neighborhoods, and type of housing were assessed for maximum
variation sampling and to achieve an equal distribution across these factors. Maximum variation
sampling was employed to enhance the understanding of the phenomenon of cessation in
supportive housing across diverse residents, diverse tobacco use patterns, and diverse housing
and social settings (Palinkas et al., 2016). Baseline data provided participants’ gender
identification (male, female, and male-to-female transgender). Drawn from the posthousing data,
participants’ addresses were used to create an indicator of single-site or scattered-site housing
settings. These addresses were also used to create a measure of neighborhood; that is, whether
residents were housed in Skid Row versus other neighborhoods, based on city-established
neighborhood borders and prior identification of Skid Row borders (Hsu, Simon, Henwood,
Wenzel, & Couture, 2015; Solari, Cortes, Henry, Matthews, & Morris, 2014).
Current Study
This study used a multiple case study approach to understand the barriers to and
facilitators of cessation across multiple socioecological levels. Multiple case studies are

126

employed when multiple cases exist across varying contexts and the researcher desires to
understand differences and similarities across cases (Yin, 2003). Evidence from this approach is
considered strong and reliable (P. Baxter & Jack, 2008). Theoretical replication is used to select
and examine cases, with the goal of building conceptual understanding beyond the patterns
initially identified to select them (Yin, 1994). Although cases in qualitative methodology are
often thought of as a place or group, making it difficult to use a multiple case study approach due
to extensive costs and effort, cases can also be individuals, allowing for these barriers to be
overcome with more ease (Yin, 1994; Zach, 2006).
This study drew on a purposive sample of Transitions participants who endorsed tobacco
use on the NIDA-modified Alcohol, Smoking and Substance Involvement Screening Test at one
or more waves and agreed to be contacted for further follow-up studies at the 12-month survey.
Tobacco use data from the baseline and 12-month surveys were required to be eligible.
Participants’ longitudinal tobacco use frequency data were used to determine whether they
increased (e.g., from weekly to daily), decreased (e.g., from daily to monthly), or maintained
their tobacco use over time. As is common in tobacco use behavior, some demonstrated mixed
change rather than a consistent decrease or increase; therefore, the final determination of their
category was based on their baseline and 12-month data.
Of 362 Transitions participants who agreed at the 12-month interview to be contacted for
follow-up studies, 277 (77%) endorsed tobacco use in at least one wave, which served as the
final number of eligible participants for the current study. Eligible participants were initially
stratified by their tobacco use pattern over time, followed by stratification by gender and
neighborhood (Skid Row versus other). Randomization was used for purposes of ordering
recruitment calls. Participant identification numbers were randomized such that the group of

127

women living in Skid Row who increased their tobacco were randomized separately from the
group of women living in other neighborhoods who increased their tobacco use to achieve the
maximum variation sampling approach. Additionally, given the limited number of interviews to
be conducted and a relatively large pool to sample from, an attempt was made to create an equal
opportunity for participation in this current study.
Participants were recruited by phone. Study personnel provided a description of study
procedures and scheduled the in-person interview. During the recruitment call, study personnel
reminded participants about previous interviewers the participant had met as part of the parent
study protocol. It was explained that participation was dependent on their willingness to engage
in a recorded interview and if they did not wish to be recorded, then they could not participate in
the current study. If participants endorsed an interest and willingness to participate, an in-person
interview was scheduled.
All qualitative interviews were conducted by the female principal investigator (PI) or a
trained male interviewer. Gender matching was used whenever possible. The PI was trained
during the parent study and has experience collecting quantitative and qualitative data with this
population and many of the participants in the parent sample. The male interviewer had
extensive experience working with persons experiencing homelessness and collecting qualitative
data. Interviews often took place in participants’ apartment or building lobby, according to
participant preference. Interviews were conducted in a private place if they were conducted
outside of the participant’s apartment. Participants were reminded that identifying information
would be removed from the transcripts and ensured that their information was confidential.
Interviewers communicated that if information was shared about residents’ violation of the
building smoking policies or rules, it would remain confidential and would never jeopardize their

128

housing status. Prior to the interview, participants were informed of their rights to confidentiality
and intended benefits of the study and provided written consent. Interviews lasted approximately
40 minutes to 1 hour and were audio recorded. A physical copy of the interview protocol was
used for each participant, allowing the PI to take individual notes throughout each interview. The
interviewer used prompts throughout the interview and collected brief field notes. All
participants received $25 for their participation in semistructured interviews.
The use of a semistructured interview allows respondents to provide open-ended
responses that can provide meaningful insight (Barnard, 1999). Interviewers used guided open-
ended questions that addressed potential barriers to and facilitators of cessation. Following
standard guidelines of ordering conversation questions, interview questions followed this order:
(a) tobacco use practices (e.g., use of electronic smoking devices) and cessation attempts in PSH;
(b) general perspective of what facilitates and hinders cessation; (c) social environment
facilitators of or hindrances to cessation (i.e., structural, functional, and provider support); (d)
physical environment facilitators of or hindrances to cessation (e.g., housing site, proximity to
services, neighborhood and building quality); (e) policies that affect cessation (e.g., recent
tobacco tax increases, building smoking policies); (f) perspectives on cessation aids; and (g)
intervention components perceived to aid in successful cessation.
Analysis
All qualitative interviews were recorded and transcribed verbatim. An initial thematic
analytic approach was used, in which portions of meaningful text were organized and
summarized by salient themes (King, 2004). The PI and a trained master’s-level social work
student independently coded three transcripts initially, followed by a face-to-face meeting to
compare codes and develop a catalog of codes. Meaningful text was determined by a process of

129

hierarchical coding, which began with identifying unifying predefined codes (e.g., functional
support, barrier to cessation) or codes derived from the data, followed by the development of
themes and subthemes that emerged through exploration. The remaining transcripts were coded
separately and then compared. Consensus regarding coding, coding procedures, and codebook
modifications occurred during regular meetings. For each participant, thematic data at each
socioecological level (housing, social, and individual) were summarized using a thematic matrix
(Miles, Huberman, & Saldaña, 2013). Emblematic quotes were included to represent common
opinions. This was followed by multiple case study analysis to examine coded segments in and
across subgroups (i.e., gender, housing model, and location). This process helped visually
present differences and decipher patterns across codes as analysis of the matrix occurred.
Results
Descriptive statistics of the subsample of residents that participated in the current study
are presented across genders in Table 4.1. Fifteen women and 18 men participated in
semistructured interviews. Most women (73%) identified as Black, as did the majority of men
(61%) who participated. Of the residents that fell into the “other” racial category, one female
participant identified as mixed race (nonwhite and Latino), as did one male participant, and
another female participant identified as Asian. Among women who participated, 47% were
housed in scattered-site units and 80% resided outside of the Skid Row neighborhood. Among
men in the subsample, 39% resided in scattered-site housing and 50% lived in housing in Skid
Row. The quantitative data gathered during residents’ initial year in PSH and used to purposively
sample indicated that across all 33 men and women in the subsample, 11 increased their tobacco
use across their first year in housing, whereas nine decreased and 12 maintained their smoking
over time. At the semistructured interview, when residents had been housed for 3–5 years, eight

130

reported that their tobacco use had increased since entering housing, 17 reported a decrease, and
eight reported their smoking had been maintained.
Analysis of the 33 semistructured interviews yielded approximately 20 codes related to
barriers to and facilitators of cessation. These codes were systematized across three broad
themes, according to level of the socioecological system (housing environment, social
environment, and individual). Subthemes identified as facilitators or barriers in each
socioecological level were distinguished and are described here, along with differences across
neighborhood, housing model, and gender.
Housing Environment: Facilitators
Accessibility. For some residents, limited access to cigarettes facilitated their ability to
cease smoking or reduce their tobacco use. Residents living in neighborhoods where stores were
not easy to access described this as a facilitator, often attributing it to a lack of transportation,
motivation, or health capacity to travel to a store. This was more common among residents living
in neighborhoods outside of Skid Row, which tended to be residential neighborhoods on the
fringes of the city. A scattered-site resident who had a recent double mastectomy described “the
task of putting on [her] prosthetics” as a deterrent to making the trip to the store to get cigarettes.
A partially blind male single-site participant living outside of Skid Row purchased cigarettes
biweekly on his trips to the store with his home health worker and would occasionally ask
neighbors to purchase cigarettes. He recognized the process involved in acquiring cigarettes
decreased his motivation to smoke: “The reason why I knew that, if I run out of cigarettes in here
and then nobody will go get me none, it don’t bother me.”
For other residents, decreased accessibility was attributed to their building and a lack of
space to smoke.

131

But then a lot of times I don’t feel like going up to the fourth floor, the fifth floor, then
see you can’t smoke down here on the patio and stuff like that, so. Kinda sorta deters me
from, if I don’t feel like going out that day or whatever. So, then I won’t smoke, yeah.
Housing as a facilitator of cessation was most often discussed in relation to smoking
policies and restrictions in supportive housing buildings. In fact, living in a building with a
nonsmoking policy was one of the most commonly reported facilitators of reductions in smoking
in the subgroup that reduced smoking in the initial year in housing. Generally, residents
expressed a deep respect for the rules outlined in their lease and understood that if they violated
them, they would risk losing their housing.
Although there may have been nearby options to smoke outside, many residents disliked
this option. Residents living in single-site housing settings with a designated smoking area often
did not like the forced social interaction that came with being among other smokers in the
building. Single-site housing residents also tended to live in larger buildings, many with
numerous levels, and the time and travel required to get outside curbed their smoking frequency.
A female resident living in single-site housing outside of Skid Row who had quit a year before
the interview discussed the impact of her housing on her cessation:
The way my housing is actually structured I believe [has had an impact]. Because they
have like a plaza area that are designated for smoking. I think maybe when we moved in,
they preferred that we smoked out, so by me not going out there, that also contributed to
me not smoking. If I were in that smoking area constantly, I would be probably involved
on wanting to smoke more.
Residents in scattered-site housing were often in buildings lacking a designated smoking
area, and they also tended to live among more nonsmokers than those in single-site housing. A

132

participant in a scattered-site building stated why she was hesitant to smoke outside: “I don’t
want my neighbors to see me smoking. You know, I’m the new girl.” The limited options for
spaces to smoke meant that their smoking would be more exposed, and this was a concern due to
fear of being judged by their neighbors.
Regardless of the type of housing model, residents in buildings with nonsmoking policies
in their units and around the building, which meant that they did not have a designated smoking
area, attributed this as a facilitator of cessation. Overall, most residents who participated in the
semistructured interviews had some sort of smoking restriction, either in their individual unit or
their building. Several participants, however, did not have such restrictions in place, but
nonetheless regarded their space highly and did not want to contaminate it with the smell of
smoking. A resident living in a scattered-site unit with her adult daughter talked about her desire
to maintain her space:
I know it’s an offensive odor, because the people back here smoke and on a nice day
they’ll be smoking, it fills the whole house up. I’m like, “Oh, what a horrendous smell.
But, see now, I’m … This is my place, I’m, this is my place of where I’m, my habitat
right now. You know and I don’t want to walk in here and smell stale old cigarette
smoke.
Whether residents chose not to smoke in their units or were restricted from doing so,
many opted not to seek an outside location because they were limited by physical health
problems.
Serenity versus chaos. Residents who lived in a neighborhood or building that they
deemed as peaceful or promoted calmness accredited their reductions in smoking to their
housing environment. Although there were no differences observed across housing models, no

133

residents living in Skid Row viewed their housing environment as peaceful or calm. Residents
who perceived their housing environments as serene often lived in newer buildings and in less
marginalized areas of the city. A female resident living in a newer single-site housing building
attributed her environment to reducing her smoking. “This is one of the better places that I’ve
lived in some aspects, because we have every trees, services around, it’s clean. We have
everything, mental health—that’s really for our advantage to become better.”
Many other residents expressed concerns about neighborhood safety, indicating that they
generally avoided going outside out of fear and avoidance of crime and chaos. This tendency to
isolate facilitated cessation and was common among residents living in Skid Row or other
neighborhoods. A resident living in Skid Row acknowledged his isolative behaviors, suggesting
they were worth the health benefits of not smoking: “I’m able to rationalize, you know, you got a
lot of less than healthy people in this area. … So, I like people, but I know some are not good for
my health.”
Housing Environment: Barriers
Accessibility. Although some residents struggled to access cigarettes or spaces to smoke,
another substantial portion of residents that felt cigarettes were too accessible in their
neighborhoods. A male participant attributed the maintenance of his tobacco use to this: “It
hasn’t changed, ‘cause when I walk out, I smell it. They’re selling it all over the place.” This was
primarily reported by residents living in or around Skid Row, because street selling of black-
market cigarettes and “loosies,” or single cigarettes, is prevalent in this area and substantially
cheaper than purchasing a pack of cigarettes from a store. It also was attributed to maintaining
tobacco use, even at low rates, because residents could easily purchase loosies. Another male

134

participant living in Skid Row described the ease and affordability of his process of getting
cigarettes:
This is downtown L.A. Skid Row … and we have sellers down here on the street that sell
cigarettes. Loose cigarettes or—they sell them loose, they sell packs, they sell cartons. I
have a lady … who I call the first of the month, and she brings me the five cartons, I pay
her, and that’s it. You’d figure … brand names, now, you’re paying a minimum of 50 to
60 dollars a carton. Through her, I’m getting them for 30.
Serenity versus chaos. Although some residents in chaotic or risky environments
isolated themselves, which facilitated cessation behaviors, others felt that life in these settings
served as a barrier to cessation. This was particularly salient for residents living in Skid Row. A
participant described the atmosphere and daily life of Skid Row:
The neighborhood is crazy. It is absolutely insane. I’ve never seen a neighborhood like
this. Anywhere I’ve lived in the United States, I haven’t seen a neighborhood like this.
This is … what they call Skid Row, and if you hang around for a little bit and watch,
you’ll understand. Police drive by … standing right on the corner, you got drug pushers,
and they ain’t afraid of the police, and they’ll yell at you and say, “Hey! Weed! You want
to try? You need weed?” With cops parked right at the curb. I was walking between this
guy and the police. They weren’t doing anything, they were just sitting there. They
weren’t even talking, because I looked at them. And the guy standing there says, “Hey,
what do you need? You need rock? You need some marijuana? I got meth.” … They
ignored him. They just sat there, never said a word. Didn’t … do anything. I mean, this
neighborhood is crazy.

135

Residents in neighborhoods with high crime described the stress associated with their
concerns about safety and risk, which subsequently increased or maintained their tobacco use
and made quitting harder. A participant described his frustration and ensuing stress related to his
environment:
You’re at 2 or 3 o’clock in the morning and somebody’s outside blasting their freaking
music. Yes, that’s causing stress. So, yes, my smoking has increased. I mean, I’ve had to
call the police innumerable times because people just don’t seem to get it.
A female participant living in Skid Row described the tumultuous environment and how
cigarettes eased the tension and fear she navigated on a routine basis:
I smoke a lots when I’m nervous. You know, a cigarette calms me down. If I’m upset, I
could jus’ smoke a cigarette, and I [exhales], I relax. They jus’ killed someone on the
corner. In the front all down the streets, they sellin’ drugs. They be jumpin’ on people.
Women’s butt naked. And it makes you feel some kinda way. Of course, I’m a human,
and—and it makes you feel some kinda way watchin’ this sadness. People nodded out
and yeah, people dyin’ in the building.
This was not limited to residents living in Skid Row. Many residents living outside of
Skid Row also lived in areas marked by high crime, drug use, and poverty. A scattered-site
participant living outside of Skid Row described his discomfort with his neighborhood:
I’m just going to say it like this. This environment makes me feel uncomfortable. It
makes me feel uncomfortable because of the activity that goes on in the neighborhood.
And by me feeling nervous and uncomfortable—that’s going to make me want to smoke
a cigarette, or chew tobacco, or something to kind of pacify the time of—helping me to
stay comfortable.

