6

Impact of Permanent Supportive Housing on Families and Youth

Previous chapters have described the effects of permanent supportive housing (PSH) for individuals with mental illnesses and substance abuse problems and described the potential benefits of housing for individuals with housing-sensitive health conditions. There is much less evidence about the extent to which PSH is a useful intervention for families, youth, and older adults experiencing homelessness. The limited evidence addressing where PSH has been used for other populations who have not experienced chronic homelessness is noted. Described below, and consistent with findings from Chapter 4, little evidence was found on addressing how PSH affects health for families, youth, and older adults. Most studies are descriptive, and few include health outcomes.

FAMILIES

PSH for families is widely used, but little studied. According to the Annual Homeless Assessment Report (AHAR) by the Department of Housing and Urban Development (HUD, 2016b) to Congress, there are nearly 123,000 PSH beds for people in families in the United States, 36 percent of the total PSH stock available across populations, and 30 percent of the total bed count in homeless programs serving families. It remains unclear to what extent this model is superior to others for serving families, and whether there are subgroups of families who might benefit more than others.

A literature review of housing and services for families by Bassuk and Geller (2006) found observational studies with no comparison groups that suggested that families who received housing subsidies with case management, including some deemed to be high risk by different criteria, were likely to be stably housed over follow-up periods of 1 to 2 years. In two studies in the same city, families who received housing subsidies without associated services were also more likely than those who did not receive subsidies to attain housing stability. In the one comparison-group study providing evidence that subsidized housing with intensive case management was superior to subsidized housing without intensive services, housing type was confounded with intensity of services. There were no randomized controlled studies found and few studies that looked at outcomes beyond housing stability. A follow-up systematic review by Bassuk and colleagues (2014) examined evidence of the effectiveness of housing interventions on ending family homelessness between 2007 and 2013. They found one randomized controlled trial (RCT), described below, and six observational studies, some with

multiple follow-up points. Of those six studies, only one addressed health outcomes associated with PSH (Building Changes, 2011). In the program’s first 6-month evaluation, it was noted that families had fewer service needs than at baseline and were more likely to have a routine source of care (93 percent of household heads and 100 percent of children, up from 78 percent and 93 percent at study entry). Adults were less likely to have moderate or severe levels of anxiety (38 percent, down from 63 percent at baseline), perhaps because 75 percent of families received mental health services, and school absences among children decreased. In analyzing the results of the systematic review more broadly, Bassuk and colleagues concluded that as it pertains to families experiencing homelessness, “research aimed at the intersection between housing and health care is especially needed” (2014, p. 472).

More recent work has largely continued the tradition of descriptive studies of families using supportive housing or suggests good outcomes for those who successfully complete programs without considering those who drop out or are excluded. Studies rarely examine the extent to which all family members (including children) might benefit. A systematic review by Speirs et al. (2013) aimed to identify interventions to improve (psychological and physical) health outcomes in homeless women in the United States and across the globe. The authors found that most interventions comprised education sessions in group settings that aimed to improve individuals’ knowledge about risk-taking sexual behavior and ways to mitigate physical abuse within domestic violence situations. Of the six studies reviewed, none focused specifically on the linkage between PSH and physical or mental health.¹

Studies that include children in families or that address the impact of PSH on health in youth are often limited by small sample size or lack child- and/or youth-specific performance measures to adequately monitor health outcomes. Others, such as the qualitative study of supportive housing in the health of 10 HIV-positive mothers and their children (Quinn et al., 2015), focus primarily on parental health and social needs without necessarily connecting these to child outcomes. Gewirtz and colleagues (2008) described the psychosocial risks and health status of 454 children living with their families in 17 supportive housing communities in the Minneapolis/St. Paul, Minnesota, metropolitan area. They suggest that insofar as supportive housing gives families residential stability and case management services, access to routine, basic, physical health services should be facilitated. In this study, child service staff in the facilities were asked to complete a 37-item child needs assessment survey, asking for parental involvement only if they were unsure of the answers. Results showed that more than 95 percent of children had health insurance, yearly physical exams, and up-to-date immunizations, although a causal link between supportive housing and these results could not be established. Over 75 percent had regular access to dental care, vision, hear-

___________________

¹ For the list of the six studies cited, see Speirs et al. (2013, pp. 1085–1086).

ing, and lead-level evaluation. Asthma rates were higher than the general population but lower than comparable low-income populations. The greatest concern was in children’s mental health, where significant numbers of children were reported to be depressed or anxious, have behavioral issues, or have a diagnosed learning disability. These issues increased significantly in children age 12–19 compared to those age 1–11 years, and children with a mentally ill parent were 1.8 times more likely to have a diagnosed or undiagnosed emotional or behavioral problem. Complicating the mental health challenges noted among children was the general absence of available mental health services (Gewirtz et al., 2008).

