Characteristics of Individuals With Mental Illness in Tokyo Homeless Shelters

The key message of the World Health Organization (WHO) World Health Report2001 (1) was that there is no justification for excluding people with mental illness from communities. In Western countries, there has been much progress toward deinstitutionalization, with mostly favorable outcomes for long-term patients discharged to the community (2). Japan has the largest number of psychiatric beds per population in the world. According to the Organization for Economic Cooperation and Development (3), Japan had 2.7 psychiatric beds per 1,000 people in 2012. Since 2004, Japan has been in the process of reforming its mental health services to move from hospital-based to community-based care (4). Considering that deinstitutionalization in the United States involved the downsizing of public hospitals (5), a key factor in Japanese mental health reform is the control of psychiatric hospitals. According to Japan’s National Institute of Mental Health, more than 80% of the country’s psychiatric hospitals are privately owned (6), and most were built in the 1960s to end home confinement of persons with mental illness. As early in 1976, “control of mental hospitals” was recommended to the Japanese government in a WHO report (7), and this issue has not been fully resolved.

Emptying psychiatric hospitals without ensuring community support for the discharged patients can be hazardous, because people with mental illness may be excluded from the community again through homelessness. Having a mental illness is an established risk factor for homelessness (8). A high rate of mental illness is reported in homeless populations (9,10), and the rate of suicide attempts among homeless persons with mental illness is high (11,12).

In Japan, elderly long-term psychiatric inpatients are especially at risk of homelessness after discharge. The proportion of psychiatric inpatients over age 65 increased from 20% in 1988% to 50% in 2010, and the proportion of all psychiatric patients with a stay of at least one year is reported to be 66% (6). Because of the increase in elderly long-term patients in psychiatric hospitals, reforming mental health services will return many elderly persons with mental illness to the community. In fact, the age of discharged patients is already increasing; the percentage of psychiatric patients over age 65 discharged from psychiatric hospitals increased from 29% in 1999 to 37% in 2010 (6).

In addition, an elderly person who does not have family faces difficulty when leasing housing, especially in Japan. Anyone leasing housing in Japan generally needs one or two guarantors, who bear the same contract obligations as the tenant. Most Japanese people ask their parents or siblings to be their guarantors. If there is no relative to rely on, employers sometimes become guarantors as long as the tenant is employed. Accordingly, an elderly person who does not have family is at great risk of homelessness.

Because of this risk of homelessness, deinstitutionalization in a rapidly aging society must include the provision of community mental health services to facilitate social inclusion of elderly persons. In particular, the society must prepare to address the risk of homelessness among elderly long-term patients after hospital discharge. However, little is known about the relationship between mental illness and homelessness in Japan. The primary aim of this study was to describe the characteristics of service users in Tokyo homeless shelters. Its secondary aim was to explore community living conditions for people with mental illness by comparing two groups: those who live in shelters and those who live in the community.

Methods

Participants

A face-to-face survey was conducted from December 2012 to March 2013 by the staff of a nonprofit organization (NPO) called Hurusato No Kai, which means “Hometown Association.” The organization, which currently has 68 staff members, was founded in 1990 in Tokyo’s biggest slum. The only medical professional on staff is a public health nurse. Provision of support begins when a person who lacks a fixed residence is referred to the organization, and support ends when the service user is able to live independently in the community. However, the relationship often continues because many users want to keep in touch with the NPO for instrumental or emotional support, even after moving into independent living. The NPO mainly provides shelters for homeless people, home care support groups for community residents, and a community café for everyone. The shelters have no exclusion criteria.

For the study reported here, NPO staff members who provide onsite services asked users of NPO’s services to complete a questionnaire administered by the staff member; no compulsion was involved. The survey questionnaire consisted of two parts. The first part was completed by the NPO staff member with information from the organization’s records. The second part was completed by the staff member with information from the interview with the user. A total of 1,056 individuals who received help from the NPO during the study period were potential study participants. Of these, 684 (65%) completed the questionnaire, including 210 persons who were living in the shelters.

The study protocol was approved by the Ethics Committee of the NPO and the Ethics Committee of the Tokyo Metropolitan Institute of Gerontology. Written informed consent was obtained from all participants prior to the investigation.

Measures

Questionnaire part 1.

The first part of the questionnaire, completed by a staff member, addressed the pathway to services, activities of daily living (ADLs), instrumental activities of daily living (IADLs), and health-related and psychosocial variables. The staff member identified where the service user lived before coming to the shelter. The questionnaire included detailed choices and space for additional comments. We divided these answers into seven categories: home, street homelessness (on the street), other homelessness (hotels, other shelters, Internet cafes, and factories), hospitals, incarceration (jails and prisons), welfare institutions (such as those for elderly people), and unknown. The data did not permit us to distinguish psychiatric hospitals from other hospitals.

