SEARCH

SEARCH BY CITATION

Keywords:

  • case–control studies;
  • mental health services;
  • patient readmission;
  • schizophrenia

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. REFERENCES

Abstract  The purpose of the present study was to identify the associated factors of rehospitalization in schizophrenic patients. A case-control study was conducted. The cases consisted of rehospitalized patients (n = 67) and controls selected from the outpatients who were matched by age, gender, and the period after the last discharge (n = 62). In the multiple logistic regression analysis, no clinic visits in the second month prior to entry, the number of clinic visits in the previous month, and junior high school graduation as education level were significantly (P < 0.01) associated with rehospitalization after controlling their present function as assessed by the Global Assessment of Functioning. Close monitoring of clinic visits and outreach service appear to be important in preventing rehospitalization of schizophrenic patients. These identified modifiable factors suggest further needs for development and implementation of integrated mental health services in the community.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. REFERENCES

The care of schizophrenic patients has been hospital-oriented in Japan, leading to the serious problem of the accumulation of long-stay inpatients in mental hospitals.1 In 1995 the Law Concerning Mental Health and Welfare for the Mentally Disabled, and the Government Action Plan for Persons with Disabilities were drawn up in Japan to promote community care of schizophrenic patients. It has been shown that community care is economically comparable2 and as cost-effective as hospital-based treatment.3 Moreover, it has been suggested that the patient's subjective quality of life is improved by community care,4,5 although there has also been a negative result on this issue.6 However, as pointed out by Weiden and Glazer, the revolving-door phenomenon is often observed when reducing beds in mental hospitals.7 The discontinuation of social life would cause deterioration of patient satisfaction and quality of life.8 Therefore, clarification of the risk factors of rehospitalization is a prerequisite for promoting community care of schizophrenic patients. However, epidemiological studies for schizophrenia especially on rehospitalization and community care are rare in Japan, where the registration system of patients with mental illness has not been established in the community.9,10 The present study was conducted to identify the risk factors for rehospitalization in schizophrenic patients in Japan.

METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. REFERENCES

In the present study the schizophrenic episode was examined using the stress–vulnerability model.11 We regarded rehospitalization not as a simple product of the exacerbation of clinical status, but as the complex outcome of illness, familial support, and mental health and welfare service utilization. Therefore, we investigated the risk factors of rehospitalization in the following domains: patient's personal factors, familial factors, factors related to psychiatric treatments, mental health and welfare service utilization.

Study area and hospitals

Patients were recruited from three private psychiatric hospitals in and around Yamagata City, where one public health center covers mental health services in the community. The number of beds in each hospital was 337, 460, and 265, respectively. The three hospitals offer similar types of outpatient clinics, inpatient wards, and rehabilitation resources such as day-care units and occupational therapy units. The average number of schizophrenic outpatients in each hospital was 554.8, 315.8, and 99.8 per month in 1998.

Patient selection

We designed a case–control study. Case subjects consisted of readmitting patients, while control subjects were chosen from the outpatient population. Inclusion criteria for case subjects were (i) rehospitalization and a chart diagnosis of schizophrenia without mental retardation; (ii) age between 20 and 69 years; and (iii) having a history of previous admission to the same hospital. During the recruitment period (July–October 1998), the number of eligible cases was 75. We excluded three patients who did not meet the diagnostic criteria of schizophrenia in Diagnostic and Statistical Manual of Mental Disorders (4th edn; DSM-IV),12 and five readmitted patients who had returned from general hospitals where they had received more intensive medical treatment for physical conditions during their long-term stay in the mental hospitals. After we selected 67 subjects, control subjects were chosen from the outpatient lists in the hospitals with matching age (± 2 years), gender, and period after the last discharge (± 90 days). Five controls were not found because there were no eligible subjects who met the matching factors. When any of the control subjects was rehospitalized during the recruitment period, they were replaced by one of the other outpatients. The final number of control subjects was 62.

Approximately a half of the subjects were female (53.7%, rehospitalized group; 51.6%, outpatient group), and most of the subjects were middle-aged. The mean age ± SD was 46.0 ± 11.4 years for the rehospitalization group and 46.0 ± 11.3 years for the outpatient group. Approximately 60% lived in communities less than 1 year from the last discharges (61.2%, rehospitalized group; 62.9%, outpatient group). The outpatient group had a higher proportion of people who spent more than 3 years in communities after the last rehospitalization (14.9%, rehospitalization group; 21.0%, outpatient group). However, there was no statistically significant difference between the two groups.

