Social Psychiatryand
Soc Psychiatry Psychiatr Epidemiol (1992)27:46-52
PsychiatricEpidemiology
9 Springer-Verlag 1992
Patterns of aftercare for psychiatric patients discharged after short inpatient treatment An Italian collaborative study A. Barbato 1, Emanuela Terzian 2, B. Saraceno 2, E Montero Barquero 2, and G. Tognoni 2
Centro Psicosociale,Unit~ Operativa di Psichiatria PoliclinicoII, Milano, Italy a Laboratorio di Farmacologia Clinica, Unit~ di Psichiatria Istituto di Ricerche Farmacologiche "Mario Negri", Milano, Italy Accepted: September 9, 1991
Summary. This is the first of a series of papers presenting the results of an Italian collaborative study on psychiatric inpatient service utilisation. Patterns of care of a large sample of patients discharged after short inpatient treatment are discussed in the light of the changes introduced by the 1978 Mental Health Act in the Italian psychiatric care delivery system. Three closely related issues are considered: a) use of psychiatric hospitalisation, b) relationships between inpatient and community treatment before and after an admission episode, c) continuity of care. The main findings are: 1) great variability between services suggests that local factors play an important role in determining the contents of care in Italian post-reform psychiatry. 2) The relationship between inpatient and community services is complex, partial integration being the most common picture. 3) Psychiatrichospitalisation is the entrypoint into the care system for a sizeable group of patients.4) Continuity of care is achieved for half the patients, mostly with diagnoses of severe mental disorders. 5) Subjects with a recent history of revolving door behaviour or a past history of mental hospital admission show the highest likelihood of remaining in community care following discharge.
The 1978 Mental Health Act introduced radical changes in the Italian mental health system. The main features of the new legislation, as well as its historical and conceptual background, have already been fully described [1-3]. Only those aspects relevant to this paper will therefore be briefly outlined here. a) No new patients could be admitted to public mental hospitals. Former patients could be readmitted voluntarily until 1981 (in some administrative regions this deadline was extended to 1983). b) Community-based services were to be set up to provide the full range of psychiatric treatments, including residential treatment and hospitalisation, for specified catchment areas. c) Small psychiatric units (Servizi Psichiatrici di Diagnosi e Cura, SPDC), linked with community psychiatric services (Servizi Psichiatrici Territoriali, SPT), were to be es-
tablished in general hospitals, to which all patients requiring both voluntary and compulsory hospitalisation were to be admitted. Over the last decade the Italian psychiatric reform has been a very controversial issue. De Girolamo [4], in a careful review of the available international literature, effectively illustrates the strong polarisation of opinions existing between enthusiastic supporters and vehement detractors. Unfortunately, this heated debate arose in the absence of reliable data on the implementation of the new model and on its effects on the provision of mental health services, not to speak of the more complex question of quality of care. In fact, not only have the substantial innovations not been accompanied by any systematic evaluative research [5, 6], but the shortage of updated national data hinders even crude assessments: to date no national data banks have yet been established to monitor the activities of old and new services. The present study was designed as a step towards filling the gap. It had two main aims: a) to investigate patterns of care and outcome of patients in a large sample of psychiatric patients, and b) to explore the role played by psychiatric hospitalisation within the new model of psychiatric care delivery. SPDCs were chosen as the basis of the investigation, in so far they were set up to serve as the main National Health Service setting for the hospitalisation of psychiatric patients. After 1978, SPDCs were established on a nation-wide basis, albeit unevenly distributed. They are small units (generally comprising at most 15 beds), usually serving well defined catchment areas; the average beds/population ratio is 5.4/100000. In 1984 there were 236 such services, with approximately 3100 beds [7]. Since virtually no information concerning their functioning was available, a large-scale collaborative study was set up in order to obtain a reliable picture. Methods
Thirty-four SPDCs located in eight of the 21 Italian regions (five Northern, one Central and two Southern), took
47 part in the study on a voluntary basis. Their catchment areas comprise 27 % of the adult population of those regions. The SPDCs were not randomly selected; an invitation to participate in the study was issued to those involved in earlier years in other research projects with the coordinating centre. The study started in February 1984; thirty patients consecutively admitted to each SPDC were to be included in the study. Data on all patients were collected during hospital stay by members of the treatment team (usually psychiatrists). The following information was recorded in an ad hoc form: | social and demographic characteristics; 9 psychiatric history; 9 type of admission (voluntary/compulsory, referral source); 9 symptoms observed on admission; 9 problems leading to admission; 9 DSM III diagnosis at discharge; 9 treatment received in the first 4 hours, between the 5th and the 12th hour after admission, and half-way throughout hospital stay; 9 overall clinical judgement at discharge (recovered, improved, unchanged, worsened); | referral for aftercare at discharge. A follow-up was planned 6 and 12 months after discharge. The services were asked to make every effort in order to trace all patients and to gather data concerning: 9 living arrangements; 9 use of aftercare services; 9 number of readmissions and length of time spent as inpatients throughout the follow-up period. Data analysis concerned three main aspects: patterns of care, outcome of care, and clinical management-i, e., pharmacological treatments and type of admission. This paper focuses specifically on the analysis of the patterns of care. In order to investigate the possible associations between the contents of care, outcomes (or interventions), and patients' characteristics, the latter were used to construct three summary variables [8], each scoring from 1 to 3 (the intermediate score value being attributed to subjects not belonging to either of the extreme categories): Social network High score. Subjects living with a partner and/or own children; living alone or with primary family only if employed. L o w score. Unemployed/economically dependent and living alone or with others (non-family members).
Psychiatric history High score. Subjects at their first contact with a psychiatric service. L o w score. Subjects with 3 or more prior admissions in SPDC, or admitted at least once in a mental hospital, whose psychiatric history dated from over one year.
The determinants of referral at discharge were analysed in two stages. In the first place, univariate correlation with prognostic variables was explored; subsequently, in order to estimate the effects of the independent variables considered, a multivariate analysis was carried out by logistic regression [9], using the EPILOG statistical package. It was conducted by steps within each block of exposure variables (patients characteristics: age, sex, social network, social status, psychiatric history, diagnosis, care provider prior to admission; admission- and treatment-related variables: problems leading to admission, type of admission, drug treatments, duration of hospital stay; SPD C of admission) including those appearing in the univariate analysis significantly associated with the dependent variable. The final model was eventually obtained by entering the variables of each block which significantly contributed to explain the outcome. 1143 patients were originally recruited in the study by 36 SPDCs. Two services withdrew from participation in the early stages, and the corresponding 61 patients were therefore excluded from all analyses. Five patients - admitted to two services during the study period - were recruited twice; only the data derived from their first admission were retained for analysis. Patients' characteristics, type of admission, inpatient treatments and referral at discharge were therefore available on 1077 patients recruited in 34 SPDCs. Since some services failed to collect follow-up data at the established intervals, and others had proved unable to trace a substantial proportion of their patients, it was decided - in order to avoid introducing a potentially important bias - to restrict the analyses requiring follow-up data to the subgroup recruited by SPDCs where a) the first and second follow-up had actually taken place 5 to 9 months and 10 to 18 months following discharge respectively, and b) at least 90 % of the patients originally recruited had been traced for follow-up. Patients deceased before the first follow-up were also excluded from analyses on follow-up data. Hence, evaluation of patterns of aftercare was conducted on 569 patients recruited by 21 SPDCs. Results The study population
Social status High score. Educational level beyond primary school qualification (if retired or housewife); skilled workers, white-collars, or in higher order of employment. L o w score. Primary school qualification or less (if retired or housewife); never-employed subjects on welfare benefits.
