The authors describe the main forms of supported accommodation available to people with mental health problems in the UK. They address the questions of who needs supported accommodation and how people may be selected for the differing types of unit. Historical trends and policy in this area are reviewed and the general research evidence and evidence of cost-effectiveness are considered. A ‘total system’ approach to assessing accommodation need in a locality is advocated. The authors also consider key quality issues that should be addressed for supported accommodation to be effective.
The provision of affordable, decent-quality, adequately supportive accommodation is a key factor in helping people with severe mental illnesses: basic physiological needs, shelter and safety are at the core of any hierarchy of needs (e.g. Maslow, 1962) and have a fundamental impact on quality of life.
The closure of the long-stay psychiatric hospitals in the UK might be seen as the most radical change in the nation’s public health policy in the 20th century, and it resulted in a major reprovision of services in what is usually known as community care. In this article we attempt to describe the ‘mixed economy’ of care that is now available for people in the UK whose mental health affects their ability to live independently.
It is important to note that we focus purely on longer-term accommodation for patients with severe mental illnesses and that we do not consider the important issue of alternatives to acute hospital care (an area that has been reviewed by Boardman & Hodgson, 2000).
What is supported accommodation?
The different forms of supported accommodation are difficult to define, owing to the existence of various systems to classify them, the multiple dimensions of staff status and staffing ratios involved and the heterogeneity of the patients who end up in such accommodation. Furthermore, categories that have traditionally operated for care of elderly people (such as private nursing homes and registered care homes), although often applied in the area of adult mental health, relate in practice to a different set of health problems and often different styles of working.
Different forms of supported accommodation
The following guide offers a breakdown of conventional forms of supported accommodation by unit type and in decreasing level of dependency. In other words, as one goes down the list the expected level of support and/or supervision typically provided reduces. It is, however, important to recognise that theoretical levels of dependency do not always equate with actual levels of need. Box 1⇓ summarises the main providers of these services.
Supported accommodation and main service providers (Lelliott et al, 1996)
Long-stay wards: usually in large NHS hospitals
High- and medium-staffed hostels (24-hour nursed-care units): provided variously – directly through the NHS, via the private and voluntary sectors and via local authority social services departments
Low-staffed hostels: mostly the private and voluntary sectors (a very few are run by local authority social services departments)
Staffed care homes: the private and voluntary sectors, with some local authority social services departments
Group homes: the voluntary sector and local authority social services departments
Core and cluster/high-dependency housing: mostly charitable organisations and housing associations
Usually part of larger National Health Service (NHS) hospitals, long-stay wards generally have lower staffing levels than acute wards. Their provision is highly variable and some districts now operate services entirely without explicit continuing-care hospital provision.
High- and medium-staffed hostels
These hostels are also known as 24-hour nursed-care units. They vary in status from hostels funded and run by the NHS to nursing homes and residential care homes provided by the private or voluntary sectors. Voluntary/charitable provision may be through mental health charities such as Mind and Rethink. These units often occupy large, older detached houses in the residential areas of cities. Staffing levels vary from 8 to 20 workers per unit, and each unit typically provides for 6 to 12 patients. Night staff may remain awake or work sleep-in rotas. For a fuller description of this type of unit see Shepherd (1998a).
The 24-hour nursed-care units within the NHS typically operate as part of a district rehabilitation service, whereas private- and voluntary-sector care homes operate outside the NHS and can receive referrals from any source.
Low-staffed hostels typically have day cover only, provided by a small number (usually two or three) staff. Far fewer of these staff have care qualifications than do staff in long-stay or acute wards (15% v. 49% and 63% respectively; Lelliott et al, 1996).
Staffed care homes
Also known as supported lodgings or adult fostering homes, these usually have a very high proportion of unqualified ‘staff’, who are appointed as carers, generally through a care scheme operated by local authority social services departments. Units vary from small, family homes with up to 3 residents to larger establishments with up to 12 residents in a type of supported hostel, with resident care home staff. Anstee (1985) has given an overview of the supported lodgings scheme.
These are not staffed, but are typically houses owned and managed by local authority social services departments, with up to 5 residents. They generally have regular visits from support workers through local social services or mental health rehabilitation services.
