Health care of homeless veterans

JGIM BRIEF REPORTS Blackwell Publishing Ltd. Health Care of Homeless Veterans O’Toole et al., Health Care of Veterans Why Are Some Individuals Falling Through the Safety Net? Thomas P. O’Toole, MD, Alicia Conde-Martel, MD, Jeanette L. Gibbon, MD, MPH, Barbara H. Hanusa, PhD, Michael J. Fine, MD, MSc It is important to understand the needs of those veterans who are homeless. We describe characteristics of homeless male veterans and factors associated with needing VA benefits from a two-city, community survey of 531 homeless adults. Overall, 425 were male, of whom 127 were veterans (29.9%). Significantly more veterans had a chronic medical condition and two or more mental health conditions. Only 35.1% identified a community clinic for care compared with 66.8% of non-veterans (P < .01); 47.7% identified a shelter-based clinic and 59.1% reported needing VA benefits. Those reporting this need were less likely to report a medical comorbidity (58.7% vs 76.9%; P = .04), although 66.7% had a mental health comorbidity and 82.7% met Diagnosic Screening Manual (DSM)-IIIR criteria for substance abuse/dependence. They were also significantly more likely to access shelter clinics compared with veterans without this need. Homeless veterans continue to have substantial health issues. Active outreach is needed for those lacking access to VA services. to compare demographic characteristics, comorbid conditions, and sources of usual care among homeless male veterans and non-veterans and to specifically look at characteristics of homeless veterans who report needing VA benefits. KEY WORDS: homelessness; mental health; needs assessment; substance abuse; veterans. Inclusion criteria were age greater than 18 years and homelessness for the majority of the previous 3 months. Participants received 5 dollars in cash or in bus tokens and a listing of area service providers upon completing the interview. J GEN INTERN MED 2003; 18:929–933. V eterans are disproportionately represented in surveys of homeless men, ranging from 38% to 42% in 1–3 community samples. The needs of homeless veterans, particularly those exposed to combat or from the Vietnam era, have been substantial, especially for physical injury, 4–6 psychiatric illness, alcohol abuse, and medical problems. It is important to understand the needs of those veterans who are currently homeless. Are they still a unique subgroup among homeless persons? How effective are current Veterans Administration (VA) efforts at serving them? We present data from a two-city, community-based survey of 531 homeless adults. The aims of this project are Received from the Johns Hopkins University School of Medicine (TPO); Hospital Universitario de Gran Canaria, Spain (AC-M); Maricopa County Department of Public Health, Phoenix, Ariz (JLG); Center for Research on Health Care, Division of General Internal Medicine, University of Pittsburgh (BHH); and the VA Center for Health Equity Research and Promotion, VA Pittsburgh Health Care System, Pittsburgh, Pa (MJF). Address correspondence and requests for reprints to: Dr. O’Toole, Welch Center, Rm 2 –513, Johns Hopkins University, 2024 E. Monument Street, Baltimore, MD 21205 (e-mail: [email protected]). METHODS We conducted a cross-sectional, community-based survey of homeless adults in Pittsburgh and Philadelphia using face-to-face interviews over a 5-month period: April to August, 1997. Approval from the Institutional Review Boards at the University of Pittsburgh and the University of Pennsylvania was obtained. The response rate was 93%. Study Population Study Sites Survey sites were clustered as: 1) unsheltered enclaves (abandoned buildings, cars, and outdoors) and congregate eating facilities; 2) emergency shelters; and 3) transitional housing or single-room occupancy (SRO) dwellings to ensure representation by all sheltering-based subgroups of homeless persons. Sites within each cluster were selected with probabilities proportional to size (pps sampling) for individual recruitment, with either random or systematic sampling at each selected site depending on capacity. The selection of interviewees varied depending on the type of site and the number of people present at the time of the interview, using 1 of 4 selection plans. The details of the selection strategy have been explained elsewhere.7 All interviews were audiotaped with 10% randomly selected for review each week to ensure data integrity. Survey Instrument The National Technical Center (NTC) Telephone Substance Dependence Needs Assessment Questionnaire was 8 used after being modified to accommodate face-to-face 929 930 O’Toole et al., Health Care of Veterans interviews and to include questions specific to homeless3,9 ness from previously administered questionnaires. Definitions of substance abuse and dependence followed the 10 Diagnostic Screening Manual (DSM)-IIIR criteria. Individuals were asked whether they were veterans of the Vietnam War or Gulf