Administration and Policy in Mental Health and Mental Health Services Research

, Volume 40, Issue 5, pp 400–405

Homeless Veterans Who Served in Iraq and Afghanistan: Gender Differences, Combat Exposure, and Comparisons with Previous Cohorts of Homeless Veterans

Authors

    • VA New England Mental Illness, Research, Education, and Clinical Center
    • Department of PsychiatryYale University School of Medicine
  • Robert H. Pietrzak
    • VA New England Mental Illness, Research, Education, and Clinical Center
    • Department of PsychiatryYale University School of Medicine
    • Clinical Neurosciences Division, National Center for Posttraumatic Stress DisorderVA Connecticut Healthcare System
  • Robert A. Rosenheck
    • VA New England Mental Illness, Research, Education, and Clinical Center
    • Department of PsychiatryYale University School of Medicine
    • Yale School of Public Health
Point of View

DOI: 10.1007/s10488-012-0431-y

Cite this article as:
Tsai, J., Pietrzak, R.H. & Rosenheck, R.A. Adm Policy Ment Health (2013) 40: 400. doi:10.1007/s10488-012-0431-y

Introduction

Ending homelessness among veterans is a national priority of the Department of Veterans Affairs (VA). Homelessness has been a public health problem for over three decades and veterans are of particular concern given their military service for the country. Yet there is little understanding of the risk of homelessness and of the unique characteristics and clinical needs of homeless veterans who served in Operations Enduring Freedom, Iraqi Freedom, and New Dawn (OEF/OIF/OND).

The extant literature on homelessness has focused primarily on two categories of risk factors: sociodemographic characteristics and psychiatric difficulties. Regarding sociodemographic characteristics, the majority of homeless veterans have been characterized as unmarried, unemployed, poor, mostly White males in their late 40’s, with histories of homelessness and incarceration (Rosenheck and Koegel 1993; Tsai et al. 2012; US Department of Housing and Urban Development and US Department of Veterans Affairs 2009; Winkleby and Fleshin 1993).

In terms of psychiatric difficulties, two commonly identified risk factors for homelessness have been severe mental illness (schizophrenia or bipolar disorder) and substance use disorders (Caton et al. 2005; Edens et al. 2011; Folsom et al. 2005; Susser et al. 1993). These psychiatric difficulties are often attributed to combat exposure, which can have deleterious effects on mental health (Hoge et al. 2004). But generally, rates of combat-related posttraumatic stress disorder (PTSD) have been found to be relatively low among homeless veterans (Kasprow et al. 2000; Tsai et al. 2012), though this has not been extensively investigated in homeless OEF/OIF/OND veterans.

We used national administrative data from the VAs largest supported housing program, the Housing and Urban Development-Veterans Affairs Supportive Housing (HUD-VASH) program, to: (1) better understand the sociodemographic characteristics of homeless OEF/OIF/OND veterans, including their level of VA service-connection, a potential protective factor against homeless (Edens et al. 2011), at the time they are referred to supported housing; and (2) examine the relation between combat exposure and psychiatric diagnoses among homeless OEF/OIF/OND veterans compared to previous cohorts of homeless veterans.

Data Source and Measures

National administrative data on homeless veterans referred to the HUD-VASH program from January 2008 to April 2011 were obtained for analysis and its use was approved by the local VA institutional review board. The HUD-VASH program is a key component of the government’s plan to end homelessness among veterans (United States Department of Veterans Affairs 2009). Veterans are referred to the program by VA mental health clinicians (psychiatrists and clinical psychologists) or HUD-VASH clinicians (social workers) who conduct community outreach to find and engage homeless veterans. Of the 44,577 homeless veterans referred to the program during the study period, 994 (2.23 %) were OEF/OIF/OND veterans.

Measures

Information on age, gender, race, marital status, military service, combat exposure, VA service-connected disability income, and homeless and incarceration histories were completed through a structured assessment form by referring clinicians based on veteran self-report, their knowledge of the veteran’s background, and review of their medical records. Combat exposure was defined by whether the veteran “ever received hostile or friendly fire in a combat zone.”

