Notes
Article history
The research reported in this issue of the journal was funded by the HTA programme as project number 14/51/01. The contractual start date was in August 2014. The draft report began editorial review in February 2015 and was accepted for publication in July 2015. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HTA editors and publisher have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the draft document. However, they do not accept liability for damages or losses arising from material published in this report.
Declared competing interests of authors
none
Permissions
Copyright statement
© Queen’s Printer and Controller of HMSO 2016. This work was produced by Paton et al. under the terms of a commissioning contract issued by the Secretary of State for Health. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.
Chapter 1 Background
Definition
There are many definitions of mental health crisis. For example, a pragmatic service-oriented approach:
Crisis brings the service user to the attention of crisis services for example through the relapse of an existing mental health condition. This results in a substantial impact on the life of the service user and their social network.
Joint Commissioning Panel for Mental Health1
Various other approaches include self-definitions of mental health crisis (i.e. the service user themselves defines their experience and recovery), a risk-focused definition (i.e. people at risk of harming themselves or others), theoretical definitions, and negotiated definitions (i.e. a decision reached collaboratively between service user, carer or professional).
Challenges of mental health crisis services
It is widely acknowledged that the quality and accessibility of care for people in crisis is highly variable. 2 Although many people in mental health crisis experience high-quality care and support when they need it, there are also a number of occasions where people find services do not respond well to their needs. 3
It is also often recognised that emergency services related to mental health treatment compare unfavourably with those related to emergency physical health services. 2 Therefore, it is a priority to improve crisis services for people with mental health problems in order to meet the objectives of parity of esteem set out in the NHS mandate.
The NHS mandate for 2014/15 identified several objectives for mental health crisis services including accessibility and quality of emergency mental health care, improving liaison psychiatric services, and for every community to plan to have sufficient resources available for crisis care. 4
In response to these issues, the Mental Health Crisis Care Concordat, Improving Outcomes for People Experiencing Mental Health Crisis, was developed which highlighted the need for a review of urgent and emergency care, with a focus on models of care for people in mental health crisis. 3
It has long been recognised that improvements are required in how health services, social services and police forces work together. Where problems exist, they often happen where these services intersect, concern how the different professional groups interact with one another and transfer from one service to another. 3
The Mental Health Crisis Care Concordat also highlighted the disproportionate rate at which some communities reach crisis point or access mental health services through involvement with the criminal justice system. People in black and minority ethnic (BME) communities were detained at higher rates under the Mental Health Act 19835 and a higher proportion admitted to hospital. 3 An independent inquiry into crisis care, carried out by Mind in 2010/11,6 as part of a crisis care campaign, suggested that people from some BME groups seemed to be treated more neglectfully or coercively in the crisis care system than other people. The inquiry also highlighted certain barriers that might be faced by different ethnic groups in relation to accessing crisis care:
-
There is variable access to crisis resolution and home teams, with lowest referral rates identified for Indian, Bangladeshi and Chinese people.
-
Once assessed by a crisis resolution and home team, BME groups are generally more likely to be admitted to hospital, particularly black Caribbean people. 6
-
Another population where there was need for improvement in crisis services is women with antenatal and postnatal mental health problems. For example, most women with psychosis in the postnatal period will experience readmission owing to a non-puerperal recurrence. For example, in a Danish study, a year after discharge 50% of women had been readmitted and 98%, in total, experienced another episode of psychosis. 7
The National Collaborating Centre for Mental Health (NCCMH)7 conducted a survey of primary care trusts in England and Wales, which found that just over half the trusts had an identified clinical lead or manager and a similar number had a protocol for the care of women with an existing disorder. It estimated that only 25% of primary care trusts had a fully developed and implemented policy for antenatal and postnatal mental health care.
The literature on antenatal and postnatal mental health services has been reviewed comprehensively in a forthcoming National Institute for Health and Care Excellence (NICE) guideline; therefore, these data will not be reviewed here (for further details see Antenatal and Postnatal Health. Clinical Management and Service Guidance by NCCMH7).
Crisis care pathway
The Crisis Concordat proposes four key stages of the mental health crisis care pathway, which are discussed in more detail in the next sections:
-
Access to support before crisis point.
-
Urgent and emergency access to crisis care.
-
Quality of treatment and care when in crisis.
-
Recovery and staying well/preventing future crises.
Access to support before crisis point
Brief policy background in England
The London Strategic Clinical Network2 has recently put forward four key standards on access to crisis care support that reflect NICE quality standards on service user experience of adult mental health services (see Quality Standard for Service User Experience in Adult Mental Health by NICE8):
-
Mental health crisis telephone helplines: available 24 hours a day, 365 days a year with links to out-of-hours alternatives such as NHS 111.
-
Self-referral: people have access to all information needed on crisis management including self-referral.
-
Third-sector organisation: engagement with voluntary organisations offers services that complement or are in place of those provided by the statutory sector.
-
General practitioner (GP) support and shared learning: training should be provided for GPs, practice nurses and other community staff on crisis assessment and management.
More broadly, there are statutory requirements of the Local Government and Public Involvement in Health Act 20079 for foundation trusts and local authorities to conduct joint strategic needs assessments of the health and well-being of their local community. 10 This includes targeting the causes of health problems such as mental health crisis in order to intervene early before crisis point.
Brief summary of current practice
Telephone triage services have been set up in England (previously NHS Direct, NHS 111) and in other countries, such as Australia and Canada.
Elsom et al. 11 reported service user experiences of a telephone mental health triage service in Australia. Most service users (67%) and carers (66%) were either satisfied or very satisfied with the triage service. The majority reported being treated with dignity and respect, receiving prompt attention and that there was good communication between themselves and the mental health triage nurse. However, there was need for improvement of the service, with 70% having difficulty accessing the triage service, including difficulty finding the triage telephone number or waiting for a call back from the triage nurse. In addition, only a minority of both service users and carers reported being asked about their preferences (for a summary of other international case examples see Improving Outcomes for People Experiencing Mental Health Crisis by the Mental Health Crisis Care Concordat12).
