ENGAGEMENT IN HIV CARE DURING COMMUNITY REENTRY: A MIXED METHODS
EXAMINATION OF INDIVIDUAL AND RELATIONAL CHALLENGES
By
Danielle Chiaramonte

A THESIS
Submitted to
Michigan State University
In partial fulfillment of the requirements
For the degree of
Psychology – Master of Arts
2017

ABSTRACT
ENGAGEMENT IN HIV CARE DURING COMMUNITY REENTRY: A MIXED METHODS
EXAMINATION OF INDIVIDUAL AND RELATIONAL CHALLENGES
By
Danielle Chiaramonte
The domestic HIV epidemic remains a serious public health concern in the United States.
The rate of HIV in prisons is three five times higher than the national average. The process of
reentering the community after release from prisons places formerly incarcerated individuals at
high risk of disengagement from HIV care. The purpose of the current study was to explore the
factors that distinguish patterns in care engagement and disengagement over a 39-month period
following release from prison. Using a socioecological framework, this project combined
archival and interview data to investigate the impact of individual- and relational-level factors on
these patterns. Rates of disengagement in this sample were very high, with only 17% of the
sample consistently engaging in care. Non-engagers made up 30% of the sample and inconsistent
engagers made up 50% of the sample. The accumulation of individual-level risk factors predicted
engagement such that those with more individual-level risk were more likely to never engage or
inconsistently engage in care when compared to consistent engagers. Poverty, history of
substance use, and heterosexual identity were associated with non-engagement and inconsistent
engagement. Consistent engagers were more likely to be satisfied with their patient-provider
relationships than inconsistent engagers. The varied patterns of engagement suggest it is doubtful
that there is a “one-size-fits-all” intervention for HIV care engagement for ex-offenders.
Practitioners should tailor interventions to individuals’ specific needs.

ACKNOWLEDGEMENTS

I would like to thank several people who have supported be throughout this process. First,
to my thesis committee, Dr. Robin Miller, Dr. Miles McNall, and Dr. Ignacio Acevedo, thank
you for your advice and support throughout this pursuit. Robin, I am so grateful for your
mentorship and all you’ve taught me during this journey. I could never have done this without
you.
I would also like to thank all my friends and family members who supported me over the
course of this project. To Zoe, Monica, Gretchen, and Lauren whose friendship has been
invaluable, even from afar. To Trevor, and Olga who assisted me when I needed them. To
Hannah and Rachael, who were there after every thesis meeting and every new draft; you two
pull me up every time I’m down and consistently inspire me as researchers, mentors and friends.
To my council, Kerri and Britt, who have supported me through every twist and turn of this
journey. I could never have done this without our weekly council calls and your unwavering
confidence in me. To my parents and my brothers, who always encouraged me to work hard,
love what I do, and to never stop laughing along the way.

iii

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................................................ vii
LIST OF FIGURES ..................................................................................................................... viii
INTRODUCTION ...........................................................................................................................1
LITERATURE REVIEW ................................................................................................................3
Understanding Engagement in Care: Disparities along the HIV Care Continuum ......................3
Community Reentry .....................................................................................................................7
Socioecological Framework .........................................................................................................9
Individual-level Influences on Engagement in Care ..................................................................11
Race....................................................................................................................................11
Income................................................................................................................................12
Employment .......................................................................................................................12
Educational attainment.......................................................................................................13
Comorbid health conditions ...............................................................................................13
Transportation ....................................................................................................................14
Housing ..............................................................................................................................14
Summary ....................................................................................................................................15
Relational-level Influences on Engagement in Care ..................................................................16
Social Support ....................................................................................................................16
Social support and health .......................................................................................16
Social support and HIV ..........................................................................................17
Satisfaction with social support .............................................................................18
Network size ..........................................................................................................19
Patient-provider relationships ................................................................................19
Summary ....................................................................................................................................21
CURRENT STUDY.......................................................................................................................23
METHODS ....................................................................................................................................24
Mixed Methods ..........................................................................................................................24
Archival data ......................................................................................................................25
Interviews ...........................................................................................................................25
AIDS Partnership Michigan’s Community Reentry Program Evaluation .................................26
Participants .................................................................................................................................27
Sample ........................................................................................................................................29
Procedure ....................................................................................................................................30
Abstracting client records ..................................................................................................30
In-person interviews...........................................................................................................31
Determining incarceration status ...........................................................................31
Determining mortality status..................................................................................31
iv

Locating current addresses .....................................................................................31
Additional data sources ......................................................................................................35
Michigan Department of Community Health CAREWare database .............................35
The Michigan State Police database...............................................................................35
APM Record ...................................................................................................................36
Measures .....................................................................................................................................36
Engagement in care.................................................................................................................36
Missing data ...................................................................................................................38
Sexual orientation ...................................................................................................................38
Social support .........................................................................................................................38
Social support network size count .....................................................................................39
Support satisfaction score ..................................................................................................39
Patient-provider relationship ..................................................................................................40
Individual-level index .............................................................................................................40
Educational attainment.......................................................................................................41
Race....................................................................................................................................41
Income................................................................................................................................41
Substance use history .........................................................................................................42
Mental health history .........................................................................................................42
Employment .......................................................................................................................42
Housing stability and satisfaction ......................................................................................42
Transportation ....................................................................................................................43
Scoring individual-level index ...............................................................................43
Analysis ......................................................................................................................................44
RESULTS ......................................................................................................................................46
Demographic Characteristics .....................................................................................................46
Descriptive Characteristics .........................................................................................................49
Indicators of individual-level risk ......................................................................................49
Indicators of relationship quality .......................................................................................49
Network size ..........................................................................................................49
Satisfaction with social support .............................................................................49
Satisfaction with medical provider ........................................................................49
Hypothesis Testing .....................................................................................................................52
Hypothesis 1.......................................................................................................................52
Hypothesis 2.......................................................................................................................52
Hypothesis 3.......................................................................................................................52
Hypothesis 4.......................................................................................................................53
Hypothesis 5.......................................................................................................................53
DISCUSSION ................................................................................................................................57
Patterns of Care Engagement .....................................................................................................57
Factors that Influenced Care Engagement .................................................................................59
Limitations .................................................................................................................................63
Implications for Research and Practice ......................................................................................65
Conclusion ..................................................................................................................................68
v

APPENDICES ...............................................................................................................................69
Appendix A: Individual-level Risk Index Validation Study ......................................................70
Appendix B: APM Abstraction Form ........................................................................................73
Appendix C: Eligibility Screening Instrument for Interviews ...................................................81
Appendix D: Interview Recruitment Materials ..........................................................................85
Appendix E: Interview Consent Form .......................................................................................90
Appendix F: Measures ...............................................................................................................92
REFERENCES ............................................................................................................................102

vi

LIST OF TABLES

Table 1. Demographic Characteristics of APM Reentry Program Clients ....................................28
Table 2. Demographic Characteristics by Care Engagement (N=60)............................................48
Table 3. Indicators of Individual-level Factors by Care Engagement (N=60) .............................49
Table 4. Indicators of Relationship Quality by Care Engagement (N=60)....................................51
Table 5. Model 1: Hypothesis Testing with Three Outcomes .......................................................55
Table 6. Model 2: Hypothesis Testing with Two Outcomes .........................................................56

vii

LIST OF FIGURES

Figure 1. Health Resources and Services Administration continuum of HIV care, describing the
spectrum of engagement in HIV care and levels of consistency in care engagement ....................3
Figure 2. Disparity along the HIV care continuum..........................................................................4
Figure 3. Results of Iroh et al., (2015) study of HIV care before, during and after incarceration
compared with the national average. Reprinted from “The HIV care cascade before, during, and
after incarceration: a systematic review and data synthesis,” by P.A. Iroh, H. Mayo, A.E.
Nijhawan, 2015, American Journal of Public Health, 105, p. e12. .................................................6
Figure 4. A socioecological perspective of HIV care engagement. Adapted from Mugavero, “The
state of engagement in HIV care in the United States: From cascade to continuum to control.
M.J. Mugavero, K.R. Amico, T. Horn, and M.A. Thompson, 2013, Clinical Infectious Diseases,
57, p. 1166......................................................................................................................................10
Figure 5. Development of Analytic Sample ..................................................................................28
Figure 6. APM Recruitment Flyer .................................................................................................89

viii

INTRODUCTION
The domestic Human Immunodeficiency Virus (HIV) epidemic represents a serious
public health concern. Estimates suggest 1.2 million youth and adults live with HIV in the
United States (Center for Disease Control and Prevention [CDC], 2015). Unfortunately, 14% are
unaware of their condition (CDC, 2015). Even more alarming, our nation systemically fails to
support and engage individuals in their HIV care.
“Engagement in care” or “care engagement” refers to a spectrum of patient care and
disease management, from initial diagnosis of HIV to full commitment to care. The degree to
which an individual is engaged in their HIV care can affect the manageability of the HIV
infection (Gardner, McLees, Steiner, del Rio & Burman, 2011). Consistent engagement in care
can lead to improved antiretroviral therapy (ART) adherence and achievement of viral
suppression (Giordano et al., 2007; Mugavero, 2008). Viral suppression signifies the virus is
undetectable and likely cannot be transmitted to others (Cheever, 2007; Gardner et al., 2011,
Mugavero, Amico, Horn & Thompson, 2013; George et al., 2009). Inconsistent or lack of
engagement in care can lead to further transmission of the virus or escalation of HIV infection to
AIDS (Gardner, McLees, Steiner, del Rio & Burman, 2011; Giordano et al., 2007; Mugavero,
2008). Despite the benefits of engagement in care, only an estimated 40% of people living with
HIV (PLWH) consistently engage (Mugavero, Norton & Saag, 2011; CDC, 2014; Althoff, et al.,
2013; Bozzette, 1988; CDC, 2015).
Certain populations have a lower likelihood of consistent care engagement, indicating a
potential disparity in health outcomes (CDC, 2015). The rate of HIV in prisons is three to five
times higher than the national average, placing formerly incarcerated individuals at high risk for
disengagement from HIV care. (Baillargeon, et al., 2010; Bureau of Justice Statistics, 2002;
Westergaard, Spaulding, & Flanigan, 2013; Brinkley-Rubinstein & Turner, 2013).
1

Every year, hundreds of thousands of inmates who receive healthcare and housing in
prison reintegrate back into the community, many of whom have an HIV-positive diagnosis
(Springer et al., 2004; Meyers, 2014; Springer et al., 2007). This process is generally referred to
as reentry and is typically represents the immediate months to 3 years after release from prison
(Visher, La Vigne & Travis, 2004; Travis, 2005; Dhami Mandel, Loewenstein & Ayton, 2006).
This time is especially critical for PLWH. In 1997, up to one-quarter of all people reentering into
the community after release from prison were PLWH or AIDS (Hammett, Roberts & Kennedy,
2001). Despite this group’s high risk for falling out of care during reentry, care engagement
remains largely under-investigated. In order to address this gap in care engagement, we must
understand the barriers and facilitators formerly incarcerated PLWH face to engaging in care
post-incarceration. Consequently, my project focuses on identifying factors that distinguish
individuals who consistently engaged in care from those who inconsistently engage in care, and
from those who do not engage in care at all over the community reentry period following release
from prison.

2

LITERATURE REVIEW
Understanding Engagement in Care: Disparities along the HIV Care Continuum
The continuum of HIV care (CoC) illustrates stages along the spectrum of engagement in
HIV care (unaware of HIV infection, aware of HIV status, may be receiving other medical care
but not HIV care, entering HIV primary medical care but dropped out, and fully engaged in HIV
care) (AIDS, 2015; Eldred & Maltiz, 2007; Cheever, 2007). In Figure 1, I display a modified
version of the continuum. The figure illustrates the stages along the spectrum of engagement as
well as the different patterns of care engagement that correspond with each stage (nonengagement, inconsistent engagement, and consistent engagement in HIV care).

Figure 1. Health Resources and Services Administration continuum of HIV care, describing the
spectrum of engagement in HIV care and levels of consistency in care engagement (adapted from
Eldred & Maltiz, 2007).
The beginning stages indicate awareness of the HIV infection through receipt of an HIV
test result. A positive HIV test should be followed by referral and linkage to an initial medical
care visit with a health care provider (Hall, Tang, Westfall & Mugavero, 2013; Torian et al.,
3

2008). Subsequently, individuals should receive routine medical monitoring, also referred to as
retention in care, and gain access to ART. Retention refers to the number of kept appointments
and appropriate spacing between appointments over time (Mugavero, et al., 2013). Retention in
care facilitates access to benefits beyond medical monitoring including prevention counseling
and supportive resources (Hall, et al., 2012; Alberg et al., 2009). If an individual is retained in
HIV care and adherent to treatment regimens, the virus can be suppressed.
Ideally, after receiving an HIV test, those who have HIV will immediately link to care,
after which they will consistently attend medical appointments, receive and adhere to ART, and
as a result, reduce their viral load. However, in reality, the HIV care continuum functions more
like a leaky pipe where movement through succeeding stages steadily declines. Some PLWH
stop care entirely or never begin in the first place (CDC, 2014; Eldred & Maltiz, 2007;
Mugavero, et al., 2013). In the United States, 86% of PLWH are aware of their diagnoses,
signifying they received an HIV test and a positive result (see Figure 2). However, only 40% of
all PLWH are consistently engaged in HIV care. Furthermore, only 37% of PLWH are receiving
ART, and merely 30% have achieved viral suppression (CDC, 2014).

Figure 2. Disparity along the HIV care continuum (CDC, 2011).
In response to the need for increased engagement in care, The White House released the
National Strategy for HIV/AIDS, the first comprehensive and coordinated effort the United States
has taken on HIV (The White House, 2010). The strategy included three goals: reducing new
infections, increasing access to care, and reducing HIV-related health disparities (The White
4

House, 2010). Furthermore, in July 2013, President Obama issued Executive Order No. 13,649
(2013) establishing the HIV Care Continuum Initiative, which prioritized addressing disparities
along the continuum of care. Finally, in July 2015, The White House released an update to the
original 2010 strategy focusing on providing support at all stages of the care continuum,
particularly to specific populations at highest risk for falling out of care (The White House,
2015).
Although lack of engagement in care is undoubtedly a problem among the general
population of PLWH, national averages mask disparities across the HIV care continuum among
specific subgroups of PLWH (Hall et al., 2013). In a systematic literature review of 92 studies,
Iroh, Mayo, and Nijhawan (2015) examined the HIV care continuum before, during, and after
incarceration. As summarized in Figure 3, taken from Iroh et al., after release, rates of linkage to
care, retention in care, and ART adherence significantly dropped. (Iroh, Mayo, & Nijhawan,
2015, p. e12). Consequently, viral suppression rates also fell after release. The graph also depicts
the disparity in care engagement and treatment adherence for recently released PLWH compared
with the general populations of PLWH.

5

Figure 3. Results of Iroh et al., (2015) study of HIV care before, during and after incarceration
compared with the national average. Reprinted from “The HIV care cascade before, during, and
after incarceration: a systematic review and data synthesis,” by P.A. Iroh, H. Mayo, A.E.
Nijhawan, 2015, American Journal of Public Health, 105, p. e12.
As noted in Iroh et al.’s (2015) study and other current literature, recently released
PLWH have a harder time consistently engaging in care and exhibit increased viral loads as
compared with currently incarcerated individuals (Stephenson, 2005; Palpeu et al., 2004; Wohl
et al., 2011; Baillargeon et al., 2009). One possible explanation for this disparity is that the
prison system permits easy access to medications as well as essentials for survival (e.g. food,
housing), potentially resulting in better control of the virus (Springer et al., 2004; Meyer, et al.,
2014 Springer, et al., 2007). Altice and colleagues (2001) found that approximately two thirds of
inmates living with HIV are diagnosed for the first time in prison, suggesting that formerly
incarcerated PLWH are not accustomed to interacting with HIV health systems in their
communities. In a retrospective cohort study of all PLWH released from the Texas Department

6

of Justice prison system from 2004 to 2007, Baillargeon and colleagues (2009) found that within
the immediate community reentry period (30 days of release), only about 20% of inmates had
refilled their ART prescriptions. A better understanding of what drives these critical disparities
during the reentry period is essential to addressing this disparity in HIV care engagement among
formerly incarcerated PLWH. The following section reviews the difficulties that arise during
community reentry.
Community Reentry
Community reentry refers to the process through which formerly incarcerated individuals
transition back into the general community. Reentry can refer to the immediate months after
release up to 3 years, although most literature only tracks reentry for a few months (Visher, La
Vigne & Travis, 2004; Travis, 2005; Dhami Mandel, Loewenstein & Ayton, 2006). Regardless
of HIV status, reentry can be physically, financially, and emotionally demanding (Richie, 2001).
Formerly incarcerated individuals may reenter communities with little or no preparation for the
demands of reentry, thus compounding challenges associated with successful reentry (Visher,
Mallik-Kane, 2007; Hammett, Roberts & Kennedy, 2001). Upon release, formerly incarcerated
individuals are expected to secure housing, obtain employment, reestablish relationships with
family and friends, and gain access to other vital resources (e.g., personal identification, food
stamps, substance use resources, transportation, mental health services) (Dennis et al., 2015;
Brinkley-Rubinstein & Turner, 2013; Harlow, 2003; Richie, 2001; La Vigne & Mamalian, 2003;
Petersilia, 2003). Oftentimes, securing vital resources must be accomplished while contending
with little access to money, limited supportive relationships, low literacy skills, low self-efficacy,
few transferable skills, and substance use and mental health disorders (Travis, Solomon & Waul,

7

2001). In addition to these challenges, formerly incarcerated PLWH face the added pressure to
engage in HIV care immediately upon release.
Navigating healthcare systems serves as a continual challenge for formerly incarcerated
PLWH, who face the burden of reestablishing care and medication access almost immediately
upon release. At release, inmates receive only 30 days of their HIV prescriptions, leaving them
roughly 1 month to find accessible care (e.g., new physician, new prescription) and identify
resources to support their care and cover medical expenses (e.g., money for prescription and
appointment co-payments, enrollment in insurance and benefit services). Completing
reenrollment applications for these services can be a laborious and complicated process.
Applications often consist of multiple steps, many of which require documents that formerly
incarcerated PLWH may not have. The AIDS Drugs Assistance Program (ADAP) application,
for instance, which provides low cost medications for PLWH, requires many forms including
proof of residence, income (or lack of it), HIV status, and insurance (or lack of it). Prisoners may
be released without accepted forms of proof of their identity and occasionally without proof of
HIV status or medication needs, requiring they undergo HIV testing to confirm their status
(Fontana & Beckerman, 2007). Procuring the proper documents to fill out forms may take
months.
Many formerly incarcerated individuals lack the skills needed to navigate these
complicated systems, particularly if these systems are brand new to inmates who were diagnosed
while incarcerated. Literacy skills, for example, are substantially lower among incarcerated
populations than the general public and can have profound implications (Petersilia, 2003).
Formerly incarcerated PLWH need to be able to understand the healthcare system, fill out health
forms, apply for benefits, and communicate adequately with physicians, all of which can be

8

stymied by low literacy skills. Moreover, formerly incarcerated individuals may lack the selfefficacy to make decisions related to employment, housing or health. While incarcerated,
inmates are stripped of their autonomy and agency (Haney, 2003). The structured nature of the
prison system limits them from making decisions related to finances, living conditions, and
health (Katzen, 2013; Visher & Malik-Kane, 2007). As such, it may be difficult for formerly
incarcerated PLWH to resume making these decisions upon release (Travis, 2001; Shinkfield &
Graffam, 2010). Formerly incarcerated PLWH may also be too slow to advocate on their own
behalf for rapid processing of ADAP, securing initial medical visits, and obtaining prescriptions
(Laufer, Arriola, Dawson-Rose, Kumaravelu & Rapposelli, 2002; Hammett et al., 1998; Travis,
2001). Moreover, HIV health care navigation during reentry may be particularly challenging for
those who have spent more time behind bars or are newly diagnosed and generally less familiar
with the changes to healthcare services since their incarceration (e.g., technological advances,
Affordable Care Act) (Katzen, 2013).
Taking all of these difficulties into consideration, the community reentry period is fraught
with challenges to securing basic needs. These challenges are exacerbated for PLWH, as they
must also manage their HIV.
Socioecological Framework
This project employs a social ecological perspective to examine factors that distinguish
individuals who consistently engage in care from those who inconsistently engage in care and
those who do not engage in care at all over a 39-month period following release from prison.
Examining factors associated with care engagement at multiple levels may explain care
engagement disparities in this population.