Social, demographic, diagnostic, and psychiatric history features of the 1077 patients recruited in the study are shown in Tables i and 2. In the analysis DSM III diagnoses have been grouped into broad categories. Data on admission- and treatment-related variables (type of admission, length of stay, drug treatments, referral for aftercare) are shown in Table 3.
48 Table 1. Social and demographic variables % Patients Age: < 25 yrs 25-44 yrs 45~54 yrs > 65 yrs Sex: Males Marital status: Married/cohabiting Social status: Low score Social network: Low score Living arrangement: Own home Residential facility Nursing home Homeless
12.6 44.6 34.4 8.4 53.4
Variability between SPDCs (%) 0-30 28~52 I , 17-50 1 0-27
33.7
7-57*
10.9
0-53*
11.1
0-27"
91.3 1.7 2.8 1.9
73-1000-14 [, 0-13 | 0-13
* P < 0.001 Table 2. Psychiatric history % Patients Variability between SPDCs (%) Diagnosis: Schizophrenic disorders Affective disorders Personality disorders Depr. neurosis/anxiety Substance abuse Problems: Symptom-related Attempted suicide Aggressive behaviour Refusal of treatment Family/housing/work First contact Revolving door ( > 2 admissions SPDC) Prior admission to mental hospital Pre-admission care: no care SPT
43.4 12.9 7.8 13.7 11.0
20-70 -
36.9 9.5 11.2 8.4 34.0 19.7 25.0
7-76 0-28 0-43* 0-33 3-70
35.0
10-60"
42.0 56.0
7-84 12-90"
0-38 0-40
7-45*
* P < 0.001
anxiety and substance abuse disorders (taken together) are unevenly distributed, representing under 25 % of patients in the whole sample, and up to 40 % of the population in SPDCs with fewer patients affected by schizophrenic disorders. Major affective disorders appear more evenly distributed between the services. The patient populations also differ significantly with respect to other variables, such as previous psychiatric history - as indicated, for instance, by the proportion of revolving door patients, or of individuals with a past history of admission to mental hospitals - and the types of problems leading to admission. All these differences, however extreme, cannot alone be held responsible for the remarkable heterogeneity of pre-admission care patterns of the population of patients, plausibly in turn a function of local differences in the availability of community resources, or of the level of integration existing between inpatient and community services. Interestingly, SPDCs show the widest ranges of variation for several key treatment indicators - such as the proportion of compulsory admissions, or of patients treated with high doses of antipsychotic drugs, or referred to community-based services. Length of hospitalisation varies to a lesser degree; a general tendency is apparent (with few exceptions) towards short-term inpatient treatment. Despite this degree of heterogeneity, overall the population described in this study seems closely to match in a number of key features most series of acute psychiatric services users described in the literature [10]. This holds true particularly for clinical severity, length of psychiatric history and degree of social disability. On the other hand, important differences emerge which ought not be overlooked. Irrespective of severity of illness and degree of impairment in social role functioning, the large majority of patients in the study show some social adjustment, and appear surrounded by supportive networks: many subjects live with their families, or (if alone) in their own house, very few are homeless or in sheltered accommodation. Homelessness or residential instability has been documented in large numbers of patients admitted to acute psychiatric inpatient units, particularly in the USA [11, 12]. Moreover, metropolitan areas have been shown to
Table 3. Patterns of inpatient care (CPZ: chlorpromazine) % Patients
Both patients' characteristics and treatment modalities vary widely between SPDCs. The differences are highly statistically significant for all variables, the only exceptions being the sex distribution and the proportion of first-onset cases. Although the majority of patients are 25 to 64 years old, in some services up to one-third of patients are younger (under 25) or older (65 or over). In the extreme age groups most patients were diagnosed as having organic mental disorders (mental retardation or dementia). Schizophrenic disorders form the largest diagnostic group (over 40 %); in some services however their proportion falls to only 20 per cent. Similarly depressive neurosis,
Compulsory admission Length of stay: < 1 week > 1 month Antipsychotic drug use: < 800 mgEq CPZ during hospital stay Referred to SPT at Discharge: * P < 0.001
23.5
Variability between SPDCs (%) 0-96*
28.2 12.3
0-57* 0-45*
8.9
0.53*
71.2
27-90*
49 % 80
60-
40-
20-
Opre-Qdrnission
referred to
no core
6 m t h s FU
~
SPT
~
12 m t h s FU
other agency
Fig.1. Patterns of care: % distribution of patients among care providers
no core pre-odmission
loo in SPT85core pre-odmission 80-
60-
40-
20-
Oreferred tD
6 mths El
12 mths no e~re
~
referred to
s1~r ~
6 mthe
12 mth~
oth~ Qeency
Fig.2. Percentage distribution of pts in SPT care/not in care pre-admission, by care providers
entail a relatively high risk of homelessness [13]. The remarkably low prevalence of homelessness observed in this study, even for patients recruited by services in northern metropolitan areas (Milano, Genova, Torino), suggests that - at least so far - homelessness does not represent a common aftermath of deinstitutionalisation in these areas. Indeed, since none of the services of the two main central-southern metropolitan areas (Rome and Naples) participated in the study this result cannot be generalised to the whole country. However, though the study design calls for some caution in the interpretation of this observation, it would probably be unreasonable to postulate that services' self-selection in Northern cities might have introduced an appreciable bias in this respect.