In terms of current provision, the NHS 111 service is an important national development in aiming to improve access to support before crisis point in England. An evaluation of the benefits of implementing the NHS 111 service is being conducted that includes examining the benefits of sharing crisis plans with NHS 111 and out-of-hours providers, and further evaluations are ongoing in 2015. A number of additional 24/7 crisis helplines exists across the country (see the London Mental Health Crisis Commissioning Guide2 for recent developments) either commissioned or voluntary. In addition, there have been, for many years, a range of listening services that enable access to support before crisis. These include services provided by the Samaritans (www.samaritains.org) and other third-sector organisations using similar models, and student counselling or listening services (such as Nightline; http://nightline.org). Mind6 also lists a number of other forms of support available, such as crisis houses or other informal support services such as drop-in centres or cafes.
Core competencies of mental health telephone triage services include opening the call, performing mental status examination, risk assessment, planning and action, termination of the call, referral, and reporting and documentation. 13
Colgate and Jones14 reported on a telephone-based mental health referral co-ordination process in old age psychiatry in Bridgend, South Wales. The system aimed to identify patient needs in a prompt and accurate manner and provide a preliminary assessment of risk. The advantages of this approach included a single point of access to services, higher degree of consistency in referral patterns and potentially more equitable service delivery. A similar approach was used by that same team in the care home sector, which resulted in a reduction in mental health admissions from local care homes over a 3-year period. 15
Another approach that could impact on access to support before crisis point is to promote mental health awareness and reduce stigma. For example, the Time-to-Change campaign (http://time-to-change.org.uk) began in England in 2009 and has reached 47 million people through advertising on television, radio, newspapers and online.
Good-quality primary care may have the potential to reduce the risk of experiencing mental health crises. However, a recent UK study found that higher Quality and Outcomes Framework scores were associated with greater risk of emergency admissions. It’s unclear whether this association reflects that Quality and Outcomes Framework is a poor measure of quality in primary care, higher-quality primary care does not prevent emergency admissions, or better-quality primary care is more able to identify unmet need for secondary care treatment. 16
Potential barriers to accessing support before crisis
This section does not constitute a comprehensive review of barriers to accessing support before crisis, but aims to briefly summarise some key issues reported in the UK.
A key theme identified by Mind6 was the need to have a ‘place to go’ for respite, safety or other reasons away from home (such as a crisis house) that can help avert a crisis. However, it was commonly reported that such support was rarely available before crisis point.
A more recent report by Mind17 also noted barriers that prevent BME communities access to mental health care, including different frames of reference and understandings of mental health, language differences, cultural taboos in the community and experiences of racism in the wider community.
A report by St Mungo’s18 highlighted that people who were homeless were among the most marginalised in society. They found many examples of homeless people reporting missed opportunities to receive the right help at the right time, which can be particularly challenging owing to the complex inter-related needs of this population. Homeless women (representing approximately 25% of the homeless population in the UK) may experience further challenges in accessing support, where services are often designed for men. St Mungo’s argued for the development of new approaches that better take into account the needs of homeless women in accessing support.
In a review of qualitative studies the NCCMH19 found that some women reported the responsibilities of providing child care as a barrier to accessing crisis support. Similarly, a survey for St Mungo’s17 of homeless women found that, of their clients who were mothers, 79% had their children taken into care or adopted, often resulting in severe trauma. It should also be acknowledged that there are many initiatives in the UK provided by third-sector organisations that seek to improve access, for example through evening sessions for parents.
The Men’s Health Forum manifesto20 also highlighted the difficulties experienced by men in accessing support early. This may reflect a tendency for men to avoid contact with health-care providers, for example in terms of attending GP appointments and attending NHS health checks. Seeking to facilitate engagement with services at the population level may have implications for suicide rates, given that males are three times more likely to commit suicide than females in the UK, accounting for 80% of all suicides in 2012. 21 However, further research is needed to investigate the benefits of such an approach.
It has also been noted that people with complex needs, such as those with personality disorders, or psychosis, and coexisting substance misuse often experience significant barriers to accessing support and services. 3,19 For example, people with a substance misuse problem may sometimes experience being excluded from receiving psychosis treatment and vice versa. Similarly, people with a personality disorder may experience exclusion from treatment services for comorbid conditions such as depression or anxiety.
Other barriers include stigma, lack of acknowledgement of problems and fear of being hospitalised. 9,22
Urgent and emergency access to crisis care
Brief summary of policy background in England
In 2013, Sir Bruce Keogh announced a comprehensive review of the NHS urgent emergency care system in England. 4 The Mental Health Crisis Care Concordat has applied these principles to the urgent and emergency access to mental health crisis care. 3 The Concordat action plans and planning groups aim to link up with (or be part of) system resilience groups to feed into the regular planning of service delivery.
A key aspect of urgent and emergency access to crisis care is providing service users and carers with a single point of access to a multidisciplinary mental health team to contact 24 hours a day, 7 days a week. 3 This also reflects the NICE quality standards8 on service user experience in adult mental health. In particular, quality statement 6 states that people should be able to access mental health services when they need them. This includes:
-
ensuring people in crisis referred to mental health secondary care services are seen within 4 hours
-
service users have access to a local 24-hour helpline staffed by mental health and social care professionals
-
ensuring crisis resolution and home treatment teams (CRHTTs) are accessible 24 hours a day, 7 days a week, regardless of diagnosis
-
ensuring people who have been admitted to a place of safety are assessed under the Mental Health Act5 within 4 hours.
The Mandate for 2014/15 outlines specific objectives for the NHS to improve mental health crisis care:4
-
NHS England to make rapid progress ensuring mental health crisis services are at parity with other health emergency services in terms of accessibility, responsiveness and quality.
-
NHS England to ensure there are adequate liaison psychiatry services in emergency departments (EDs).
-
Every community to have plans to ensure access to mental health crisis services based on principles required by the Concordat.
In addition, there have been increasing concerns regarding the inappropriate use of restrictive interventions for people with mental health problems and others who present with behaviour that challenges. Recent guidance on positive and proactive care has been issued by the Department of Health (DH) to reduce the need for restrictive interventions. 23
Police officers in England and Wales are responsible, under Section 136 of the Mental Health Act (1983),5 to intervene when a person with mental health problems is in immediate need of care or control in a public place (in the interests of that person or the protection of other persons) and take them to a place of safety. 20,24 Section 135 of the Mental Health Act provides a number of examples of places of safety such as a hospital, residential accomodation, a police station. 5
The place of safety should normally be a health-care setting (preferably a mental health-care setting), but only in very exceptional circumstances a police station. 24 A person may be detained in a place of safety for not exceeding 72 hours to enable an assessment by a registered medical practitioner and approved mental health professional.