Activities of daily living were measured with the Barthel Index (13). Participants were divided into three impairment categories: severe (≤50 points), moderate (51–74 points), and mild or no impairment (≥75 points) (14). Staff noted the user’s ability regarding food preparation, housekeeping, shopping, handling of finances, responsibility for his or her own medications, and use of the telephone and modes of transportation. The items were adapted, in part, from the Lawton Instrumental Activities of Daily Living Scale (15).

The staff member noted (yes or no) whether the service user had a mental illness according to whether information from the cover sheet of the support record included this medical information from the referring party. The staff member also noted whether the user had a history of certain medical conditions—in particular, cancer, tuberculosis, chronic infectious diseases (such as HIV and viral hepatitis), allergies, lifestyle diseases (such as hypertension and diabetes mellitus), musculoskeletal problems (such as a slipped disk, rheumatism, gout, osteoporosis, and arthritis), cerebrovascular problems (such as intracranial hemorrhage, cerebral infarction, and subarachnoid hemorrhage), circulatory diseases (such as myocarditis, angina, myocardial infarction, arrhythmia, and heart failure), and renal diseases (such as nephritis, renal failure, urinary tract stone, acute cystitis, and benign prostatic hyperplasia). Users were asked whether they were registered with a doctor.

Family ties of the service user were assessed by the staff member. A user was considered to have poor family ties if the user and the staff had no relationship with the user’s family and if staff were allowed to contact the family only if the user was dying.

Educational status was recorded as low if the user did not graduate from junior high school. This information was obtained from the face sheet of the record. The user’s literacy was noted by the staff member as either “has difficulty with reading or writing” or “has no problem with reading or writing.” The staff also noted the user’s history of trouble with alcohol (yes or no).

Questionnaire part 2.

The second part of the questionnaire, completed with information from the user interview, addressed mental well-being and psychosocial variables. The user completed the Japanese version of the WHO-Five Well-Being Index (16) to assess mental well-being. The standard cutoff criterion was used; persons who scored <13 were regarded as having poor mental well-being. The user’s desire to help others was measured by the question, “Do you have a wish to help others with your skills and experiences if there is a chance to do so?” We adopted this item because an altruistic attitude is related to interdependency in the community, which is an important component of recovery. Users who reported that they did not have anyone to talk to or consult with about little things in everyday life were regarded as being in a state of social isolation. The service user was asked to respond to the statement, “I like the region where I live now,” with four possible responses. Those who answered “not so much” and “no” were regarded as lacking attachment to the region.

Statistical Analysis

The first analysis compared the demographic characteristics of the inhabitants of homeless shelters by dividing them into two groups: persons without mental illness (N=138) and persons with mental illness (N=72). The second analysis compared two groups of people with mental illness and mild or no impairment in regard to ADLs: those living in shelters (N=45) and those living at home in the community (N=48). Because the role of physical impairments in the process of exclusion from the community was beyond the scope of this study, 32 participants with mental illness who had severe to moderate impairments in ADLs (Barthel Index ≤75) were excluded from the second analysis—five in the community and 27 in shelters. A multivariate regression analysis was conducted to examine risk factors for homelessness. Only variables identified as significant in bivariate analyses were included in the multivariate analysis. IADLs were represented by housekeeping ability, which landlords consider very important when leasing housing.

Student’s t test was used for comparison of continuous variables, and the chi square test was used for categorical data. Continuous variables are reported as means and standard deviations. All tests were two-tailed. Statistical analyses were performed with SPSS, version 20. A p value of >.05 was regarded as statistically significant.

Results

Eighteen percent of the 684 survey participants had a mental illness. Table 1 presents descriptive statistics for NPO service users with and without mental illness who were staying at homeless shelters. Of the 210 individuals in our sample who lived in shelters, 72 (34%) were reported to have mental illness by the staff. The mean age of this group with mental illness was 64.9, and 57% were age 65 or over. Of the 138 individuals who were not reported to have mental illness, the mean age was 67.7 and 64% of them were age 65 or over. A number of pathways to the shelters were identified. Among those with mental illness, 36% came from hospitals and 21% homelessness; among those without mental illness, 31% came from homelessness and 24% from hospitals.

With or without mental illness, the service users who lived in shelters had poor social relationships. More than half of both groups had poor family ties, a lack of desire to help others, and a lack of attachment to the region. Compared with those without mental illness, those with mental illness were more likely to have a history of trouble with alcohol and to report poor mental well-being. None of the service users with mental illness were reported to have a general medical illness. No significant difference was found between the two groups in terms of age, education, ADL, or IADL.

Table 2 summarizes characteristics of service users with mental illness living in shelters and living at home in the community. Both groups reported poor mental well-being. Those who were living in the community—and therefore, more socially included—did not have a higher level of well-being, as was expected. Again, more than half of both groups reported a lack of desire to help others and a lack of attachment to the community. Multivariate logistic regression analysis indicated that being older and having a history of trouble with alcohol, poor family ties, and impaired IADLs were independent risk factors for homelessness (Table 3).

