BRITISH JOURNAL OF PSYCI-IIATRY (2004), 185, 203-290 REVIEW ARTICLE

Components of a modern service: and international organisations and institu- tions. It comprises two services: answers to a pragmatic balance of community questions to support the decision-making process, and ready access to sources of evidence such as databases, documents and hospital care and networks of experts. Overview of systematic evidence METHOD GRAHAM THORNICROFT and MICHELE TANSELLA This paper focuses upon the following key questions: (a) How far should mental Background There is controversy The public health impact of mental dis- health services be provided in community about whether mental health services orders is profound (Murray & Lopez, and/or hospital settings? (b) What service should be provided in community or 1996; World Health Organization, 2001a). components are necessary and which are The estimated disability-adjusted life-years optional? (c) What are the differing service hospital settings.There is no worldwide in 2000 attributable to mental disorders development priorities for areas (countries consensus on which mental health service represents 11.6% of total disability in the and regions) with low, medium and high models are appropriate in low-, medium- world - more than double the level of dis- levels of resources? and high-resource areas. ability caused by all forms of cancer The recent growth of mental health (5.3%) and higher than the level of disabil- services research has provided substantial Aims To provide an evidence base for ity due to cardiovascular disease (10.3%). evidence in relation to these questions, but this debate, and present a stepped care Historically, the response of the mental few attempts have been made to review health services can be seen in three periods: these results as a whole and to put them model. the rise of the asylum, the decline of the in a resource context so that they are usable Method Cochrane systematic reviews asylum and the reform of mental health for the planning and provision of services at services (Wing & Brown, 1970; Grob, national and regional levels. The aim of this and other reviews were summarised, 1991; Desjarlais et al, 1995; Thornicroft review is therefore to summarise such evi- Results The evidence supports a & Tansella, 1999). In the third period, dence, and to propose a stepped care model community-based and hospital-based services that contextualises the relevance of this balanced approach, including both comm- commonly aim to provide treatment and evidence to areas at different stages of eco- unity and hospital services. Areas with low care that are close to home, including acute nomic development. It refers to mental levels of resources may focus on improving hospital-care and long-term residential health services for adults of working age, primary care, with specialist back-up. facilities in the community; respond to dis- and does not directly address other import- abilities as well as to symptoms; are able to ant groups, such as children, older people Areas with medium resources may offer treatment and care specific to the diag- or those whose primary problem is drug additionally provide out-patient clinics, nosis and needs of each individual; are con- or alcohol misuse. We appreciate, however, community mental health teams (CMHTs), sistent with international conventions on that for regions with fewer resources, acute in-patient care, community human rights; are related to the priorities where the majority of service provision is residential care and forms of employment of service users themselves; are coordinated at the primary care level, these distinctions between mental health professions and may be less relevant. and occupation. High-resource areas may agencies; and are mobile rather than static. The procedure used was that first we provide all the above, together with more We have described this as the `balanced care' searched Medline for the period 1980 to specialised services such as specialised out- approach (Thornicroft & Tansella, 2002). April2003, using the search terms MENTAL patient clinics and CMHTs, assertive This paper summarises and extends a and COMMUNITY and HOSPITAL (3177 review prepared for the Health Evidence community treatment teams, early records were extracted). Only English- Network of the World Health Organization language articles were examined to include intervention teams, alternatives to acute European Regional Office (WHO-EURO) those relevant journals with higher impact in-patient care, alternative types of (Thornicroft & Tansella, 2003). The factors (1810 records); of these 141 were community residential care and alternative Health Evidence Network is an information review articles, which were considered in occupation and rehabilitation. service initiated and coordinated by WHO- preparing this paper. In addition, the authors EURO which provides the best evidence searched the Cochrane Library and included Conclusions Both community and available in the field of public health other relevant systematic reviews. This pro- hospital services are necessary in all areas (http://www.who.dk/hen) . Working with cedure allowed us to summarise the evidence over 30 partner organisations, it aims to for distinct service components, and to regardless oftheir level of resources, deliver timely information to recommend three particular blends of these according to the additive and sequential decision-makers in the WHO European components as suitable for areas with low, stepped care model described here. Region by providing summaries from a medium and high level of resources, as a wide range of existing sources, including contribution to the debate about resource- Declaration of interest None. websites, databases, documents, national appropriate models of care.

283 THORNICROFT & TANSELLA

RESULTS in fact represents widespread neglect of Vazquez-Barquero, 2001), whereas in the mentally ill people. Where asylums do African continent 80% of countries spend The results of this review are organised in exist, policy makers face choices about less than 1% of their limited total health relation to the level of resources available, whether to upgrade the quality of care budget on mental health. These and other as proposed by the WI-10 World Health offered (Njenga, 2002) or to use the relevant comparative data are available Report (World Health Organization, resources of the larger hospitals to estab- from the WHO Project Atlas website 2001a: pp. 112-115). Table 1 indicates that lish decentralised services instead (Alem, (World Health Organization, 2001b) and areas with a low level of resources are likely 2002). from the World Bank (2002). For example, to need to provide most or all of their men- Differences in mental health services although health spending represents some tal health care in primary health care set- between low-resource and high-resource 7.9% of global gross domestic product, tings, delivered by primary care staff, with countries are vast. In Europe, for example, with an average expenditure expressed in specialist back-up to provide training, con- there are 5.5-20.0 psychiatrists per international dollars (based on purchasing sultation for complex cases, and in-patient 100 000 population, whereas the figure is power parities) of I$523 on health assessment and treatment of cases that 0.05 per 100 000 in African countries services, this average varies significantly cannot be managed in primary care (Njenga, 2002); the average number of psy- across countries and regions, ranging from (Mubbashar, 1999; Saxena & Maulik, chiatric beds is 8.70 in the European region I$82 per person in Africa to I$2078 in the 2003). Some low-resource countries may and 0.34 in Africa (Alem, 2002). About Organization for Economic Cooperation in fact be in a pre-asylum stage (Njenga, 5-10% of the total health budget is spent and Development (OECD) countries 2002) in which apparent community care on mental health in Europe (Becker & (Poullier et al, 2002). Further, for both