Psychiatric diagnoses were made by referring clinicians at the time of program referral through an assessment form that listed major psychiatric diagnoses, which differentiated between PTSD from combat (combat-related PTSD) and PTSD from non-combat trauma (non-combat-related PTSD). Self-report clinical measures included a mental health symptoms score, which is a sum of eight items from the Psychiatric subscale of the Addiction Severity Index (McLellan et al. 1980); and a social quality of life score, which is the mean score of three items from the Heinrich-Carpenter Quality of Life Scale (Heinrichs et al. 1984). Clinician-rated measures included an observed psychosis rating scale (Dohrenwend 1982); the Global Assessment of Functioning (GAF; American Psychiatric Association 2000); and the Clinician Alcohol and Drug Use Scales (Drake et al. 1996).

Data Analysis

Descriptive statistics were used to summarize the sociodemographic characteristics, psychiatric diagnoses, and clinical status of homeless OEF/OIF/OND veterans by gender. Then, homeless OEF/OIF/OND veterans who reported combat exposure were compared to those who reported no combat exposure on these variables using Mann–Whitney U and Chi-square tests.

Results

Sociodemographic Characteristics

The majority of homeless OEF/OIF/OND veterans referred to HUD-VASH were unmarried, White males, in their 30’s, who had never been incarcerated and been homeless less than twice in the past 3 years (Table 1). These sociodemographic characteristics are similar to previous cohorts of homeless veterans and non-veterans from other data (Rosenheck and Koegel, 1993; Tsai et al. 2012; US Department of Housing and Urban Development, and US Department of Veterans Affairs 2009; Winkleby and Fleshin 1993), except that homeless OEF/OIF/OND veterans were younger, had shorter homeless histories, and appeared to be less likely to have a history of incarceration. The observed differences in homeless and incarceration history are likely due to age, i.e., OEF/OIF/OND veterans are younger and have had less time to be homeless or incarcerated. While we are not able to compare OEF/OIF/OND veterans and previous cohorts of veterans in the same time period after discharge, we are able to provide information on their current differences, i.e., how are homeless OEF/OIF/OND veterans currently different from other homeless veterans of previous eras, and how might their current needs be different.
Table 1

Sociodemographic characteristics and clinical status of homeless male OEF/OIF veterans by combat exposure

 

All (n = 789)

No combat exposure (n = 140)

Combat exposure (n = 649)

Test of differencea

Age

30.87 (7.01)

31.26 (6.50)

30.78 (7.11)

Z = −1.35

Race

 White

363 (46.36 %)

54 (39.42 %)

309 (47.83 %)

χ2(3) = 3.25

 Black

230 (29.37 %)

46 (33.58 %)

184 (28.48 %)

 Hispanic

155 (19.80 %)

30 (21.90 %)

125 (19.35 %)

 Other

35 (4.47 %)

7 (5.11 %)

28 (4.33 %)

Marital status

 Never married

282 (35.92 %)

55 (39.57 %)

227 (35.14 %)

χ2(2) = 1.81

 Divorced/widowed

270 (34.39 %)

49 (35.25 %)

221 (34.21 %)

 Married

233 (29.68 %)

35 (25.18 %)

198 (30.65 %)

Times homeless, past 3 years

 None or one

516 (65.90 %)

91 (66.91 %)

425 (65.69 %)

χ2(2) = 0.28

 Two to four

220 (28.10 %)

36 (26.47 %)

184 (28.44 %)

 Five or more

47 (6.00 %)

9 (6.62 %)

38 (5.87 %)

Lifetime incarceration

 None

446 (57.03 %)

80 (58.39 %)

366 (56.74 %)

χ2(2) = 1.08

 1 year or less

316 (40.41 %)

52 (37.96 %)

264 (40.93 %)

 More than 1 year

20 (2.56 %)