Core and cluster (high-dependency) housing
In this more recent model of care, individual flats or bedsits are overseen by a ‘core’ staffed unit or by visiting support staff, who may be employed by statutory services (the NHS or local authority social services). However, they are more commonly run by charitable organisations such as Rethink or by various housing associations. Provision of this form of accommodation, which is described by Carling (1993), has increased greatly in the past 10 years.
Who needs supported accommodation?
The broad range of patients who may benefit from sheltered accommodation has often been divided into the following three main groups.
Old long-stay patients
The ‘old long-stay patients’ are those whose hospitalisation pre-dated the deinstitutionalisation movement that started in the 1960s, but have remained in-patients. This group is now very small, and is non-existent in many districts.
New long-stay patients
The term ‘new long-stay patients’ was created by Mann & Cree (1976) to refer to patients who, despite modern treatment approaches and the ideology of community care, cannot be discharged from hospital, owing to their level of psychopathology, disability or behavioural disturbance.
The community care generation
The health care system now includes a large group of patients who have responded to treatment, but have residual symptoms and require ongoing care and support, but do not have access to family or similar help. Historically, these patients’ care pathways would have often involved lengthy periods in psychiatric hospitals. However, in the current health care system some of these patients may never have been psychiatric in-patients.
Lamb (1998) has described consequences of the loss in the USA of long-term institutions, including an increase in rates of severe mental illness among prison and homeless populations. He cites high rates of disturbance in the residual psychiatric hospital population, and suggests that present-day psychiatric patients have not been ‘institutionalised to passivity’ by lengthy in-patient treatment. The position in the UK is less clear, perhaps in part because alternative supported community placements have been developed, albeit in an inconsistent, variable form.
Although attempts have been made (Lelliott & Wing, 1994) to assess the numbers of new long-stay patients, and indeed ‘match’ these to existing residential care facilities, there is no clear method to assess the need for supported accommodation among the broader psychiatric population – the community care generation – and this area remains unclear and insufficiently researched. Local provision is highly varied and largely determined by historical patterns of development (Lelliott, 1996). At best, imprecise norms for differing forms of supported accommodation have been proposed (Johnson et al, 1996). There is no simple system to assess overall need in a locality and no generally recognised instrument to assess housing need at an individual level (Strathdee & Jenkins, 1996).
Patient selection for different types of unit
The literature contains few examples of systematic approaches to allocating special needs accommodation at a local level.
In theory, patients who have better living skills (able to manage self-care, budgeting and shopping) would be likely to cope in more independent settings. There is also evidence of higher levels of psychopathology (Simpson et al, 1989) and dependency (Robson, 1995) in hostel residents and rehabilitation patients in hospital than in residents of group homes. Shepherd et al(1996) confirmed this in a detailed analysis of 25 residential units in London, but interestingly found that staffing ratios in community homes bore no relationship to levels of dependency.
A key issue in selection is patient choice, and a developing literature indicates that patients are increasingly dissatisfied with the traditional, ‘institutional’ model of supported accommodation. When asked, they tend to express a preference for independent accommodation that allows for privacy. They also prefer to have access to flexible levels of support when needed, as opposed to support being provided as part of their accommodation (Tanzman, 1993; Rose & Muijen, 1997).
In practice, there is considerable overlap across different types of supported accommodation in terms of the types of patient being cared for. Ideally, a locality-based accommodation system should include a broad range of facilities with differing levels of support that allows individual choice to be accommodated as much as possible. The reality is rather different and presents a picture of extensive geographical variation, with typically fragmented, poorly organised systems for allocating special needs housing and a stark lack of awareness of unmet need within the system.
Recently, more systematic approaches to assessing need for accommodation have been developed in some countries. Durbin et al(2001) reported a Canadian project commissioned by the Ontario Ministry of Health to identify alternative placements for current hospital in-patients. The process involved multidimensional assessment of need incorporating the Colorado Client Assessment Record and use of stakeholder panels to assess need across the whole system, rather than in a single programme. Five levels of care were developed, and it was concluded that only 10% of current in-patients needed to remain in hospital and that 60% could live independently in the community with appropriate support. The results of this project will be interesting to follow.
