Social support and housing transitions among homeless adults with serious mental illness and substance use disorders

Abstract

Objective

Research suggests that social supports are associated with housing retention among adults who have experienced homelessness. Yet, we know very little about the social support context in consumers find and retain housing. We examined the ways and identified the junctures in which consumers' skills and deficits in accessing and mobilizing social supports influenced their longitudinal housing status.

Methods

We performed semi-structured qualitative interviews with VA Greater Los Angeles consumers (n=19) with serious mental illness (SMI), substance use disorders (SUD), and a history of homelessness; interviews explored associations between longitudinal housing status (categorized as: stable, independent housing; sheltered housing, continually engaged in structured housing programs; and unstable housing) and social supports. We compared data from consumers in these three mutually exclusive categories.

Results

All participants described social support as important for finding and maintaining housing. However, participants used formal (provider/case managers) and informal (family/friends) supports in different ways. Participants in stable housing relied on formal and informal supports to obtain/maintain housing. Participants in sheltered housing primarily used formal supports, e.g., case management staff. Unstably housed participants used formal and informal supports, but some of these relationships were superficial or of negative valence. Interpersonal problems were prevalent across longitudinal housing status categories.

Conclusions and Implications for Practice

Social context, including patterns of formal and informal support, was associated with participants' longitudinal housing status. Within interventions to end homelessness, these findings suggest the value of future research to identify, tailor, and implement practices that can help consumers improve their social resources.

Introduction

Adults with serious mental illness (SMI) are 10-20 times more likely to experience homelessness than the general population. (Kuno, Rothbard, Averyt, & Culhane, 2000) The presence of a co-occurring substance use disorder (SUD) further escalates consumers' risk of experiencing homelessness. (Balshem, Christensen, Tuepker, & Kansagara, 2011; Padgett, Stanhope, Henwood, & Stefancic, 2010) Though substantial research examines factors conveying risk for homelessness, (Balshem et al., 2011; Hamilton, Poza, & Washington, 2011) few have studied factors influencing exits from homelessness. (Gabrielian, Bromley, et al., 2015a) In particular, though research suggests that social supports are associated with housing retention among adults who have experienced homelessness,(Nelson et al., 2015; Rosenheck, Morrissey, & Lam, 2001; Wong & Stanhope, 2009) we know very little about the social support context in which consumers with a history of homelessness find and retain housing.

Interventions to end homelessness generally follow one of two paradigms. Traditionally, services were offered along a linear “continuum,” with consumers progressing from shelters, to transitional housing, to residential treatment, and eventually independent housing as they grew adherent with mental healthcare. (Greenwood, Schaefer-McDaniel, Winkel, & Tsemberis, 2005; Lamb & Bachrach, 2001) More recently, services shifted towards a “Housing First” approach, in which consumers receive independent, permanent housing with community-based supportive services, including non-mandated treatment referrals. (Greenwood et al., 2005; Hopper & Barrow, 2003)

Current services span both models and provide a path for many consumers to exit homelessness. (McGuire, Rosenheck, & Kasprow, 2010; O'Toole & Pape, 2015; Stergiopoulos et al., 2015) However, despite such resources, many consumers still struggle to secure and retain housing; in the Department of Veterans Affairs' (VA) Housing First program—one of the nation's largest (>70,000 participants) permanent supportive housing initiatives—25% of participants exit the program each year,(O'Connell, Kasprow, & Rosenheck, 2012) usually with negative outcomes, like incarceration or returns to homelessness. (Gabrielian, Burns, et al., 2015b) Moreover, many formerly homeless consumers who achieve housing (often via Housing First services) remain socially isolated, with negligible improvements in their social integration. (Tsai, Mares, & Rosenheck, 2012) The rates of recidivism into homelessness (O'Connell et al., 2012) and limited social integration of recently housed consumers (Tsai et al., 2012) point to active performance gaps within current services for persons who have experienced homelessness. To this end, examining the understudied relationships between social supports and longitudinal housing status may inform the implementation of innovative services with in current interventions for persons with a history of homelessness.