Patterns o f care
As mentioned earlier, an important aim of the study was to identify and describe prevalent patterns of psychiatric services utilisation. To this end, three closely related aspects have been considered: a) use of psychiatric hospitalisation; b) relationships between inpatient treatment and treat-
ment in the community before and after the admission episode; c) continuity of care. Prior to the index admission, 39 % of patients were not in psychiatric care, and 59 % were treated by SPTs (Fig. 1). At discharge, 72 % of patients are referred for aftercare to SPTs and 21% to other care providers (general practitioners, private psychiatrists, social services or other agencies). At the time of the second follow-up, the proportion of subjects in care of SPTs fell to pre-admission levels, while the proportion of those treated by others remained stable. Of the latter group, 31% of patients were in the care of private psychiatrists, 27 %, of general practitioners, and 20 %, of social services. The picture appears quite different if two subgroups of subjects are examined separately, namely, patients in SPT care and those not in care prior to admission (Fig. 2). Almost all patients previously treated by SPTs were referred to SPTs at discharge, and by and large they were still in treatment 12 months later. On the contrary, only 56% of previously untreated patients were referred to SPTs at discharge, while one third were referred to other agencies. Moreover the proportion of previously untreated subjects in SPT care steadily declined during the follow-up period, more markedly than that of subjects in care of other agencies. In other words, past history of treatment by community psychiatric services is an important predictor of further treatment by SPTs after discharge. Once more, the degree of heterogeneity among SPDCs is considerable and several instances of departure from the general pattern can be identified. The proportion of patients in SPT care prior to admission ranged from 12 to 90 %; accordingly, the proportion of previously untreated subjects varied between 7 and 84 per cent. One year after discharge, however, the differences between services appear somewhat narrowed: the proportions of patients in SPT care range from 34 to 86 % and those of patients not in care from 0 to 43 per cent. At the time of the second follow-up, general practitioners were by far the most important care providers for patients discharged by one particular service. In addition, diagnosis contributes significantly to the prediction of patterns of aftercare (Table 4). Patients with severe mental disorders - such as schizophrenic or major affective disorders - were more often referred to SPTs, and were more likely to remain in SPT care during the follow-up period. Subjects with less severe disorders followed a quite different pattern. One year after discharge one third were in SPT care, one third were treated by other agencies, and one third were not in care. Patients presenting substance abuse disorders were least likely of all to use any aftercare service. Within six months following discharge, 42% of the patients were readmitted for psychiatric treatment (33 % to SPDCs, 5 % to private hospitals and 4 % to other inpatient care settings). Low social network scores and previous psychiatric history are significantly associated with readmission. A detailed analysis of readmissions and of other outcome indicators will be presented in the second paper of this series.
50 Table 4. Percentage distribution of patients by diagnostic groups and type of care providers at 6 and 12 months Diagnostic groups
Schizophrenic disorders Major affective disorders Depressive neurosis, anxiety Substance abuse disorders
Care providers SPT Other agency 6 12 6 12 Mths Mths Mths Mths 75 61 16 17
Not in care 6 12 M t h s Mths 10 11
67
64
24
26
9
10
47
42
36
30
16
27
40
37
25
26
35
37
Continuity of care According to Bachrach [14], continuity of care can be defined as a process involving the uninterrupted movement of patients over time through the diverse elements of the service delivery system. For several years the continuity of care model has gained wide acceptance as the best answer to the problems of service delivery in the community for the severe, long-term mentally ill. Continuity of care within public mental health services has been evaluated in this study by examining the pathways taken by patients through the care delivery system. To this end the interaction between patients and services at three points in tinre (prior to admission, referral at discharge, and 6 months later) was assessed. At each point three alternatives were explored, i.e., at discharge: no referral, referral to SPT and referral to other care providers, and in the other two occasions: no care, SPT care and care provided by other agencies. Continuity of care was deemed to have been achieved for patients not in care, or in SPT treatment prior to admission, who had been referred to SPT at discharge and were still SPT users after six months. 51% of the patients experienced continuity of care thus defined. Two different explanations could apply to the lack of continuity of care experienced by the remaining 49 % : inadequate integration between inpatient and community services, or inappropriate hospitalisations. The latter seems particularly plausible for patients diagnosed as having anxiety or substance abuse disorders, often discharged within a week, who would probably benefit more from interventions other than hospitalisation. Discontinuation of care is most likely soon after discharge, due to non-referral or to poor coordination between hospital and community providers. Relatively few patients among those in SPT care six months after discharge appear subsequently to abandon treatment.
referral for aftercare is an important indicator of the degree of integration between hospital and community services, critically relevant to the continuity of care. The potential risk factors for non-referral to SPTs after discharge - i.e. the variables positively associated with this choice on the part of the hospital team - were therefore explored in depth. The variables jointly comprising statistically significant independent determinants of outcome (non-referral to SPTs) in the multivariate analysis are listed in the logistic regression model reported in Table 5, along with the relevant relative risks (RR), confidence limits (CL) and P-values. Among patients characteristics, age > 65, a diagnosis of substance abuse disorders or of "other" disorders (mostly mental retardation and dementia), and a low social network score appear to contribute significantly to non-referral. Taken together, these factors probably indicate to a large extent a policy of referral of elderly patients, lacking family support, to services other than SPTs (mainly social services). Interestingly, a past history of SPT care is the strongest predictor of referral to SPT, while diagnosis - though significantly contributing to the model - loses most of its explanatory power after adjusting for the remaining independent variables. Among the problems leading to admission, the best predictor of referral to SPT is treatment refusal. Of the second block of variables, only duration of hospital stay is significantly associated with the referral decision. Patients hospitalised for 1 to 4 weeks were those most likely to be referred to SPTs at discharge. This finding could be partly explained by the fact that the majority Table 5. Determinants of non-referral to SPT (logistic regression model; log-likelihood: - 488.72;Z2:286.43,49 df) RR b
P
0.16-0.33 0.43-3.93
0.000 0.644
1.55-4.83
0.000
1.03-2.10 1.34-3.89
0.003 0.002
0.34-0.74 0.48-1.38
0.000 0.450
0.45-1.39 0.41-1.31 0.13-0.61 0.48-1.10
0.420 0.288 0.001 0.128
0.87-2.56 0.69-2.43 1.31-3.76 0.67-1.94 1.40-4.21
0.141 0.414 0.003 0.626 0.002
Multivariate analysis. Since the transition from the hospi-
No care (pre-admission) In SPT care Other agencies Age < 65 65 + Social network High Medium Low Hospital stay (wks) < 1 1- < 4 > =4 Problems Symptom-related Attempted suicide Aggressive behaviour Treatment refusal Work/housing/family Diagnosis Schizophrenic d. Affective disorders Personality disorders Substance abuse d. Depressive neurosis Other
tal to the community after an admission episode is a key aspect of the long-term treatment of mental disorders,
a Reference categories; b Estimates from LogReg equations including term for SPDCs.
Determinants of referral at discharge
a 0.23 1.23 a 2.74 a 1.47 2.29 a 0.50 0.82
CL
0.79 0.73 0.28 0.72 a
1.50 1.30 2.22 1.14 2.42
51 of patients with minor mental disorders, often not referred to SPTs, constitute a large proportion of very short hospitalisations. Finally, the service itself appears to play a critical role in determining non-referral after adjustment for the other variables; for eleven SDPCs the relative risks significantly depart from unity.
