Elderly Homeless Veterans in Los Angeles: Chronicity and Precipitants of Homelessness

Abstract

Objectives. We compared the characteristics of chronically homeless and acutely homeless elderly veterans to better understand precipitants of homelessness.

Methods. We conducted interviews with 33 chronically and 26 acutely homeless veterans aged 65 years and older receiving transitional housing services in Los Angeles, California, between 2003 and 2005. We asked questions regarding their sociodemographic characteristics and other social status measures. Other precipitants of homelessness were acquired via observation and open-ended and structured questions.

Results. Both veterans groups were more similar than different, with substantial levels of physical, psychiatric, and social impairment. They differed significantly in homelessness history, with chronically homeless veterans having more homelessness episodes and more total time homeless. They were also less educated and had smaller social networks. In response to open-ended questioning, elderly homeless veterans revealed how health and substance use issues interacted with loss of social support and eviction to exacerbate homelessness.

Conclusions. Assessment of a range of factors is needed to address risk factors and events leading to homelessness. Further research with larger samples is needed to confirm the characteristics and needs of the elderly homeless veteran population.

In 2007, the US Departments of Housing and Urban Development and Veterans Affairs conducted the first comprehensive census of homeless people and found that a disproportionate number of the homeless population were veterans.¹ The population of veterans overall has been aging: veterans older than 65 years accounted for 38% of the total veteran population in 1999, and the most current estimate projected this rate rose to 42% in 2011.² Elderly veterans are approximately twice as likely to be homeless as elderly civilians: 39% of homeless veterans were aged 51 to 61 years and 9% of homeless veterans were aged 62 years or older (compared with 19% and 4%, respectively, for same-aged groups of homeless nonveterans).¹

Researchers have proposed a cumulative risk model composed of risk and protective factors to explain homelessness among elderly people.^3–6 Risk can include life events (death of spouse, marital breakdown, exiting employment, evictions), problem conditions (mental illness or medical conditions), and internal and external factors (minority status and higher levels of disruptive events during childhood, including parental incarceration or history of substance abuse).^4,7,8 Multiple risk factors, including substance abuse and increased mental health problems, increase the likelihood of negative outcomes.⁹ To date, little research has been done on the relationship between such risk factors and their impact on the trajectory of homelessness for either elderly homeless nonveterans or elderly homeless veterans.¹⁰ Relatedly, in recent years, risk for chronic homelessness has been a major concern in the literature on homelessness. Chronically homeless individuals have been found to be more likely to have more severe mental health, physical, and substance abuse issues.^11–13 Little research has been done on chronic homelessness among elderly individuals,¹⁴ with no attention paid to homeless elderly veterans.

The research that does exist has found that—in comparison with nonelderly homeless people—homeless elderly people experience a wider, more intense array of medical, psychological, and social problems¹⁵; cognitive impairments and dementia¹⁵; greater sensitivity to the effects of alcohol and drug use¹⁶; dramatically lower rates of being married¹⁷; and weaker social ties.¹⁸ These issues can affect elderly people’s ability to make housing decisions and can increase the likelihood of eviction or abandonment by family or other support persons.¹⁹ Other research has suggested that the length of time elderly people are homeless has been found to vary in relation to the timing of events during different life stages, for example, disruptive events (e.g., foster care, parental separation, incarceration) during childhood; the impact of limited education, health, and substance abuse problems during middle age on occupational functioning; and significant loss in older age.⁷

For veterans across all age groups, a large body of research comparing the characteristics of homeless veterans with those of homeless nonveterans has indicated that their profiles contrast: homeless veterans are slightly more likely to be White, be better educated, be more likely to have a history of marriage (as well as divorce), and be more likely to have serious alcohol problems than drug problems.^1,7,13,20 The cumulative risk model has also been applied to homeless veterans and has shown that additional risk factors such as combat exposure, wartime trauma, and posttraumatic stress disorder (PTSD) increase vulnerability to homelessness.²¹

Cohen³ has concluded that older homeless people are invisible to researchers, policymakers, and the public at large. To provide visibility to elderly homeless veterans as a significant homeless elderly population and contribute to the understanding of homeless chronicity in this group, we conducted interviews with a sample of elderly homeless veterans receiving transitional housing services between 2003 and 2005. Given the specific vulnerabilities of veterans, we had 2 goals in examining issues relevant to homelessness for this sample. The first was to examine their characteristics, comparing those who were chronically homeless with those who were acutely homeless. The second was to better understand what precipitated homelessness through the veterans’ own narratives and structured questioning across a broad range of potentially precipitating factors.