136

Social Environment: Facilitators
Social support. Although residents generally reported limited contact with peers and
family, many indicated that these relationships served as facilitators for reducing their tobacco
use. For some residents, simply the presence of someone, often a close family member, child, or
partner, made quitting easier. A male resident attributed his reductions in smoking to his children
and grandchildren coming around more often. “Just basically my grandkids and my kids—when
they come, that helps me out a whole lot.”
For others, it was more than the presence of a person. The words that residents’ peers and
particularly their loved ones expressed to them often had an impact on their smoking. Often these
were not deep conversations or expressions of emotional support, but rather a sort of gentle
nagging or encouragement. A male resident talked about his children’s reminders to quit:
I see one of my daughters and one of my grandkids often and then, the other one, she stay
in Vegas, so I see her when she come down. But she, it’s like she here ‘cause she calls me
all the time, so she’s still here. They just say, “No, Daddy, stop. Don’t smoke around me,
just, Daddy, stop.” So, I took it as, I’ll stop smoking around them.
Several residents described receiving informational support, including advice or
assistance with tasks or challenges, as a facilitator. In the case of informational support, residents
reported that advice and assistance specific to cessation was a facilitator. Some residents
indicated that the informational support they found particularly useful came from a peer or loved
one who was a former smoker and had successfully quit or someone who was a current smoker
who had committed to quitting with them. A participant who had decreased his tobacco use in
housing explained the role of the advice he received from a family member:

137

My mother encourages me to stop ‘cause she used to smoke a lot when she drinks, but
she still drinks now but like only at night. But she stopped, quite a few years ago. She’s
all, “Yeah, you need to stop smoking.” Gives me tips. Yeah. It is supportive, ‘cause I
used to smoke almost a pack a day.
Socialization versus distancing. A contrasting mixture of residents’ responses to their
social environments was identified, with some residents reporting that socialization—feeling part
of a community—and interacting with others facilitated their ability to cease tobacco use. A male
resident living in a scattered-site apartment talked about how he smoked less before he entered
housing because life in the shelter meant that he was around others he interacted with regularly:
“I smoke less because of the … communication with someone.” He also described the sense of
awareness that came with socialization because he did not want to smoke around other people in
the shelter. “Having the respect not to smoke around others, not to smell like smoke around
others, and dealing with the classes you have to take.” These were among the reasons why this
socialization was beneficial to his ability to quit and why isolation posthousing increased his
smoking. However, other residents felt that many of their interactions and socialization hindered
their cessation and therefore, distancing themselves from their peers and family worked as a
facilitator. A female participant talked about her tactic of distancing herself from family
members and peers:
I had cut out a lot of people. I try to sort out and, and not deal with the negative people in
my life that always, you know. There’s always people that try to just … all they do is
thrive on negativity and never have anything positive to say. I try to eliminate those
people from my life and just limit my associations, who I deal with. And that helps me to
not smoke.

138

Providers. When asked about support from providers, including medical doctors, mental
health providers, and case managers, many residents reported they did not feel supported
regarding their efforts to quit smoking. For those who did experience provider support as a
facilitator of their ability to quit smoking, support was commonly described as care and concern
from providers, rather than medication or a cessation aid. Residents said when providers took the
time to check in with them, educate them, and express care about their health, this felt
supportive. A resident who had decreased her smoking attributed the path toward cessation to her
doctor who worked with her for many years:
It was my primary doctor—that be on me a lot—about the smoking, so that’s who really
kinda changed my mind—over the whole smoking situation. So, when I first started out
with her, I was like, a terrible smoker, then, like … it was like, I needed them all the time,
so [my doctor] taught me how to gradually slow myself down—one day, eventually, I
imagine I’m gonna quit completely.
Talking about her providers, including her primary care and mental health provider, a
female resident expressed the depth of their impact on her life.
They allow me to be me. And by them allowing me to be me, the things that I don’t even
know are bothering naturally come out and I’m like, “Darn, I didn’t know that was
inside.” You know what I’m sayin’? So, yeah. They bring me great relief because they
allow me to be me. And I not only, not only just that. They look, they look, they, they
listen, like they’re really interested.
For many of these residents, providers’ concerns about their smoking was part of the
broader display of care about their overall health. This meant that providers did not harp on their
smoking habits, which was associated with feelings of guilt and shame, but rather considered

139

many aspects of their health and well-being. This was true not only for medical professionals,
but also for case managers. A resident living in Skid Row described how his previous case
manager, who had since left the agency, had an impact on his tobacco use:
Well, when we first met, she kind of asked me serious questions like what were
challenges, and one was my vision and the other one, one was that I used cigarettes as
kind of a twisted habit, so we kind of talked through that and I got different activities you
know, I’m a musician—so we talked about those type things and spouse and that type of
stuff that kind of affected my habit.
The fear of a disapproving provider acted as a facilitator for a portion of residents.
Routine visits to providers meant there were frequent health check-ins, and providers may be
more likely to ask about or detect residents’ smoking behavior. Some residents perceived this
doctor–patient interaction as more of a nuisance. A male participant living in single-site housing
in Skid Row who had an asthma diagnosis stated, “I just don’t want to hear it from [my doctor],
so I cut down.” Nonetheless, this “pestering” seemed to facilitate reductions in tobacco use.
Other residents expressed some fear regarding interactions with their providers. A male
participant who had asthma and attended bimonthly doctor appointments to monitor his
breathing attributed his decreased smoking in part to concerns that his doctor would “detect [his]
smoking” during an appointment. A female participant living in single-site housing in a
neighborhood outside Skid Row who had a bipolar diagnosis had once been told by a professor
that “if you’re smoking, you must be suicidal.” This ingrained belief led her to fear that if her
medical doctor found out she was smoking, this would be communicated to her mental health
provider, who would have her hospitalized.

140

Cessation aids. Residents had an array of responses to perceptions of cessation aids and
responses to utilizing them. Of four residents who had ceased smoking, two of them used
cessation aids—nicotine patches, specifically—to facilitate their ability to quit. For some
accessing and using nicotine replacement therapy (NRT), most commonly nicotine patches, was
easy and made them “more confident in [their] ability to quit.” A male participant living in
scattered-site housing who had ceased smoking discussed the role of nicotine patches in this
process:
First, they put a patch on me. Stick a derm patch on me, all that crap. It was for most of a
year and a half. [Veterans Affairs] is pretty good about helping you stop whatever you
want, whatever you’re trying to, whatever it is, again, alcohol and drugs, cigarettes. Sex,
gambling, whatever you want to call it. Once a month, I went to the doctor. They wanna
know how you doing with the patch, and if you get the patch, because what they do is
they decrease the amount of nicotine they give you in the patch. Instead of smoking an
actual cigarette, you put the patch on. … They take that puff away. Yeah, they were
helpful. I don’t crave cigarettes anymore.
Some residents wanted to use the patch but experienced barriers in access. A female
resident in Skid Row wanted to use nicotine patches, but she experienced barriers with insurance
coverage:
My doctor wanted me to cut back actually. Wanted me to know if I ever wanted it, I can
get the patch. Offered it, yeah. But Medicare doesn’t pay for it, so. Unfortunately, I
couldn’t afford it. Wouldn’t mind trying to, you know, quit but—it’s kinda hard right
now.

141

A male resident used patches inconsistently, partially due to the difficulty of getting them. The
combination of working full time, the 15-minute walk to his doctor, and the long wait for them to
be ready discouraged him. He explained:
I would say the nicotine patches are good though, ‘cause I tried, like, I think three times,
but what it is, I procrastinate on getting a refill. It’s very easy to get ‘em from my doctor.
But it’s, I guess, it’s the process of getting them though. That’s what it is. It’s just like,
“Ah, later on, tomorrow,” you know?
A substantial number of residents were opposed to NRT options, indicating it did not
make sense to them to “substitute smoking with nicotine.” Many residents had received patches
from their doctors, but they lacked the education of how to use them appropriately and therefore,
did not use them. A resident discussed her reservations: “I really don’t know how to work them.
I really don’t. I put one on, and they said, ‘Don’t smoke with them on, it’s dangerous.’ So, I just
leave them alone now.” Others were afraid of NRT options, mostly due to the side effects they
had experienced or heard about. A male scattered-site resident explained that the patch helped
him, yet the side effects he experienced led him to use them inconsistently:
The patch helps, because like when I said when I’m trying to quit, I go through this
detrimental depression. This miserable state of, “I’m no good. I’m a failure, I’m a loser, I
deserve to go through … this uncomfortable feeling.” But the patch does help. It …
decreases that sensational depression, but the patch does give me a migraine.
A female resident in scattered-site housing had not used the patches and was afraid to do so:
I did read the fine print, and said, “No thanks, I’ll go cold turkey.” There were a list of
side effects—I know I would have problems with. I would probably have been at risk.
Seizures, mental disorder, nightmares—I’m just now finally starting sleeping overnight.

142

Being homeless, you know, you can’t sleep, so I’m in my own place and I can sleep
again. Oh yeah, I don’t want no flashback and have bad dreams all over again.
Social Environment: Barriers
Ubiquitous smoking. Residents said living in a community where smoking was
abundant and normalized made cessation a challenge. This was commonly discussed in
relationship to accessibility, such that living in a neighborhood where purchasing cigarettes was
very easy often meant that the community featured many smokers. This was especially true for
residents living in Skid Row. A resident living in Skid Row who had increased his smoking
attributed the increase to the ubiquitous smoking around him. “Everyone in the building
smoking. I’m around, I’m surrounded around smoking people smoking cigarettes.” When asked
what makes it difficult to stop smoking, another Skid Row resident explained:
What makes it difficult is because I mean, pssht … damn near wherever you go. You
know, I’m amongst the people that’s continuing to have, continue to smoke. And you
know, it’s kind of like, you know, what they call a trigger. If I go over here and I talk
with this cat for a few, 10 or 15 minutes, I know, I’m already know what he’s doing over
there, but my dumb ass still go over there anyway.
However, for many residents living in other neighborhoods, particularly those in single-
site housing, being in a building with formerly homeless individuals meant that there were high
rates of smoking among their neighbors. A resident living in single-site housing outside of Skid
Row discussed the temptation when living among many neighbors who also smoke:
All I can say is I’ve slowed down lately, and the apartment and the place … there’s
smokers around here, and if I don’t have a cigarette at the time I might go to my friend

143

Leonard. “Hey, Larry can I get a cigarette,” or “Hey, Betsy can I get a cigarette?” You
know?
A female resident in single-site housing in Skid Row described why living among
smokers was a barrier, especially when attempting to quit.
It’s just, it’s a little hard. You know, like we were saying when you see other people
smoking, you might get the urge, you know, to have a cigarette, and then that just
intensifies. Then once you start smoking one, then there’s another one, so. All my
friends, basically, are smokers themselves. And I would say at least 70% of the people
that live in this building are smokers. So, yeah. Staying away from a lot of smokers. And
being a loner, keeping to myself a lot. That really helps out a lot.
Stigma and conflict. When asked about whether their relationships helped or hindered
their ability to quit, many residents expressed frustration with their loved ones’ responses to their
tobacco use. Residents’ family members often judged them for smoking, which led to residents
feeling stress and strain, making cessation challenging. A resident who had quit smoking
discussed how he found it harder while he had a partner:
I won’t do it with somebody else going, “Come on, dude, come on, dude.” My ex-wife
made it worse. Actually. She was trying to help, but it made it worse. That made it into a
fight, and then we got into a fight, and then no one is happy.
Outside of their tobacco use, residents had many relationships characterized by conflict.
Often these relationships were with their neighbors, particularly those living in single-site
housing models. These negative interactions generated much internal tension, and they used
smoking to cope. A female resident living in a scattered-site unit explained: “One of my sister-
in-laws, actually, made me very upset—and I was crying and crying, and after I calmed down, I

144

actually had to have a cigarette. Because I was that stressed out.” Although he had substantially
reduced his tobacco use, a resident living in Skid Row recalled why it was hard to quit. Smoking
was an alternative to “getting heated” in arguments with his former partner, and then it became a
way to cope:
Well, if I got in an argument or got pissed off, I smoked a cigarette—it would relax me,
calm me down. I got used to doing it instead of participating in an argument. Then we go
back to that level of verbal abuse. … And then, I’m feeling bad about it. I just smoked.
Providers. All residents who participated in the semistructured interviews were
connected to at least a primary care provider, and many were connected to other care providers
and saw them regularly. An alarming portion of residents reported that their health care providers
had not intervened in their smoking behavior and many expressed discontent with their
providers, which hindered them from initiating conversations with their providers about their
tobacco use or seeking support to quit. A resident explained that although he saw his doctor
regularly, he had never felt the support he needed to quit: “I don’t know, really. I really don’t
know. Because I’ve never experienced any kind of support like that. I really wouldn’t know what
to—to tell you, what to expect.”
As with primary care providers, all residents had case managers, and some residents saw
their case mangers monthly, whereas most interacted with their case managers more frequently,
up to daily. Many residents also expressed that they had tumultuous relationships with their case
managers or cited the high turnover of case managers. When asked about communicating their
cessation goals and challenges, residents frequently stated that their case managers were not a
source for support or linkages, given these strained relationships. A female resident in scattered-
site housing said she felt like case managers were not genuinely concerned about her issues:

145

“Case managers making that 30 dollars an hour. She don’t give a shit what you’re really going
through at home. She got her own problems. She got her money. Her home is cool.” She went on
to explain that she felt if providers exhibited more genuine concern for residents, they would
have greater ease in reducing their smoking:
A real counselor or professional, they work from here [points to heart], and go out their
way and do what they can to help that situation. If people did more of that heart stuff,
there would be a lot less drug users.
Others did not perceive case managers in this way, because their case managers were also
smokers and smoked regularly in front of or with residents. A resident living in a scattered-site
apartment talked about the nature of his relationship with his case manager:
Whenever we get together or whatever, it’s not really based on cigarettes or tobacco use.
It’s just based on my program, my housing, you know, stuff like that. And, well, she’s a
smoker. And I’m a smoker, and sometimes we’ll go out on the patio at the place where
she works and smoke a cigarette.
Another resident chuckled when asked about his case manager’s role in his tobacco use,
exclaiming, “They all smoke, too!”
Individual: Facilitator
Willpower. Among residents, particularly those who reduced their tobacco use or were
trying to, the most commonly identified facilitator of reducing or ceasing smoking was
willpower. This perception of an innate trait that residents had or could cultivate was consistent
across gender, housing model, and neighborhood. A female resident in single-site housing
acknowledged her faith in this process, but ultimately attributed it to her own doing: “I thank
God that I do have a strong willpower. That I can do it on my own.” For residents, this meant

146

that although other factors might have influenced their smoking behavior, ultimately the decision
and commitment were internal. “It’s up to the individual. Same with losing weight, with
religion—you can dedicate yourself to it.”
A female participant living in a scattered-site apartment described the barriers she
experienced while homeless, yet indicated that ultimately, the power to quit was in her hands:
When you’re under duress or stress, or in a situation that you can’t control—you use
tobacco, or drugs, or alcohol to make up for that situation. And yet you can control that,
“I’ll smoke if I want to, I’ll stop if I want,” or, “This is something that … it’s in my
hands. No one else is making the decision but myself.”
For some residents, this internal or intrinsic motivation they experienced came with the
perception that housing had shifted their lives in some way. A female resident exclaimed, “For
the first time in my life, really, I’m just now starting to think about me.” Housing was a
launching pad for recovery and pursuing bigger things.
It’s just that fact that, you know, now that I have a home, a security. That played the most
important role as far as, I have no worries, and [smoking] would be my outlet—I have no
need anymore, thankfully, for that.
This notion of intrinsic motivation as a facilitator of cessation was seen in other aspects
of residents’ lives, such that many of the residents who had reduced their smoking often reported
working toward other goals. Many had gone back to school (two of the four who had ceased
smoking were in school), whereas others were involved in a career or hobby, including art, fixing
cars, and culinary training. These residents often acknowledged that working toward these goals
meant they were quite busy and did not have time to smoke. A female resident who had quit

147

smoking explained, “I used to have a lot more free time. Now I don’t. I’ve been in school. I just
actually completed my second year, so, that contributed to it a lot because I don’t have time.”
For some, this was a more active process of replacing smoking with other activities. A
man living in Skid Row described the active restraint it sometimes required:
I’m trying to stop smoking, ‘cause I’ve been trying several times, you know. Say I’m
gonna just go sit outside and smoke or just think about [smoking]. Instead, I go do
something positive, something good for me. Go for a walk. And maybe go, like, library.
Or this bookstore right here, I go to the bookstore, like, you know, or sometimes go to
McDonald’s and buy me a coffee. So, that’s usually what I do.
Others acknowledged that smoking got in the way of pursuing other goals, either because
it made them feel physically restrained or because it took up time, so it naturally decreased as
their foci shifted. A resident who had gone back to school full time felt that eventually he got to
the point where “even one cigarette—I could feel it.” That interfered with his school schedule: “I
have to be up and on the bus at 7. Then in school ‘til 4. I started to feel the difference when I
smoked—I couldn’t have them weighing me down no more.”
Other residents seemed to view quitting as a challenge, and this was internally
motivating. This was often facilitated by internal frustration after years of telling friends and
family that they going to quit, yet not following through. It also was a product of building
internal strength so that residents gained a renewed sense of confidence in their ability to quit.
This internal strength was attributed to achieving occupational goals (e.g., going back to school,
getting a job), ceasing use of other substances, and reuniting with and being there for family
members, particularly children.