Brief summary of current practice
Urgent and emergency access to crisis care needs to be a commissioning-led strategy that works across a range of sectors (such as social care, mental health care, acute care, ambulance services, police, etc.). 24 Values-based commissioning has also been suggested as a promising strategy with service users in collaboration with professional groups commissioning mental health services. 25,26
Access to crisis care services varies widely across different countries and health-care systems. Current initiatives in England, such as the Sunderland and South of Tyne Initial Response Team, have been developed to seek to improve access to crisis services. 27 Other examples of current service developments can be found in the London Mental Health Crisis Commissioning Guide. 2
An independent inquiry by Mind6 reported that access to crisis care services across the UK varied in terms of the types of crisis care available, staffing levels and in the range of options available for those who need a safe place to go that is not a hospital.
A review carried out by the Care Quality Commission (CQC)28,29 reported that some places of safety were working effectively, but that others were not fully complying with the recommended national standards. This may impact on people’s experiences relating to point of contact and being detained by police, through to being transported to hospital and then transferred to a place of safety before they are assessed under the Mental Health Act. 28,29 Four key areas were identified by the review: (1) too many places of safety were turning people away, which might result in long waits with the police; (2) too many policies did not include young people, people who are intoxicated, and people with disturbed behaviour; (3) commissioners might not be adequately fulfilling their responsibilities; and (4) too many providers of safe places were not adequately monitoring their service provision, making it difficult to evaluate if provision was meeting the needs of people in their local area. 28,29
Urgent and emergency care services are under great pressure to provide high-quality health care to ensure that people with mental health problems receive the best care. Immediate challenges include capacity issues, such as too many people experiencing a mental health crisis being turned away from health-based places of safety because of a lack of capacity. Some health settings refuse to help people who are intoxicated or exhibiting disturbed behaviour. 30 A report by Her Majesty’s Inspectorate of Constabulary/CQC/Her Majesty’s Inspectorate of Prisons found that, depending on the areas sampled, between 6% and 76% of the total number of people detained under Section 136 were taken into police custody. 20,24 The most common reasons were insufficient staff at health-based place of safety, absence of available beds, person had consumed alcohol or the person was displaying (or had a history of) violent behaviour. 20,24
A recent report by the Independent Commission on Mental Health and Policing identified various shortcomings in police performance regarding systems and procedures as well as behaviours of individual police officers. These shortcomings included Central Communications Command failing to deal effectively with calls in relation to mental health, lack of mental health awareness among staff and lack of training in suicide prevention. 30 In addition, police stations were not designed with the needs of people with acute mental health problems at the forefront: the physical environment might be seen as antitherapeutic, and there could be some delay in contacting appropriate mental health professionals. 25,31
Improvements in multiagency working between police and health and social care professionals are a key recommendation of recent reports. 24,30 The implementation of a crisis intervention team (CIT) model for police officers has increased over the past decade, particularly in the USA and with some case examples in Australia. Street triage is a more common model in England, where health or social care professionals attend Mental Health Act-related calls to provide triage and assistance to police officers. Alternatively, health- or social-care professionals may provide telephone triage (see Edmondson and Cummins32 for a recent example in Oldham). The aims of these approaches are to help police officers in dealing with apparent mental health problems and to reduce the number of people taken into police custody under Section 136 as a place of safety.
A further concern relates to the much higher proportion of black service users experiencing police involvement on the pathway to mental health care, found in several systematic reviews. For example, Anderson et al. 33 found a twofold increased odds of police involvement for black service users compared with white service users [odds ratio (OR) 2.05, 95% confidence interval (CI) 1.63 to 2.59]. However, a recent UK study of 863 individuals by Singh et al. 34 found that when age, diagnosis, risk and level of social support were accounted for, ethnicity no longer remained a predictor of Mental Health Act detention.
Brief summary of service user experience
The NCCMH19 conducted a review of 133 qualitative studies on the experiences of service users with a range of mental health disorders. They identified waiting times as a key barrier to accessing support. Many service users described having to wait too long for an assessment to receive acute mental health treatment, with some feeling they had to resort to violence to receive treatment. 17
The NCCMH19 found that health-care professionals were perceived to be a barrier to accessing services by some and showed a lack of willingness to provide psychological support. Mind6 also reported that service users felt ‘batted away’ or ‘deflected’ from receiving support when needed.
In addition, the NCCMH19 also included data from service users reporting their experience of detention under the Mental Health Act. For example, some felt they were not provided with sufficient information and on occasion were unaware they were detained until trying to leave the inpatient ward.
A recent qualitative study,22 conducted in the UK, investigated the experiences of people who had been involuntarily admitted to psychiatric treatment. The accounts of those who retrospectively agreed that their involuntary admission was right were analysed separately from those who retrospectively thought their involuntary admission was wrong.
Both groups agreed that they did not receive sufficient information and involvement in their treatment decisions.
Those individuals who thought their involuntary admission was right acknowledged that they found it difficult to recognise they needed help when unwell. In addition, they felt they were at risk of harming themselves or others and saw their admissions as necessary to prevent further harm.
Those who thought their involuntary admission was wrong believed they needed help from services, but considered it unnecessarily coercive and an unjust infringement of their human rights.
Mind6 and the NCCMH19 both found that people who self-harmed often had traumatic experiences accessing care in EDs, including some staff refusing to use anaesthetic when stitching self-harm wounds. However, there were also positive experiences of supportive staff, particularly in psychiatric liaison services.
Mind6 conducted an inquiry into crisis services. A key factor mentioned by service users was the need for a ‘timely and effective response’, 24-hour help, before reaching crisis point to avoid escalation of a crisis. Similar findings were identified in a review of qualitative studies conducted for the purposes of the NICE guideline on experience of mental health services. 19
Mind6 found a commonly cited problem was difficulty making contact as a result of either not being able to get through to the crisis teams by telephone or having to wait a long time for someone to visit. Some people found it difficult to express themselves over the telephone, particularly during crisis, and were afraid they would not receive a helpful or supportive response. Therefore, a number of people wanted other options of contact such as text lines.