METHODS

Participants were recruited at 2 Department of Veterans Affairs (VA)–funded specialized transitional living programs for homeless elderly veterans located in Los Angeles, CA. Veterans can receive services in these programs for as long as 2 years. To be eligible for the study, veterans had to be homeless at program entry and aged least 65 years.

Staff from these programs identified veterans and determined their interest in participating in study. Graduate social work research interviewers screened participants with a short series of questions designed to ensure that participants were able to understand and to provide informed consent for the study.

Measures

Variables measured—counting those identified in Cohen’s³ framework—included sociodemographic characteristics, combat status, disruptive events during youth, social support, criminal history, health and mental health status, service needs, barriers to housing, and factors leading to homelessness.

Chronic homelessness status was identified using the US Department of Housing and Urban Development (HUD) definition of chronic homelessness: continuous homelessness for a year or more or 4 or more episodes of homelessness over the past 3 years, accompanied by a disabling condition.²² All veterans in the study reported a disabling medical or psychiatric condition defined as a serious medical condition or psychiatric diagnosis, so we differentiated chronicity by homeless history.

Sociodemographic characteristics measured included age, race/ethnicity, marital status, years of education completed, and income in the past 30 days. As an indicator of combat status, veterans were asked how many months they had spent in a war zone.

Other social status measures included troubles in childhood, criminal history, and social support. We calculated disruptive childhood events, measured as historical stressors, as a sum across 11 events, including, for example, living in foster care, physical abuse, or parental estrangement.^23,24 We assessed criminal history through number of arrests and number of months of lifetime incarceration.

Interviewers used the Lubben Social Network scale to assess instrumental and emotional social support among family and social support among neighbors or friends.²⁵ In addition, they used a general social network measure to assess the number of relatives, friends, and acquaintances respondents had in the Los Angeles area.

Veterans’ physical health status was assessed by asking whether the veteran had ever been told by a doctor or nurse practitioner that he had any of 22 serious health problems such as high blood pressure, hypertension, or hepatitis A.²⁶ A summary score of the number of chronic and acute conditions identified made up the current health status score.

For mental health status, interviewers asked veterans whether a doctor had ever given them a psychiatric diagnosis of schizophrenia, PTSD, depression, bipolar disorder, or other serious emotional problem. Past and present alcohol and illicit drug use problems were assessed by asking, “Do you have a problem with alcohol or drug dependency now?” and “Have you had a problem with alcohol or drug dependency in the past?” Interviewers also assessed for cognitive impairment using the Mini-Mental State Examination, a brief 30-point questionnaire testing arithmetic, memory, and orientation functions.²⁷

Service needs scores were derived by first asking respondents to pick services they currently needed (including those services currently being received) from a list of services. We then collapsed the services identified as needed into 7 categories: (1) mental health services, (2) health services, (3) substance abuse services, (4) financial services, (5) housing services, (6) case management services, and (7) nursing services.

Interviewers assessed barriers to becoming rehoused and the veteran’s explanation of current homeless status through a series of questions. Veterans were asked what was preventing them from living in permanent housing. We collapsed responses to this open-ended question into the following categories: (1) financial problems, (2) record of institutionalization in prison (or psychiatric hospital), (3) eviction record, (4) health problems, (5) not interested in permanent housing, and (6) other.

Interviewers assessed precipitants of homelessness using observation and open-ended and structured questions. Veterans’ responses to open-ended questions (“How did you first end up in the program?”) specifically referenced the homelessness episode before program entry and provided veterans’ perceptions of entry into homelessness. The responses to open-ended questions were transcribed, along with observational notes on each respondent, and examined for patterns using Atlas.ti version 7 (Atlas.ti Scientific Software Development GmbH, Berlin, Germany). Initial thematic analysis revealed 7 different factors leading to homelessness. One author (C. v. d. B.-C.) coded responses and observations into 7 categories of veteran-explained reasons for becoming homeless: (1) financial, (2) landlord, (3) health, (4) personal crisis, (5) social support, (6) substance abuse, and (7) mental health. Further thematic analysis provided a deeper understanding of how the categories interacted.

