How many long-stay schizophrenia patients can be discharged in Japan?


Yoshio Mino, MD, Mental Health Section, School of Humanities and Social Science, Osaka Prefecture University, 1-1 Gakuen-cho, Sakai, Osaka, 599-8531, Japan. Email:


Abstract  The mental health-care system in Japan remains hospital-based, and has the largest number of psychiatric beds per capita in the world. However, serious discussion about deinstitutionalization has recently begun. This study attempts to determine the proportion of inpatients that would benefit from community-based programs, as judged by hospital psychiatrists, and to evaluate the need for community resources for their community placement. Inpatients with schizophrenia from 139 hospitals were randomly selected. Data on the psychiatrists’ judgment of discharge and required resources for community placement were obtained for 2758 subjects. Among the subjects, 1097 (39.8%) were judged to have the possibility of being discharged using community resources (possible discharge group; PDG). Provided that the proportion of PDG was 40%, controlling for the hospital background variables, the number of schizophrenia inpatients with a hospital stay of ≥1 year who could be discharged from psychiatric hospitals in Japan was estimated to be 66 000. For the PDG, the required community resources, including accommodation, daytime activity, and daily living support services, were calculated. The numbers of governmental targets for community resources, including community accommodation, daytime activities, and daily living support services may have been underestimated.


Deinstitutionalization is an ongoing process throughout the world.1,2 Many countries have been searching for more appropriate social policies for people with severe mental illnesses.1–6 In contrast, few workable attempts at deinstitutionalization have as yet been made in Japan. The number of psychiatric beds had been on the rise until sometime in the last few years.7,8 The number of psychiatric beds per 100 000 people in 1998 was 287, which was the highest rate in the world.8,9 In addition, the circumstances with regard to psychiatric hospitals in Japan are known to be deprived.8,10 Nevertheless, some serious discussion about possible plans to establish community care programs as a substitute for psychiatric inpatient care has recently begun.11

It is important to explore the current situation regarding psychiatric inpatients in Japan who might benefit from community-based care, as it is developed. The present study attempts to estimate the proportion of inpatients capable of being discharged if appropriate community-based programs were available, as judged by hospital psychiatrists using a large-scale nationwide cross-sectional survey, and to estimate the required community resources, including accommodation, daytime activity facilities, and daily living support services.

Community mental health and inpatient care in Japan

The characteristics of psychiatric inpatients and psychiatric hospitals are as follows.12,13 Among psychiatric inpatients, 71% remain in the hospital for ≥1 year, and the average length of stay is 8.5 years. In addition, there has been a gradual aging of inpatients, with 29.5% of patients being over the age of 65, and 64.6% over the age of 50. The proportion of involuntary admission cases is no less than 32%, whereas that of unlocked wards is only 40%. The inpatient wards are very restrictive environments. Specifically, in a nationwide survey, we found that the quality of life within psychiatric hospitals is much poorer than that of sheltered accommodations under the Livelihood Protection Law, which was enacted to ensure a minimum level of quality of life in Japan.10

In contrast, progress has been made since the mid-1980s in community mental health policy. Social rehabilitation facilities, in line with the new mental health law, amended in 1987, have been established in several districts across Japan. Recently, various types of community mental health programs have been put in place, although the provisions allocated for these programs are as yet insufficient.


Subjects and hospitals

Psychiatric hospitals from across Japan, whose administrators understood the aim and importance of this research, took part in the survey. In order to introduce the study, a prospectus and a sample of the research kits were sent to 534 hospitals (32.1% of psychiatric hospitals in Japan). The prospectus letter explained the study purpose and methods, and asked the recipients to participate. The subject hospitals included public hospitals that were members of the Association of Local Authority Hospitals, private hospitals in a certain mutual aid organization, and hospitals that had family self-support organizations within them. Favorable replies were received from 139 hospitals (26.0%). The total number of psychiatric beds within these hospitals was 42 063, which accounted for 12.0% of the psychiatric beds in Japan.

To recruit subjects into the study, each hospital made a list of all patients meeting the following criteria: diagnosis with schizophrenia according to DSM-III-R,14 and hospital inpatient stay of ≥1 year on the day of data collection. Each hospital then sent this list to the investigator, who randomly selected 20 subjects from each list based on patient identification numbers. We selected 30 random samples from 20 hospitals. For six hospitals in which the total number of cases did not reach 20, all cases that satisfied the criteria were selected. All subjects provided written informed consent.

A total of 2898 patients fulfilled the inclusion criteria, and satisfactory replies were obtained from 2758 patients (95.2%). As illustrated in Table 1, 57.8% of subjects were male, and the average age and mean length of stay of all subjects were 51.9 years and 13.5 years, respectively. These figures in the sample are consistent with national estimates in Japan.12 In terms of the features of participating hospitals, the proportion of public hospitals (39.6%), the average number of psychiatric beds (303 beds), and the proportion of unlocked wards (44.1%) included in the sample were all greater than the national averages. In contrast, the characteristics of the inpatients and the proportion of medical staff within the participating hospitals were comparable to those reported based on national data. There were no significant differences between the hospitals participating the study and the hospitals that refused to take part.