148

Health. Residents often faced numerous health problems, but for many, getting off the
streets and being housed fostered a prioritization of their health. No longer worrying about
finding housing and having routine access to doctors meant that residents could focus on their
health in a way that was not possible while homeless. A female resident who had reduced her
smoking described the shift in focusing on her health: “It does play a major role on my health.
When I’m homeless, I have nothing to worry about. The environment itself will kick me to the
curb. So, it’s really doesn’t make a difference whether I smoke or not.”
This focus on health led some residents to realize that their own morbidity and mortality
was heightened as a result of smoking. This realization often came following a physical health
diagnosis or the deteriorated health or death of a loved one. A female resident living in a
scattered-site apartment acknowledged her relationship with smoking, following a diagnosis she
received once housed: “You know, my desire to smoke was removed when I had my breasts
removed. I survived cancer. I don’t want, really take that risk.” For others, it was a function of
their doctors educating them about the state and trajectory of their health. Entering a new phase
of life and the reality of their aging bodies was also a source of motivation to implement
healthier behaviors. A male resident living in Skid Row who had smoked since age 13 and had
not smoked tobacco in more a month attributed quitting to his awareness of his health and age.
“You know, for me, it was just the will to stop altogether because I know it’s not doing anything
health-wise to my body but destroying it. So, I just stopped altogether, because I know that, that
I’m 52 now.”
Financial. Interviews were conducted in the months following a tax increase on
cigarettes in Los Angeles. For many residents, the increase in cigarette prices ensured their
“dedication to quitting smoking [was] so much higher.” Residents felt their financial

149

circumstances were already dire and this would only worsen them. Although residents expressed
frustration about the tax increase, the financial gains from quitting were nonetheless an incentive
to quit. This tended to be the case for residents who were still smoking and living in other
neighborhoods, because street selling in Skid Row made the price increase of cigarettes less
influential. A male resident living in a scattered-site apartment stated:
When I run out, I stop for a week, definitely, because I don’t have no more money to buy
the cigarettes. They $9.25 a pack. Yeah, so, that stops me too for at least a week until I
get to where I can buy some. That helps a lot, the price of them going way up, that helps
a lot.
Another male scattered-site resident attributed finances as the primary facilitator of cessation:
One of the things is I had to be completely flatly broke. No money at all. That is the key
thing right there. Like, I think that almost one at the top of the lines of me quitting. You
know? If I got finance and I go through the withdrawals, I’m going to go spend the
finance. If I don’t have any finance and I go through the withdrawals and I don’t have
anyone that I can borrow money from, or ask for money, or whatever, I have to go
through this without no … no ifs, ands, or buts about it.
Individual: Barrier
Willpower. Although residents who had reduced or ceased their tobacco use attributed it
to internal willpower, those who were struggling to cut down or had maintained their smoking
behavior also cited willpower, or a lack thereof, as a barrier to quitting. This idea that “it has to
come from me” came in conjunction with the dismissal that anyone else could intervene in their
smoking behaviors. A resident suggested he “don’t really need no support. It’s all in my mind.”
While struggling to identify what would help him get motivated, he said, “I either make it

150

difficult or easy, or I could just say I stop and then I try to stop. But yeah, nobody gonna help me
quit smoking.”
In fact, many residents in this position struggled to identify what would help them feel
more motivated. Some said the motivation would eventually come, that they would be ready at
some point but simply were not at that time. This was primarily the case for residents who had
maintained their tobacco use. This sense of complacency in their lack of motivation to quit
characterized this group. A male participant in a scattered-site apartment shrugged as he was
asked about what he needed to quit:
I think it comes down to … up to now, I haven’t made a conscious effort to really stop
smoking because I enjoy smoking cigarette. I would think that, you know, we all wanna
live as long as we can. But at the same time there’s no, there’s no guarantees for a
tomorrow for any of us.
Receipt of health education and information about the negative effects of health were
noted as a needed resource—some desired a scare tactic approach to prompt their motivation.
However, many acknowledged they knew the negative effects of tobacco use, yet it was not
enough to stop. A woman in single-site housing outside of Skid Row said that “being educated
on the pros and cons of smoking” would “encourage me even more to want to quit.” A male
scattered-site resident who considered himself a closet smoker explained:
Well, as far as being helpful, I’m … if you want to call it a closet smoker—so, no one can
help me as far as encourage me about the cigarettes. But from my years of reading and
studying and listening and experiencing others behavior on smoking, what it has done to
them, the commercials, the pamphlets, the books, the courses, the classes, they have

151

educated me, per se, on long-term effect on the nicotine or the secondhand smoke or
smoking.
Many residents seeking this information did not seem to see providers as the conduit for this
information.
I think that advertising, all that—if the negative repercussions to cigarette smoking was
advertised more, I think that the only think that will really grab our attention. Because,
number one, in our time where we old, we get set in our way. We not willing to change.
Then when it comes to young people, it’s cool. But when you really see what cigarettes
smokes really can do and really do, I think that’s really the only effective measure.
For others, their prolonged use of tobacco meant that it was an established habit.
Breaking that habit was a challenge. A female resident identified the hand-to-mouth motion as
barrier, which was brought up by many residents:
I still think it’s the hand motion and the getting in the car and just light one. I just will
stop because half the times I would smoke two or three puffs and go, “Yuck, what did I
do that for?” Put it out, and then do it again later. So, I think it’s just the habit. A bad
habit.
Boredom. Many residents discussed the relationship between time and smoking—that
smoking filled the time and that boredom was one of their biggest barriers to cessation. The
absence of a job, hobby, or strong social ties meant they spent a lot of time alone, which
increased boredom and subsequently increased smoking. A female scattered-site resident
explained the role of boredom in her tobacco use:
Well, you know, it’s if you want to say loneliness—being bored, nothing to do. It’s sort
of like sucking your thumb, you know. When we don’t have anything to do, you want

152

something to do. You want something to kind of comfort you, to give you ease to the
boredom, or stress, or being uncomfortable. The clock’s ticking and you’re watching the
clock and you’re sitting there and not doing anything, so, smoking a cigarette gives you
something to do with your hands, with moving around.
Working, being active, or keeping busy was not always possible due to mental and
physical health limitations. A female resident livening in scattered-site housing was a former
caterer and had to stop due to physical conditions.
Well, when I first moved into housing, I wasn’t smoking. But as time moved on, you
know, after a year or so, then I started back to smoking again. Because I can’t go out and
do what I want to do. I’d like to go out and cook. I’m a caterer, I know how to cook, I
know how to cater. I used to walk up here to the [track field] and walk around the lot four
or five times every day. I don’t do that no more. It hurts me to walk.
Although such limitations were acknowledged by residents, particularly those in
scattered-site and single-site housing in other neighborhoods, they also voiced a desire for
appropriately matched activities (e.g., exercise, art support groups) that would reduce boredom
and facilitate their intrinsic motivation. A female resident wanted to take an exercise class to fill
her time and remind her of the physical effects of tobacco.
I feel if they had, like, a Zumba class or something else that would show the harm that it
does because you’re out of breath and certain things like that. You know, just activities
that are, you know, a little higher impact.
Mental health. Overall, the primary aspect of mental health that hindered residents’
ability to cease smoking was stress. Residents used smoking to cope with their stress, and the
soothing effects of cigarettes were considered a barrier to cessation. Acknowledging the ease and

153

calm cigarettes brought, smoking represented a simple pleasure in their chaotic lives. A male
resident living in Skid Row did not foresee himself quitting: “People say, ‘Well, that’s just a
crutch, it doesn’t really help.’ I’ve been smoking since I was 16. It does help. It really does.”
Another male resident who had decreased his smoking acknowledged his lingering affection for
smoking: “I like coffee and cigarettes, and it’s—it’s just an enjoyable thing, you know? Like the
commercial says, you know? It really cools you down, but I know that it’s bad for my lungs.”
Residents experienced stress from several factors including housing, relationships, and
health. Many residents were not satisfied with their housing environment, including apartment
size, location, safety, and neighbors. Many residents expressed a desire to move and struggles to
navigate that process, which contributed to stress. A male resident living in Skid Row explained
that he did not perceive management of his stress being possible until he found a new apartment.
If you’re not down here experiencing it, you don’t understand it. You won’t comprehend
it. If you’re hanging out here for a while, you’ll see what I’m talking about. So, that’s
basically what I need to do. I need to get up out of here.
Relationship stressors enhanced smoking, and residents struggled with alternatives or to
manage the overwhelming feelings that these ties brought. In some cases, the resident was bound
to the relationship in some way (e.g., children, partner living with them, neighbors). A female
resident with several young adult children struggled with the demands of parenting discussed this
issue:
Just daily stuff that you wanna do and complete, they turn around. If it ain’t one of your
kids with their problems, it’s somebody else and it’s just too much, like everybody comes
to me with their problems and—make it harder for me.

154

A participant who struggled with his relationships with his neighbors and street life in Skid Row
shared similar sentiments:
I truly believe that stress creates a great—high stress creates a great deal of
complications. We’re already read in the tabloids that stress causes cancer and all these,
you know, physical ailments and everything, and it’s true. It’s absolutely true. But I
believe in an environment where I’m not being pulled left and right and stressed out and
having people blast their frickin’ music at 3:30 in the frickin’ morning or come right here
into the community area here—and sit there with those frickin’ boxes and blast their
frickin’ music. I wouldn’t smoke as much.
Some residents expressed difficulty identifying alternative coping mechanisms to smoking,
stating things like, “Sometimes my coping skills are not the best and the tobacco does bring, it
brings my stress level down.”
Very few residents spoke about the specific mental health symptomatology that drove
their tobacco use other than stress; however, several residents expressed fears of quitting due to
the negative mental health symptoms that would ensue. These residents tended to be men who
had maintained their tobacco use over time. Symptoms of depression were the primary concern,
which were a function of experiences with previous attempts to quit. A male scattered-site
resident explained his fears of quitting due to the resulting mental health symptoms he had:
Literally, [quitting] is detrimental to my mind, body, and soul. Because trying to quit, the
body was going through all types of depression, thoughts of, “You’re no good. You
deserve to go through this—this withdrawal stage,” you know? Just terrible and, you
know, I tried to quit. Smoking—what it does to me, it’s like a friend, or a buddy, or a
companion, or gives you something to do. It occupies your mind. When your mind is

155

occupied on the smoking, it’s not occupied on your behavior, your actions, your thoughts,
you know.
Substance use. Residents who had histories of substance use or were current substance
users generally said they felt that using substances was a barrier to cessation. Many residents
with histories of illicit substance use or binge drinking had ceased or substantially reduced their
use of these substances and had subsequently ceased or reduced their smoking. A female resident
in a scattered-site apartment who was now sober discussed how her cessation of other substances
facilitated her own ability to quit smoking:
You can’t really move forward, you know. Drugs are always gonna stagnate you. At one
point, I looked at myself as a, a functional user and a functional addict and all that shit,
but somewhere down the line, you know, it’s gonna show—it might be your health or
your relationships—and you’re not gonna be able to function.
A participant who was still using illicit substances occasionally acknowledged the relationship
between using and his tobacco use:
I use it when it’s, when I’m using other substances and stuff like that. But then again, you
know, that’s my only constant continue to add on to my depression. And I’m already
depressed financially, I don’t have the money to put a type of crap but, you know, it’s just
that—I have this urge or desire to try to deal with my depression.
The relationship between cannabis use and smoking was substantially more varied for
residents. Some residents found that when they used cannabis, their cigarette cravings increased.
A few residents who were particularly heavy cannabis users reported using cannabis and tobacco
simultaneously. When they reduced their cannabis use, often for financial reasons, their tobacco
use also decreased. A resident in this situation was saving up to move out of his apartment:

156

I’m trying to move out here. I’m trying to move out of this environment. You know, and I
come up in December for my Section 8 voucher—I want to be able to get that and I want
to be able to move next year. Weed takes my money away. It takes money away from the
purpose of me having the need to, to get out of here.
For others, cannabis use aided cessation, and these residents found themselves
substituting cannabis for tobacco or smoking less tobacco by mixing it with cannabis (via blunts
and joints). A resident described this process and how specific strains of cannabis affected his
desire to smoke in differential ways.
Well, I kinda use more hemp products instead of tobacco so, it seems like I don’t smoke
as much. I noticed that with the hemp blunts that I smoke maybe three—maybe three
blunts a day, maybe two. It depends on what kind you smoke, like if you smoke sativa.
Sativa makes me want to smoke more. Okay but if I smoke indica, then I just—you
know, it helps my back pain, and it just helps me to calm down and relax and not be so
[uptight].
Discussion
This study explored the physical, social, and individual facilitators of and barriers to
tobacco cessation among residents in scattered-site and single-site supportive housing settings
across various neighborhoods in Los Angeles County. Findings from 33 semistructured
interviews with residents living in differing PSH models across various Los Angeles
neighborhoods and who presented with different tobacco use trajectories revealed they perceived
barriers and facilitators at each socioecological level. Moreover, many of these barriers and
facilitators combined to create compounded barriers and facilitators at multiple socioecological
levels, ultimately affecting residents’ ability to cease using tobacco.