9

A socioecological framework (Figure 4) helps conceptualize factors contributing to
inconsistent or lack of engagement in care (Mugavero, Amico, Horn & Thompson, 2013;
DiClemente, Salazar & Crosby, 2007; McLeroy, Bibeau, Steckler & Glanz, 1988). The
framework illustrates dynamic interrelations among individual, interpersonal, community and
structural factors, and their impact on an individual’s ability to, in this case, adequately engage in
HIV care (Bronfrenbrenner, 1979; Mugavero et al., 2013). Mugavero’s (2013) adaptation of the
socioecological perspective considers five levels of influence on engagement in care; of those
five levels, the individual- and relational-levels are the focus of this project.
The limited literature on reentry challenges for PLWH suggest that individual and
relationship-level factors can facilitate or hinder engagement in care. The literature also suggests
that the individual- and relationship-level factors may interact to affect engagement in care. In
the current project, I use the socioecological framework to frame the synergistic impact of
individual and relational factors on engagement in care for formerly incarcerated PLWH.

Figure 4. A socioecological perspective of HIV care engagement. Adapted from Mugavero,
“The state of engagement in HIV care in the United States: From cascade to continuum to
control. M.J. Mugavero, K.R. Amico, T. Horn, and M.A. Thompson, 2013, Clinical Infectious
Diseases, 57, p. 1166.

10

Individual-level Influences on Engagement in Care
Individual characteristics impact the ease with which people reenter a community and
engage in care. Racial minority status, poverty, unemployment, low educational attainment, the
presence of co-occurring health conditions, housing instability, and lack of access to
transportation are each associated with poor engagement in HIV care and ART nonadherence
(Israelski et al., 2001; Traeger, O’Cleirigh, Skeer, Mayer, Safran, 2012; Giordano et al., 2003;
Cargill, 2012; Torian et al., 2008; Ulett et al., 2009; Rebeiro et al., 2013; Hall, Gray, Tang,
Shouse & Mermin, 2012; Mugavero et al., 2009; Pettus-Davis, Scheyett, Hailey, Golin & Wohl.,
2009; Rajabiun et al., 2007; Wolf et al., 2007; Dennis, 2015; Brinkley-Rubinstein & Turner,
2013).
Race. Black individuals account for almost half of all new HIV infections every year
(CDC, 2015), and have the highest imprisonment rates compared to people of other races (CDC,
2015; Bureau of Justice Statistics “BJS” 2014). Further, Blacks disproportionately experience
gaps in care along the engagement spectrum (CDC, 2012; Mugavero, Amico, Horn &
Thompson, 2013; Cargill, 2012). Studies consistently observe Black men experience delayed
linkage to care (Torian et al., 2008; Ulett et al., 2009), poor retention in care (Rebeiro et al.,
2013; Hall, et al., 2012), increased likelihood of missing appointments (Mugavero, et al., 2007;
Mugavero et al., 2009; Traeger, et al., 2012), delayed ART initiation and adherence (Giordano et
al., 2010), and longer time to achievement of viral suppression (Hall et al., 2013). Approximately
half of all PLWH are Black, a majority of whom are also young men who have sex with men
(YMSM) (CDC, 2015). Black YMSM are least likely to know their HIV status, with more than
half unaware of their infection (CDC, 2015). Among those who know their status, young Blacks
are significantly less likely to be in care, adherent, or virally suppressed (CDC, 2010; CDC,

11

2014; Hall et al., 2014). These disparities are, in part, a result of systematic discrimination and
racism historically experienced by Black communities (Delgado & Stefancic, 2012; Williams,
1999, Cooper & David, 1986).
Income. Income plays an important role in formerly incarcerated PLWH’s ability to
engage in their HIV care. Previous research finds low-income PLWH tend to miss more HIVrelated healthcare appointments than their higher-income counterparts (Israelski et al., 2001;
Traeger, et al., 2012; Giordano et al., 2003). Other studies have found that lacking adequate
finances to secure basic needs also hindered patients’ ability to engage in care (Dennis et al.,
2015; Haley et al., 2014)
Employment. Unemployment contributes to an inability to secure basic needs required to
consistently engage in HIV care (Dennis et al., 2015; Haley et al., 2014; Springer, Spaulding,
Meyer, & Altice, 2011). Obtaining employment, however, is especially difficulty for formerly
incarcerated individuals, yet crucial to successful reentry (Petersilia, 2003). Employment aids in
reentry by facilitating the development of life skills, financial stability, productivity, and social
ties (Petersilia, 2003). However, while incarcerated, inmates are oftentimes unable to cultivate
work histories or learn valuable skills necessary for most work forces and job applications
(Schnittker & John, 2007). Former inmates may also experience prejudice in hiring due to their
criminal histories. Living with HIV only exacerbates these challenges. PLWH have also
historically experienced discrimination from employers in the work force (Malcolm, et al.,
1998). Engagement in HIV care necessitates several doctors’ appointments a year and potential
side effects from ART medications can make it difficult to hold a steady job (Haley et al., 2014).
Collectively, the literature suggests employment is crucial for consistent access to care, however
formerly incarcerated PLWH face additional struggles in attaining employment, exacerbating

12

this challenge (Hall et al., 2013; Cargill, 2012; Rajabuin et al., 2007; Dennis, et al., 2015; MalikKane & Visher, 2008).
Educational attainment. Lower education and health literacy in PLWH has been
associated with nonadherence to HIV medication, increased emotional distress, less social
support, and increased barriers to accessing medical care (Wolf et al., 2007; Dennis et al., 2015;
Kalichman, Catz, Ramachandran, 1999; Cargill, 2012). Wolf and colleagues (2007) found that
PLWH with less than a high school education or who were functionally illiterate were least
likely to be adherent to HIV medications and attend HIV care appointments.
Comorbid health conditions. Many formerly incarcerated PLWH leave prison with
physical, mental health, or substance use conditions that can impact engagement in care (Traeger
et al., 2012; Mugavero et al., 2013; Cashman et al., 2004; Rajabiun, 2007; Stephenson, 2005;
Katzen, 2013; Mallik-Kane & Visher, 2008; Torian, 2008; Rebeiro, 2013, Ulett, et al., 2009). In
fact, it is rare for former prisoners to have only one type of health condition. Approximately half
of incarcerated individuals used substances prior to incarceration (Mumola & Karberg, 2004).
Mallik-Kane and Visher (2008) found four in ten men reported multiple co-occurring physical,
mental health, and substance use conditions during prisoner reentry. Thus, former inmates leave
prison sometimes doubly or triply diagnosed with serious health conditions on top of HIV. These
conditions alone may require specialized treatment and management (Mallik-Kane & Visher,
2008). Comorbid conditions may make it more difficult to acquire housing or employment and
maintain interpersonal relationships (Mallik-Kane & Visher, 2008). Comorbid psychosocial
health issues have also been associated with lower ART adherence and viral suppression
(Blashhill et al., 2014; Friedman et al., 2015). Consequently, comorbid physical, mental health,

13

and substance use issues in PLWH may hinder engagement in care among formerly incarcerated
individuals.
Transportation. Patients need to have reliable transportation to attend health care
appointments. Lack of access to quality transportation is an often-reported hindrance to
successful reentry, medical appointment attendance, and ART adherence (Dennis, 2015;
Buchberg et al., 2015; Scheyett et al., 2010; Althoff et al., 2012). Althoff and collegues (2012)
found that formerly incarcerated PLWH who received transportation assistance were more likely
to follow up on their post-release HIV medical appointments, suggesting that access to quality
transportation can influence the likelihood of care engagement.
Housing. Securing safe and stable housing is one of the most difficult challenges postincarceration for PLWH (Brinkley-Rubinstein & Turner, 2013; Haley et al., 2014; Visher,
LaVigne & Travis, 2004; Petersilia, 2003; Clear, Rose, & Ryder, 2001; Rich et al., 2001). The
National Alliance to End Homelessness estimated that homelessness rates in 2006 among
recently released inmates in major urban areas was as high as 30-50% (National Alliance to End
Homelessness, 2006). An estimated 3.4% of homeless individuals were PLWH, compared to
0.4% of adults and adolescents with HIV in the general population (CDC, 2008; National
Alliance to End Homelessness, 2006).
Acquiring housing for certain subgroups of PWLH is particularly challenging. Current or
former substance users, for instance, struggle to obtain housing because most low-income
housing does not accept residents with drug-related criminal histories (Brinkley-Rubinstein &
Turner, 2013; U.S. Department of Health and Human Services, n.d.). Sex offenders requiring
lifetime registration may face additional parole stipulations. They are restricted from living
within 1,000-2,000 feet of a schools or parks and are limited in their access to federally assisted

14

housing (Grubesic, Murray & Mack, 2011). Obtaining housing is especially difficult if former
prisoners cannot return to the home of a spouse, partner, or relative and are released to the streets
or a homeless shelter. Housing restrictions placed on sex offenders and substance users also
apply to friends’ and relatives’ housing, further curbing housing opportunities. Applying for
housing requires money for applications, deposits, and first months’ rent, which many people
may lack due to limited earnings during incarceration or due to post-incarceration
unemployment.
Consequently, formerly incarcerated individuals may be forced to move to unstable
neighborhoods not conducive to successful reentry or substance use recovery (Haley et al., 2014;
Visher, LaVigne & Travis, 2004). These communities are often burdened by high rates of
community unemployment, poverty, substance use, and are generally situated in resource-sparse
areas with poor transportation (Hill, Ross & Angel, 2005). Further, these communities are not
stable enough to support individuals reentering with HIV properly and thereby contribute to poor
engagement in care (Petersilia, 2000; Rich et al., 2001; Arnold, Hsu, McFarland & Rutherford,
2009; Shacham, Lian, Onon, Donovan & Overton, 2013; Latkin, German, Vlahov & Galea,
2013).
Summary
Current literature provides ample evidence of the relationship between individual-level
factors and lack of engagement in care. What remains to be understood is what happens when an
individual is burdened by several individual-level challenges. Few studies have investigated how
care engagement is impacted by an accumulation of risk factors. This accumulation of
individual-level risks may be more important in predicting care engagement than any specific
individual-level risk alone. I propose to examine the additive effect of individual-level risk

15

factors’ impact on engagement in care. The current study examines the impact of a composite
individual-level risk score on engagement in care. I expect consistent engagers to have lower
individual-level risk scores than inconsistent and non-engagers.
Relational-level Influences on Engagement in Care
Relationships can be a critical source of support and motivation for PLWH (Pierce,
Sarason & Sarason, 1996). Absence of or stress within relationships can hinder PLWH’s ability
to stay engaged in care (Galvan, Davis, Banks & Bing, 2008). The role of supportive
relationships in promoting engagement in care among formerly incarcerated PLWH remains
relatively unexplored.
Social Support
Social support encompasses the delivery or exchange of social resources, actual or
perceived (Cohen, Underwood & Gottlieb, 2000). Social support benefits PLWH directly and
indirectly. Directly, social support can influence the adoption of behaviors and social norms,
healthy or unhealthy (Cohen & Syme, 1985). Indirectly, social support can buffer the negative
effects of stressful life events (Cohen & Wills, 1985).
Social support and health. There is an extensive literature demonstrating the positive
influence of social support on health (House et al., 1985; Gallant, 2003; Cohen & Syme, 1985).
Empirical evidence links perceived social support to improved mental and physical health,
positive outlook, coping skills, and the ability to handle stressful life events. A lack of perceived
social support is associated with depression and psychological impairment during stressful life
events (Pierce, Sarason & Sarason, 1996; Serovich et al., 2001); Galvan, Davis, Banks & Bing,
2008). Cohen & Wills’ (1985) classic work on social support demonstrated the association
between health and support. Social support may be of particular importance to formerly

16

incarcerated PLWH, as they experience an increased amount of stress associated with the
practical and psychological aspects of reentry and HIV care management.
Social support and HIV. Never-incarcerated PLWH perceive social support from family
and friends as critical to HIV stress management and survival (Alfonso et al., 2006; Rajabiun,
2007; Edwards, 2006; Knowlton, Hua & Latkin, 2005). A strong network of relationships
facilitates engagement in HIV care (Kelly, Hartman, Graham, Kallen & Giordano, 2014; Alfonso
et al., 2006; George et al., 2009; Edwards, 2006; Giordano et al., 2007; Catz, Kelly, Bogart,
Benotsch & McAuliffe, 2000; Byrgoyne, 2005; Hagedoorn et al., 2000; DiMatteo, 2004). The
general social support literature does not, however, provide a sense of how social support affects
care engagement for groups prone to atypical social interactions and networks (Rajabuin et al.,
2007; Catz, et al., 2001; DiMatteo, 2004; Simoni, Mason & Marks, 1997; Edwards, 2006;
George, et al., 2009).
Dennis and colleagues’ (2015) qualitative study of reentry for PLWH found social
support to be one of the most crucial facilitators of successful HIV care engagement and reentry.
At a personal and institutional level, social support was considered instrumental in helping
respondents manage their HIV and navigate the reentry process. Similarly, Pettus-Davis and
colleagues (2009) utilized a social support framework to examine the role of social support in
post-release plans for prisoners living with HIV. Participants discussed barriers they would face
upon leaving prison (e.g., housing, employment, criminal history, stigma, restrictions) and
expected social support to be essential in overcoming these barriers. The majority of participants
acknowledged the importance of informational and instrumental social support and interpersonal
relationships, employment, and faith-based communities in managing their HIV care. Some
participants thought having at least one supportive person in their networks could aid in avoiding

17

substance use and other negative outcomes. Other participants anticipated little or no social
support upon release. Many participants also thought their HIV status would be a significant
barrier to achieving a normal life with friends and family.
Satisfaction with social support. Maintenance of supportive relationships is not easy in
the best of circumstances. The process may be severely strained by incarceration and reentry
(Alexander, 2012; Baillargeon et al., 2009). Incarceration of any length can disrupt important
relationships due to distance and a lack of regular communication and intimacy (Schnittker &
John, 2007; Rindfuss and Stephen 1990; Wheeler 1961). Incarceration can produce shame,
anger, and mistrust, further straining relationships with social networks (Hagan and Dinovitzer
1999; Braman 2004; Schnittker & John, 2007). The social costs of incarceration can affect entire
communities (Roberts, 2004). Removing key family members, partners, friends, or parents from
a social network can negatively influence those who depended on them as key support members
(Harawa & Adimore, 2008). Loved ones may lose important financial or emotional resources
necessary for day-to-day living and suffer as a result. Therefore, upon reentry, loved ones may
hold negative attitudes toward or feel ambivalent about former inmates.
In other cases, formerly incarcerated PLWH might return to the relationships that
contributed to their initial incarceration and that could be detrimental to their successful reentry.
Indeed, former inmates sometimes prefer to avoid this possibility by establishing new
relationships in different communities or enduring the hardships of social isolation rather than
take up old relationships upon release (Schnittker & John, 2007). Dennis and colleagues (2015)
found many formerly incarcerated PLWH reentering into the community returned to an
environment full of alcohol and drugs, negatively impacting their ability to adhere to ART
medications. Overall, although social networks are present, individuals may not be satisfied with

18

those networks’ positive contribution to HIV care. In this study, I expect consistent engagers to
be more satisfied with support received than inconsistent and non-engagers.
Network size. Research findings are not yet clear about whether there is an optimum
network size or composition for supporting PLWH in care engagement (Schaefer, Coyne, &
Lazarus, 1981). Support networks can be of various sizes and can come from sources inside or
outside of family, friends, and partner networks (Lewin, 2000; George et al., 2009). Two studies
suggest large social networks may better aid in health care management (Semple, 1996;
McDowell & Serovich, 2007). In this study, I expect consistent engagers to report larger
networks than inconsistent and non-engagers.
Patient-provider relationships. One key support member of a well-functioning network
for PLWH are HIV care providers. Patient-provider relationships are of critical importance for
PLWH, as maintenance of HIV requires long-term biomedical treatment (including
approximately two to three visits a year) (Mugavero et al., 2013; McCoy, 2005). As such, the
quality of patient-provider relationships for formerly incarcerated PLWH may influence
engagement in care.
Relevant literature on never-incarcerated PLWH consistently finds that patients want to
feel supported by their provider in their HIV care. Positive relationships with providers increase
the likelihood of medication adherence (Rajabiun et al., 2007; Laws, Bradshaw & Safren et al.,
2011; George, et al., 2009; Bakken et al., 2000; Cederfjall et al., 2002; Preau et al., 2004;
Roberts, 2002; Schneider, Kaplan Greenfield & Wilson, 2004), while more contentious patientprovider relationships are associated with reduced adherence (Roberts, 2002). Formerly
incarcerated PLWHs’ likelihood of encountering a contentious relationship may be higher due to
their individual characteristics such as substance use, medical distrust, stigma, and stress. There

19

is also evidence to suggest physician bias against HIV positive injection-drug users, patients of
belonging to ethnic or racial minorities, and patients of low SES (O’Connor, 1994; Gerbert,
Maguire, Bleecker, Coates & McPhee, 1991; Wong et al., 2004; King et al., 2004).
In a qualitative study, Rajabiun et al., (2007) investigated the process of engagement in
care in never incarcerated PLWH through in-depth interviews, finding those who were not stably
in care reported more negative experiences of health care providers than those with stable
engagement patterns. Unengaged participants disliked when providers exercised control over
health decisions without considering their experiences or needs. In addition, patients felt
providers’ attitudes or behaviors served as an obstacle to engaging in HIV care. By contrast,
participants who consistently engaged in care reported positive relationships with their providers,
which they valued.
Appointment aversion is common when patients perceive negativity or judgment from
providers (George et al., 2009; Fontana & Beckerman, 2007). For example, McCoy (2005) found
those who were socially or economically marginalized were more likely to describe issues with
providers, such as being seen too briefly, pressured to take ART meds, or not receiving
information in an intelligible way (McCoy, 2005). Dennis and colleagues’ (2015) report
previously incarcerated participants felt stigma surrounding their incarceration record and HIV
status, and this stigma contributed to negative experiences with providers and reluctance in
attending appointments and adhering to medication. These patient-provider dynamics were also
influenced by health insurance status; people who lacked health insurance reported limited
access to supportive and respectful providers. Poor patient-provider relationships may explain
some of the disparity along the spectrum of engagement for formerly incarcerated PLWH. In this

20

study, I expect those consistent engagers to report greater satisfaction with medical providers
than inconsistent engagers
Taken together, relevant literature suggests social support facilitates engagement in care
while lack of social support hinders engagement in care. Formerly incarcerated PLWH face
unique barriers to reconnecting with support systems and maintaining optimal support systems to
aid in their HIV care. Satisfactory social support and positive patient provider relationships may
ease individual-level challenges. Understanding supportive relationships and the impact they can
have on engagement in care is crucial to addressing the HIV epidemic among formerly
incarcerated individuals.
The bulk of previous work in this field has largely focused on individual challenges. The
few studies that do examine relational challenges have been qualitative in nature. These studies
suggest that the relationships formerly incarcerated inmates keep may affect their access to vital
individual-level resources such as employment, transportation, money, and housing (Dennis et
al., 2015; Haley et al., 2014). Due to the complexities of incarceration and reentry, it seems that
ex-offenders may also rely on support networks more than the average individual to re-adjust and
reintegrate back into the community after release (Visher, Mallik-Kane, 2007; Hammett, Roberts
& Kennedy, 2001). This has implications the importance of relationships in care engagement.
Given what qualitative research has found about relationships in this population, I expect
relationship factors to explain more variance in care engagement patterns than individual-level
factors.
Summary
The relevant literature suggests that formerly incarcerated PLWH may experience
challenges because of their specific needs at the time of reentry. Reentering PLWH must manage

21

co-occurring individual-level challenges that can impact their ability to consistently engage in
care. In addition, they may experience trouble reconnecting with support networks and building
relationships with HIV care providers, which can be vital assets to successful reentry and
consistent HIV care engagement. The limited literature on reentry challenges for PLWH suggests
that individual- and relational-level factors facilitate or hinder engagement in care. Most research
examines either specific individual factors or specific relationship variables. Rarely are these
examined simultaneously. In the current study, I use the socioecological framework to examine
the synergistic impact of individual and relational factors on engagement in care for formerly
incarcerated PLWH.

22

CURRENT STUDY
This project will contribute to the care engagement literature in the following ways. First,
this study adds to the literature by examining formerly incarcerated PLWH at 39 months since
release from prison. This population remains under-investigated and at high risk for poor
engagement in care. Few studies have examined engagement in care beyond the initial months
following release from prison. Second, engagement in care is dynamic. This study explored three
engagement patterns: 1) no engagement in care, 2) inconsistent engagement, and 3) consistent
engagement in care. Finally, this study adopts a socioecological perspective, which has the
potential to elucidate mechanisms that may explain the multilevel nature of factors that impact
on engagement in care.
The following hypotheses guided my investigation:
Hypothesis 1: Those who are consistently engaged in care will have significantly lower
individual-level risk scores as compared to those who are inconsistently engaged and those who
are not engaged in care.
Hypothesis 2: Those who are consistently engaged in care will report larger social
networks than those who are inconsistently engaged and those who are not engaged.
Hypothesis 3: Those who are consistently engaged in care will report greater satisfaction
with the support they receive as compared to those who are inconsistently engaged and those
who are not engaged.
Hypothesis 4: Those who are consistently engaged in care will report greater satisfaction
with providers as compared to those who are inconsistently engaged in care.
Hypothesis 5: Relationship factors will explain more variance in care engagement
patterns than individual-level factors.