There is also some question as to whether permanent housing subsidies without dedicated services attached to them are sufficient for most families. The Family Options Study is a 12-site RCT that examined different housing and service programs for families recruited after a stay of 7 days in a homeless shelter (Gubits et al., 2015, 2016). The 2,282 families were randomized to offers of (1) permanent housing subsidies (usually in the form of a Housing Choice Voucher that limits housing costs to 30 percent of income) without additional services; (2) temporary rapid re-housing subsidies with low-intensity case management (averaging 36 families per case manager) focused on housing and self-sufficiency; (3) transitional housing with higher-intensity case management (averaging 20 families per case manager) and extensive psychosocial services; or (4) usual care, starting with the shelter at which they were recruited. All families were free to seek additional services in their communities. Analyses were on an intent-to-treat basis, including all families offered a particular intervention, irrespective of the type of housing they actually took up. Offers of housing subsidies dramatically reduced homelessness and doubling up with other households because the family could not find or afford a place of their own at both the 20-month and the 37-month follow-up points; offers of transitional housing decreased homelessness more modestly during the period when some families remained in transitional housing programs; and offers of rapid re-housing led families to leave shelter more quickly but had no other effect on housing outcomes. With respect to health outcomes, offers of housing subsidies reduced adult psychological distress at both points and alcohol dependence or drug abuse by a quarter (4.5 percentage points) at 20 months only, and children’s behavioral problems as reported by mothers at 37 months only. Housing subsidies reduced recent intimate partner violence by over half (6.8 percentage points) at 20 months and over one-third (4 percentage points) at 37 months. Neither of the other interventions affected these outcomes at either time point, and no intervention affected global reports of adult or child health. Thus, permanent housing subsidies alone had more impact on the psychosocial problems that can sometimes cause homelessness than did psychosocial services in time-limited transitional housing programs.

Gubits and colleagues (2015, 2016) also examined whether the interventions were differentially effective for two subgroups of families: those with more psychosocial challenges (a count of 9 issues such as intimate partner violence, poor health, psychological distress, and substance problems reported at the study outset), and those with more housing barriers (a count of 15 issues such as lack of

money to pay rent, lack of employment, poor credit, or past evictions). Although statistical power for these tests was limited, the patterns of scattered differences across outcomes did not exceed what would be expected by chance. Over the 37-month period, families in the permanent housing subsidy group cost only 9 percent more than usual care, because costs of the subsidies were offset by greater costs for shelter and transitional housing programs in the usual-care group.

One criticism of the study is that, because of low program take-up among families offered transitional housing (53 percent) and rapid re-housing (59 percent) compared to those offered permanent subsidies (88 percent considering all forms of subsidy), the experimental contrast was weaker in these comparisons. Crossover from usual care to the assigned intervention was largely equivalent for the three active interventions (largest, at 38 percent for the permanent subsidy group). However, at 37 months there were no significant differences in homelessness and doubling up among families assigned to rapid re-housing or transitional housing by whether or not they took up the intervention. Among families assigned to permanent housing subsidies, those who took up the intervention were significantly less likely to be homeless or doubled up (Gubits et al., 2016).

The study did not examine PSH, and thus provides no evidence about the extent to which psychosocial services in addition to permanent subsidies might enhance their effects. An observational study in Philadelphia suggests that some families may need such services. Culhane et al. (2011) found that both inpatient behavioral health services and foster care placements among families, which were reduced during the period that families were in shelter, rebounded afterward, regardless of whether families were discharged to permanent housing subsidies. The rebound was smaller, but not significantly smaller, for the relatively small group of families who received permanent subsidized housing placements. The authors suggest screening for behavioral health and foster care needs while families are engaged with shelters or transitional housing and linkages to community services afterward.