Interviewers obtained information explaining the onset of homelessness later in the interview when veterans responded to a series of structured queries preceded by “Now I am going to ask you about some reasons that people become homeless. Which of the following led you to become homeless most recently?” Eighteen possible reasons were queried. To provide comparison with the open-ended question responses, we coded these reasons into the same 7 veteran-explained reason categories.

Analyses

We compared characteristics of chronic and acutely homeless veterans using the χ² test for categorical variables and the t test for continuous variables. We then used the χ² test (Fisher exact test when appropriate) to compare the frequency of structured and unstructured responses within each category; a paired t test was used to examine differences in the total number of precipitants across unstructured and structured responses.

RESULTS

A little more than half (56%) of the veterans were chronically homeless at admission. Veterans in both chronic and acute groups were, on average, aged 74 years. Non-Hispanic Black veterans made up the highest percentage of the sample of respondents (42%), but we found no differences in race/ethnicity between those who were chronically homeless and those who were acutely homeless. Similarly, we found no differences in marital status. The percentage married, 5% across the 2 groups, is similar to that in other homeless populations.¹⁷ For the overall sample, the percentage without a high school diploma was lower than national averages for elderly people aged 65 years and older (27% for elderly homeless veterans compared with 35% for the general population).²⁸ Acutely homeless veterans were more likely to have more years of education than chronically homeless veterans. Most of the veterans in the sample had served in the World War II through the post–Korean and pre–Vietnam War eras. Average war zone exposure was slightly more than 6 months, with no differences between the 2 groups (Table 1).

TABLE 1—

Sociodemographic and Health Characteristics, Service Needs, and Housing Barriers of Elderly Homeless Veterans (n = 59): Los Angeles, CA, 2003–2005