Fitz & Evenson (1999) reported on a new tool, the St Louis Inventory of Community Living Skills, developed to help clinicians recommending residential settings appropriate for people with mental health problems. They concluded that community living skills, social skills and problem behaviour were primary characteristics affecting adjustment to residential settings. In a related area, Bartlett et al(2001) analysed 730 acute admissions, finding that 35% of patients had been inappropriately placed at some time and that many of the patients might have benefited from alternative, mostly community-based, services. In 24% of cases divertible to community care it was considered that specialist accommodation supported by nurses or care workers would provide an effective alternative to acute hospital care.
Historical patterns of residential provision
Between 1954 and 1996, 110000 psychiatric hospital beds were closed, but only about 13000 places in hostels and homes were opened to replace them; this resulted in a reduction of two-thirds in the number of residential places available for people with severe mental illness (Lelliott, 1996). Together with the shift towards local authority and voluntary-sector provision, this has resulted in an even greater loss of 24-hour nursed-care units (Faulkner et al, 1992). A survey examining patterns of use of community and hospital units (Lelliott et al, 1996) suggested that individuals who pose a high risk of violence tend to be excluded from community placement, with a resulting tendency for the creaming off of the least difficult patients, while the most difficult are left in increasingly disturbed acute or long-stay NHS wards.
A UK audit (Lelliott & Wing, 1994) revealed an average point prevalence of 6.1 new long-stay patients per 100000 population, while many English services had few long-stay psychiatric beds: 31% of new long-stay patients were housed on acute wards and, of the 47% of these patients who were thought to require a community placement, half remained on acute wards owing to a lack of available resources. Concerns about blockage of acute beds and the increasing pressures on acute hospital services have been most apparent in London and other large cities; the Monitoring Inner London Mental Illness Services (MILMIS) Project Group reported ‘true’ bed occupancy levels of 130% (MILMIS Project Group, 1995). The Government was aware of such pressures on overall provision, and as early as 1991 had issued guidance proposing that more 24-hour nursed-care beds should be commissioned (Department of Health, 1991). Whether this guidance has been heeded remains unclear: these relatively expensive units would require considerable additional funding, estimated at between £25000 and £50000 per bed annually (Department of Health, 1996). Where hospital closure has already occurred, there has been no clear system to identify new funding for this UK Government priority. It appears that further research to identify current levels of provision of 24-hour nursed care and to set this in the context of the increasing levels of low secure provision would be of value.
The problem of pressure on acute beds cannot be viewed in isolation (Shepherd, 1998b), and in order to carry out a meaningful assessment of a locality, an understanding of the total provision of hospital, hostel and supported accommodation should be attempted: levels of different forms of accommodation vary greatly and adequate levels of one form can to some extent compensate for deficiencies in others. It is important that homeless people are not overlooked, as there is known to be a higher level of psychiatric illness in this population and they may gain particular long-term benefit from specialist supported accommodation.
The mentally ill prison population
It has been argued (Murphy, 1992) that the closure of psychiatric hospital beds has led to an increase in mental health problems dealt with in the prison system, because of a reluctance among general psychiatrists working in modern acute wards, with typically short admission periods, to accept high-risk patients from the courts and prisons. Brooke et al(1996) estimated that 880 men in England and Wales needed transfer from remand prisons to hospital for psychiatric treatment, and it has been estimated that such prisoners wait on average 11 months for transfer (Reed & Lyne, 2000).
It is unclear whether the deinstitutionalisation movement has led to real increases in the numbers of severely mentally ill people in the UK prison system. However, Lamb (1998) is sure that in the USA it has. He presents two arguments to support this view: the large numbers of mentally ill prison residents, and the observation that a high proportion of mentally ill people found in the criminal justice system resemble in most aspects those who used to be in long-term psychiatric institutions. Within the UK system it is clear that prisons are sometimes used to accommodate difficult patients with challenging behaviour who do not readily fit into modern short-stay acute NHS facilities, and it also appears that a small number of people with severe mental illness are inappropriately placed in the criminal justice system.