Table 1.  Subject and hospital characteristics
 Present study n (%)National statistics
  • † 

    1996 Patients Survey.

  • ‡ 

    1999 Department of Mental Health and Welfare, Ministry of Health and Welfare.

Subjects (n = 2758)
 Male1 594 (57.8)56.9%
 Age (mean years)51.9 ± 11.654.0
 >60 years753 (27.3)33.4%
 Length of current stay (mean years)13.5 ± 10.713.5
 Length of illness (mean years)25.3 ± 10.9 
 No. previous admissions (1 or 2)1 070 (38.8) 
Hospitals (n = 139)
 Public hospitals55 (39.6)18.6%
 Hospitals, psychiatry only90 (64.7)71.1%
 No. psychiatric beds (mean)302.6 ± 194.8209.7
 Unlocked wards (%)44.1 (16 863/38 250; n = 119 hospitals)33.7%
 Stay for <1 year (%)28.8 (11 323/39 279; n = 137 hospitals)27.9%
 Stay for ≥5 years (%)46.8 (18 514/39 598; n = 137 hospitals)45.4%
 Involuntary admission (%)37.8 (14 516/38 419; n = 132 hospitals)28.6%
 Age ≥65 years (%)34.1 (13 526/39 650; n = 138 hospitals)29.6%
 No. medical doctors per 100 patients (mean)2.6 ± 1.22.8
 No. registered nurses per 100 patients (mean)31.7 ± 9.029.8

Procedures and measures

Research kits, including various rating scales, questionnaires and research manuals, were sent to participating hospitals. Hospital psychiatrists assessed the clinical conditions of the study participants and provided judgments about the possibility of each study participant. The survey was conducted from January to February, 1995.

Judgment of the possibility of being discharged

The criterion for the judgment of the possibility of being discharged was that the hospital psychiatrist felt that they would be able to discharge the patient within 6 months if three different community resources were made available to the patient. Specifically, resource areas included: (i) housing programs (e.g. supported housing, group home, and welfare accommodation); (ii) daytime activity programs (e.g. supported employment, sheltered workshop, and medical daycare services); and (iii) daily living support services (e.g. home help services, meal delivery services, and short-stay services). A list of six to seven specific services were identified within each resource area (Appendix I), and when a psychiatrist made a judgment of discharge, s/he was asked to choose the most appropriate community resources in each of the three areas. This procedure was more sophisticated than that in previous studies.


We calculated the proportions of ‘possible discharge group (PDG)’ inpatients according to the sociodemographic variables and clinical features of the subjects and subject hospitals’ attributions. We mainly used data from the standard survey in the analysis of hospital attributions and set the significance level at 1%, because of the large number of subjects. Numbers of PDG among all schizophrenia inpatients and among all inpatients in mental hospitals all over the country were estimated. Numbers of community accommodation facilities, daytime activities, and daily life support services, which the PDG required for community life, were also calculated.

All statistical analyses were carried out using SPSS for Windows (version 9; SPSS Japan, Tokyo, Japan).


Judgment of the possibility of being discharged

Psychiatrists determined that a total of 1097 study participants (39.8% of all study participants) would be able to be discharged given appropriate community resources (PDG). In contrast, the psychiatrists judged 1631 (59.1%) of the study participants as requiring continuing treatment in an inpatient setting (necessary to be treated in hospital group; NTG).

We hypothesized that the proportion of the PDG would vary according to the backgrounds of subjects and hospitals. Table 2 illustrates how the proportion of the PDG decreased as an individual’s length of illness and the length of the current hospital stay increased. With regard to the background variables of hospitals shown in Table 1, only two variables were significantly related to the proportion of the PDG (P < 0.01), including the number of psychiatric beds and the type of hospital.

Table 2.  Proportion of patients judged to be discharged by subject and hospital backgrounds
 nProportion judged discharged (%)Test
  • † 

    χ2 test.