5 (3.65 %)

15 (2.33 %)

 Any VA-service connection

358 (46.19 %)

53 (38.41 %)

305 (47.88 %)

χ2(1) = 4.10

Mental health diagnoses

 Alcohol Abuse/dependency

249 (31.60 %)

43 (30.71 %)

206 (31.79 %)

χ2(1) = 0.06

 Drug abuse/dependency

198 (25.16 %)

30 (21.43 %)

168 (25.97 %)

χ2(1) = 1.26

 Psychotic disorder

37 (4.7 %)

7 (5.00 %)

30 (4.62 %)

χ2(1) = 0.04

 Mood disorder

436 (55.26 %)

67 (47.96 %)

369 (56.86 %)

χ2(1) = 3.77

 Posttraumatic stress disorder

553 (70.09 %)

56 (40.00 %)

497 (76.58 %)

χ2(1) = 73.50**

 Other disorder

284 (35.99 %)

43 (30.71 %)

241 (37.13 %)

χ2(1) = 2.06

Clinical statusb

 Mental health symptom score

1.86 (1.63)

1.31 (1.38)

1.98 (1.65)

Z = −3.58**

 Observed psychosis rating

0.25 (0.35)

0.16 (0.26)

0.26 (0.36)

Z = −2.97*

 GAFc score

59.59 (9.97)

64.25 (11.13)

58.66 (9.46)

Z = −4.16**

 Alcohol abuse rating scale

1.51 (0.74)

1.40 (0.66)

1.54 (0.75)

Z = −1.65

 Drug abuse rating scale

1.25 (0.66)

1.17 (0.54)

1.26 (0.68)

Z = −1.20

 Social quality of life score

3.92 (1.41)

4.32 (1.43)

3.83 (1.39)

Z = −3.12*

* p < 0.01, ** p < 0.001

at tests and Chi-square tests were used to compare veterans with and without combat exposure. Kolmogorov–Smirnov Z tests were used to compare non-normally distributed dependent variables

bSample sizes for clinical status measures were smaller, n = 544 for all male veterans; n = 93 for non-combat veterans, and n = 451 for combat veterans

cGAF Global Assessment of Functioning scale

Among homeless female OEF/OIF/OND veterans, the majority were Black, in their 30’s, unmarried, had been homeless less than twice in the past 3 years and never been incarcerated (Table 2), similar to previous samples of homeless female veterans (Gamache et al. 2003; Leda et al. 1992; Washington et al. 2010), although homeless female OEF/OIF/OND veterans in the current sample were younger and appeared to be more likely to be black.
Table 2

Sociodemographic characteristics and clinical status of homeless female OEF/OIF veterans by combat exposure

 

All (n = 205)

No combat exposure (n = 85)

Combat exposure (n = 120)

Test of differencea

Age

30.62 (7.70)

29.51 (5.98)

31.40 (8.65)

Z = −2.34

Race

 White

57 (27.94 %)

20 (23.53 %)

37 (31.09 %)

χ2(3) = 1.83

 Black

99 (48.53 %)

42 (49.41 %)

57 (47.90 %)

 Hispanic

35 (17.16 %)

17 (20.00 %)

18 (15.13 %)

 Other

13 (6.37 %)

6 (7.06 %)

7 (5.88 %)

Marital Status

 Never married

77 (37.75 %)

35 (41.67 %)

42 (35.00 %)

χ2(2) = 2.53

 Divorced/widowed

103 (50.49 %)

37 (44.05 %)

66 (55.00 %)

 Married

24 (11.76 %)

12 (14.29 %)

12 (10.00 %)

Times homeless, past 3 years

 None or one

134 (66.01 %)

56 (66.67 %)

78 (65.55 %)

χ2(2) = 10.48*

 Two to four

55 (27.09 %)

17 (20.24 %)

38 (31.93 %)

 Five or more

14 (6.90 %)

11 (13.10 %)

3 (2.52 %)