Quality of treatment and care when in crisis
Policy background in England
The Crisis Concordat states that the function of this element of the care pathway is to provide support and treatment by the right people with the right skills in a setting that suits the service user and their needs. 3
The London Strategic Clinical Network commissioning guide focuses on two key areas of treatment provision – (1) crisis housing, and (2) CRHTTs – which are primarily aimed at alternatives to inpatient treatment. These priorities on alternatives to inpatient admission reflect some of the downsides of inpatient care including the unpopularity of overcrowded wards, involuntary aspects of mental health care provided there and high costs. 35
There have also been increasing concerns regarding the inappropriate use of restrictive interventions for people with mental health problems who require crisis care. Recent guidance on positive and proactive care has been issued by the DH to reduce the need for restrictive interventions. 23
Current practice: alternatives to inpatient treatment
Home-based alternatives offer the potential benefits of greater autonomy and better preservation of social functioning than if a service user was admitted. In addition, this may enable service users to develop skills, such as involving social networks, for coping with future crises in the community. 35 It has been a national requirement since 2000 for all trusts to develop CRHTTs, although it is no longer mandatory to provide these services. Standards for home treatment teams have recently been developed by the Royal College of Psychiatrists’ Centre for Quality Improvement. 36
The National Confidential Inquiry into Suicide37 found that the average rate of suicides fell by 18% between the first and last 2 years of its data. However, the suicide rate (14.6 per 10,000 under crisis care compared with 8.8 per 10,000 admissions) was almost twice as high as that found in inpatient treatment. 37 There was also evidence that people with greater social deprivation were less likely to have positive outcomes when receiving care from crisis resolution teams (CRTs). 38,39
Home treatment during a crisis may not always be possible; therefore, other community residential services (such as crisis houses) may be of benefit for people in crisis and are recommended by NICE as an alternative to inpatient admission. 35 Crisis houses are popular with service users and are often viewed as more calm and personal than inpatient wards. 6
Current practice: inpatient treatment
Although service users may often prefer alternatives, inpatient treatment is still sometimes needed and requested when providing crisis treatment. Bowers et al. 40 argued that, based on current UK practice, there are four elements that distinguish inpatient treatment from that provided in community services: (1) legitimate authority; (2) presence; (3) containment; and (4) treatment and management.
Once admitted, the service user is under the legitimate authority and control of the inpatient staff. This authority provides staff with the ability to readily persuade and sometimes compel service users to engage with treatment, adhere to medication, etc.
The 24-hours-a-day, 7-days-a-week presence of staff (which Bowers et al. 40 referred to as ‘presence+’) has a number of benefits, including providing continuous multidisciplinary care and the ability to provide more intensive and more risky treatments that require a high level of monitoring.
In addition, inpatient services allow for the containment of the service user in order to reduce the risk to self and others. This is achieved through three main factors: (1) intrusion (breaking usual norms of privacy, bodily integrity, etc.); (2) separation (from people or objects); and (3) restriction (of freedom of physical movement).
Treatment and management includes delivering the ‘primary admission task’, which is based on the rationale for admission. For example, if the reason for admission was risk then the primary aim is to keep the service user or others safe. There are also secondary tasks such as being able to tackle long-standing problems that although not necessarily the reason for admission might have an important role in promoting recovery. Finally, staff also must seek to ensure admission does not result in harm or negative side effects for the service user (e.g. institutionalisation, loss of job or welfare benefit problems).
Discharge from hospital can be a particular challenge. According to the National Confidential Inquiry into Suicide, 19% of all patient suicides are accounted for by suicides occurring 3 months post discharge from inpatient care. 37 Most post-discharge suicides occurred 2 weeks after leaving hospital; therefore, reducing self-harm and suicide post discharge is a key priority.
Current experiences of treatment and care in crisis
Crisis resolution and home treatment teams
Mind6 found that there were many positive experiences of CRHTTs, but also a number of frustrations. An expressed concern for those who wanted to go into hospital was that they often saw CRHTTs as obstructive, setting high thresholds for access to hospital. Others mentioned high thresholds for receiving CRHTT services.
The need for multiple assessments owing to a lack of information sharing and access to records out of hours was also a major complaint by service users. This was very hard for service users, who had to demonstrate their need for help multiple times in order to access services.
Inpatient wards
Mind6 reported that many people gave examples of receiving good hospital care. Many spoke of kind and supportive staff, being treated with respect and efforts to make the wards homely and relaxing.
However, the majority of comments about hospitals were negative. Some felt that the wards were just holding and containing them without trying to treat the cause of their problems (e.g. a lack of access to psychological therapies or one-to-one conversations with support workers). Similar themes were identified by the NCCMH19 in their review of the qualitative literature relating to the need for therapeutic relationships during inpatient care and how this was often lacking. The NCCMH19 and Mind6 both found that there was often a perceived unavailability of staff because of either busyness or apathy.
Boredom was also a key source of frustration, and a lack of structure led people to feel anxious (particularly people with learning disabilities). Both Mind6 and the NCCMH19 found that many reported a lack of scheduled activities led to boredom, which led to some service users ‘acting out’. Consistent with this qualitative data, a cross-sectional study in the UK found that more intense programmes of service user activities were associated with reduced self-harm. 40
Mind6 also found that lack of safety was a key concern for many service users. The problems of mixed-sex wards and adolescent service users being treated in adult wards were common causes of stress and anxiety. 6,19
A lack of capacity (lack of inpatient beds and high occupancy rates) was also a significant issue for many service users. 6 This led to overcrowding and frequently having to move from ward to ward.
Crisis houses
There were many positive comments about crisis houses and the need for more to be available. 6 People liked that crisis houses were smaller than inpatient wards and felt calmer and more personal. There was a perception of a better environment and sense of mutual support in contrast to the feelings of boredom and frustration often found on hospital wards.
However, some people mentioned not receiving adequate support when in a crisis house and the need for extended emotional support when leaving. 6
Recovery and staying well/preventing future crises
Brief policy background in England
The two main emphases of the Crisis Concordat on this stage of the care pathway are on crisis planning, and providing services to support recovery and staying well. 3 The London Strategic Clinical Network,2 in their commissioning guide for crisis services, similarly suggested the need for integrated care and adopting a holistic approach to promoting recovery.
The importance of effective transitions between services (e.g. between children and young people’s and adult services, or adult and older adult services) has been noted in the recent CQC standards. 29 They also emphasised the need for integrated working to promote recovery and well-being and to prevent or respond appropriately to crisis.