In the general population, positive social support, i.e., help or comfort that improves one's emotions, promotes improved health and lessens emotional strain from life stressors. (Berry & Welsh, 2010; Carton, Young, & Kelly, 2009; Hwang et al., 2009) Supports can be formal (paid, e.g., case managers or other clinicians) or informal (natural, e.g., family/friends). (Cantor, 1979; Pichler & Wallace, 2007) They can address emotional needs; financial problems; and/or perform instrumental tasks,(Hwang et al., 2009) e.g., assistance with apartment rentals. To this end, some suggest that vulnerability towards homelessness derives in part from deficits in accessing and/or mobilizing positive supports, (Hwang et al., 2009; Wong & Stanhope, 2009) or reliance on supports that are negative or undermining, (Rosenheck et al., 2001) i.e., supports that cause emotional discomfort. In addition, among persons with a history of homelessness, positive social supports are associated with subjective well-being, (Barczyk, Thompson, & Rew, 2014) improved mental health, decreased use of substances,(Hwang et al., 2009) and better housing retention. (Nelson et al., 2015; Rosenheck et al., 2001; Wong & Stanhope, 2009)

Important questions remain about the roles and characteristics of supports used to obtain and/or retain a residence. Most literature on homeless consumers' social networks describes relationships between negative supports with risky behaviors, e.g., drug use, and these behaviors in homeless persons;(Nyamathi, Leake, & Keenan, 2000; Rhoades et al., 2011; Wenzel, Golinelli, Zazzali, & McCarty, 2009) less attention is devoted to associations among the types, quantity, and nature of supports and individuals' housing trajectories. (Rosenheck et al., 2001) This paper uses qualitative methods to examine the ways and identify the junctures in which consumers' skills and deficits in accessing and mobilizing positive social supports—both formal and informal—influence longitudinal housing status. We use this information to generate hypotheses about the roles of social support in exiting homelessness and to suggest clinical interventions that may complement and improve current services for this vulnerable population. Data were part of a larger study that—among consumers with SMI, SUD, and a history of homelessness—identified categories of longitudinal housing status and highlighted patient- and environmental-level variables that best-differentiated consumers into these categories. (Gabrielian, Bromley, et al., 2015a)

Methods

We used a grounded theory approach to generate insights about the role of social support in longitudinal housing status. (Starks & Trinidad, 2007) Accordingly, we employed an iterative process during data collection and analyses for the parent study in which we refined our interview guide, sampling strategy, and analytic approach in order to explore emergent findings.

Participants

We used a theoretical sampling approach in which findings from the parent study informed the accrual of the interview sample. Inclusion criteria for the parent study included: 1) a history of enrollment in a residential rehabilitation program on VA Greater Los Angeles' grounds (called the “Domiciliary”) between 12/1/2008 and 11/30/2011; 2) homelessness at the time of Domiciliary admission; 3) at least one SMI diagnosis (broadly defined, including major depressive disorder, bipolar disorder, posttraumatic stress disorder, or psychotic disorder); and 4) at least one co-occurring SUD diagnosis (alcohol or drug abuse or dependence in DSM IV). We queried an administrative dataset of VA healthcare utilization (the Veterans Health Administration Medical SAS Data Set) to identify consumers who met these inclusion criteria.

The larger study (Gabrielian, Bromley, et al., 2015a) used the Residential Time-Line Follow-Back (TLFB) Inventory (Tsemberis, McHugo, Williams, Hanrahan, & Stefancic, 2006) to identify three mutually exclusive categories of longitudinal housing status. The TLFB is a validated instrument to collect consumers' housing event histories, categorizing each residence as “stable” (permanent housing), “institutional” (e.g., residential rehabilitation) or “unstable” (temporary settings or places not meant for human habitation). (Tsemberis et al., 2006) We used the TLFB to gather housing history between Domiciliary discharge and the day of data collection (mean=2.5 years), calculating the percent of total days spent in stable, institutional, and unstable housing. (Gabrielian, Bromley, et al., 2015a)

The parent study aimed to identify factors that best-predicted achievement of independent housing outside the context of a subsidized housing voucher, i.e., successful exits from homelessness without the overt provision of housing. As such, we were specifically interested in consumers' housing trajectories outside the context of subsidized housing, e.g., the VA's Housing First program. As such, in developing our longitudinal housing status categories, we analyzed TLFB housing data from Domiciliary discharge to the day of data collection or, if applicable, the day of subsidized apartment move-in. As such, all participants fell into one of the following mutually exclusive categories of longitudinal housing status: 1) stable, independent housing (70-100% of days in stable housing, e.g., one's own home or permanent residence in a family member's home); 2) sheltered housing (70-100% of days in residential rehabilitation or other structured programs); and 3) unstable housing (stable and independent or sheltered housing for <70% of days). (Gabrielian, Bromley, et al., 2015a) Across these categories, participants engaged in diverse VA housing services, though a minority of participants obtained housing on their own, without using services intended for persons who have experienced homelessness.

