Correlates of Homelessness At Baseline and Becoming Homeless At 1-Year Follow-up Among Patients in the Veterans Affair Healthcare System

Background: On any given day, approximately 75,000 veterans are homeless. The U.S. government has recently made the eradication of homelessness among veterans a priority. Compared to veterans in the general population, those seeking medical treatment in the Veterans Affairs Health Care System (VA) tend to be poorer, with higher levels of mental illness and substance use. Given that these are significant correlates of homelessness in other populations, it follows that VA clients are at elevated risk of homelessness. However, there is a dearth of scholarship examining homelessness in this vulnerable population. Moreover, few studies examine the factors associated with transitioning into homelessness over time. This research seeks to address these gaps by conducting a longitudinal study of homelessness and its correlates among 6,232 veterans seen in 8 VA sites across the U.S.

Methods: Drawing on the Veterans Aging Cohort Study (2002-2004), a longitudinal, prospective multisite study of VA patients, we conducted: 1) a baseline assessment of factors associated with risk of homelessness, and 2) a longitudinal assessment of factors associated with transitioning into homelessness over the course of a year. Logistic regression and proportional hazards models were used to analyze the data.

Results: At baseline, we found an 11% prevalence of current homelessness, significantly higher than the 4% prevalence recorded among all veterans in previous studies. Substance use, depression, being African American and being a man were significantly associated with risk of homelessness (p<.01). Age had a curvilinear relationship, with risk peaking at 50 years (p<.01), indicating that the cohort who joined the forces voluntarily immediately after the Vietnam war were at highest risk. People with HIV were significantly less likely to become homeless, indicating that services for seropositive clients were succeeding in keeping them housed. Longitudinally, the most significant risk factor for becoming homeless over the course of a year was increasing consumption of alcohol to hazardous levels (p<.001), a threshold of risky alcohol consumption that is lower than clinical abuse or dependence. 

Implications: Given that VA clients are at higher risk of homelessness than other veterans, and are already in care, initiatives to remove homelessness among veterans should start with this population. Additionally, the results help to construct a profile for those at highest risk of homelessness. Finally, the results shed light on the kind of interventions that need to be extended to VA clients in order to secure stable housing: psychosocial services like those offered to seropositive clients, especially those that target depression were successful. The results indicate that early and brief alcohol treatment can prevent later homelessness. The VA needs to screen for hazardous alcohol consumption routinely and implement very brief, five-minute psycho-educational interventions that have been shown to be effective in reducing consumption among hazardous consumers of alcohol. Given the consistency with which veterans return for care, and the high prevalence of homelessness and its risk factors among its patients, the VA constitutes a setting tailor-made to successfully engage veterans at risk of homelessness.