Discussion
The study was undertaken on the assumption that the analysis of the patterns of care received by a large, unselected sample of psychiatric patients experiencing an admission episode could contribute relevant information on psychiatric care delivery services in Italy after the changes introduced in 1978. The model underlying the Mental Health Act clearly entails continuity of care, local availablility of the full range of treatments, long-term support for patients and their families, and integration among services providing different components of care. The existence of a wide consensus around these aims is also evinced in the international literature [15]. Recent studies support the view that integration between hospital and community services is a viable, economically convenient solution to improving delivery of care by mental health services, and have led to suggestions that hospitalisation should only exceptionally represent the entry point in the treatment system-ideally, patients should be given inpatient care only as a last resort after the failure of community treatment [16]. To what extent do the patterns of care described above reflect the underlying model? The repeatedly stressed variability between services indicates that local factors whether related to the social or geographical characteristics of the catchment areas, or to different organisational and clinical models, or to the availability of community resources, or to any combination of the above are very likely to play a major role in determining the contents of care. Furthermore, there is currently no widespread consensus on the role and specific functions attributed to SPDCs; nor has an agreement been reached concerning the categories of patients pertaining to SPDCs, or indeed to psychiatry. The different cultural determinants of the Italian post-reform psychiatry (as has generally happened in post-deinstitutionalisation psychiatry worldwide) leave important "grey areas" to the discretion of individual services, which in practice are left to define for themselves the boundaries of their assignment (typical examples are alcoholics, drug abusers, vagrants, or homeless people). The decision of being or not involved in the care of a substantial fraction of mentally disturbed patients therefore is not necessarily (or not only) a result of scarce financial and human resources; often it reflects explicit choices of individual services. Further research is needed to clarify this point, possibly with the aim of assessing the impact of service characteristics on patients' outcome. Indeed, the heterogeneity observed strengthens previous observations on the different ways in which the legislation was implemented [17]. In this respect, further efforts to collect broad administrative
data at a national or regional level seem to be of limited value. The general picture resulting from this study can be summarised as follows: 1. The relationships existing between SPDCs and SPTs are complex, and cannot be explained according to one model. In some instances full integration is achieved; more commonly features indicating partial integration are observed. On the whole, integration is more likely to materialise for patients with severe mental disorders and a previous history of psychiatric treatment. Although limited availability of SPTs in some areas might be an obvious explanation of the discontinuity of care following discharge, it cannot be held to represent the only cause, particularly since a significant association with non-referral has been shown for SPDCs located in areas where community services are available and adequately equipped. 2. SPDCs are the entry point into the care system for a sizeable group of subjects; often they appear to relate to community services other than SPTs, particularly general practice and social services. A large proportion of patients with minor mental disorders are referred to GPs, and there are evidences indicating that such patients are more easily kept in treatment by the latter than by SPTs. This calls for a thorough exploration of the role of general practitioners in the care of a broad range of minor mental disorders, a field so far virtually unexplored in Italy. 3. SPTs represent with few exceptions the only providers far chronic or clinically severe patients, for 50 % of whom continuity of care was achieved. The private sector has a very limited function in the care of these patients, as opposed to the results of observations carried out in other European countries [18]. The subjects presenting a recent history of revolving door behaviour or a past history of mental hospital admissions show the greatest likelihood of remaining in SPT care in the follow-up period; in this respect the Italian model seems again to be at odds with the pattern resulting from international experience [12]. As expected, the inherent character of the Italian organisation and structure of mental health care services determines unique care delivery patterns - to an extent which hinders any exercise of international comparison. Whether and how this peculiar structure also determines different outcomes will form the object of a subsequent publication.
References
1. Mosher LR (1982) Italy's revolutionary mental health law:an assessment. Am J Psychiatry 139:199-203 2. Tansella M, WilliamsP (1987) The Italian experience and its implications. Psychol Med 17:283-289 3. Saraceno B, Tognoni G (1989) Methodological lessons from the Italian psychiatricexperience. Int J Soc Psychiatry 35:98-109 4. De Girolamo G (1989) Italian psychiatry and reform law: a review of the international literature. Int J Soc Psychiatry 35:21-37 5. Morosini PL, Repetto F, De Salvia D, Cecere F (t985) Psychiatric hospitalisation in Italy before and after 1978. Acta Psychiatr Scand 71 [Suppl 316]: 15-26
52 6. Bollini E Mollica RF (1989) Surviving without the asylum: an overview of the studies on the Italian reform movement. J Nerv Ment Dis 177:607~515 7. Crepet P (1990) A transition period in psychiatric care in Italy ten years after the reform. Br J Psychiatry 156:27-36 8. Avison WR, Speechley KN (1987) The discharged psychiatric patient: a review of social, social-psychological, and psychiatric correlates of outcome. Am J Psychiatry 144:10-18 9. Kleinbaum DG, Kupper LL, Morgenstern H (1982) Epidemiologic research: principles and quantitative methods. Van Nostrand, New York 10. Bassuk E, Winter R, Apsler R (1983) Cross-cultural comparison of British and American psychiatric emergencies. Am J Psychiatry 140:180-184 11. Chafetz L, Goldfinger SM (1984) Residential instability in a psychiatric emergency setting. Am J Orthopsychiatry 56:20-34 12. Surles RC, McGurrin MC (1987) Increased use of psychiatric emergency services by young chronic mentally ill patients. Hosp Comm Psychiatry 38:401-405 13. Bachrach L (1987) The homeless mentally ill. In: Menninger W, Hannah G (eds) The chronic mental patient-II. American Psychiatry Press, Washington, pp 65-92
14. Bachrach L (1981) Continuity of care for chronic mental patients: a conceptual analysis. Am J Psychiatry 138:1449-1556 15. Fuller Torrey EF (1986) Continuous treatment teams in the care of the chronic mentally ill. Hosp Comm Psychiatry 37:1243-1247 16. Tyrer R Turner R, Johnson AL (1989) Integrated hospital and community psychiatric services and use of inpatient beds. BMJ 299:298-300 17. Tognoni G, Saraceno B (1989) Regional analysis and implementation. Int J Soc Psychiatry 35:38-45 18. Gaebel W, Pietzcker A (1987) Prospective study of the course of illness in schizophrenia: III. Treatment and outcome. Schizophr Bull 13:307-316
Dr. A. Barbato Centro Psicosociale Unit~ Operativa di Psichiatria Policlinico II, USSL 75/4 Via Ravenna 13 1-20139 Milano Italy
Announcement XIIth International Conference on the Social Sciences and Medicine 14-18 September 1992, Peebles, UK F o r f u r t h e r i n f o r m a t i o n , p l e a s e contact: Dr. P. J. M. M c E w a n , G l e n g a r d e n , Ballater, A b e r d e e n s h i r e A B 3 5 5 U B , S c o t l a n d , U K
Soc Psychiatry Psychiatr Epidemiol (1992)27:46-52
PsychiatricEpidemiology
9 Springer-Verlag 1992
Patterns of aftercare for psychiatric patients discharged after short inpatient treatment An Italian collaborative study A. Barbato 1, Emanuela Terzian 2, B. Saraceno 2, E Montero Barquero 2, and G. Tognoni 2
Centro Psicosociale,Unit~ Operativa di Psichiatria PoliclinicoII, Milano, Italy a Laboratorio di Farmacologia Clinica, Unit~ di Psichiatria Istituto di Ricerche Farmacologiche "Mario Negri", Milano, Italy Accepted: September 9, 1991
Summary. This is the first of a series of papers presenting the results of an Italian collaborative study on psychiatric inpatient service utilisation. Patterns of care of a large sample of patients discharged after short inpatient treatment are discussed in the light of the changes introduced by the 1978 Mental Health Act in the Italian psychiatric care delivery system. Three closely related issues are considered: a) use of psychiatric hospitalisation, b) relationships between inpatient and community treatment before and after an admission episode, c) continuity of care. The main findings are: 1) great variability between services suggests that local factors play an important role in determining the contents of care in Italian post-reform psychiatry. 2) The relationship between inpatient and community services is complex, partial integration being the most common picture. 3) Psychiatrichospitalisation is the entrypoint into the care system for a sizeable group of patients.4) Continuity of care is achieved for half the patients, mostly with diagnoses of severe mental disorders. 5) Subjects with a recent history of revolving door behaviour or a past history of mental hospital admission show the highest likelihood of remaining in community care following discharge.