23

METHODS
The purpose of this study was to improve the field’s understanding of engagement in care
and factors that influence engagement in care for formerly incarcerated individuals living with
HIV. I used mixed methods as the methodological approach to investigate this complex topic.
Mixed Methods
Rooted in pragmatism, mixed methods research uses multiple methods to make sense of
the social world (Tashakkori & Teddlie, 1998; Johnson & Onwuegbuzie, 2004). Mixed-methods
are useful when one method insufficiently captures multiple facets of a phenomenon (Creswell &
Clark, 2011). Mixed-methods research capitalizes on the strengths of multiple methods for
purposes of complementarity (Greene, 2007). Complementarity allows researchers to dissect
multiple dimensions of a phenomenon by combining different methods to provide a broader
understanding of one topic (Greene, 2007).
This project examined a tremendously complex social phenomenon—engagement in HIV
care during the community reentry period post incarceration—as it is experienced by a poorly
understood population, previously incarcerated PLWH. As such, multiple complementary
methods can deepen our understanding of reentry and engagement in HIV care, multiple
complementary methods are necessary. For example, a primary goal of this project was to gain a
multi-level understanding of the challenges PLWH face when engaging in care postincarceration. Complementary methods provided me with diverse insights into these multiple
ecological levels, which is especially important in a population that is overlooked in research
(Dennis et al., 2015; Haley et al., 2014). Little is known about previously incarcerated
individuals and even less about previously incarcerated individuals living with HIV. This
population is often difficult to incorporate into community-based research due to the transient

24

nature of their lifestyles, stigma, institutional distrust, and constant cycling in and out of
incarceration. Utilizing multiple methods allowed for the combination of complementary sources
to create a simultaneously broad and deep picture of PLWH.
Mixing two methods, such as interviews and archival data, can enhance a study by
offsetting the weaknesses from one method with the strengths of the other method; they
compensate for each other (Creswell & Clark, 2011; Patton, 2002; Marshall & Rossman, 1989).
In this project, I integrated archival data and interview data. Although archival data are
invaluable in the context of health research, they fail to tell a complete story. Archival data
provided an important record of each individual’s life (e.g., medical, criminal), while interview
data provided context to these records. Mixing methods in this way allowed me to gain a richer
understanding of a population that is underrepresented in the literature (Creswell & Clark, 2011).
Archival data. The use of archival data has become increasingly popular among
researchers as it offers economical and methodological advantages (Webb & Campbell, 2000;
Boslaugh, 2007; Vartanian, 2010). Archival data refer to documents or records that provide a
large, rich source of information (Patton, 2002; Hodder, 2000). Specifically, in health research
archival data may consist of client records from service providers that offer a unique perspective
on health and represent information that is challenging, if not impossible, to obtain through selfreport (Boslaugh, 2007; Patton, 2002). Client records are particularly useful for the proposed
project to examine variables including HIV-related health status, criminal histories, mental
health, and place of release.
Interviews. The purpose of interviewing is to capture a participant’s knowledge and
perspective (Patton, 2002). Interviews provide an opportunity to discover information about
individual’s perceptions and experiences in their own words (Patton, 2002; Oakley, 1988). In this

25

study, semi-structured in-person interviews were conducted to examine feelings, behaviors, and
experiences of people living with HIV post-incarceration.
Semi-structured interviews follow a standardized interview guide; however, the
interviewer has the freedom to probe or follow up on any topic throughout the interview (Patton,
2002; Singleton & Straights, 2010). Consequently, semi-structured interviews are ideal for
situations in which it is unlikely the researcher will get the chance to interview the population
again. One advantage of semi-structured interviews is that they allow flexibility to explore
unanticipated topics. In this project, this flexibility was especially important, as there is a dearth
of information on this population. In addition, previously incarcerated PLWH are generally very
transient and researchers must go to great lengths to locate and recruit the participants. Finally,
interviews serve as a space for researchers and participants to engage in an emotionally safe
dialogue (Campbell, Sefl, Wasco & Ahrens, 2004). Researchers have noted that qualitative
interviews provide a rare opportunity for the participant to talk freely about experiences and
receive validation (Wilson & Wilson, 1994; Campbell, et al., 2004).
AIDS Partnership Michigan’s Community Reentry Program Evaluation
This study used data from the AIDS Partnership Michigan’s (APM) Community Reentry
Program Evaluation. The APM project evaluated long-term effectiveness of the APM
Community Reentry Program. The purpose of the Community Reentry Program is to link HIVpositive people to medical case management prior to their release from prison, which should help
link clients to medical care and other AIDS-related services after release.
The evaluation consisted of telephone interviews with case managers and physicians and
in-person interviews with clients. Archival data were also obtained for the evaluation from: APM
program records, The Michigan Department of Community Health’s CAREWare database, The

26

Department of Corrections database (OTIS), The Michigan State Police database (ICHAT), the
state insurance registry, the Social Security Death Index, and The Michigan Drug Assistance
Program’s (MIDAP) records. This study analyzed a subset of these records to examine factors
that influence patterns of care engagement for formerly incarcerated people living with HIV.
Participants
The larger evaluation study examined the post-release experiences of 190 former inmates
living with HIV who used the APM community reentry program. For the purposes of the current
study, I focused only on those who completed in-depth, in-person client interviews with the
APM evaluation team (n=60). The demographics of both the original APM evaluation sample
and the sample included for the current study are displayed in Table 1. In the tables, transgender
individuals are included as male or female, based on their lived gender identity. Race was
collapsed into “Black/Mixed” and “White”. As shown, there were no significant differences
between clients who completed an interview and clients who were not interviewed.

27

Table 1.
Demographic Characteristics of APM Reentry Program Clients
Primary Study
(N=190)
n(%)
Gender
Male
Female
Race
Black/Mixed
Other
Ethnicity
Not Hispanic
Hispanic
On Parole
Yes
No
Absconded from Parole
Yes
No
History of Mental Illness
Yes
No
History of Substance Use
Yes
No
Age
Length of Incarceration

Analytic Sample
(n=60)
n(%)

Chi-Square
(X2)
.0045

184 (96.8)
6 (3.2)

58 (96.7)
2 (3.3)

157 (84.4)
29 (15.6)

53 (86.7)
7 (13.3)

.181

.6163
181 (96.2)
7 (3.7)

59 (98.3)
1 (1.7)
1.5667

23 (21.1)
86 (78.9)

11 (31.4)
24 (68.6)

9 (8.3)
100 (91.7)

1 (2.9)
34 (97.1)

1.1954

.161
60 (32.8)
123 (67.2)

18 (30.0)
42 (70)

139 (73.2)
51 (26.8)
Average (Range)
47.64 (27-68)
5.18 (0.25-30)

45 (75.0)
15 (25.0)
Average (Range)
48.17 (31-68)
6.17 (0.50 -30)

.0796

28

t-test
0.445
1.256

Sample
The primary study took a census of all people who were enrolled in the APM Community
Reentry program between May 2003 and May 2008 (N=190). This timeframe was chosen so that
each client would have had the opportunity to receive care for at least 36 months’ post release.
The analytic sample is the subset of former clients who participated in client interviews with the
APM evaluation team (n=60). To be eligible to participate in the client interviews, clients had to:
1) be located by the APM evaluation team (i.e., not incarcerated, alive, and residing in Michigan)
and, 2) agree to be interviewed by the APM evaluation team. Figure 5 depicts how the final
sample was developed.

Figure 5. Development of Analytic Sample.

29

Procedure
The APM Community Reentry Program evaluation was comprised of several steps: 1)
abstraction of client records; 2) location of former clients for in-person interviews; 3) recruitment
of former clients for in-person interviews; 4) screening of former clients for in-person
interviews; 5) in-person client interviews; and 6) receipt of customized additional data sources.
Each step will be explained in detail in the following section.
Abstracting client records. The evaluation team sent researchers to APM in Detroit,
Michigan to abstract data from the client records of those who received APM services between
the dates of May 2003 and May 2008. To identify these clients, the team searched all APM
program files for unduplicated clients enrolled during the specified time-period.
During the APM data abstraction, the evaluation team followed standardized procedures
to ensure clients’ confidentiality and the safe transportation of documents. A standard data
abstraction form was designed to aid data collection. The form was pre-printed with randomly
generated ID numbers. The first page included descriptive client information such as name,
demographics and contact information. The pages that followed included HIV status, housing,
legal, and other health information. The first page containing the identifying information was
immediately detached from the other client information and stored separately in a locked box.
All abstracted data were then transported to Michigan State University. Data were immediately
transferred to locked file cabinets and stored in locked offices with restricted access. Data were
entered into a password-protected file on a secure network. The list linking data to names was
not destroyed immediately, enabling the evaluation team to match the data to all other data they
would collect.

30

In-person interviews. After identifying which clients received APM services during the
specified period, the evaluation team sought to locate these persons for in-person interviews.
Determining incarceration status. In order to determine whether or not former clients
had been re-incarcerated, the evaluation team accessed OTIS. OTIS is a public record of offender
information including parole, probation, and current incarceration data. The database only
displays information for the offenders under the jurisdiction of the Michigan Department of
Community Health (MDOC) within the last 3 years. The evaluation team used this database to
determine which clients were re-incarcerated and therefore could not participate in the in-person
interview (n=43).
Determining mortality status. Public vital record searches were used to determine which
clients were deceased and therefore unable to participate in an in-person interview (n=32). First,
the evaluation team needed to obtain known aliases of the clients. To do so, the evaluation team
accessed the ICHAT database. This database includes felony and misdemeanor crimes
punishable by 93 days or more, parole, probation, sentencing, and known alias information. The
majority of former clients had an ICHAT Record (n=180). Mortality status was obtained from
The Social Security Death Index. Social security numbers for the former clients were selected
from the APM records. The Social Security Death Index contains data on deceased individuals in
the United States. Causes of death were confirmed through the Michigan Department of
Community Health, when possible. The Michigan Department of Community Health conducted
a search on AIDS-related deaths using the list of deceased individuals provided by the evaluation
team.
Locating current addresses. Once a list of clients who were living and not incarcerated
was established, the evaluation team took steps to locate these clients. To begin locating clients,

31

the evaluation team pulled addresses from the abstracted APM records. The records included the
address to which the client was released post-incarceration and addresses of any relatives or
friends to whom the client gave APM permission to contact.
The evaluation team sent 463 letters to former clients, friends, and relatives using the
addresses in APM records to recruit eligible clients to the in-person interview. If an address was
to a shelter or institution, no letter was sent. Letters included discretely worded information
about the study, a toll-free number to call, and a business card. One hundred and two of these
letters were returned, indicating a wrong address.
To obtain better addresses for the 102 returned letters, the evaluation team used four
other locating records: Michigan’s PSOR, MIDAP, and Michigan’s insurance database. First, the
evaluation team conducted a search on Michigan’s PSOR. This databased contained names,
demographics, and address information for individuals convicted of a sex-related crime. The
evaluation team used this database to locate clients who had been convicted of a sex-related
crime (n=12).
Since many former APM clients enrolled in MIDAP for HIV medication assistance, the
evaluation team requested the last updated address of individuals registered to receive MIDAP
benefits. This database records information for all persons who qualify to receive HIV-related
medication assistance. The evaluation team made the same request for those registered for state
insurance benefits. The evaluation team also paid to use a database that searches utility records
for the last known addresses of a client and all relatives named in their APM record. Phone calls
were also made to the last known phone number of clients and their relatives.
If a client could not be reached by the above strategies, the evaluation team contacted
parole officers (for those still on probation or parole) and requested they disseminate the letter to

32

parolees with whom they still had contact. For those clients who were still known to use APM
services, a letter was distributed to them by APM caseworkers. The evaluation team also sent out
posters to drug treatment facilitates with information about the study. Finally, as a last resort to
locate addresses, the evaluation team requested public driver’s license records through Michigan
Secretary of State Driver’s Records, if none of these other sources produced an address. Through
these methods, the evaluation team could find usable addresses for 81 of the 112 former clients
who were eligible for an interview. For clients whose letters were not returned and who did not
call the evaluation office (n=25), the evaluation team made home visits. At home visits, the
evaluation team left business cards or spoke with a resident. The evaluation team successfully
spoke with 63 of clients; 3 refused an interview.
Prior to scheduling the interviews, all former clients who agreed to participate went
through a screening process that confirmed individuals’ identity. The evaluation team prepared a
screening instrument containing questions that would allow them to be confident they were
interviewing the former client to whom they intended to speak. The screening instrument
included questions about incarceration (e.g., years, prison name, release date), demographics
(e.g., name, date of birth) and the last 4 digits of their social security number. Interviewers then
compared the answers to data abstracted from APM records to verify callers’ identities. Only
after identities were verified did the team reveal the nature of the interview.
In-person interviews were conducted and audio recorded at a place and time convenient
to the participant. The interviews lasted 50 to 210 minutes (100 minutes, on average).
Participants were given several breaks during the interview. The interview addressed experiences
with APM intake and case management services, services received at time of release, social
support, experiences with medical providers, experiences living with HIV, risk and protective

33

behaviors, employment, housing, and current health status. The interview used recall prompts
such as calendars to aid participants’ memory. Although the interview was structured,
interviewers conducted the interview as semi-structured, allowing participants to speak freely. If
the participant displayed misinformation about a topic, suggesting that they did not understand
their disease or how to access available services, an educational intervention followed the
completion of the interview. All participants were provided with referrals to additional services,
if needed. Participants were paid $50 for participating in the interviews.
Due to the sensitive nature of the data needed for this evaluation, extensive measures
were taken to protect client’s confidentiality in the collection and maintenance of all data
sources. All former APM clients signed a HIPPA waiver at intake to allow their health
information to be accessed for the purposes of a program evaluation. In addition, the entire
evaluation team signed a confidentiality agreement.
The evaluation team made efforts to minimize risk for clients participating in the inperson interview. To protect participants from signing their name to a document with the study
description, which revealed the study is examining PLWH, the evaluation team applied for and
received a waiver of documentation of consent. In addition, when providing incentives and
reimbursing for interviews and home visits, the evaluation team received an exception so that a
participant’s signature was not required and the exact address of the participant’s home was not
disclosed.
After the different data sources were obtained and linked, each participant’s random
identification code replaced their name to protect their confidentiality. Identifying information
was destroyed after the client interviews. All data were stored in locked files on password
protected computers and networks. Audio recordings were immediately removed from the digital

34

recorders and saved onto a secure network with password protection for transcription. Audio
from the client interviews was transcribed verbatim and transcripts were de-identified. Data
collection procedures for the evaluation were reviewed and approved by Michigan State
University’s Social Science and Education Institutional Review Board.
Additional data sources. Additional data sources were created by the evaluation team to
examine former clients’ health and criminal history. These data sources were used in the current
study to examine factors impacting HIV care engagement.
Michigan Department of Community Health CAREWare database. The evaluation
team requested CAREWare data from MDCH. The CAREWare database tracks Ryan Whitefunded services provision to persons living with HIV in Michigan. CAREWare is a national
software program distributed by HRSA. The database contains medical records including
lifetime history of CD4 counts, viral loads, and STI test results. MDCH created a customized
database for the evaluation team by hand-searching for CAREWare records with the name and
social security numbers of former clients abstracted from the APM. In sum, 119 former clients
had health-related data in the CAREWare database, and 70 of those clients had CAREWare data
for 3 or more years after receiving services from APM.
The Michigan State Police database. The evaluation team used ICHAT to create a
database containing client aliases, criminal histories, and other information necessary to locate
clients. The database includes felony and misdemeanor crimes punishable by 93 days or more.
The database also includes parole, probation, and sentencing information. The majority of former
clients had an ICHAT Record (n=180).

35

APM record. Data abstracted from APM client records included basic demographics,
residential status, anticipated needs, HIV health information, mental health history, substance use
history, and final case disposition (n=190).
Measures
Only measures pertinent to the current study are described.
Engagement in care. Engagement in care was measured as one variable with three
levels: never engaged in care, inconsistently engaged in care, and consistently engaged in care.
In this project, presence of a viral load or CD4 test, obtained from the CAREWare database,
served as a proxy for kept medical care visits, a common measure of retention in care (Mugavero
et al., 2013). The CAREWare database includes viral load and CD4 test information for 52 of the
60 former APM clients.
There are several widely-used measures of retention in care. The most common measures
examine missed appointment visits, appointment adherence, visit constancy, and gaps in care
(Mugavero et al., 2012; Baligh et al., 2012). Recent reviews of retention measures have
established that there is currently no gold standard for measuring retention in care with each
measure having specific advantages and limitations (Mugavero et al., 2012; Baligh et al., 2012).
Decisions to use a specific retention measure depend on the availability of data and the goals of
the research (Mugavero et al., 2012; Baligh et al., 2012). This project measured retention in care
based on the HRSA medical visits performance measure, which considers patients retained in
care if they attend two or more visits with their medical care provider in a 12-month period with
greater than 3 months between visits (Mugavero et al., 2012; Hall et al., 2012; HRSA “HAB
Core Clinical Performance Measures for Adult/Adolescent Clients”, 2008).

36

To calculate how and when former clients received CD4 or viral load tests, I took the
APM enroll date and determined 3-year period following their APM enrollment. To do this, I
calculated 12-month intervals for each participant who had a record of a CD4 or viral load test
during the 3-year period. Any tests that did not fall into the designated time-period were filtered
out. Therefore, if a client had two or more CD4 or viral load tests in every 12-month period with
3 or more months between each test, they were classified as “consistently engaged in care”. If a
client had no CD4 or viral load tests within 3 years of his APM enroll date, the client was
classified as “not engaged in care”. All other clients were classified as “inconsistently engaged in
care”.
Following the HRSA definition exactly, only 10% of the sample (6 respondents) would
have been considered consistently engaged in care, which is extremely low compared to previous
literature. Therefore, given that the HRSA performance measure is typically meant to only
measure 1-year intervals of retention and has been critiqued for being too rigid in certain
samples, I chose to slightly adapt the measure for this project (Mugavero et al., 2012; Baligh et
al., 2012). Rather than looking at exactly 3 years, I chose to add an extra month to each 12month interval and examine a 39-month period (3 years and 3 months rather than 3 years).
Therefore, the revised definition is as follows: a patient is considered “consistently engaged in
care” if they attend two or more visits with their medical care provider in every 13-month period
with greater than 3 months between visits (Mugavero et al., 2012; Hall et al., 2012, (HRSA
“HAB Core Clinical Performance Measures for Adult/Adolescent Clients”, 2008). Nonengagement and inconsistent engagement definitions stayed the same, such that, if a client had
no CD4 or viral load tests within 39 months of his APM enroll date, the client was classified as
“not engaged in care” and all other clients were classified as “inconsistently engaged in care”.