One RCT compared two housing and service models for families in which the mother had a diagnosable mental health or substance abuse condition. One was a Family Critical Time Intervention (FCTI) in which families received more intensive case management (maximum ratio 12 families to 1 caseworker) from a single worker for a 9-month period as they moved from shelter to housing. This structured program of linking families to community services was compared to treatment as usual, which involved less intensive casework with different workers in shelter (ratio 24:1) and community (ratio 50:1). Families in both groups received affordable housing, but those in the FCTI group received it about 3 months faster without meeting criteria for “housing readiness.” Psychiatric symptoms improved substantially over time in both groups, and there was no difference between them. The authors pointed out that among families selected for high levels of distress, distress may decrease over time regardless of services received (Samuels et al., 2015). Children in the FCTI group showed modest improvements

in behavior problems at home and at school relative to children in the treatment-as-usual group (Shinn et al., 2015), but again, all children improved over time.

FAMILIES INVOLVED IN THE CHILD WELFARE SYSTEM

The considerable overlap among families who experience homelessness and families who are involved in the child welfare system (Culhane et al., 2003; Harburger and White, 2004; Park et al., 2004; Rodriguez and Shinn, 2016) has led to calls for supportive housing for this group. Most studies remain descriptive. For example, Farrell and colleagues (2010) assessed 1,720 families with 3,779 children (52 percent male, 48 percent female; mean age = 10.1 years) involved in child welfare who participated in Connecticut’s Supportive Housing for Families (SHF) Program between 1999 and 2008. The SHF provided access to scattered-site PSH, coordination of mental health, parenting interventions, and access to child welfare resources. The authors found improvements from intake to discharge in obtaining permanent housing, employment, and access to health care. Reports of substance abuse outcomes were deemed unreliable and child welfare outcomes were not tracked. While child well-being scores were reported to have improved, there were no specific health outcomes included in the analysis. Length of stay in the program was significantly associated with success at discharge, which could reflect the impact of longer exposure or could reflect sorting of clients who may have left or been asked to leave due to “dissatisfaction, or noncompliance with program requirements, arrest, etc.” (Farrell et al., 2010, p. 150). One quasi-experimental study examined reunifications with children among high-needs families given supportive housing and two matched comparison groups. Of the 189 supportive housing families, 20 percent had experienced a separation and 11 percent had been reunified a year later. The proportion of reunifications was significantly higher than for a matched sample who entered shelter, but not significantly higher than for a matched sample who entered public housing (Rog et al., 2016). Health outcomes were not reported.

The Family Unification Program (FUP) is a federally funded partnership between Public Housing Authorities and Child Welfare Agencies that offers subsidized housing, in the form of Housing Choice Vouchers, to families whose inadequate housing threatens out-of-home placement or impedes reunification. Youth ages 18 to 21 who left foster care after age 15 are also eligible. A small experiment in Chicago randomized families who were enrolled in a Housing and Cash Assistance Program designed to prevent family separation due solely to living circumstances to additionally receive FUP vouchers or not (Fowler and Chavira, 2014). After 10 months, receipt of FUP vouchers reduced both homelessness and rates of out-of-home placement, albeit with a p-value of .11 for the latter outcome. The small sample size (n = 31 families in the FUP group at followup) meant statistical power to detect effects was low, but the result is promising. A larger quasi-experimental study in two sites examined FUP relative to a waiting-list control for both family preservation and family reunification cases (Pergamit et al., 2017). The authors found small reductions in repeat reports and

substantiated reports of abuse or neglect and faster case closings in both sites for families in preservation cases, but no difference in removals of children from families. There was also an increase in the likelihood of reunification and probability of case closure for reunification cases in one of the sites, with mixed findings on repeat reports. The low levels of child placements and high levels of reunification for the waiting list controls suggest that the program may not have been targeted to families at high risk. Neither study reported on other aspects of physical or mental health. A five-site RCT of PSH for families involved in the child welfare system with some measures of adult and child health and psychosocial functioning is under way at the Urban Institute (Cunningham et al., 2014).