Current practice
Many of the fundamental concepts of recovery originated from the work of service users, and organisations representing service users, and have been influential on mental health treatment and services. Promoting recovery includes the treatment of the mental health condition to improve quality of life of service users through reduction of symptoms of the mental health condition, preventing relapse and improving social functioning. Interventions recommended by NICE across a range of conditions are important for improving recovery. Examples of such services might include early intervention services (EISs) and improving access to psychological treatment services.
Leamy et al. ,41 based on a systematic review of the recovery literature, developed a conceptual framework of personal recovery based on five categories: (1) connectedness; (2) hope and optimism about the future; (3) identity; (4) meaning in life; and (5) empowerment.
Priebe et al. 42 have also conceptually reviewed a number of resource-oriented psychiatric models (such as befriending, peer support, systemic family therapy, therapeutic communities and open dialogue) that seek to tap into the strengths of service users and to utilise their positive personal and social resources.
These strength-based approaches are often provided by third-sector organisations, but are also available in traditional NHS services. Peer support is a particularly important model of promoting recovery, and such programmes have been available primarily through third-sector organisations for many years in the UK, USA, Canada, Australia and New Zealand. Peer support is popular among service users and many call for further availability of such services (e.g. Mind6).
Supported employment [such as individual placement and support (IPS)] interventions are also becoming more widely implemented in England and across the UK. This is reflected by NICE guidance, which recommends supported employment. 43 The Centre for Mental Health has also selected 13 sites across England as Centres of Excellence to act as exemplars of how to implement interventions to support people with mental health problems into employment.
In addition, the IPS in Improving Access to Psychological Therapies (IAPT) pilot study is being conducted in four areas of England (Durham/Tees Valley, Wolverhampton, Shropshire/Telford, and Sussex) before being implemented more widely.
Equality of access
Leamy et al. 41 found substantial similarity in defining recovery in both ethnic minority and majority populations, but there was a greater emphasis on spirituality/religion and collectivist notions of recovery in ethnic minority accounts of recovery. The importance of spirituality and stigma was also found in service user-led qualitative studies of the experience of black women44 and ethnic minority men and women. 45 A particularly strong theme that emerged from Kalathil’s44 study was the contribution of oppression and racism to mental distress experienced by black women. Moreover, participants expressed the need for recovery approaches to include support in overcoming such oppression as part of a holistic package of treatment. The impact of discrimination on refugees and other ethnic minority groups was also found in another user-led qualitative study. 45
There was consistent evidence that both mental distress and recovery needed to be interpreted within the context of identity, race and culture rather than within the conventional Western medical approach. Service users felt discouraged to talk about the influences of race and sexuality [particularly lesbian, gay, bisexual and transgender (LGBT) participants] on their mental distress. 45
Differences between ethnic minority and majority communities in terms of access and engagement with mental health and primary care services have received much attention. A number of systematic reviews have found that black service users were less likely to experience GP involvement and more likely to experience police involvement than white service users (e.g. Anderson et al. 46), but no differences were found between service users from an Asian background and white service users. 33 People from BME communities were also less likely to receive psychological interventions. 17 In a survey by Mind,17 only 10% of BME respondents felt that their cultural needs were taken into account when services were offered (although for most this did not matter).
Experiences of promoting recovery
A recent service user-led study was conducted by Gould47 on service users’ experience of the 2008 Care Programme Approach in promoting recovery, using questionnaires (n = 81) and focus groups (n = 22).
One of the key findings was that most service users thought there was a large discrepancy between what they defined as recovery and what professionals defined as recovery and that this was unhelpful.
Female participants, as a whole, were less satisfied with services than men, particularly in terms of professionals’ unwillingness to accept non-medical explanations of mental distress. African and African Caribbean men and women were even less satisfied with professionals’ lack of openness to accept non-medical interpretations. Many felt factors such as racial stereotypes and racist treatment affected their care. However, Asian and Asian British participants’ responses did not differ much from participants in general.
Service users advocated an increased emphasis on holistic approaches. They also stressed the need for professionals to move further away from unhelpful emphases on risk and compulsion and to put more focus on rights and control for service users.
Service users mentioned that factors, such as hope, listening, respect, compassion, and humility, when professionals interact with them, were as important as any treatments they received. Although they did report some examples of good practice, they felt there was a great deal of improvement needed.
Chapter 2 Aims and objectives
The aim of this project was to conduct a rapid synthesis of the evidence on the clinical effectiveness and cost-effectiveness of models of care for providing treatment and support for people experiencing mental health crisis.
The project addressed four main objectives regarding the evaluation of models of care at each of the four stages of the mental health crisis care pathway identified by the Crisis Concordat:3
-
access to support before crisis point
-
urgent and emergency access to crisis care
-
quality of treatment and care when in crisis
-
promoting recovery/preventing future crises.
Chapter 3 Review methods
A rapid review of the best evidence was conducted to inform the aims stated above. The review was undertaken systematically following established principles set out in Centre for Reviews and Dissemination (CRD) guidance for undertaking systematic reviews, adapted for the requirements of a rapid evidence synthesis. 48
An advisory group of people was formed with expertise in mental health crisis treatment who oversaw all stages of the review. Expertise in the group included psychiatry, psychotherapy, social care, development of mental health crisis services, mental health research and evidence synthesis.
Hearing service users’ perspectives on mental health crisis models of care was an important aspect of this report. The advisory group therefore included two service users (Nigel Ayre and Ceri Dare) and the Director of External Relations for Mind (Sophie Corlett). Their input was crucial in ensuring that the needs and views of service users were taken into account throughout the project, including designing the inclusion criteria for the review and commenting on the draft report and the Plain English summary. In addition, the Managing Director of National Survivor User Network (Sarah Yiannoullou) provided comments on the draft report.
The methods for the rapid evidence synthesis are summarised in the protocol registered with PROSPERO49 and discussed in more detail below.
Identification of studies
Search for guidelines and systematic reviews
Literature searching was conducted as an iterative process for each of the pathway components. The first stage identified relevant guidelines using the NHS Evidence search interface and the ‘guidance’ filter. Systematic reviews were searched for using electronic databases [e.g. the Cochrane Database of Systematic Reviews, the Database of Abstracts of Reviews of Effects (DARE), the NHS Economic Evaluation Database (NHS EED) and the Health Technology Assessment (HTA) database]. Searches were performed on 25 and 26 June 2014. The PROSPERO register was searched to identify unpublished and ongoing systematic reviews. As an additional check the National Institute for Health Research (NIHR) HTA and Health Services and Delivery Research programme webpages were also scanned. For the full search strategy, see Appendix 1.