37

Missing data. Some clients did not have a CAREWare record (n=8). If a client’s
CAREWare record was missing, this could be a result of the client receiving care from a private
physician, a physician at the Veteran’s Administration, or another non-Ryan White supported
provider who failed to record test results in the CAREWare database, as legally required.
Alternatively, CD4 tests and viral load counts could be truly missing. For those who were
missing a CAREWare record entirely, I examined the in-person client interview data and
transcripts to determine if the data were truly missing because the person was not engaged in
care or if they were in care.
The majority of these clients reported receiving viral load and CD4 tests during the inperson interview. However, the test they reported in the interview occurred after the 39-month
period post-release window of interest and after the creation of the CAREWare database (late
June 2011). Given this, I had to assume they were not in care during the period captured in the
CAREWare database because there is no evidence they received a test prior to June 2011. If, for
instance, they reported a test in the in-person interview that fell into the CAREWare database
period, we would have evidence that the CAREWare database was inaccurate for that person.
However, this was not the case for all 8 clients. Therefore, all 8 clients without a CAREWare
record were classified as not engaged in care.
Sexual orientation. Sexual orientation was obtained from the in-person interview.
Sexual orientation was collapsed into two categories for analysis, heterosexual and gay, lesbian
or bisexual. The majority of the sample was heterosexual (65%).
Social support. The Arizona Social Support Interview Schedule (Barrera, 1980, 1981)
was used to gather information about people in participants’ support network. This measure has

38

shown to be internally consistent in multiple studies (a = .70 - .78) (Barrera, 1980, 1985;
Hernandez-Plaza, Pozo & Alonso-Morillego, 2004).
The 25-item measure collects information about five different types of social support:
intimate support, physical support, material support, social recreational support, and sexual
support. Intimate support pertains to support and advice about personal or private feelings.
Physical support pertains to physical acts of support, such as going to the store for someone.
Material support pertains to tangible belongings, such loaning money. Social recreational support
pertains to people to hang out with. Sexual support pertains to support for sexual health, such as
someone you ask about condoms. For each of these support types, the measure collects
information about the number of people respondents received support from and satisfaction with
support. Two scores were calculated from this measure: support network size and satisfaction
with support received.
Social support network size count. This subscale collects information on the number of
members in the support network respondents used for intimate support, physical support,
material support, social recreational support, and sexual support. For the purposes of this study, a
continuous variable was created based on a count of the unique number of people respondents
named across all types of support. Number of members in social support network ranged from 0
to 15 with a mean of 6.7 (SD = 3.9).
Support satisfaction score. This subscale collects information on satisfaction with
received intimate support, physical support, material support, social recreational support, and
sexual support. Response choices range from 1 – “very dissatisfied” to 4 – “very satisfied”. I
summed individual responses to create a single score, which can range from 4-20. Lower scores

39

indicate dissatisfaction. Satisfaction with social support ranged from 4 to 20 with a mean of 13.4
(SD = 4.5).
Patient-provider relationship. To assess satisfaction with patient-provider relationship,
Bakken et al.’s (2000) Engagement with Health Care Provider scale was used. The scale is a 13item measure of patient-provider relationship quality. An example item was: “My healthcare
provider respects me.” Response choices range from 0 – “never” to 3 – “always”. I summed
responses to create a summary score. Summed scores range from 0-39. Higher scores indicate
greater provider relationship satisfaction. The mean score for this sample was 32.61 (SD=8.7).
Cronbach’s alpha for this sample was 0.96.
Individual-level risk index. To assess individual-level factors that may influence care
engagement, a composite index was constructed. It is difficult to accurately measure a complex
concept with a single indicator. Rarely are we able to capture all facets of a concept. To account
for this, I created an index of individual-level risk factors by combining several indicators into
one composite measure to capture the cumulative impact of individual-level risk. The composite
index was calculated following Trochim & Donnelly’s (2001) steps for constructing an index.
First, I used recent literature to conceptualize the components of the individual-level
index. Individual-level factors that emerged as influential to care engagement included race,
poverty, current employment status, transportation access, history of substance use, history of
mental illness, incarceration history, educational attainment, and history of housing instability
(Cargill, 2012; Traeger, et al., 2012; Mugavero et al., 2009, 2007; Rajabiun et al., 2007; MalikKane & Visher, 2008; Petersilla, 2003, Rose & Clear, 2002; Dennis, et al., 2015; Haley et al.,
2014). Next, I operationalized each of these variables to create a composite individual-level risk

40

score. I created each variable as a dichotomy so that items would be equally weighted. I coded
each variable as 1 or 2, with higher scores indicating higher individual-level risk.
Educational attainment. Clients were asked to identify their level of educational
attainment as graduate or professional degree, bachelor’s degree, associate’s degree, some
college, high school graduate/GED and less than high school. Educational attainment was coded
as: (1) some college, associate’s degree, bachelor’s degree or graduate degree (n=27), or (2) less
than high school, high school or GED (n=33). Although there is no literature to support this
specific cutoff, relevant literature supports lower educational attainment may be related to poor
engagement in HIV care (Kalichman, Catz, Ramachandran, 1999; Wolf et al., 2007; Cargill,
2012).
Race. Race was obtained from the APM client record. Race was collected as:
Black/African American, White, Asian/Pacific Islander, Arab/Chaldean, Indian/Native American
and other. For the analysis, race was coded as: (1) White (n=7), or (2) Black/Mixed (n=53).
Relevant literature supports racial minority status, and Black specifically, may be related to poor
engagement in HIV care (Traeger, et al., 2011; Mugavero et al., 2009, 2007; CDC, 2012;
Rebeiro, et al., 2013; Ulett, et al., 2009; Hall, et al., 2012; Torian et al., 2008; Hall, et al., 2013).
Income. Clients were asked to report their current monthly household income and how
many people that income supports. In the analytic sample, income ranged from $0-$5200 with a
mean of $1031.92 (SD = 939.86). The majority of clients only supported themselves (n=44); 9
clients supported two people, 2 clients supported three people, and 2 clients supported five
people. Income was coded as: (1) current income is above the federal poverty level (n=21), or (2)
current income is below the federal poverty level (n=38). I accounted for the number of people
the income supports in the household when assigning these codes. Federal poverty level

41

information was obtained from the DHHS measure of income level in 2011. Relevant literature
suggests lower income may be related to poor engagement in HIV care (Rajabiun et al., 2007;
Rebeiro, et al., 2013; Ulett, et al., 2009; Torian et al., 2008; Giodano, et al., 2003; Petersilla,
2003, Clear et al., 2001).
Substance use history. Substance use history was obtained from the APM client record
and was coded such that: (1) APM client record indicated no history of substance use (n=15), or
(2) APM client record indicated a history of substance use (n=45). Relevant literature suggests
substance use may be related to poor engagement in HIV care (Dennis et al., 2015; Haley, et al.,
2014; Malik-Kane & Visher, 2008; Ulett, et al., 2009; Torian et al., 2008; Giodano, et al., 2003).
Mental health history. Mental health history was obtained from the APM client record
and was coded such that: (1) APM client record indicates no history of mental illness (n=42), or
(2) APM client record indicates a history of mental illness (n=18). Relevant literature supports
mental illness may be related to poor engagement in HIV care (Malik-Kane & Visher, 2008;
Rajabiun et al., 2007; 2013; Scheyett et al., 2010; Traeger, et al., 2012; Cashman et al, 2004).
Employment. Clients were asked to report their current employment status during the
client interview: employed full time (>34 hours per week), employed part-time, sporadically
employed. Employment was coded as: (1) currently employed (n=11), or (2) currently
unemployed (n=49). Relevant literature suggests unemployment may be related to poor
engagement in HIV care (Cargill, 2012; Rajabuin et al., 2007; Dennis, et al., 2015; Malik-Kane
& Visher, 2008).
Housing stability and satisfaction. To assess housing stability, clients were asked
whether or not they had been homeless since their release. Homelessness was defined as not
having your own house or apartment, living on the streets or in a shelter, or staying with friends

42

or family in their apartment without paying rent. For the proposed analysis, homelessness was
coded as: (1) never homeless since release (n=21), or (2) ever homeless since release (n=39).
Clients were asked how satisfied they were with their current housing situation during the
client interview. Response options included extremely satisfied, satisfied, mixed, dissatisfied,
and extremely dissatisfied. Current housing satisfaction was collapsed so it was dichotomous: (1)
extremely satisfied or satisfied (n=32) or (2) extremely dissatisfied, mixed, or dissatisfied
(n=28). The “mixed” responses (n=14) were classified as dissatisfied because mixed satisfaction
implies some dissatisfaction with their housing. Relevant literature suggests housing instability
and homelessness may be related to poor engagement in HIV care (Dennis et al., 2015; Haley et
al., 2014; Malik-Kane & Visher et al., 2008; Brinkely-Rubenstien & Turner, 2013).
Transportation. Clients were asked if they had access to reliable transportation (e.g., car,
friend’s car, bus). Access to transportation was coded as (1) reliable access to transportation
(n=45) or (2) no access to reliable transportation (n=15). Relevant literature suggests access to
transportation may be related to poor engagement in HIV care (Buchberg et al., 2015; Dennis,
2015).
Scoring individual-level risk index. To create a single composite risk score, each variable
in the index was unit weighted, as these are as accurate as empirical (optimal) weights (Wainer,
1978). Scores on each of the variables were summed. Scores could range from 9-18, with lower
scores indicating lower individual-level risk and higher scores indicating higher individual-level
risk. In the analytic sample, scores ranged from 10-15 with a mean of 12.43 (SD = 1.6). The
index was normally distributed. (See Appendix A for a description of the validation of the
individual-level risk index.)

43

Analysis
Prior to data analysis, I ensured the data were clean. I randomly checked the paper data
against the electronic data by generating a randomly ordered participant ID number list. I
checked every fifth file. Next, I calculated frequency distributions and descriptive statistics
(means, standard deviations, and ranges) for all pertinent variables to check for outliers and
missing data. In addition, I computed a correlation matrix of all of the individual-level factors.
Outliers were examined to assess whether data entry errors or failure to specify missing values
accounted for missing data. There were no outliers in the main variables used for hypothesis
testing. Given this, no transformations were needed to deal with outliers. As a result of the size
of the analytic sample, I used egan row total imputation rather than remove missing cases to
calculate the sum scores for the satisfaction with social support variable (Tabachnick & Fidell,
2013). This approach is appropriate for this type of data because the data was missing truly at
random and because it allowed me to retain all cases for analysis. This imputation technique was
only necessary for two cases.
To address hypotheses one to five, I conducted bivariate analyses, followed by a
multinomial logistic regressions. I used a nested sequential multinomial logistic regression to test
hypothesis 5. Logistic regression was an appropriate for these data because it is a flexible
analytic tool. It does not assume normally distributed data, linearity or equal variance among
groups (Tabachnick & Fidell, 2013). Furthermore, it has the capacity to mix several types of
predictors and outcomes, which is advantageous for this dataset (Tabachnick & Fidell, 2013).
First, I conducted bivariate analyses (t-test, ANOVA, Chi-Square) to test the relationships
between each predictor and the dependent variable, care engagement status. Subsequently, I
conducted a multinomial logistic regression with relative risk ratios to predict membership in the

44

three care engagement categories (consistently engaged in care, inconsistently engaged in care,
and not engaged in care) in three blocks (Tabachnick & Fidell, 2013). In the first block, I entered
the control variable (sexual orientation). In the second block, I added the individual-level risk
score. In the third block, I included network size and satisfaction with social support. I assessed
model adequacy or goodness of model fit by comparing Akaike’s (AID) and Bayesian (BIC)
information criteria between models. Models with smaller AIC and BIC indicate better fit
(Tabachnick & Fidell, 2013).
To assess the role of patient-provider relationships on engagement, I repeated this
analysis for those who were consistently and inconsistently engaged in care, regressing
engagement status first on sexual orientation, followed by the composite individual-level risk
score, and then by social support, network size, and provider satisfaction. All bivariate and
multivariate analyses were conducted in Stata’s Data Analysis and Statistical Software (Stata
Corp, 2015).

45

RESULTS
I will begin this section by describing the characteristics of the entire sample. Then, I
will describe characteristics for each of the three care engagement groups. Finally, I will present
the results of hypothesis testing. Due to the small sample size, I report conventionally nonsignificant results as findings trending toward significance, as these marginally significant effects
may have been significant given a larger sample size (Turner, Miller & Henderson, 2008). For
the purposes of this project, I considered p < .05 as significant and p < .10 as trending toward
significant.
Demographic Characteristics
The sample included 60 formerly incarcerated individuals who received APM services
between May 2003 and May 2008. The majority of the sample was Black (86.7%), non-Hispanic
(98.3%), and currently identified as male (96.7%). Respondents ranged in age from 31 to 68
(M=48.17, SD=7.41). Most of the respondents identified as straight or heterosexual (65.0%).
Respondents faced a significant number of individual-level challenges. Over half of the
sample had a high school education or less (55.0%) and nearly all were unemployed at the time
of research (81.7%). The average monthly income ranged from $0 - $5,000 (Median=$874.0).
Many respondents reported lifetime histories of substance use (75.0%) and homelessness
(65.0%). A minority of respondents reported a history of mental illness (30.0%). Length of time
spent incarcerated ranged from 6 months to 30 years (M=6.17, SD= 5.54).
For the purposes of this research, participants were classified as not engaged in care,
inconsistently engaged in care, or consistently engaged in care. The most common care
engagement pattern was inconsistent engagement, accounting for 50.0% of respondents (n=30).
The least common was consistent engagement, accounting for 17.0% of respondents (n=10).
Approximately 33.0% were not engaged in care (n=20). To compare demographic characteristics
46

across the three care engagement groups, I used chi-square likelihood ratio tests, t-tests, and
ANOVA, depending on the nature of the variable being reviewed. Only those variables with
sufficient variance to perform statistical tests were assessed. Examining demographics among the
three groups revealed differences on sexual identity (p = .03) and current income (p = .01).
Individuals who identified as gay or bisexual were more likely to be members of the consistent
care group than were individuals who identified as straight or heterosexual. The consistent care
engagement group had the largest monthly income and the non-engagement group had the
smallest monthly income (Table 2). No other significant differences emerged on demographics
among the three engagement groups.

47

Table 2.
Demographic Characteristics by Care Engagement (N=60)
Not Engaged
in Care
(n=20)
n
%
Gender
Male
Female
Race
Black
Other
Sexual
Orientation
Straight
Gay
Hispanic
No
Yes

Inconsistently
Engaged in
Care
(n=30)
n
%

Consistently
Engaged in
Care
(n=10)
n
%

Total
(N=60)
n
%

10
0

58
2

-19
1

95.0
5.0

29
1

96.7
3.33

100.0
0

Current
Monthly
Income (in
dollars)

96.7
3.3
--

19
1

95.0
5.0

25
5

83.3
16.7

9
1

90.0
10.0

53
7

88.3
11.7

13
7

65.0
35.0

23
7

76.7
23.3

3
7

30.0
70.0

39
21

65.0
35.0

7.2*

-20
0
M

100.0
0
Range

30
0
M

100
0
Range

9
1
M

90.0
10.0
Range

46.9

31 - 58

48.6

33 - 63

49.5

42 - 68

5.1

0.8 15.0

6.2

0.5 –
30.0
Range

8.1

2.0 –
15.0
Range

Med

Range

Med

716.
4

(02,400)

998.9

Age
Length of
Incarceration
(in years)

χ2

Med

(03,200)

48

1,758

59
1
M
48.1
7

5.9

98.3
1.7
Range

F

31 - 68

0.5

0.50 –
15.0
Range

1.2

Med

(05,200)

874.
0

F

(05,200)

4.7*

Descriptive Characteristics
To understand the association between individual and relational indictors on care
engagement, I conducted one-way between subject’s ANOVAs. When necessary, Bonferonni
post-hoc tests were conducted to compare engagement statuses against each other (Tabachnick &
Fidell, 2013).
Indicators of individual-level risk. In Table 3, I display the individual-level risk score
and its corresponding indicators by care engagement status. As shown, there were significant
differences in substance use history by care engagement status (F (2,59) = 8.07, p = .000). Posthoc tests revealed the consistent engagers were significantly less likely to possess substance use
histories than non-engagers (p = .005) and inconsistent engagers (p = .001).
The effect of the total individual-level index score on care engagement status trended
toward significance [F(2,59) = 2.8, p = .07]. Post-hoc tests revealed a trend toward significant
mean differences on the individual-level risk score between the non-engaged group and the
consistently engaged group; the consistent engagers had lower individual-level risk scores than
the non-engaged group (p = .077). There were no significant differences between the nonengaged and the inconsistently engaged group (p = n.s.).

49

Table 3.
Indicators of Individual-level Risk Factors by Care Engagement (N=60)
Not
Engaged
(n=20)
n
%
Educational Attainment
High School, GED or
less
Associates or more
Below Poverty Level
Yes
No
Substance Abuse History
Yes
No
Mental Illness History
Yes
No
Currently Employed
Yes
No

Inconsistently Consistently
Engaged
Engaged
(n=30)
(n=10)
n
%
n
%

Total
(N=60)
n
%

12
8

60.0
40.0

16
14

53.3
45.7

5
5

50.0
50.0

33
27

55.0
45.0

14
6

70.0
30.0

21
9

70.0
30.0

4
5

40.0
60.0

39
21

65.0
35.0

F
0.16

1.7

8.1*
16
4

80.0
20.0

26
4

86.7
13.3

3
7

30.0
70.0

45
15

75.0
25.0

5
15

25.0
75.0

8
22

26.7
73.3

5
5

50.0
50.0

18
42

30.0
70.0

3

15.0
85.0

5

16.7
83.3

3

30.0
70.0

11
49

18.3
81.7

60.0
40.0

5
5

50.0
50.0

39
21

65.0
35.0

1.1

0.5
17

Ever Homeless
Yes
16
No
4
Satisfied with Housing
Yes
11
No
9
Access to Transportation
Yes
13
No
7
M
Total Individual-level
Index Score (10-15)
12.80
Note. * signifies p < .05

25

7

1.7
80.0
20.0

18
12

0.8
55.0
45.0

14
16

46.7
53.3

7
3

70.0
30.0

32
28

53.3
46.7
0.8

65.0
35.0
SD

24
6
M

80.0
20.0
SD

8
2
M

80.0
20.0
SD

45
15
M

75.0
20.0
SD

F

1.6

12.53

1.5

11.40

1.7

12.43

1.6

2.8*

49

Indicators of relationship quality. In Table 4, I display the indicators of relationship
quality by care engagement status. I examined network size, relationship satisfaction, and
patient-provider relationship satisfaction.
Network size. Network size for this sample ranged from 0 to 15 members (M=6.7,
SD=3.9). Although the non-engagers reported the smallest social network, compared with the
other two groups (M=5.8, SD=3.5) and the inconsistent and consistent engagers reported the
larger social networks (inconsistent - M=7.2, SD=4.0; consistent - M=7.0, SD=4.6), the mean
network size difference was not statistically significant (F (2, 59) = 0.77, p = n.s.).
Satisfaction with social support. Respondents reports of satisfaction with the social
support they received ranged from four to 20 on a 20-point scale (high scores indicate being
satisfied). Although the non-engagers were least satisfied with the social support they received
(M = 12.6, SD = 5.7) and the consistent (M = 13.1, SD = 5.3) and inconsistent engagers (M =
13.9, SD = 3.3) reported higher total satisfaction with social support, the mean differences for
satisfaction with social support between groups was not statistically significant (F (2, 59) = 0.56,
p = n.s.).
Satisfaction with medical provider. To examine satisfaction with the patient-provider
relationship, I compared those who were consistently engaged in care with those who were
inconsistently engaged in care. The consistent engagement group reported greater satisfaction (M
= 37.5, SD = 7.4) with their patient-provider relationship than the inconsistent engagers (M =
34.2, SD = 11.1). T-tests revealed that the mean difference between the consistently engaged and
inconsistently engaged groups on patient-provider satisfaction trended toward significance (t (0,
40) = -0.87, p = 0.1).

50

Table 4.
Indicators of Relationship of Quality by Care Engagement (N=60)
Not Engaged in
Care
(n=20)
M
SD
Network Size
Avg # of People in Network
Network Satisfaction
Network Satisfaction w/ Support
Provider Relationship
Satisfaction w/ Medical Provider

Inconsistently
Engaged in
Care (n=30)
M
SD

Consistently
Engaged in
Care (n=10)
M
SD

Total
(N=60)
M
SD

F/t

5.8

3.5

7.2

4.0

7.0

4.6

6.7

3.9

0.77

12.6

5.7

14.0

3.3

13.1

5.3

13.4

4.5

0.56

--

--

34.20

11.1

37.5

7.4

35.9

12.4

-0.87*

Note. *signifies p < .05

51

Hypothesis Testing
As shown in Table 5, I tested all hypotheses using multinomial logistic regression
models. I used sexual orientation as a covariate due to its’ systematic association with
engagement status (X2 (2) = 7.18, p = .03). The consistently engaged group served as the
reference group for hypotheses 1-4.
Hypothesis 1. My first hypothesis was that those with low individual-level risk scores
would be most likely to engage in care consistently. This hypothesis was supported. A one-unit
increase in individual-level risk was associated with a 1.90 increase in the relative log-odds of
being non-engaged versus consistently engaged. Similarly, a one-unit increase in individual-level
risk was associated with a 1.7 increase in the relative log odds of being inconsistently engaged
versus consistently engaged. In other words, lower individual-level risk predicted classification
as a consistent engager (X2 (4) = 12.70, p = .013).
Hypothesis 2. My second hypothesis was that those who reported a greater number of
people in their social network would be most likely to engage in care consistently. This model
trended toward statistical significance, indicating partial support for the hypotheses. A oneperson increase in network size was associated with a .95 decrease in the log-odds of being
unengaged versus consistently engaged. Similarly, a one-person increase in network size was
associated with a 1.07 increase in the log-odds of being consistently engaged versus
inconsistently engaged. In other words, these data show a trend toward finding a larger social
network was associated with classification as a consistent engager and an inconsistent engager as
compared to a non-engager (X2 (4) = 9.10, p = .059).
Hypothesis 3. My third hypothesis was that those who reported higher levels of
satisfaction with the support they received would be most likely to engage in care consistently.