Recent trends in supported accommodation
Since the initial wave of ward closures, in which patients mostly moved to hostels, group homes or supported lodgings, the pattern of reprovision has continued to evolve in line with changing national provision and the changing expectations of the general population. In the past 10 years there has been a move away from group home provision (Abrahamson et al, 1995), with patients increasingly voicing concerns about sharing their ‘home’ and problems of ‘ghettoisation’ and stigmatisation. New forms of high-dependency housing schemes (e.g. see Keck, 1990; Carling, 1993; Middleboe et al, 1998) appear to have the potential to meet some of the need in this population, while being more acceptable to users. The key issue is the support of the more challenging patients, who may have comorbid substance misuse and/or forensic problems. In theory, the assertive community treatment approach may facilitate this type of initiative (Mueser et al, 1998), by underpinning the support provided within the non-statutory housing sector and offering variable levels of support which can be targeted appropriately, according to need and fluctuating health.
Current approaches to clinical risk management, underpinned by centrally driven policies in this area, may also have a significant effect on the overall deployment and balance of supported accommodation for people with severe mental illnesses. The increasing numbers of medium and low secure beds, often provided through private organisations, to deal with patients who present challenging behaviour should also be seen as a component of the total provision.
The NHS hospital plan (Ministry of Health, 1962) started the hospital closure programme and promoted the development of acute units in general hospitals. Care in the Community (Department of Health and Social Security, 1981) passed the responsibility for managing residential care from regional health authorities to local authorities. Furthermore, the National Health Service and Community Care Act 1990 enabled and encouraged non-statutory agencies to operate residential facilities. Partnerships in Action (Department of Health, 1998) promoted joint health and social services commissioning and ‘cross-management’ arrangements. The Supporting People initiative (Department of the Environment, Transport and the Regions, 2001) aims to provide housing-related support services to vulnerable people (including those with mental illnesses), robustly funded and planned using a coordinated multi-agency approach. The advantage of this should be greater ability to assess and plan for need at a local level, but it is possible that the level of need identified will exceed the funding available and the new system may be compromised by a rigid financial framework.
Changes to the Mental Health Act 1983 may also have had an impact on accommodation issues. The Patients in the Community Act 1994 empowered doctors to place patients in defined accommodation following compulsory treatment in hospital. However, use of this legislation appears to have been limited by practical difficulties, in particular in dealing with crisis. It appears that levels of use of this part of the Mental Health Act are in practice patchy and inconsistent.
The research evidence
The most researched area of supported accommodation is the 24-hour nursed care provided in high- and medium-staffed hostels, reviewed by Shepherd (1998a) and also by Macpherson & Jerrom (1999). Such units can effectively support most patients referred to them, some of whom can be resettled to less dependent community placements over a period of years. Units providing 24-hour nursed care tend to be associated with an improvement in social functioning, higher levels of social networks and a reduced level of negative symptoms in schizophrenia, but they typically do not affect positive symptoms. Patients and relatives have tended to report a higher satisfaction with these units than with hospitals, but they may be less attractive to some patients owing to perceptions of stigma and restrictiveness. The research in this area varies in quality, and only one randomised controlled trial has been undertaken (Hyde et al, 1987). In both research and practice, the above-mentioned bias in the selection of patients for placement that leaves the most disturbed individuals in hospital environments makes an overall evaluation of the work of such units within the total health care system more difficult.
The evidence base for staffed care homes or core and cluster accommodation is less well established, although a review in this area demonstrated a strong user preference for independent private flats and for flexible outreach support (Tanzman, 1993). A North American study (Keck, 1990) suggested that an approach that aimed to provide ‘normal housing’, together with ‘practical assistance’ was largely effective and was associated with a dramatic decline in hospitalisation. Other studies, from the USA and Scandinavia respectively, have found high levels of independent social functioning (Segal & Kofler, 1993) and improved quality of life (Middleboe et al, 1998) in core and cluster units. A comparative evaluation in the USA undertaken by Nelson et al(1997) found that residents of supported apartments, group homes and board and care homes (similar to supported lodgings in the UK) all had positive outcomes over time in terms of work and education. Residents in the group facilities reported that they experienced greater support and lower levels of abuse than those in the other settings. Those in supported apartments and group homes spent less on rent and made more decisions about various aspects of their life.