Table I Mental health service components relevant for countries and regions with low, medium and high levels of resources

Low level of resources Medium level of High level of resources resources

Step A Step A + step B Step A + step B + step C Step A: Primary care with specialist back-up Step B: Mainstream mental health care Step C: Specialised/differentiated mental health services Screening and assessment by primary Out-patient/ambulatory clinics Specialised clinics for specific disorders or patient groups, including: care staff e eating disorders Talking treatments, including counselling e dual diagnosis and advice e treatment-resistant affective disorders Pharmacological treatment e adolescent services Liaison and training with mental health specialist staff, when available Community mental health teams Specialised community mental health teams, including: Limited specialist back-up available for: e early intervention teams e training e assertive community treatment e consultation for complex cases • in-patient assessment and treatment for Acute in-patient care Alternatives to acute hospital admission, including: cases that cannot be managed in primary • home treatment/crisis resolution teams care, for example in general hospitals • crisis/respite houses • acute day hospital

Long-term community-based Alternative types of long-stay community residential care, residential care including: • intensive 24 h staffed residential provision • less intensively staffed accommodation e independent accommodation

Employment and occupation Alternative forms of occupation and vocational rehabilitation: e sheltered workshops e supervised work placements • cooperative work schemes e self-help and user groups • club houses/transitional employment programmes e vocational rehabilitation • individual placement and support service

204 COMMUNITY AND/OR HOSPITAL CARE

Europe and Africa there are also consider- treatments within primary health care in (d) whether direct or indirect payment is able and often growing variations both these countries is feasible (World Health made; between countries and between regions Organization, 2001a). (e) methods used to enhance attendance within countries, not only in health expen- rates; diture but also in social care. As a conse- Step B: Mainstream mental quence the forms of service provision health care (f) how the clinic responds to non- relevant to low-resource areas will be Mainstream mental health care refers to a attenders; very different from those relevant to range of service components, which may (g) the frequency and duration of clinical medium- and high-resource areas. be necessary in areas that can afford more contacts. Areas (countries or regions) with a than a primary care-based system with spe- There is surprisingly little evidence on medium level of resources may first estab- cialist back-up. However, the recognition any of these key characteristics of out- lish the service components shown in and treatment of the majority of people patient care (Becker, 2001), but there is a column 2 of Table 1, and later, as resources with mental illnesses, especially depression strong clinical consensus in many countries allow, choose to add some of the wider and anxiety-related disorders, remains a that such clinics are a relatively efficient range of more differentiated services indi- task that falls mostly to primary care. Von way of organising the provision of assess- cated in column 3. The choice of which of Korff & Goldberg (2001) reviewed 12 dif- ment and treatment, provided that the these more specialised services to develop ferent randomised controlled trials of clinic sites are accessible to local popula- first depends upon local factors, including enhanced care for major depression in tions. Nevertheless, these clinics are simply service traditions and specific circum- primary care settings. They found that methods of arranging clinical contact stances; consumer, carer and staff prefer- interventions directed solely towards train- between staff and patients, and so the key ences; existing service strengths and ing and supporting general practitioners issue is the content of the clinical inter- weaknesses; and the way in which evidence have not been shown to be effective. They ventions: namely, to deliver treatments that is interpreted and used. This stepped care argued that interventions should focus on are known to be evidence-based (Roth & model also indicates that the forms of care low-cost case management, coupled with Fonagy, 1996; Nathan & Gorman, 2002; relevant and affordable in areas with a high flexible and accessible working relation- BMJ Publishing Group, 2003). level of resources will include elements ships between the case manager, the pri- from column 3, in addition to the compo- mary care doctor and the mental health nents in columns 1 and 2 which will usually specialist. In other words, the whole pro- Community mental health teams (CMHTs) already be present. The model is therefore cess of care needs to be enhanced and reor- Community mental health teams are the both additive and sequential, in that new ganised to include the following key basic building block for community mental resources allow extra levels of service to elements: active follow-up by the case health services. The simplest model of pro- be provided over time, in terms of mixtures manager, monitoring treatment adherence vision of community care is for generic of the components within each step, when and patient outcomes, adjustment of treat- (non-specialised) teams to provide the full the provision of the components in the ment plan if patients do not improve, and range of interventions (including the contri- previous step is complete. referral to a specialist when necessary butions of psychiatrists, community psychi- Decisions on the planning and invest- (Von Korff & Goldberg, 2001). This could atric nurses, social workers, psychologists ment of funds to improve mental health be seen as a major reversal of what is con- and occupational therapists), prioritising will need to include a wide range of stake- sidered by many to be the conventional adults with severe mental illness, for a local holders, often bringing divergent or even approach: enhancing the training of family defined geographical catchment area conflicting perspectives to this task. It is doctors. Rather, the evidence now strongly (Thornicroft at al, 1999; Department of now increasingly common in many coun- suggests that improving outcomes of Health, 2002). A series of studies and sys- tries for service users and family members chronic diseases such as depression does tematic reviews, comparing community or carers to participate routinely in such appear to require more than changing the mental health teams with a variety of local decision-making. skills of one profession alone: namely, the usual services, suggests that there are clear combination of several concurrent active benefits to the introduction of generic, Step A: Primary care ingredients. community-based multidisciplinary teams: mental health with specialist Mainstream mental health care can be they can improve engagement with services, back-up considered to be an amalgam of the core increase user satisfaction, increase met Well-defined psychological problems are components described below. needs and improve adherence to treatment, common in general health care and primary although they do not improve symptoms or health care settings in every country, and Out-patient and ambulatory clinics social function (Tyrer et al, 1995, 1998, cause disability which is usually in propor- Out-patient and ambulatory clinics vary 2003; Thornicroft at al, 1998; Burns, tion to the number of symptoms present according to: 2001; Simmonds at al, 2001). In addition, (Ormel at al, 1994). In areas with a low (a) whether patients can self-refer, or need continuity of care and service flexibility level of resources (Table 1, column 1), the to be referred by other agencies such have been shown to be more developed large majority of cases of as primary care; where a community mental health team should be recognised and treated within (b) there are fixed appointment times or model is in place (Sytema at al, 1997). primary health care (Desjarlais at al, open access assessments; 1995). The WHO has shown that the (c) doctors alone or other disciplines also Case management. Within community integration of essential mental health provide clinical contact; mental health teams, case management is