Characteristic	Total Sample (n = 59), % or Mean ±SD	Chronically Homeless Elderly Veterans (n = 33), % or Mean ±SD	Acutely Homeless Elderly Veterans (n = 26), % or Mean ±SD	P
Age	74 ±5.8	74 ±5.7	74 ±6.0	.4
Male	100	100	100	NS
Race/ethnicity
Alaskan Native or Native American	2	3	0	.37
Black, non-Hispanic	42	49	35	.28
Hispanic	7	3	12	.2
White, non-Hispanic	37	27	50	.07
Other	12	18	4	.11
Married	5	3	8	.56
Education
GED	27	32	19	.34
Years of education	13.0 ±2.4	12.5 ±2.1	13.8 ±2.6	.05
Military
Service era
Pre-WW II	2	0	2	.44
WW II	17	15	19	.47
Post–WW II to pre–Korean War	12	12	12	.64
Korean War	48	52	42	.48
Post–Korean War to pre–Vietnam War	46	42	50	.56
Vietnam War	10	9	12	.54
War zone exposure, mo	7.5 ±11.6	6.6 ±10.3	8.7 ±13.1	.24
Social
Disruptive childhood events	2.8 ±2.0	2.7 ±2.1	3.1 ±1.7	.42
Homelessness
Lifetime homelessness events	3.1 ±4.0	4.4 ±5.1	1.6 ±0.9	< .01
Lifetime homelessness, y	3.3 ±5.6	5.8 ±6.6	0.3 ±0.2	< .01
Criminal history or status
Lifetime arrests	6.1 ±9.5	6.4 ±9.4	5.8 ±9.8	.82
Lifetime mo in jail or prison	25.0 ±84.7	16.9 ±39.4	37.6 ±123.2	.43
On parole or probation	11	13	9	.67
Income and employment
Income received past mo, $	962 ±358	936 ±31.6	998 ±412	.43
Longest job, y	15.8 ±12.3	14.0 ±10.0	18.3 ±14.7	.11
Social network and support
Close friends or relatives nearby	5.7 ±8.2	5.2 ±8.9	6.5 ±7.3	.56
People could rely on for instrumental support	1.8 ±1.7	1.4 ±1.4	2.3 ±1.8	.05
People could rely on for emotional support	1.6 ±1.5	1.3 ±1.3	2.0 ±1.7	.09
Physical health
Serious health problems	4.5 ±3.1	4.6 ±3.0	4.5 ±3.3	.87
Hypertension	60	56	65	.48
Chest infection or bronchitis	17	13	23	.29
Pneumonia	5	3	8	.44
Lung problem	22	22	23	.91
Hearing problem	33	34	31	.77
Ear, nose, or throat problem	21	22	19	.81
Eye or vision problem	41	47	35	.35
Cancer	12	16	8	.43
Heart trouble	24	28	19	.73
Stroke	12	6	19	.13
Kidney or bladder trouble	24	28	19	.43
Arthritis or rheumatism	45	47	45	.73
HIV positive or AIDS	0	0	0	NS
Hepatitis A, B, or C	12	13	12	.91
Diabetes	19	19	19	.96
Stomach or digestive disorder	24	25	23	.87
Anemia	12	9	15	.49
Pancreatitis	2	0	4	.26
Thyroid disease	2	3	0	.37
Skin disorders	16	13	19	.48
Seizures	5	6	4	.68
Back or neck problems	46	52	39	.32
Mental health
Serious psychiatric problem	1.1 ±1.2	1.3 ±1.4	0.9 ±1.0	.15
Schizophrenia	7	6	8	.77
PTSD	14	16	13	.74
Bipolar disorder	13	9	17	.41
Depression	32	41	21	.12
Alcohol abuse	25	31	17	.21
Alcohol dependency	43	47	38	.48
Drug abuse	23	28	17	.32
Drug dependency	41	50	29	.12
MMSE score	26.5 ±3.0	26.3 ±2.8	26.9 ±3.4	.74
Service need
Mental health	38	44	30	.32
Health	95	100	87	.04
Substance abuse	22	28	13	.18
Financial	73	72	74	.87
Housing	100	100	100	NS
Case management	71	66	78	.31
Nursing	27	34	17	.16
Housing barriers
Financial	88	97	77	.04
Criminal or psychiatric Institutionalization record	16	16	16	.97
Eviction	19	15	23	.44
Health	20	18	23	.64
Does not want to be housed	16	15	16	.93
Other	37	42	31	.36

Note. GED = general equivalency diploma; MMSE = Mini-Mental State Examination; NS = not significant; WW II = World War II.

With regard to background, the 2 groups had a mean of 3 disruptive childhood events or stressors, with no difference in the summary scores between the 2 groups. The most frequently experienced events before age 18 years included loss of 1 or both parents (52%), family income low enough that it was sometimes not possible to meet monthly expenses (42%), and having had friends in trouble with the law or school authorities (29%; data not shown). The 2 groups did differ significantly in homelessness history: on average, chronically homeless veterans had 3 times the number of homelessness episodes and 5.5 more years of total time homeless than the acutely homeless group. Both groups had histories of substantial numbers of arrests and time incarcerated, but differences between the 2 groups were not significant.

Current parole or probation status indicated no difference in monitoring by the criminal justice system. We also identified no differences in current income from all sources: both chronic and acutely homeless veterans earned approximately $1000 a month. Although the size of social networks was similar, acutely homeless veterans had a larger social support system to rely on for instrumental help and a trend toward a larger emotional support system.

Physical and mental health variables revealed a range of clinical problems, but we found no significant differences between the 2 groups. The groups did not differ in cognitive impairment as measured by Mini-Mental State Examination. The participants’ scores (27 ±1.5) were similar to those of seniors of the same age and education.²⁹

For service needs among the overall sample, housing was reported as the top need for all veterans, followed by needs for health, financial, and case management services. Chronically homeless veterans were more likely to report service needs related to health and financial barriers to being rehoused.