157

Findings from this qualitative inquiry suggest that for many residents, housing is a
launching pad for health and recovery, and that this contributes to their confidence and follow
through with cessation from tobacco. Several studies that have examined longitudinal
quantitative data on supportive housing residents’ health outcomes suggest that housing
improves some aspects of health for formerly homeless residents (A. J. Baxter, Tweed,
Katikireddi, & Thomson, 2019; National Academies of Sciences, Engineering, and Medicine,
2018). Supportive housing residents’ goals shift in the transition from homelessness to housing,
with research indicating that after 1 year in housing, residents have prioritized their health and
financial goals (Wenzel et al., 2018). This current study adds to the body of literature on
supportive housing residents’ outcomes of health and well-being, because findings suggest that
for many residents, health remains a prevalent concern and goal several years after entering
housing.
Although housing itself may be a facilitator of cessation for some, residents’ housing
context also plays a role. Smoke-free policies in housing were one of the most frequently
reported facilitators for reducing and ceasing tobacco use in this sample. Following the passage
and implementation of the U.S. Department of Housing and Urban Development’s requirement
of smoke-free public housing units, there has been extensive debate about whether smoke-free
policies in public housing violates smokers’ right to privacy (Fagundes & Roberts, 2018).
Moreover, some have argued that requiring smoke-free homes is a discriminatory practice for an
underserved population like formerly homeless supportive housing residents, because violation
of such rules could lead to eviction and a return to homelessness (Fagundes & Roberts, 2018).
Although this study cannot resolve this debate, results from these in-depth interviews suggest an
overwhelming appreciation of smoke-free policies among those living in buildings with these

158

policies, not only because they facilitated cessation, but because they also reduced secondhand
smoke from other units and maintained a clean, smoke-free environment. Residents in this
sample had been living in their units for 3 to 5 years and generally vocalized their gratitude for
their unit, expressing little to no difficulty or frustration with upholding the policies. Although a
few residents reported violating these rules on occasion, none of these residents reported any
consequential punishment or risk of eviction at the time of the semistructured interview.
Accessibility of cigarettes beyond smoke-free housing policies also affected residents’
smoking behavior, either by facilitating cessation for those living in more remote, fringe
neighborhoods or by virtue of living in Skid Row, where street vendors are abundant. Regarding
accessibility as a barrier, this was predominantly true for residents living in Skid Row. Due to the
availability of cigarettes on the streets, the increase in cigarette prices was less impactful for
residents in this neighborhood. Living near many tobacco product retailers decreases travel
burden and subsequently increases exposure to marketing of tobacco products (Kong, Myers,
Isgett, & Ribisl, 2020; Luke, Hammond, & Combs, 2017). Neighborhoods characterized by high-
density tobacco retailers have higher smoking rates and lower rates of cessation than those with
fewer tobacco retailers (Halonen, Kivimaki, & Kouvonen, 2014; Kong et al., 2020). For
residents in scattered-site housing in other neighborhoods where tobacco retailers are less
available, this may seem like an opportune way to bolster cessation; however, this was often
accompanied by isolative behaviors. This warrants concern as this population often struggles
with socialization and further marginalization posthousing (Ecker & Aubry, 2017). Substituting
tobacco for another maladaptive behavior like isolation may continue to be devasting to health
(Cacioppo & Cacioppo, 2014). Community integration has been an ongoing struggle for
providers serving supportive housing residents, and neighborhood safety, location, and housing

159

quality are associated with community integration for this population (Ecker & Aubry, 2017).
These findings suggest providers may need to employ strategies that help residents navigate
social barriers that inhibit successful cessation while supporting much-needed community
integration outcomes for this population.
Although certainly not a new finding, residents in this study reported a great deal of
conflict and strain in their personal relationships. Research among supportive housing residents
has found that their social networks decrease in size and support in the transition to housing
(Harris, Rhoades et al., 2019). Residents are challenged with leaving unhealthy, unhelpful social
relationships following entry into housing (Macnaughton et al., 2015) and may resist forming
connections with their formerly homeless neighbors in an effort to distance themselves from
their homeless identities (Henwood, Lahey, et al., 2018). Encouragingly, for some residents,
connecting with familial relationships and close peers aided cessation. Overall, these connections
provided structural support, such that the presence of such a person affected cessation, rather
than anything that person said or did or any emotional support they provided. This might be a
positive finding because it suggests even minimal support can promote smoking reduction and
cessation. Given residents’ deficits in socialization and social relationships, it is difficult to
determine the specific forms of support that may be most effective in reducing their tobacco use.
However, residents’ responses regarding providers and their care and concern that facilitated
cessation, receipt of support from peers and loved ones would also facilitate cessation. Future
research would benefit from exploring how residents’ relationships, including peer models, can
be improved to enhance cessation.
It is of grave concern that so many residents had primary care providers who had not
intervened in their tobacco use. The Clinical Practice Guideline for Treating Tobacco Use and

160

Dependence recommends that care providers ensure all patients who screen positive for tobacco
use receive, at minimum, brief advice on quitting (U.S. Department of Health and Human
Services, 2020). Even brief advice to quit from providers has demonstrated cessation rates
between 5% to 10% a year, whereas combining behavioral counseling and pharmacologic
treatment can achieve cessation rates up to 25% per year (Prochaska & Benowitz, 2019). Persons
experiencing homelessness often face numerous chronic health and mental health conditions, and
providers may be prioritizing these issues over their tobacco use. Given research on the
relationship between tobacco use and morbidity, the desire to quit expressed by this population
(Okuyemi et al., 2006), and the opportunity housing presents to overcome barriers to cessation
while homeless (Baggett et al., 2013), supportive housing residents would greatly benefit from
prioritization of cessation from their providers. The provider–patient relationship may be
especially important for this population, because mistrust of providers can be high (Kerman,
Gran-Ruaz, Lawrence, & Sylvestre, 2019). Findings from this study suggest that this population
may be served best by providers who spend adequate time understanding their patients’ needs
and guiding them through intervention options. The 2020 Surgeon General Report also identified
the need to adjust policies and procedures to ensure providers have the time to deliver quality
care and adequately promote cessation treatments (U.S. Department of Health and Human
Services, 2020). These findings also suggest that more time spent with clients to educate them
about how to use cessation aids might be a way to show concern and may be a necessary, while
also assessing the appropriateness of NRT and aiding them with navigating side effects.
By far the most common facilitator of cessation perceived by residents was willpower.
Theories of motivation suggest intrinsic motivation, or engagement in a behavior because it is
personally rewarding, rather than external rewards or punishment avoidance is more likely to

161

result in behavior change (Ryan & Deci, 2000). For residents who reduced or ceased smoking,
motivation or willpower were often discussed in conjunction with their broader life goals,
including employment, education and health goals. Yet residents struggling to find the
willpower, motivation, and skills needed to quit frequently indicated that they needed
extrinsically motivating approaches such as scare tactics to facilitate their ability to quit. This
thematic finding highlights a strong need for interventions that target intrinsic motivation such as
motivational interviewing to facilitate residents’ cessation. Motivational interviewing draws on a
collaborative provider–client relationship and clients’ goals or values to target clients’ behavior
change, resolve ambivalence, and support self-efficacy for change (Caponnetto et al., 2019).
With this approach, residents can enhance their intrinsic motivation by generating their own
reasons for cessation in addition to empathic support to identify ways to navigate the many
environmental barriers to cessation they face (Caponnetto et al., 2019).
Limitations
This study comes with some limitations. Regarding generalizability of these findings, this
study is limited in that it may not reflect the perceived barriers and facilitators to cessation for
PSH residents living outside of Los Angeles County. Attempts were made to purposively sample
across critical factors including housing model, neighborhood, gender, and tobacco use
trajectories during the initial year in housing; however, other factors may affect cessation
outcomes that were not included in the sampling strategy. Although this sample reflects the
broader homeless population and PSH resident population in Los Angeles, regarding
demographic characteristics, other racial compositions may differ in findings. Moreover, the
research team involved in the parent study was ethnically diverse, yet the semistructured
interviews were conducted by two individuals. Therefore, the gender and race and ethnicity of

162

the interviewers for the current study may have influenced the perceptions and reactions of
participants. Last, supportive housing and services are known to vary across geographic regions.
This in addition to the uniqueness of Skid Row as the densest neighborhood of persons
experiencing homelessness and homeless service providers in the United States may limit
generalizability.
Conclusion
The social ecological model purports that levels of the ecological system are nested and
work in conjunction with one another, ultimately affecting individual health behavior (Macintyre
& Ellaway, 2000; Stokols, 1996). This was evident in this study because the housing
environment often set the tone for residents’ social environments and created a compounded
barrier for many residents. For example, residents living on Skid Row were frequently
dissatisfied with their housing, specifically the physical state of their housing and neighborhood,
and concerned about the safety of their surroundings. Simultaneously, Skid Row residents are
tasked with navigating a chaotic social climate on a regular basis, and environmental stressors
culminated to exacerbate stress. Navigating these routine stressors and identifying internal
willpower to achieve successful cessation were further challenged by the ubiquitous smoking in
the Skid row community, which was tied to the ease of accessing cigarettes in the neighborhood.
The identified web of factors contributing to a multilayered health problem found in this study
supports the need to address health behavior via multiple socioecological level efforts.
The results of the current study have health policy and treatment implications across
several socioecological levels. Although the dearth of available housing is recognized in Los
Angeles, the city with the highest number of unsheltered persons experiencing homelessness,
improved attention and strategic placement of PSH is needed, because these findings

163

demonstrate how for many people, the location of their housing negatively affects cessation,
exacerbates stress, and hinders community integration. Given their cost effectiveness and
broader public health implications, smoke-free policies in supportive housing appear to be an
effective method for reducing tobacco use for PSH residents. However, this comes with the
recommendation to simultaneously increase and improve efforts to address socialization and
relational issues that commonly characterize this population. More cessation services, in the form
of support groups or health education programs, are needed at PSH buildings, and receipt of
NRT should be accessible and streamlined, ideally delivered directly to residents’ units.
Providers of all types should prioritize smoking. Health care providers can be an essential tools
in improving cessation outcomes—this study exemplifies the need to train providers to enhance
their doctor–patient relationships with residents and delivering health education information and
recommendations tailored to this population, in addition to a need for effectiveness research that
identifies ways to implement these practices in real-world settings. Given that care coordination
is a goal of service delivery in PSH, nurses and mental health care providers should also be
considered as key facilitators of cessation, especially because they may have more routine
interaction with residents. More research is needed to understand how case manager–resident
relationships can be improved and utilized to facilitate cessation. With this comes addressing
case managers’ own tobacco use and a need to provide alternative methods for building rapport
between case managers and residents. Providers of all types can be conduits for delivering
motivational interviewing interventions and improving intrinsic motivation, given the strong role
it plays in residents’ ability to cease smoking. Additionally, improving intrinsic motivation may
be possible through supporting residents’ employment and educational opportunities, along with
increasing volunteer opportunities and alternative activities to smoking, specifically those that

164

allow for creative expression, give residents a sense of purpose, and target health. Above all,
policies, interventions, and targeted efforts should consider the social and environmental
contexts that frame residents’ health behaviors—such an approach is necessary for reducing the
stark health disparities and disproportionate cessation outcomes for this highly vulnerable
population.

165

References
Antoon, B. G. (2017). Reducing tobacco-related health disparities: exploring the barriers and
facilitators to smoking cessation among individuals experiencing homelessness (Doctoral
dissertation). University of Louisville, Louisville, KY.
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.
doi:10.1056/NEJMp1301935
Barnard, A., McCosker, H., & Gerber, R. (1999). Phenomenography: a qualitative research
approach for exploring understanding in health care. Qualitative health research, 9(2),
212-226.
Baxter, A. J., Tweed, E. J., Katikireddi, S. V., & Thomson, H. (2019). Effects of Housing First
approaches on health and well-being of adults who are homeless or at risk of
homelessness: Systematic review and meta-analysis of randomised controlled trials.
Journal of Epidemiology and Community Health, 73, 379–387. doi:10.1136/jech-2018-
210981
Baxter, P., & Jack, S. (2008). Qualitative case study methodology: Study design and
implementation for novice researchers. Qualitative Report, 13, 544–559.
Billi, J. E., Pai, C. W., & Spahlinger, D. A. (2007). The effect of distance to primary care
physician on health care utilization and disease burden. Health Care Management
Review, 32, 22–29. doi:10.1097/00004010-200701000-00004
Bronfenbrenner, U. (1977). Toward an experimental ecology of human development. American
Psychologist, 32, 513–531. doi:10.1037/0003-066X.32.7.513

166

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. doi:10.2105/AJPH.2013.301336
Businelle, M. S., Ma, P., Kendzor, D. E., Reitzel, L. R., Chen, M., Lam, C. Y., … Wetter, D. W.
(2014). Predicting quit attempts among homeless smokers seeking cessation treatment:
An ecological momentary assessment study. Nicotine & Tobacco Research, 16, 1371–
1378. doi:10.1093/ntr/ntu088
Cacioppo, J. T., & Cacioppo, S. (2014). Social relationships and health: The toxic effects of
perceived social isolation. Social and Personality Psychology Compass, 8, 58–72.
doi:10.1111/spc3.12087
Caponnetto, P., DiPiazza, J., Cappello, G. C., Demma, S., Maglia, M., & Polosa, R. (2019).
Multimodal smoking cessation in a real-life setting: Combining motivational
interviewing with official therapy and reduced risk products. Tobacco use insights.
Advance online publication. doi:10.1177/1179173X19878435
Centers for Disease Control and Prevention. (2020). Office on Smoking and Health. Washington,
DC: National Center for Chronic Disease Prevention and Health Promotion.
Chen, J. S., Nguyen, A. H., Malesker, M. A., & Morrow, L. E. (2016). High-risk smoking
behaviors and barriers to smoking cessation among homeless individuals. Respiratory
Care, 61, 640–645. doi:10.4187/respcare.04439
Creswell, J. W., Shope, R., Plano Clark, V. L., & Green, D. O. (2006). How interpretive
qualitative research extends mixed methods research. Research in the Schools, 13, 1–11.

167

Ecker, J., & Aubry, T. (2017). A mixed methods analysis of housing and neighbourhood impacts
on community integration among vulnerably housed and homeless individuals. Journal of
Community Psychology, 45, 528–542. doi:10.1002/jcop.21864
Fagundes, D., & Roberts, J. L. (2018). Housing, healthism, and the HUD smoke-free policy.
Northwestern University Law Review, 113, 97–119.
Halonen, J. I., Kivimaki, M., & Kouvonen, A. (2014). Proximity to a tobacco store and smoking
cessation: A cohort study. Tobacco Control, 23, 146–151. doi:10.1136/tobaccocontrol-
2012-050726
Harris, T., Rhoades, H., Duan, L., & Wenzel, S. L. (2019). Mental health change in the transition
to permanent supportive housing: The role of housing and social networks. Journal of
Community Psychology, 47, 1834–1849. doi:10.1002/jcop.22230
Harris, T., Winetrobe, H., Rhoades, H., & Wenzel, S. (2019). The role of mental health and
substance use in homeless adults’ tobacco use and cessation attempts. Journal of Dual
Diagnosis, 15, 76–87. doi:10.1080/15504263.2019.1579947
Henwood, B. F., Harris, T., Woo, D., Winetrobe, H., Rhoades, H., & Wenzel, S. L. (2018).
Availability of comprehensive services in permanent supportive housing in Los Angeles.
Health & Social Care in the Community, 26, 207–213. doi:10.1111/hsc.12510
Henwood, B. F., Lahey, J., Harris, T., Rhoades, H., & Wenzel, S. L. (2018). Understanding risk
environments in permanent supportive housing for formerly homeless adults. Qualitative
Health Research, 28, 2011–2019. doi:10.1177/1049732318785355
Henwood, B. F., Matejkowski, J., Stefancic, A., & Lukens, J. M. (2014). Quality of life after
Housing First for adults with serious mental illness who have experienced chronic
homelessness. Psychiatry Research, 220, 549–555. doi:10.1016/j.psychres.2014.07.072

168

Hsu, H.-T., Simon, J. D., Henwood, B. F., Wenzel, S. L., & Couture, J. (2016). Location,
location, location: Perceptions of safety and security among formerly homeless persons
transitioned to permanent supportive housing. Journal of the Society for Social Work and
Research, 7, 65–88. doi:10.1086/685034
Karasek, D., Ahern, J., & Galea, S. (2012). Social norms, collective efficacy, and smoking
cessation in urban neighborhoods. American Journal of Public Health, 102, 343–351.
doi:10.2105/AJPH.2011.300364
Kerman, N., Gran-Ruaz, S., Lawrence, M., & Sylvestre, J. (2019). Perceptions of service use
among currently and formerly homeless adults with mental health problems. Community
Mental Health Journal, 55, 777–783. doi:10.1007/s10597-019-00382-z
King, N. (2004). Using templates in the thematic analysis of text. In C. Cassell & G. Symon
(Eds.), Essential guide to qualitative methods in organizational research (pp. 256–270).
London, United Kingdom: Sage.
Kong, A. Y., Myers, A. E., Isgett, L. F., & Ribisl, K. M. (2020). Neighborhood racial, ethnic,
and income disparities in accessibility to multiple tobacco retailers: Mecklenburg County,
North Carolina, 2015. Preventive Medicine Reports, 17, 101031.
doi:10.1016/j.pmedr.2019.101031
Lee, J. G., Henriksen, L., Rose, S. W., Moreland-Russell, S., & Ribisl, K. M. (2015). A
systematic review of neighborhood disparities in point-of-sale tobacco marketing.
American Journal of Public Health, 105, e8–e18. doi: 10.2105/AJPH.2015.302777
Luke, D. A., Hammond, R. A., & Combs, T. (2017). Tobacco town: Computational modeling of
policy options to reduce tobacco retailer density. American Journal of Public Health,
107, 740–746. doi:10.2105/AJPH.2017.303685