28S THORNICROFT & TANSELLA

a method of delivering care, rather than health budget (Knapp et al, 1997). There- residential long-term care that will be being a clinical intervention in its own fore, minimising the number of bed-days needed in any particular area is also highly right, and at this stage the evidence suggests used, for example by reducing the average dependent upon which other services are that it can most usefully be implemented length of stay, may be an important goal, available locally, and upon social and cul- within the context of the community men- if the resources released in this way can be tural factors, such as the amount of family tal health team (Holloway & Carson, used for other service components. A care that is provided (van Wijngaarden 2001). It is a style of working that has been related policy issue concerns how to pro- et al, 2003). described as the `coordination, integration vide acute beds in a humane and less insti- and allocation of individualised care within tutionalised way that is acceptable to Employment and occupation limited resources' (Thornicroft, 1991). patients, for example in general hospital Rates of unemployment among people with There is now a considerable literature to units (Quirk & Lelliott, 2001; Tomov, mental disorders are usually much higher show that this style of working can be mod- 2001). than in the general population (Warr, erately effective in improving continuity of 1987; Warner, 1994). Traditional methods care, quality of life and patient satisfaction, Long-term community-based residential care of occupation and day care have been but there is conflicting evidence as to It is important to know whether patients provided by day centres or a variety of psy- whether it has any impact on the use of with severe and long-term disabilities chiatric rehabilitation centres (Shepherd, in-patient services (Saarento et al, 1996; should be cared for in larger, traditional 1990; Rosen & Barfoot, 2001). There has I-Iansson et al, 1998; Mueser et al, 1998; institutions, or be transferred to long-term been little scientific research into these Ziguras & Stuart, 2000; Ziguras at al, community-based residential care. The traditional forms of day care, and a review 2002). Case management needs to be evidence here, for areas with medium and of over 300 papers found no relevant carefully distinguished from the much high resource levels, is clear. When deinsti- randomised controlled trial (Marshall at al, more specific and more intensive assertive tutionalisation is done carefully for those 2001). Non-randomised studies have given community treatment (see below). who had previously received long-term conflicting results, and for areas with med- in-patient care for many years, the out- ium levels of resources it is reasonable at Acute in-patient care comes are more favourable for most this stage to make pragmatic decisions patients who are discharged to community about the provision of rehabilitation and There is no evidence that a balanced system care (Tansella, 1986; Thornicroft & day care services if the more differentiated of mental health care can be provided with- Bebbington, 1989; Shepherd & Murray, and evidence-based options discussed out acute beds. Some services (such as 2001). The Team for the Assessment of below are not affordable (Marshall et al, home treatment teams, crisis houses and Psychiatric Services study in London (Leff, 2001; Catty at al, 2003). acute day hospital care, see below) may 1997), for example, completed a 5-year be able to offer realistic alternative care follow-up of over 95% of 670 people with- Step C: Specialised for some voluntary patients. Nevertheless, out dementia discharged from long-stay and differentiated mental health people who need urgent medical assess- residential care and found that: services ment, or those with severe and comorbid medical and psychiatric conditions, or (a) two-thirds of the patients were still The stepped care model suggests that areas those experiencing severe psychiatric re- living in their new residence; with a high level of resources may already lapse and behavioural disturbance, or those (b) there was no increase in the death rate provide all or most of the service compo- with high levels of suicidality or assaulta- or the suicide rate; nents in steps A and B, and are then able to offer additional components from the tiveness, or with an acute neuropsychiatric (c) very few patients became homeless, and following options (step C; Table 1). condition, or elderly patients with concomi- none was lost to follow-up from a tant severe physical disorders, will usually staffed home; require high-intensity immediate support Specialised out-patient and ambulatory (d) over a third were briefly readmitted, in acute in-patient hospital units. clinics and at follow-up 10% of the sample There is a relatively weak evidence base were in hospital; Specialised out-patient facilities for specific on many aspects of in-patient care, and disorders or patient groups are common in (e) patients' quality of life was greatly most studies are descriptive accounts many high-resource areas and may include improved by the move to the (Szmukler & Holloway, 2001). There are services dedicated, for example, to those community; few systematic reviews in this field, one of with eating disorders; patients with dual which found no difference in outcomes (f) there was little difference between total diagnosis (psychotic disorder and substance between routine admissions and planned hospital and community costs, and misuse); people with treatment-resistant short hospital stays (Johnstone & Zolese, overall community care was more cost- affective or psychotic disorders; those re- 1999). More generally, although there is a effective than long-stay hospital care. quiring specialised forms of psychotherapy; consensus that acute in-patient services are However, there is less evidence available on mentally disordered offenders; mentally ill necessary, the number of beds required is the treatment and care needs of the never- women with babies; and those with other highly contingent upon what other services institutionalised group of long-term specific disorder groups (such as post- exist locally and upon local social and patients (Holloway et al, 1999), and so traumatic stress disorder). Local decisions cultural characteristics (Thornicroft & careful local assessment of the needs of this about whether to establish such specialist Tansella, 1999). Acute in-patient care population will be especially important. clinics will depend upon several factors, commonly absorbs most of the mental The range and capacity of community including their relative priority in relation