The 1978 Mental Health Act introduced radical changes in the Italian mental health system. The main features of the new legislation, as well as its historical and conceptual background, have already been fully described [1-3]. Only those aspects relevant to this paper will therefore be briefly outlined here. a) No new patients could be admitted to public mental hospitals. Former patients could be readmitted voluntarily until 1981 (in some administrative regions this deadline was extended to 1983). b) Community-based services were to be set up to provide the full range of psychiatric treatments, including residential treatment and hospitalisation, for specified catchment areas. c) Small psychiatric units (Servizi Psichiatrici di Diagnosi e Cura, SPDC), linked with community psychiatric services (Servizi Psichiatrici Territoriali, SPT), were to be es-
tablished in general hospitals, to which all patients requiring both voluntary and compulsory hospitalisation were to be admitted. Over the last decade the Italian psychiatric reform has been a very controversial issue. De Girolamo [4], in a careful review of the available international literature, effectively illustrates the strong polarisation of opinions existing between enthusiastic supporters and vehement detractors. Unfortunately, this heated debate arose in the absence of reliable data on the implementation of the new model and on its effects on the provision of mental health services, not to speak of the more complex question of quality of care. In fact, not only have the substantial innovations not been accompanied by any systematic evaluative research [5, 6], but the shortage of updated national data hinders even crude assessments: to date no national data banks have yet been established to monitor the activities of old and new services. The present study was designed as a step towards filling the gap. It had two main aims: a) to investigate patterns of care and outcome of patients in a large sample of psychiatric patients, and b) to explore the role played by psychiatric hospitalisation within the new model of psychiatric care delivery. SPDCs were chosen as the basis of the investigation, in so far they were set up to serve as the main National Health Service setting for the hospitalisation of psychiatric patients. After 1978, SPDCs were established on a nation-wide basis, albeit unevenly distributed. They are small units (generally comprising at most 15 beds), usually serving well defined catchment areas; the average beds/population ratio is 5.4/100000. In 1984 there were 236 such services, with approximately 3100 beds [7]. Since virtually no information concerning their functioning was available, a large-scale collaborative study was set up in order to obtain a reliable picture. Methods
Thirty-four SPDCs located in eight of the 21 Italian regions (five Northern, one Central and two Southern), took
47 part in the study on a voluntary basis. Their catchment areas comprise 27 % of the adult population of those regions. The SPDCs were not randomly selected; an invitation to participate in the study was issued to those involved in earlier years in other research projects with the coordinating centre. The study started in February 1984; thirty patients consecutively admitted to each SPDC were to be included in the study. Data on all patients were collected during hospital stay by members of the treatment team (usually psychiatrists). The following information was recorded in an ad hoc form: | social and demographic characteristics; 9 psychiatric history; 9 type of admission (voluntary/compulsory, referral source); 9 symptoms observed on admission; 9 problems leading to admission; 9 DSM III diagnosis at discharge; 9 treatment received in the first 4 hours, between the 5th and the 12th hour after admission, and half-way throughout hospital stay; 9 overall clinical judgement at discharge (recovered, improved, unchanged, worsened); | referral for aftercare at discharge. A follow-up was planned 6 and 12 months after discharge. The services were asked to make every effort in order to trace all patients and to gather data concerning: 9 living arrangements; 9 use of aftercare services; 9 number of readmissions and length of time spent as inpatients throughout the follow-up period. Data analysis concerned three main aspects: patterns of care, outcome of care, and clinical management-i, e., pharmacological treatments and type of admission. This paper focuses specifically on the analysis of the patterns of care. In order to investigate the possible associations between the contents of care, outcomes (or interventions), and patients' characteristics, the latter were used to construct three summary variables [8], each scoring from 1 to 3 (the intermediate score value being attributed to subjects not belonging to either of the extreme categories): Social network High score. Subjects living with a partner and/or own children; living alone or with primary family only if employed. L o w score. Unemployed/economically dependent and living alone or with others (non-family members).
Psychiatric history High score. Subjects at their first contact with a psychiatric service. L o w score. Subjects with 3 or more prior admissions in SPDC, or admitted at least once in a mental hospital, whose psychiatric history dated from over one year.
The determinants of referral at discharge were analysed in two stages. In the first place, univariate correlation with prognostic variables was explored; subsequently, in order to estimate the effects of the independent variables considered, a multivariate analysis was carried out by logistic regression [9], using the EPILOG statistical package. It was conducted by steps within each block of exposure variables (patients characteristics: age, sex, social network, social status, psychiatric history, diagnosis, care provider prior to admission; admission- and treatment-related variables: problems leading to admission, type of admission, drug treatments, duration of hospital stay; SPD C of admission) including those appearing in the univariate analysis significantly associated with the dependent variable. The final model was eventually obtained by entering the variables of each block which significantly contributed to explain the outcome. 1143 patients were originally recruited in the study by 36 SPDCs. Two services withdrew from participation in the early stages, and the corresponding 61 patients were therefore excluded from all analyses. Five patients - admitted to two services during the study period - were recruited twice; only the data derived from their first admission were retained for analysis. Patients' characteristics, type of admission, inpatient treatments and referral at discharge were therefore available on 1077 patients recruited in 34 SPDCs. Since some services failed to collect follow-up data at the established intervals, and others had proved unable to trace a substantial proportion of their patients, it was decided - in order to avoid introducing a potentially important bias - to restrict the analyses requiring follow-up data to the subgroup recruited by SPDCs where a) the first and second follow-up had actually taken place 5 to 9 months and 10 to 18 months following discharge respectively, and b) at least 90 % of the patients originally recruited had been traced for follow-up. Patients deceased before the first follow-up were also excluded from analyses on follow-up data. Hence, evaluation of patterns of aftercare was conducted on 569 patients recruited by 21 SPDCs. Results The study population
Social status High score. Educational level beyond primary school qualification (if retired or housewife); skilled workers, white-collars, or in higher order of employment. L o w score. Primary school qualification or less (if retired or housewife); never-employed subjects on welfare benefits.