Precipitants Leading to Homelessness

Three quarters of the sample (n = 44) answered the open-ended question regarding factors relating to their homelessness. Veterans identified a range of homelessness precipitants in their responses to the structured and unstructured questions (Table 2). We found no significant differences between responses to structured and unstructured questions in each of the precipitant categories, but respondents did offer more precipitants to homelessness when given structured responses. However, the average number of precipitating factors within structured and unstructured response categories did not differ between the chronically and acutely homeless groups (data not shown).

TABLE 2—

Factors Precipitating Current Homelessness Episode by Unstructured or Structured Responses (n = 44): Los Angeles, CA, 2003–2005

Precipitants	Unstructured, % or Mean ±SD	Structured, % or Mean ±SD	P
Financial	23	77	.58
Eviction	26	42	.48
Physical health	26	33	.49
Crisis	5	2	.95
Social support	37	30	.59
Substance abuse	33	33	.09
Mental health	7	24	.43
Legal or incarceration	11	14	.54
Average no. of factors	1.6 ±0.6	2.5 ±0.6	< .01

Interaction of Precipitants Leading to Homelessness

The narratives of veterans’ unstructured responses to open-ended questions provide helpful details about how precipitants leading to homelessness interact with each other. Richard and Jameson (all respondents’ names are fictitious) provided typical responses, listing loss of social support as a reason for homelessness. Respondents commonly answered that they became homeless after leaving (or being evicted from) the home of a son or daughter. However, they generally seemed to do so willingly, seeing their substance abuse, health issues, or mental health issues as a cause of additional burden to their relatives. For example, Richard, an alcoholic who occasionally convinces his daughter to bring him to a sober living facility, illustrated this by sadly commenting on the caretaking role reversal implicit in this arrangement.

Among those living on their own, health and substance use often interrelated with the ability to keep housing and eviction. When eviction occurred, it often exacerbated health problems. Vincent, aged 74 years, has been homeless since 1989. He explained that he first became homeless when his landlady sold the apartment building in which he was living. The loss of housing resulting from eviction appeared to cause significant stress on the part of respondents such as Vincent. Respondents commonly had cardiovascular disease, a condition known to be a major cause of mortality among people who are homeless, which is exacerbated by stressful conditions associated with eviction.³⁰ Other, more commonly reported ongoing problems made more difficult with eviction included cancer, severe swelling of limbs, and severe asthma.

Mental illness, a factor interrelated with all these precipitating events, often sat at the back of respondents’ minds and rarely came up during open-ended interviews. However, the relationship between the course of mental illness and homelessness is captured in the story of Phil, who thought he would make a career out of being in the Air Force but was later discharged for speaking up about the treatment of fellow Blacks in the South. Phil did not directly call himself mentally ill. Rather, he spoke about racially motivated violent incidents, including a Black man being hung by the Ku Klux Klan, that traumatized him and have stayed with him his whole life. Throughout the interview, he had a hard time controlling his emotions about these events. Although Phil had worked for about 15 years as an aircraft mechanic, he had also been admitted to psychiatric hospitals approximately 20 times. At age 65 years, he continued to emphasize work and wanting to work. His ability to work, however, had been interrupted by paranoia and suspicion that made it difficult for him to maintain a stable job and housing.

The same thing happened with Jack, who reported substance abuse and social support issues as the cause of his homelessness. Before the formal interview, Jack spoke about having encountered extreme trauma from being raped at gunpoint when he was young. His resulting PTSD had led to drug use, which originally led him to go to the VA for assistance. At the time, Jack had been living on the streets or on the beach, in his car, or with his daughter or a friend. He spoke about his substance use problems in relation to the legal system and social welfare policy. His experience seems to resonate with those of many homeless people who have high arrest rates as a result of petty violations, which keep them homeless longer.³¹ Jack had lost his Social Security benefits while serving time to clear a warrant because currently Social Security is suspended if an individual is admitted for more than 30 continuous days in jail or prison and can only be reinstated 1 month after the month in which the individual is released.^32,33

DISCUSSION

The veterans’ narratives speak to the range of factors that are important to this group’s onset of homelessness. In our study of 59 homeless elderly veterans, roughly divided between those who were chronically homeless and those who were acutely homeless, we found substantial levels of physical, psychiatric, and social impairment across both groups. Both groups were more similar than different; the only differences between veterans with relatively little homeless history and those who had been chronically homeless were that chronically homeless veterans were less educated and had a smaller social network. Both groups identified financial barriers as the primary challenge to rehousing and listed housing, health, and financial service needs for assistance to rehouse in a city known for its high cost of housing. Detailed questioning identified a wider range of factors precipitating homelessness than were subjectively volunteered. Personal statements by veterans emphasized the personal importance of the loss of social support as a precipitant leading to homelessness.