Search results were limited to the time period 1999 (the date of the National Service Framework for Mental Health) to June 2014. No language restrictions were applied to the searches. Where a foreign-language article represented the best evidence available it was translated into English. Reference lists of retrieved articles, reviews and evaluations were scanned to identify additional studies.
Search for primary studies on urgent and emergency access to crisis care
The advisory group identified substantive gaps regarding systematic reviews on urgent and emergency access to crisis care. Primary studies on this topic were searched for with no restrictions on study design using MEDLINE, PsycINFO and the Criminal Justice Abstracts from inception to November 2014. Searches were performed on 11 November 2014. No language restrictions were applied to the searches. Where a foreign-language article represented the best evidence available, it was translated into English. For the full search strategy see Appendix 1.
Study selection
Inclusion criteria: population, intervention, comparator, outcome
Access to support before crisis point
-
Population: people with a mental health problem requiring support before crisis point.
-
Intervention: telephone helplines, self-referral by service users for treatment, mental health awareness/antistigma programmes.
-
Comparator: training for GPs and other general practice staff to improve outcomes for people requiring support before crisis point.
-
Outcome: improved access to services before crisis point, self-harm, violence, hospital admission, mental health outcomes.
Emergency access to crisis care
-
Population: people with a mental health problem accessing ED care, police, fire, ambulance or other emergency services.
-
Intervention: service level interventions (including training) to improve service and service user outcomes for people with mental health problems.
-
Comparator: quantitative studies were required to have some form of control, either a separate comparison group or a before-and-after comparison. Studies that only reported quantitative descriptive data were excluded. Qualitative studies were not required to have a comparator, as they were primarily providing data on acceptability rather than clinical effectiveness.
-
Outcome: waiting times, hospital admissions, reduction of use of force/restraint, self-harm, violence, mental health outcomes and service user experience.
Quality treatment and care in crisis
Alternatives to inpatient treatment
-
Population: people experiencing mental health crisis.
-
Intervention: three types of crisis intervention were focused on as alternatives to inpatient treatment:
-
CRHTTs defined as ‘any type of crisis-orientated treatment of an acute psychiatric episode by staff with a specific remit to deal with such situations’ (p. 508). 35
-
Crisis houses defined as ‘residential alternative to acute admission during crisis’ (p. 515). 35
-
Acute day hospital care defined as ‘diagnostic and treatment services for acutely ill individuals who would otherwise be treated in traditional psychiatric inpatient units’ (p. 518). 35
-
-
Comparator: treatment as usual, active control, waitlist, no treatment.
-
Outcome: hospital admissions, reduction of use of force/restraint, self-harm, violence, mental health outcomes and service user experience.
Inpatient treatment
-
Population: people experiencing mental health crisis.
-
Intervention: inpatient mental health treatment.
-
Comparator: no restrictions were applied.
-
Outcome: hospital admissions, absconding, aggression, conflict, reduction of use of force/restraint, self-harm, violence, mental health outcomes and service user experience.
Promoting recovery/preventing future crises
-
Population: people with mental health problems. Though people who have experienced a mental health crisis are of primary relevance, we acknowledge that most studies will not have reported this as part of their inclusion criteria.
-
Intervention: service-level interventions to improve service utilisation and service user outcomes for people with mental health problems.
In addition, in discussion with the advisory group it was judged appropriate to broaden the scope of interventions to include individual-level interventions to promote recovery and reduce the risk of relapse. This is to reflect that many of the interventions for promoting recovery suggested by the Crisis Care Concordat are conducted at the individual level. This is an extremely broad scope with almost all interventions for mental health conditions potentially relevant. Therefore, in order to make the review manageable within the time and resources of the project we limited such interventions to those recommended in relevant NICE guidance.
We also examined the effectiveness of interventions to improve access and engagement to mental health treatment for BME groups.
-
Comparator: treatment as usual, active interventions, waiting-list and no treatment.
-
Outcome: hospital admission, admission under the Mental Health Act,5 self-harm, violence, relapse, other mental health outcomes (the primary outcome identified in the systematic review), quality of life, employment and education.
Inclusion criteria: study design
Relevant evidence was included in the synthesis according to the following hierarchy (with preference given in ascending order (1–4):
-
Guidelines: guidelines produced or accredited by NICE. This included UK guidelines produced by NICE and by UK bodies accredited by NICE such as the Royal College of Physicians. It also included guidelines produced in English by non-UK guidance producers who had received NICE accreditation.
-
Systematic reviews of reviews.
-
Systematic reviews of primary studies and economic evaluations.
-
Good-quality primary studies: when no relevant guidelines, review of reviews, or systematic reviews of primary studies were not available we included primary studies [both randomised controlled trials (RCTs) and non-RCTs] and economic evaluations.
Data extraction
Studies were managed using EndNote X7 (Thompson Reuters, Philadelphia, PA, USA) and data were extracted in review software (EPPI Reviewer 4.0, EPPI Centre, London, UK). Data extraction forms were designed by two researchers, piloted on a small selection of studies and adjusted as necessary. Data extraction was undertaken by one researcher and checked by another, with discrepancies resolved by consensus or recourse to a third researcher, if necessary. Where necessary, authors were contacted for missing or unclear data.
Quality assessment
Quality assessment of guidelines was based on their NICE accreditation; NICE-accredited guidance meets criteria set out in the Appraisal of Guidelines for Research and Evaluation statement. 50 In addition, we utilised the Grading of Recommendations Assessment, Development and Evaluation (GRADE)51 ratings concerning strength of evidence provided in the guidelines. Assessment of systematic reviews was based on using the Risk Of Bias In Systematic reviews tool,52 designed for the critical appraisal of systematic reviews as well as existing critical appraisals provided by DARE and NHS EED. For the appraisal of reviews of reviews, we adapted the criteria used to assess systematic reviews. The quality of primary studies was assessed using criteria appropriate to the study design: controlled trials were appraised using the Cochrane risk-of-bias tool, the Critical Appraisal Skills Programme tool53 was used for qualitative studies and the NICE tool for prognostic studies was used for studies without a control group. 54
Quality assessment was carried out by one researcher and checked by a second, with discrepancies resolved by consensus or recourse to a third researcher, if necessary.
