Early pioneers of admission diversion: Arie Querido and Joshua Carse

The wider context for CRTs is the deinstitutionalisation movement, the quest for effective alternatives to hospital in-patient care that has dominated mental health policy, service planning and service research for much of the past half century. Diversion from acute hospital admission has been an element in deinstitutionalisation from its beginning: the first admission-diversion service to stimulate widespread interest and discussion was that established by the psychiatrist Arie Querido in Amsterdam in the 1930s (Reference QueridoQuerido 1935). Querido instituted a city-wide system of home visiting by a psychiatrist and a social worker whenever a patient was referred for admission, with an alternative treatment plan, sometimes involving follow-up home visits, implemented whenever possible. In the UK, community visits in crises were instituted in some areas as early as the 1950s, as in the Worthing experiment: initiated in 1956, this involved home visits by a psychiatrist and a social worker to all those referred for acute admission, and was reported to result in falls in admissions to two local hospitals of 55% and 79% (Reference Carse, Panton and WattCarse 1958).

The Denver system and the Barnet Family Service

The early home-visiting initiatives in the UK generally formed part of a community-oriented reform of working practices throughout the psychiatric services of a catchment area – they were not separate teams dedicated solely to managing crises and preventing admission. In the 1960s and early 1970s, specialist admission teams with a distinct identity, staff team and budget were established and evaluated in various parts of the English-speaking world. One of the most extensive of these initiatives was the network of services developed by Paul Polak in Denver, Colorado, in the 1970s (Reference Polak and KirbyPolak 1976). Polak's innovations included a team which assessed all individuals referred for admission at home and offered 24-hour home treatment whenever feasible, integration of hospital and community treatment services, and a network of family sponsor homes, in which families were paid to accommodate up to two patients in crisis, supported by the home treatment team. Distinctive characteristics of the Denver system included the elimination of clinical staff offices suitable for interviewing patients, so that all patient contact had to take place in homes or other community settings. Dennis Scott's Barnet Family Service in north London broke new ground as a specialist team dedicated to admission diversion and home treatment, although patients could not be seen very frequently (Reference Scott and StarrScott 1985).

The Madison and Sydney teams

Some of the working practices of current CRTs can be traced back to two services that share the somewhat confusing distinction of being cited in support of two different innovative models: CRTs and assertive outreach teams (AOTs). The Training in Community Living service established by Leonard Stein and colleagues in Madison, Wisconsin, in the late 1970s (Reference Stein and TestStein 1980) and the service established by John Hoult and his colleagues in Sydney in 1979 (Reference Hoult, Hall and BrockingtonHoult 1991) resembled current CRTs in recruiting patients at the point of acute admission during a crisis and diverting them wherever possible to home treatment. However, like AOTs and unlike CRTs, the initial Madison and Sydney teams continued to treat people intensively in the community once the initial crisis had resolved, with the long-term goals of improving their stability in the community and their social functioning.

Following these initial experiments, Stein and Hoult both concluded that the crisis treatment function would be better split off from continuing care, as it seemed to them difficult for a single team to have both roles (Reference Johnson, Thornicroft, Johnson, Needle and BindmanJohnson 2008a). The Dane County, Wisconsin, crisis teams were established by Stein in the 1970s and have continued to operate ever since, with 24-hour availability, screening of all patients prior to admission, and provision of visits several times a day if needed over a short period of treatment lasting until the crisis has stabilised and the patient is discharged. Similar teams were established in Australia from the early 1980s and some have survived long term, especially in the state of Victoria, which introduced a requirement for such teams in 1994, prefiguring the NHS modernisation (Reference Carroll, Pickworth and ProtheroeCarroll 2001).

Early UK acute home treatment services

In the UK in the late 1980s and 1990s, community mental health teams were the primary providers of crisis response in the community. Most, however, operated only during ‘office hours’ (09.00 to 17.00), five days a week, which considerably limited their capacity to respond to crises and substitute for acute admission.

Experimentation with models involving more intensive and specialised crisis care began in a few centres. For example, Christine Dean and colleagues established an innovative home treatment service in Birmingham focused especially on the local Asian community. They found that it reduced admissions and was preferred by relatives (Reference Dean, Phillips and GaddDean 1993). At around the same time at the Maudsley hospital in London, the Daily Living Programme was a replication of Stein's and Hoult's original combined crisis and longer-term care models and demonstrated reduced bed use and some clinical benefits over an 18-month follow-up (Reference Marks, Connolly and MuijenMarks 1994).