169

Macintyre, S., & Ellaway, A. (2000). Ecological approaches: Rediscovering the role of the
physical and social environment. Social Epidemiology, 9, 332–348.
Macnaughton, E., Stefancic, A., Nelson, G., Caplan, R., Townley, G., Aubry, T., … Tsemberis,
S. (2015). Implementing Housing First across sites and over time: Later fidelity and
implementation evaluation of a pan-Canadian multi-site Housing First program for
homeless people with mental illness. American Journal of Community Psychology, 55,
279–291. doi:10.1007/s10464-015-9709-z
Miles, M. B., Huberman, A. M., & Saldaña, J. (2013). Qualitative data analysis: A methods
sourcebook. Thousand Oaks, CA: Sage.
National Academies of Sciences, Engineering, and Medicine. (2018). Permanent supportive
housing: Evaluating the evidence for improving health outcomes among people
experiencing chronic homelessness. Washington, DC: National Academies Press.
National Institute on Drug Abuse. (2012). Resource guide: Screening for drug use in general
medical settings: The NIDA Quick Screen. Retrieved from
https://www.drugabuse.gov/publications/resource-guide-screening-drug-use-in-general-
medical-settings/nida-quick-screen
Okuyemi, K. S., Caldwell, A. R., Thomas, J. L., Born, W., Richter, K. P., Nollen, N., …
Ahluwalia, J. S. (2006). Homelessness and smoking cessation: Insights from focus
groups. Nicotine & Tobacco Research, 8, 287–296. doi:10.1080/14622200500494971
Padgett, D. K., Henwood, B., Abrams, C., & Drake, R. E. (2008). Social relationships among
persons who have experienced serious mental illness, substance abuse, and homelessness:
Implications for recovery. American Journal of Orthopsychiatry, 78, 333–339. doi:
10.1037/a0014155

170

Palinkas, L. A., Garcia, A. R., Aarons, G. A., Finno-Velasquez, M., Holloway, I. W., Mackie, T.
I., ... & Chamberlain, P. (2016). Measuring use of research evidence: The structured
interview for evidence use. Research on Social Work Practice, 26, 550–564.
doi:10.1177/1049731514560413
Patterson, F., Seravalli, L., Hanlon, A., & Nelson, D. B. (2012). Neighborhood safety as a
correlate of tobacco use in a sample of urban, pregnant women. Addictive Behaviors, 37,
1132–1137. doi:10.1016/j.addbeh.2012.05.011
Pearce, J., Hiscock, R., Moon, G., & Barnett, R. (2009). The neighbourhood effects of
geographical access to tobacco retailers on individual smoking behaviour. Journal of
Epidemiology & Community Health, 63, 69–77. doi:10.1136/jech.2007.070656
Pratt, R., Pernat, C., Kerandi, L., Kmiecik, A., Strobel-Ayres, C., Joseph, A., … Okuyemi, K.
(2019). “It’s a hard thing to manage when you’re homeless”: The impact of the social
environment on smoking cessation for smokers experiencing homelessness. BMC Public
Health, 19, 635. doi:10.1186/s12889-019-6987-7
Prochaska, J. J., & Benowitz, N. L. (2019). Current advances in research in treatment and
recovery: Nicotine addiction. Science Advances, 5, eaay9763.
doi:10.1126/sciadv.aay9763
Pulakka, A., Halonen, J. I., Kawachi, I., Pentti, J., Stenholm, S., Jokela, M., … Kivimäki, M.
(2016). Association between distance from home to tobacco outlet and smoking cessation
and relapse. JAMA Internal Medicine, 176, 1512–1519.
doi:10.1001/jamainternmed.2016.4535

171

Ryan, R. M., & Deci, E. L. (2000). When rewards compete with nature: The undermining of
intrinsic motivation and self-regulation. In Intrinsic and extrinsic motivation (pp. 13–54).
San Diego, CA: Academic Press.
Schneider, J. E., Reid, R. J., Peterson, N. A., Lowe, J. B., & Hughey, J. (2005). Tobacco outlet
density and demographics at the tract level of analysis in Iowa: Implications for
environmentally based prevention initiatives. Prevention Science, 6, 319–325.
doi:10.1007/s11121-005-0016-z
Solari, C. D., Cortes, A., Henry, M., Matthews, N., & Morris, S. (2014). The 2013 Annual
Homeless Assessment Report (AHAR) to Congress: Part 2, estimates of homelessness in
the United States. Washington, DC: U.S. Department of Housing and Urban
Development, Office of Community Planning and Development.
Somers, J. M., Moniruzzaman, A., Patterson, M., Currie, L., Rezansoff, S. N., Palepu, A., &
Fryer, K. (2017). A randomized trial examining Housing First in single-site and scattered
site formats. PLOS One, 12, e0168745. doi:10.1371/journal.pone.0168745
Stokols, D. (1996). Translating social ecological theory into guidelines for community health
promotion. American Journal of Health Promotion, 10, 282–298. doi:10.4278/0890-
1171-10.4.282
Tailor, T. D., Choudhury, K. R., Tong, B. C., Christensen, J. D., Sosa, J. A., & Rubin, G. D.
(2019). Geographic access to CT for lung cancer screening: A census tract-level analysis
of cigarette smoking in the United States and driving distance to a CT facility. Journal of
the American College of Radiology, 16, 15–23. doi:10.1016/j.jacr.2018.07.007

172

Tsemberis, S., Gulcur, L., & Nakae, M. (2004). Housing First, consumer choice, and harm
reduction for homeless individuals with a dual diagnosis. American Journal of Public
Health, 94, 651–656. doi:10.2105/AJPH.94.4.651
Tucker, J. S., Shadel, W. G., Golinelli, D., Mullins, L., & Ewing, B. (2015). Sniping and other
high-risk smoking practices among homeless youth. Drug and Alcohol Dependence, 154,
105–110. doi:10.1016/j.drugalcdep.2015.06.036
U.S. Department of Health and Human Services. (2014). The health consequences of smoking —
50 years of progress: A report of the surgeon general. Atlanta, GA: Author.
U.S. Department of Health and Human Services. (2020). Smoking cessation: A report of the
surgeon general. Atlanta, GA: Author.
Vijayaraghavan, M., Elser, H., & Apollonio, D. (2019). Interventions to reduce tobacco use in
people experiencing homelessness. Cochrane Database of Systematic Reviews, 9,
CD013413. doi:10.1002/14651858.CD013413
Vijayaraghavan, M., Hurst, S., & Pierce, J. P. (2017). A qualitative examination of smoke-free
policies and electronic cigarettes among sheltered homeless adults. American Journal of
Health Promotion, 31, 243–250. doi:10.4278/ajhp.150318-QUAL-781
Vijayaraghavan, M., Olsen, P., Weeks, J., McKelvey, L., Ponath, C., & Kushel, M. (2018). Older
African American homeless-experienced smokers’ attitudes toward tobacco control
policies: Results from the HOPE HOME study. American Journal of Health Promotion,
32, 381–391. doi:10.1177/0890117117729928
Vijayaraghavan, M., Tieu, L., Ponath, C., Guzman, D., & Kushel, M. (2016). Tobacco cessation
behaviors among older homeless adults: Results from the HOPE HOME Study. Nicotine
& Tobacco Research, 18, 1733–1739. doi:10.1093/ntr/ntw040

173

Visser, M. (2007). The social ecological model as theoretical framework in community
psychology. In N. Duncan & B. Bowman (Eds.), Community psychology: Analysis,
context and action (pp. 102–116). Cape Town, South Africa: Juta.
Wenzel, S. L., Rhoades, H., Moore, H., Lahey, J., Henwood, B., La Motte ‐Kerr, W., & Bird, M.
(2018). Life goals over time among homeless adults in permanent supportive housing.
American Journal of Community Psychology, 61, 421–432. doi:10.1002/ajcp.12237
Yanos, P. T., Felton, B. J., Tsemberis, S., & Frye, V. A. (2007). Exploring the role of housing
type, neighborhood characteristics, and lifestyle factors in the community integration of
formerly homeless persons diagnosed with mental illness. Journal of Mental Health, 16,
703–717. doi:10.1080/09638230701496378
Yin, R. K. (1994). Case study research: Design and methods. London, United Kingdom: Sage.
Yin, R. K. (2003). Case study research: Design and methods (3rd ed.). Thousand Oaks, CA:
Sage.
Zach, L. (2006). Using a multiple-case studies design to investigate the information-seeking
behavior of arts administrators. Library Trends, 55, 4–21.

174

Tables
Table 4.1. Descriptive Statistics of Purposive Sample (N = 33) That Participated in
Semistructured Interviews
Women Men
(n = 15) (n = 18)

n % n %
Race and ethnicity

Black 11 73 11 61
White 1 7 5 27
Other 2 13 1 5
Hispanic or Latino 1 7 1 5
Scattered site 7 47 7 39
Skid Row 3 20 9 50
Tobacco use change across first year in housing
Increased 5 33 6 33
Decreased 5 33 4 23
Maintained 5 33 8 44
Tobacco use change at semistructured interview
Increased 5 33 3 17
Decreased 9 60 8 44
Maintained 1 7 7 39

175

Table 4.2. Summary of Thematic Findings at Multiple Socioecological Levels

Facilitators to Cessation Barriers to Cessation
Accessibility Being far away from accessing
cigarettes (e.g. stores are not easy
to get to; no convenient place to
smoke, etc) and/or non-smoking
policies in place reduce smoking
Easy access to cigarettes (e.g.
people selling right outside
apartment, living within close
proximity to store, or can buy
cigarettes at any hour)
Serenity or chaos Aspects of neighborhood/ building
helped reduce smoking, or staying
inside to maintain peace and
quiet/safety (e.g. the quiet/ peace
of the neighborhood design of the
building, apartment as sanctuary)
Chaos, risk prevalence, and
the stress associated with
living in this setting is a
barrier to cessation
Support The presence of a person in their
lives reduced smoking (participant
does not specify any particular
support from that person, e.g. “My
grandkids coming over” or “Just
having my girlfriend around”) or
support from a person (participant
specifies the type of support- e.g.
“he told me what helped him quit”
or “she checked in with me often”
--
Socialization or
distancing
Avoiding others (either due to risk,
conflict or location) helped to
reduce smoking
Reduced interactions
(typically compared to when
homeless/prior to entry into
housing)/being alone often
made quitting harder
Cessation aids General discussion of cessation
aids including related doctor
interactions, used them and
positive/ negative experiences,
fears and/or ease of accessing them
--
Ubiquitous smoking -- Reason for continual smoking
is due to neighbors or
community members that
smoke, e.g. “everyone around
here/me smokes”
Stigma or conflict Reduced smoking as a result of the
stigma associated with smoking;
smoking seen as “bad” or overall
negatively
Social interactions that make
quitting challenging, mention
of peers/ relationships that
lead to smoking to cope
Providers Support from a provider facilitates
their reduction or cessation; doctor
advised to quit; doctor
Lack of provider support (e.g.
Provider has not asked or
intervened); Providers are

176

demonstrated care and concern
about health
also smokers; Difficulty
accessing providers/services
Health Concerns about health or aging
was a reason to reduce or stop
smoking
--
Financial Finances motivated them to reduce
or eliminate smoking/ discussion
of increase in cigarette price/taxes
--
Willpower Quitting/ reducing was attributed
or would be attributed to their
internal self-regulatory system and
the ability to quit without anyone
else’s help (e.g. “it has to come
from me”)
Lack of will power/self-
control/self-regulation is a
barrier to quitting
Distraction or boredom Reduced smoking by engaging in
alternative activities or became
focused on other goals, feeling too
busy with other things to do
Feelings of boredom, few
activities, limited engagement
in the community leads to
smoking (e.g. “to fill the
time”)
Mental health -- Stress (may or may not be
attributed to a specific
stimulus) or their mental
health/ specific mental
disorder makes it difficult to
quit; fear of mental health
symptoms that may ensure if
cessation is attempted
Substance use Using other substances (cannabis)
helps to reduce smoking tobacco;
Reducing/ ceasing other substance
use facilitates tobacco cessation
Use of other substances
drives tobacco use/cravings;
Reduction/cessation of other
substances leads to increased
smoking

177

Chapter 5: Conclusion
Introduction
This dissertation aimed to understand how smoking behavior changes from homelessness
to during the initial year in supportive housing and up to 5 years posthousing and the
socioecological factors that affect cross-sectional and longitudinal smoking outcomes. Prior to
this body of work, the extant literature had demonstrated that rates of tobacco use among adults
experiencing homelessness were disproportionately high compared to that of the general
population (Soar, Dawkins, Robson, & Cox, 2020). Additionally, numerous studies indicated
that interventions targeting cessation with homeless adults had poor postintervention
improvements in smoking outcomes (Businelle et al., 2014; Okuyemi et al., 2013; Rash, Petry, &
Alessi, 2018; Segan, Maddox, & Borland, 2015). Although this population’s high smoking rates
and poor cessation outcomes have been attributed to the abundance of institutional, social, and
individual barriers they face, few studies have examined these barriers or incorporated them into
cessation treatment for adults experiencing homelessness (Baggett et al., 2018; Okuyemi et al.,
2006; Pratt et al., 2019; Reitzel et al., 2014). Many of these barriers have the potential to be
addressed via housing and support in permanent supportive housing (PSH), yet there is minimal
research on tobacco use among supportive housing residents. What research does exist has
primarily focused on the implementation of smoke-free housing (Durazo et al., 2020; Petersen,
Stewart, Walters, & Vijayaraghavan, 2018), leaving little understanding of whether these barriers
are addressed posthousing and their impact on smoking for this vulnerable population.
The empirical studies depicted in this dissertation are the first of their kind. This body of
work reflects the first study to examine differential changes in tobacco use over time in a sample
of adults who entered PSH, the first study to examine physical and social environmental
correlates of smoking behavior change in the transition from homelessness to housing, and the

178

first study to qualitatively examine multilevel barriers and facilitators of cessation among
residents living in PSH for at least 3 years. Prior to this, only one study had examined changes in
smoking from homelessness to housing. Tsai and Rosenheck (2012) examined smoking behavior
across a year in a cohort of veterans, a portion of whom entered housing, and found that smoking
did not change over time. The study by Tsai and Rosenheck is distinct from the studies included
in this dissertation in that these studies drew from a cohort of civilians and veterans, all of whom
entered supportive housing, and found that smoking behavior did change over time and that
barriers and facilitators across many socioecological levels affected smoking behavior cross-
sectionally and longitudinally.
The data this dissertation drew on (NIDA Grant No. R01DA036345; NIDA Grant No.
F31DA045429-02) are particularly singular for several reasons. The first being the longitudinal
nature of this information gathered with a population that is typically challenging to engage with
for long periods. The research team worked closely and diligently with participants in this study
to maintain routine contact. These efforts proved valuable—12 months posthousing, 91% of the
baseline sample was interviewed. Second, PSH residents interviewed for these studies were
housed in diverse settings across varying housing model types and neighborhoods. Rarely are
such comparisons possible with this population and to the author’s knowledge, this is the only
study in the United States that contains such rich, diverse housing information. Last, these data
are transdisciplinary, cutting across several disciplines (Guimarães, Pohl, Bina, & Varanda,
2019). Transdisciplinary data enhance the understanding of complex phenomena, incorporate
diverse perspectives and permit conclusions that link broad and case-specific information of the
issue at hand (Guimarães et al., 2019). The value of these transdisciplinary data will be
exemplified as the findings from this body of work are summarized and triangulated, guiding

179

theoretical frameworks are applied, and descriptive, normative conclusions are drawn to make
policy and practice recommendations.
Review of Major Findings and Integration with Existing Research
Using quantitative, social network, and qualitative data from formerly homeless PSH
residents, this dissertation identified barriers and facilitators to smoking behavior outcomes
across multiple socioecological levels. This work demonstrates that these levels often work in
conjunction and subsequently affect tobacco use, tobacco risk, and cessation in unique ways.
Specifically, this work highlights the varying patterns of smoking behavior that occur as adults
transition from homelessness to housing and the individual-level demographic and clinical
characteristics associated with differing trajectories. Further, findings indicate that residents’
social environment play a critical role in their tobacco risk and highlight the interconnectedness
of their housing and support systems associated with their tobacco risk. Last, this dissertation
gives voice to those with lived experience and provides context for how these socioecological
levels come together and remain salient barriers and facilitators several years posthousing. The
following subsections review the findings of Chapters 2–4, highlighting their contributions to the
existing literature on tobacco use among homeless adults and supportive housing residents;
explore the application of the guiding theoretical frameworks; and discuss practice and policy
implications of this work.
Chapter 2: Trajectories of Smoking Behavior and Change in the Transition from
Homelessness to Supportive Housing
Chapter 2 used longitudinal tobacco frequency data from the Transitions to Housing
study to conduct a repeated-measures latent class analysis (RMLCA) that identified varying
patterns of smoking behavior from homelessness to housing. Findings revealed three latent