There is little evidence in the literature of differing effectiveness between the various forms of community provision. This is perhaps not surprising in a health care system where different units are perceived to cater for different types of patient (i.e. with different levels of dependency) and levels of challenging behaviour. Some patients benefit from the support and increased contact of group living and may otherwise face loneliness, isolation and neglect. Older patients are vulnerable to physical decline and poor quality of life without support. In practice, it seems that access to a range of different forms of supported accommodation, through which patients may move according to need as well as by choice, at different times in their lives or phases of their illness, is an ideal worth aspiring towards.
Evidence from the TAPS study
An important consideration in research in this area is the overall effect of hospital closure in a locality. The Team for the Assessment of Psychiatric Services (TAPS) generated extensive evidence regarding the progress of long-stay hospital patients leaving Friern and Claybury hospital in North London (Leff, 1997). In a carefully planned process, which involved the allocation of a ‘funding dowry’ to each discharged patient, patients were carefully followed up and evaluated over an extended period after they moved from long-term hospital to community care. One year after discharge, 49% of the patients were living in large hostels, residential or nursing homes; 15% were in community in-patient accommodation (which we take to mean directly provided NHS hospital hostels); 12% were living independently; 6% were in staffed group homes; and 4% were in unstaffed group homes (Beecham et al, 1997). The remainder were in sheltered housing or foster care. Compared with matched controls remaining in hospital, the community group had reduced negative symptoms, improved social functioning, increased social networks and greatly increased levels of satisfaction. There was no difference, however, in positive symptoms, physical health status or rates of suicide and crime. Overall, costs were slightly lower for the community group.
The results of the TAPS programme, together with the results of research into 24-hour nursed care and a large number of uncontrolled studies of patients in the community (e.g. Borge et al, 1999), have been widely viewed as supporting the value of alternative community provision for long-term hospital in-patients.
The cost of community care
The cost of changing the balance between hospital and alternative forms of long-term care provision was explored by Chisholm & Hallam (2001), with reference to the TAPS project and a large cross-sectional survey in England and Wales. They concluded that reprovision of services through community care should be seen not as a cheap substitute for in-patient care, but as a spectrum of provision that meets the diverse needs of people with mental health problems. They also noted that caring for those with greatest need in the community may actually be more expensive than in the old institutions.
The need for a ‘total system’ approach
It is evident from an examination of individual forms of supported accommodation that none will function independently and that there is a complex dynamic of overall provision that encompasses specialist community mental health and primary care services, in addition to the support services attached to any residential care units. Thus, although a review of new long-stay patients (Lelliott & Wing, 1994) concluded that many individuals remain in hospital because residential needs are not being met by existing community provision, it has been argued (Shepherd et al, 1997) that simply providing more acute beds is not the best solution. The problem requires a spectrum of solutions that include home-based intensive support and a range of hospital and community facilities, including properly funded 24-hour nursed care and assertive outreach teams. In order to meet the overall need for specialist accommodation in a population, it has been argued that a ‘systems perspective’ should be taken, the aim of which is to generate a ‘well co-ordinated, clearly targeted and efficient system for delivering appropriate housing’ (Shepherd, 1998b).
The development of systems to identify existing provision and met and unmet accommodation needs in a local region will help to establish the case for further resources. It has been recognised (Audit Commission, 1994) that budgeting for the needs of patients following psychiatric hospital closure was never really adequate and there appears to be a recognition in the Supporting People initiative that increased resources will be needed to fund this area properly. Franklin (1998) has stressed the importance of real integrated approaches, good communication between different agencies and shared approaches to assessing housing need.
In a study that formally assessed individually rated patient need against costs and care being provided for tenants of a housing association in London, Järbrink et al(2001) found that there was no simple relationship between measured needs and the level of care provided, or between the level of care provided and improvement in basic living skills. However, interestingly, cost variations were related to need and to the quality of the housing environment. Kinane & Gupta (2001) have studied the relationship between usage of health services and the costs of providing different forms of health care, noting that although vulnerable residents in the care homes they studied were costing the health services relatively little, there was a need for multi-agency planning at a local level to avoid gross inequities of provision and the problem of localised proliferation of services, which can lead to a transfer of morbidity from the original catchment area.
Box 2⇓ lists the types of accommodation that should be available within any local area, and this list represents the spectrum of accommodation that should be demanded by general and rehabilitation psychiatrists. The need for development in a particular area would depend on current provision across the spectrum, and identifying this need could be seen as a ‘gap analysis’.