In 1995, John Hoult, recently arrived from Australia, established the Yardley Psychiatric Emergency Team (Reference Minghella, Ford and FreemanMinghella 1998). This can be seen as the first full UK implementation of the CRT model, in which Hoult drew on his observations regarding the organisational features associated with greatest effectiveness in Australian crisis teams. The model was replicated in a variety of centres, including Bradford and Islington, London, before it was adopted in 2000 as national policy in the NHS Plan (Department of Health 2000). This and the subsequent Mental Health Policy Implementation Guide (Department of Health 2001) mandated the development of 335 CRTs across England. Each was expected to carry a case-load of 20 to 30 at a time, to see around 300 people a year in total, and to be available 24 hours a day, 7 days a week.

The remit of CRTs

The original Mental Health Policy Implementation Guide (Department of Health 2001) suggested that CRTs should be available for adults of working age (18 to 65 years), and also that people with learning disabilities (the term commonly used in UK health services to mean intellectual disability) and primary diagnoses of personality disorder should not be included. However, this raises issues of equity as there is no evidence that these groups would not benefit. Some CRTs extend their client age range beyond 65, and specialist home treatment services for elderly people have begun to develop, with some evidence that they can prevent admissions (Reference Dibben, Saheed and StagiasDibben 2008). Preliminary evidence suggests that CRTs can also prevent admission among people with personality disorders, and many services do work with this group (Reference Cotton, Johnson and BindmanCotton 2007).

Gatekeeping and around-the-clock help

A gatekeeping role, with patients admitted to acute beds only if the CRT has assessed them and agreed that this is necessary, is considered key to success in reducing admissions (Reference Glover, Arts and BabuGlover 2006). The question arises of whether the CRT should ever agree to the admission of a patient whom they have not seen face to face: the default should be that they see every patient, though in practice some flexibility is likely to be needed from time to time if insisting on this would introduce an unacceptable delay in a situation of high risk.

Twenty-four-hour availability is also important if severely ill people are to be managed at home, as carers need to be confident that help is available at any time. However, keeping a community office open 24 hours a day may be impractical if the volume of night-time work is low. To overcome this problem, some CRTs provide night-time cover by small number of staff located in a hospital, often in or near the accident and emergency department, or arrange for staff to be on call from home. Further evidence on the impact of different policies on gatekeeping and 24-hour cover would be useful.

The composition of the team

A full multidisciplinary team seems desirable if a full range of psychological, social and biological perspectives on assessment and interventions is to be available, although in practice some professions (nurses, support workers and doctors) seem to be much better represented than others (occupational therapists and psychologists) (Reference Onyett, Linde and GloverOnyett 2008). Many of the skills required by CRT staff are in any case specific to the CRT worker role rather than to a particular profession (Reference Ramsey, Shaw, Johnson, Needle and BindmanRamsey 2008). In some CRTs, consultant psychiatrist input comes from senior doctors who are also responsible for local community mental health teams and/or patients on the wards; in other cases dedicated CRT consultants are embedded within the team. The former arrangement has some advantages for continuity of care, but dedicated CRT consultants are favoured by many experts, with some evidence that their integration in the CRT may help effective gatekeeping and good joint working with other parts of the system (Reference Middleton, Glover and OnyettMiddleton 2008).

Integration with other healthcare services

Given that they intervene short term with patients whose mental health needs are often very long term, a key challenge for CRTs is to achieve good integration with other parts of the service system. Good communication and working relationships with in-patient services, community mental health teams, casualty department liaison teams and specialist services such as assertive outreach and early intervention are essential.

Crisis resolution teams need to maintain clear delineation of their role, and of admission and discharge protocols and thresholds, but at the same time they must avoid unnecessary conflict and be perceived by other mental health services as helpful and useful (Reference Flowers, Hoult, Johnson, Needle and BindmanFlowers 2008). This is more easily achieved if CRT staff have a very clear concept of what their main tasks and target groups are, and explain their decisions about these as clearly as they can to other professionals.

The position of CRTs in the service system means that they are well placed for identification of perceived failings of other components of the service: when they identify these, they need to resist being unduly critical, but instead work constructively with other services to achieve smooth and effective care pathways. Good relationships with in-patient services are important in order to identify ward patients who are still symptomatic but are now more cooperative and less ‘risky’, making them candidates for early discharge.