In the much larger Family Options Study, where families were not selected for child welfare risk, offers of permanent housing subsidies, without additional services, reduced child separations at the 20-month follow-up from 16.9 percent in the usual-care group to 9.8 percent in the group receiving subsidies and more than halved foster care placements (5.0 percent versus 1.9 percent). Transitional housing and temporary rapid re-housing had no effect on these outcomes, and differences in the permanent housing subsidy group were no longer detectable at the 37-month follow-up point (Gubits et al., 2015, 2016).

UNACCOMPANIED HOMELESS CHILDREN AND YOUTH

Using HUD’s definition, in January 2016, there were 3,800 unaccompanied children under 18 years of age and 35,686 youth, that is, young adults under age 25 experiencing homelessness, across the United States (HUD, 2016b). Unaccompanied children and youth who are not part of a family or a multichild household make up approximately 10 percent of people who experience homelessness as individuals; that is, they are not accompanied by family member (s) or other individual(s) on a single night (HUD, 2016c).

An accurate prevalence of youth homelessness is difficult to determine, however, due to a number of factors, including the lack of a consistent definition of youth homelessness with respect to both age and housing condition, as well as the population’s transient nature and the impermanence of their homeless status. For example, there were 89,000 unaccompanied homeless youth (and 1.3 million total homeless youth) enrolled in school (preschool to grade 12) in 2013–2014 by the broader U.S. Department of Education definition, which includes “children and youth who are sharing the housing of other persons due to loss of housing, economic hardship, or similar reason; are living in motels, hotels, trailer parks, or camping grounds due to lack of alternative adequate accommodations; are living in emergency or transitional shelters; are abandoned in hospitals; or are awaiting foster care placements” (Public Health Service Act, 42 U.S.C. § 11434a(2), 2001) in addition to the categories recognized by HUD. Over three-quarters of these children, most of whom are homeless with their families, are living in “doubled-up” situations with other households (Endres and Cidade, 2015).

The term “homeless youth” is used to describe a host of individuals including runaways (i.e., youth who have spent more than one night away from home without parental permission), youth forced to leave their homes, street youth (i.e., youth living in locations such as under bridges and in abandoned buildings), and systems youth (i.e., youth who have previously been involved in foster care or juvenile justice) (Edidin et al., 2012).

The Runaway and Homeless Youth Act, which authorizes funding through the Family and Youth Services Bureau of the U.S. Department of Health and Human Services (HHS), defines a homeless youth as one who is “not more than 21 years of age . . . for whom it is not possible to live in a safe environment with a relative and who has no other safe alternative living arrangement” (42 U.S.C. § 5732a). The regulations implementing this act define a runaway as someone “under 18 years of age who absents himself or herself from home or place of legal residence without the permission of parents or legal guardians” (45 CFR § 1351.1(l)). HUD defines youth as ages 18-24 (HUD, 2016b,c). The McKinney-Vento Homeless Education Assistance Improvements Act of 2001 applies to students eligible for public education services under state and federal law and defines unaccompanied youth as “those who are not in the physical custody of a parent or guardian” (42 U.S.C. §§ 11431 et seq.). This can include “runaways living in runaway shelters, abandoned buildings, cars, on the streets, or in other inadequate housing; children and youth denied housing by their families; and school-age unwed mothers living in homes for unwed mothers because they have no other housing available” (Popp et al., 2007, p. 11).

Factors contributing to youth homelessness include family conflict, interpersonal violence and trauma, socioeconomic factors, and mental health and substance use disorders, among others (Mallett et al., 2005; Ferguson, 2009; Coates and McKenzie-Mohr, 2010; Edidin et al., 2012). Lesbian, gay, bisexual, transgender, and questioning youth are thought to represent 20–40 percent of youth experiencing homelessness (National Alliance to End Homelessness, 2012; Ray, 2006).

Numerous studies identify the risks of youth homelessness for various health and social outcomes and the short-term benefits of some approaches in ameliorating them (Ferguson and Maccio, 2016; Slesnick et al., 2009). In their systematic review of effective interventions for youth experiencing homelessness, however, Altena et al. (2010) found no compelling evidence that specific interventions are beneficial for improving quality of life among youth experiencing homelessness, due to moderate study quality, and no evidence of the health impacts of PSH for youth experiencing homelessness.