286 COMMUNITY AND/OR HOSPITAL CARE

to the other specialist services described psychosis'; other studies have placed more A wide variety of respite houses, havens below, identified services gaps and the emphasis upon providing family interven- and refuges have been developed, but the financial opportunities available. tions when a young person's psychosis is term `crisis house' is used here to mean first identified (Addington et al, 2003; facilities that are alternatives to non- Specialised community mental health teams Raune et al, 2004). There is now emerging compulsory hospital admission. The little Specialised community mental health teams evidence that longer duration of untreated available research evidence suggests that are by far the most researched of all the psychosis is a predictor of worse outcome they are very acceptable to their residents components of balanced care, and most for the disorder; in other words, if patients (Davies et al, 1994; Sledge et al, 1996a,b; recent randomised controlled trials and sys- wait a long time after developing a psy- Szmukler & Holloway, 2001), may be able tematic reviews in this field refer to such chotic condition before they receive treat- to offer an alternative to hospital admission teams (Mueser et al, 1998). Two types of ment, then they may take longer to for about a quarter of those who would specialised community mental health team recover and have a less favourable long- otherwise be admitted, and may be more have been particularly well developed as term prognosis. Few controlled trials of cost-effective than hospital admission adjuncts to generic teams: assertive such interventions have been published, (Sledge et al, 1996a,b; Mosher, 1999). community treatment teams and early and a recent Cochrane systematic review Nevertheless, there is emerging evidence intervention teams. (Marshall & Lockwood, 2004) has con- that female patients in particular prefer cluded that there are `insufficient trials to non-hospital alternatives (such as crisis Assertive community treatment teams. Asser- draw any definitive conclusions, .. . the sub- houses) to acute in-patient treatment, and tive community treatment teams provide a stantial international interest in early inter- this may reflect the lack of perceived safety form of specialised mobile outreach treat- vention offers an opportunity to make in hospital (Killaspy et al, 2000). ment for people with more disabling mental major positive changes in psychiatric prac- disorders, and have been clearly charac- tice, but this opportunity may be missed Home treatment and crisis resolution teams. terised (Deci et al, 1995; Teague et al, without a concerted international pro- Home treatment and crisis resolution teams 1998; Scott & Lehman, 2001). There is gramme of research to address key unan- are mobile community mental health teams now strong evidence that assertive com- swered questions'. It is therefore currently offering assessment for patients in psychi- munity treatment can produce the follow- premature to judge whether specialised atric crises and providing intensive treat- ing advantages in areas with high levels of early intervention teams should be seen as ment and care at home. A Cochrane resources: a priority (Larsen et al, 2001; McGorry systematic review (Catty et al, 2002) found (a) reduced admissions to hospital and use & Killackey, 2002; McGorry et al, 2002; that most of the research evidence comes of acute beds; Warner & McGorry, 2002; Friis et al, from the USA and the UK, and concluded 2003; Harrigan et al, 2003). that home treatment teams reduce days (b) improved accommodation status and spent in hospital, especially if the teams occupation; make regular home visits and have respon- Alternatives to acute in-patient care (c) increased service user satisfaction. sibility for both health and social care (Joy In recent years three main alternatives to et al, 2002). Assertive community treatment has not acute in-patient care have been developed: been shown to produce improvements in acute day hospitals, crisis houses and home Alternative types of long-stay community mental state or social behaviour. It can treatment/crisis resolution teams. residential care reduce the cost of in-patient services, but These are usually replacements for long- does not change the overall costs of care Acute day hospitals. Acute day hospitals stay wards in psychiatric institutions (Latimer, 1999; Phillips et al, 2001; offer programmes of day treatment for (Shepherd et al, 1996; Trieman Marshall & Lockwood, 2003). Neverthe- et al, those with acute and severe psychiatric 1998; Shepherd & Murray, 2001). Three less, it is not known how far this approach problems, as an alternative to admission categories of such residential care can be is cross-culturally relevant and indeed there to in-patient units. A recent systematic identified: is evidence that it may be less effective review of nine randomised controlled trials where usual services already offer high (a) 24 h staffed residential care (high- has established that acute day hospital care levels of continuity of care, for example in staffed hostels, residential care homes is suitable for about 30% of people who or nursing homes, depending on the UK, than in settings where the 'treat- would otherwise be admitted to hospital, whether the staff have professional ment as usual' control condition may offer and offers advantages in terms of faster qualifications); little to patients with severe mental illness improvement and lower cost. It is reason- (b) day-staffed residential places (hostels or (Burns et al, 1999, 2001; Fiander et al, able to conclude that acute day hospital residential homes which are staffed 2003). care is an effective option when demand during the day); Early intervention teams. There has been for in-patient beds is high (Wiersma et al, (c) lower supported accommodation (mini- considerable interest in recent years in the 1995; Marshall et al, 2001). mally supported hostels or residential prompt identification and treatment of homes with visiting staff). first- or early-episode cases of psychosis. Crisis houses. Crisis houses are houses in There is limited evidence as to the cost- Much of this research has focused upon community settings which are staffed by effectiveness of these types of residential the time between the first clear onset of trained mental health professionals and care, and no completed systematic review symptoms and the beginning of treatment, offer admission for some patients who (Chilvers et al, 2003). It is therefore reason- referred to as the `duration of untreated would otherwise be admitted to hospital. able for policy makers to decide upon the

