Unsheltered homelessness among veterans: correlates and profiles

We identified correlates of unsheltered status among Veterans experiencing homelessness and described distinct subgroups within the unsheltered homeless Veteran population using data from a screening instrument for homelessness that is administered to all Veterans accessing outpatient care at a Veterans Health Administration (VHA) facility. Correlates of unsheltered homelessness included male gender, white race, older age, lower levels of VHA eligibility, substance use disorders, frequent use of VHA inpatient and infrequent use of VHA outpatient services, and residing in the West. We identified six distinct subgroups of unsheltered Veterans; the tri-morbid frequent users represented the highest need group, but the largest group was comprised of Veterans who made highly infrequent use of VHA healthcare services. Differences between sheltered and unsheltered Veterans and heterogeneity within the unsheltered Veteran population should be considered in targeting housing and other interventions. Introduction More than 170,000 people experiencing homelessness on one night in January 2014 were staying in unsheltered situations, which includes places not meant for human habitation, such as the streets, abandoned buildings, vehicles, or parks (U.S. Department of Housing and Urban Development, 2014). However, surprisingly little is known about unsheltered homelessness. Previous work has found that persons staying in unsheltered situations are often chronically homeless (Cousineau, 1997; O’Toole, Gibbon, Hanusa, & Fine, 1999; Shern et al., 2000; Tsai, Kasprow, Kane, & Rosenheck, 2014) and have high rates of behavioral health and medical needs (Shern et al., 2000), but less access to care (Lam & Rosenheck, 1999; Levitt, Culhane, DeGenova, O’Quinn, & Bainbridge, 2009). Given the findings of this limited research, it is often assumed that people living in unsheltered situations comprise the highest-need and hardest-to-serve homeless subpopulation. Research to date has identified many correlates of unsheltered homelessness—that is, characteristics that are more commonly observed among persons experiencing unsheltered homelessness than those who are sheltered. Compared with their sheltered counterparts, who are accessing emergency shelter or transitional housing, individuals experiencing unsheltered homelessness are more often White (Early, 2005; Gelberg & Linn, 1989; Larsen, Poortinga, & Hurdle, 2004), older (Levitt et al., 2009), identify as a Veteran (Gelberg & Linn, 1989; Levitt et al., 2009), have lower education, are unemployed (O’Toole et al., 1999), and have a history of incarceration (Levitt et al., 2009). The literature indicates that individuals experiencing unsheltered homelessness more often have both lifetime and recent serious mental illness (SMI) or mental health issues (Levitt et al., 2009; Nyamathi, Leake, & Gelberg, 2000), lifetime and recent substance use or substance use disorders (SUD) (Gelberg & Linn, 1989; Nyamathi et al., 2000), and comorbid SMI and SUD (Levitt et al., 2009). Poorer general health is also more common among individuals living in an unsheltered situation, as is recent experience of victimization and injury (Gelberg & Linn, 1989; Nyamathi et al., 2000) and recent lack of preventive healthcare (Nyamathi et al., 2000). While these findings provide some evidence of broad differences between persons experiencing homelessness based on their sheltered status, there has been little in-depth exploration of the characteristics of individuals within sheltered or unsheltered groups, particularly among unsheltered individuals. Unsheltered homelessness is more prevalent among Veterans than among non-Veterans. At one point in time in 2014, 36% of homeless Veterans were unsheltered compared with 30% of non-Veterans (U.S. Department of Housing and Urban Development, 2014). Research regarding unsheltered homelessness among Veterans specifically is limited. The few prior studies on this topic have found that Veterans staying in unsheltered homeless situations were significantly different than Veterans staying in a sheltered situation in terms of demographics and housing outcomes. These findings mirror the results of similar studies using non-Veteran samples. Most strikingly, the odds of Veterans experiencing persistent homelessness (i.e., for more than 6 months) were 2.7 times greater for those staying in an unsheltered situation as opposed to a sheltered situation (Montgomery, Byrne, Treglia, et al., unpublished data). While these findings provide some evidence that Veterans living in an unsheltered situation have significant needs and are at increased risk of longer-term homelessness, the assumption that they represent the most “hard-core” group of individuals experiencing homelessness may ignore important heterogeneity within this population. Therefore, the purpose of this study was to examine more closely the characteristics of Veterans who reported being homeless and staying in an unsheltered situation using administrative data from more than 4,000 individuals self-identifying as such. The following aims were devised to address this purpose: (1) to identify correlates of Veterans’ unsheltered homelessness status and (2) to identify and describe distinct subgroups within the unsheltered homeless Veteran population.