Social, demographic, diagnostic, and psychiatric history features of the 1077 patients recruited in the study are shown in Tables i and 2. In the analysis DSM III diagnoses have been grouped into broad categories. Data on admission- and treatment-related variables (type of admission, length of stay, drug treatments, referral for aftercare) are shown in Table 3.
48 Table 1. Social and demographic variables % Patients Age: < 25 yrs 25-44 yrs 45~54 yrs > 65 yrs Sex: Males Marital status: Married/cohabiting Social status: Low score Social network: Low score Living arrangement: Own home Residential facility Nursing home Homeless
12.6 44.6 34.4 8.4 53.4
Variability between SPDCs (%) 0-30 28~52 I , 17-50 1 0-27
33.7
7-57*
10.9
0-53*
11.1
0-27"
91.3 1.7 2.8 1.9
73-1000-14 [, 0-13 | 0-13
* P < 0.001 Table 2. Psychiatric history % Patients Variability between SPDCs (%) Diagnosis: Schizophrenic disorders Affective disorders Personality disorders Depr. neurosis/anxiety Substance abuse Problems: Symptom-related Attempted suicide Aggressive behaviour Refusal of treatment Family/housing/work First contact Revolving door ( > 2 admissions SPDC) Prior admission to mental hospital Pre-admission care: no care SPT
43.4 12.9 7.8 13.7 11.0
20-70 -
36.9 9.5 11.2 8.4 34.0 19.7 25.0
7-76 0-28 0-43* 0-33 3-70
35.0
10-60"
42.0 56.0
7-84 12-90"
0-38 0-40
7-45*
* P < 0.001
anxiety and substance abuse disorders (taken together) are unevenly distributed, representing under 25 % of patients in the whole sample, and up to 40 % of the population in SPDCs with fewer patients affected by schizophrenic disorders. Major affective disorders appear more evenly distributed between the services. The patient populations also differ significantly with respect to other variables, such as previous psychiatric history - as indicated, for instance, by the proportion of revolving door patients, or of individuals with a past history of admission to mental hospitals - and the types of problems leading to admission. All these differences, however extreme, cannot alone be held responsible for the remarkable heterogeneity of pre-admission care patterns of the population of patients, plausibly in turn a function of local differences in the availability of community resources, or of the level of integration existing between inpatient and community services. Interestingly, SPDCs show the widest ranges of variation for several key treatment indicators - such as the proportion of compulsory admissions, or of patients treated with high doses of antipsychotic drugs, or referred to community-based services. Length of hospitalisation varies to a lesser degree; a general tendency is apparent (with few exceptions) towards short-term inpatient treatment. Despite this degree of heterogeneity, overall the population described in this study seems closely to match in a number of key features most series of acute psychiatric services users described in the literature [10]. This holds true particularly for clinical severity, length of psychiatric history and degree of social disability. On the other hand, important differences emerge which ought not be overlooked. Irrespective of severity of illness and degree of impairment in social role functioning, the large majority of patients in the study show some social adjustment, and appear surrounded by supportive networks: many subjects live with their families, or (if alone) in their own house, very few are homeless or in sheltered accommodation. Homelessness or residential instability has been documented in large numbers of patients admitted to acute psychiatric inpatient units, particularly in the USA [11, 12]. Moreover, metropolitan areas have been shown to
Table 3. Patterns of inpatient care (CPZ: chlorpromazine) % Patients
Both patients' characteristics and treatment modalities vary widely between SPDCs. The differences are highly statistically significant for all variables, the only exceptions being the sex distribution and the proportion of first-onset cases. Although the majority of patients are 25 to 64 years old, in some services up to one-third of patients are younger (under 25) or older (65 or over). In the extreme age groups most patients were diagnosed as having organic mental disorders (mental retardation or dementia). Schizophrenic disorders form the largest diagnostic group (over 40 %); in some services however their proportion falls to only 20 per cent. Similarly depressive neurosis,
Compulsory admission Length of stay: < 1 week > 1 month Antipsychotic drug use: < 800 mgEq CPZ during hospital stay Referred to SPT at Discharge: * P < 0.001
23.5
Variability between SPDCs (%) 0-96*
28.2 12.3
0-57* 0-45*
8.9
0.53*
71.2
27-90*
49 % 80
60-
40-
20-
Opre-Qdrnission
referred to
no core
6 m t h s FU
~
SPT
~
12 m t h s FU
other agency
Fig.1. Patterns of care: % distribution of patients among care providers
no core pre-odmission
loo in SPT85core pre-odmission 80-
60-
40-
20-
Oreferred tD
6 mths El
12 mths no e~re
~
referred to
s1~r ~
6 mthe
12 mth~
oth~ Qeency
Fig.2. Percentage distribution of pts in SPT care/not in care pre-admission, by care providers
entail a relatively high risk of homelessness [13]. The remarkably low prevalence of homelessness observed in this study, even for patients recruited by services in northern metropolitan areas (Milano, Genova, Torino), suggests that - at least so far - homelessness does not represent a common aftermath of deinstitutionalisation in these areas. Indeed, since none of the services of the two main central-southern metropolitan areas (Rome and Naples) participated in the study this result cannot be generalised to the whole country. However, though the study design calls for some caution in the interpretation of this observation, it would probably be unreasonable to postulate that services' self-selection in Northern cities might have introduced an appreciable bias in this respect.
