Characteristics and Use of Services Among Literally Homeless and Unstably Housed U.S. Veterans With Custody of Minor Children

Psychiatric conditions are common among U.S. veterans, particularly homeless veterans (1). This pattern has worsened recently throughout the wars in Iraq and Afghanistan. The demographic makeup of the U.S. military has also been changing. More women are entering the military (2,3), and more younger veterans have developed posttraumatic stress disorder (PTSD) because of recent wars in Iraq and Afghanistan (4,5). As a result, a growing number of homeless veterans with psychiatric conditions have young children. There is no recent literature identifying the prevalence of this subpopulation, their mental health needs, and the services that are available to them.

A literature search of homelessness among veterans with young children found only one previous study, published about a decade ago (6). Of 9,444 homeless veterans surveyed, 37% were parents with minor children, and 11% of children resided with their parents in transitional housing. It is not well understood how the U.S. Department of Veterans Affairs (VA), an institution that has historically almost exclusively served males, is providing services to homeless female veterans with children in its ambitious goal to end homelessness among veterans (7).

Outside the VA, families have been recognized as a growing segment of the homeless population since the 1980s (8–10). Most studies of homeless families have been conducted with homeless women and their children. These studies have shown that compared with housed mothers, homeless mothers are more likely to have chronic general medical and psychiatric problems and often face greater struggles in obtaining housing, social support, and health services, which can affect their parenting ability (8,11–13). In fact, high rates of developmental delay, cognitive deficits, mental and behavioral problems, and various health problems have been found among children of homeless parents (14–17). Thus homeless parents may experience a “double crisis” of homelessness and child rearing (18).

We conducted an observational study to examine the prevalence of homeless veterans with custody of minor children (herein referred to as “children in custody”), compared the sociodemographic and clinical characteristics of this subgroup with those of other veterans, and observed the extent to which this subgroup was referred and admitted to VA programs. We focused on veterans because there are limited VA administrative data about their children, but the results may have implications for their children. We hypothesized that a higher proportion of homeless female veterans would have children in custody compared with men, that high rates of general medical and mental health problems would be found among this subgroup overall, and that this subgroup would be more likely than other veterans to receive VA supported housing services.

Methods

The Homeless Operations Management and Evaluation System (HOMES) is one of several data streams for a comprehensive homeless registry that offers near real-time information on the population of VA homeless service users who have been referred and admitted for services. This study used client-level data from HOMES for all veterans who were referred to VA homeless programs between April 2011 and November 2012. Because of eligibility criteria, program capacity, and other constraints, not all veterans who were referred to VA homeless programs were admitted.

The main analyses focused on all referred veterans, but additional analyses on referral and admission patterns were conducted. The original data set included 120,852 veterans, but only veterans who were “literally homeless” (19) or “unstably housed” were included in the study. “Literally homeless” was understood to mean living on the streets or in places not meant for human habitation (20). “Unstably housed” was defined as being in “imminent risk of losing housing” or being “unstably housed/at risk of losing housing.” Veterans who were incarcerated were excluded from the study.

The resultant data set included 89,142 veterans across 142 sites, of whom 75.6% (N=67,410) were literally homeless and 24.4% (N=21,732) were unstably housed, 13.2% (N=11,777) had children in custody, and 86.8% (N=77,365) did not have children in custody. To put these numbers in context, of the 5.6 million veterans who used VA health services in fiscal year 2012, 1.2% were literally homeless, .4% were unstably housed, and .2% were literally homeless or unstably housed and had children in custody.

Program Descriptions

Referral and admission to five main VA homeless programs were examined, including Housing and Urban Development–Veterans Affairs Supportive Housing (HUD-VASH), a permanent supported-housing program; the grant and per diem program, a transitional housing program; Health Care for Re-Entry Veterans, a prisoner reentry program; Veterans Justice Outreach, a jail diversion or outreach program; and Domiciliary Care for Homeless Veterans, a residential treatment program.