287 THORNICROFT & TANSELLA

need for such services with local stake- mental health care, where the main role planning mechanisms may be weak in some holders (Hafner, 1987; Nordentoft et al, for the relatively few specialist mental areas. Second, this model implies that the 1992; Rosen & Barfoot, 2001; Thornicroft, health staff is to support primary care staff training of mental health staff should be 2001). (step A, column 1, Table 1). Areas that can fit for purpose according to the service afford a more differentiated model of care stage reached (A, B or C) and the level of Alternative forms of employment may first consolidate their mainstream resources in the area of practice (high, and occupation mental health care (step B), with the capa- medium or low). In practice it is likely that city of each service component decided as in any particular area some but not all of Although vocational rehabilitation has a balance between the known local needs the service components described here will been offered in various forms to people (Thornicroft, 2001), the resources available be present, and that such identified gaps with severe mental illness for over a and the priorities of local stakeholders. In may inform local planning for service century, its role has weakened because of general, as mental health systems develop developments. discouraging results, financial disincentives away from an asylum-based model, the In recent years there has been a debate to work and pessimism about outcomes proportion of the total budget spent on between those who are in favour of the pro- for these patients (Lehman et al, 1995; the large asylums gradually decreases. In vision of mental health treatment and care Polak & Warner, 1996; Wiersma et al, other words, new services outside hospital in hospitals, and those who prefer to use 1997). However, recent alternative forms can only be provided by using extra re- primarily or even exclusively community of occupation and vocational rehabilitation sources (which is uncommon) or by using settings, in which the two forms of care have again raised employment as an out- the resources that are transferred from the are often seen as incompatible. This false come priority. Consumer and carer advo- hospital sites and staff (which is the more dichotomy can now be replaced by an cacy groups have set work and occupation usual case). Interestingly, the evidence from approach that balances both community as one of their highest priorities, to enhance cost-effectiveness studies of deinstitutiona- services and modern hospital care. How- both functional status and quality of life lisation and the provision of community ever, since this framework cannot be (Becker et al, 1996; Thornicroft et al, mental health teams is that the quality of applied in the same way in settings with 2002). There are recent indications that care is closely related to the expenditure different resources, the stepped care model it is possible to improve vocational and upon services, and overall community- presented in this paper suggests a sequential psychosocial outcomes with supported based models of care are largely equivalent view of how to develop a balance of ser- employment models, which emphasise in cost to the services that they replace. vices in any specific context, moving over rapid placement in competitive jobs and Over time, and as resources allow, each time from the left column to the right support from employment specialists of the components of the mainstream column in Table 1. In this way, implement- (Drake et al, 1999). This individual place- model can be complemented by additional ing the components of a modern mental ment and support model emphasises com- and differentiated options, described here health service can be seen as a pragmatic petitive employment in integrated work as specialised differentiated mental health exercise undertaken by all those with an settings with follow-up support (Priebe et services (step C). Notably, the evidence interest in improving care. al, 1998); studies of such programmes have base for these more recent and innovative been encouraging in terms of increased forms of care is stronger than for any of rates of competitive employment (Marshall ACKNOWLEDGEMENT the service components in steps A or B, et al, 2001; Lehman et al, 2002). described above in relation to lower This review is based upon an evidence synthesis pre- resource countries. Indeed, few high-quality pared for the WHO Regional Office for Europe's DISCUSSION scientific studies have been carried out Health Evidence Network (http://www.euro. in low-income countries (Patel & who.int/document/hen/mentalhealth.pdf). This review makes clear that there is no Sumathipala, 2001; Isaakidis et al, 2002). compelling argument and no scientific evi- Consequently, the relevance of most pub- REFERENCES dence favouring the use of hospital services lished research in this field to less econom- alone. On the other hand, there is also no ically developed countries may be low. This Addington, J., Coldham, E. L., Jones, B., et al (2003) evidence that community services alone schema therefore places the evidence of The first episode of psychosis: the experience of relatives, Acta Psychiatrica Scandinavica, 108, 285-289. can provide satisfactory and comprehensive effective services within the appropriate re- care. Both the evidence available so far, and source context; `resource' here refers not Alem, A. (2002) Community-based vs. hospital-based accumulated clinical experience, therefore only to the monetary investments made, mental health care: the case of Africa. World Psychiatry, I, 99-100. support a balanced approach, incorpor- but also to the available numbers of staff, ating elements of both hospital and their levels of experience and expertise, Becker,T. (2001) Out-patient psychiatric services. In community care (Thornicroft & Tansella, their therapeutic orientation and the contri- Textbook of Community Psychiatry (eds G. Thornicroft & G. Szmukler), pp. 277-282. Oxford: Oxford University 2002). butions available from the wider social and Press. The material resources available will family networks (Desjarlais et al, 1995). severely constrain how this approach is Two important implications arise from Becker, T. & Vazquez-Barquero, J. L. (2001) The European perspective of psychiatric reform. Acta applied in practice. In low-resource areas this approach. First, the stepped care model Psychiatrica Scandinavica Supplementum, 8-14. it may be unrealistic to invest in any of suggests that there should be a degree of co- the components described here as main- ordination between service components, in Becker, D. R., Drake, R. E., Farabaugh, A., et al (1996) Job preferences of clients with severe psychiatric stream mental health care (step B), and particular between the provision of primary disorders participating in supported employment the focus will need to be upon primary and specialist care. We recognise that such programs. Psychiatric Services, 47, 1223-1226.