Data synthesis
A narrative synthesis was structured by the pathway component addressed and the type of care model assessed. The generalisability of the studies to UK practice was considered in relation to both the populations included and the national and local service-level context of model delivery.
The type and range of evidence identified precluded meta-analysis. However, where possible we assessed heterogeneity measured either by the I2 or Q-statistic as well as visual assessment of the variability of effect estimates and overlap of CIs.
A summary of the most relevant outcomes at each stage of the pathway formed part of the synthesis. Cost-effectiveness evidence was included in the synthesis at each stage of the pathway, but no de novo model of cost-effectiveness was developed.
Following the evidence synthesis, where appropriate, we identified key gaps in the literature and provided recommendations for further primary research.
Chapter 4 Description of studies
The electronic search for reviews and guidelines was performed on 24 June 2014 for guidelines and on 25 and 26 June 2014 for systematic reviews, yielding 2750 records. After removal of duplicates (n = 23) articles were screened and 191 articles were identified as being potentially relevant to the review (Figure 1 contains study flow information). A third reviewer screened titles and abstracts to confirm study selection, and this resulted in 111 articles being excluded. Fifty-two full-text articles were obtained for full assessment by two reviewers. A further 16 articles were excluded on full assessment owing to the evidence either not being the most relevant to the review or not being of the highest level.
Nine UK guidelines produced by NICE were identified and included in the review as being relevant to mental health care in England and Wales. 35,55–62 Guidelines were published between 2009 and 2014. Mental health conditions, covered by the guidelines, included psychosis in adults, psychosis in young people, psychosis and substance misuse, bipolar disorder, self-harm, borderline personality disorder (BPD), depression in adults, depression in people with chronic physical health problems, and experience of mental health services.
Assessment of the nine guidelines identified comprehensive and up-to-date appraisals of current evidence for many aspects of the review. Therefore, of the 27 reviews identified as relevant to the review criteria, 26 were published46,63–87 and one was unpublished [Professor Sonia Johnson and Dr Brynmor Lloyd-Evans, University College London (UCL), 26 January 2015, personal communication]. Of these 27 relevant reviews, seven were included in the synthesis to supplement the findings of the guidelines, six63–66,70,84 were published and one was unpublised (Professor Sonia Johnson and Dr Brynmor Lloyd-Evans, personal communication). Characteristics for these reviews are provided in Appendix 2. Risk-of-bias assessments for systematic reviews and reviews of reviews are summarised below in Table 1. As can be seen, most reviews were rated as at low risk of bias, although several were rated at an unclear risk of bias, because of a lack of reporting.
Study | Study eligibility | Study selection | Data collection | Synthesis and findings | Review risk of bias |
---|---|---|---|---|---|
Archer et al., 201263 | Low | Low | Low | Low | Low |
Bowers et al., 201464 | Unclear | Unclear | Unclear | Unclear | Unclear |
Professor Sonia Johnson and Dr Brynmor Lloyd-Evans, personal communication | Low | Low | Low | Low | Low |
Ibrahim et al., 201470 | Low | Unclear | Low | Unclear | Unclear |
Lloyd-Evans et al., 200966 | Low | Low | Low | Low | Low |
Shepperd et al., 200965 | Low | Low | Low | Low | Low |
Stoffers et al., 201284 | Low | Low | Low | Low | Low |
Following the search for guidelines and systematic reviews, consultation with the advisory group identified substantive limitations regarding interventions to improve urgent and emergency access to crisis care. Therefore, we conducted a search in November 2014 for primary studies, which identified 4657 records (after duplicates removed). Full-text articles were obtained for 64 records and 15 studies were included in the synthesis (Figure 2 shows the study flow diagram). See Appendix 3 for full study characteristics and risk-of-bias assessments.
Chapter 5 Results of the review
Access to support before crisis point
Telephone helplines
Although reviews of telephone triage across a range of conditions are generally supportive of telephone triage (e.g. CRD,88,89 Blank et al.,90 Huibers et al.,91 Ismail et al. 92 and Ramos-Rios et al. 93), we did not identify any systematic reviews relevant to access to mental health support before crisis point.
Early detection programmes for reducing duration of untreated psychosis
Early access to interventions before crisis point is an important aim of the Crisis Concordat. Early detection programmes aim to reduce the duration of untreated psychosis (DUP), which is associated with poor clinical outcomes and poorer quality of life at first contact with services.
The NCCMH35 updated an existing review which resulted in 13 studies of 10 early detection programmes (five multifocus public awareness campaigns, three GP education programmes, one specialist EIS and one online education campaign for parents of high school children).
There was no convincing evidence that early detection programmes improved access to support before crisis point. All three GP education programmes impacted on GPs’ referral behaviour including being more likely to refer people with first-episode psychosis (FEP) to mental health services, and a reduced time from first contact with GP to referral to EISs. However, these changes in referral behaviour did not translate into a reduction of the DUP. The NCCMH35 concluded that there was no evidence of benefit for any type of early detection programme in reducing the DUP or increasing presentations of people with FEP to treatment services. We rated the quality of evidence according the GRADE criteria as low because of limitations regarding the risk of bias of the included studies.
Training health-care professionals who work with people who self-harm
There was no convincing evidence that training health-care professionals improved outcomes for people who self-harm. In addition, most studies did not have sufficient controls and so were vulnerable to confounding.
The NCCMH56 found 18 studies (14 before-and-after studies and four controlled studies) on training health-care professionals who work with people who self-harm. Most studies found a positive effect on knowledge, skills, attitudes and the psychological impact of suicide and self-harm based on self-report data from health-care professionals receiving the training. However, most studies did not report data more directly related to service user outcomes.
There were two studies in the NCCMH review56 that included training of GPs or general practice staff and provided data on either change in behaviour/practice of staff or service user outcomes. One study examined the impact of a 1-day training course for GPs to recognise and respond to suicidal ideation in young people. The authors compared consultations conducted with 203 service users 6 weeks before the workshop and 220 service users after the workshop. Another study found improved recognition rates of psychological distress (e.g. 48% increased odds of recognition based on the General Health Questionnaire-12 score) and suicidality (e.g. 130% increased odds based on Depressive Symptom Inventory – Suicidality Subscale score). However, no significant changes in GP service user management strategies were found.