War, and if they were currently receiving VA benefits, their health care at a VA facility, or had VA health insurance. Respondents were also asked to rate potential needs from a list of 18 categorical options. Eligibility for veterans programs was not assessed as part of the survey. The Behavioral Model for Vulnerable Populations, which groups factors into categories of predisposing, illness, and need, was used as the theoretic framework for 11 assessing our findings. Methods of Analysis Only 2 of the 129 veterans identified in this sample were women. Therefore, analyses are presented comparing only homeless men. Differences between veterans and nonveterans were assessed with chi-squared or Fisher’s exact tests for categorical data and Student’s t test for continuous data. Independent variables for being a homeless veteran were assessed in a multiple logistic regression model. SPSS 10.0 (Chicago, IL) and StatXact (Cambridge, MA) statistical software packages were used for analyses. RESULTS Overall, 531 homeless adults were interviewed in Pittsburgh (N = 267) and Philadelphia (N = 264), PA, of whom 425 were male. Among male veterans, 127 (45.0%) were Vietnam War veterans and 11.6% were Gulf War veterans. Demographic Characteristics Homeless male veterans were significantly older (43.4 vs 39.2 years, P < .01), better educated (>12th grade education; 86.6% vs 64.3%, P < .01), and less likely to have been arrested in the previous 12 months (10.2% vs 20.5%, P = .01). In the multiple logistic regression model, only age > 40 years (OR 1.1; 95% confidence interval [CI] 1.0 to 1.1) and having at least a 12th grade education (OR 3.5; 95% CI, 1.9 to 6.4) were independently associated with veteran status. Comorbidities, Health Insurance, and Source for Usual Care Homeless male veterans were significantly more likely to report a chronic medical condition (66.1% vs 55.4%; P = .04) and 2 or more mental health conditions (33.1% vs 22.2%; P = .02), with higher rates of hepatitis/cirrhosis (18.9% vs 7.0%, P < .01) and posttraumatic stress disorder (18.1% vs 8.1%, P < .01) reported. There was no difference in DSM-IIIR defined alcohol or drug abuse/dependence between veterans and non-veterans, although rates for both groups were extremely high (79.5% and 82.6%, respectively) (Table 1). Homeless male veterans were significantly more likely to report going to a shelter-based clinic or street outreach JGIM team for usual care (47.7% vs 0%, P < .001), while nonveterans were significantly more likely to report going to a community clinic (66.8% vs 35.1%, P < .01) or an emergency department (17.0% vs 9.0%, P = .05). Homeless male veterans were also significantly more likely to have health insurance (67.5% vs 54.4%, P = .01) with most of this difference due to VA coverage. Overall, 41.1% of veterans reported accessing the VA health system, including VA emergency departments, hospital- or community-based clinics, or, in Pittsburgh, a nurse-administered homeless outreach program. Daily Sustenance and Self-Identified Needs Money from odd jobs or steady employment was the most frequently identified income source (veterans: 50.4%; non-veterans: 45.3%). Non-veterans more likely reported income from hustling / stealing (21.1% vs 8.7%, P < .01) while veterans more likely reported income from selling plasma (17.3% vs 11.4%, P = .01). Overall, 59.1% of veterans surveyed reported needing VA benefits. There were no demographic differences between these individuals and those not needing benefits. They were significantly less likely to report a medical comorbidity (58.7% vs 76.9; P = .04) although they had comparable rates of mental health comorbidities (66.7%, 53.8%, P = .19) and alcohol or drug abuse/dependence (82.7%, 75.0%, P = .37). This group was significantly less likely to access a community clinic (25.0% vs 48.9%, P = .02) but more likely to access a shelter-based clinic or street outreach team (56.3% vs 36.2%, P = .04). (Table 2) They were also more likely to report needs in 10 of 18 service categories, including physical health care, substance abuse treatment, and entitlement assistance. DISCUSSION The data presented suggest homeless male veterans continue to have greater medical and mental health needs than non-veteran homeless men. They are also significantly less likely to access community health centers, instead relying on shelter-based and street outreach services for care. These findings support the importance of continuing to distinguish homeless persons by veteran status and the ongoing need to tailor interventions and services for this subgroup. The VA has developed some of the most comprehensive and successful models for homeless care and has made a substantial commitment to their needs.12,13 In 1998, 20% of the VA annual inpatient mental health budget ($404 12 million) was spent on homeless persons. However, while homeless veterans without public assistance are more likely to access VA programs, researchers also found that