Discussion

Of the 210 individuals in our sample who lived in homeless shelters in Tokyo, 34% were reported to have a mental illness. These individuals tended to be referred from hospitals and to have a history of trouble with alcohol. They also reported poor mental well-being. They seemed to be facing a very difficult situation in the shelters, in that the shelters were not meeting their mental health needs. Our results are consistent with findings from other countries reporting a high rate of mental illness among homeless populations (9,10). They are also consistent with a previous report that patients are at high risk of homelessness in the period immediately after discharge from a psychiatric hospital (17).

The homeless persons in our study were much older than homeless persons in Western countries. The mean age of shelter users in Tokyo was 64.9 for those with mental illness, which is higher than the mean age of 40.1 for shelter users in Western countries, according to a systematic review (10). In addition, being age 65 or older was significantly associated with homelessness (odds ratio [OR]=3.50). There are three possible explanations for this finding. First, it reflects the rapid aging of the Japanese society. The percentage of people over age 65 in the Japanese population was 25% in 2013, the highest percentage in the world, and the percentage is estimated to reach 40% in 2060 (18). The number of elderly people living alone is also increasing. In Japan, the proportion of elderly persons (age 65 or over) living alone rose from 4% for males and 11% for females in 1980 to 11% and 20%, respectively, in 2010 (19). The proportion of people with impairment in IADLs inevitably increases with age. Therefore, more elderly people with mental illness are at risk of homelessness as the society ages.

Second, it is not uncommon for Japanese young people to remain in the family home, which provides socioeconomic protection from homelessness. However, this has led to an unfavorable result: there is a group of young people who have withdrawn from social life and have had no relationships outside the family for more than six months, which is referred to by the Japanese term “hikikomori” (20). A nationwide survey conducted as a part of the WHO World Mental Health Initiatives estimated that there are 232,000 ongoing hikikomori cases in Japan (21). Thus, many young persons who leave their homes are unable to support themselves or to seek help appropriately, and it is natural for them to return home rather than remain homeless.

Third, the finding may result from sampling bias. The NPO serves persons in the largest slum in Tokyo, where many low-skilled day laborers rushed from all over the country during the high growth of Japanese economy (1950s–1970s). These aging individuals remain in the slum, forgotten by or excluded from the postindustrial society. Men may have been overrepresented in our study for of the same reason.

Poor family ties were significantly associated with homelessness (OR=3.88). Our results suggest that families play an important role in the social inclusion of people who are mentally ill. Although family ties are difficult to measure, in Japanese society they seem to be weakening. For example, the number of “solitary deaths” (death at home by natural causes of someone who lives alone, without anyone knowing) is reported to be increasing in government-subsidized apartments, from 207 cases in 1999 to 613 cases in 2008 (22).

There is little information about the dynamics of the lives of homeless people and the relationship between mental illness and homelessness in Japanese society. According to Culhane and Metraux (23), there is a group of homeless persons, described as “episodic homeless,” who move repeatedly between the streets, hospitals, jails, and shelters, and members of this group tend to have a mental illness. The literature emphasizes the need for specialized services that focus on continuity of care during a critical time of transition (24–26). Our results suggest that implementation of critical time intervention could be effective in a society in the early stages of mental health care reform.

Contrary to our expectations, service users who were living in the community did not report a higher level of mental well-being. This finding suggests that emptying psychiatric beds and offering former patients a place to live is not enough for social integration. To move toward deinstitutionalization, mental health care must be more integrated into a broad-based social services package delivered in the community, not in the hospital. Further research with a large-scale, representative sample of persons in homeless shelters is urgently needed to create effective policies to realize social inclusion and end homelessness.

This study had several limitations. The determination of mental illness was based on records, rather than on a structured interview. In addition, the staff had no special training in conducting the survey, and their involvement was voluntary. Therefore, test-retest reliability was not assessed. The survey was limited to users of one NPO in Tokyo who were willing to answer the questionnaire. Thus the study is limited in terms of its representativeness. In addition, although our interest was in the deinstitutionalization of persons with mental illness, the data did not allow us to distinguish between those referred from psychiatric hospitals or from general hospitals. Neurological illness, such as traumatic brain injury, was not assessed and not reported separately from mental illness. Substance misuse, except for alcohol, was not assessed. Finally, a cross-sectional design was used, which has inherent limitations for determining causal relationships.

Conclusions

One-third of individuals living in Tokyo shelters were reported to have a mental illness. Mental health care reform from hospital-based to community-based services is urgently needed. However, evidence-based policy is essential. Our results show the need for a support system to integrate patients discharged from hospitals and prevent them from being excluded again through homelessness. This support system is especially necessary for elderly, long-term patients in Japan’s rapidly aging society.