Assessment

We investigated 58 variables in the following six domains.

  • 1
    The patient's background and history including marital status, age at onset of schizophrenia, the legal form of the previous hospitalization, and education level were collected from the medical records.
  • 2
    The patient's personal factors, for example, employment, medication compliance, and experience of life events were examined. Medication compliance were evaluated by the primary psychiatrists of the subjects using the compliance scale developed by Kemp et al.13 Each of the levels were referred to as follows: (i) complete denial; (ii) partial refusal, for example, refusing depot drugs or accepting only the minimum dose; (iii) reluctant acceptance: accepting only because treatment is compulsory or questioning the need for treatment often (every 2 days); (iv) occasional reluctance about treatment: questioning the need for treatment once a week; (v) passive acceptance; (vi) moderate participation: some knowledge of, and interest in, treatment and no prompting needed to take drugs; and (vii) active participation, ready acceptance, and taking some responsibility for treatment. The subjects were asked if they experienced life events in the past 3 months according to the categories of axis IV in DSM-IV.12 The problems were grouped into the following categories: problems with primary support group; problems related to the social environment; educational problems; occupational problems; housing problems; economic problems; problems with access to health-care services; and problems related to interaction with the legal system or crime.
  • 3
    Information about cohabitants and about their mood of relationships (e.g. critical or supportive comments) were obtained by the social workers in each hospital.
  • 4
    The information on psychiatric treatments was obtained from the medical charts. It contained the number of clinic visits, the types and routes of medication therapy, neuroleptic dosage converted in chlorpromazine equivalents,14 use of anxiolytic drugs and mood stabilizers, and other comorbid illness.
  • 5
    Information on the welfare service utilization was gathered by the social workers in each hospital.
  • 6
    The information on psychiatric treatments and health services utilization was obtained by one of the authors (YS) with face-to-face interviews on service contacts in the past 1 year. Readmitted patients were interviewed in the admission wards after their symptoms became steady enough for the acceptance of the interview. The outpatients were interviewed on their regular clinic visits.

The primary psychiatrists assessed the level of psychiatric symptoms according to the Clinical Global Impressions (CGI) scale,15 and the present function and the best function in the past 1 year by the Global Assessment of Functioning (GAF) scale.12 In CGI a greater score indicates a worse clinical state of the patient (range: 0–7). The GAF has a range of 0–100, and a full point means superior functioning and no symptom. The time point for the assessment of the present function was at the admission for the rehospitalized patients and at the latest clinic visit for the outpatients. Maximal attention to privacy was paid during data collection, and written informed consent was obtained before interviewing the subjects.

Statistical analyses

Sixty-seven cases and 62 controls underwent statistical analysis. On interviewing subjects, there were four drop-outs: two subjects (one case and one control) moved to other hospitals from reason not related to the present study; and two subjects (one case and one control) were lost to follow up. Therefore, 65 cases and 60 controls were analyzed for the interview. Means and standard deviations were calculated for the following variables: age at onset of schizophrenia; psychotic symptom; and GAF score. For the bivariate analyses, χ2 tests were used for categorical variables and t-tests were used for continuous variables. The variables showing significant association in the bivariate analyses were entered into the unconditional multiple logistic regression analysis to obtain the adjusted odds ratios (OR) and 95% confidence intervals (95%CI). We put the present function as assessed by GAF into the model as a controlling variable because the  score  of  GAF  was  considered  as  a  confounder. The backward stepwise procedure of multiple logistic regression analysis was employed for the purpose of including a wide number of potential variables in the relatively small number of samples. The criterion for removal in the multiple logistic regression analysis was P > 0.1. All analyses were performed using spss 6.1 (SPSS, Chicago, IL, USA) of the Macintosh statistical software.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. REFERENCES

Bivariate analyses

The indicators of clinical status in each group are shown in Table 1. Twenty-two patients in the case group were diagnosed as having the paranoid type of schizophrenia, while there were 11 such patients in the control group. The mean of the CGI scores was higher (4.4 vs 4.0), and the mean of the present GAF score was lower (40.1 vs 54.9) in the rehospitalized group than in the outpatient group at the recruitment period. However, the mean of the best GAF scores in the past 1 year showed no significant difference between the two groups (56.0 vs 59.4).

Table 1.  Comparison of the clinical status in the rehospitalized and outpatient groups
Variables Rehospitalized patientsOutpatientsP
n = 67%n = 62%
  1. DSM-IV, Diagnostic and Statistical Manual of Mental Disorders, 4th edn; CGI, Clinical Global Impressions Scale; GAF, Global Assessment of Functioning Scale; NS, not significant.