In one small study, Kisely and colleagues (2008) examined the association between supportive housing and health among fifteen 16- to 25-year-old youth in Halifax, Nova Scotia, comparing them with 30 youth experiencing homelessness using a drop-in center in the same agency. Youth in both groups had access to identical services but controls lived in shelters (57 percent) or independent apartments (27 percent) or couch-surfed with family or friends (13 percent). Using a survey design, youth in supportive housing reported higher levels of health and

lower rates of substance abuse compared to controls; 40 percent of youth in supportive housing rated their health as “excellent” while none in the control group did so. There were no measurable indicators of what “health” constituted in the study.

More recently, Gilmer (2016) estimated the health service costs associated with PSH enrollment among 2,609 youth (ages 18–24) with serious mental illness (defined as schizophrenia, bipolar disorder, or major depressive disorder) receiving public mental health services in California between 2004 and 2010.

Gilmer (2016) compared health service costs among youth in PSH and youth with serious mental illness receiving public mental health services in California from January 1, 2004, through June 30, 2010. The primary variables of interest in the models comparing PSH participants with the propensity score–matched control group were participation in the PSH program for the post-period and for the interaction between the PSH and the post-period. The author included age, gender, race/ethnicity, clinical diagnosis, comorbid substance use disorders, and Medi-Cal coverage as additional control covariates. Results indicated that youth in PSH programs had increased inpatient, crisis residential, and mental health outpatient costs, suggesting that PSH programs and connected services may be inadequately designed or implemented to meet the needs of this population.

A subanalysis of the At Home/Chez Soi RCT conducted across five Canadian cities (Kozloff et al., 2016) evaluated the effect of housing stability among 156 young adults age 18–24 who were randomly assigned to receive Housing First (housing combined with ACT or intensive case management) or treatment as usual (not defined) for 24 months. The study’s primary outcome was housing stability, defined as the percentage of days one remained housed as a proportion of days for which residence data were available. While youth in the Housing First group were, on average, stably housed at a higher percentage over youth in usual treatment (65 percent compared to 31 percent), there were no statistically significant differences in quality of life or self-rated physical or mental health symptoms between the groups.

Non-PSH Interventions: Youth in Foster Care

Youth in the foster care system sometimes have experience with homelessness. Focus group and interview data collected by the Family and Youth Services Bureau Street Outreach Program (2016) showed that nearly 5 percent of youth experiencing homelessness had been in foster care prior to their first homeless episode. Study participants who reported being in foster care also had significantly longer periods of time being homeless when compared to their non-foster care peers. Youth aging out of the foster care system have a high probability of becoming homeless (between 31 percent and 46 percent by age 26 in a study in three midwestern states), and having symptoms of mental health disorders places youth at higher risk. Extended foster care delays but does not prevent the onset of

homelessness, suggesting that a more robust housing intervention may be necessary (Dworsky et al., 2013). A literature review indicates that foster youth aging out of the system often face housing instability and homelessness due to lack of education, lack of preparation for entering the labor force, and lack of financial support needed for renting a unit during their transition into adulthood (Dworsky et al., 2012).

CONCLUSIONS

A convening of experts by the HUD concluded that “considerable research supports targeting permanent supportive housing to those who experience chronic homelessness,” (HUD, 2015a, p. 5) which, by HUD’s definition, requires an ongoing disability. Beyond that, however, “little evidence exists to support targeting interventions to specific families or individuals” and “existing assessment tools do not have a strong evidence base” (HUD, 2015a, p. 5). The committee’s assessment of the literature indicates that while families who obtain PSH do well—in terms of reducing child behavioral problems and depression and improving parenting competencies (Gewirtz et al., 2015)—the evidence is not clear that they do better than families who obtain ongoing rental subsidies (Gubits et al., 2015, 2016). Likewise, it is not clear how to target a subgroup that might benefit from case management and additional services linked to housing. There is also suggestive evidence that PSH may reduce child placements for some families involved in the child welfare system, but again it is not clear how to target this resource (Gewirtz et al., 2015). Although unaccompanied youth and those who age out of the foster care system are at high risk for adverse health and social outcomes, there is little evidence as to whether PSH might be a useful intervention for them. It is plausible that permanent supportive housing would support both housing and health outcomes for high-risk members of all of these populations, but evidence is largely descriptive and ranges from weak to nonexistent. Given this, it is unclear whether other, less-intensive interventions might do as well, or how subpopulations who might benefit from PSH should be identified.