180

classes: (a) low-frequency increasers, a subgroup of participants who presented with little to no
smoking at baseline (while homeless) and significantly increased their smoking over time; (b)
high-frequency maintainers, comprising the largest subgroup of participants who were daily or
almost daily smokers and maintained this level of smoking across time; and (c) mixed-frequency
decreasers, who presented with varying smoking frequencies and significantly reduced their
tobacco use over time. This is the first study to identify any change in smoking from
homelessness to housing, as the study by Tsai and Rosenheck (2012) noted smoking significantly
changed over the year participants were observed. Chapter 2 highlights the value of using an
RMLCA, because the use of a more common multilevel modeling approach considered all
tobacco users in aggregate and found that smoking frequency significantly declined over time.
We know now that this was only the case for a portion of residents.
Chapter 2 also examined demographic and clinical correlates of class membership. Race,
mental health conditions, and substance use behaviors emerged as significant correlates of class
membership. Specifically, when compared to Black participants, Latino participants had more
representation in the mixed-frequency decreaser and low-frequency increaser classes. Research
has indicated Latino adults have high rates of nondaily smoking behavior (Kaplan et al., 2014),
which aligns with findings from Chapter 2. Importantly, research has also suggested Latino
adults are less likely to use cessation aids to support their quit efforts (Kaplan et al., 2014).
Qualitative research has found the primary reasons Latinos attempt cessation include the impact
and health concerns of family (Carter-Pokras et al., 2011), which may imply there are unique
cultural factors at play in Latino PSH residents’ tobacco use and cessation outcomes. Although
this was the only correlate of race significant in the multivariable models, other racial categories

181

approached significance, which may insinuate considering race in cessation efforts for this
population could be of importance.
Regarding clinical characteristics, comparisons across subgroups indicated high-
frequency maintainers were more likely to have a diagnosis of bipolar disorder. This being the
sole mental health diagnosis associated with class membership is surprising, given research has
found other mental health conditions are associated with smoking (Harris, Winetrobe, Rhoades,
& Wenzel, 2019; Tsai & Rosenheck, 2012; Vijayaraghavan, Tieu, Ponath, Guzman, & Kushel,
2016). A systematic review of the relationship between bipolar disorder and tobacco use
cessation indicated cessation is particularly challenging for persons with this diagnosis because
other substances are used at higher rates, rapid cycling of mood is a major barrier, medication
adherence is poor, and providers have expressed concerns about the interactions of cessation
medications with bipolar medications (George, Wu, & Weinberger, 2012). Changes in cannabis
use were associated with membership in the high-frequency maintainers and mixed-frequency
maintainers subgroups, which may highlight an integral relationship between tobacco and
cannabis. Overall, research has indicated the relationship between cannabis and tobacco is
unique in that these substances are often used in conjunction, yet are also commonly replaced by
the other (Agrawal, Budney, & Lynskey, 2012). Because this swapping of substances can
prolong tobacco use across the lifespan, dual abstinence has been recommended to maximize
cessation outcomes (Agrawal et al., 2012), although this may conflict with guiding principles of
harm reduction in PHS (Tsemberis, Gulcur, & Nakae, 2004).

182

Chapter 3: Tobacco Risk Change in the Transition from Homelessness to Housing: The
Role of Housing and Support
Chapter 3 examined changes in tobacco risk across residents’ initial year in supportive
housing. The examination of tobacco risk was particularly important for this population given the
goals of recovery and harm reduction in PSH (Tsemberis et al., 2004; Watson, Shuman,
Kowalsky, Golembiewski, & Brown, 2017) in addition to labeling smoking as a risk, which
acknowledges the immense harm it can have on health and the environmental influences that
play a role (Alemanno, 2012). Using latent growth curve modeling, a significant quadratic
relationship was identified indicating that tobacco risk decreases in the first 9 months or so, with
a slight increase at 12 months posthousing. This study also examined residents’ housing and
support as correlates of tobacco risk change. Support measures, encompassing structural,
functional, and provider support, were time-variant measures, and unfortunately, many of these
support measures significantly decreased over time. The deficits in supportive housing residents’
social relationships have been an ongoing concern and challenge for providers (Harris, Dunton,
et al., 2019; Henwood et al., 2015; Padgett, Tiderington, Smith, Derejko, & Henwood, 2016;
Patterson, Moniruzzaman, & Somers, 2014; Tsai, Mares, & Rosenheck, 2012; Yanos, Barrow, &
Tsemberis, 2004).
This paper also examined the cross-sectional effects of housing and support measures, in
addition to the longitudinal effect of social networks measures, and the interactive effect between
housing and support associated with tobacco risk. Main-effect models revealed that residents in
downtown Los Angeles (DTLA) had higher rates of tobacco risk, cross-sectionally. Main effects
also indicated that many of the potentially prosocial support factors were associated with
increased tobacco risk. The interactive effects demonstrated that change some housing model

183

and neighborhood factors moderated the relationship between several support factors and
tobacco risk. Many of the significant interactive effects indicated that changes in support
measures and their effect on tobacco risk were distinctly opposite across housing models and
neighborhoods. Specifically, results suggested that scattered-site and DTLA residents’ may
reduce their tobacco risk as a result of increasing their isolative behaviors, whereas single-site
and Skid Row residents may face the very real challenge of navigating the risk in which they are
entrenched. Although minimal studies have compared single-site and scattered-site housing
models, those that have found few distinctions (Harris, Dunton, et al., 2019). However,
Whittaker, Dobbins, Swift, Flatau, and Burns (2017) found single-site residents’ injection drug
use increased over time, which supports findings from Chapter 3 that when residents are located
in high-risk environments, their risk is exacerbated. A prior cross-sectional examination of these
support measures in this sample indicated Skid Row residents reported higher rates of functional
support (Harris, Dunton, et al., 2019). This study elucidated the salience of these findings,
demonstrating that more support may also mean greater risk longitudinally.
Chapter 4: Barriers and Facilitators to Cessation among Formerly Homeless Supportive
Housing Residents
The third paper in this dissertation drew on primary qualitative data from semistructured
interviews that explored supportive housing residents’ perceptions of the barriers and facilitators
to cessation across physical, social, and individual socioecological levels. This study found that
residents identified barriers and facilitators across each socioecological level and provided rich
context for how these levels work in conjunction to facilitate or hinder cessation for this
population. Moreover, this study identified novel information about barriers and facilitators not
examined in the prior quantitative studies. These data are particularly unique given that residents

184

were purposively sampled based on their quantitative tobacco use trajectories in their initial year
in housing, which provides the ability to advance understanding of how these differential tobacco
use trajectories ensue several years posthousing.
There are two findings that stand out as particularly salient given their additive
contribution to the understanding of PSH residents’ tobacco use. One is the role of nonsmoking
policies in and around residents’ buildings, which was a facilitator to reducing smoking
identified by many residents. This is a particularly important finding because it was not a
primary focus of the initial qualitative interviews, yet emerged as a strong and salient facilitator.
Additionally, there has been a great deal of debate surrounding this issue, with one side arguing
residents’ housing stability and security are at risk if these rules are imposed and subsequently
violated, and the other side arguing for the benefits of smoke-free housing to the tobacco user in
addition to their neighbors’ rights to health (Levy, Adams, & Adamkiewicz, 2017). Efforts to
increase and implement smoke-free housing policies in homeless service settings (e.g., shelter,
transitional housing) and supportive housing have been gaining popularity particularly in recent
years. Qualitative focus groups with transitional housing residents in San Diego revealed
tobacco-using residents were receptive to smoke-free housing because it limited secondhand
smoke exposure among nonsmokers and children (Vijayaraghavan, Hurst, & Pierce, 2017).
Another study of homeless adults staying in shelters found that 75% of those in shelters with
smoke-free policies attributed their reduced smoking to the policy (Vijayaraghavan et al., 2017).
Chapter 4 findings affirm the results of these studies with currently homeless adults, indicating
these issues remain relevant in PSH. They add to this existing research with findings that suggest
residents are concerned about and committed to upholding smoke-free policies because they
align with their desire to maintain the cleanliness and security of their apartments.

185

The second critical finding is that of willpower, which was by far the most commonly
reported thematic finding, identified as both a facilitator and a barrier. Residents’ discussion of
willpower often came in conjunction with rejections of providers or other social influences
intervening in or supporting their cessation. This finding aligns with a study by Collins et al.
(2018) that found adults experiencing homelessness preferred engaging in self-defined and
directed cessation efforts (e.g., vaping, replacing smoking with alternative activities, self-talk).
Homeless participants in the Collins et al. (2018) study also indicated they did not perceive
simple advice to quit as helpful and desired providers to use more nonjudgmental and
compassionate approaches, in addition to allowing homeless adults to choose cessation
mechanisms. Perceived control and autonomy among homeless adults and PSH residents
intersect with other relevant considerations for this population. The first being that the
experience of homelessness, which inherently strips away perceived control and autonomy, may
lead this population to view smoking as one of their few expressions of autonomy and may be
driving the desire for autonomous control of cessation (Baggett, Tobey, & Rigotti, 2013).
The second crucial factor that may be involved in the willpower finding is race and
experienced racism. Black and African American persons are disproportionately represented in
the homeless and supportive housing populations (U.S. Department of Housing and Urban
Development, 2018), and 72% of residents who participated in the semistructured interviews
identified as Black, African American, or biracial. There are likely cultural responses to
structural racism at play in the willpower finding (Powell, 2007). Studies that have explored
facets of structural racism and their impact on health outcomes and health care utilization have
found Black and African American people receive lower-quality health care, view health care
settings as overly institutionalization, report higher mistrust of health care providers, and feel left

186

out of personal health decision processes (Freeman et al., 2017). Consequently, these issues
heighten feelings of dehumanization surrounding Black and African American people’s health
(Freeman et al., 2017). Together the issues of willpower, autonomy and racial inequities point to
a dire need for heightened empathy and autonomous support from providers serving this
population, along with further research that explores these issues in greater detail.
Integration, Triangulation, and Theoretical Application
Together, Chapters 1–3 provide a robust understanding of barriers and facilitators across
several socioecological levels, including the physical, social, and individual levels. Moreover,
the triangulation of cases in Chapter 4, in addition to the triangulation of the varying types of
data and methodological approaches used across these chapters, enriches the understanding of
how these socioecological levels work in unique ways to affect smoking outcomes for supportive
housing residents (Flick, 2004). Chapter 2 used quantitative data to identify three latent classes
of residents who increased, decreased, or maintained their smoking from homelessness to
housing and examined individual-level correlates of these tobacco use trajectories. Chapter 3
added to Chapter 2 by examining correlates from two additional socioecological levels, the social
and physical levels, and began to examine how these levels work in conjunction with one another
by examining egocentric network data, which are rich and rare sources of information related to
residents’ support system (DeJordy & Halgin, 2008). Chapter 4 drew on the quantitative tobacco
use frequency data used in Chapter 2 to purposively sample across patterns of smoking behavior
in the transition to housing. Although the patterns of behavior identified for purposive sampling
purposes were preliminary, they aligned with the significant smoking trajectories identified in
the RMLCA used in Chapter 2, further validating these phenomena of change and stability from
homelessness to housing (Mark & Shotland, 1987). Chapter 4 provides the milieu for how

187

multilevel factors come together, including how unique pairings of cross-level barriers and
facilitators differentially affect smoking and cessation.
The application of the guiding theoretical frameworks elucidates the benefits of the
triangulated data and methods employed in this dissertation. Social learning theory (SLT)
highlights the integral role of the social environment as the avenue for behavior acquisition,
maintenance, and change. However, SLT also places great emphasis on the role of individual
cognitive processes in behavior and behavior change (Rosenstock, Strecher, & Becker, 1988).
For example, expectancies about personal competence, referred to as self-efficacy, can influence
individual behavioral outcomes (Rosenstock et al., 1988) and has been identified as a critical
component to changing substance use behaviors, including tobacco use (Kadden & Litt, 2011).
Poor mental health and use of other substances affect cognition and are associated with poor
smoking cessation self-efficacy (Bandura & Walters, 1977; Nordgren, van der Pligt, & van
Harreveld, 2008). Although fewer mental health and substance use correlates were associated
with class membership than expected, Chapter 2 highlights the role of mental health and use of
other substances in smoking behavior change and maintenance.
According to SLT, the social environment is the primary function of behavior, because
social interactions reinforce or punish behaviors (Bandura & Walters, 1977; Perry, Baranowski,
& Parcel, 1990). Chapter 3 findings imply that residents are likely have high-risk networks and
many network components decreased from homelessness to housing, which was associated with
a subsequent decline in tobacco risk over time, supporting SLT. Additionally, this study found
the physical environment moderated the relationship between the social environment and
tobacco risk, thus aligning with SLT’s position that the physical environment “sets the stage” for
the social environment (Bandura & Walters, 1977; Perry et al., 1990). Chapter 4 bridges findings

188

from Chapters 2 and 3, illustrating the application of SLT. In Chapter 4, we gain clarity
regarding the role of mental health and substance use in residents’ cognition, finding that few
residents attributed their smoking to a specific mental health condition or use of another
substance, but nearly all residents’ struggling to quit smoking endorsed a lack of willpower, a
term that is interchangeable with self-efficacy (Rosenstock, Strecher & Becker 1988). Further
many residents attributed stress as a barrier to cessation as well as a hindrance to enhancing
willpower. Stress was driven by social relationships with family, peers or neighbors and the
social settings one was housed in, such as those living in Skid Row who described the chaos,
crime and risk within this neighborhood as a source of stress and barrier to cessation.
Application of the social ecological model to this work furthers the understanding of the
interrelation of these socioecological levels and how they come together to affect smoking
behavior (Macintyre & Ellaway, 2000). While SLT acknowledges the physical environment as
playing a part in the social environment and individual behavior, the social ecological model
purports that these levels are nested, such that one does not exist without the outer level that
encompasses it (Visser, 2007). In Chapter 2 we come to understand that individual, or
intrapersonal factors contribute to differential trajectories of tobacco use behavior. The
associations within the individual level could be considered in isolation, however the
socioecological suggests that behavior change is maximized when outer level barriers are also
considered and addressed (Macintyre & Ellaway, 2000). Certainly, this rings true with findings
from Chapter 3, as the interplay between the physical and social environments, or institutional
and interpersonal levels, was reflected in housing mode and neighborhood moderating the
relationships between changes in support measures and change in tobacco risk. Moreover, these
significant interactions highlight the nested nature of these socioecological levels. Many of the

189

support measures were not significant main effects implying that understanding the impact of
changes in support on tobacco risk change can only be achieved by considering the interrelation
between the between the physical and social environment.
Again, Chapter 4 brings an enriched understanding of the unique pairings of these socio-
ecological levels and their subsequent impact on smoking behavior. In Chapter 4 clarity is gained
surrounding the ways residents in similar settings can vary in responses to their changing social
environments, leading to distinct smoking behavior changes. For residents in scattered-site
housing the combination of being far from social relationships and consequent isolation
facilitated boredom and increased or maintained smoking. Contrastingly, for some scattered-site
residents’ being distanced from their social relationships meant that their exposure to risk was
reduced and this facilitated smoking reductions. Chapter 4 also elucidates how differing settings
can lead to similar behavioral responses. For residents outside of Skid Row the decreased
accessibility to tobacco, positive provider relationships and increased engagement in goal-
oriented activities reduced smoking, while for residents’ in Skid Row distancing oneself from the
chaos and crime in their neighborhoods led to seeking refuge in their apartment and remove
themselves from the abundant access to tobacco products, facilitating a focus on health and
subsequent reductions in smoking. Disentangling these nuanced pathways are integral for
achieving a holistic understanding and maximizing tailored efforts that cut across levels to
address the widespread tobacco use among this population.
Policy and Practice Implications, Recommendations, and Future Research
Findings from this dissertation draw implications that cut across socioecological levels.
These conclusive thoughts are discussed in combination with recommendations for changes in
practice, policy, and future research in the following subsections.