  2. χ2 tests or t-tests were used.

Type (DSM-IV)Paranoid2232.81117.7NS
Disorganized 6 9.01117.7 
Catatonic 6 9.0 4 6.5 
Undifferentiated1217.9 812.9 
Residual2131.32845.2 
Symptom (CGI)Mean (SD) 4.40 (1.0) 4.00 (1.2)< 0.05
Present function (GAF)Mean (SD)40.1(16.1)54.9(17.7)< 0.01
Best function in the past 1 year (GAF)Mean (SD)56.0(17.6)59.4(17.5)NS

The results of the bivariate analyses of potential associated factors are shown in Table 2. Junior high school graduation as educational level; total number of admissions of more than 11 times; number of clinic visits in the previous month; no clinic visit in the second month prior to the entry; and no utilization of psychiatric day-care units were positively associated with rehospitalization. The dose of neuroleptics and poor medication compliance (score <2 on the scale, i.e. complete refusal or partial refusal) were associated with rehospitalization. Recent experience of life events showed a significant difference between the two groups. A familial supportive comment toward the patient was negatively associated with rehospitalization (statistically at a borderline level; P = 0.052). Those who participated in the day-care unit at the public health center (P = 0.063) and attended workshops (P = 0.073) tended to be rehospitalized. The other psychiatric treatments and mental health service utilization showed no association statistically.

Table 2.  Results of potential risk factors of rehospitalization of schizophrenia by bivariate analyses
Variable Rehospitalized patientsOutpationtsP
n = 67%n = 62%
  • The χ2 tests of Fisher's exact test were used in the nominally scaled variables, and the Mann–Whitney test was used in the ordinally scaled variables.

  • Chlorpromazine equivalents.

Education levelJunior high school2029.9 914.5 inline image<0.05
High school4059.74674.2 inline image 
College and others 710.4 711.3 
Total no. admissions1–104871.65588.7<0.05
11+1928.4 711.3 
Clinic visits in the previous monthNone 5 7.5 0 0.0<0.05
1–21826.93862.3 
3–42841.82134.4 
5+1623.9 23.3 
Clinic visits in the second month prior to the entryNone1322.8 3 5.1 inline image<0.01
1–22645.63966.1 inline image 
3–41119.31627.2 
5+ 712.4 1 1.7 
Neuroleptic dose (mg)1–5001931.13253.3<0.05
501–15003455.72135.0 
1501–2500 813.1 711.7 
Medication compliance (score)1–21319.4 3 4.8<0.05
3–75480.65995.2 
Use of
 Paychiatric day-careYes 710.41725.4<0.05
 Day-care at the public health centerYes1320.0 5 8.3<0.1
 WorkshopYes1726.2 813.3<0.1
Supportive commentsYes2131.32961.7<0.1
Life eventsYes4163.12338.3<0.01

Multivariate analysis

Backward stepwise unconditional multiple logistic regression analysis was performed using the variables that showed significant associations in the bivariate analyses. Because the psychiatric symptom and functional level were highly correlated, we employed the score of GAF as a confounder to represent the clinical status. The following variables were tested in the model: present GAF score; education level (junior high school/higher); experience of life events (yes/no); number of clinic visits in the previous month; clinic visits in the second month prior to the entry (yes/no); total number of admissions; use of psychiatric day-care units (yes/no); compliance score (2/3 or more); and dose of neuroleptics. The adjusted OR (95%CI) were 29.2 (4.5–188.4) in no clinic visit in the second month prior to the entry, 3.6 (1.2–14.7) in the number of clinic visits in the previous month, and 4.8 (1.6–14.7) for junior high school graduation as education level (P < 0.01; Table 3).

Table 3.  Adjusted OR and 95% CI of association of rehospitalization in the multiple logistic regression analysis
VariablesOR95% CIP
  • OR, odds ratio; 95% CI, 95% confidence interval; GAF, Global Assessment of Functioning Scale.

  • The present function (GAF score) was entered as a controlling variable.