Measures

Staff of VA homeless programs conduct in-person assessment interviews with all homeless veterans who are prospective clients of VA homeless programs by using a structured form that collects information on sociodemographic characteristics, homeless history, and health characteristics. Veterans are asked whether they have children under the age of 18 and if so, whether any of the children and how many are in their legal custody. Administrative data for program referrals and admissions are documented by VA staff and collected by HOMES.

Housing status and incarceration history.

Veterans are asked how long their most recent episode of homelessness had lasted and how many episodes of homelessness they have experienced in the past three years. Veterans are categorized as chronically homeless if they meet the federal definition of being continuously homeless for one year or more, having four or more episodes of homelessness in the past three years, or both (21).

Incarceration history is assessed by asking veterans how much total time they have spent in jail or prison during their lifetime. For this study, their responses were then coded dichotomously as having or not having an incarceration history.

Health status.

Medical history is assessed by asking clients whether a doctor or nurse has ever told them they had a list of medical conditions. Client are also asked to rate their general medical health in the past month on a 5-point scale from poor to excellent; responses were categorized as excellent or very good, good or fair, and poor.

Psychiatric diagnoses are assessed by clinicians in their assessment interview and through review of existing medical records. Furthermore, clients are asked whether they have ever been hospitalized for a psychiatric problem (not including residential treatment or hospitalization for a substance use problem).

Data Analysis

First, the number and proportion of veterans who had minor children and the subgroup with children in custody were examined by homeless status (literally homeless or unstably housed) and gender. Second, the sociodemographic and health characteristics of literally homeless and unstably housed veterans with children in custody, stratified by gender, were compared with those of other veterans. We limited the sample of veterans with minor children to those with children in custody because their children presumably have a greater role in their lives than is the case for veterans whose children are not in their custody. It may be notable that veterans with children in custody included those with only partial custody (whose children did not reside with them), and we were not able to differentiate between those with partial and full custody of their children. Given the size of the samples, notable effect sizes (Cohen’s d of ≥±.5 for continuous variables and a percentage change of >5% for dichotomous variables) were indicated. A log transformation was conducted on the total monthly income variable because of a nonnormal distribution.

Third, differences in referral and admission patterns for various VA homeless programs among veterans with children in custody were compared. A series of logistic regressions was conducted to analyze the extent to which independent variables explained the differences in percentages of veterans referred or admitted to various programs. In the logistic regression models, the dependent variable was referral or admission to one of the VA homeless programs. The independent variables of interest were gender, child-in-custody status, and a gender × child-in-custody interaction term. The interaction term was included because there are distinct differences in the proportion and needs of female and male homeless veterans with children. Other independent variables entered as covariates into the models included sociodemographic and health characteristics found during the second stage of the data analysis to be notably different among veterans with and without children in custody; it was important to control for these variables because they may influence referral and admission patterns.

Results

Table 1 shows that among men, nearly one-tenth of literally homeless veterans and nearly one-fifth of unstably housed veterans had children in custody. Unstably housed veterans were more likely than literally homeless veterans to have children in custody, and much more so among women than men. Among literally and unstably housed male veterans with minor children, only about one-third to one-half had custody of their children, whereas among women, 83% had custody.

Table 2 shows the sociodemographic and general medical characteristics of literally homeless and unstably housed male veterans with children in custody and male veterans without children in custody. Compared with other male veterans, male veterans with children in custody were younger; more likely to have served in recent theaters of operations, such as Iraq and Afghanistan; more likely to be married; and less likely to have a history of incarceration. They were also less likely to have a number of general medical and psychiatric conditions—including hepatitis C, chronic obstructive pulmonary disease, and substance use disorders—and prior psychiatric hospitalizations. However, 72% of male veterans with children in custody had a psychiatric diagnosis, and 11% were diagnosed as having a psychotic disorder; they were more likely than other male veterans to have PTSD and other anxiety disorders.

