270), Baltimore (n = 295), and Chicago (n = 486) in 1985-1987. It assessed whether veterans were more likely to be homeless than men who were not veterans. Among homeless men, 41% reported military service, compared with 34% of men in the general U.S. population. Comparing veterans with nonveterans showed that those 35-44 years old (those who were most likely to have served in Vietnam) had an odds ratio (OR) for homelessness of 1.01 (95% CI 0.85-1.21). Among both white and black veterans, the cohorts 20-34 and 45-54 years old had a greater likelihood of homelessness (OR 3.95, 95% CI 3.39-4.58; and OR 1.75, 95% CI 1.45-2.15, respectively). The youngest veterans in the 1987 Third Survey of Veterans, a nationally representative sample of 9442 noninstitutionalized veterans, had served after Vietnam in the all-volunteer Army, and only 7% had seen combat. Most of the veterans 45-54 years old had served between the Korean War and the Vietnam War, and 17% of them had seen combat; among veterans who served during the Vietnam War era, 40% reported combat exposure. The much higher prevalence of psychiatric illness, substance abuse, and antisocial personality disorder among white veterans 20-34 years old seemed to be a more likely contributor to the greater vulnerability of this group than combat exposure. Homeless veterans in the most vulnerable groups were the least likely to have served during wartime or combat, and this reduces the role that could be attributed to combat stress in the genesis of homelessness among veterans.
Gamache et al. (2001) conducted a followup of the study by Rosenheck et al. (1994) to determine whether the exceptionally high risk of homelessness among post-Vietnam-era, largely noncombat veterans first observed in 1987 was still evident one decade later. Data on 1841 current male clients of homeless assistance programs throughout the U.S. gathered by the 1996 National Survey of Homeless Assistance Providers were compared with information on the general population from the 1996 Current Population Survey. The homeless clients, who were interviewed face-to-face, came from the 28 largest metropolitan statistical areas (MSAs), 24 randomly chosen small and medium-sized MSAs, and 24 randomly chosen rural counties. The relative probability of homelessness for veteran vs nonveteran was determined for defined age and race groups. As of 1996, the proportion of veterans in the adult male homeless population had declined from 41% to 33%, although this proportion was still greater than the 28% in the general population. The especially high risk of homelessness among veterans of the immediate post-Vietnam era who had been aged 20-34 in 1987 was sustained 10 years later (OR 3.17, 95% CI 2.69-3.73), although the youngest cohort of veterans also showed a significant risk of homelessness (OR 2.04, 95% CI 1.59-2.64).
The secondary studies that were reviewed provided mixed findings as to whether deployment to a war zone increased the risk of homelessness.
Winkleby and Fleshin (1993) compared three groups of homeless men residing in three California shelters: 173 combat-exposed veterans, 250 veterans without combat exposure, and 585 nonveterans. The prevalence of psychiatric hospitalizations and physical injuries before homelessness was approximately 1.5-2 times as high in combat veterans as in nonveterans or veterans who had not experienced combat. The time between military discharge and first being homeless was more than a decade for 76% of the homeless combat veterans.
Rosenheck et al. (1991) compared data on 10,524 veterans participating in the VA Homeless Chronically Mentally III Veterans Program with veterans in the general population. Of the participating veterans, 50% had served in the military during the Vietnam War era, 45% had served in the Vietnam theater, and 41% had been exposed to combat. Only 29% of the total U.S.