Lifetime incarceration

 None

161 (79.31 %)

69 (81.18 %)

92 (77.97 %)

χ2(2) = 1.39

 1year or less

39 (19.21 %)

14 (16.47 %)

25 (21.19 %)

 More than 1 year

3 (1.48 %)

2 (2.35 %)

1 (0.85 %)

 Any VA service-connection

78 (38.05 %)

29 (34.12 %)

49 (40.83 %)

χ2(1) = 0.95

Mental health diagnoses

 Alcohol abuse/dependency

39 (19.21 %)

16 (19.05 %)

23 (19.33 %)

χ2(1) = 0.00

 Drug abuse/dependency

37 (18.23 %)

16 (19.05 %)

21 (17.65 %)

χ2(1) = 0.07

 Psychotic disorder

9 (4.39 %)

2 (2.35 %)

7 (5.83 %)

χ2(1) = 1.44

 Mood disorder

128 (62.44 %)

51 (60.00 %)

77 (64.17 %)

χ2(1) = 0.37

 Posttraumatic stress disorder

109 (53.17 %)

34 (40.00 %)

75 (62.50 %)

χ2(1) = 10.12*

 Other disorder

68 (33.17 %)

31 (36.47 %)

37 (30.83 %)

χ2(1) = 0.71

Clinical statusb

 Mental health symptom score

1.60 (1.53)

1.33 (1.48)

1.78 (1.54)

Z = −1.93

 Observed psychosis rating

0.22 (0.32)

0.20 (0.31)

0.23 (0.32)

Z = −1.54

 GAFc score

62.05 (9.26)

63.05 (8.62)

61.39 (9.66)

Z = −1.29

 Alcohol abuse rating scale

1.39 (0.65)

1.40 (0.64)

1.39 (0.67)

Z = −0.19

 Drug abuse rating scale

1.11 (0.37)

1.10 (0.35)

1.11 (0.38)

Z = −0.24

 Social quality of life score

4.18 (1.25)

4.33 (1.27)

4.07 (1.23)

Z = −1.47

p < 0.01, ** p < 0.001

at tests and Chi-square tests were used to compare veterans with and without combat exposure. Kolmogorov–Smirnov Z tests were used to compare non-normally distributed dependent variables

bSample sizes for clinical status measures were smaller, n = 155 for all female veterans; n = 63 for non-combat veterans, and n = 92 combat veterans

cGAF Global Assessment of Functioning scale

The majority of homeless OEF/OIF/OND male and female veterans did not report receiving VA service-connected disability payments at the time of referral (46 and 38 %, respectively), which is somewhat lower than then the 50 % found among all VA mental health service users (Edens et al. 2011). Among homeless OEF/OIF/OND veterans who received VA service-connected disability payments, male veterans reported receiving $641.94 (340.97) a month and female veterans $553.86 (319.30) a month, which both equate to a service-connection disability rating of 40–50 % (US Department of Veterans Affairs 2011). This finding contrasts with a recent study that reported higher VA service-connection levels among VA homeless program users, implying such benefits may paradoxically, be a risk for homelessness (Blackstock et al. 2012). This difference may reflect examination of homeless veterans already engaged in VA homeless services instead of at the time of referral.

Psychiatric Diagnoses and Clinical Status

Of the total sample of homeless OEF/OIF/OND veterans, 662 (66.60 %) were diagnosed with PTSD, 564 (56.74 %) with a mood disorder, 373 (37.53 %) with a substance use disorder, and 46 (4.63 %) with a psychotic disorder at the time of referral.

As shown in Tables 1, 2, the majority of homeless OEF/OIF/OND male and female veterans had a diagnosis of PTSD and/or a mood disorder. Additionally, the most common comorbidity among homeless OEF/OIF/OND male and female veterans with a diagnosis of PTSD was a co-occurring mood disorder (62.75 % among males and 77.06 % among females). Of the 553 males with a diagnosis of PTSD, 512 (92.59 %) reported combat-related PTSD and 41 (7.69 %) reported non-combat-related PTSD (with no combat-related PTSD). Of the 109 females with diagnosis of PTSD, 82 (75.23 %) reported combat-related PTSD and 27 (24.78 %) only reported non-combat related PTSD.