288

COMMUNITY AND/OR HOSPITAL CARE

BMJ Publishing Group (2003) Clinical Evidence. Isaakldis, P., Swingler, G. H., Pienaar, E., at a1(2002) Murray, C. & Lopez, A. (1996) The Global Burden of London: BMJ Books. Relation between burden of disease and randomised Disease Vol. I. A Comprehensive Assessment of Mortality evidence in sub-Saharan Africa: survey of research. BMJ, and Disability From Diseases, Injuries and Risk Factors in Burns,T. (2001) Generic versus specialist mental health 324, 702. 1990, and Projected to 2020. Cambridge, MA: Harvard teams. In Textbook of Community Psychiatry (eds G. University Press. Thornicroft & G. Szmukler), pp. 231-241. Oxford: Johnstone, P. & Zolese, G. (1999) Systematic review of Oxford University Press. the effectiveness of planned short hospital stays for Nathan, P. & Gorman, J. (2002) A Guide to Treatments mental health care, BMJ, 318, 1387-1390. That Work. Oxford: Oxford University Press Burns,T., Creed, F„ Fahy,T., et al (1999) Intensive versus standard case management for severe psychotic Joy, C., Adams, C. & Rice, K. (2002) Crisis Njenga, F. (2002) Challenges of balanced care in Africa. illness: a randomised trial. UK 700 Group. Lancet, 353, intervention for people with severe mental illness. World Psychiatry, I, 96-98. 2185-2189. Cochrane Library, issue 2, Oxford: Update Software. Nordentoft, M., Knudsen, H. C. & Schulsinge , F. Burns,T., Fioritti, A„ Holloway, F., et a/ (2001) Case Killaspy, H., Dalton, J., McNicholas, S., at a1(2000) (1992) Housing conditions and residential needs of management and assertive community treatment in Drayton Park, an alternative to hospital admission for psychiatric patients in Copenhagen. Acta Psychiatrica Europe. Psychiatric Services, 52, 631-636. women in acute mental health crisis. Psychiatric Bulletin, Scandinavica, 85, 385-389. 24, Catty, J., Burns, T„ Knapp, M., et al (2002) Home 101-104. Ormel, J.,Von Korff, M., Ustun, B., at al (1994) Common mental disorders and disability across cultures. treatment for mental health problems: a systematic Knapp, M., Chisholm, D., Astin, J., at al (1997) The Results from the WHO Collaborative Study on review. Psychological Medicine, 32, 383-401. cost consequences of changing the hospital-community Psychological Problems in General Health Care. JAMA, balance: the mental health residential care study Catty, J., Burns,T. & Comas, A. (2003) Day centres 272, 1741-1748. for severe mental illness (Cochrane Review). Cochrane Psychological Medicine, 27, 681-692, Library, issue I.Oxford: Update Software. Patel,V. & Sumathipala, A. (2001) International Larsen,T. K., Friis, S., Haahr, U., at al (2001) Early representation in psychiatric literature: surveys of six detection and intervention in first-episode Chilvers, R., Macdonald, G. & Hayes, A. (2003) leading journals. British journal of Psychiatry, 178, schizophrenia:a critical review. Acta Psychiatrica Supported I-lousing for People With Severe Mental 406-409. Disorders (Cochrane Review). Oxford: Update Software. Scandinavica, 103, 323-334. Phillips, S. D., Burns, B. J., Edgar, E. R., at al (2001) Davies, S., Presilla, B., Strathdee, G., at al (1994) Latimer, E. A. (1999) Economic impacts of assertive Moving assertive community treatment into standard Community beds: the future for mental health care? community treatment: a review of the literature. practice. Psychiatric Services, 52, 771-779. Social Psychiatry and Psychiatric Epidemiology, 29, Canadian Journal of Psychiatry, 44, 443-454. Polak, P. & Warner, 241-243, R. (1996) The economic life of Leff, J. (1997) Care in the Community. Illusion or Reality? seriously mentally ill people in the community. Psychiatric Decl, P. A., Santos, A. B., Hiott, D.W., at al (1995) London: Wiley. Services, 47, 270-274. Dissemination of assertive community treatment Lehman, A. F., Carpenter, W.T., Goldman, H. H., et al Poullier, J-P., Hernandez, P„ Kawabata, K., at al programs. Psychiatric Services, 46, 676-678, (1995) Treatment outcomes in schizophrenia: (2002) Patterns of Global Health Expenditures: Results for Department of Health (2002) Community Mental implications for practice, policy, and research. 191 Countries.World Health Organization Discussion Health Teams. Mental Health Policy Implementation Schizophrenia Bulletin, 21, 669-675. Paper No. 51.Geneva: WHO. Guide, London: Department of Health. Lehman, A. F., Goldberg, R., Dixon, L. B., at al Priebe, S., Warner, R., Hubschmld,T., at al (1998) Desjarlais, R., Eisenberg, L., Good, B., at al (1995) (2002) Improving employment outcomes for persons Employment, attitudes toward work, and quality of life World Mental Health. Problems and Priorities in Low with severe mental illnesses. Archives of General among people with schizophrenia in three countries. Income Countries. Oxford: University Press. Psychiatry, 59, 165-172. Schizophrenia Bulletin, 24, 469-477.