52

This hypothesis was partially supported by the analyses, as the model trended toward statistical
significance. A one-unit increase in satisfaction with support received was associated with a 0.98
decrease in the relative log-odds of being non-engaged versus consistently engaged. Similarly, a
one-unit increase in satisfaction with support received was associated with a 1.05 increase in the
relative log-odds of being consistently engaged versus inconsistently engaged. In other words,
greater satisfaction with support predicted classification as a consistent engager and an
inconsistent engager, as compared to non-engagers (X2 (4) = 8.26, p = .083).
Hypothesis 4. My fourth hypothesis was that consistent engagers would report higher
levels of satisfaction with their medical provider than inconsistent engagers. This hypothesis was
supported. A one-unit increase in satisfaction with one’s medical provider was associated with a
0.95 decrease in the relative log-odds of being inconsistently engaged versus consistently
engaged. In other words, satisfaction with medical provider predicted classification as a
consistent engager (X2 (2) = 8.40, p = .015).
Hypothesis 5. My final hypothesis was that relational-level factors would explain more
variation in engagement in care than individual-level factors. This hypothesis was not supported.
As shown in Table 5, adding relational-level factors did not significantly improve model fit once
individual-level risk was taken into account. The final models did explain more variance overall
and were statistically significant when compared to the model included only sexual orientation
and individual-level risk, however the parsimonious model including only the covariate and the
individual-level factors best predicted care engagement.
The second regression model included only two engagement patterns as outcomes
(inconsistent and consistent). These results were similar to the first model. As shown in Table 6,
adding in relational-level factors, including patient-provider relationship, did not significantly

53

improve model fit once individual-level risk was considered. This model did explain more
variance than the other model and was statistically significant, however the parsimonious model
including only the covariate and the individual-level factors best predicted care engagement.

54

Table 5.
Model 1: Hypothesis Testing with Three Outcomes
Block 1
Not Engaged
Sexual
orientation
Individual-level
Risk
Network Size
Network
Satisfaction
Consistently
Engaged
Sexual
Orientation
Index Score
Network Size
Network
Satisfaction

Block 2

Block 3

RRR

SE

P

CI

RRR

SE

P

CI

RRR

SE

P

CI

0.57

0.36

0.37

.16-1.97

0.57

0.36

0.38

0.16-2.0

0.47

0.32

0.27

0.13-1.78

-

-

-

-

1.11
-

0.21
-

0.58
-

0.77-1.61
-

1.07
0.91

0.21
0.86

0.74
0.29

0.72-1.57
0.75-1.09

-

-

-

-

-

-

-

-

0.97

0.07

0.67

0.83-1.13

0.13
-

0.11
-

0.01
-

0.02-0.64
-

0.12
0.58
-

0.10
0.17
-

0.01
0.06
-

0.02-0.62
0.33-1.02
-

0.10
0.53
0.96

0.91
0.16
0.12

0.01
0.04
0.75

0.02-0.59
0.29-0.97
0.76-1.22

-

-

-

-

-

-

-

-

0.91

0.98

0.36

0.732-1.12

0.05
Pseudo R2
2
0.03
X
-57.2
LL
0.031
LR Test
122.4
AIC
130.8
BIC
Note. Reference group is “inconsistently engaged in care”

0.10
0.01
-54.3
0.057
120.7
133.2

55

0.13
0.05
-52.9
0.564
125.7
146.6

Table 6.
Model 2: Hypothesis Testing with Two Outcomes
Block 1
Consistently
Engaged
Sexual
Orientation
Individual-level
Risk
Network Size
Network
Satisfaction
Satisfaction
with Medical
Provider
Pseudo R2
X2
LL
LR Test
AIC
BIC

Block 2

Block 3

RRR

SE

P

CI

RRR

SE

P

CI

RRR

SE

P

CI

0.13

0.11

0.01

.03-0.64

0.12

0.11

0.02

0.02-0.71

0.12

0.11

0.02

0.02-0.70

-

-

-

-

0.61
-

0.18
-

0.08
-

0.35-1.07
-

0.56
1.00

0.19
0.13

0.08
0.97

0.30-1.07
0.78-1.30

-

-

-

-

-

-

-

-

0.87

0.06

0.32

0.67-1.14

-

-

0.16
0.008
-19.0
0.008
42.0
45.4

-

-

-

0.23
0.006
-17.4
0.072
40.8
45.8

-

1.03

0.56

0.49
0.28
0.03
-16.3
0.540
44.6
54.7

0.93-1.15

Notes. Reference group is inconsistently engaged in care

56

DISCUSSION
Understanding engagement in HIV care among at-risk populations is crucial to improving
health outcomes for these vulnerable groups, particularly given the national shift toward
treatment-focused efforts to address the HIV epidemic. PLWH are especially vulnerable to
falling out of care in the years immediately after release from prison, and this disengagement
may have fatal consequences (Mugavero, 2010; Baillargeon et al., 2009; Dennis, et al., 2015).
This study identified a number of factors that may be particularly influential on engagement in
care among formerly incarcerated PLWH over a 39-month community reentry period.
Patterns of Care Engagement
Findings indicated that most respondents failed to engage in care adequately following
their release from prison. The federal government defines engagement in care as being linked to
a HIV care provider and attending two visits per year, at least 3 months apart (HRSA HAB,
Mugavero, 2010). In this study, I classified respondents as consistently engaged in care if they
received two CD4 or viral load tests within every 13-month period, with at least 3 months
between tests. According to this definition, only 17% of respondents in this sample were engaged
in care over the 39 months immediately following incarceration. This rate is roughly half of what
one might expect to see in the general population of PLWH and is substantially less than
previous studies of formerly incarcerated individuals (Iroh et al., 2015; Stephenson, 2005; Palpeu
et al., 2004; Wohl et al., 2011; Baillargeon et al., 2009).
There are several possible explanations for why I found higher rates of care
disengagement in this study. First, this finding may be a result of the dearth of research in this
area that follows formerly incarcerated PLWH over years. The current study benefitted from
examining substantially longer follow-up periods than previous investigations, increasing my
ability to identify disengagement over time, as the longer the period under investigation, the
57

greater the possibility of missed medical appointments and tests. In other words, because
individuals have continual opportunities to fall out of care, longer observation periods may be
necessary to fully capture all possible trajectories of disengagement. The current study may
provide a more accurate reflection of care engagement in the years following release from
prison. Existing studies, due to their short follow-up periods, may actually underestimate the rate
at which formerly incarcerated PLWH fall out of care because they only examine behavior over
a few months. Considering the challenges of engaging in care last a lifetime and may be uniquely
difficult for formerly incarcerated PLWH, future research should examine engagement
longitudinally to characterize how rates of engagement vary over time accurately.
Alternatively, the lower rates of consistent care engagement in this study may be a result
of other unexamined factors. For example, in Iroh and colleagues’ (2015) review of engagement
in care for incarcerated persons, they suggest that rates of engagement may vary based on
location. Depending on the city, state, or prison, there may be more resources or programming
devoted to helping people engage in care (Iroh et al., 2015). Of all the studies included in Iroh
et al.’s, review (2015) none examined formerly incarcerated PLWH in the state of Michigan.
Therefore, it may be that community, institutional, and legislative factors in the state of
Michigan contribute to lower rates of care engagement. Future research should build upon this
project and continue exploring factors that influence different patterns of engagement in light of
state policies and programs.
In this study, I identified three distinct care engagement patterns – non-engagers,
inconsistent engagers, and consistent engagers. The consistent engagers were people who
attended a care appointment approximately twice a year for 39 consecutive months with visits at
least 3 months apart. Non-engagers did not attend any appointment for 39 consecutive months.

58

Inconsistent engagers started care long after their release, started and stopped care and then
restarted again, or started care and then stopped entirely over a period of 39 consecutive months.
Inconsistent engagement was the most common pattern and the most complex. The CoC suggests
that movement along the continuum is a simple, linear process (Mugavero, et al., 2013; Eldred &
Maltiz, 2007). People test, find a doctor, visit their doctors consistently, take their medications
and become virally suppressed. The predominance of inconsistent engagers in the sample
suggests that some individuals’ movement through the continuum may be more complex than the
CoC implies. (Figure 1; Gardner, et al., 2011; Cheever, 2007). This complex movement, perhaps
best characterized as sporadic engagement, may be overlooked in current approaches to
measuring engagement because engagement is often treated as a simple dichotomy: engaged, not
engaged. Simplifying engagement in this way fails to capture the nuances of inconsistent
engagement. Furthermore, dichotomous measurement assumes inconsistent engagers never
interact with physicians or other systems of care, which is not always the case. Findings from
this study suggested inconsistent engagers did interact with physicians and had distinct needs
from non-engagers and consistent engagers. If inconsistent engagers continue to be overlooked
in typical engagement measures we are not only losing important information about engagement
rates, but also failing to capture interactions with health care systems, sporadic or otherwise, that
may be useful for intervention. Future research should continue to explore the optimal
measurement of care engagement and approaches to avoid obscuring nuanced patterns of
interaction with medical care.
Factors that Influenced Care Engagement
Engagement in care for individuals reentering the community post-incarceration remains
tremendously difficult (Brinkley-Rubinstein & Turner, 2013; Iroh, Mayo & Nijhawan, 2015;

59

Hall, et al., 2013). In this study, respondents experienced co-occurring poverty, comorbid health
disorders, unemployment, unstable housing, inconsistent access to transportation, and lack of
education. Consistent with the bulk of literature in this area, I found that these individual-level
challenges burdened PLWH during reentry and inhibited their ability to engage in care. As the
presence of individual-level risks increased, the odds of care engagement decreased. Individuallevel factors were the best predictors of engagement in care. Although many studies have
established that PLWH and ex-offenders face competing individual demands during reentry, only
a handful of studies have examined the cumulative effects of these demands (Dennis et al., 2015;
Haley et al., 2014). In this study, I found that the additive effects of histories of mental health,
substance use, housing instability, current poverty, lack of education, unemployment, and
inconsistent access to transportation predicted poor care engagement. These results support the
National HIV/AIDS Strategy of addressing the basic needs of PLWH as a means of addressing
the HIV epidemic. This study’s results suggest that addressing the basic needs of recently
released PLWH may be an effective way to bolster their ability to engage in care.
Of all the individual-level risk indicators examined, two seemed to drive the relationship
between individual-level risk and lack of engagement: substance use and poverty. In bivariate
analyses, I observed living below the poverty line was associated with not engaging in care. Prior
studies on the role of economics in care engagement have similarly found that, even with
entitlements, people who live below the poverty line tend to struggle with retention in care
(Dennis, et al., 2015; Haley et al., 2014). This study adds to the growing body of evidence that
entitlements appear insufficient to meet the needs to support routine engagement in care for
people living below the poverty line.

60

Second, in bivariate analyses, respondents who were not engaged in care were more
likely to have a history of substance use disorders than consistent engagers. This finding is
consistent with other evidence that numerous PLWH leave prison with substance conditions,
many of which can impact engagement in care (Ulett, et al., 2009; Traeger et al., 2012; Rajabiun,
2007; Stephenson, 2005; Mallik-Kane & Visher, 2008; Rebeiro, 2013). Scholars speculate that
substance users may have more difficulty managing and attaining financial resources in ways
that undermine their ability to engage consistently (Malta, Strathdee, Magnanini & Bastos, 2008;
Gonzalez, Barinas & O’Clerigh, 2011). Furthermore, as is consistent with this study, researchers
have found that substance users may experience comorbidities that disrupt healthcare access and
treatment outcomes and vulnerability to poor treatment in the medical system (Krishnan,
Wickersham, Chitsaz, Springer, Jordan, Zaller & Altice, 2013; Springer, Spaulding, Meyer,
2011). In a qualitative focus group, Dennis and colleagues (2015) found that many individuals
reentering simply wanted to celebrate their homecoming and returned to an environment
occupied by alcohol and drugs. These individuals chose not to adhere to ART in the presence of
their “partying” peers for fear of stigma and inadvertent disclosure of their serostatus. Injection
drug use, in particular has been associated with less timely linkage to care (Torian, 2008) and
poor retention (Rebeiro, 2013, Ulett, et al., 2009), including missed appointments and nonadherence to ART (Giordano, 2003). Additionally, those who used substances prior to
incarceration are more likely to use again upon reentry (Mallik-Kane & Visher, 2008). Findings
from the current study highlight the continuing need to address substance use disorders after
release from prison.
Researchers have suggested that social support and high-quality patient provider
relationships are uniquely relevant in the lives of formerly incarcerated PLWH, however,

61

scholars are just beginning to explore the role of relationships in supporting engagement in care
for reentering community life post-incarceration (Dennis, et al., 2015; Pettus-Davis, et al., 2009).
In these data, the only relational factor that clearly distinguished between care engagement
patterns was satisfaction with the medical provider relationship. Although it is possible that
relational factors could have been significant in a better-powered study, the relational items that
were included in the final model (e.g., network size, satisfaction with social support) failed to
distinguish among engagement patterns. In this study, consistent engagers were significantly
more satisfied with their providers than those who did not consistently engage in care. Given the
stark differences between consistent engagers and the inconsistent engagers in their histories of
substance abuse and poverty, these data raise the question of whether physician biases affect the
quality of relationships between formerly incarcerated PLWH who are burdened by these
challenges. Qualitative research has found that some patients feel providers may exhibit bias
against HIV-positive drug users, limit provision of ART prescriptions to drug users, and are not
understanding of drug users’ competing needs (Gerbert, Maguire, Bleecker, Coates & McPhee,
1991, Wong et al., 2004; King et al., 2004; McCoy, 2005). Previous research has also suggested
that substance use and poverty co-occur with stigma and medical distrust, which may interfere
with a patient’s positive interactions with providers (Wong, et al., 2004; Bakken, et al., 2000;
Dennis, et al., 2015). Future research should examine the role of physician bias and their
relationships with formerly incarcerated PLWH.
Sexual orientation emerged as a significant factor in distinguishing those who engaged in
care successfully from those who did not. Respondents who identified as gay or bisexual were
most likely to engage care consistently. Research on formerly incarcerated PLWH has failed to
examine sexual orientation. The HIV prevention literature, however, suggests that young gay

62

men are more likely to seek out services than straight youth because the infrastructure
surrounding HIV prevention and testing historically developed for gay men (Harper, Hosek &
Robinson 2002; Harper et al., 2013). It may be that HIV awareness and use of HIV-care services
is normative among the gay and bisexual male community, making it easier for those who are
gay or bisexual to pursue care and feel comfortable doing so. Additionally, the HIV-infected
people in this study who are not gay or bisexual may be prone to greater HIV stigma than those
who are gay or bisexual. They may fear people will presume they are gay or bisexual because
they have HIV. As such, they may be unwilling to participate in HIV care due to fear, stigma,
and confidentiality concerns (Braithwaite & Arriola, 2003). Future research should strongly
consider in depth examinations into sexual orientation when evaluating engagement in care for
formerly or currently incarcerated PLWH. In sum, findings suggest that the reentry is a critical
period fraught with an abundance of challenges that can impact engagement in HIV care, many
of which remain under investigated.
Limitations
Several limitations of the current study should inform interpretation of the findings. First,
although the overall sample was 190, we could only locate 60 people for in-person interviews.
Sample size limited the complexity of analyses that could be performed and accuracy of
findings. Analyses with small sample sizes can suffer a decline in statistical power and an
increase in the likelihood of a type II error (false negative) (Pett, 2015; Tabachnick & Fidell,
2013). Given the risk of using such a small sample, I took measures to reduce the likelihood of
error by selecting statistical analyses that are appropriate for small samples (Pett, 2015;
Tabachnick & Fidell, 2013). To ensure additional accuracy, I verified information in the
interviews in other data sources. Despite the small sample and possibility of biased statistical

63

tests, I am confident of the findings because they are logical and consistent with previous
literature. Nevertheless, future research should attempt to replicate these findings with larger
samples.
Secondly, as a primary data source in this study was archival, the limitations that come
with archival data apply to this study. I relied on the CAREWare database to create a measure of
engagement in care. There were several advantages to using medical records from this source.
However, I did not personally collect or clean these data, necessitating I accept that some errors
in the data may be unknown to me (i.e. omissions in CAREWare database). Further, I cannot be
certain that I correctly classified each respondents’ engagement status given unknown data errors
or omissions in CAREWare. I did take care to mitigate these limitations by verifying the
CAREWare records with client’s self-reports whenever possible, lending me greater confidence
in the accuracy of these records.
Additionally, this study examined only factors associated with care engagement at two of
four socio-ecological levels. Mugavero’s original (2013) model of the socioecological influences
on care engagement (Figure 4) suggests several individual, relational, community, and policylevel factors can impact care engagement. Although this study is limited because only a subset of
factors suggested by Mugavero were investigated, the scope of this study did not allow for a full
examination of each. As is typical when using a socioecological framework, research is generally
only able to focus on one or two levels due to methodological constraints (DiClemente et al.,
2005; Mugavero et al., 2013). Future research should move beyond these limitations and
examine care engagement at all socioecological levels.
Moreover, the interview data may not have perfectly mapped onto the archival data. For
example, the CAREWare data allowed me to classify people based on their CD4 and viral load

64

counts over 39 months following release. The interview data, however, only looked at current or
recent behaviors. Therefore, we cannot know whether those current behaviors were stable
features and representative of what occurred over the 39-month period. Future research might
improve upon this design by following formerly incarcerated PLWH longitudinally, in order to
assess relational and individual-level risks concurrent with care engagement and nonengagement behaviors.
Finally, engagement in care was based from the HRSA medical performance measure
(HRSA, 2008). Recent reviews of measures of retention in care have suggested that only one
measure of retention may not sufficient (Mugavero et al., 2012, Hall et al., 2012, Baligh et al.,
2012). Previous research suggests accuracy of measure may be different depending on the
sample and data available and that some measures may be too rigid. Therefore, they recommend
researchers combine more than one measure when examining retention or engagement in care.
This study only used one measure of engagement and adapted this measure to be less rigid for
this sample. Future research should continue to explore optimal measures of engagement and
ideal combinations of measures, particularly over longer periods of investigation.
Implications for Research and Practice
This study has several important implications for practitioners and researchers. Findings
highlight that investigation into the specific needs for people who are inconsistently and not
engaged in care is sorely needed. The varied patterns of inconsistent engagement suggests it is
doubtful that there is a “one-size-fits-all” intervention for promoting care engagement. Ideally,
practitioners could consider tailoring interventions to individuals’ specific needs, grounding
interventions in an understanding of patients’ care engagement pattern.

65

For many patients who are out of care, stigma, lack of provider support, substance use,
and unmet basic needs due to financial impoverishment seem to be the most salient impediments
to engagement (Avery, et al., 2012; Pettus-Davis et al., 2009 Arriola et al., 2007). Given the
results of this study, interventions that focus on more than one of these barriers simultaneously
may be particularly promising for this population. Oxford Houses, for instance, are
democratically run substance use recovery homes based on the Housing First model. These
interventions prioritize access to housing as a gateway to other basic needs and resources
(Wolitski, et al., 2010; Wang, Hong, Samuels, Savit, Sanders & Kushel, 2010; Jason & Ferrari,
2010). Oxford houses help address drug addiction and community reentry by providing stable
housing without many restrictions or requirements. Most notably, these houses do not limit
length of stay, provide a network of job opportunities, and a network of other reentering
individuals in recovery. (Jason & Ferrari, 2010; Jason, Stone, Stevens & Light, 2014). Several
papers have been published on the success of Oxford Houses and other Housing First
interventions in increasing access to housing and reducing rates of relapse and recidivism (Jason,
Davis, Ferrari, 2006; Krishnan, et al., 2013; Wolitski, et al., 2010; Tsemberis, 2004). Holistic
interventions of this sort have never been examined in the context of PLWH during the
community reentry period. Housing interventions such as these, which attend to housing
instability, substance use recovery, social support, and employment, may be particularly useful in
addressing the needs former inmates living with HIV. These types of resources, paired with
financial incentives (HIV Prevention Trials Network Study, 2014) and integrated comprehensive
HIV and substance use treatment and care may best address to challenges facing PLWH during
reentry (Nunn, et al., 2009).