From the parent study, we purposively selected a subsample of participants (n=19) for qualitative assessments, maximizing the sample's variation across SMI diagnoses and the three categories of longitudinal housing status. Of note, participation in the VA's Housing First program (or other subsidized voucher program) did not influence the selection of participants for the qualitative interviews described in this paper; distinct from the parent study, we were interested in consumers' connections between social supports and housing status regardless of housing services received. The research team invited participants to participate in qualitative data collection after the completion of quantitative assessments, continuing sampling until data review suggested thematic saturation. The VA Greater Los Angeles Institutional Review Board approved these procedures. Informed consent was obtained from all participants.

Interview Structure

All individual interviews (30-40 minutes/each) were conducted by one of two study authors. Interviews followed administration of the aforementioned TLFB. (Tsemberis et al., 2006) We developed a semi-structured interview guide to query each participant about his current and immediately preceding residence. Of note, though we excluded subsidized apartments (generally obtained through VA's Housing First program) when deriving our longitudinal housing status categories, we interviewed participants about such housing arrangements (if such was their current or immediately preceding residence in the TLFB).

We designed the interview to provide detailed information about the relationships between social supports and housing; in so doing, we aimed to develop a more comprehensive understanding of these relationships than would be generated by survey data alone. (Curry, Nembhard, & Bradley, 2009) We posed the same questions for each of these residences (which ranged from unsheltered settings, e.g., abandoned buildings, to permanent housing in the community). First, we asked participants to describe help they received in finding and securing the residence. Next, we asked participants to describe people who made it difficult and/or easier to live at the residence. Last, we queried participants about problems faced in each residence.

Analyses

Interviews were digitally recorded and professionally transcribed; written transcripts were checked against audio-recordings for accuracy. Analyses were conducted using ATLAS.ti, (“Atlas.ti,” n.d.) a qualitative data software tool. In developing a codebook grounded in the perspectives of participants, we approached the qualitative data as a potential means to elaborate and enhance results from our quantitative analyses with regards to distinctions between the described housing status categories. (Greene, Caracelli, & Graham, 1989)

Three authors blinded to housing status categories independently reviewed the transcripts and developed a set of codes to identify concepts in the data related to the relationships between social supports and housing status. (Gabrielian, Bromley, et al., 2015a) Next, we iteratively refined the codebook to capture essential features of interpersonal processes, including narratives describing formal vs. informal supports; positive vs. negative supports; and support functions (emotional, financial, and instrumental). We defined supports broadly, encompassing all individuals (including housing program staff, family, friends, and other acquaintances) described by participants in discussing their housing transitions. We conceptualized conflicts to include all interpersonal interactions with negative valence, including discrimination experiences.

At each stage, coders overlapped in coding at least 10% of the interviews, comparing responses, reconciling disagreements, discussing with other authors, and refining the codebook until agreement was reached; the finalized codebook was subsequently applied to the entire dataset. A constant comparative approach was used to link codes across and within interviews, as well as to search for distinctions between concepts marked by codes. Finally, two authors used constant comparison to review all transcripts to search for salient similarities and differences in experiences across all participants, as well as by participants' longitudinal housing status category (using document families in ATLAS.ti).

Results

Table 1 describes our sample. Participants ranged in age from 32 to 69 years. All were male. Very few (10.5%) were married or partnered; the majority (57.9%) were divorced or separated. Most (63.2%) self-identified as non-Hispanic black, with fewer (15.8% each) Hispanic or non-Hispanic white participants.

Table 1. Sample Demographics, Diagnoses, and Longitudinal Housing Status.