Patterns o f care
As mentioned earlier, an important aim of the study was to identify and describe prevalent patterns of psychiatric services utilisation. To this end, three closely related aspects have been considered: a) use of psychiatric hospitalisation; b) relationships between inpatient treatment and treat-
ment in the community before and after the admission episode; c) continuity of care. Prior to the index admission, 39 % of patients were not in psychiatric care, and 59 % were treated by SPTs (Fig. 1). At discharge, 72 % of patients are referred for aftercare to SPTs and 21% to other care providers (general practitioners, private psychiatrists, social services or other agencies). At the time of the second follow-up, the proportion of subjects in care of SPTs fell to pre-admission levels, while the proportion of those treated by others remained stable. Of the latter group, 31% of patients were in the care of private psychiatrists, 27 %, of general practitioners, and 20 %, of social services. The picture appears quite different if two subgroups of subjects are examined separately, namely, patients in SPT care and those not in care prior to admission (Fig. 2). Almost all patients previously treated by SPTs were referred to SPTs at discharge, and by and large they were still in treatment 12 months later. On the contrary, only 56% of previously untreated patients were referred to SPTs at discharge, while one third were referred to other agencies. Moreover the proportion of previously untreated subjects in SPT care steadily declined during the follow-up period, more markedly than that of subjects in care of other agencies. In other words, past history of treatment by community psychiatric services is an important predictor of further treatment by SPTs after discharge. Once more, the degree of heterogeneity among SPDCs is considerable and several instances of departure from the general pattern can be identified. The proportion of patients in SPT care prior to admission ranged from 12 to 90 %; accordingly, the proportion of previously untreated subjects varied between 7 and 84 per cent. One year after discharge, however, the differences between services appear somewhat narrowed: the proportions of patients in SPT care range from 34 to 86 % and those of patients not in care from 0 to 43 per cent. At the time of the second follow-up, general practitioners were by far the most important care providers for patients discharged by one particular service. In addition, diagnosis contributes significantly to the prediction of patterns of aftercare (Table 4). Patients with severe mental disorders - such as schizophrenic or major affective disorders - were more often referred to SPTs, and were more likely to remain in SPT care during the follow-up period. Subjects with less severe disorders followed a quite different pattern. One year after discharge one third were in SPT care, one third were treated by other agencies, and one third were not in care. Patients presenting substance abuse disorders were least likely of all to use any aftercare service. Within six months following discharge, 42% of the patients were readmitted for psychiatric treatment (33 % to SPDCs, 5 % to private hospitals and 4 % to other inpatient care settings). Low social network scores and previous psychiatric history are significantly associated with readmission. A detailed analysis of readmissions and of other outcome indicators will be presented in the second paper of this series.
50 Table 4. Percentage distribution of patients by diagnostic groups and type of care providers at 6 and 12 months Diagnostic groups
Schizophrenic disorders Major affective disorders Depressive neurosis, anxiety Substance abuse disorders
Care providers SPT Other agency 6 12 6 12 Mths Mths Mths Mths 75 61 16 17
Not in care 6 12 M t h s Mths 10 11
67
64
24
26
9
10
47
42
36
30
16
27
40
37
25
26
35
37
Continuity of care According to Bachrach [14], continuity of care can be defined as a process involving the uninterrupted movement of patients over time through the diverse elements of the service delivery system. For several years the continuity of care model has gained wide acceptance as the best answer to the problems of service delivery in the community for the severe, long-term mentally ill. Continuity of care within public mental health services has been evaluated in this study by examining the pathways taken by patients through the care delivery system. To this end the interaction between patients and services at three points in tinre (prior to admission, referral at discharge, and 6 months later) was assessed. At each point three alternatives were explored, i.e., at discharge: no referral, referral to SPT and referral to other care providers, and in the other two occasions: no care, SPT care and care provided by other agencies. Continuity of care was deemed to have been achieved for patients not in care, or in SPT treatment prior to admission, who had been referred to SPT at discharge and were still SPT users after six months. 51% of the patients experienced continuity of care thus defined. Two different explanations could apply to the lack of continuity of care experienced by the remaining 49 % : inadequate integration between inpatient and community services, or inappropriate hospitalisations. The latter seems particularly plausible for patients diagnosed as having anxiety or substance abuse disorders, often discharged within a week, who would probably benefit more from interventions other than hospitalisation. Discontinuation of care is most likely soon after discharge, due to non-referral or to poor coordination between hospital and community providers. Relatively few patients among those in SPT care six months after discharge appear subsequently to abandon treatment.
referral for aftercare is an important indicator of the degree of integration between hospital and community services, critically relevant to the continuity of care. The potential risk factors for non-referral to SPTs after discharge - i.e. the variables positively associated with this choice on the part of the hospital team - were therefore explored in depth. The variables jointly comprising statistically significant independent determinants of outcome (non-referral to SPTs) in the multivariate analysis are listed in the logistic regression model reported in Table 5, along with the relevant relative risks (RR), confidence limits (CL) and P-values. Among patients characteristics, age > 65, a diagnosis of substance abuse disorders or of "other" disorders (mostly mental retardation and dementia), and a low social network score appear to contribute significantly to non-referral. Taken together, these factors probably indicate to a large extent a policy of referral of elderly patients, lacking family support, to services other than SPTs (mainly social services). Interestingly, a past history of SPT care is the strongest predictor of referral to SPT, while diagnosis - though significantly contributing to the model - loses most of its explanatory power after adjusting for the remaining independent variables. Among the problems leading to admission, the best predictor of referral to SPT is treatment refusal. Of the second block of variables, only duration of hospital stay is significantly associated with the referral decision. Patients hospitalised for 1 to 4 weeks were those most likely to be referred to SPTs at discharge. This finding could be partly explained by the fact that the majority Table 5. Determinants of non-referral to SPT (logistic regression model; log-likelihood: - 488.72;Z2:286.43,49 df) RR b
P
0.16-0.33 0.43-3.93
0.000 0.644
1.55-4.83
0.000
1.03-2.10 1.34-3.89
0.003 0.002
0.34-0.74 0.48-1.38
0.000 0.450
0.45-1.39 0.41-1.31 0.13-0.61 0.48-1.10
0.420 0.288 0.001 0.128
0.87-2.56 0.69-2.43 1.31-3.76 0.67-1.94 1.40-4.21
0.141 0.414 0.003 0.626 0.002
Multivariate analysis. Since the transition from the hospi-
No care (pre-admission) In SPT care Other agencies Age < 65 65 + Social network High Medium Low Hospital stay (wks) < 1 1- < 4 > =4 Problems Symptom-related Attempted suicide Aggressive behaviour Treatment refusal Work/housing/family Diagnosis Schizophrenic d. Affective disorders Personality disorders Substance abuse d. Depressive neurosis Other
tal to the community after an admission episode is a key aspect of the long-term treatment of mental disorders,
a Reference categories; b Estimates from LogReg equations including term for SPDCs.
Determinants of referral at discharge
a 0.23 1.23 a 2.74 a 1.47 2.29 a 0.50 0.82
CL
0.79 0.73 0.28 0.72 a
1.50 1.30 2.22 1.14 2.42
51 of patients with minor mental disorders, often not referred to SPTs, constitute a large proportion of very short hospitalisations. Finally, the service itself appears to play a critical role in determining non-referral after adjustment for the other variables; for eleven SDPCs the relative risks significantly depart from unity.