Drake, R. E., McHugo, G. J., Bebout, R. R., at al Marshall, M. & Lockwood, A. (2003) Assertive Quirk, A. & Lelliott, P. (2001) What do we know (1999) A randomized clinical trial of supported community treatment for people with severe mental about life on acute psychiatric wards in the UK? A employment for inner-city patients with severe mental disorders (Cochrane Review). Cochrane Library, issue I. review of the research evidence. Social Science and disorders. Archives of General Psychiatry, 56, 627-633. Oxford: Update Software. Medicine, 53, 1565-1574. Raune, D., Kuipers, E. & Bebbington, P. E. (2004) Flander, M., Burns, T., Mc Hugo, G. J., et of (2003) Marshall, M. & Lockwood, A. (2004) Early Expressed emotion at first-episode psychosis: Assertive community treatment across the Atlantic: intervention for psychosis (Cochrane Review). Cochrane investigating a carer appraisal model. British Journal of comparison of model fidelity in the UK and USA. British Library, issue 2. Chichester: John Wiley. Journal of Psychiatry, 182, 248-254. Psychiatry, 184, 321-326. Marshall, M., Crowther, R., Almaraz-Serrano, A., at Rosen, A. & Barfoot, K. (2001) Day care and Friis, S., Larsen, T. K., Melle, I., et al (2003) at (2001) Systematic reviews of the effectiveness of day Methodological pitfalls in early detection studies - the occupation: structured rehabilitation and recovery care for people with severe mental disorders: (I) acute programmes NAPE Lecture 2002. Nordic Association for Psychiatric and work. In Textbook of Community day hospital versus admission; (2) vocational Psychiatry Epidemiology. Acta Psychiatrica Scandinavica, 107, 3-9. (eds G. Thornicroft & G. Szmukler), rehabilitation: (3) day hospital versus outpatient care. pp. 296-308. Oxford: Oxford University Press. Grob, G. (1991) From Asylum to Community. Mental Health Technology Assessment, 2, I -75. Roth, A. & Fonagy, P. (1996) What Works for Whom? A Health Policy in Modern America. Princeton, NJ: Princeton McGorry, P. D. & Klllackey, E. J. (2002) Early Critical Review of Psychotherapy Research. New York: University Press. intervention in psychosis: a new evidence based Guilford Press. Hafner, H. (1987) Do we still need beds for psychiatric paradigm. Epidemiologia e Psichiatria Sociale, II, 237-247. Saarento, O., Hansson, L., Sandlund, M., et al (1996) patients? An analysis of changing patterns of mental McGarry, P. D., Yung, A. R., Phillips, L. J., et al (2002) The Nordic comparative study on sectorized psychiatry. health care, Acta Psychiatrica Scandinavica, 75, 113-126. Randomized controlled trial of interventions designed to Utilization of psychiatric hospital care related to amount Hansson, L., Muus, S.,Vinding, H. R., at al (1998) The reduce the risk of progression to first-episode psychosis and allocation of resources to psychiatric services. Social Nordic Comparative Study on Sectorized Psychiatry: in a clinical sample with subthreshold symptoms. Archives Psychiatry and Psychiatric Epidemiology, 31, 327-335. contact rates and use of services for patients with a of General Psychiatry, 59, 921-928. Saxena, S. & Maulik, P. (2003) Mental health services functional psychosis. Acta Psychiatrica Scandinavica, 97, Mosher, L. R. (1999) Soteria and other alternatives to in low and middle income countries: an overview. 315-320, acute psychiatric hospitalization: a personal and Current Opinion in Psychiatry, 16, 437-442, professional review Journal of Nervous Harrigan, S. M., McGorry, P. D. & Krstev, H. (2003) and Mental Scott, J. & Lehman, A. (2001) Case management and Disease, 187, 142-149. Does treatment delay in first-episode psychosis really assertive community treatment. InTextbook ofCommunity matter? Psychological Medicine, 33, 97-I 10. Mubbashar, M., (1999) Mental health services in rural Psychiatry (eds G. Thornicroft & G. Szmukler), Holloway, F. & Carson, J. (2001) Case management: Pakistan, In Common Mental Disorders in Primary Care pp. 253-264. Oxford: Oxford University Press. an update. International journal of Social Psychiatry, 47, (eds M.Tansella & G.Thornicroft), pp.67-80. London: Shepherd, G. (1990 Theory and Practice of Psychiatric 21-31. Routledge. Rehabilitation. Chichester: John Wiley. Holloway, F.,Wykes,T., Petch, E., at at (1999) The Mueser, K.T., Bond, G. R., Drake, R. E., at at (1998) Shepherd, G. & Murray, A. (2001) Residential care. In new long stay in an inner city service: a tale of two Models of community care for severe mental illness: a Textbook of Community Psychiatry (eds G. Thornicroft & cohorts. International Journal of Social Psychiatry, 45, review of research on case management. Schizophrenia G. Szmukler), pp.309-320.Oxford: Oxford University 93-103. Bulletin, 24, 37-74, Press.

289 THORNICROFT & TANSELLA

Shepherd, G., Muijen, M., Dean, R., et al (1996) Residential care in hospital and in the community - quality of care and quality of life. British Journal of CLINICAL IMPLICATIONS Psychiatry, 168, 448-456. Slmmonds, S., Cold, J., Joseph, P., at a1(2001) n Countries with low levels of resources should focus on (a) establishing and improving Community mental health team management in severe services in primary care settings, along with (b) the provision of specialist back-up. mental illness: a systematic review. British Journal of Psychiatry, 178, 497-502. n Medium-resource countries, in addition to (a) and (b), may then develop Sledge,W. H.,Tebes, J., Rakfeldt, J., et al (1996a) Day 'mainstream' mental health care with the following components; (i) out-patient/ hospital/crisis respite care versus inpatient care. Part I: Clinical outcomes. American Journal of Psychiatry, 153, ambulatory clinics; (ii) community mental health teams; (iii) acute in-patient care; 1065-1073. (iv) long-term community-based residential care; and (v) employment and occupation, Sledge, W. H.,Tebes, J., Wolff, N., et al (19966) Day hospital/crisis respite care versus inpatient care, Part II: n High-resource countries may include (a) and (b), and then develop evidence-based Service utilization and costs. American Journal of specialised/differentiated care for each of (i)-(v). Psychiatry, 153, 1074-1083. Sytema, S., Mlcciolo, R. & Tansella, M. (1997) LIMITATIONS Continuity of care for patients with schizophrenia and related disorders: a comparative south Verona and n Within countries there may be wide variations between regions, so that different Groningen case-register study. Psychological Medicine, 27, 1355-1362. resource levels apply when planning services.