Characteristic	Sample (N=19)
Age (years)
Range	32 – 69
Mean ± SD	53.0 ± 8.2
Male, n (%)	19 (100)
Marital status, n (%)
Never married	4 (21.1)
Married	2 (10.5)
Divorced or separated	11 (57.9)
Widowed	2 (10.5)
Race/ethnicity, n (%)
Non-Hispanic white	3 (15.8)
Non-Hispanic black	14 (63.2)
Hispanic, any race	3 (15.8)
Decline to state	1 (5.3)
Serious mental illness, n (%)
Major depressive disorder	4 (21.1)
Bipolar disorder	3 (15.8)
Posttraumatic stress disorder	7 (36.8)
Psychotic disorder	5 (26.3)
Substance abuse or dependence, n (%)
Alcohol	5 (26.3)
Cannabis	1 (5.3)
Cocaine	4 (21.1)
Opioid	1 (5.3)
Polysubstance	8 (42.1)
Longitudinal housing status, n (%)
Stable Housing	8 (42.1)
Sheltered Housing	5 (26.3)
Unstable Housing	6 (31.6)

Primary SMI and SUD diagnoses were captured from the medical record at Domiciliary discharge; participants had a range of SMI diagnoses, most commonly PTSD (36.8%), followed by psychotic disorders (26.3%). Most (42.1%) participants were diagnosed with polysubstance abuse or dependence. Similar numbers (5 to 8) fell into each of the three categories of longitudinal housing status.

Below, we describe the sample's social support context in finding/maintaining housing and discuss stated interpersonal problems. We detail similarities and differences by longitudinal housing status category, (Gabrielian, Bromley, et al., 2015a) summarizing these findings in Table 2. Exemplary narratives from participants by longitudinal housing status category are captured in Table 3.

Table 2. Summary of findings by longitudinal housing status.

	Supports for finding housing	Supports for maintaining housing	Interpersonal problems
Stable, independent housing	Used formal and informal support, with robust support from family and friends Supports performed emotional and instrumental functions	Used formal and informal supports, with a predominance of informal supports that carried out emotional functions Family, twelve-step sponsors and peers were often described as key informal supports	Experienced negative emotions (like loneliness) when facing interpersonal problems Sought formal and informal support to help with interpersonal problems
Sheltered housing	Predominantly used formal supports When informal supports were described, they were limited to friends (no participants in this group described support from family)	Predominantly used formal supports A few participants described informal support from peers, generally limited to instrumental support in housing settings	Few emotional consequences of interpersonal conflicts Relied predominantly on formal supports when seeking help for interpersonal problems
Unstable housing	Used formal and informal support, but support was superficial from both groups Some informal supports had negative valence, e.g., encouraged risky behaviors Very few descriptions of emotional support from formal or informal supports	Predominantly used formal supports A few descriptions of informal support from family, generally limited to instrumental support in housing settings	Experienced negative emotions when facing interpersonal problems Sought formal and informal support to help with interpersonal problems

Table 3. Exemplary narratives by longitudinal housing status.

	Supports for finding housing	Supports for maintaining housing	Interpersonal problems
Stable, independent housing	“My cousin [knew we were] looking for a house…she found a house that was basically a quick sell…she approached my fiancé and [said], ‘you guys, I've got a house.’”	“My wife [keeps] me sane…she's just great. I mean, she's my right hand…I can't say enough about her.”	“I just want [my girlfriend] to be more open-minded…just understanding what really was going on…I just don't think that she knows how to be as supportive as [I] need [her to be].”
Sheltered housing	“My [case] worker helped me with the process of checking out the apartment…like about the guidelines they give you in the manual [for the Housing First program].”	“My friend downstairs…he had a vehicle. So, whenever I needed to go to the laundromat, go grocery shopping, or if I need to find someplace that I wasn't familiar with…he's always help me out. He was great.”	“If need help or I don't know what to do…I always can call my [case] worker. That's what they're there for…reasoning with the [property managers]…he pushed the schedule around to make time [to help] me.”
Unstable housing	“…one of my closest friends was a former member of [the residential treatment program that gave me housing]…[he] supplied me with an understanding of what the program consists of and the rules, regulations, and policies of that program.”	“…my [case worker] helped me get myself settled, [always] telling me what I need to do, directing me to the right source.”	“I feel kind of lonely over there…people [probably weren't] corresponding to me because I probably wasn't reaching out…it's a two way street.”