Subject background
 Male159440.7P = 0.316
 Female114938.6χ2 = 1.007 d.f. = 1
Age (years)
 <40 35840.10.794
 40–59164239.8χ2 = 0.462
 ≥60 75341.2d.f. = 2
Duration of illness (years)
 <20 84144.8P = 0.001
 ≥20184738.1χ2 = 10.712 d.f. = 1
Length of current stay (years)
 <10133942.8P = 0.005
 10–20 69140.2χ2 = 10.556
 ≥20 72835.4d.f. = 2
Hospital background
Years since establishment
 <40135741.7P = 0.035
 ≥40138338.0χ2 = 4.45 d.f. = 1
No. psychiatric beds
 <300150642.5= 0.005
 ≥300123236.9χ2 = 7.903 d.f. = 1
Organization that manages the hospital
 Private169738.0= 0.015
 Public106142.6χ2 = 5.864 d.f. = 1
Psychiatric or general hospital
 Psychiatric183737.8= 0.002
 General 90444.0χ2 = 9.746 d.f. = 1
Totaln = 275839.8 

Estimated proportion of dischargeable patients and estimated numbers

In order to estimate the number of patients nationwide who would be able to receive community-based mental health services, it is necessary to correct for the proportion of the PDG according to three hospital background variables: the size of the hospital; whether the hospital is private or public; and whether it is a general or psychiatric hospital.

Consequently, we multiplied the proportion of the PDG in each attribution class of the three variables shown in the lower part of Table 2 by the number in each class, and obtained the estimated number of patients who could be discharged in each class. Then we calculated the total estimated numbers in the three variables, divided them by the total number (2758), and obtained the corrected proportions according to the three variables. The obtained values were 40.4% for the size of hospitals, 39.5% for general hospitals versus psychiatry only, and 38.8% for private versus public hospitals.

On the assumption that the proportion of the PDG was 40%, we can estimate the number of inpatients with schizophrenia staying for ≥1 year who could be discharged to be 66 000 people. When applying the same proportion to all inpatients staying for ≥1 year in Japan, the estimated number who could be discharged was found to be 94 000 people.

Estimated numbers of required community resources

Table 3 shows the required accommodation services. Among 1097 PDG patients, 32.6% were assumed to require community hostels, and 19.1%, large-scale welfare accommodation. Psychiatrists judged that 26.3% were able to live with their family members. The estimated number of various accommodation services required for all nationwide schizophrenia patients staying for ≥1 year was 21 539 places for community hostels, and 12 574 for large-scale welfare accommodation. Applying the same proportion to all inpatients staying for ≥1 year in Japan, they required 30 676 and 17 909 places, respectively.

Table 3.  Estimates of required accommodation services
 n%Required numbers for schizophrenia (places)Required numbers for all inpatients (places)
Supported housing766.94 5726 512
Group homes with staff on-site15013.79 02512 853
Community hostels35832.621 53930 676
Large-scale welfare accommodations20919.112 57417 909
Other housing programs60.5361514
Living with family members28826.317 32724 678

Results for required daytime activity services are shown in Table 4. Among the PDG patients, 22.0% required social club activities and 20.7%, daily day-care programs. The estimated number of the daytime activity programs was 14 500 places in social clubs and 13 657 in daily day-care programs for all schizophrenia inpatients, and 20 651 and 19 451 places for all inpatients, respectively.

Table 4.  Estimates of required daytime activity programs
 n%Required numbers for schizophrenia (places)Required numbers for all inpatients (places)
Supported or regular employment11310.36 7999 683
Sheltered workshops19417.711 67216 624
Daily day care programs22720.713 65719 451
Day care programs available 1–2 days/week827.54 9337 026
Social club activities24122.014 50020 651
Other day activities181.61 0831 542
Roles with housework or family business11710.77 03910 026
Not necessary595.43 5505 056
Others322.92 7723 948
Unknown141.38581 092

Table 5 shows the required daily living support services for PDG patients. Among the PDG patients, 58.2% required home visit services, 46.3% social club services, and 43.9% meal delivery services. The estimated number of places for daily living support services was 38 445 for home visit, 30 624 for social club services, and 28 999 meal services for all schizophrenia inpatients staying for ≥1 year, and 54 755, 43 615, and 41 302 for all inpatients, respectively.

Table 5.  Estimates of required daily living support services
 n%Required numbers for schizophrenia (places)Required numbers for all inpatients (places)
Required very much
 Home visits to support daily living63958.238 44554 755
 24-h counseling services30027.318 04925 706
 Providing places and support for social clubs50946.430 62443 615
 Evening care programs17916.310 76915 338
 Meal delivery services48243.928 99941 302
 Short stay services32229.419 37327 592
 Bathing and laundry facility services31228.418 77126 735
Required to some extent
 Home visits to support daily living31728.919 07227 163
 24-h counseling services49945.530 02242 758
 Providing places and support for social clubs40136.624 12634 361
 Evening care programs40637.024 42734 789
 Meal delivery services22120.113 29618 937
 Short stay services37434.122 50132 047
 Bathing and laundry facility services31028.318 65126 563


In the current study, we demonstrated that the proportion of long-stay patients who were judged to be capable of being discharged was approximately 40%, and we ascertained the validity of the judgment in a nationwide survey. Similar nationwide surveys using a uniform method were conducted four times, including the present study.15–17 In these surveys, the proportions of dischargeable inpatients ranged from 32.5% to 44.9%. The current study, however, was the most systematic in that it utilized standardized psychiatric symptom scales and social functioning scales, and demonstrated judgment validity.