Szmukler, G. & Holloway, F. (2001) In-patient n The model proposed is most directly applicable to health systems that have a high treatment, In Textbook of Community Psychiatry (eds G. Thornicroft & G. Szmukler), pp. 321-337. Oxford: degree of coordinated planning and commissioning of services. Oxford University Press. n There remain many areas of mental health services, most Tansella, M. (1986) Community psychiatry without notably in-patient care mental hospitals - the Italian experience: a review and out-patient/ambulatory services, where relatively little research on effectiveness Journal ofthe Royal Society of Medicine, 79, 664-669. or cost-effectiveness has been conducted. Teague, G. B., Bond, G. R. & Drake, R. E. (1998) Program fidelity in assertive community treatment: development and use of a measure. American Journal of Orthopsychiatry, 68, 216-232. GRAHAM THORNICROFT, MD, Health Services Research Department, Institute of Psychiatry, King's College Thornicroft, G. (1991) The concept of case London, UK; MICHELE TANSELLA, MD, Department of Medicine and Public Health, Section of Psychiatry, management for long-term mental illness. International University of Verona, Italy Review of Psychiatry, 3, 125-132. Thornicroft, G. (2001) Measuring Mental Health Needs Correspondence: Professor GrahamThornicroft, Health Service Research Department, Institute of (2nd edn). London: Gaskell. Psychiatry, King's College London, De Crespigny Park, London 5E5 8AF, UK.Tel: +44 (0)207 848 0735; Thornicroft, G. & Bebbington, P. (1989) fax: +44 (0)2072771462; e-mail: g.thornicroft@iop,kcl.ac.uk - from hospital closure to service development. British Journal of Psychiatry, 155, 739-753. (First received 7 October 2003, final revision 21 May 2004, accepted 24 May 2004) Thornicroft, G. & Tansella, M. (1999) The Mental Health Matrix: A Manual to Improve Services. Cambridge: Tyrer, P., Morgan, J., Van Horn, E., et al (1995) A Cambridge University Press. Wiersma, D., Klulter, H., Nienhuis, F. J., et al (1995) randomised controlled study of close monitoring of Costs and benefits of hospital and day treatment with Thornicroft, G. & Tansella, M. (2002) Balancing vulnerable psychiatric patients, Lancet, 345, 756-759. community care of affective and schizophrenic disorders, community-based and hospital-based mental health British Journal of Psychiatry 167 (suppl. 27), 552-559, care. World Psychiatry, I, 84-90, Tyrer, P., Evans, K., Gandhi, N., et al (1998) Randomised controlled trial of two models of care for Wiersma, D., Nlenhuls, F. J., Slooff, C. J., et al (1997) G. Thornicroft, & Tansella, M. (2003) What Are the discharged psychiatric patients. BMJ, 316, 106-109. Assessment of needs for care among patients with Arguments for Community-Based Mental Health Care? schizophrenic disorders 15 and 17 years after first onset Copenhagen: World Health Organization (European Tyrer, P., Cold, J., Slmmonds, S., at al (2003) of psychosis. Epidemiologia e Psichiatria Sociale, 6, 21-28. Region) Health Evidence Network. Community mental health teams (CMHTs) for people Wing, J. K. & Brown, G. (1970) Institutionalism Thornicroft, G., Wykes,T., Holloway, F., et al (1998) with severe mental illnesses and disordered personality and Schizophrenia. Cambridge: Cambridge University Press. From efficacy to effectiveness in community mental (Cochrane Review). Cochrane Library, issue 2.Oxford: Update Software. health services. PRiSM Psychosis Study, 10. British Journal World Bank (2002) World Development Report 2002. of Psychiatry, 173, 423-427. Van Wijngaarden, G. K., Schene, A., Koeter, M., et al Building Institutions for Markets. Washington, DC: World Thornicroft, G., Becker, T., Holloway, F., et al (1999) (2003) People with schizophrenia in five countries: Bank. Community mental health teams: evidence or belief? conceptual similarities and intercultural differences in World Health Organization (2001a) World Health British Journal of Psychiatry, 175, 508-513. family caregiving. Schizophrenia Bulletin, 29, 573-586. Report 2001. Mental Health: New Understanding, New Thornicroft, G., Rose, D., Huxley, P., et al (2002) Hope. Geneva: WHO. http: / /www.who.int/whr200I/ Von Korff, M. & Goldberg, D. (2001) Improving What are the research priorities of mental health service 2001/main/en/ outcomes in depression. The whole process of care users? Journal of Mental Health, II, I-5, needs to be enhanced. BMJ, 323, 948-949. World Health Organization (2001b) Atlas: Mental Tomov,T. (2001) Central and Eastern European Health Resources in the World 2001, Geneva: WHO. Countries. In The Mental Health Matrix: A Manual to Warner, R. (1994) Recovery From Schizophrenia (2nd http: / /www.who.int/mental - health/media/en/244, pdf Improve Services (eds G. Thornicroft & M. Tansella), edn). London: Routledge, pp.216-227. Cambridge: Cambridge University Press. Ziguras, S. J. & Stuart,G,W. (2000) A meta-analysis of Warner, R. & McGorry, P. D. (2002) Early intervention the effectiveness of mental health case management Trleman, N., Smith, H. E., Kendal, R., et al (1998) in schizophrenia: points of agreement, Epidemiologia e over 20 years. Psychiatric Services, 51, 1410-1421. TheTAPS Project 41: homes for life? Residential stability Psichiatria Sociale, II, 256-257. five years after hospital discharge. Team for the Ziguras, S. J., Stuart, G.W. & Jackson, A. C. (2002) Assessment of Psychiatric Services. Community Mental Warr, P. (1987) Work, Unemployment and Mental Health, Assessing the evidence on case management, British Health Journal, 34, 407-417. Oxford: Oxford University Press. Journal of Psychiatry, 181, 17-21,

290