Together, results of this study suggest that homeless OEF/OIF/OND veterans have substantially higher rates of PTSD (67 % in this sample) compared to previous cohorts of homeless veterans and non-veterans in which only 8–13 % had PTSD (Kasprow et al. 2000; Mares et al. 2004; Tsai et al. 2012). Further, the current sample had lower rates of psychosis and substance abuse (5 % had a psychotic disorder and 38 % had a substance use disorder), whereas previous cohorts reported 7–21 % for psychotic disorders and 28–80 % for substance use disorders (Kasprow et al. 2000; Mares et al. 2004; Tsai et al. 2012).

Currently, most homeless services focus on substance abuse treatment (Wenzel et al. 2001), “harm reduction” (Tsemberis et al. 2004), and transitional housing to help individuals maintain sobriety (Dordick 2002). This is not surprising given the high prevalence of substance abuse problems among most homeless adults. However, our findings suggest that in contrast to previous cohorts of homeless veterans, a larger proportion of homeless OEF/OIF/OND veterans have PTSD, underscoring the importance of increased clinical attention to this disorder in this population.

Comparison Between Veterans with and Without Combat Exposure

The majority of homeless male (82.26 %) and female (58.54 %) OEF/OIF/OND veterans reported combat exposure. There were no significant sociodemographic differences between those with and without combat exposure for both male and female veterans (Tables 1, 2), except for a significant association between combat exposure and homeless history among female veterans.

With regard to clinical status, male veterans who had combat exposure were more likely to be diagnosed with PTSD and reported greater psychopathology than male veterans who had no combat exposure (Table 1). Female veterans who had combat exposure were also more likely to be diagnosed with PTSD than female veterans who had no combat exposure (Table 2). These findings are consistent with studies on previous cohorts of homeless veterans, which found that combat exposure was associated with greater psychopathology (Rosenheck et al. 1991; Winkleby and Fleshin 1993).

Study Limitations

This study was based on administrative HUD-VASH referral data, which may represent only a subsample of homeless OEF/OIF/OND veterans so the representativeness of the sample is unknown. Comparisons of sociodemographic characteristics and diagnoses between this sample and previous published samples are tentative because of differing study methodologies and varying years of assessment after military discharge. Moreover, combat exposure was not comprehensively assessed and relied on one item and the clinical judgment of referring clinicians. Administrative data on service branch was not available because this information was not considered important for delivery of HUD-VASH services, so further research is needed to examine differences by service branch, including reserve components such as the National Guard.

Discussion

OEF/OIF/OND veterans constitute less than 3 % of all homeless veterans referred to the VAs flagship supported housing program, as compared to 12 % of OEF/OIF/OND veterans in the general population (Westat, and Department of Veterans Affairs 2010) and 11 % of OEF/OIF/OND veterans among VA service users (Edens et al. 2011). Thus, OEF/OIF/OND veterans appear to be at lower risk for homelessness than veterans from other military service eras (Kasprow and Rosenheck 2011). Besides being younger, homeless OEF/OIF/OND veterans were sociodemographically similar to previous cohorts of homeless veterans. However, unlike previous cohorts of homeless veterans, the majority of homeless OEF/OIF/OND veterans report combat exposure and have been diagnosed with PTSD. Most homeless OEF/OIF/OND veterans also do not receive any VA-service connected disability (although many do), suggesting that VA service-connection is a protective factor against homelessness. Taken together, these findings underscore the importance of promoting treatment for PTSD and efforts to help homeless OEF/OIF/OND veterans obtain VA service-connected disability or pension benefits when they meet eligibility criteria.

Copyright information

© Springer Science+Business Media, LLC (outside the USA) 2012