Supports for Finding Housing

Participants universally described the strong role of supports in finding housing. Formal and informal supports most commonly provided instrumental assistance to find housing; emotional support was also frequently mentioned.

Some participants moved into the homes of significant others, family members, or friends; others relied on supports for instrumental functions, e.g., identifying rentals and negotiating terms. One participant described his mother's help finding, moving into, and furnishing his residence: “she's the one that took me [to find apartments]…the week before I was going to be discharged [from rehabilitation] she went looking for places for me to live.” He continued, “I don't think I could've found this place without my mom.” Formal supports helped other participants find housing. One participant described, “when I was released [from jail] it was a snag with me finding housing…I fortunately had support from [my mental health treatment program] staff to help me find a place that best suited my needs.”

Emotional support for finding housing entailed encouragement and help with decision-making; participants found it helpful to discuss options, e.g., moving in with significant others, and the implications of such with informal and formal supports. One participant sought help from his twelve-step sponsor to decide between several housing options: “…every time he talks about or advises [me]…he gives an example from his lifestyle, from his past and which means a lot. He's the man, you know…a very important person next to staying sober in my life at present and for maybe a long time to come.” Another described strong emotional support from his Housing First case manager during his housing search, “the lady is very comprehensive, and she comes out and talks to you and treats you like family…She's your big sister or your friend.”

In comparing participants by longitudinal housing status category, we noted differences in the use of supports to procure housing. Participants with stable housing relied on both formal and informal supports; informal supports were most robust in this group, e.g., offering the participant to move in, finding suitable residences, or helping with indispensable aspects of the housing process. For example, one participants' sister was a real estate agent and effectively negotiated rental terms. Participants with stable housing cited the value of both emotional and instrumental support.

In contrast, most supports identified by participants in sheltered housing were formal in nature; no participants in sheltered housing mentioned informal support from family in finding housing, contrasting from stably housed participants who usually had such support. Though persons with unstable housing described both formal and informal supports to find housing, such support had less depth; the roles of supports mentioned by both unstably housed and sheltered participants were generally limited to learning the existence of housing options. For example, one unstably housed participant identified a friend as important in helping him find a sober living home; his description of the friend's role was limited to: “…he just told me it was a decent place.”

Throughout the narratives of participants with unstable housing, there was less mention of emotional support. When emotional support was described, it was usually limited to enjoying others' company, e.g., a participant's grandson kept his “mind together and strong” through the challenges he faced finding housing. Moreover, some informal supports identified by unstably housed participants had negative valence; for example, on a peer's advice, one participant with cannabis use disorder feigned cocaine use to find housing: “…I may have done a lot of coke is what I tell the lady…they put me in a drug program [to find housing]…that was an experience because I never honestly considered myself a dope addict.”

Supports for Maintaining Housing

Participants also described instrumental and emotional functions of their supports in maintaining housing. Instrumental support included cleaning, cooking, mediating conflicts with landlords/roommates/other tenants, and managing finances. These functions were taken on by formal supports, such as case managers, informal supports, i.e., family and friends, or both. As one participant described the role of formal supports, “If I need help [with something in my apartment], I always call my [case manager]. That's what they're there for…they're still backing me up, [helping me] reason with the [property] manager.” In sheltered housing, peers were often credited with orienting participants to policies that helped them maintain housing in residential settings: “…my roommate…he told me no, you can't go up and down the hall while using your phone…you have the rules that are written and unspoken rules…[he] enlightened me to let me know what not to do.”

Many participants sought emotional support to help maintain housing. One participant said, “my [twelve-step] sponsor lived next door, so if I ever had any problems, I'd just knock on his door and he was there.” Others relied on formal supports for this function: “[my Housing First case manager] comes once every 90 days…She has an open-door policy, so I can call her and talk to her if I need to…it's been a long, hard road that turned out with [sic] a great experience.”

Similar to patterns seen in the use of supports to find housing, participants in stable housing employed the most informal supports (with instrumental and emotional roles) to stay housed. In particular, twelve-step sponsors and peers were commonly mentioned among the stably housed, but absent from narratives from participants with other housing outcomes. Family support for maintaining housing was more often described by participants with stable housing than unstably housed participants (who had a few examples of family support), and only mentioned by one participant with sheltered housing (who relied on peers when they did employ informal support). Participants in sheltered housing and unstable housing relied more on formal supports, with a preponderance of narratives focused on instrumental—as opposed to emotional—functions, e.g., working with property managers.