66

For PLWH post-incarceration, addressing only one socioecological layer of challenges is
not enough. Relationships, particularly, quality of patient-provider relationship, may affect care
engagement. Connection to a competent and trusted doctor immediately upon release is crucial.
Providers must be capable of effectively responding to this population’s co-occurring challenges.
Many treatment providers are ill equipped to identify, let alone treat, substance use and mental
health issues. McCoy (2005) found needs for respect and confidentiality were more commonly
emphasized in socially marginalized and drug-using groups compared to highly educated and
professionally employed groups. Given this, training physicians to be respectful of all patients
and prepared to work with patients with several compounding and often stigmatized health
challenges may be particularly important.
Several promising coordinated-care interventions could be adapted for this population.
These interventions combine high-quality providers who prioritize patient-provider relationships,
while connecting clients to ancillary medical and social services within close proximity. These
“medical home” models are patient centered and ensure care is well coordinated and satisfactory
to patients (Beane, Culyba, DeMayo & Armstrong, 2014). For example, the Bronx Transition
Clinic has demonstrated preliminary success in linking people to these comprehensive care
clinics (Fox et al., 2014; Teixeira, Jordan, Zaller, Shah, & Venters, 2014). The benefit of
coordinated care, particularly for this population, is the ease of access to other crucial services
and the identification of a personalized physician to facilitate care (Beane, Culyba, DeMayo &
Armstrong, 2014; Starfield & Shi, 2004). Using this approach post-incarceration could capitalize
on service accessibility, while also addressing quality of patient-provider relationships.
Findings also highlight the need to continue to pull apart the complexities of HIV stigma,
particularly among individuals who do not identify as gay or bisexual. There appear to be unique

67

factors at play that affect engagement in care in this population as a function of sexual
orientation. The care engagement literature among formerly incarcerated PLWH is in its infancy
and has yet to begin looking at subgroups differences. Future research might benefit from
examining post-incarceration experiences through the lens of intersectionality theory, which
suggests that social identities such as race, gender, and sexual orientation interact to form
qualitatively different meanings and experiences (Crenshaw, 1991; Bowleg, 2012; Warner,
2008).
Conclusion
Formerly incarcerated PLWH are at a critical disadvantage when released from prison.
Their health, safety, and wellness are at-risk. This population is more likely than the general
population of PLWH to never link to care (Baillargeon et al., 2010) or to fall out of care
(Baillargeon et al., 2009). With thousands of HIV-infected inmates being released from prison
each year, intervening to support their engagement in care during the reentry is crucial for their
individual health and the health of the community as a whole. Toward the aim of successfully
connecting formerly incarcerated PLWH to care, this study offers important contributions to the
literature. First, my work shows that care engagement does not follow one universal pattern
following release from prison. To address distinctive patterns of suboptimal interaction and their
unique predictors, interventions addressing engagement in care may need to be tailored.
Additionally, I provide preliminary evidence for the cascading effects of multiple burdens on
care engagement and the need for development of long-term multi-component and multi-level
comprehensive interventions. Future work must further elucidate the multi-level mechanisms
that prevent formerly incarcerated PLWH from securing health resources over the years
following their release.

68

APPENDICES

69

APPENDIX A: Individual-level Risk Index Score: Validation Study
It is difficult to measure a complex concept with a single indicator accurately, as rarely
are we able to capture all facets of a concept. By creating a composite risk index, errors deflate
and yield a more precise measure (Singleton & Straits, 2010; Davis, Smith & Marsden, 2007).
To assess individual-level factors that may influence care engagement, I constructed a composite
index to capture the cumulative impact of individual-level risk. The composite index was
constructed following Trochim & Donnelly’s (2001) steps for constructing an index. The
purpose of this appendix is to describe validation of the index. For a description of its
construction, please see the methods section in the main document.
To validate the index, I assessed two different types of construct validity, convergent and
discriminant validity (Campbell & Fiske, 1959; Trochim & Donnelly, 2001). Construct validity
refers to the idea that if you really measured what you think you measured your construct should
be related to a similar construct (Cronbach & Meehl, 1955; Trochim & Donnelly, 2001). In order
to claim that the index has construct validity, I needed to establish both convergent and divergent
validity. Convergent validation refers to the idea constructs that are theoretically associated with
one another are in fact statistically associated. Divergent validation refers to the idea that
measures of constructs that are theoretically unrelated are indeed unrelated statistically. To
establish convergent validity of my index, I examined associations between my individual-level
index and individual-level risk measures not included in the index (i.e. measures of depression,
drug and alcohol use, and HIV stress). I chose these three measures because they are all
individual-level risk factors for engagement in HIV care, they were collected with the same
analytic sample, they are not currently included in the individual-level risk index, and they are
theoretically similar to the index. Higher levels of depression, HIV stress, and substance use

70

should affect the ability to engage in care in the same way higher scores on the individual-level
risk index should effect the ability to engage in care (Traeger, O’Cleirigh, Skeer, Mayer &
Safren, 2012; Wohl et al., 2010). Depression was measured using the Center for Epidemiologic
Studies Depression Scale (CESD) (Radloff, 1977; Eaton, et al., 2004), HIV stress using the The
HIV/AIDS Stress Scale (Pakenham & Rinaldis, 2002), and substance use using the Alcohol and
Other Drug Abuse (AOD) (Center for Substance Abuse Treatment, 1994; Winters & Zenilman,
1994). I expected the individual-level risk index to correlate positively with measures of
HIV/AIDS stress, depression, and drug and alcohol use. There is no conventional criterion of
statistical significance to asses construct validity, therefore I used the standard p < .05
significance level to determine assessment of convergent and divergent validity.
To establish divergent validity, I examined associations between the individual-level
index and sexual activity (i.e. number of sexual partners in the last 90 days and number of times
engaging in unprotected sex in the last 90 days). I chose sexual activity because the literature
indicates the factors in my individual-level risk index bear no consistent relationship to these
aspects of sexual activity. Sexual activity was measured using a sexual risk and protective
behaviors measure for people living with HIV. I expected individual-level risk scores to be
uncorrelated with all measures of sexual activity.
Construct Validity: Convergent Validity.
To establish convergent validity, I examined associations between my individual-level
index and individual-level risk measures not included in the index (i.e. measures of depression,
drug and alcohol use, and HIV stress). I used the Pearson correlation coefficient to test
associations. The individual-level risk score was significantly and positively correlated with all
three subscales of the HIV/AIDS Stress Scale (social – r = .30; p = .02; instrumental – r = .51; p

71

= .00; emotional – r = .31; p = .02). The individual-level risk score trended toward significance
with depression (r = .25; p = .06). The individual-level risk score was positively and significantly
related to alcohol and other drug use (r = .31; p = .02).
Construct Validity: Divergent Validity.
To establish divergent validity, I examined the associations between the individual-level
index and sexual activity (number of sexual partners in the last 90 days and 12 months and
number of times engaging in unprotected sex in the last 90 days and 12 months) using the
Pearson correlation coefficient. The individual-level risk score was not significantly associated
with number of sex partners in past 90 days (r = .27; p = .12), number of sex partners in past 12
months (r = .25; p = .11), number of times engaging in unprotected anal sex (r = .18; p = .95), or
number of times engaging in unprotected vaginal sex (r = .77; p = .70).
Conclusion
This study reported the results of analyses to validate an individual-level risk index in a
sample of formerly incarcerated individuals recently released from prison. Results showed
evidence of convergent and divergent validity. The individual-level index was positively
associated with all subscales of the HIV stress scale, depression and drug and alcohol use. The
strongest correlation occurred between the index and the HIV instrumental stress subscale. This
is unsurprising given the economic themes present in both scales. Furthermore, as expected, the
individual-level index was not associated with any of the sexual activity measures. Overall, the
individual-level risk index was associated with the measures it theoretically should be, and the
index was not associated with measures it theoretically should not be. Given this evidence, I
consider the individual-level risk index to be a valid measure of cumulative individual-level risk.

72

APPENDIX B: APM Abstraction Form
DEMOGRAPHICS & CONTACT INFORMATION
1. Name of Client: First _____________________Middle _______________ Last

_____________________
2. Date of intake: ______/______/_______
3. Nicknames/Aliases________________________________________________
4. Date of Birth:

______/______/_______
5. Sex :

 Male



6. Hispanic or Latino?  YES

 NO

Female
7. Race:

 Black
/African
American


Asian/Pacific
Islander


Indian/Nativ
e American

 White


Arab/Chalde
an

 Other:
__________
______

8.

Current Address:
Street_________________________ Apt. #________ City
_______________State ______ Zip____________
a.
Whose home is this (Specify relationship to client)
___________________________
b.
Can mail be sent here?  Yes, Complete
 Yes, Discreet
 No

9.

Mailing Address:
Street_________________________ Apt. #________ City
_______________State ______ Zip____________
a.
Whose home is this (Specify relationship to client)
_____________________________________
b.
Can mail be sent here?  Yes, Complete
 Yes, Discreet
 No

10. Phone Numbers:

a.
 No

Home ____ Can Message be left?

 Yes, Complete  Yes, Discreet

b.
 No

Cell

 Yes, Complete  Yes, Discreet

____ Can Message be left?

73

c.
Work ________ Can Message be left?
Discreet
 No

 Yes, Complete  Yes,
 Yes, Complete 

11. Email _________________ Can Message be left?
Yes, Discreet
 No
12. Client Locator Information:

a.
Contact #1: First Name ________________Last
Name___________________
Home Phone ________-__________ Work Phone ________-__________
Cell Phone________-__________
Relationship__________________________
Aware of HIV Status?
Can Message be left?

 Yes

 No

 Yes, Complete  Yes, Discreet

 No

Street:______________________ Apt. #________ City
_______________State: ______ Zip________
b.
Contact #2: First Name _________________________Last Name
_________________________
Home Phone (________) ________-__________ Work Phone (________)
________-__________
Cell Phone(________) ________-__________
Relationship_______________________________
Aware of HIV Status?  Yes
 No
Yes, Complete  Yes, Discreet  No

Can Message be left?



Street:_________________________ Apt. #________ City
_______________State: ______ Zip__________
REMOVE THIS PAGE BEFORE LEAVING SITE AND STORE SEPARATELY DURING
CLINICAL REVIEW

13. HIV Status:

 HIV+ Not AIDS
 HIV+ AIDS Status Unknown
 CDC Defined AIDS
 HIV Negative (Affected)
14. Was HIV status verified?

 No

 Yes – Medical/Labs

Affected Client

74

 N/A

15. Date of HIV Diagnosis: ______/______/_______
16. Date of AIDS Diagnosis: ______/______/_______
17. ADAP?

 Yes

 No

18. Are you taking medications?

 Yes

 No

19. Are you new to taking HIV Medications?

 Yes

 No

20. When did you first start taking HIV Medications? ______/______/_______
21. What is your CD4? __________  Unknown
22. When were your last labs done? ______/______/_______
23. What is your viral load? _________  Unknown
24. Any Opportunistic infections?

 Yes

 No

a.
If yes, List
here___________________________________________________________
____
25. Medical concerns at this time:

__________________________________________________________________
__
26. Dental concerns at this time:

__________________________________________________________________
__
27. Optical concerns at this time:

__________________________________________________________________
_
HOUSEHOLD/LIVING INFORMATION
28. Homeless?

 Yes

 No

 Unknown/Unreported

29. Number in Household: _______
30. Annual Household Income: $__________
31. Source of Income: _________________________________
32. Food Stamps?

 Yes

 No

33. Are you currently employed?

 Yes
75

 No

a.

If yes:  Full Time

 Part Time

b.

If yes, where? ____________________

c.

If yes, how long? __________________

34. When was client last employed? ______/______/_______
35. Education? __________________________________________

a. Living Arrangements:  Permanent  Non-Permanent 
Other (Specify)______________________________
36. Marital Status:

 Single Never Married
 Partnered
 Married
 Married but separated
If separated, how long? __________________
 Widowed
 Divorced
37. Risk Factors

 Heterosexual
 MSM
 IDU
 Prenatal Transmission
 Hemophilia/Coagulation or Blood Transfusion
 Recipient of transfusion of blood, blood components, or tissue
 Other: Explain ______________________________________

76

CLIENT SERVICE

List of Client Needs :
38. Advocacy:
39. BHS/ Mental Health Counseling:
40. Benefits Assistance:
41. Chore Services:
42. Clothing:
43. Criminal Charges:
44. Crystal Clear Support Group:
45. Day/Respite Care:
46. Dental:
47. Education / GED:
48. Emergency Financial Assistance:
49. Employment:
50. Ex-Offenders Group Referral:
51. Furniture:
52. Food:
53. Guardianship:
54. HIV Education:
55. HIV Support Group:
56. Healthy Relationships APM Support Group:
57. Home Health Care:
58. Housing:
59. Insurance:
60. Legal Assistance:
61. Medical HIV:
62. Medical Non-HIV:
63. Medications:
64. Optical:
65. Parenting Group Referral:
66. Replace Lost/ Stolen ID:
67. Risk Reduction:
68. Spiritual Support:
69. Substance Abuse:
70. Transportation:
71. Utilities:
72. WAGS Referral for Pets:
73. Other:
77

 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes
 Yes

 No
 No
 No
 No
 No
No
 No
 No
 No
 No
 No
 No
 No
 No
 No
 No
 No
 No
 No
 No
 No
 No
 No
 No
 No
 No
 No
 No
 No
 No
 No
 No
 No
 No
 No
 No

If yes, specify:

78

CO-MORBIDITIES

74. TB:  Yes

 No

a.
If yes, receiving
treatment?  Yes
 No
75. Developmentally Delayed:
 Yes
 No
77. Hepatitis B:  Yes
79. Diabetes:  Yes

 No
 No

76. Cancers:  Yes

78. Hepatitis C:  Yes

 No

80. Hypertension (High Blood

Pressure):  Yes
81. STIs:  Yes

 No

 No

 No

a.
f yes, which STI’s (List
all):___________________________________
82. Mental Health History:

 Yes, Active in last 3 months
 Yes, Not Active in last 3
months
 Yes, No History

83. If client has a history of mental
health:
 In Treatment
 Dropped
 Waiting List for out of
Treatment
Treatment

 Refused
Treatment
 Completed
Treatment

 No Active
Treatment or
Counseling
 Not
Applicable

84. Does the Client want referral for BHS or other counseling?
 No
85. Substance Abuse History :

 Yes, Active in last 3 months
 Yes, Not Active in last 3
months
 No History

 Yes

86. If client has a history of
Substance Abuse:

 In Treatment
 Waiting List for
Treatment
 Refused
Treatment
 Completed
Treatment

79

 Dropped
out of
Treatment
 No Active
Treatment or
Counseling
 Not
Applicable

87. Are you presently using drugs/ alcohol?

 Yes

 No

88. Do you want drug/ alcohol treatment?

 Yes

 No

89. Names of drugs using/ used and other notes:

_______________________________________________________
______________________________________________________________
___
LEGAL INFORMATION
90. Has client been convicted or held

91. If the client has a history of or

for a crime?

any pending legal concerns, or needs
referral for legal advice, please
explain:

 No History
 Yes, In a correctional facility in the
last 3 months
 Yes, In a correctional facility 4
months ago or prior

_______________________________
____________
_______________________________
____________

92. Parole/ Probation Officer Name: First Name ____________________Last

Name ___________________________
93. Parole/ Probation Officer Phone Number : (________) ________-__________
94. Length of Parole/ Probation:___________
FOR APM – OFFICE USE ONLY

95. Date Intake Completed: ______/______/______
96. Name of Case Manager Assigned to Case:

First ______________________Last _________________________
97. Assessment Scheduled :

a.

Date: ______/______/_______

b.

Time: ___________________

c.

Location:_____________________________________________

98. Letter of Reminder Sent:_____________________________________________

80

APPENDIX C: Eligibility Screening Instrument for Interviews

Thanks for calling to find out about the study. My name is [name]. Michigan State University is
conducting a research study for the Michigan Department of Community Health. The study is of
Partnership Michigan’s community re-entry program. The purpose of the research is to find out
how people who used the program are doing now. We contacted you because you may have
participated in the re-entry program.
As a former client of this program, we would like to hear about your experiences with the
program and how you’ve been doing since then. If you are interested in participating, we would
like to interview you, in person, at a time and location that is most convenient for you. The
interview will take about 1 to 2 hours and we will pay you $50 for your time. Everything you tell
us will be completely confidential and private.
Would you be interested in participating in the project?
If no, say “Okay, thanks for your time”
If yes, proceed:
Can I ask you a few questions to confirm that you are eligible to participate? All your responses
will be kept completely confidential.
If no, “Unfortunately, I am unable to determine your eligibility without asking you a few
questions. Thank you for your time.”
If yes, proceed:
1. Have you been in prison in the State of Michigan at any time in the past 10 years?
___Yes
___No (skip to question 5)
___Not sure (skip to question 5)
___Refused “Unfortunately, I am unable to determine your eligibility. Thank you for your
time.”
2. How many times have you been in prison in the past 10 years? ______________
In which Michigan prison(s) were incarcerated during the last 10 years?
81

a. Alger
b. Baraga Maximum
c. Bellamy Creek
d. Boyer Road
e. Carson City
f. Central Michigan
g. Charles Egeler
h. Chippewa
i. Cooper Street
j. Earnest C. Brooks
k. Florence Crane
l. G. Robert Cotton
m. Gus Harriosn
n. Ionia Maximum
o. Kinross
p. Lakeland
q. Macomb
r. Marquette
s. Michigan Reformatory
t. Mid-Michigan/Pine River
u. Mound
v. Muskegon
w. Newberry
x. Oaks
y. Ojibway
z. Parnall
aa. Parr Highway
bb. Pugsley
cc. Richard A. Handlon
dd. Ryan
ee. Saginaw
ff. Special Alternative Incarceration
gg. St. Louis
hh. Straits
ii. Thumb
jj. West Shoreline
kk. Women’s Huron Valley
ll. Woodland Center
If none of these facilities are named, go to question 5.

82

3. When were you released from (facility named on list above)?
_____ (Month) ______ (Year)
___Refused “Unfortunately, I am unable to determine your eligibility. Thank you for your
time.”
4. Did you receive help on the telephone from Partnership Michigan to set up health care
and other services when you were about to be released?
___Yes
___No
___Not sure
___Refused “Unfortunately, I am unable to determine your eligibility. Thank you for your
time.”
5. What is your name?_______________________________________
___Refused “Unfortunately, I am unable to determine your eligibility. Thank you for your
time.”
6. What is your date of birth?
_____ (Month) ______ (Year)
___Refused “Unfortunately, I am unable to determine your eligibility. Thank you for your
time.”
7. What are the last 4 digits of your social security number?
_____/______/______/_______
COMPARE ANSWERS TO ELIGIBILITY LIST. Is this person eligible to participate?
_____ Yes
_____ No
_____ Unable to determine due to missing information
For INELIGIBLE (and unable to determine) participants, say:

83

“Participants for this research project are selected based on the questions you were just asked.
Based on your answers, it turns out you’re not eligible to participate . Thank you for taking the
time to speak with me.”
For ELIGIBLE participants, say:
“Thank you very much for the information you provided. Based on your answers to these
questions, you are eligible to participate in the pilot study. Are you interested in setting up a time
to meet?”
What I’d like to do is set up a time and a place that works best for you. We can come to your
home or we could meet in a public place such as a library. Since we’ll be asking you some
personal questions, it would be best if we could do the interview somewhere that we can speak
privately. When we meet, I’ll tell you more about the purpose of the study. At that time, I’ll ask
for your consent to participate or you can decide not to do the interview. Do you have any
additional questions about the project right now?
Where would you like to meet?___________________________________________________
What day and time is best?______________________________________________________
In case we need to reschedule, I’m going to give you an 800 number where we can be reached.
Give participant phone number
Is there a good phone number to contact you in case the interviewer is lost or delayed in making
the interview appointment? _______________________
Could you provide another contact number where I can reach you? ______________________
If there are no more questions, we’ll see you at [time] on [day] at [location].

84

APPENDIX D: Interview Recruitment Materials
Letters to Former Clients

85

Business Cards Left for Former Clients

86

Letter to Parole Officer

87

Letter to AIDS Service Agencies

88

Figure 6. APM Recruitment Flyer.

89

APPENDIX E: Interview Consent Form

90

91

APPENDIX F: Measures
A.
Patient/Provider Relationships (Bakken, Holzemer, Brown, et al., 2000)
[Interviewer: If respondent said they are not receiving medical care for HIV on page 7, question
4, skip section F.] Now we want to ask you some questions about the healthcare provider or
doctor who is the main person you see for care related to HIV and your relationship with him or
her. Please rate the degree to which each statement is true for you.
Always Usually Sometimes Never
Refuse
1. My healthcare provider or
doctor listens to me.
2. My healthcare provider or
doctor cares about me.
3. My healthcare provider or
doctor answers my
questions.
4. My health care provider or
doctor spends enough time
with me.
5. My healthcare provider or
doctor involves me in
decisions.
6. My healthcare provider or
doctor respects my
decisions.
7. My healthcare provider or
doctor deals with my
problems
8. My healthcare provider or
doctor engages me in my
care.
9. My healthcare provider or
doctor is helpful to me.
10. My healthcare provider or
doctor respects me.
11. My healthcare provider or
doctor supports my
decisions.
12. My healthcare provider or
doctor sees me when I ask.
13. My healthcare provider or
doctor provides me with
information.
14. My healthcare provider or
doctor explains things to

3

2

1

0

-9

3

2

1

0

-9

3

2

1

0

-9

3

2

1

0

-9

3

2

1

0

-9

3

2

1

0

-9

3

2

1

0

-9

3

2

1

0

-9

3

2

1

0

-9

3

2

1

0

-9

3

2

1

0

-9

3

2

1

0

-9

3

2

1

0

-9

3

2

1

0

-9

92

me in a way that I can
understand.
The next few questions are about your visits to the main healthcare provider or doctor who
manages your HIV care. Please indicate how strongly you agree or disagree with the following
statements.
Strongly
Strongly
Disagree Agree
Refuse
Disagree
Agree
15. I make or keep my
appointments with my
healthcare provider
every 3 to 4 months.
16. I don’t ask my
healthcare provider
about my treatment
options.
17. I keep a list of
symptoms to tell my
healthcare provider.
18. I don’t write down
questions before seeing
my healthcare provider.
19. I ask my healthcare
provider to explain a
word that I don’t know.
20. I’m sometimes too high
or dope sick when I go
to see my healthcare
provider.
21. I tell my healthcare
provider my symptoms.
B.