190

Physical Environment
Findings from this body of work point to the integral role of neighborhood and housing
model, and their interrelation, in PSH residents’ tobacco use and socialization. Overall, results
indicate that although smoke-free housing policies appear to be a primary facilitator to smoking
cessation, other housing policies and practices in addition to deficits in community resources
culminate to negatively affect residents’ health and well-being. These issues include the
tendency to house PSH residents in fringe, marginalized neighborhoods, the practice of
concentrating risk to certain locales, a lack of public transportation, a lack of available affordable
housing, and a lack of choice in housing. Although scattered-site housing placements, which
largely reflect housing Black of African American adults experiencing homelessness in more
affluent White or Caucasian neighborhoods, this practice is rare in Los Angeles and may be a
declining trend nationally. Due to housing availability and “not in my backyard” discriminatory
practices, there is an increasing tendency to house scattered-site residents on the outskirts of the
city. Such housing placement often reflect greater housing satisfaction by PSH residents and
have demonstrated little neighbor opposition based on race or class (Palmer, 2016), but also lead
Black and African American residents to report low satisfaction with their neighbors and
subsequent hindrances in socialization and integration (Palmer, 2016). For Los Angeles
scattered-site residents, placement in marginalized areas far from resources, services, and
prosocial relationships and with few public transportation options also hindered their community
integration, and as seen in Chapters 3 and 4, for some residents, this isolation drives tobacco use.
The lack of affordable, available housing in Los Angeles and across other metropolises
with high density and poverty have led to the practice of concentrating homeless and formerly
homeless persons together, via their neighborhood (e.g., Skid Row) or housing model (i.e.,

191

single-site). Although a cross-sectional examination of data with Transitions to Housing Study
sample found Skid Row residents reported higher support, findings from this dissertation suggest
that being housed in an environment where persons with shared lived experiences of
homelessness are abundant may also equate to being entrenched in risk. Despite the numerous
proximal services in Skid Row, efforts to address individual health behaviors may be moot if
environmental barriers are so prevalent. Although the Housing First model endorses scattered-
site housing to maximize community integration for marginalized homeless persons, single-site
housing was born out of the need to increase available housing for the growing homeless
population in this nation (Tsemberis et al., 2004; U.S. Department of Housing and Urban
Development, 2018). Both housing models have demonstrated strong retention outcomes, and
there are proponents of the implementation of Housing First in single-site housing models
(Malone, Collins, & Clifasefi, 2015; Montgomery et al., 2019; Parsell, Petersen, & Moutou,
2015). This is now the second empirical finding that substance use and associated risk increases
over time for single-site residents (Whittaker et al., 2017). Although the balancing act of
increasing PSH housing and choice in housing in addition to managing adverse risk outcomes is
by no means easily achieved, improvements to further understand and address these concerns are
warranted.
Social Environment
Social support. Although changes in PSH residents’ support systems posthousing were
largely unknown prior to this work, the principles of PSH warranted an assumption that some
barriers to prosocial relationships would be overcome post housing (Tsemberis et al., 2004).
Instead, results suggest residents likely have a dearth of prosocial persons in their lives,
especially in close proximity to their housing. Moreover, many residents appear to be resorting to

192

isolation to reduce risk, which defies the core principle of recovery in PSH. Understanding how
critical support can be in facilitating smoking cessation and other health improvements may
increase efforts to address this issue. Research has suggested and this dissertation supports the
role of structural support, which is simply the existence of a social tie (Holt-Lunstad & Uchino,
2015), as a key facilitator to cessation. In urban community samples, familial relationships,
specifically relationships with children and grandchildren, are associated with cessation
(Rosenthal et al., 2013). Because the experience of homelessness and the high rates of mental
health and substance use among homeless adults often lead to severed ties with family and loved
ones (De Venanzi, 2008), supporting family reunification and enhancing opportunities for
familial engagement may be critical to improve cessation, as they have for the general population
(Holt-Lunstad & Uchino, 2015).
Functional support, which describes the actual and perceived functions (e.g.,
instrumental, emotional, informational) delivered in a social relationship, is a longstanding
correlate of morbidity and mortality (Holt-Lunstad & Uchino, 2015; Uchino, 2009; Wills &
Sinar, 2000). The important facet of functional support that can be gleaned to address PSH
residents’ deficits in this area is the heightened role of perceived support, beyond that of received
support in health outcomes. That is, the perception that someone will be there to talk through an
emotionally challenging situation is a greater predictor than that of the person directly delivering
support (Holt-Lunstad & Uchino, 2015). Although the deficits in this population’s support
system may seem abundant, there are tangible, practical ways that perceived support can be
improved. These include increased social interactions and engagement and greater relational
quality, which can be generated by increasing positive interactions and consistency in social
relationships (Martire & Franks, 2014). The finding that providers’ demonstrating care for

193

residents’ health, sometimes without direct cessation intervention, was a facilitator to cessation
aligns with this idea. These support resources are invaluable and much needed in residents’
personal networks and relationships with their providers. In addressing these support deficits, it
is important to acknowledge the interrelation between the physical and social environments. For
PSH residents in settings with exacerbated tobacco rates and risk, social support may be
particularly lacking, with social norms bolstering smoking behavior (Rosenthal et al., 2013).
Providers. Adults experiencing homelessness have reported they perceive that providers
place smoking cessation as a low priority (Okuyemi et al., 2006). Survey research with
homelessness service providers including health care providers, mental health counselors, and
frontline providers has suggested prioritization of tobacco use varies across clinical disciplines,
with health care providers placing it much higher than frontline providers (Baggett et al., 2012).
Although health care providers acknowledged the importance of cessation, they attributed
inconsistencies in addressing it to competing priorities, including other mental health and health
conditions (Baggett et al., 2012). It is known that health care providers have demanding jobs,
often battling limited timeframes in which they work; therefore, increasing time with patients to
provide thorough counseling and education related to cessation is likely a more challenging
recommendation, although a much needed one nonetheless (Greene, Hibbard, Alvarez, &
Overton, 2016). Other potentially more accessible, less burdensome ways to facilitate health care
providers’ intervening in PSH residents’ tobacco use include forming partnerships with patients
and allowing them to own their treatment plan, identifying small, manageable steps of behavior
change, following up frequently, problem solving, showing caring and concern for patients,
providing self-help materials, and utilizing telehealth options including telephone support calls or
text message check-in and reminders (Greene et al., 2016; West et al., 2015). Additionally

194

providers may not understand the desire to quit smoking expressed by this population, because
PSH residents tend to be long-term, high-frequency tobacco users, which like race is associated
with disparities in providers’ delivery of cessation interventions and should be considered as
provider support is adapted and improved (Danesh, Paskett, & Ferketich, 2014).
Healthcare providers are more likely to perceive cessation as a priority than frontline
providers (e.g. case managers; Baggett et al., 2012). This disparity may be attributed to several
issues within frontline provider services that may require attention to improve cessation
outcomes for PSH residents. Frontline providers have far more frequent and direct exposure to
residents’ daily activities and behaviors than other service providers. The challenges of serving a
population characterized by high risk and poor health and mental health are associated with high
rates of burnout (Olivet, McGraw, Grandin & Bassuk 2010; Reyes, 2016; Schiff, & Lane, 2019).
cite). Moreover, case managers within PSH often lack training in evidence-based practices, are
paid low wages, and are commonly positioned as disciplinarians, blurring the lines between
service providers and landlords/ property managers (Tiderington, 2019). Considering this, the
high turnover of frontline providers in PSH is not surprising, although is nonetheless deserving
of attention given the prior recommendations to increase functional support by improving
relational consistency (Martire & Franks, 2014). Findings from Chapter 4 also support prior
research that has found frontline providers smoke at high rates and use smoking as tool to build
rapport within client interactions (Baggett et al., 2012; Okuyemi et al., 2006), and incentivize
clients with cigarettes (Baggett et al., 2012; Prochaska, Hall & Bero, 2008). Intervening in
frontlines providers’ tobacco use may be valuable for building their desire and willingness to
address PSH residents smoking via cessation methods or harm reduction. Attending to some of
the previously described organizational factors within PSH services, may also improve frontline

195

providers motivation to address PSH residents smoking, and their own (Olivet, McGraw,
Grandin & Bassuk 2010).
Individual
Mental health and substance use. Chapter 2 examined individual level demographic
and clinical characteristics correlated with varying smoking trajectories. These findings in
addition to body of research that has identified mental health and substance use drive smoking
and hinder cessation (Harris, Winetrobe, et al., 2019; Tsai & Rosenheck, 2012; Vijayaraghavan
et al., 2016), along with the racial disparities in smoking, cessation rates, and services to address
tobacco use (Danesh, Paskett & Ferketich, 2014) warrant the incorporation of these issues into
tailored interventions for PSH residents. This mandates future research that disentangles the over
the overlap of these issues to adapt or establish best practices.
Person-centered efforts. The findings from this work draw several implications for
improving and increasing broad individual-level interventions that target PSH residents’
smoking. The first being the clear need for efforts that prioritize providers’ empathy and
acceptance toward their clients. Efforts guided by such principle emphasize providers
confirmation that they understand or attempt to understand the client’s perspective and express
unconditional positive regard (Champassak, Catley, et al, 2014). The second being interventions
that promote client choice and autonomous support, which is reflected in providers’ elicitation of
clients’ health goals, allowing the client to explore avenues of achievement, in addition to
conveying the notion that change lies within the client and is not be imposed by others
(Champassak, Catley, et al., 2014). These principles align with those of motivational
interviewing and patient activation, which are directive patient‐centered counseling approaches
that target self-efficacy and have demonstrated effectiveness in improving cessation outcomes

196

(Hibbard & Greene, 2013; Lindson-hawley, Thompson & Begh, 2015). Last this work points to
the need for harm reduction efforts that incorporate tobacco use as a risk behavior. This may
seem counter-intuitive to promoting motivational interviewing and patient activation, which may
aim to extinguish tobacco use, however these approaches rest on many shared principles and
acknowledge behavior change as a stage process (Watson, Shuman, Kowalsky, Golembiewski &
Brown, 2017). Future work is needed to draw the parallels across these efforts and support
providers, especially those on the front lines, to identify and exercise these tools to address much
needed health issues and become allies for a population that feels voiceless and marginalized.
Conclusion
Findings from this work are rich, yet the stark reality of the level and magnitude of
barriers formerly homeless supportive housing residents’ face in achieving cessation comes with
a sense of heaviness. There is no single solution to address this population’s smoking behavior as
this research demonstrates the pathways to smoking and cessation are vast and varied. This body
of work has nonetheless carved out crosscutting socio-ecological factors that negatively affect
cessation outcomes and warrant improved practice, policy and further research efforts.
Specifically, increased smoke-free housing and improvements in housing policies that consider
residents community integration, improved support including from providers, and incorporating
mental health and substance use challenges into adapted humanitarian focused efforts and
evidence-based practices are greatly needed. Smoking among homeless and formerly homeless
populations is widespread and multi-faceted and the efforts to their health should reflect that.

197

References
Agrawal, A., Budney, A. J., & Lynskey, M. T. (2012). The co‐occurring use and misuse of
cannabis and tobacco: A review. Addiction, 107, 1221–1233. doi:10.1111/j.1360-
0443.2012.03837.x
Alemanno, A. (2012). The EU’s Tobacco Products Directive seeks to ‘nudge’ citizens whilst
preserving individual choice about smoking. British Politics and Policy. Retrieved from
https://blogs.lse.ac.uk/politicsandpolicy/smoking-nudge-eu-alemanno/
Baggett, T. P., Anderson, R., Freyder, P. J., Jarvie, J. A., Maryman, K., Porter, J., & Rigotti, N.
A. (2012). Addressing tobacco use in homeless populations: A survey of health care
professionals. Journal of Health Care for the Poor and Underserved, 23, 1650–1659.
doi:10.1353/hpu.2012.0162
Baggett, T. P., Chang, Y., Yaqubi, A., McGlave, C., Higgins, S. T., & Rigotti, N. A. (2018).
Financial incentives for smoking abstinence in homeless smokers: A pilot randomized
controlled trial. Nicotine and Tobacco Research, 20, 1442–1450. doi:10.1093/ntr/ntx178
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.
doi:10.1056/NEJMp1301935
Bandura, A., & Walters, R. H. (1977). Social learning and personality development. London,
United Kingdom: Holt, Rinehart, and Winston.
Businelle, M. S., Kendzor, D. E., Kesh, A., Cuate, E. L., Poonawalla, I. B., Reitzel, L. R., …
Wetter, D. W. (2014). Small financial incentives increase smoking cessation in homeless
smokers: A pilot study. Addictive Behaviors, 39, 717–720.
doi:10.1016/j.addbeh.2013.11.017

198

Carter-Pokras, O. D., Feldman, R. H., Kanamori, M., Rivera, I., Chen, L., Baezconde-Garbanati,
L., ... & Noltenius, J. (2011). Barriers and facilitators to smoking cessation among Latino
adults. Journal of the National Medical Association, 103(5), 423-431.
Champassak, S. L., Catley, D., Finocchario-Kessler, S., Farris, M., Ehtesham, M., Schoor, R., &
Goggin, K. (2014). Physician smoking cessation counseling and adherence to a clinical
practice guideline. European journal for person centered healthcare, 2(4), 477.)
Collins, S. E., Orfaly, V. E., Wu, T., Chang, S., Hardy, R. V., Nash, A., … Clifasefi, S. L.
(2018). Content analysis of homeless smokers’ perspectives on established and
alternative smoking interventions. International Journal of Drug Policy, 51, 10–17. doi:
10.1016/j.drugpo.2017.09.007
Danesh, D., Paskett, E. D., & Ferketich, A. K. (2014). Disparities in receipt of advice to quit
smoking from health care providers: 2010 National Health Interview Survey. Preventing
Chronic Disease, 11, E131. doi:10.5888/pcd11.140053
De Venanzi, A. (2008). The institutional dynamics of homelessness. International Journal of
Sociology and Social Policy, 28, 129–145. doi:10.1108/01443330810862197
DeJordy, R., & Halgin, D. (2008). Introduction to ego network analysis. Retrieved from
http://www.analytictech.com/e-net/pdwhandout.pdf
Durazo, A., Hartman-Filson, M., Perez, K., Alizaga, N. M., Petersen, A. B., & Vijayaraghavan,
M. (2020). Smoke-free home intervention in permanent supportive housing: A
multifaceted intervention pilot. Nicotine & Tobacco Research. Advance online
publication. doi:10.1093/ntr/ntaa043
Flick, U. (2004). Triangulation in qualitative research. In U. Flick, E. von Kardoff, & I. Steinke
(Eds.), A companion to qualitative research (pp. 178–183). Thousand Oaks, CA: Sage.