Therefore, in beginning to plan a community-based mental health system, it is prudent to consider the scope of current inpatients who would benefit from a less restrictive treatment setting.

The proportion of the PDG in the current study was smaller than that in British studies using similar research methods conducted in the 1970s, in which 56% of new long-stay patients with schizophrenia were found to be dischargeable,18 and 71% of long-stay patients were under retirement age for all types of diagnoses.19 However, it is expected that the proportion of the PDG would vary according to the judgment criteria. In the present study, the judgment criteria were to evaluate the possibility of discharge given that appropriate community resources were available. Consequently, we believe that the present proportion does not include more severe inpatients who might be considered dischargeable if more intensive care programs were available (e.g. assertive community treatment; ACT)20 or long-stay hostels or sheltered accommodation.18,19

Based on these findings, we propose the development of community-based mental health-care programs that would enable approximately 40% of the current schizophrenia inpatients with ≥1 year hospital stay to be treated in the community. To prevent against institutionalism,21 this PDG should be returned to the community as soon as possible. In the current study, the estimated numbers of community resources, including accommodation, daytime activities, and daily living support services for PDG inpatients, were calculated. The estimated numbers were larger than those proposed in the Plan for the Disabled by the government.22 In the Plan, the government intended to increase the number of community accommodation places by 2007, indicating the following target numbers: 6740 places in community hostels, 4010 places in welfare homes, and 11 968 places in group homes. As for the daytime activities, the target numbers were 7160 places in sheltered workshops. For the daily living support, the target number was 3305 for home-help services. Considering the results of the current study, the target numbers in the Plan for the Disabled might have been underestimated.

A number of limitations to the present study need to be considered. This study obtained its subjects from a convenient sample of hospitals, although the majority of hospital and patient characteristics are comparable to national estimates in Japan. In addition, because the same psychiatrist determined the judgment of discharge and clinical assessment, a potential bias is present. We held a training seminar on the rating scales for psychiatrists involved with patients from the detailed survey, but we did not do so for the additional psychiatrists assessing patients from the general survey. The current study was conducted approximately 10 years ago, but the critical status of psychiatric hospitals in Japan has remained unchanged over the past decade. Last, in the estimates regarding all inpatients, we used the same figures as in the schizophrenia inpatients. Despite these limitations, it is reasonable to expect that the results of the present study will continue to inform public policy makers regarding the reform of psychiatric hospitals in Japan, and we believe that the estimated proportion of psychiatric long-stay inpatients capable of being discharged using the validated method will encourage movement away from hospital care and promote the development of community mental health programs in Japan.


The nationwide survey was conducted by the research committee (Y. Inomata, Chairperson; Y. Mino, S. Inoue, A. Yoshizumi, Y. Kawazoe, K. Okagami, I. Oshima, K. Inazawa, Y. Igarashi, Y. Ueki, K. Kon, Y. Saito, Y. Sawa, T. Takayama, T. Takizawa, I. Terada, T. Yamada) funded and organized by the National Federation of Families with the Mentally Ill in Japan (Zenkaren). The current study was partly supported by the Bureau for the Elderly, Ministry of Health and Welfare in Japan.



Judgment of possibility of being discharged by hospital psychiatrists

Q. Select one in terms of the possibility of being discharged for this inpatient in the near future (within 6 months).

  • 1He/she needs to remain in the hospital for additional medical treatment
  • 2He/she will be able to be discharged within 6 months if the following three community resources are made available to him/her (select from each of the three lists (a–c) below)

(a) Housing programs (select one):

  • 1Supported housing that includes the availability of home visiting services by psychiatric hospital or public health center staff, if necessary.
  • 2Group homes with staff on-site for a portion of each day where meals are provided.
  • 3Community hostels where 10–20 consumers live together and which includes daily care by professional staff, meal services, and involvement of psychiatrists.
  • 4Large-scale welfare accommodation for elderly people or others with the involvement of psychiatrists.
  • 5Other housing programs (please provide additional details: )
  • 6Living with family members.

(b) Daytime activity programs (select one):

  • 1Supported employment or regular employment
  • 2Sheltered workshops
  • 3Daily day care programs available (e.g. medical day care programs)
  • 4Day care programs available once or twice a week (e.g. day care programs at public health centers)
  • 5Social club activities
  • 6Other day activities
  • 7Roles with housework or family business
  • 8Not necessary

(c)  Daily living support services (select as many as necessary):

  • 1Home visits to support daily living (e.g. home help services)
  • 224-h counseling services available for support of daily living
  • 3Providing places and support for social clubs
  • 4Evening care programs
  • 5Meal delivery services
  • 6Short stay services
  • 7Bathing and laundry facility services