1

2

3

4

-9

1

2

3

4

-9

1

2

3

4

-9

1

2

3

4

-9

1

2

3

4

-9

1

2

3

4

-9

1

2

3

4

-9

Social Support
In the next few minutes, I would like to get an idea of the people who are important to you in
a number of different ways. I will be reading descriptions of the ways that people can be
important to us. After I read each description, I will be asking you to give me the first name,
initials or nicknames of the people who fit the description. These people might be friends,
family, ministers, doctors, or other people you know. I am interested in people you have
known for at least a week. So don’t mention someone you have just met. If you have
questions about the descriptions, please ask me to try to make it clearer.
Intimate Support:

93

1. If you wanted to talk to someone about things that are very personal and private or if a
situation came up where you needed advice, who would you talk to? (If none, record
“00” on the sheet and go to Question 4).
2. During the last 3 months, which of these people did you actually talk to about things that
were personal and private? Which people gave you advice? (Check about those not
mentioned and record total number on answer sheet.)
3. How satisfied or dissatisfied were you when you did talk to these people about your
private feelings or when you were given advice? Were you very satisfied, satisfied,
dissatisfied, or very dissatisfied?
4. During the past 3 months, how much to you think you needed people to talk to for advice
or about things that were very personal or private?
Physical Support:
5. Who are the people who would give up some of their time and energy to help you –
things such as driving you someplace you needed to go, going to the store for you and
things like that? (If none, record “00” on the sheet and go to Question 8).
6. During the last 3 months, how many of these people actually pitched in to help you with
things that you needed some help with? (Check about those not mentioned and record
total number on answer sheet.)
7. How satisfied or dissatisfied were you with the help you received? Were you very
satisfied, satisfied, dissatisfied, or very dissatisfied?
8. During the past 3 months, how much do you feel that you need people who would pitch
in to help you do things?
Material Support:
9. If you needed to borrow $25 or something valuable, who are the people who would lend
you money or something of value? (If none, record “00” on the sheet and go to Question
12).
10. During the last 3 months, how many of these people actually loaned you something that
you needed? (Check about those not mentioned and record total number on answer
sheet.)

94

11. How satisfied or dissatisfied were you with the things you were loaned? Were you very
satisfied, satisfied, dissatisfied, or very dissatisfied?
12. During the past 3 months, how much do you feel that you needed people who could loan
you money or give you things you needed?
Recreational Support:
13. Who are the people you get together with to have fun or relax? (If none, record “00” on
the sheet and go to Question 16).
14. During the last 3 months, which of these people did you actually get together with?
(Check about those not mentioned and record total number on answer sheet.)
15. How satisfied or dissatisfied were you with the times that you got together with people
just to relax or have fun? Were you very satisfied, satisfied, dissatisfied, or very
dissatisfied?
16. During the past 3 months, how much do you feel that you needed people who you could
just relax with?
Sexual Support:
17. Who are the people you could go to if you needed information about sex, condoms, and
so on? (If none, record “00” on the sheet and go to Question 20).
18. During the last 3 months, which of these people did you actually go to for information
about something related to sex? (Check about those not mentioned and record total
number on answer sheet.)
19. How satisfied or dissatisfied were you with the times that people gave you information
related to sex? Were you very satisfied, satisfied, dissatisfied, or very dissatisfied?
20. During the past 3 months, how much do you feel that you needed people who you could
give you information related to sex?
21. Now, I’d like to get some information about each of the people we have talked about.
How old is [name]? (Record ages for up to 15 persons. If uncertain, ask for best guess. If
no information given, record “888”.)

95

22. What is [name]’s relationship to you?
Father/stepfather ……..………………………………

1

Mother/stepmother.…………………………………….

2

Sibling…………………………….…………………….

3

Extended family member……………………………..

4

Peer…………………..…………………………………

5

Acquaintance…………………………………………..

6

Formal helper/care provider…………………………..

7

Spouse/partner…………………………………………

8

(Don’t Know)…….………………………………….…..

-8

23. Is [name] male or female?
24. You’ve mentioned a total of __ ___ people. How many of these people do you think are
well informed about AIDS?

____ _____ (“-88” = doesn’t know”)
25. How many of these people have you told that you are HIV positive?

____ _____ (“-88” = doesn’t know”)

96

Support Members

Intimate
could did

Physical
could did

Material
could did

1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
TOTALS
SATISFACTION
NEED

97

Social
could did

Sex
could did

Age

Rel.

Sex
M
F

C.

Demographics
Last, I’d like to ask you some general questions about yourself.
1. How many children do you have? ________
IF MORE THAN ZERO ASK: How many children are from each age group?
_____ UNDER ONE YEAR OLD
_____ 1 TO 5 YEARS OLD
_____ 6 TO 12 YEARS OLD
_____ OVER 12 YEARS OLD
2. Do you consider yourself to be:
HETEROSEXUAL/STRAIGHT

1

LESBIAN/GAY ………………………………………...

2

BISEXUAL …………………………..…………………

3

OTHER ( __________________________________)

4

REFUSED……………………………………………….

-9

3. What’s your educational level now?
LESS THAN HIGH SCHOOL..…………….………….

1

HIGH SCHOOL GRAD/GED………………………….

2

SOME COLLEGE………………………………….........

3

ASSOCIATE’S DEGREE (2-year)………………………

4

BACHELOR’S DEGREE (4-year)………………………

5

GRADUATE OR PROFESSIONAL DEGREE…………

6

REFUSED

-9

4. Were you employed before your most recent incarceration in the state prison?
YES……..………………………………

1

NO..…………………………………….

2 (skip to question 5)

(Unsure)………………………..

-8 (skip to question 5)

IF YES, ASK:

a. At that time, were you working part-time, full-time, or sporadically (off and on,
temporary)?
(FULL-TIME = 35 HOURS PER WEEK OR MORE)
PART-TIME……………………………..

1

FULL-TIME…………………………….

2

SPORADICALLY………………………

3

(Not Applicable)………………………..

-8

b. What was your monthly household income from all sources before your last
incarceration? (employment, TANF, disability, etc)?
[INCLUDE ONLY RELIABLE INCOME] ___________

c. How many people did that income support? _______________
9. Are you employed right now?
YES……..………………………………

1

NO..…………………………………….

2

(Not Applicable)………………………..

-8

IF YES, ASK:
a. Do you work part-time, full-time, or sporadically (off and on, temporary)?
(FULL-TIME = 35 HOURS PER WEEK OR MORE)
PART-TIME……………………………..

1

FULL-TIME…………………………….

2

SPORADICALLY………………………

3

(Not Applicable)………………………..

-8

10. What is your current monthly household income from all sources (employment, TANF,
disability, etc)
[INCLUDE ONLY RELIABLE INCOME] ___________________
99

a. How many people does this income support? _______________
11. Since participating in the re-entry program on (Interviewer: Remind participant of month and
year of their FIRST intake interview with APM) and since staying at the place where you first
lived when you were released, how many times have you moved? ____________
12. Since that time, were you ever homeless? By homeless I mean: (1) that you did not have your
own house or apartment, or (2) that you were living on the streets or in a shelter or (3) that you
were staying with friends or family in their apartment and did not pay them rent.
YES……..………………………………

1

 If yes, ask 15

NO..…………………………………….

2

 If no, skip to 16

13. Now I’d like to know how many times you were homeless, for how long, and where you
stayed when you were homeless.
a. What was the month and year that you were homeless? (Write response in column a
below).
b. Where were you staying? (Write response in column b below).
c. How many days did this episode of homeless last? (Write response in column c below).
(Record all episodes of homelessness between the date of intake and TODAY)
a. Start Date

b. Place

c. Number of days

1st Episode
2nd Episode
3rd Episode
4th Episode
5th Episode

14. As far as where you're living right now, do you:
RENTAPT OR HOME…………......

1

RENT A ROOM…………………….

2

OWN A HOME……………………

3

100

STAY WITH FRIENDS/RELATIVES…………….………

4

OTHER (_________________________________).

5

15. How do you feel about your current housing situation?
EXTREMELY SATISFIED……………………………….......

1

SATISFIED…………………………………………………....

2

MIXED (EQUALLY SATISFIED & DISSATISFIED)…….

3

DISSATISFIED………………….……………….………....

4

EXTREMELY DISSATISFIED…………………….………....

5

16. Could you tell me why you feel that way? [PROBE STABILITY & AFFORDABILITY]:
__________________________________________________________________
17. Do you currently have access to reliable transportation, e.g., your car, a friend’s car, or the
bus?
YES……..……………………………

1

NO..…………………………………….

2

101

REFERENCES

102

REFERENCES

Aberg, J. A., Kaplan, J. E., Libman, H., Emmanuel, P., Anderson, J. R., Stone, V. E., ... &
Gallant, J. E. (2009). Primary care guidelines for the management of persons infected with
human immunodeficiency virus: 2009 update by the HIV medicine Association of the
Infectious Diseases Society of America. Clinical Infectious Diseases, 49(5), 651-681.
Alfonso, V., Geller, J., Bermbach, N., Drummond, A., & Montaner, J. S. G. (2006). Becoming a
“treatment success”: what helps and what hinders patients from achieving and sustaining
undetectable viral loads. AIDS Patient Care and STDs, 20(5), 326–34.
http://doi.org/10.1089/apc.2006.20.326
Althoff, A. L., Zelenev, A., Meyer, J. P., Fu, J., Brown, S. E., Vagenas, P., … Altice, F. L.
(2013). Correlates of retention in HIV care after release from jail: Results from a multi-site
study. AIDS and Behavior, 17(SUPPL. 2), 156–170. http://doi.org/10.1007/s10461-0120372-1
Arnold, M., Hsu, L., Pipkin, S., McFarland, W., & Rutherford, G. W. (2009). Race, place and
AIDS: The role of socioeconomic context on racial disparities in treatment and survival in
San Francisco. Social Science and Medicine, 69(1), 121–128.
http://doi.org/10.1016/j.socscimed.2009.04.019
Baillargeon, J., Giordano, T. P., Harzke, A. J., Spaulding, A. C., Wu, Z. H., Grady, J. J., … Paar,
D. P. (2010). Predictors of Reincarceration and Disease Progression Among Released HIVInfected Inmates. AIDS Patient Care and STDs, 24(6), 389–394.
http://doi.org/10.1089/apc.2009.0303
Baillargeon, J., Giordano, T. P., Rich, J. D., Wu, Z. H., Wells, K., Pollock, B. H., & Paar, D. P.
(2009). Accessing antiretroviral therapy following release from prison. JAMA : The Journal
of the American Medical Association, 301(8), 848–857.
http://doi.org/10.1001/jama.2009.202
Bakken, S., Holzemer, W. L., Brown, M. a, Powell-Cope, G. M., Turner, J. G., Inouye, J., …
Corless, I. B. (2000). Relationships between perception of engagement with health care
provider and demographic characteristics, health status, and adherence to therapeutic
regimen in persons with HIV/AIDS. AIDS Patient Care and STDs, 14(4), 189–197.
http://doi.org/10.1089/108729100317795
Barrera, M., (1981). Preliminary development of a scale of social support. American Journal of
Community Psychology, 9, 435-447.
Beane, S. N., Culyba, R. J., DeMayo, M., & Armstrong, W. (2014). Exploring the medical home
in Ryan White HIV care settings: a pilot study. Journal of the Association of Nurses in
AIDS Care, 25(3), 191-202.
103

Bhatia, R., Hartman, C., Kallen, M. a., Graham, J., & Giordano, T. P. (2011). Persons newly
diagnosed with HIV infection are at high risk for depression and poor linkage to care:
Results from the steps study. AIDS and Behavior, 15(6), 1161–1170.
http://doi.org/10.1007/s10461-010-9778-9
Blashill, A. J., Lassiter, J., Wilson, J. M., Safren, S. A., & Parsons, J. T. (2014). Mental health
and substance use in the scale-up of HIV prevention. In Biomedical Advances in HIV
Prevention (pp. 139-159). Springer New York.
Bouris, A., Voisin, D., Pilloton, M., Flatt, N., Eavou, R., Hampton, K., ... & Schneider, J. A.
(2013). Project nGage: network supported HIV care engagement for younger black men
who have sex with men and transgender persons. Journal of AIDS & clinical research, 4.
Bowleg, L. (2012). The problem with the phrase women and minorities: intersectionality—an
important theoretical framework for public health. American Journal of Public Health,
102(7), 1267-1273.
Bozzette, S. A., Berry, S. H., Duan, N., Frankel, M. R., Leibowitz, A. A., Lefkowitz, D., ... &
Shapiro, M. F. (1998). The care of HIV-infected adults in the United States. New England
Journal of Medicine, 339(26), 1897-1904.
Braithwaite, R. L., & Arriola, K. R. (2003). Male prisoners and HIV prevention: a call for action
ignored. American Journal of Public Health, 93(5), 759-763.
Brinkley-Rubinstein, L., & Turner, W. L. (2013). Health impact of incarceration on HIV-positive
African American males: a qualitative exploration. AIDS Patient Care and STDs, 27(8),
450–8. http://doi.org/10.1089/apc.2012.0457
Broffrenbrenner, U. (1979). The Ecology of Human Development: Experiments by Nature and
Design.
Buchberg, M. K., Fletcher, F. E., Vidrine, D. J., Levison, J., Peters, M. Y., Hardwicke, R., …
Bell, T. K. (2015). A Mixed-Methods Approach to Understanding Barriers to Postpartum
Retention in Care Among Low-Income, HIV-Infected Women. AIDS Patient Care and
STDs, 29(3), 126–132. http://doi.org/10.1089/apc.2014.0227
Cargill, V. a. (2013). Linkage, engagement, and retention in HIV care among vulnerable
populations: “I”m sick and tired of being sick and tired". Topics in Antiviral Medicine,
21(4), 133–7. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/24225079
Carson, E. A. (2014). Prisoners in 2013. Bureau of Justice Statistics, (September), 1–34.
Cashman, S. B., Savageau, J. a, Lemay, C. a, & Ferguson, W. (2004). Patient health status and
appointment keeping in an urban community health center. Journal of Health Care for the
Poor and Underserved, 15(3), 474–488. http://doi.org/10.1353/hpu.2004.0037
Catz, Kelly, Bogard Benotsch, M. (2001). Patterns, Correlates and Barriers to Medication
104

Adherence Among Persons Prescribed New Treatments for HIV. Health Psychol, 19(2),
124–133. http://doi.org/10.1037//0278Centers for Disease Control and Prevention. (2008). Communities in Crisis: Is there a
Generalized HIV Epidemic in Impoverished Urban Areas of the United States? Retreived
from http://www.cdc.gov/hiv/group/poverty.html
Centers for Disease Control and Prevention. (2010). HIV by group. Retrieved from
http://www.cdc.gov/hiv/group/msm/index.html
Centers for Disease Control and Prevention. (2011). National HIV Surveillance System and
Medical Monitoring Project. Retrieved from https://www.aids.gov/federalresources/policies/care-continuum/
Centers for Disease Control and Prevention. (2014). HIV Care Saves Lives. Retrieved from
http://www.cdc.gov/vitalsigns/hiv-aids-medical-care/index.html
Centers for Disease Control and Prevention. (2015). HIV/AIDS. Retrieved from
http://www.cdc.gov/hiv/
Cheever, L. W. (2007). Engaging HIV-infected patients in care: their lives depend on it. Clinical
Infectious Diseases : An Official Publication of the Infectious Diseases Society of America,
44(11), 1500–1502. http://doi.org/10.1086/517534
Clear, T. R., Rose, D. R., & Ryder, J. A. (2001). Incarceration and the Community: The Problem
of Removing and Returning Offenders. Crime & Delinquency, 47(3), 335–351.
http://doi.org/0803973233
Cohen, S. E., & Syme, S. (1985). Social support and health. Academic Press.
Cohen, S., & Wills, T. A. (1985). Stress, social support, and the buffering hypothesis.
Psychological bulletin, 98(2), 310.
Cooper, R., & David, R. (1986). The biological concept of race and its application to public
health and epidemiology. Journal of Health Politics, Policy and Law, 11(1), 97–116.
http://doi.org/10.1215/03616878-11-1-97
Crenshaw, K. (1991). Mapping the margins: Intersectionality, identity politics, and violence
against women of color. Stanford law review, 1241-1299.
Delgado R, Stefancic J. Critical Race Theory: An Introduction. New York New York University
Press; 2001
Dennis, A. C., Barrington, C., Hino, S., Gould, M., Wohl, D., & Golin, C. E. (2015). “You’re in
a World of Chaos”: Experiences Accessing HIV Care and Adhering to Medications After
Incarceration. Journal of the Association of Nurses in AIDS Care, 26(5), 542–555.
http://doi.org/10.1016/j.jana.2015.06.001
105

Dhami, M. K., Mandel, D. R., Loewenstein, G., & Ayton, P. (2006). Prisoners’ Positive Illusions
of Their Post-Release Success. Law and Human Behavior, 30(6), 631–647.
http://doi.org/10.1007/s10979-006-9040-1
DiClemente, R. J., Salazar, L. F., & Crosby, R. a. (2007). A review of STD/HIV preventive
interventions for adolescents: sustaining effects using an ecological approach. Journal of
Pediatric Psychology, 32(8), 888–906. http://doi.org/10.1093/jpepsy/jsm056
DiClemente, R. J., Salazar, L. F., Crosby, R. A., & Rosenthal, S. L. (2005). Prevention and
control of sexually transmitted infections among adolescents: the importance of a socioecological perspective—a commentary. Public health, 119(9), 825-836.
DiMatteo, M. R. (2004). Social Support and Patient Adherence to Medical Treatment: A MetaAnalysis. Health Psychology, 23(2), 207–218. http://doi.org/10.1037/0278-6133.23.2.207
Edwards, L. V. (2006). Perceived Social Support and HIV/AIDS Medication Adherence Among
African American Women. Qualitative Health Research, 16(5), 679–691.
http://doi.org/10.1177/1049732305281597
Eldred, L. & Maltiz, F. (2007). Introduction. Aids Patient Care and STDs, 21(s1): S-1-S-2.
Elkington, K. S., Bauermeister, J. A., & Zimmerman, M. A. (2011). Do parents and peers
matter? A prospective socio-ecological examination of substance use and sexual risk among
African American youth. Journal of adolescence, 34(5), 1035-1047.
Exec. Order No. 13,649, 3 C.F.R. (2013).
Fontana, L., & Beckerman, A. (2007). Recently released with HIV/AIDS: primary care treatment
needs and experiences. Journal of Health Care for the Poor and Underserved, 18(3), 699–
714. http://doi.org/10.1353/hpu.2007.0058
Friedman, M. R., Stall, R., Silvestre, A. J., Wei, C., Shoptaw, S., Herrick, A., ... & Plankey, M.
W. (2015). Effects of syndemics on HIV viral load and medication adherence in the
multicentre AIDS cohort study. AIDS (London, England), 29(9), 1087-1096.
Fox, A. D., Anderson, M. R., Bartlett, G., Valverde, J., Starrels, J. L., & Cunningham, C. O.
(2014). Health outcomes and retention in care following release from prison for patients of
an urban post-incarceration transitions clinic. Journal of health care for the poor and
underserved, 25(3), 1139.
Gallant, M. P. (2003). The Influence of Social Support on Chronic Illness Self-Management: A
Review and Directions for Research. Health Education & Behavior, 30(2), 170–195.
http://doi.org/10.1177/1090198102251030
Galvan, F. H., Davis, E. M., Banks, D., & Bing, E. G. (2008). HIV stigma and social support
among African Americans. AIDS patient care and STDs, 22(5), 423-436.
106