the majority of individuals in domiciliary care programs were 14 already established patients in the VA. How well positioned this system is to provide outreach to those veterans not engaged in VA-based care requires further investigation. In our study, the majority of homeless male veterans reported needing veterans benefits. Those reporting this need were significantly less likely to be accessing community 931 Volume 18, November 2003 JGIM Table 1. Demographics, Comorbidities, and Sources for Usual Care of Veteran and Non-veteran Homeless Males Veterans (%) N = 127 Non-veterans (%) N = 298 P Value Mean age, y (± SD) 43.4 (±8.2) 39.2 (±9.0) <.01 Race African-American 88.3 (106) 82.2 (326) .14 Education ≥12th grade or equivalent 86.6 (110) 64.3 (191) <.01 Marital status Single/ divorced/ widow 97.6 (124) 95.3 (284) .42 Duration homelessness <12 months 61.2 (79) 61.1 (245) .98 Health insurance 67.5 (86) 54.4 (163) .01 Employment status Unemployed 63.0 (80) 64.4 (192) .83 Monthly income < $250 37.0 (47) 32.9 (98) .44 Arrested (last 12 months) 10.2 (13) 20.5 (61) .01 Chronic medical condition 66.1 (84) 55.4 (164) .04 Two or more chronic medical conditions Hypertension Hepatitis / cirrhosis Diabetes Heart disease 36.2 24.4 18.9 7.1 7.1 29.1 22.8 7.0 6.7 4.4 Any psychiatric condition 61.4 (78) 54.2 (161) Two or more psychiatric conditions Depression Anxiety disorder Posttraumatic stress disorder Bipolar disorder 33.1 37.8 16.5 18.1 10.2 22.2 34.3 12.8 8.1 6.4 Any alcohol or drug abuse /dependence Source for usual care Emergency department Community clinic Hospital clinic Shelter-based clinic/street outreach (46) (31) (24) (9) (9) .17 .71 <.01 .84 .24 .20 (66) (102) (38) (24) (19) .02 .51 .36 <.01 .23 79.5 (101) 82.6 (246) .49 9.0 35.1 7.2 47.7 17.0 (39) 66.8 (153) 13.1 (30) 0 .05 <.01 .14 <.01 clinics, relying instead on shelter clinics and street outreach for their care. While they were less likely to have a medical condition than those without a need for veterans benefits, over half had at least one chronic medical condition, two thirds reported a chronic mental health condition, and 82.7% met criteria for alcohol or drug abuse/ dependence. The need for VA benefits is relevant to receipt of health services, and especially to the scope of services indicated for this population. Earlier research demonstrated that facilitating enrollment in VA programs and receipt of VA benefits is likely to increase utilization of VA services and improve quality of life.15 Gamache et al. found that “enabling” factors, including entitlement to VA services and proximity to VA facilities, were more important than “predisposing” (demographics, wartime experience) or “illness” 16 factors in predicting VA service use. Wenzel et al. identified “need” factors (chronic medical problems, mental health needs, substance abuse) as more strongly related to health services use in a cohort of homeless veterans in Los (42) (48) (21) (23) (13) (86) (68) (21) (20) (13) (10) (39) (8) (53) 17 Angeles. Finally, health insurance, recent medical need, and comorbid conditions were all independently associated with identifying an ambulatory clinic for usual care among 18 Pittsburgh, PA homeless persons. It is likely that increasing enrollment in VA services alone will not be enough. An exploratory study of homeless veterans’ perspectives on social service use identified high 19 levels of stress and frustration with the delivery system. How effectively services are tailored to homeless veterans’ unique needs will also be important. This is likely to be relevant to the effectiveness of the Homeless Veterans Comprehensive Services Program Act, which allows the VA to broaden the 20 pool of providers serving homeless veterans. Over half of the respondents who reported needing veterans benefits were accessing shelter clinics and street outreach teams. This argues for enhancing outreach and engagement at sites where homeless congregate as critical to any care plan. The scope of need identified by homeless veterans needing VA benefits further supports the importance of wrap-around and 932 O’Toole et al., Health Care of Veterans JGIM Table 2. Demographics, Comorbidities, and Sources for Usual Care of Veterans Needing and Not Needing Veterans Benefits Perceived Need (%) N = 75 No Perceived Need (%) N = 52 P Value Mean age, y (± SD) 43.1 (7.7) 43.8 (8.9) .65 Race African-American 88.4 (61) 88.2 (45) .98 Education ≥12th grade or equivalent 84.0 (63) 90.4 (47) .43 Marital status Single/ divorced/ widow 97.3 (73) 98.1 (51) .99 Duration homelessness <12 months 64.0 (48) 55.8 (29) .36 Health insurance 68.0 (51) 66.7 (35) .88 Employment status Unemployed 64.0 (48) 61.5 (32) .85 Monthly income < $250 34.7 (26) 40.4 (21) .58 Arrested (last 12 months) 9.3 (7) 51.4 (38) 11.5 (6) 38.5 (20) .77 .20 58.7 (44) 76.9 (40) .04 Vietnam veteran Chronic medical condition 30.7 (23) 44.2 (23) .14 66.7 (50) 53.8 (28) .19 36.0 (27) 28.8 (15) .45 82.7 (62) 75.0 (39) .37 Source for