Interpersonal Problems

Participants across longitudinal housing status categories described similar types of interpersonal problems. In particular, transitional housing and rehabilitation settings (along the continuum of care) were laden with conflicts with peers. One participant described his experience in drug rehabilitation: “it's kind of a prison mentality…I felt very uncomfortable…if you're going to walk through them to go eat, you're going to have to put up with something.” Others described discrete conflicts, e.g., with roommates or friends/family they lived with. As one participant described his roommate: “By the time I'd come home he'd be drunk or passed out or in his room…it was like three or four times before I said I've got to move.”

Many participant narratives described discrimination experiences from formal or informal supports, due to race/ethnicity, sexual orientation, or Veteran status. One participant perceived discrimination from a case manager at his residential rehabilitation program: “…every time [she] sees me…[she] want[s] me to come to [her] office…and to me that's harassment.” Another participant described discrimination from peers in his residential rehabilitation program: “…they're combat Veterans, so they're very macho…so my lifestyle—being gay—it's very, very hard to live with them because they call [me] names.”

Though interpersonal problems were similar across longitudinal housing status categories, participants experienced these problems differently. Persons in stable and unstable housing commonly attached negative emotional consequences, e.g., isolation or loneliness, to interpersonal problems. As one participant said, “…it does get lonely. When I get lonely I just take walks…but then you gotta face reality and go home sooner or later.” In contrast, persons in sheltered housing rarely attached emotional consequences to interpersonal conflicts. For example, one participant described a conflict with his substance abuse counselor about twelve-step meeting attendance (a requirement of his housing program). He expressed no remorse or negative feelings about the conflict, describing: “I'm saying…you can't tell me what I need to do…I knew [sic] what I'm gonna do. I'm a grown man. They got their degrees and I'm a happy guy.”

In addition, participants in stable and unstable housing turned to both informal and formal supports to address conflicts, while participants in sheltered housing relied most heavily on formal supports. For example, a participant in unstable housing described turning to a friend from rehabilitation to help resolve a conflict with a roommate: “[My friend] has known him longer and I asked his advice on it.” In contrast, one sheltered housing participant described turning primarily to the assistant director of his housing program, stating: “you could approach her at any time, ask her a question, and she'll give you an answer.”

Consistency of Support Experiences

Many participants used consistent supports across their housing experiences, e.g., relied on family support to find housing, maintain housing, and manage interpersonal problems. Specifically, among participants who described help from family or friends (informal supports), about half relied on these supports to find and maintain housing. Similarly, among participants who received assistance from professional staff (formal supports), e.g., case managers, more than two-thirds relied on such supports to find and maintain housing. This trend was strongest for family support in the stably housed group; for peers or friends in the unstably housed group; and for formal supports in the sheltered group.

Moreover, persons who helped participants find and maintain housing were often also helpful in managing interpersonal problems. For example, one participant described his case manager as instrumental in helping him find and keep his apartment. He additionally described her importance in managing a conflict with his landlord: though he was having trouble getting his sink repaired, the situation improved when his “[case] worker came to help out…mediating between me and the [landlord].”

Discussion

These findings suggest associations between social support context and longitudinal housing status for adults with SMI, SUD, and a history of homelessness. To find housing and stay housed, participants in stable and independent housing described roles for informal (family and friends) and formal supports (case managers/other staff) of positive valence. These persons provided emotional and instrumental support (financial support was essentially absent), such as with apartment searches. In contrast, participants in sheltered housing described relying primarily on formal supports, particularly for instrumental tasks. Unstably housed participants discussed using informal and formal supports; some of these relationships were more superficial or of negative valence. Among participants with unstable housing, some described supports as simply instrumental while others experienced emotional support. Interpersonal problems were prevalent across longitudinal housing status categories; in particular, many participants cited interpersonal difficulties in sheltered housing. Participants with stable and unstable housing sought help from informal supports (family or friends) to resolve problems; sheltered housing participants often relied on formal supports, e.g., staff, to assist with conflict resolution. Many participants described consistent supports across their housing experiences (finding housing, maintaining housing, and help with interpersonal problems).