No clinic visit in the second month prior to entry29.24.5–188.4<0.01
No. clinic visits in the previous month 3.61.2–14.7<0.01
Junior high school graduation 4.81.6–14.7<0.01

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. REFERENCES

It has been shown that approximately 30–40% of patients with mental illness including schizophrenia have several clinically severe episodes during their long course of illness in Europe.16,17 Meanwhile, the relapse rate of schizophrenia has been reported to be 50–60% within 5–6 years in Japan.18,19 The majority of these patients are rehospitalized, indicating that rehospitalization is very common for schizophrenic patients in Japan. However, there have been very few studies on the risk factors of rehospitalization and the efficacy of community care in schizophrenic patients in Japan.

A study on the readmission of schizophrenic patients discharged after long-term hospitalization was conducted by Higuchi and Hayashi.18 This study was conducted in one of the largest psychiatric hospitals in Japan but its subjects did not have a comparison group. The present study was conducted in the more common, middle-sized hospitals with a case–control design.

Risk factors of rehospitalization

The multiple logistic regression analysis revealed that the pattern of psychiatric treatment is associated with rehospitalization. Those who did not visit clinics in the second month prior to the entry and those who visited clinics more frequently than five times in the previous month were liable to readmit to the hospitals. These results may be interpreted such that absence of clinic visit induces clinical exacerbation, resulting in frequent clinic visits followed by rehospitalization. Ohara et al. reported that those who changed doctors or took medication discontinuously were readmitted more significantly.9 The importance of continuous clinic visits to continuance of living in the communities was emphasized even though clinical status was steady.20 This result indicates the importance of regular clinic visits. Therefore, to prevent the irregular pattern of clinic visits, close monitoring of the number of clinic visits and use of the outreach team would be of help to those patients who are reluctant to come to the hospital.

The efficacy of the aforementioned approaches is still controversial in Europe, especially in England,21,22 where mental health service plays a major role in community care for the mentally ill. However, the community care by mental health services has not been developed sufficiently in Japan, and the aftercare for those who are discharged from psychiatric hospitals is mostly organized by the outpatient units in the hospitals. The close monitoring of clinic visits or outreach by nurses are mostly hospital-based. Therefore, these approaches may produce outcomes that are different from those in Europe, especially England.

Interestingly, the utilization of the day-care unit at the public health center and workshops in the community were related to rehospitalization. Presumeably, when patients attending these facilities deteriorate clinically, the staff have few choices other than to advise them to visit an outpatient clinic or to admit to hospital, which then leads to rehospitalization. There are options other than rehospitalization such as home visits by public health nurses. Although these systems are not functioning well at present, they should be integrated with medical approaches.

Junior high school graduation as education level was revealed to be another risk factor of rehospitalization. In other words, educational level is an important factor in secondary prevention of schizophrenic episodes. While most Japanese go to senior high school, those who did not enter senior high school or those who quit senior high school were at risk of rehospitalization in the present study. It is possible that this reflects a problem of social function such as school refusal or being the object of bullying, as well as being a problem of intelligence. This social dysfunction may be ascribable to a premorbid state or onset of schizophrenia. From a point of view of primary prevention, the early detection of this population is required because treatment delay may affect the course of schizophrenia. In the study by Johnstone et al., untreated illness was the strongest predictor of relapse.23 Moreover, Loebel et al. showed that the time to remission as well as the degree of remission were closely related to duration of untreated psychosis when other prognostic variables were controlled.24 As stated by Birchwood et al., these findings may be the result of type of illness (those with an inherently high risk of relapse may include symptoms that delay treatment) or the result of psychosocial disruption arising from a long period of untreated psychosis.25 In the present study, age of onset and type of schizophrenia were not related to rehospitalization, suggesting that psychosocial disruption resulting from untreated illness is more important than type of illness in influencing the course of schizophrenia including rehospitalization. Therefore, we suggest that early intervention is important for those who show any signs during school age. The education level itself, however, would not be a modifiable factor. We should intervene to expose such vulnerable people to a better social experience.

Poor medication compliance was not significant in the multivariate analysis, although previous studies by Sullivan et al. and Corrigan et al. revealed it to be a strong indicator of rehospitalization.26,27 One may wonder if there is a limitation in the way we assessed compliance (i.e. the use of compliance scale of Kemp et al.13). However, as pointed out by Kemp et al., there are problems associated with more direct measures.13 Urine tests may overestimate compliance when drugs have a long half-life. Serum assays are invasive and are of limited value in assessing partial compliance. Pill counts  are  widely  considered  a  useful  measure  but this method has a potential for dissimulation and inaccuracy.