199

Freeman, R., Gwadz, M. V., Silverman, E., Kutnick, A., Leonard, N. R., Ritchie, A. S., …
Martinez, B. Y. (2017). Critical race theory as a tool for understanding poor engagement
along the HIV care continuum among African American/Black and Hispanic persons
living with HIV in the United States: A qualitative exploration. International Journal for
Equity in Health, 16, 54. doi:10.1186/s12939-017-0549-3
George, T. P., Wu, B. S., & Weinberger, A. H. (2012). A review of smoking cessation in bipolar
disorder: implications for future research. Journal of Dual Diagnosis, 8, 126–130.
doi:10.1080/15504263.2012.671717
Greene, J., Hibbard, J. H., Alvarez, C., & Overton, V. (2016). Supporting patient behavior
change: approaches used by primary care clinicians whose patients have an increase in
activation levels. Annals of Family Medicine, 14, 148–154. doi:10.1370/afm.1904
Guimarães, M. H., Pohl, C., Bina, O., & Varanda, M. (2019). Who is doing inter-and
transdisciplinary research, and why? An empirical study of motivations, attitudes, skills,
and behaviours. Futures, 112, 102441. doi:10.1016/j.futures.2019.102441
Harris, T., Dunton, G., Henwood, B., Rhoades, H., Rice, E., & Wenzel, S. (2019). Los Angeles
housing models and neighbourhoods’ role in supportive housing residents’ social
integration. Housing Studies, 34, 609–635. doi:10.1080/02673037.2018.1462308
Harris, T., Winetrobe, H., Rhoades, H., & Wenzel, S. (2019). The role of mental health and
substance use in homeless adults’ tobacco use and cessation attempts. Journal of Dual
Diagnosis, 15, 76–87. doi:10.1080/15504263.2019.1579947
Henwood, B. F., Stefancic, A., Petering, R., Schreiber, S., Abrams, C., & Padgett, D. K. (2015).
Social relationships of dually diagnosed homeless adults following enrollment in housing

200

first or traditional treatment services. Journal of the Society for Social Work and
Research, 6, 385–406. doi:10.1086/682583
Hibbard, J. H., & Greene, J. (2013). What the evidence shows about patient activation: better
health outcomes and care experiences; fewer data on costs. Health affairs, 32(2), 207-
214.
Holt-Lunstad, J., & Uchino, B. N. (2015). Social support and health. In K. Glanz, B. K. Rimer, &
K. Viswanath (Eds.), Health behavior: Theory, research and practice (pp. 183–204). San
Francisco, CA: Jossey-Bass.
Kadden, R. M., & Litt, M. D. (2011). The role of self-efficacy in the treatment of substance use
disorders. Addictive Behaviors, 36, 1120–1126. doi:10.1016/j.addbeh.2011.07.032
Kaplan, R. C., Bangdiwala, S. I., Barnhart, J. M., Castañeda, S. F., Gellman, M. D., Lee, D. J.,
… Giachello, A. L. (2014). Smoking among US Hispanic/Latino adults: The Hispanic
Community Health Study/Study of Latinos. American Journal of Preventive Medicine,
46, 496–506. doi:10.1016/j.amepre.2014.01.014
Levy, D. E., Adams, I. F., & Adamkiewicz, G. (2017). Delivering on the promise of smoke-free
public housing. American Journal of Public Health, 107, 380–383.
doi:10.2105/AJPH.2016.303606
Lindson‐Hawley, N., Thompson, T. P., & Begh, R. (2015). Motivational interviewing for
smoking cessation. Cochrane Database of Systematic Reviews, (3).
Macintyre, S., & Ellaway, A. (2000). Ecological approaches: rediscovering the role of the
physical and social environment. Social epidemiology, 9(5), 332-348.

201

Malone, D. K., Collins, S. E., & Clifasefi, S. L. (2015). Single-site Housing First for chronically
homeless people. Housing, Care and Support, 18, 62–66. doi:10.1108/HCS-05-2015-
0007
Mark, M. M., & Shotland, R. L. (1987). Alternative models for the use of multiple methods. New
Directions for Program Evaluation, 1987, 95–100. doi:10.1002/ev.1461
Martire, L. M., & Franks, M. M. (2014). The role of social networks in adult health: Introduction
to the special issue. Health Psychology, 33, 501–504. doi:10.1037/hea0000103
Montgomery, A. E., Gabrielian, S., Cusack, M., Austin, E. L., Kertesz, S. G., & Vazzano, J.
(2019). Applying the Housing First approach to single-site permanent supportive
housing. Journal of Social Distress and the Homeless, 28, 24–33.
doi:10.1080/10530789.2018.1546796
Nordgren, L. F., van der Pligt, J., & van Harreveld, F. (2008). The instability of health
cognitions: Visceral states influence self-efficacy and related health beliefs. Health
Psychology, 27, 722–727. doi:10.1037/0278-6133.27.6.722
Okuyemi, K. S., Goldade, K., Whembolua, G. L., Thomas, J. L., Eischen, S., Sewali, B., …
Resnicow, K. (2013). Motivational interviewing to enhance nicotine patch treatment for
smoking cessation among homeless smokers: A randomized controlled trial. Addiction,
108, 1136–1144. doi:10.1111/add.12140
Okuyemi, K. S., Thomas, J. L., Hall, S., Nollen, N. L., Richter, K. P., Jeffries, S. K., …
Ahluwalia, J. S. (2006). Smoking cessation in homeless populations: A pilot clinical trial.
Nicotine & Tobacco Research, 8, 689–699. doi: 10.1080/14622200600789841

202

Olivet, J., McGraw, S., Grandin, M., & Bassuk, E. (2010). Staffing challenges and strategies for
organizations serving individuals who have experienced chronic homelessness. The
journal of behavioral health services & research, 37(2), 226-238.
Padgett, D. K., Tiderington, E., Smith, B. T., Derejko, K. S., & Henwood, B. F. (2016). Complex
recovery: Understanding the lives of formerly homeless adults with complex needs.
Journal of Social Distress and the Homeless, 25, 60–70.
doi:10.1080/10530789.2016.1173817
Palmer, G. L. (2016). Examining the effects of scattered site supportive housing on the social
and economic integration of men who are formerly homeless and primarily Black/African
American. Journal of Black Studies, 47, 846–868. doi:10.1177/0021934716653351
Parsell, C., Petersen, M., & Moutou, O. (2015). Single-site supportive housing: Tenant
perspectives. Housing Studies, 30, 1189–1209. doi:10.1080/02673037.2015.1009874
Patterson, M. L., Moniruzzaman, A., & Somers, J. M. (2014). Community participation and
belonging among formerly homeless adults with mental illness after 12 months of
Housing First in Vancouver, British Columbia: A randomized controlled trial.
Community Mental Health Journal, 50, 604–611. doi:10.1007/s10597-013-9672-9
Perry, C. L., Baranowski, T., & Parcel, G. S. (1990). How individuals, environments, and health
behavior interact: Social learning theory. In K. Glanz, F. M. Lewis, & B. K. Rimer (Eds),
Health behavior and health education: Theory, research, and practice (pp. 161–186).
San Francisco, CA: Jossey-Bass.
Petersen, A. B., Stewart, H. C., Walters, J., & Vijayaraghavan, M. (2018). Smoking policy
change within permanent supportive housing. Journal of Community Health, 43, 312–
320. doi:10.1007/s10900-017-0423-7

203

Powell, J. A. (2007). Structural racism: Building upon the insights of John Calmore. North
Carolina Law Review, 86, 791–816.
Pratt, R., Pernat, C., Kerandi, L., Kmiecik, A., Strobel-Ayres, C., Joseph, A., … Okuyemi, K.
(2019). “It’s a hard thing to manage when you’re homeless”: The impact of the social
environment on smoking cessation for smokers experiencing homelessness. BMC Public
Health, 19, 635. doi:10.1186/s12889-019-6987-7
Prochaska JJ, Hall SM, Bero LA. (2008.) Tobacco use among individuals with schizophrenia:
What role has the tobacco industry played? Schizophr. Bull. 34:555–67
Rash, C. J., Petry, N. M., & Alessi, S. M. (2018). A randomized trial of contingency
management for smoking cessation in the homeless. Psychology of Addictive Behaviors,
32, 141–148. doi:10.1037/adb0000350
Reitzel, L. R., Kendzor, D. E., Nguyen, N., Regan, S. D., Okuyemi, K. S., Castro, Y., …
Businelle, M. S. (2014). Shelter proximity and affect among homeless smokers making a
quit attempt. American Journal of Health Behavior, 38, 161–169.
doi:10.5993/AJHB.38.2.1
Reyes, K. (2016). Exploring Compassion Fatigue, Burnout, Compassion Satisfaction and
Mindfulness in Direct Service Providers (Doctoral dissertation, California State
University, Northridge).
Rosenstock, I. M., Strecher, V. J., & Becker, M. H. (1988). Social learning theory and the health
belief model. Health Education Quarterly, 15, 175–183.
doi:10.1177/109019818801500203
Rosenthal, L., Carroll-Scott, A., Earnshaw, V. A., Sackey, N., O’Malley, S. S., Santilli, A., &
Ickovics, J. R. (2013). Targeting cessation: understanding barriers and motivations to

204

quitting among urban adult daily tobacco smokers. Addictive Behaviors, 38, 1639–1642.
doi:10.1016/j.addbeh.2012.09.016
Schiff, J. W., & Lane, A. M. (2019). PTSD symptoms, vicarious traumatization, and burnout in
front line workers in the homeless sector. Community mental health journal, 55(3), 454-
462.
Segan, C. J., Maddox, S., & Borland, R. (2015). Homeless clients benefit from smoking
cessation treatment delivered by a homeless persons’ program. Nicotine and Tobacco
Research, 17, 996–1001. doi:10.1093/ntr/ntv062
Soar, K., Dawkins, L., Robson, D., & Cox, S. (2020). Smoking amongst adults experiencing
homelessness: A systematic review of prevalence rates, interventions and the barriers and
facilitators to quitting and staying quit. Journal of Smoking Cessation, 15, 94–108.
doi:10.1017/jsc.2020.11
Tiderington, E. (2019). Practice Dilemmas, Successes, and Challenges in the Delivery of
Homeless Services: Voices from the Frontline. In Homelessness Prevention and
Intervention in Social Work (pp. 379-394). Springer, Cham.
Tsai, J., Mares, A. S., & Rosenheck, R. A. (2012). Does housing chronically homeless adults
lead to social integration? Psychiatric Services, 63, 427–434.
doi:10.1176/appi.ps.201100047
Tsai, J., & Rosenheck, R. A. (2012). Smoking among chronically homeless adults: Prevalence
and correlates. Psychiatric Services, 63, 569–576. doi:10.1176/appi.ps.201100398
Tsemberis, S., Gulcur, L., & Nakae, M. (2004). Housing First, consumer choice, and harm
reduction for homeless individuals with a dual diagnosis. American Journal of Public
Health, 94, 651–656. doi:10.2105/AJPH.94.4.651

205

Uchino, B. N. (2009). Understanding the links between social support and physical health: A
life-span perspective with emphasis on the separability of perceived and received support.
Perspectives on Psychological Science, 4, 236–255. doi:10.1111/j.1745-
6924.2009.01122.x
U.S. Department of Housing and Urban Development. (2018). The 2018 Annual Homeless
Assessment Report (AHAR) to Congress: Part 1: Point-in-time estimates of homelessness.
Washington, DC: U.S. Department of Housing and Urban Development. Retrieved from
https://files.hudexchange.info/resources/documents/2018-AHAR-Part-1.pdf
Visser, M. (2007). The social ecological model as theoretical framework in community
psychology. Community psychology: Analysis, context and action, 102-116.
Vijayaraghavan, M., Hurst, S., & Pierce, J. P. (2017). A qualitative examination of smoke-free
policies and electronic cigarettes among sheltered homeless adults. American Journal of
Health Promotion, 31, 243–250. doi:10.4278/ajhp.150318-QUAL-781
Vijayaraghavan, M., Tieu, L., Ponath, C., Guzman, D., & Kushel, M. (2016). Tobacco cessation
behaviors among older homeless adults: Results from the HOPE HOME study. Nicotine
& Tobacco Research, 18, 1733–1739. doi:10.1093/ntr/ntw040
Watson, D. P., Shuman, V., Kowalsky, J., Golembiewski, E., & Brown, M. (2017). Housing First
and harm reduction: A rapid review and document analysis of the US and Canadian open-
access literature. Harm Reduction Journal, 14, 30. doi:10.1186/s12954-017-0158-x
West, R., Raw, M., McNeill, A., Stead, L., Aveyard, P., Bitton, J., … Borland, R. (2015).
Health ‐care interventions to promote and assist tobacco cessation: A review of efficacy,
effectiveness and affordability for use in national guideline development. Addiction, 110,
1388–1403. doi: 10.1111/add.12998

206

Whittaker, E., Dobbins, T., Swift, W., Flatau, P., & Burns, L. (2017). First examination of
varying health outcomes of the chronically homeless according to Housing First
configuration. Australian and New Zealand Journal of Public Health, 41, 306–308.
doi:10.1111/1753-6405.12631
Wills, T. A., & Shinar, O. (2000). Measuring perceived and received social support. In S. Cohen,
L. G. Underwood, & B. H. Gottlieb (Eds.), Social support measurement and intervention:
A guide for health and social sciences (pp. 86–135). New York, NY: Oxford University
Press.
Yanos, P. T., Barrow, S. M., & Tsemberis, S. (2004). Community integration in the early phase
of housing among homeless persons diagnosed with severe mental illness: Successes and
challenges. Community Mental Health Journal, 40, 133–150.
doi:10.1023/b:comh.0000022733.12858.cb

Abstract (if available)

Abstract Permanent supportive housing (PSH) offers permanent housing and supportive services to formerly homeless tobacco users with high rates of mental health and substance use conditions, enhancing opportunities to overcome the numerous barriers to cessation experienced while homeless, including physical (e.g., housing instability, proximity to services) and social (e.g., emotional support, provider support) environmental barriers, yet no empirical research has explored whether these characteristics of PSH affect tobacco use behavior change. This body of work is the first to examine associations between the physical, social and individual characteristics of PSH and tobacco use behavior change, and the differential tobacco use change pathways across chronic mental health conditions and risky substance use behaviors, among homeless adults transitioning to housing, using multi-level social ecological data (i.e., quantitative, social network, qualitative). Results from this research enhance the identification of new directions in efforts needed to improve cessation among supportive housing residents. Specifically, identifying how housing, social support, mental health and substance use factors can be altered or incorporated into efforts to improve cessation outcomes for PSH residents.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Homelessness and substance use treatment: using multiple methods to understand risks, consequences, and unmet treatment needs among young adults

PDF

Tobacco and marijuana surveillance using Twitter data

PDF

U.S. Latinx youth development and substance use risk: adversity and strengths

PDF

Couch-surfing among youth experiencing homelessness: an examination of HIV risk

PDF

Homeless female veterans—silent epidemic

PDF

Engaging homeless men and shelter providers to adapt an existing evidence-based HIV prevention intervention

PDF

Technology enhanced substance use disorder treatment

PDF

Understanding emotional regulation and mood of young adults in the context of homelessness using geographic ecological momentary assessment

PDF

Rethink Homelessness project

PDF

Social network engagement and HIV risk among homeless former foster youth

PDF

Increasing access to mental health counseling for homeless youth: Peer2Peer Counseling Supports

PDF

Pain and multi-morbidity among veterans: theory-guided, data-driven, and narrative approaches

PDF

Building a trauma-informed community to address adverse childhood experiences

PDF

WISER women’s program: well-being innovation with support and education for resilience—a homelessness prevention intervention

PDF

Substance use disorder treatment clinician and director challenges working with collaborative justice court and child welfare systems: A multisite qualitative study

PDF

Unto the least of these homeless ministry: ending homelessness within the co-occurring population

PDF

Social network norms and HIV risk behaviors among homeless youth in Los Angeles, California

PDF

Behavioral Health for All Kids

PDF

Integrative care strategies for older adults experiencing co-occurring substance use and mental health disorders (I-CARE)…

PDF

The algorithm of Black homelessness: a classification tree analysis

Asset Metadata

Creator Harris, Taylor (author)

Core Title Tobacco use change among formerly homeless supportive housing residents: socioecological barriers and facilitators to cessation

School Suzanne Dworak-Peck School of Social Work

Degree Doctor of Philosophy

Degree Program Social Work

Publication Date 08/01/2020

Defense Date 06/22/2020

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag cessation,health,Homelessness,Housing,Mental Health,Motivation,neighborhood,OAI-PMH Harvest,smoking,social networks,social support,substance use,tobacco use

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Wenzel, Suzanne (committee chair), Henwood, Benjamin (committee member), Unger, Jennifer (committee member)

Creator Email tayeharris@gmail.com,taylorh@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-354124

Unique identifier UC11666479

Identifier etd-HarrisTayl-8843.pdf (filename),usctheses-c89-354124 (legacy record id)

Legacy Identifier etd-HarrisTayl-8843.pdf

Dmrecord 354124

Document Type Dissertation

Rights Harris, Taylor

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA