Housing outcomes among homeless-experienced Veterans engaged in vocational services

Abstract

Though unemployment and homelessness are closely intertwined, vocational services are rarely prioritized for homeless-experienced individuals engaging in housing services. Our goal was to examine associations between vocational service use and housing outcomes among homeless-experienced Veterans engaged in permanent supportive housing. We obtained data from Veterans Health Administration (VA) medical record and homelessness registry data for homeless-experienced Veterans engaged in VA Greater Los Angeles’ permanent supportive housing program from 10/2016–9/2017 (n=1,200). We used multivariate logistic regression to examine whether vocational service use was associated with housing attainment and/or premature permanent supportive housing exits. We found that Veterans in permanent supportive housing who used vocational services were more likely to attain housing (OR=2.52, p-value<0.001) than their peers who did not use these services. There were no between-group differences in the odds of premature exits from the permanent supportive housing program (OR=1.92, p-value=0.425). Our study suggests that, among homeless-experienced Veterans engaged in permanent supportive housing programs, those who use vocational services potentially may be more likely to attain housing. However, future research can better elucidate the pathways underlying vocational service use and housing outcomes for individuals in permanent supportive housing programs. Greater integration of vocational services and permanent supportive housing programs, and encouragement of vocational service use may enhance housing outcomes among permanent supportive housing participants.

Introduction

Though unemployment and homelessness are closely intertwined (Metraux et al., 2017), vocational services are rarely prioritized for homeless-experienced individuals engaging in housing services. This gap stems from common beliefs among providers that this population is not ready and able to work: e.g., they cannot stay sober; they lack skills; or that they are not reliable (Shaheen & Rio, 2007). Yet research, though limited, suggests that employment may support exits from homelessness (Burt, 2012; Johnson et al., 2015; Nilsson et al., 2019; Shaheen & Rio, 2007). For example, demonstration projects including the U.S. Department of Labor’s Job Training for the Homeless Demonstration Program, the Employment Intervention Demonstration Project, and Los Angeles’ Homeless Providing Opportunity illustrate that coupling job training and employment programs with supportive services and housing, as well as rapid access to jobs after program enrollment may support exits from homelessness (Burt, 2012; Shaheen & Rio, 2007).

The Veterans Health Administration (VA) offers robust housing and vocational services, providing an opportunity to examine relationships between vocational service use and housing outcomes for Veterans who have experienced homelessness (VEH). VA’s permanent supportive housing program, the U.S. Department of Housing and Urban Development-VA Supportive Housing (HUD-VASH) program, is the linchpin of VA’s plan to end Veteran homelessness. HUD-VASH provides subsidized permanent housing and field-based supportive services for VEH (U.S. Department of Housing and Urban Development, 2013). It also links VEHs to other VA services, including vocational services (e.g., supported employment, job training).

However, there is limited research on the relationships between vocational service use and housing outcomes among VEH receiving VA homeless services. Several studies have found positive housing outcomes among VEH participating in Individualized Placement and Support, Supported Employment, which is a vocational program where that focuses on rapid placement into competitive employment, ongoing support, individual preferences and tailored job support, integration of employment and clinical services, and openness to anyone willing/interested in work (LePage et al., 2021; Robert A. Rosenheck & Alvin S. Mares 2007). HUD-VASH—the largest permanent supportive housing program in the United States and one of few permanent supportive housing programs that also offers vocational services—is a unique setting to further study relationships between vocational services and housing outcomes. Our study examined whether vocational service use was associated with better housing outcomes among VEH engaged in HUD-VASH. We hypothesized that VEHs who used vocational services would have better housing outcomes compared to those who did not used vocational services.

Methods

Setting and sample.

We conducted this observational study among VEH enrolled in VA Greater Los Angeles’ (VAGLA) HUD-VASH program, the largest HUD-VASH program in the nation. HUD-VASH employs a “Housing First” model to provide Veterans with subsidized housing, including field-based case management, housing navigation, and supportive services, including non-mandated linkages to care and VA’s vocational services. Typically, after Veterans are enrolled in the HUD-VASH program, they receive case management and referral to a local public housing authority (PHA, in this case, run by the city or County of Los Angeles) to apply for a HUD-VASH voucher. Vouchers are assigned in the order that Veterans enroll in the program; assignments occur when the PHA has an available voucher and a case manager is available to case manage the Veteran. While Veteran-level characteristics do not determine prioritization of vouchers, we do note that Veterans with fewer vulnerabilities often find and rent housing faster, due to past experiences with housing, better credit histories, and overall higher levels of social functioning.

Vouchers equate to housing subsidies, with Veterans contributing between 30–40% of their monthly income—often from disability payments—to their rent (Gabrielian et al., 2014). If a Veteran identifies a rental unit (with or without the assistance of a housing navigator or case manager), the public housing authority will conduct an inspection to determine if the selected unit is voucher-eligible (i.e., meets specified safety requirements and is priced at a rate deemed appropriate given local comparable units) (Ann Elizabeth Montgomery et al., 2016; Los Angeles County Develoment Authority). Our study included both project-based (dedicated multi-unit public housing developments with multiple veterans living in the same building, with co-located supportive services) and tenant-based housing (subsidized rental units located in the community and are rented through the private rental market) (Montgomery et al., 2020). Of note, though Veterans with HUD-VASH vouchers have potential access to housing, i.e., a subsidy that can assist with rental payments, some vulnerable Veterans in the program never attain housing; others who attain housing struggle to retain it.

VAGLA offers HUD-VASH participants a variety of vocational rehabilitation intervention strategies and programming (described below) based on their unique competitive employment goals. Veterans can access these services at any point after enrollment in HUD-VASH. VAGLA’s vocational rehabilitation team will conduct an assessment with any Veteran who is interested in accessing vocational services to determine the program or services that best fit Veterans’ needs. An important component of VAGLA’s vocational services is that an integrated, unified team works with HUD-VASH to provide a variety of services and programs that may encompass different philosophies and approaches to vocational rehabilitation. This allows for a flexible approach to meet Veterans’ vocational needs as they progress through their recovery (e.g., Veterans may switch between programs as they progress through their recovery process). VA Greater Los Angeles Healthcare System Institutional Review Board approved this study.

We used VA’s Homeless Operations Management and Evaluation System (HOMES)—an administrative dataset of VA homeless service utilization—and VA electronic medical record to identify VEH enrolled in VAGLA’s HUD-VASH program between October 2016 to November 2017 (n=1,200). HOMES includes data on dates of HUD-VASH program entry and exit, reasons for exit, and some demographic, military and homelessness history, and social support information collected during initial enrollment into the program. We obtained data on HUD-VASH program entry dates, HUD-VASH program outcomes, demographics, homelessness history, and relationship status of VEHs in HUD-VASH from HOMES. We gathered data for one-year from each VEH’s HUD-VASH enrollment date to identify VEHs who attained housing (n=688), and, in this VEH subsample who attained housing, one-year from each VEH’s move-in date to identify those with premature HUD-VASH exists.

Variables.

Our dependent variables were: 1) a dichotomous indicator of attaining housing through HUD-VASH within one-year of program entry; and 2) among those who attained HUD-VASH housing, a dichotomous indicator of premature program exit (noncompliant with HUD-VASH case management, evicted from HUD-VASH housing, unhappy with HUD-VASH housing, or could not be located) (VHA Homeless Program Office, 2019) within one-year of moving in.

Our independent variable was a dichotomous indicator of whether VEHs participated in the following vocational services during the one-year after HUD-VASH enrollment for the full sample, and during the one-year after move-in to HUD-VASH housing for the subsample who attained housing:

• Compensated Work Therapy (CWT) (Veterans Health Administration, 2020): provides evidence-based and evidence-informed vocational services for Veterans recovering from chronic mental illness or substance use disorder, regardless of housing status. CWT includes three subprograms: Transitional Work (a pre-employment program that places Veterans in real-live work assignments, often at VA, for a time-limited basis); Community-Based Employment Services (job placement assistance as a bridge to competitive employment in the community); and Supported Employment (evidence-based practice based on the Individual Placement and Support model (Bond, 2004) that enables competitive employment for persons with serious mental illness). CWT requires active engagement in VA behavioral health treatment.

• Homeless Veteran Community Employment Services (HVCES) (U.S. Department of Veterans Affairs, 2020): includes a range of VEH-tailored employment services, including job training, social and resource support, connections to job-related resources, and VA and community employment opportunities through collaboration with VA and non-VA partners; importantly, HVCES uses a zero-exclusion approach i.e., these services are available to any VEH in HUD-VASH who indicates an interest in employment; at VAGLA, employment specialists are embedded within HUD-VASH teams,

We controlled for potential confounders, including age, gender, race/ethnicity (non-Hispanic Black, non-Hispanic White, Hispanic, Other), history of chronic homelessness (ever experienced chronic homelessness [i.e., being homeless for 1+ years (Department of Housing and Urban Development, 2015)] vs. never), relationship status (in a relationship vs. not), prior serious mental illness diagnosis within 2-years prior to HUD-VASH enrollment date (any/none), and substance use disorder diagnosis during the observation period (any/none). We identified serious mental illness with ICD-10 codes based on VA’s Strategic Analytics for Improvement and Learning Value Model quality metrics (ICD-10 code list in Appendix) (U.S. Department of Veterans Affairs, 2021), and substance use disorder from ICD-10 codes (code list in Appendix). These variables were selected because they are associated with housing outcomes among VEH (Gabrielian et al., 2014; Montgomery et al., 2016).

Statistical analysis.

We calculated means and proportions for key variables in the full VEH sample enrolled in HUD-VASH, and the VEH sub-sample who received housing within one-year of their enrollment date. We fit two multivariate logistic regression models to determine 1) adjusted associations between vocational service use and housing attainment in the full sample, and 2) among those who attained housing, adjusted associations between vocational services use and premature HUD-VASH exits, adjusting for control variables above. Analyses were conducted in Stata version 15.1 (StataCorp, College Station, TX).

Results

In our sample of 1,200 VEH enrolled in VAGLA’s HUD-VASH program from 10/2016–9/2017, 11% were female, 55% had a history of chronic homelessness, and 87% were not in a relationship (Table 1). Non-Hispanic Blacks comprised the largest racial/ethnic group (39%), followed by non-Hispanic Whites (32%), and Hispanics (18%). The average age was 52 years (standard deviation = 14.7 years). Only 9% of the sample used VA vocational services, averaging 6.4 visits (standard deviation [SD]=7.1 visits), and 57% attained housing during the year following their HUD-VASH enrollment date. Among this subsample of VEHs who attained housing, 8% had previously used VA vocational services, with an average of 7.6 visits (SD=7.0); 8% had a premature exit from the HUD-VASH program within a year of their move-in date. Other characteristics of this subsample of VEH who attained housing were similar the to the main sample of VEHs enrolled in HUD-VASH.

Table 1. Sample characteristics among VEH enrolled in the VA Greater Los Angeles HUD-VASH program from 10/2016–9/2017.

	Mean or %
	Total Sample (n = 1200)	Moved in (n = 688)
Demographic Characteristics
Age, years (SD)	51.6 (14.7)	52.2 (14.5)
Gender, %
Male	88.8	88.5
Female	10.6	10.8
Other/missing	0.6	0.7
Race/ethnicity, %
Non-Hispanic White	31.9	29.1
Non-Hispanic Black	39.4	41.0
Hispanic	17.9	18.6
Other1	3.8	4.4
Missing	6.9	7.0
Prior episode of chronic homelessness2, %	54.7	54.9
Relationship status, %
Not in a relationship3	87.4	86.9
In a relationship4	9.5	10.2
Missing	3.1	2.9
Serious mental illness diagnosis in the past 2 years, %	16.9	15.1
Concurrent substance use disorder diagnosis, %	40.2	36.2
Vocational service use within 1-year of HUD-VASH enrollment
Any vocational service use, %	6.4	8.3
Number of visits, mean (SD)	4.4 (4.6)	7.6 (6.9)
Housing Outcomes
Attained Housing, %	57.3	N/A
Premature HUD-VASH Exit, %	N/A	8.0

Notes:

¹-Other racial/ethnic groups include non-Hispanic American Indian/Alaskan Natives; Asians; Native Hawaiians/Other Pacific Islanders;

²-chronic homelessness defined as being homeless for 1+ year;

³- divorced, never married, separated, widowed;

⁴-committed relationship/partnered, married

In multivariate analyses, VEH who used VA vocational services within a year of program entry had 152% higher odds of attaining housing than those who did not use these services (odds ratio [OR] = 2.52, 95% Confidence Interval [CI]: 1.59, 4.01), p-value< 0.001) (Table 2). Additionally, compared to non-Hispanic White Veterans, non-Hispanic Black (OR = 1.33, 95% CI: 1.00, 1.76, p-value = 0.047) and Hispanic Veterans (OR = 1.40, 95% CI: 0.98, 1.98) were more likely to attain housing, with the latter group trending towards statistical significance (p-value = 0.062). Additionally, the relationship between several covariates and housing attainment also reached non-significant trend levels: Veterans with a previous SMI diagnosis were less likely to attain housing compared to those without a prior SMI diagnosis (OR = 0.73, 95%CI: 0.53, 1.00, p-value = 0.050). Veterans who had a substance use disorder diagnosis within one year of program entry were also less likely to attain housing (OR = 0.73, 95%CI: 0.63, 1.03, p-value = 0.081. The odds of attaining housing increased by 1% for each year increase in age (OR = 1.01, 95% CI: 1.00, 1.02, p-value = 0.079).

Table 2. Adjusted associations between use of VA vocational services and HUD-VASH housing outcomes among VEH enrolled in the VA Greater Los Angeles HUD-VASH program from 10/2016 to 9/2017.

	Attained Housing (n = 1200)		Premature HUD-VASH Exit (n = 688)
	OR (95% CI)	p-value	OR (95% CI)	p-value
VA vocational services
No use	1.00 (ref)		1.00 (ref)
Any use	2.52 (1.59, 4.01)	<0.001**	1.92 (0.39, 9.60	0.425
Age	1.01 (1.00, 1.02)	0.079*	0.99 (0.95, 1.03)	0.711
Gender
Female	1.00 (ref)		1.00 (ref)
Male	1.11 (0.75, 1.64)	0.608	1.78 (0.34, 9.45)	0.336
Race/ethnicity
Non-Hispanic White	1.00 (ref)		1.00 (ref)
Non-Hispanic Black	1.33 (1.00, 1.76)	0.047**	0.83 (0.24, 2.90)	0.767
Hispanic	1.40 (0.98, 1.98)	0.062*	0.85 (0.19, 3.93)	0.840
Other1	1.81 (0.94, 3.46)	0.074*	-- 5
Prior episode of chronic homelessness2
Never	1.00 (ref)		1.00 (ref)
1+ episode	0.86 (0.57, 1.28)	0.448	0.58 (0.15, 2.26)	0.431
Relationship status
Not in a relationship3	1.00 (ref)		1.00 (ref)
In a relationship4	1.20 (0.79, 1.78)	0.400	--5
Serious mental illness diagnosis in the past 2 years
No diagnosis	1.00 (ref)		1.00 (ref)
Any diagnosis	0.73 (0.53, 1.00)	0.050*	0.90 (0.19, 4.31)	0.898
Substance use disorder diagnosis
No diagnosis	1.00 (ref)		1.00 (ref)
Any diagnosis	0.80 (0.63, 1.03)	0.081*	1.41 (0.44, 4.49)	0.563

Notes:

^*denotes p<0.10

^**denotes p<0.05;

¹-Other racial/ethnic groups include non-Hispanic American Indian/Alaskan Natives; Asians; Native Hawaiians/Other Pacific Islanders;

²-chronic homelessness defined as being homeless for 1+ year;

³- divorced, never married, separated, widowed;

⁴- committed relationship/partnered, married;

⁵- insufficient sample size and variation to estimate this association

Individuals with missing data were included in the model but results are not presented.

We did not find an association between vocational service use and premature exits from HUD-VASH among VEH who attained housing (OR = 1.92, 95%CI: 0.39, 9.60, p-value = 0.425). None of the other covariates in our analyses were associated with premature exits from HUD-VASH.

Discussion

We found that among VEH enrolled in a large HUD-VASH program, VA vocational service was associated with increased likelihood of attaining housing within one-year of program enrollment. However, VA vocational service use was not associated with premature HUD-VASH exists among those who attained housing.

Prior research has identified both Veteran and program characteristics associated with HUD-VASH enrolled Veterans attaining housing, including not having a service connected disability or a diagnosis of schizophrenia or other psychosis, engagement with HUD-VASH and supportive services, a strong and trusting relationship with HUD-VASH case managers and program staff, and case manager assistance with the housing search (Ann Elizabeth Montgomery et al., 2016). We are encouraged that our findings suggest a relationship between vocational services and housing attainment among HUD-VASH participants. This finding aligns with limited research on the benefits of providing employment support and opportunities to homeless-experienced individuals (Burt, 2012; Gray et al., 2017; Poremski et al., 2016). However, our findings cannot be interpreted causally (i.e., that accessing vocational services leads to housing attainment), because we did not account for the temporal ordering of vocational service use relative to housing attainment. While these services may facilitate housing attainment for some Veterans through increased income from successfully gaining employment (e.g., increased funds may enable on-time rental payments, via fewer competing needs for food/other belongings), it is also possible that Veterans who use these services have other characteristics to help them attain housing (e.g., greater motivation, further in their recovery).

Despite this, this exploratory research suggests a potentially important benefit of encouraging greater vocational service use among homeless-experienced individuals, and an area that can benefit from more research, particularly to elucidate the mechanisms that underly the observed vocational service use and housing attainment relationships in our analysis. Understanding the underlying mechanisms can guide permanent supportive housing programs in targeting and tailoring who should be encouraged to use vocational services. Future studies may consider conducting analyses in other populations and housing programs, considering the temporal ordering, accounting for selection issues between those who do and do not participate in vocational services, and including employment outcomes (i.e., whether use of services lead to gaining employment). If future studies find few differences between those who do and do not participate in these services, then this might suggest that permanent supportive housing programs that also have vocational services, such as VA HUD-VASH programs, should encourage widespread participation in these programs. Alternatively, future studies may find that those who participate in vocational services are different from those who do not, which would suggest that permanent supportive housing programs should motivate homeless-experienced individuals to participate in these programs as well as identify and target individuals who are most likely to benefit from vocational services. Understanding whether use of vocational services leads to better employment outcomes, such as gaining and retraining employment, can further elucidate underlying mechanisms, such as whether participation in these programs leads to better housing outcomes via increased income from employment.

Vocational services can help homeless-experienced individuals overcome these challenges (e.g., age, disability, homelessness stigma) (Fleury et al., 2021) to finding employment. VAGLA HUD-VASH vocational services match Veterans’ skills with employer needs, and foster relationships with employers. VAGLA’s vocational services also have a zero-exclusion policy: this means that any VEH who is interested can participate in vocational services, regardless of their recovery stage, clinical status, or prior work experience. However, providers often underutilize vocational services among homeless-experienced individuals, in part, because they believe that VEHs are not “work-ready” (Shaheen & Rio, 2007).

The lack of association between VA vocational service use and premature HUD-VASH exits was surprising given our finding of increased housing attainment and existing literature of the benefits of employment services (Burt, 2012; Gray et al., 2017; Poremski et al., 2016). It is possible that while use of employment services benefit some VEHs, others may struggle to find or maintain employment after using vocational services and then become discouraged, leading to mental health issues (e.g., depression) or substance use, and eventually, premature program exits, resulting in an overall lack of association.

We also found that some demographic characteristics were associated with better housing outcomes. Specifically, in our sample, non-Hispanic Black and Hispanic VEHs were more likely to attain housing compared to their non-Hispanic White counterparts. We were surprised by this finding, given well-documented–albeit more subtle–racial and ethnic discrimination in the rental housing market (U.S. Department of Housing and Urban Development). These experiences may make it more difficult for non-Hispanic Black and Hispanic VEHs to attain tenant-based housing, since these units are rented through the private rental market. Prior research on housing outcomes among racial/ethnic minoritized VEHs is mixed. One study of Veterans with substance use histories found that HUD-VASH had less of an impact on housing outcomes, including reducing days of homelessness, among non-Hispanic Black Veterans compared to non-Hispanic White Veterans (O’Connell et al., 2012). In another study, conducted among a national cohort of HUD-VASH enrolled VEHs, Black Veterans were more likely to attain HUD-VASH housing, while Hispanic Veterans were less likely to attain housing compared to White and non-Hispanic Veterans, respectively (Montgomery et al., 2016). It is possible that HUD-VASH’s case management and supportive services – particularly housing search assistance or triaging to project-based settings with on-site support – can help racial/ethnic minorities individuals overcome barriers to finding housing, such as navigating racial/ethnic discrimination in the housing market, whereas White counterparts may navigate the housing search independently. Additionally, Black VEHs may receive more social support from their families (Novacek et al., 2020) compared to other racial/ethnic groups, which may contribute to improved housing outcomes among in this group.

Additionally, older Veterans were more likely to attain housing and less likely to have premature HUD-VASH exits than younger Veterans, although the effect size is small and may have limited clinical relevance. This aligns with prior studies conducted in national cohorts of VEHs in HUD-VASH that found that older Veterans were both more likely to attain HUD-VASH housing (Ann Elizabeth Montgomery et al., 2016), and less likely to be evidenced from HUD-VASH housing compared to younger VEHs (Montgomery et al., 2017). We also found that VEHs who had a a serious mental illness or substance use disorder diagnosis may be less likely to attain housing, which is consistent substance use disorders and serious mental illnesses being risk factors for homelessness (Greenberg & Rosenheck, 2010). While permanent supportive housing may help to mitigate these risk factors (Tsai et al., 2014), VEHs with these conditions typically have worse housing outcomes (Montgomery et al., 2017). Of note, our analysis accounted for having a substance use disorder diagnosis, but this likely underestimates the prevalence of substance use disorder in this sample, as some individuals may not use VA services for substance use disorder.

Our study had several limitations. First, there is the potential for omitted variable bias, particularly VEH characteristics (prior employment history, internal motivation) that may confound the relationship between vocational service use and housing outcomes. Second, we cannot make causal inferences from our data due to our use of cross-sectional data and potential selection bias in who chose to use vocational services. Third, we do not have information about whether VEHs eventually attained employment. Fourth, our findings may have limited generalizability to smaller, less well-resourced and staffed HUD-VASH programs. Finally, our independent variable did not capture more nuanced aspects of employment service use, such as the number of visits and duration of using vocational services. Using measures that better capture the quality and quantity of engagement with vocational services may potentially result in even stronger associations among individuals who were most engaged with vocational services. By dichotomizing vocational service use in our analysis, we may have biased our results towards the null, as those included in the any vocational use group also includes individuals who may have only had one or a few encounters. Of note, since these data were collected, VAGLA has increased vocational services embedded within HUD-VASH through HVCES; our analysis does not capture these improvements. We were inadequately powered to examine CWT and HVCES separately since HVCES comprised a small proportion of encounters (27%) at the time when the data were collected. Future studies can examine housing outcomes for CWT and HVCES participants separately and identify the vocational services most appropriate to meet a particular VEH’s needs.

Our findings suggest the value of providing vocational services to homeless individuals in permanent supportive housing programs. However, only a small proportion of HUD-VASH enrolled-VEH used them, so it will be important to explore factors that influence their use and which VEHs may benefit the most from participation in vocational services. Embedding vocational services in permanent supportive housing, and identifying and articulating benefits derived from vocational services – including those beyond housing and employment stability, such as social integration (Gray et al., 2017; Shaheen & Rio, 2007) – can potentially expand use of these services.

Public Policy Relevance:

Vocational services are rarely prioritized for homeless-experienced individuals engaged in housing services. Our study, conducted among homeless-experienced Veterans engaged in VA’s permanent supportive housing program, suggests that Veterans who were engaged with vocational services were more likely to attain housing. Encouraging permanent supportive housing program participants to use vocational services may enhance their housing outcomes.