The range of residential support which should be available in a locality for people with mental health problems
Unstaffed group homes
Core and cluster (high-dependency) flats
Adult placement schemes
Residential care schemes
Nursing homes registered to take psychiatric patients
24-hour nursed NHS accommodation
Acute and longer-term in-patient care
Medium secure units
Quality issues in supported accommodation
In the new, dispersed, multi-agency system of community care, the issues of care quality and staff training are now more complex and also, arguably, more important. There is often little information about the quality of care in residential settings in the public, private or voluntary sectors, and little or no coordination of residential care at the local level. Allen et al(1989) showed that care can be highly institutionalised in ostensibly forward-looking hostel-based settings and Shepherd et al(1996) found that the quality of care in community homes is strongly dependent on the personality and orientation of project leaders. Good leadership and external management support are particularly important in individual units, which can become isolated and develop idiosyncratic styles of practice. Real problems can develop when trying to supervise poorly paid, untrained staff in relatively isolated settings, and there is a potential for abuse, particularly of older people. Approaches to reduce isolation and develop professional/supervisory support systems across units of this type are likely to reduce these risks.
In-patient facilities provided by the NHS are now so clogged up with the most difficult patients (Carson et al, 1989) that they have limited capacity to respond to crisis in patients in private-sector accommodation, who may need to be admitted following a deterioration in their level of disability or disturbance. This can lead to friction between statutory and voluntary/independent providers, which perceive a lack of support from statutory services. These issues can only be challenged by real partnership working, which might include NHS outreach support to independent-provider units and shared training programmes. Similarly, if problems of stigma are to be tackled effectively, community-based units must adopt a consistent and prolonged approach to engaging the minority of potentially destructive members of the general public who have negative, prejudiced attitudes (Penn et al, 1994).
Key factors for ensuring the quality of care in community units are summarised in Box 3⇓.
Key elements of high-quality care in community units
Ongoing monitoring of care practices through a supported management system
Staff training that includes risk management, dealing with challenging behaviour and the concept of expressed emotion
Care packages that are patient-focused, promote opportunities and give basic choices about how and where to live
Proactive attention to, and energetic approaches to dealing with, stigma where this affects a unit’s residents
Real partnership working between statutory and other providers
The encouragement of a culture that promotes strong leadership and an outward-looking, evidence-based approach within individual units
In their review, Shepherd & Murray (2001) stated powerfully that ‘housing should be at the centre of community psychiatry’. In any local area a variety of forms of supported accommodation is required, including 24-hour nursed care and various lower levels of support. There is limited evidence of the value of core and cluster (high-dependency) flats with variable levels of support, although these are often more attractive to potential residents. However, in many areas there appears to be a lack of clarity about overall levels of need, or even the extent and type of specialist accommodation provided for those with severe mental illnesses. The Government’s recent Supporting People initiative may provide an opportunity for more systematic assessment of need and provision in this important area, which has a direct impact on all other aspects of mental health care provision.
Multiple choice questions
The ‘new long-stay patient’:
is a concept developed by Mann & Cree
refers to patients who have remained in the long-stay wards of asylums
no longer exists
may be placed on an acute psychiatric ward
can generally be discharged to fully independent accommodation.
The following factors may be important in determining the quality of supported accommodation:
good inter-agency working
links to day-service provision
support systems for staff
the existence of an effective complaints system
Research into supported accommodation for people with mental illnesses:
includes a large number of randomised controlled trials
indicates that 24-hour nursed care is not cost-effective
includes an important contribution from the TAPS study
has consistently questioned its effectiveness
has tended to be descriptive and qualitative regarding core and cluster (high-dependency) flats.
24-hour nursed care:
is always directly provided within the NHS
units are typically 6–12 bedded
is provided in all areas of the UK
was recognised as deficient in some areas of the UK in recent Department of Health guidance
units may be registered to take patients detained under the Mental Health Act 1983.
Types of supported accommodation for people with severe mental illnesses include:
adult fostering schemes
private nursing homes
core and cluster flats
24-hour nursed-care units
- © The Royal College of Psychiatrists 2004