Assessment

Assessment is necessarily a core task for CRT practitioners: teams need to have members who are confident in assessing and re-assessing risk, suitability for home treatment, symptoms and their response to treatment, substance misuse, social difficulties that may have triggered or perpetuate the crisis, and psychological and social resources for coping with the crisis.

Engagement

Talking through difficulties with patients and members of their social networks and offering emotional and practical support is another essential component in the role of CRT staff. A particular challenge is the need to maintain warm relationships and conversations that make some progress when shift systems mean that several different people will usually be visiting each CRT client. However, the intensity of contact, with most CRTs offering twice-daily visits if needed, can help in establishing a therapeutic alliance. Simple psychoeducation about mental health problems and their management will usually be included in meetings with both patients and carers.

Practical help

Help with practical problems is a further important focus for many teams, although opinions vary about the extent to which team members should themselves help with practical tasks (Reference Bindman, Flowers, Johnson, Needle and BindmanBindman 2008b). Home treatment requires an acceptable home environment, access to food and money, and freedom from serious social pressures such as the threat of eviction or harm from others living with or close to the patient. For some, hospital admission may in fact be the best short-time respite from a very unsuitable home environment; for others, advocacy with issues such as benefits and housing problems, help tidying up and getting food, and placement in temporary accommodation or in a community residential facility such as a crisis house may facilitate home treatment.

Medication monitoring

Since CRTs focus on people with severe mental health problems, medication will often be an important intervention, especially in trying to reduce quickly the severity of initial symptoms and disturbance. As well as initiating or adjusting prescriptions, an important role for CRTs is in administering and monitoring medication, which is generally feasible on a twice-daily basis when needed.

Wider interventions

Beyond these simple but essential activities of engaging, assessing, monitoring, supporting, educating, and ensuring that appropriate medication is received, a standard array of CRT interventions has not been established, and practice appears to vary depending on the skills, interests and approaches of clinicians and managers in each team. As already discussed, the idea that the antecedents to crises are often social and can more readily be addressed in patients' own homes has been important in the development of CRTs. Many CRTs therefore aim to intervene with patients' social networks in some way, identifying and addressing some of these social triggers.

Reference Bridgett and PolakBridgett & Polak (2003a,Reference Bridgett and Polakb) describe a relatively structured approach to social systems intervention, involving the early convening of meetings of key individuals in the patient's social system so that problems in the system can be identified and the participants encouraged to find solutions.

Other types of intervention that may be useful within CRTs include brief psychological interventions focusing on symptoms or substance use, structured work on relapse prevention or developing crisis plans to be implemented in any future crisis, and interventions focusing on problem-solving or medication adherence.

Initiating complex interventions

An important question is when CRTs should initiate an intervention themselves and when request that it be provided by another professional or agency (Reference Hurcom, Parnham and HorlerHurcom 2008). The very short duration of most periods of CRT treatment make it difficult for CRTs to deliver treatment of any complexity themselves, although there may be scope for them to collaborate with other teams on interventions such as relapse prevention. More evidence on what interventions and ways of working are most effective within a CRT framework would be useful: currently, the model is in many ways a vehicle for service delivery rather than a specific treatment approach.

Evidence: the positive

When CRTs first became national policy, they were criticised for their scanty evidence base (Reference Pelosi and JacksonPelosi 2000), derived mainly from older studies in which neither experimental nor control groups were very comparable with current models. What do we now know about their impact? Some positive findings can be reported. When the model is implemented with relatively high fidelity, including 24-hour cover and gatekeeping, congruent evidence from national bed-use data (Reference Glover, Arts and BabuGlover 2006), naturalistic investigations of the effects of implementing the model within catchment areas (Reference Johnson, Nolan and HoultJohnson 2005a; Reference Jethwa, Galappathie and HewsonJethwa 2007; Reference Keown, Tacchi and NiemiecKeown 2007; Reference Barker, Taylor and KaderBarker 2011) and a randomised controlled trial (Reference Johnson, Nolan and PillingJohnson 2005b) shows that reductions in numbers of admissions occur, with accompanying falls in costs (Reference McCrone, Johnson and NolanMcCrone 2009a,Reference McCrone, Johnson and Nolanb). Greater satisfaction among service users has also been reported in some CRT studies (Reference Johnson, Nolan and HoultJohnson 2005a,Reference Johnson, Nolan and Pillingb; Reference Winness, Borg and KimWinness 2010; Reference Barker, Taylor and KaderBarker 2011), although we know little about the views of carers regarding the current UK model. Where they have been examined, other outcomes, such as symptoms and social functioning, appear similar after an episode of acute care with or without a home treatment team involved. The workforce implications of this reorganisation of the acute care system are also important: a survey of London CRT staff was reassuring, suggesting fairly good satisfaction and low burnout (Reference Nelson, Johnson and BebbingtonNelson 2009), subsequently confirmed in a national investigation of staff morale (Reference Johnson, Osborn and ArayaJohnson 2012).