Logistic regressions adjusted for age were used to further explore differences between male veterans with and without custody of children; the results indicated that men with children in custody were still significantly more likely to have PTSD (odds ratio [OR]=1.45, 95% confidence interval [CI]=1.38–1.53) and other anxiety disorders (OR=1.11, CI=1.06–1.18) compared with other male veterans. The same health differences were found when the analyses examined literally homeless and unstably housed male veterans separately. [The results of these analyses are available in an online supplement to this article.]

Table 3 presents a comparison of the sociodemographic and health characteristics of literally homeless and unstably housed female veterans with children in custody and female veterans without children in custody. Similar to the male veterans, female veterans with children in custody were younger; more likely to have served in a recent theater of operations; more likely to be married; less likely to have a history of incarceration; less likely to have numerous general medical and psychiatric conditions, including hepatitis C, diabetes, chronic pain, a psychotic disorder, and substance use disorders; and less likely to have prior psychiatric hospitalizations. Compared with other women, female veterans with children in custody were also less likely to report chronic homelessness and more likely to be black. Notably, 67% of female veterans with children in custody had a psychiatric diagnosis, and 11% were diagnosed as having a psychotic disorder. The same health differences between female veterans with and without children in custody were found when the analyses examined literally homeless and unstably housed female veterans separately [see online supplement].

Table 4 shows differences in program referral and admission patterns among literally homeless and unstably housed veterans with and without children in custody (combined by gender). The largest differences (>20% change) for both referral and admission patterns were found in HUD-VASH and the grant and per diem program. After adjustment for differences in sociodemographic and health characteristics, the analyses showed that veterans with children in custody had more than three times greater odds of being referred and admitted to HUD-VASH compared with other veterans; however, their odds of being referred and admitted to the grant and per diem program were less than one-third of the odds among other veterans.

There was also a significant gender effect in terms of referral and admission patterns (Table 4). Female veterans were more likely than male veterans to be referred and admitted to HUD-VASH, and male veterans were more likely to be referred and admitted to the grant and per diem program. Furthermore, a significant gender × children-in-custody interaction effect revealed that female veterans with children in custody were more likely to be referred and admitted to the HUD-VASH than to any other program; they were least likely to be referred and admitted to the grant and per diem program. There were no notable differences between veterans with or without children in custody in rates of referrals to general medical or mental health services, but rates of referrals to mental health services were relatively low (22% and 25% for veterans with and without children in custody, respectively) given the high prevalence of psychiatric diagnoses in the sample.

Discussion

This is the first study in the past decade, to our knowledge, to examine general medical and mental health problems among homeless veterans with children. A substantial proportion of literally homeless veterans (men, 9%; women, 30%) and unstably housed veterans (men, 18%; women, 45%) had minor children in their custody, although it is also important to recognize they are a small subgroup (.2%) of the 5.6 million VA service users in fiscal year 2012. Nonetheless, these findings suggest that the typical image of a homeless veteran as a single, older male may be changing. Many homeless veterans are women, and many veterans have children in custody, making them homeless families rather than individuals. Unfortunately, we could not determine how many homeless veterans with children in custody had only partial custody and did not reside with their children. Nonetheless, the same factors that impede a parent’s ability to maintain a stable residence may also impair their capacity to nurture children (18).

Although both literally homeless and unstably housed veterans with children in custody were less likely than veterans without children in custody to have various chronic general medical conditions and psychiatric disorders, they still had high rates of psychiatric disorders, consistent with reports of psychiatric illness among homeless parents who are not veterans (12,22). Most notably, we found that 11% of homeless male and female veterans with children in custody had been diagnosed as having a psychotic disorder. Studies of parents with psychosis have found that a significant proportion experience impairments in parenting, which may negatively affect their children’s development (23–25). Although we did not have data about the children, there may be cause for concern about the children of homeless veterans, given that homeless children have been found to have high rates of mental illness, to have poor general medical health, and to often experience developmental and academic delays (14–17).