usual care Emergency department Community clinic Hospital clinic 9.4 (6) 25.0 (16) 7.8 (5) 8.5 (4) 48.9 (23) 6.4 (3) 1.0 .02 1.0 Shelter-based clinic/ street outreach 56.3 (36) 36.2 (17) .04 Two or more chronic medical conditions Any psychiatric condition Two or more psychiatric conditions Any alcohol or drug abuse /dependence coordinated services. VA “Stand Downs” held in several cities for homeless veterans are one example of how these services 21 can be concentrated and tailored for this population. There are several limitations to consider when interpreting these findings. First, the data are self-reported and not validated by any collaborating sources. It is possible that respondents underreported the amount of care they were receiving at the VA or overreported their VA eligibility. We did not verify veteran status, including whether they had an honorable or dishonorable discharge from the service. Given the paucity of veteran women in our sample, we only report on male veterans and our results cannot be assumed to apply to women also. Finally, the data presented are on an urban homeless population and cannot be generalized to suburban or rural settings. In summary, veterans are disproportionately represented in homeless samples and continue to have substantial needs. Special attention must also be given to engaging homeless veterans not currently accessing services or receiving benefits. This project was funded by a contract from the Center for Substance Abuse Treatment, 270-95-0009. Dr. O’Toole is funded by an NIDA career development award K23DA13988-01. REFERENCES 1. Rosenheck R, Frisman L, Chung AM. The proportion of veterans among homeless men. Am J Public Health. 1994;84:466 – 9. 2. Winkleby MA, Fleshin D. Physical, addictive, and psychiatric disorders among homeless veterans and nonveterans. Public Health Rep. 1993;108:30– 6. 3. O’Toole TP, Gibbon JL, Hanusa BH, Fine MJ. Utilization of health care services among subgroups of homeless and housed poor. J Health Polit Policy Law. 1999;24:91–114. 4. Rosenheck R, Fontana A. A model of homelessness among male veterans of the Vietnam War generation. Am J Psychiatry. 1994;151:421– 7. 5. Rosenheck R, Gallup P, Leda CA. Vietnam era and Vietnam combat veterans among the homeless. Am J Public Health. 1991;81:643 – 6. 6. Rosenheck R, Leda C, Gallup P. Combat stress, psychosocial adjustment, and service use among homeless Vietnam veterans. Hosp Community Psychiatry. 1992;43:145– 9. 7. O’ Toole TP, Gibbon JL, Seltzer D, Hanusa BH, Fine MJ. Urban homelessness and poverty during economic prosperity and welfare reform: changes in self-reported comorbidities, insurance, and sources for usual care, 1995–97. J Urban Health. 2002;79:200– 10. 8. National Technical Center for Substance Abuse Needs Assessment. Available at: www.ntc.org. Accessed July 17, 1998. 9. Robertson MJ, Ropers RH, Boyer R. The homeless of Los Angeles: an empirical evaluation. The Federal Response to the Homeless Crisis. Hearing before the Subcommittee of the Committee on Government Operations, U.S. House of Representatives. Washington, DC: Government Printing Office 1016A 1016B MF; 1985. JGIM Volume 18, November 2003 10. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, 3rd edn. Revised. Washington, DC: American Psychiatric Association; 1987. 11. Gelberg L, Anderson RM, Leake BD. The Behavioral Model for Vulnerable Populations: applications to medical care use and outcomes for homeless people. Health Serv Res. 2000;34:1273– 302. 12. Rosenheck R, Seibyl CL. Homelessness: health service use and related costs. Med Care. 1998;36:1256– 64. 13. Kizer KW, Fonseca ML, Long LM. The veterans healthcare system: preparing for the twenty-first century. Hosp Health Serv Adm. 1997;42:283– 98. 14. Rosenheck R, Leda C. Who is served by programs for the homeless? Admission to a domicilliary care program for homeless veterans. Hosp Community Psychiatry. 1991;42:176– 81. 15. Rosenheck RA, Dausey DJ, Frisman L, Kasprow W. Outcomes after initial receipt of social security benefits among homeless veterans with mental illness. Psychiatr Serv. 2000;51:1549 – 54. 933 16. Gamache G, Rosenheck R, Tessler R. Factors predicting choice of provider among homeless veterans with mental illness. Psychiatr Serv. 2000;51:1024– 8. 17. Wenzel SL, Bakhtiar L, Caskey NH, et al. Homeless veterans’ utilization of medical, psychiatric, and substance abuse services. Med Care. 1995;33:1132 – 44. 18. O’Toole TP, Gibbon JL, Hanusa BH, Fine MJ. Preferences for sites of care among urban homeless and housed poor adults. J Gen Intern Med. 1999;14:599 – 605. 19. Applewhite SL. Homeless veterans: perspectives on social services use. Soc Work. 1997;42:19– 30. 20. VA homeless providers grant and per diem program clarification of per diem eligibility—VA Final rule. Fed Regist. 1997;62 :6121 – 6. 21. Information on Veterans Administration-sponsored homeless ‘standdowns.’ Available at: http://www.nchv.org/standdown.cfm. Last accessed May 2, 2003.