Discussion
The study was undertaken on the assumption that the analysis of the patterns of care received by a large, unselected sample of psychiatric patients experiencing an admission episode could contribute relevant information on psychiatric care delivery services in Italy after the changes introduced in 1978. The model underlying the Mental Health Act clearly entails continuity of care, local availablility of the full range of treatments, long-term support for patients and their families, and integration among services providing different components of care. The existence of a wide consensus around these aims is also evinced in the international literature [15]. Recent studies support the view that integration between hospital and community services is a viable, economically convenient solution to improving delivery of care by mental health services, and have led to suggestions that hospitalisation should only exceptionally represent the entry point in the treatment system-ideally, patients should be given inpatient care only as a last resort after the failure of community treatment [16]. To what extent do the patterns of care described above reflect the underlying model? The repeatedly stressed variability between services indicates that local factors whether related to the social or geographical characteristics of the catchment areas, or to different organisational and clinical models, or to the availability of community resources, or to any combination of the above are very likely to play a major role in determining the contents of care. Furthermore, there is currently no widespread consensus on the role and specific functions attributed to SPDCs; nor has an agreement been reached concerning the categories of patients pertaining to SPDCs, or indeed to psychiatry. The different cultural determinants of the Italian post-reform psychiatry (as has generally happened in post-deinstitutionalisation psychiatry worldwide) leave important "grey areas" to the discretion of individual services, which in practice are left to define for themselves the boundaries of their assignment (typical examples are alcoholics, drug abusers, vagrants, or homeless people). The decision of being or not involved in the care of a substantial fraction of mentally disturbed patients therefore is not necessarily (or not only) a result of scarce financial and human resources; often it reflects explicit choices of individual services. Further research is needed to clarify this point, possibly with the aim of assessing the impact of service characteristics on patients' outcome. Indeed, the heterogeneity observed strengthens previous observations on the different ways in which the legislation was implemented [17]. In this respect, further efforts to collect broad administrative
data at a national or regional level seem to be of limited value. The general picture resulting from this study can be summarised as follows: 1. The relationships existing between SPDCs and SPTs are complex, and cannot be explained according to one model. In some instances full integration is achieved; more commonly features indicating partial integration are observed. On the whole, integration is more likely to materialise for patients with severe mental disorders and a previous history of psychiatric treatment. Although limited availability of SPTs in some areas might be an obvious explanation of the discontinuity of care following discharge, it cannot be held to represent the only cause, particularly since a significant association with non-referral has been shown for SPDCs located in areas where community services are available and adequately equipped. 2. SPDCs are the entry point into the care system for a sizeable group of subjects; often they appear to relate to community services other than SPTs, particularly general practice and social services. A large proportion of patients with minor mental disorders are referred to GPs, and there are evidences indicating that such patients are more easily kept in treatment by the latter than by SPTs. This calls for a thorough exploration of the role of general practitioners in the care of a broad range of minor mental disorders, a field so far virtually unexplored in Italy. 3. SPTs represent with few exceptions the only providers far chronic or clinically severe patients, for 50 % of whom continuity of care was achieved. The private sector has a very limited function in the care of these patients, as opposed to the results of observations carried out in other European countries [18]. The subjects presenting a recent history of revolving door behaviour or a past history of mental hospital admissions show the greatest likelihood of remaining in SPT care in the follow-up period; in this respect the Italian model seems again to be at odds with the pattern resulting from international experience [12]. As expected, the inherent character of the Italian organisation and structure of mental health care services determines unique care delivery patterns - to an extent which hinders any exercise of international comparison. Whether and how this peculiar structure also determines different outcomes will form the object of a subsequent publication.
References
1. Mosher LR (1982) Italy's revolutionary mental health law:an assessment. Am J Psychiatry 139:199-203 2. Tansella M, WilliamsP (1987) The Italian experience and its implications. Psychol Med 17:283-289 3. Saraceno B, Tognoni G (1989) Methodological lessons from the Italian psychiatricexperience. Int J Soc Psychiatry 35:98-109 4. De Girolamo G (1989) Italian psychiatry and reform law: a review of the international literature. Int J Soc Psychiatry 35:21-37 5. Morosini PL, Repetto F, De Salvia D, Cecere F (t985) Psychiatric hospitalisation in Italy before and after 1978. Acta Psychiatr Scand 71 [Suppl 316]: 15-26
52 6. Bollini E Mollica RF (1989) Surviving without the asylum: an overview of the studies on the Italian reform movement. J Nerv Ment Dis 177:607~515 7. Crepet P (1990) A transition period in psychiatric care in Italy ten years after the reform. Br J Psychiatry 156:27-36 8. Avison WR, Speechley KN (1987) The discharged psychiatric patient: a review of social, social-psychological, and psychiatric correlates of outcome. Am J Psychiatry 144:10-18 9. Kleinbaum DG, Kupper LL, Morgenstern H (1982) Epidemiologic research: principles and quantitative methods. Van Nostrand, New York 10. Bassuk E, Winter R, Apsler R (1983) Cross-cultural comparison of British and American psychiatric emergencies. Am J Psychiatry 140:180-184 11. Chafetz L, Goldfinger SM (1984) Residential instability in a psychiatric emergency setting. Am J Orthopsychiatry 56:20-34 12. Surles RC, McGurrin MC (1987) Increased use of psychiatric emergency services by young chronic mentally ill patients. Hosp Comm Psychiatry 38:401-405 13. Bachrach L (1987) The homeless mentally ill. In: Menninger W, Hannah G (eds) The chronic mental patient-II. American Psychiatry Press, Washington, pp 65-92
14. Bachrach L (1981) Continuity of care for chronic mental patients: a conceptual analysis. Am J Psychiatry 138:1449-1556 15. Fuller Torrey EF (1986) Continuous treatment teams in the care of the chronic mentally ill. Hosp Comm Psychiatry 37:1243-1247 16. Tyrer R Turner R, Johnson AL (1989) Integrated hospital and community psychiatric services and use of inpatient beds. BMJ 299:298-300 17. Tognoni G, Saraceno B (1989) Regional analysis and implementation. Int J Soc Psychiatry 35:38-45 18. Gaebel W, Pietzcker A (1987) Prospective study of the course of illness in schizophrenia: III. Treatment and outcome. Schizophr Bull 13:307-316
Dr. A. Barbato Centro Psicosociale Unit~ Operativa di Psichiatria Policlinico II, USSL 75/4 Via Ravenna 13 1-20139 Milano Italy
Announcement XIIth International Conference on the Social Sciences and Medicine 14-18 September 1992, Peebles, UK F o r f u r t h e r i n f o r m a t i o n , p l e a s e contact: Dr. P. J. M. M c E w a n , G l e n g a r d e n , Ballater, A b e r d e e n s h i r e A B 3 5 5 U B , S c o t l a n d , U K