There was a surprisingly low rate of referral to outpatient VA mental health services among homeless veterans, despite their high rate of psychiatric disorders, suggesting that VA mental health services may be underutilized or that housing and other psychosocial needs are taking precedence over mental health needs. Providers of homelessness services may need to make greater and more immediate linkages to outpatient mental health services. The importance of these linkages is underscored by the Veterans Access, Choice, and Accountability Act of 2014, which allows veterans to use community mental health services that are reimbursable by the VA and may help increase utilization of mental health services in this population.

It is also notable that male veterans with children in custody were more likely to have PTSD than male veterans without children in custody, which is mostly explained by their younger age and exposure to recent wars. The exact effect of PTSD on parenting is unknown, although a previous study found that among homeless female veterans with children, PTSD symptoms were associated with depression and anxiety symptoms among the veterans’ children (17). Further study of homeless veterans with children is needed, and emerging interventions, such as supported parenting (26,27), may be warranted to help homeless veterans with child rearing. The VA has not traditionally addressed the parenting needs of homeless veterans, and addressing these needs may represent a new opportunity to provide comprehensive services for this population.

Veterans with children in custody were more likely than other veterans to be referred and admitted to HUD-VASH. That was also true for female veterans with children in custody compared with male veterans with children in custody. Thus the supported-housing program is one of the VA’s main programs for homeless families, offering case management services and independent housing that may be particularly suitable for the needs and preferences of homeless veterans with children in custody. Features of HUD-VASH make it accommodating for veteran families; for example, as a tenant-based voucher program, it allows veterans to secure apartments suitable for their families. In contrast, the grant and per diem program mostly offers transitional housing arrangements that are populated mainly by men, offer few resources for parents, largely focus on substance abuse, and are not always sensitive to privacy needs (28,29). Thus there may be a lack of female-friendly and family-friendly VA transitional housing options.

The VA has moved toward a policy of providing immediate, permanent supported housing for individuals by using the Housing First model (30). HUD-VASH has formally adopted the Housing First model, which has been used by non-VA programs for over a decade to successfully place homeless families immediately into permanent subsidized housing(31). However, there has not been a systematic evaluation of the Housing First approach for homeless veteran families. In addition, rapid rehousing interventions, such as Supportive Services for Veteran Families, may prove to be highly effective and stabilizing for homeless and at-risk veteran families, but rigorous evaluation of these programs is needed.

This study had several limitations of note. One, the data were based on the HOMES database, which relies on the accurate and diligent documentation of VA clinicians and is limited to veterans in the VA system. In addition, the medical history of veterans was based on self-report, which may have resulted in underreporting. Second, this study was cross-sectional, so causation cannot be determined. Third, the number of veterans considered unstably housed was only a small proportion of the actual population, given that the sample was based only on veterans who were engaged with VA homelessness services. It is unlikely that there are more literally homeless veterans than unstably housed veterans, as was the case in this study. Fourth, specific parenting needs and the needs of children were not assessed. Strengths of the study included data on the population of VA homeless service users, examination of service patterns for a variety of programs, separate analyses by gender, and new findings on a relatively unexplored topic.

Conclusions

Veterans and their families are a growing segment of the homeless veteran population and have great mental health needs. Homeless veterans with children may need additional services to support their role as parents and to help them find suitable housing for themselves and their children. Various practical and ethical dilemmas need to be considered, such as whether participation in a parenting program should be required or encouraged as part of supported housing for homeless families, how mental illness may affect parenting, and how case managers can manage caseloads of high-risk families with complex needs. These issues are especially relevant for participants in HUD-VASH and for the growing population of homeless female veterans. Further research is needed to determine how homeless parents can be supported in their roles as parents and to identify mental health and psychosocial services that are needed to ensure the healthy development of their children and end the cycle of homelessness.