Gardner, E. M., McLees, M. P., Steiner, J. F., Del Rio, C., & Burman, W. J. (2011). The
spectrum of engagement in HIV care and its relevance to test-and-treat strategies for
prevention of HIV infection. Clinical Infectious Diseases, 52(6), 793–800.
http://doi.org/10.1093/cid/ciq243
George, S., Garth, B., Wohl, A. R., Galvan, F. H., Garland, W., & Myers, H. F. (2009). Sources
and types of social support that influence engagement in HIV care among Latinos and
African Americans. Journal of Health Care for the Poor and Underserved, 20(4), 1012–35.
http://doi.org/10.1353/hpu.0.0210
Gerbert, B., Maguire, B. T., Bleecker, T., Coates, T. J., & McPhee, S. J. (1991). Primary care
physicians and AIDS: attitudinal and structural barriers to care. JaMa, 266(20), 2837-2842.
Giordano, T. P., Gifford, A. L., White, a C., Suarez-Almazor, M. E., Rabeneck, L., Hartman, C.,
… Morgan, R. O. (2007). Retention in care: a challenge to survival with HIV infection.
Clinical Infectious Diseases : An Official Publication of the Infectious Diseases Society of
America, 44(11), 1493–1499. http://doi.org/10.1086/516778
Gonzalez, A., Barinas, J., & O’Cleirigh, C. (2011). Substance use: impact on adherence and HIV
medical treatment. Current HIV/AIDS Reports, 8(4), 223.
Haley, D. F., Golin, C. E., Farel, C. E., Wohl, D. a, Scheyett, A. M., Garrett, J. J., … Parker, S.
D. (2014). Multilevel challenges to engagement in HIV care after prison release: a theoryinformed qualitative study comparing prisoners’ perspectives before and after community
reentry. BMC Public Health, 14(1), 1253. http://doi.org/10.1186/1471-2458-14-1253
Hall, H. I., Gray, K. M., Tang, T., Li, J., Shouse, L., & Mermin, J. (2012). Retention in Care of
Adults and Adolescents Living With HIV in 13 US Areas. JAIDS Journal of Acquired
Immune Deficiency Syndromes, 60(1), 77–82.
http://doi.org/10.1097/QAI.0b013e318249fe90
Hammett, T. M., Roberts, C., & Kennedy, S. (2001). Health-Related Issues in Prisoner Reentry.
Crime & Delinquency, 47(3), 390–409. http://doi.org/10.1177/0011128701047003006
Harper, G. W., Fernandez, I. M., Bruce, D., Hosek, S. G., Jacobs, R. J., & Adolescent Medicine
Trials Network for HIV/AIDS Interventions. (2013). The role of multiple identities in
adherence to medical appointments among gay/bisexual male adolescents living with HIV.
AIDS and Behavior, 17(1), 213-223.
Haney, C. (2003). The psychological impact of incarceration: Implications for post-prison
adjustment. Prisoners once removed: The impact of incarceration and reentry on children,
families, and communities, 33, 66.
Harawa, N., & Adimora, A. (2008). Incarceration, African Americans and HIV: Advancing a
Research Agenda. Journal of the National Medical Association, 100(1), 57–62. Retrieved
from
http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3171166&tool=pmcentrez&ren
107

dertype=abstract
Hosek, S. G., Harper, G. W., & Robinson, W. L. (2002). Identity development in adolescents
living with HIV. Journal of Adolescence, 25(4), 355-364.
House, J. S., Kahn, R. L., McLeo2015d, J. D., & Williams, D. (1985). Measures and concepts of
social support.
HPTN 065. TLC-Plus: A Study to Evaluate the Feasibility of an Enhanced Test, Link to Care,
Plus Treat Approach for HIV Prevention in the United States. U.S. National Institute of
Allergy and Infectious Diseases; 2010.
Iroh, P. A., Mayo, H., & Nijhawan, A. E. (2015). The HIV care cascade before, during, and after
incarceration: a systematic review and data synthesis. American Journal of Public Health,
105(7), e5-e16.
Israelski D, Gore-Felton C, Power R,Wood MJ, & Koopman C. Sociodemographic
characteristics associated with medical appointment adherence among HIV-seropositive
patients seeking treatment in a county outpatient facility. Prev Med. 2001;33(5):470–5
Jason, L. A., & Ferrari, J. R. (2010). Oxford house recovery homes: Characteristics and
effectiveness. Psychological services, 7(2), 92.
Jason, L. A., Ferrari, J. R., & Davis, M. (2014). Creating communities for addiction recovery:
The Oxford House model. Routledge.
Jason, L. A., Davis, M. I., & Ferrari, J. R. (2007). The need for substance abuse after-care:
Longitudinal analysis of Oxford House. Addictive Behaviors, 32(4), 803-818.
Kalichman, S. C., Catz, S., & Ramachandran, B. (1999). Barriers to HIV/AIDS treatment and
treatment adherence among African-American adults with disadvantaged education.
Journal of the National Medical Association, 91(8), 439–446.
Katzen, A. L. (2013). Commentary African American Men ’ s Health and Incarceration : Access
to Care upon Reentry and Eliminating Invisible Punishments. Berkeley Journal of Gender,
Law & Justice, 26(2), 221–253.
Kelly, J. D., Hartman, C., Graham, J., Kallen, M. a., & Giordano, T. P. (2014). Social support as
a predictor of early diagnosis, linkage, retention, and adherence to HIV care: Results from
the steps study. Journal of the Association of Nurses in AIDS Care, 25(5), 405–413.
http://doi.org/10.1016/j.jana.2013.12.002
Knowlton, a. R., Hua, W., & Latkin, C. (2005). Social support networks and medical service use
among HIV-positive injection drug users: implications to intervention. AIDS Care, 17(4),
479–492. http://doi.org/10.1080/0954012051233131314349
Krishnan, A., Wickersham, J. A., Chitsaz, E., Springer, S. A., Jordan, A. O., Zaller, N., & Altice,
108

F. L. (2013). Post-release substance abuse outcomes among HIV-infected jail detainees:
results from a multisite study. AIDS and Behavior, 17(2), 171-180.
Laufer, F. N., Arriola, K. R. J., Dawson-Rose, C. S., Kumaravelu, K., & Rapposelli, K. K.
(2002). From jail to community: innovative strategies to enhance continuity of HIV/AIDS
care. The Prison Journal, 82(1), 84-100.
Laws, M. B., Bradshaw, Y. S., Safren, S. a., Beach, M. C., Lee, Y., Rogers, W., & Wilson, I. B.
(2011). Discussion of sexual risk behavior in HIV care is infrequent and appears ineffectual:
A mixed methods study. AIDS and Behavior, 15(4), 812–822.
http://doi.org/10.1007/s10461-010-9844-3
Malta, M., Strathdee, S. A., Magnanini, M. M., & Bastos, F. I. (2008). Adherence to
antiretroviral therapy for human immunodeficiency virus/acquired immune deficiency
syndrome among drug users: a systematic review. Addiction, 103(8), 1242-1257.
Maruschak, L. Department of Justice. Bureau of Justice Statistics. (2012). HIV in Prisons, 20012010. Retrieved from http://www.bjs.gov/content/pub/pdf/hivp10.pdf
Mallik-Kane, K., & Visher, C. (2008). Health and Prisoner Reentry. Health and Prisoner
Reentry: How Physical, Mental, and Substance Abuse Conditions Shape the Process of
Reintegration. Retrieved from http://www.urban.org/research/publication/health-andprisoner-reentry/view/full_report
Mantel, N. (1963). Chi-square tests with one degree of freedom; extensions of the MantelHaenszel procedure. Journal of the American Statistical Association, 58(303), 690-700.
Mantel, N., & Haenszel, W. (1959). Statistical aspects of the analysis of data from retrospective
studies. J natl cancer inst, 22(4), 719-748.
McCoy, L. (2005). HIV-positive patients and the doctor-patient relationship: perspectives from
the margins. Qualitative Health Research, 15(6), 791–806.
http://doi.org/10.1177/1049732305276752
McLeroy, K. R., Bibeau, D., Steckler, a, & Glanz, K. (1988). An ecological perspective on
health promotion programs. Health Education Quarterly, 15(4), 351–377.
http://doi.org/10.1177/109019818801500401
Meyer JP, Cepeda J, Wu J, Trestman RL, Altice FL, Springer SA: Optimization of human
immunodeficiency virus treatment during incarceration: viral suppression suppression at the
prison gate. JAMA Intern Med 2014, 174:721–729.
Mowbray, C. T., Holter, M. C., Teague, G. B., & Bybee, D. (2003). Fidelity criteria:
Development, measurement, and validation. American Journal of Evaluation, 24(3), 315340.

109

Mugavero, M. J. (2008). Improving engagement in HIV care: what can we do? Topics in HIV
Medicine : A Publication of the International AIDS Society, USA, 16(5), 156–161.
Mugavero, M. J., Amico, K. R., Horn, T., & Thompson, M. A. (2013). The state of engagement
in HIV care in the United States: From cascade to continuum to control. Clinical Infectious
Diseases : An Official Publication of the Infectious Diseases Society of America, 57(8),
1164–71. http://doi.org/10.1093/cid/cit420
Mugavero, M. J., Davila, J. a, Nevin, C. R., & Giordano, T. P. (2010). From access to
engagement: measuring retention in outpatient HIV clinical care. AIDS Patient Care and
STDs, 24(10), 607–613. http://doi.org/10.1089/apc.2010.0086
Mugavero, M. J., Norton, W. E., & Saag, M. S. (2011). Health care system and policy factors
influencing engagement in HIV medical care: Piecing together the fragments of a fractured
health care delivery system. Clinical Infectious Diseases, 52(SUPPL. 2).
http://doi.org/10.1093/cid/ciq048
Mugavero, M. J., Westfall, A. O., Zinski, A., Davila, J., Drainoni, M.-L., Gardner, L. I., …
Giordano, T. P. (2012). Measuring retention in HIV care: the elusive gold standard. J
Acquir Immune Defic Syndr, 61(5), 574–580.
http://doi.org/10.1097/QAI.0b013e318273762f
Mumola, C.; Karberg, J. Drug Use and Dependence, State and Federal Prisons, 2004. US
Department of Justice, Bureau of Justice Statistics; 2006 [June 2, 2010]. Report No.: NCJ
213530. Washington, DC. Available at:
http://bjs.ojp.usdoj.gov/index.cfm?ty=pbdetail&iid=778
National Alliance to End Homelessness. 2006. Homelessness and Prisoner Reentry.
http://www.nationalhomeless.org/factsheets/hiv.html
Nunn A, Zaller N, Dickman S, Trimbur C, Nijhawan A, Rich JD. Methadone and buprenorphine
prescribing and referral practices in US prison systems: Results from a nationwide survey.
Drug Alcohol Depend. 2009; 105(1-2):83–88. [PubMed:
Palepu, a, Tyndall, M. W., Chan, K., Wood, E., Montaner, J. S. G., & Hogg, R. S. (2004).
Initiating highly active antiretroviral therapy and continuity of HIV care: the impact of
incarceration and prison release on adherence and HIV treatment outcomes. Antiviral
Therapy, 9(5), 713–719. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15535408
Petersilia, J. (2003). When prisoners come home: Parole and prisoner reentry. Oxford University
Press.
Pett, M.A. (2015). Nonparametric statistics for health care research: Statistics for small samples
and unusual distributions. Sage Publications.
Pettus-Davis, C., Scheyett, A. M., Hailey, D., Golin, C., & Wohl, D. (2009). From the “Streets”
To “Normal Life”: Assessing the Role of Social Support in Release Planning for HIV110

Positive and Substance-Involved Prisoners. Journal of Offender Rehabilitation, 48(5), 367–
387. http://doi.org/10.1080/10509670902979447
Rajabiun, S., Mallinson, R. K., McCoy, K., Coleman, S., Drainoni, M.-L., Rebholz, C., &
Holbert, T. (2007). “Getting me back on track”: the role of outreach interventions in
engaging and retaining people living with HIV/AIDS in medical care. AIDS Patient Care
and STDs, 21 Suppl 1(October 2015), S20–S29. http://doi.org/10.1089/apc.2007.9990
Rebeiro, P., Althoff, K. N., Buchacz, K., Gill, M. J., Horberg, M., Krentz, H., & Moore, R.
(2013). Retention among north american HIV infected persons in clinical care, 2000-2008.
Journal of Acquired Immune Deficiency Syndromes, 62(3), 356–362.
http://doi.org/10.1097/QAI.0b013e31827f578a.
Reed, S. J., & Miller, R. L. (2016). Thriving and adapting: resilience, sense of community, and
syndemics among young Black gay and bisexual men. American journal of community
psychology, 57(1-2), 129-143.
Richie, B. E., Freudenberg, N., & Page, J. (2001). Reintegrating women leaving jail into urban
communities: a description of a model program. Journal of Urban Health, 78(2), 290-303.
Robbins, G. K., Daniels, B., Zheng, H., Chueh, H., Meigs, J. B., & Freedberg, K. a. (2007).
Predictors of antiretroviral treatment failure in an urban HIV clinic. Journal of Acquired
Immune Deficiency Syndromes (1999), 44(1), 30–7.
http://doi.org/10.1097/01.qai.0000248351.10383.b7
Roberts, K. J. (2002). Physician-patient relationships, patient satisfaction, and antiretroviral
medication Adherence among HIV-infected adults attending a public health clinic. AIDS
Patient Care and STDs, 16(1), 43–50. http://doi.org/10.1089/108729102753429398
Rose, D. R., & Clear, T. R. (1998). Incarceration, social capital, and crime: Implications for
social disorganization theory. Criminology, 36(3), 441-480.
Scheyett, A., Parker, S., Golin, C., White, B., Davis, C. P., & Wohl, D. (2010). HIV-infected
prison inmates: Depression and implications for release back to communities. AIDS and
Behavior, 14(2), 300–307. http://doi.org/10.1007/s10461-008-9443-8
Schneider, J., Kaplan, S. H., Greenfield, S., Li, W., & Wilson, I. B. (2004). Better physicianpatient relationships are associated with higher reported adherence to antiretroviral therapy
in patients with HIV infection. Journal of General Internal Medicine, 19(11), 1096–1103.
http://doi.org/10.1111/j.1525-1497.2004.30418.x
Schnittker, J., & John, A. (2007). Enduring stigma: the long-term effects of incarceration on
health. Journal of health and social behavior, 48(2), 115-130.
Serovich, J. M., Kimberly, J. A., Mosack, K. E., & Lewis, T. L. (2001). The role of family and
friend social support in reducing emotional distress among HIV-positive women. Aids Care,
13(3), 335-341.
111

Shinkfield, A. J., & Graffam, J. (2010). The relationship between emotional state and success in
community reintegration for ex-prisoners. International Journal of Offender Therapy and
Comparative Criminology, 54(3), 346-360.
Simoni, J. M., Mason, H. R., & Marks, G. (1997). Disclosing HIV status and sexual orientation
to employers. AIDS Care, 9(5), 589–599. http://doi.org/10.1080/713613192
Singer, M. (1994). AIDS and the health crisis of the US urban poor; the perspective of critical
medical anthropology. Social science & medicine, 39(7), 931-948.
Springer, S. a, Pesanti, E., Hodges, J., Macura, T., Doros, G., & Altice, F. L. (2004).
Effectiveness of Antiretroviral Therapy among HIV- Infected Prisoners : Reincarceration
and the Lack of Sustained Benefit after Release to the Community. Clin Inf Dis, 38, 1754–
60. http://doi.org/10.1086/421392
Springer, S. A., Friedland, G., Doros, G., Pesanti, E., & Altice, F. (2007). Antiretroviral
Treatment Regimen Outcomes Among HIV-Infected Prisoners Sandra. HIV Clinical Trials,
8(4), 205–212.
Springer, S. a., Spaulding, a. C., Meyer, J. P., & Altice, F. L. (2011). Public Health Implications
for Adequate Transitional Care for HIV-Infected Prisoners: Five Essential Components.
Clinical Infectious Diseases, 53(5), 469–479. http://doi.org/10.1093/cid/cir446
Stall, R., Mills, T. C., Williamson, J., Hart, T., Greenwood, G., Paul, J., ... & Catania, J. A.
(2003). Association of co-occurring psychosocial health problems and increased
vulnerability to HIV/AIDS among urban men who have sex with men. American journal of
public health, 93(6), 939-942.
StataCorp. 2015. Stata Statistical Software: Release 14. College Station, TX: StataCorp LP.
Starfield, B., & Shi, L. (2004). The medical home, access to care, and insurance: a review of
evidence. Pediatrics, 113(Supplement 4), 1493-1498.
Stephenson, B. L., Wohl, D. a, Golin, C. E., Tien, H.-C., Stewart, P., & Kaplan, A. H. (2005).
Effect of release from prison and re-incarceration on the viral loads of HIV-infected
individuals. Public Health Reports, 120(1), 84–88.
Stone, A., Jason, L. A., Stevens, E., & Light, J. M. (2014). Factors affecting the stability of
social networks during early recovery in ex-offenders. The American journal of drug and
alcohol abuse, 40(3), 187-191.
Surratt, H. L., Kurtz, S. P., Levi-Minzi, M. a, & Chen, M. (2015). Environmental Influences on
HIV Medication Adherence: The Role of Neighborhood Disorder. American Journal of
Public Health, 105(8), 1660–1667. http://doi.org/10.2105/AJPH.2015.302612
Tabachnick, B. & Fidell L. (2013). Using Multivariate Statistics. Pearson; 6th edition.
Teixeira, P. A., Jordan, A. O., Zaller, N., Shah, D., & Venters, H. (2015). Health outcomes for
112

HIV-infected persons released from the New York City jail system with a transitional carecoordination plan. American journal of public health, 105(2), 351-357.
Torian, L. V, Wiewel, E. W., Liu, K.-L., Sackoff, J. E., & Frieden, T. R. (2008). Risk factors for
delayed initiation of medical care after diagnosis of human immunodeficiency virus.
Archives of Internal Medicine, 168(11), 1181–7.
http://doi.org/10.1001/archinte.168.11.1181
Traeger, Lara; O’cleirigh, Conall; Skeer, Margie R; Mayer, Kenneth H; Safren, S. A. (2012).
Risk factors for missed HIV primary care visits among men who have sex with men ProQuest Psychology Journals - ProQuest. Journal of Behavioral Medicine, 35, 548–556.
Retrieved from
http://search.proquest.com.proxy2.cl.msu.edu/psychology/docview/1039014636/589CFB52
CEB843B1PQ/22?accountid=12598
Travis, J. C. (2005). But they all come back: facing the challenges of prisoner reentry. Urban
Institute Press, 391.
Travis, J., Solomon, A. L., & Waul, M. (2001). From prison to home: The dimensions and
consequences of prisoner reentry.
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(4), 651-656.
Ulett, K. B., Willig, J. H., Lin, H.-Y., Routman, J. S., Abroms, S., Allison, J., … Mugavero, M.
J. (2009). The therapeutic implications of timely linkage and early retention in HIV care.
AIDS Patient Care and STDs, 23(1), 41–49. http://doi.org/10.1089/apc.2008.0132
Visher, C. a., & Mallik-Kane, K. (2007). Reentry experiences of men with health problems.
Public Health Behind Bars: From Prisons to Communities, 434–460.
http://doi.org/10.1007/978-0-387-71695-4_25
Visher, C., La Vigne, N., & Travis, J. (2004). Returning home: Understanding the Challenges of
Prisoner Reentry. Urban Institute. http://doi.org/10.1037/e720382011-001
Wainer, H. (1978). On the sensitivity of regression and regressors.
Warner, L. R. (2008). A best practices guide to intersectional approaches in psychological
research. Sex roles, 59(5-6), 454-463.
Westen, D., & Rosenthal, R. (2003). Quantifying construct validity: two simple measures.
Journal of Personality and Social Psychology, 84(3), 608.
Westergaard, R. P., Spaulding, A. C., & Flanigan, T. P. (2013). HIV among persons incarcerated
in the US: A review of evolving concepts in testing, treatment and linkage to community
care. Current opinion in infectious diseases, 26(1), 10.
113

The White House Office of National AIDS Policy. (2010). National HIV AIDS Strategy for the
United States. Washington, DC.
The White House Office of National AIDS Policy. (2015). Update to the National HIV AIDS
Strategy for the United States. Washington, DC.
Williams, D. R. (1999). Race, socioeconomic status, and health. The added effects of racism and
discrimination. Annals of the New York Academy of Sciences, 896, 173–188.
http://doi.org/10.1111/j.1749-6632.1999.tb08114.x
Wohl, A. R., Galvan, F. H., Myers, H. F., Garland, W., George, S., Witt, M., … Lee, M. L.
(2011). Do social support, stress, disclosure and stigma influence retention in HIV care for
Latino and African American men who have sex with men and women? AIDS and
Behavior, 15(6), 1098–1110. http://doi.org/10.1007/s10461-010-9833-6
Wolf, M., Davis, T.C., Osborn, C., & Bennett., C.L. (2007). Literacy, self-efficacy, and HIV
medication adherence. Patient Education and Counseling (65)253-260.
Wolitski, R. J., Kidder, D. P., Pals, S. L., Royal, S., Aidala, A., Stall, R., ... & Housing and
Health Study Team. (2010). Randomized trial of the effects of housing assistance on the
health and risk behaviors of homeless and unstably housed people living with HIV. 3), 493503.
Yehia, B. R., Fleishman, J. A., Metlay, J. P., Korthuis, P. T., Agwu, A. L., Berry, S. A., &
Moore, R. D. (2012). Comparing different measures of retention in outpatient HIV care.
AIDS (London, England), 26(9), 1131.

114