We were surprised to find little difference between chronic and acute groups given the current emphasis in the literature on the importance of differences between these groups in homeless populations in general and in prioritizing services for those who are chronically homeless. An additionally noteworthy finding was the veterans’ lack of cognitive dysfunction, whose average age was 74 years. Both findings run counter to conceptual speculations regarding the older homeless population and may indicate that older homeless people are generally more equivalent in disability burden regardless of homeless chronicity, yet may or may not necessarily be less cognitively impaired than nonhomeless elderly populations of similar age.

Moreover, our findings suggest that assessment of a range of factors or domains is needed to address risk factors and events leading to homelessness and that assessment should be conducted using open-ended questions to allow veterans to emphasize what they view as the most important causes of their homelessness. The importance of individualizing intervention once the range of pertinent issues is identified is underscored by this veteran’s comments:

It just appears to me that they have a one size fits all mindset when it comes to handling veterans’ problems. One size fits all. And this is absurd. I mean, people come into a place like this for various reasons. Either they’re depressed because their marriage broke up, they’re depressed about their financial situation, they’re depressed about their medical situation … I know this from life’s experiences here in Southern California and the Los Angeles area—unless you’ve got a lot of money, it isn’t possible for the average person to find a really decent, safe place to live.

This veteran’s statement resonates with the housing issues raised among other elderly veterans in our sample, a group with significant health, mental health, legal, benefits-related, and substance use issues, along with other conditions, and whose homelessness is often associated with loss of social support. This portrait calls for supportive housing, in which the strong peer support networks among veterans could play critical roles in constructive feedback as well as companionship and instrumental assistance. Affordable housing programs, which provide services to elderly people, such as HUD’s Section 202 program, make up only 20% of affordable housing available to elderly individuals and house only approximately 300 000 of the 3.8 million very-low-income elderly people.³⁴ If these individuals are anything like our sample, many could be contenders for long-term care; however, because of their economic status, cost-containment strategies on the part of states (which limit functional eligibility and create long-term care waiting lists), they are often only considered for housing-first supportive housing, which often focuses services on much younger populations.³⁵

Long-term housing options are available for veterans who have sufficient income to live in a house or apartment, veterans who have sufficient income to live in a board-and-care home, or veterans who are disabled enough to require nursing or other long-term care. The 2 prominent options that exist for veterans not in these circumstances are the HUD–Veterans Affairs Supportive Housing voucher plus case management program specifically designed for veterans and a newly implemented strategy for permanent housing in place funded by the VA’s Grant and Per Diem program. Both programs emphasize case management services, and both have provisions for congregate living that are veteran specific. How much access elderly homeless veterans, particularly those who are not chronically homeless, have to these programs and with what outcomes should be examined in the evaluation of these programs.

This issue raises the study’s most important limitations. The sample size was small, consisting solely of male veterans receiving transitional residential services, and participants were recruited via convenience sampling methods through a VA transitional housing program. Findings here are thus not generalizable to elderly homeless veterans (or nonveterans) across the homelessness spectrum. Future study of homeless elderly veterans and homeless elderly nonveterans older than age 65 years should use larger and more diverse samples across a number of different sites and include objective measures of patient diagnoses (i.e., from patient medical records) to confirm or challenge these findings across larger elderly populations in the Grant and Per Diem and HUD–Veterans Affairs Supportive Housing programs. Further study of this population is also needed to identify the range of available income supports and access strategies needed to assist elderly homeless veterans to secure stable housing. Accurate assessment and appropriate services will be important in addressing this specific subpopulation of homeless veterans to ensure that the VA Secretary’s goal of eliminating veteran homelessness is realized.