Most importantly, these data challenge assertions that this population lacks social resources. (Eyrich & Pollio, 2003; Hwang et al., 2009; Nelson, Clarke, Febbraro, & Hatzipantelis, 2005) Rather, our findings are reminiscent of the multi-dimensional relationships between social support and physical/mental health; (Veiel, 1985) the quality and depth of supports, and by whom they are offered—and in what settings—influences health in different ways. Here, we suggest similar dimensionalities in the interactions between supports and housing status. Better longitudinal housing status categories were associated with support from family and friends and reliance on social resources for emotional, as well as instrumental support. Moreover, participants with better longitudinal housing status described less social undermining, (Eyrich & Pollio, 2003; Skeem, Eno Louden, Manchak, Vidal, & Haddad, 2008; Vinokur, 1993) i.e., fewer negative supports. These associations may reflect differential consumer preferences, needs, resources, and/or abilities to access supports.

Additionally, these findings parallel prior work suggesting complex, distinct, and important roles of both formal and informal supports for consumers with SMI. (Pahwa et al., 2014) These consumers in sheltered housing may rely exclusively on formal supports in certain housing contexts—or across all housing experiences—it is uncertain if informal supports are displaced by formal supports or if consumers lacked supports altogether prior to sheltered housing. Regardless, it may be disadvantageous for formal supports to displace informal supports, as informal supports more commonly filled certain roles, e.g., emotional support.

These findings are particularly relevant in light of an alarming (25% a year in VA) (O'Connell et al., 2012) rate of exits from Housing First programs (often resulting in transition to settings like incarceration or street homelessness, (Gabrielian, Burns, et al., 2015b) along with poor social integration among formerly homeless consumers who do gain housing. (Tsai et al., 2012) Our data suggest that consumers with a history of homelessness may benefit from evidence-based practices that are effective among persons with SMI but uncommonly implemented within services for persons who have experienced homelessness. Among such practices are social skills training, (Bellack, Mueser, Gingerich, & Agresta, 2004) which improves social skills and functioning; social cognition training, which trains consumers to better perceive and use social information, interpret social cues, and interpret social events;(Kurtz & Richardson, 2012) and family-focused treatments, e.g., psychotherapy or psychoeducation, which can help individuals improve relationships with family members (who provide informal support). (Dixon et al., 2009) These interventions may help consumers build or repair their informal support networks, differentiate between positive vs. negative relationships, and ultimately influence longitudinal housing status and quality of life outcomes.

This study has limitations. First, these data are from a small sample of male Veterans in an urban area who have experienced homelessness; men vs. women, Veterans vs. non-Veterans, and urban vs. rural populations may employ different social supports in housing transitions. These data may not extrapolate well to the broader homeless population with SMI and SUD. Second, these data were cross-sectional, relying on retrospective recounts about the use of social supports in housing contexts. If participants were queried about their social supports at the time of apartment search, for example, described supports might differ from these data. Third, individuals in sheltered housing can more easily access formal supports; often, e.g., in rehabilitation, these supports are present on-site. This differential accessibility may influence our findings. Related to such, it is unclear if social support differences by group are in themselves responsible for the different housing outcomes, or rather a result of differential opportunities created in contrasting housing environments. However, the consistency of supports described—by participant, across housing experiences—suggests that consumers' existing social supports may influence their housing outcomes. Moreover, given our small sample size, we were unable to examine subsample differences by diagnoses or symptom severity. This study would benefit from replication in a larger and more diverse sample, with data collected prospectively, as individuals find and retain housing.

Conclusions and Implications for Practice

Current housing services help many consumers exit homelessness. (McGuire et al., 2010; O'Toole & Pape, 2015; Stergiopoulos et al., 2015) However, these services have performance gaps: many consumers prematurely disengage from housing programs (O'Connell et al., 2012) and others remain socially isolated despite gaining permanent housing. (Tsai et al., 2012) For persons with SMI, SUD, and a history of homelessness, our findings suggest the importance of social support—in its various dimensions, types, and relationship contexts—in finding and retaining housing. Within existing housing programs, further research is needed to identify, tailor, and implement practices that can help this vulnerable population improve its social resources.