… and the not so positive

Alongside these accounts of reasonably successful services, considerable reservations have emerged regarding the nationwide implementation and impact of the model. Two reports on the national picture by bodies with responsibility for monitoring public sector provision have suggested considerable variation between trusts in ways of working and effectiveness (National Audit Office 2007; Healthcare Commission 2008), and recent reports by the national mental health charity Mind (2011) and the Centre for Social Justice (2011), an influential think tank, have questioned the effectiveness of these teams in meeting service users' needs.

Implementation of the gatekeeping principle appears to have varied greatly between trusts. Discontinuities of care are reported at various points in the acute care system, and service users and carers, although in the main positive about the possibility of receiving care at home, also report some unsatisfactory experiences of CRT care. These relate especially to relationships with staff, with reports that contacts are fleeting and superficial and that too many staff are involved in each episode of care (Reference Hopkins and NiemiecHopkins 2007), and to the range of interventions offered, with complaints that teams focus too much on prescribing and dispensing medication and too little on emotional and practical support (Reference Lyons, Hopley and BurtonLyons 2009).

A common feature of the cited national reports is an emphasis on the need to develop integrated acute care pathways within catchment areas, with good continuity between components in the system. In Norway, the other country where CRT introduction has been national policy for several years, implementation studies suggest both substantial divergence from the English model, with a greater focus on less severe disorders, and considerable variation between Norwegian teams (Reference Hasselberg, Grawe and JohnsonHasselberg 2011).

Recent innovations

The variable implementation of this model and the doubts that persist about it are not surprising, given the speed at which this innovation has been introduced nationwide, its origins in many areas in government policy rather than local enthusiasm or assessment of needs, and the limited evidence about how best to implement CRTs and what interventions to deliver within them. Innovative service development work in various parts of the country indicates that the model is still evolving.

Linking the model closely with community residential accommodation is a fairly common innovation: capacity to manage relatively severe crises outside hospital may be substantially greater when a 24-hour staffed residential facility and a full multidisciplinary CRT work closely together (Reference Johnson, Lloyd-Evans and MorantJohnson 2010). Integration of day services and CRTs has also been described and may address the criticism that patients receiving CRT care may be spending much of their days alone between team visits and may have relatively little structure to their days or activity (Reference Allen, Blaylock and MieczkowskiAllen 2009). Strengthening links with the in-patient ward, for example by rotating staff between ward and CRT, is another innovation that may improve the continuity of local acute care pathways, especially in the increasing number of areas with acute in-patient assessment units that aim, whenever feasible, to achieve early discharge, often with CRT support.

What of the future?

Following a decade of gradually rising investment (Department of Health 2011), the establishment of a national network of CRTs has been a significant achievement. Nonetheless, there remains considerable scope for further development and evaluation of the model and methods for implementing it. Getting CRTs right is very important for the NHS if this is to remain the predominant way of delivering acute care in the community, although a new UK government and widespread changes in health policy mean that the model is no longer mandatory, and local variations and even decisions not to maintain it may in future be more widespread than before.

Given the effects of model services on admissions, costs and service user satisfaction, the model is also a promising one for implementation in other high-income countries, making it all the more important to achieve an understanding of how to optimise it within an effective overall acute care pathway. One potential way forward in further specifying and standardising the model is to adopt the methods of the evidence-based practices programmes increasingly prevalent in the USA (Reference Drake, Bond and EssockDrake 2009). This involves developing a fidelity scale intended to measure adherence to best practice in delivering the model, usually accompanied by an implementation resource kit of materials and methods to support achieving high fidelity on the fidelity scale. A current research programme, Crisis Team Optimisation and Relapse Prevention (CORE), on which I am the lead, aims to develop and test a fidelity scale and implementation resource kit of this kind: it is hoped that this will support implementation of CRT in its most effective form, both in the UK and internationally.