Sullivan et al. indicated that non-compliance was associated with lack of insight (i.e. denial of mental illness), a history of fewer outpatient appointments, and being single.26 In another study the most powerful indicators of compliance were reported as being a good attitude to treatment and insight into illness.28 The average age of the patients in the study by Sullivan et al. study was 34.5 years,26 which is younger than that in the present study by more than 10 years. Taking this age difference into account, it is suggested that the subjects in our study accepted the illness and took medication, in comparison with the patients of Sullivan et al. In further studies, not only the therapeutic behaviors as evaluated in the present study but also the attitude toward medication therapy including insight into the illness, should be examined.

Incidentally, poor medication compliance was associated with rehospitalization in the bivariate analyses. Because compliance was moderately correlated with the present GAF scores (r = 0.31), the possibility that controlling the latter factors produced the negative result in the multivariate analysis cannot be excluded entirely.

We asked the subjects if they experienced life events in the past 3 months. On the interview we categorized the life events according to axis IV in DSM-IV, which was developed to describe patients in a biological, psychological and social model in a clinical situation. The answers to this question may reflect subjective experience of the interviewees. Because we employed the stress–vulnerability model of schizophrenia, in which stress plays an important role in understanding the episode of schizophrenia, the results of this life event are still informative for risk factors of rehospitalization.

In the multiple logistic regression models, life events did not remain at a significant level. The previous study suggested that life events were associated with relapse irrespective of the compliance problem,29 although it was not clear as to which domain of life events was more influential. Many of the life events in the present study were problems in the primary support group, especially families. These may be the results of patient–family interactions such as over-involvement in patient symptoms or an excess burden of patient care. According to the study by Tanaka et al., emotions of the family influence the course of schizophrenia.30

Finally, because the present study was conducted with a case–control design, we could not clarify the causal relationship. A further prospective study should be conducted to elucidate the causal relationship between the factors presented here and rehospitalization of schizophrenic patients.

ACKNOWLEDGMENTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. REFERENCES

We would like to acknowledge the invaluable assistance of patients, social workers, and clinical staff of the hospitals. We also acknowledge the contribution of the psychiatrists in the participating hospitals: Nihonmatsu-kai Yamagata Hospital, Director, Dr Tobisawa, Dr Hirose, Dr Oyama, Dr Miura, Dr Uchigasaki, Dr Tanaka, Dr, Ichikawa, Dr Ito, Dr Kyono, Dr Iwabuchi, Dr Takahashi; Nihonmatsu-kai Kaminoyama Hospital, Director, Dr Yokokawa, Dr Tomita, Dr Igarashi, Dr Eguchi, Dr Ohori, Dr Kuwayama, Dr Yamada, Dr Goto; Shinoda Kosei-kai Chitose Shinoda Hospital, Director, Dr Yoshida, Dr Kato, Dr Yoshimura, Dr Kimura, Dr Kizu. Dr Ihara and Dr Oiji provided helpful comments and suggestions on an earlier draft of this paper. Dr Imuta, Dr Abe, Dr Goto, and Dr Nishise kindly commented on the study. We thank Ms Joy Norton, who reviewed the English manuscript of this article.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. REFERENCES
  • 1
    Mino Y, Kodera R., Bebbington P. A comparative study of psychiatric services in Japan and England. Br. J. Psychiatry 1990; 157: 416420.
  • 2
    Fenton WS, Mosher LR, Herrell JM, Blyler CR. Randomized trial of general hospital and residential alternative care for patients with severe and persistent mental illness. Am. J. Psychiatry 1998; 155: 516522.
  • 3
    Knapp M, Marks I, Wolstenholme J et al. Home-based versus hospital-based care for serious mental illness. Br. J. Psychiatry 1998; 172: 506512.
  • 4
    Clarkson P, McCrone P, Sutherby K, Johnson C, Johnson S, Thornicroft G. Outcomes and costs of a community support worker service for the severely mentally ill. Acta Psychiatr. Scand. 1999; 99: 196206.
  • 5
    Henderson C, Phelan M, Loftus L, Dall’Agnola R, Ruggeri M. Comparison of patient satisfaction with community-based vs. hospital psychiatric services. Acta Psychiatr. Scand. 1999; 99: 188195.
  • 6
    Taylor RE, Leese M, Clarkson P, Holloway P, Thornicroft G. Quality of life outcomes for intensive versus standard community mental health services. PRiSM psychosis study 9. Br. J. Psychiatry 1998; 173: 416422.
  • 7
    Weiden P, Glazer W. Assessment and treatment selection for ‘revolving door’ inpatients with schizophrenia. Psychiatr. Q 1997; 68: 377392.
  • 8
    Salokangas RKR. Living situation, social network and outcome in schizophrenia: a five-year prospective follow-up study. Acta Psychiatr. Scand. 1997; 96: 459468.
  • 9
    Ohara H, Mimura K, Suzuki J, Nagamatsu K, Nagamatsu I, Osaka T. A follow-up study of schizophrenia (Report 1). Characteristics of the patients and social factors influenced on readmission of schizophrenics. Nippon Koshu Eisei Zassi (Jpn J. Public Health) 1981; 28: 522532 (in Japanese).
  • 10
    Masaki N, Fujita T, Kai S et al. Readmission among discharged psychiatric patients and its correlates. Nippon Koshu Eisei Zassi (Jpn J. Public Health) 1997; 44: 372383 (in Japanese).
  • 11
    Nuechterlein KH, Dawson ME. A heuristic vulnerability/stress model of schizophrenic episodes. Schizophr. Bull. 1989; 10: 300312.
  • 12
    American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, 4th edn. American Psychiatric Association, Washington, DC, 1994.
  • 13
    Kemp R, Hayward P, Applewhaite G, Everitt B, David A. Compliance therapy in psychotic patients: randomized controlled trial. BMJ 1996; 312: 345349.
  • 14
    American Psychiatric Association. Practice guideline for the treatment of patients with schizophrenia. Am. J Psychiatry 1997; 154 (Suppl.).
  • 15
    Guy W. CGI. Clinical Global Impressions. In: Guy W (ed.) ECDEU Assessment Manual for Psychopharmacology. National Institute of Mental Health, Rockville, MD, 1976; 218222.
  • 16
    Ciompi L. The natural history of schizophrenia in the long term. Br. J. Psychiatry 1980; 136: 413420.
  • 17
    Huber G. The heterogeneous course of schizophrenia. Schizophr. Res. 1997; 28: 177185.
  • 18
    Higuchi M, Hayashi N. A study of the readmission of schizophrenic patients after long-term hospitalization. Seisin Igaku 1996; 38: 245251 (in Japanese).
  • 19
    Ohmori T, Ito K, Abekawa T, Koyama T. Psychotic relapse and maintenance therapy in paranoid schizophrenia: a 15 year follow up. Eur. Arch. Psychiatry Clin. Neurosci. 1999; 249: 7378.
  • 20
    Masuno H. A study on relapse of schiaophrenia. Seisin Igaku 1976; 18: 11471154 (in Japanese).
  • 21
    Burns T, Teadsmoore A, Bhat AV, Oliver A, Mathers C. A controlled trial of home-based acute psychiatric services. I. clinical and social outcome. Br. J. Psychiatry 1993; 163: 4554.
  • 22
    Tyrer P, Morgan J, Van Horn E et al. A randomised controlled study of close monitoring of vulnerable psychiatric patients. Lancet 1995; 345: 756759.
  • 23
    Johnstone EC, Crow TJ, Johnson A, MacMillan JF. The Northwich Park Study of first episode, schizophrenia. I. Presentation of the illness and problems relating to admission. Br. J. Psychiatry 1986; 148: 115120.
  • 24
    Loebel AD, Lieberman JA, Alvir JMN. Duration of psychosis and outcome in first episode schizophrenia. Am. J. Psychiatry 1992; 149: 11831188.
  • 25
    Birchwood M, McGorry P, Jackson H. Early intervention in schizophrenia. Br. J. Psychiatry 1997; 170: 25.
  • 26
    Sullivan G, Wells KB, Morgenstern H, Leake B. Identifying modifiable risk factors for rehospitalization: a case-control study of seriously mentally ill persons in Mississippi. Am. J. Psychiatry 1995; 152: 17491756.
  • 27
    Corrigan PW, Liberman RP, Engel JD. From noncompliance to collaboration in the treatment of schizophrenia. Hosp. Community Psychiatry 1990; 190: 12031211.
  • 28
    Kelly GR, Mamon JA, Scott JE. Utility of the health belief model in examining medication compliance among psychiatric outpatients. Soc. Sci. Med. 1987; 11: 12051211.
  • 29
    Hirsch S, Bowen J, Emami J et al. A one year prospective study of the effect of life events and medication in the aetiology of schizophrenic relapse. Br. J. Psychiatry 1996; 168: 4956.
  • 30
    Tanaka S, Mino Y, Inoue S. Expressed emotion and the course of schizophrenia in Japan. Br. J. Psychiatry 1995; 167: 794798.