One study conducted a training programme in mental health, primary care and accident and emergency (A&E) departments in Lancashire, UK. Although there was evidence of improvement in staff’s ability to conduct risk and needs assessments for people who self-harm, a follow-up study found no differences between suicide rates before (8.8 per 100,000 in 1994–6) and after (8.6 per 100,000 in 1998–2000) the training programme.
The NCCMH56 concluded that there were multiple limitations to the included studies. First, there were no RCTs and very few controlled studies. Second, most studies assessed self-reported data from health-care professionals on their knowledge and attitudes with very limited data on whether training impacted on the behaviours of health-care professionals and whether or not there were any benefits of this intervention on service user outcomes. A further limitation from the perspective of this review was that such interventions were not aimed specifically at staff support before crisis point but also at providing longer-term care.
On the basis of these conclusions, we rated the quality of evidence according to GRADE criteria as very low.
Urgent and emergency access to crisis care
Nine studies evaluated interventions aimed at improving access to crisis care in the ED (Table 2). 97–105 Of these, five were assessing the effectiveness of liaison psychiatry models (a psychiatric nurse, psychiatrist, or liaison psychiatry team providing assessments, support, or triage in the ED). 97–101 Two studies involved providing training to ED staff102,103 and two studies examined process changes in the ED. 104,105 Two studies were conducted in the UK,97,100 two in the USA,98,104 two in Australia,99,101 one in Germany,103 one in Canada105 and one in Switzerland. 102
Author Country |
Study design | Interventions | Comparator | Outcomes |
---|---|---|---|---|
Tadros et al., 201397 UK (England) n = 6413 |
Before and after One ED (Birmingham) Follow-up period: 7 months |
RAID liaison psychiatry model | Pre RAID (before intervention implemented) RAID-influenced services |
Length of stay Readmission rates |
Woo et al., 200798 USA n = 200 |
Before and after One ED Follow-up period: 12 months |
Psychiatric emergency service | Psychiatric consultant to ED | Admission to inpatient unit Waiting time Emergency medication provided Seclusion and restraint Readmission within 30 days |
McDonough et al., 200499 Australia n = 604 |
Before and after One ED Follow-up period: 12 months |
Mental health nursing triage service in ED (21.00 to 07.30, 7 nights a week) | Before intervention | Patient contact rate Waiting time |
Sinclair et al., 2006100 UK (Scotland) n = 6097 |
Crossover Two EDs (Glasgow) Follow-up period: 12 months |
Psychiatric nursing service in ED | Before intervention | Waiting times Service user satisfaction |
Wand et al., 2011101 Australia n = 23 |
Mixed methods One ED |
Psychiatric nurse in ED | None | Service user satisfaction and experiences |
Cailhol et al., 2007102 Switzerland n = 479 |
Before and after One ED Follow-up period: 12 months |
ED staff education intervention:
|
Before intervention | Violent behaviour (limited information on how this is defined) of patients admitted for suicide attempt |
Pajonk et al., 2008103 Germany n = 646 |
Quasi-experimental Two EDs Follow-up period: 12 months |
ED staff received an educational intervention | ED staff did not receive an educational intervention | Classification of psychiatric emergency situations Documentation of suicidal behaviour Medication use |
Adams and Nielson, 2012104 USA n ≈ 6000 |
Before and after One ED Follow-up period: 12 months |
Process improvement plan for people entering ED with 30 days of inpatient discharge | Before intervention | Readmission rates |
Strike et al., 2008105 Canada n = 25 |
Qualitative One ED |
Development of psychiatric assessment rooms | None | Service user experiences |
Service-level interventions to improve urgent and emergency access to crisis care in the emergency department
Psychiatric liaison models
Five studies98–102 on psychiatric liaison models were identified in our search. Two were conducted in the UK, two in Australia and one in the USA (see Table 2).
There was largely positive evidence for the Rapid, Assessment, Interface and Discharge (RAID) model conducted in the UK. Tadros et al. 97 investigated the impact of introducing a liaison psychiatry service (RAID) on the mean length of hospital stay. There was a saving of 0.9 bed-days per patient (797 bed-days saved over 8 months) after the implementation of RAID, but this difference was not statistically significant (p = 0.31).
However, there was a substantial reduced risk of readmission after RAID had been implemented compared with controls [e.g. RAID vs. pre RAID: hazard ratio (HR) 2.45, 95% CI 2.33 to 2.57], which was estimated to be equivalent to 22 beds saved daily (20 out of the estimated 22 bed-days saved were attributable to geriatric wards).
Tadros et al. 97 reported that 113 out of 124 A&E referrals (91%) per month were assessed by RAID within 1 hour (average time 24 minutes), and 788 out of 886 ward referrals (89%) were assessed within 24 hours (average 16 hours). 97,106 RAID led to an increase in the detection and diagnosis of mental illness. Of the referrals to the RAID team, 32% were for self-harm and 9% for psychosis. Some of the biggest increases included an 8% increase in the diagnosis of schizophrenia. 97
However, this was only based on one before-and-after study. Although analyses were adjusted for confounding these findings need replication with more rigorous study designs (such as cluster RCTs) to reduce the risk of selection bias. Woo et al. 98 also investigated the impact of implementing a psychiatric emergency service, but in the USA. They did not find any differences between groups on admission or readmission to inpatient treatment, but there were reductions in waiting times (before, 639 minutes; after, 291 minutes; p < 0.01), seclusion and restraint (before, 15%; after, 6%; p < 0.05), provision of emergency medication (before, 74%; after, 53%; p = 0.01) and elopement (before, 13%; after, 5%; p = 0.05). This is largely consistent with the benefits of setting up a psychiatric liaison service found from the RAID study discussed above, but their results are vulnerable to potential confounding.
Three studies investigated the impact of psychiatric nursing services in the ED. 99–101 There was limited evidence of benefit in two Australian studies based on a limited range of outcomes but neither of these was a controlled trial. McDonough et al. 99 found a reduction in mean waiting time (before, 235 minutes; after, 36 minutes) and an increase in nightly service user contact rate (before, 2.9; after, 3.75). A mixed-methods study by Wand et al. 101 found largely positive feedback about the psychiatric nursing service in the ED, but there was no control group so it was difficult to determine the impact of the intervention.
A study by Sinclair et al. 100 in the UK did not find statistically significant differences in waiting times or service user satisfaction associated with a psychiatric nursing service implemented in two EDs.
Providing educational interventions to emergency department staff
Two studies investigated the impact of providing educational interventions to ED staff. 102,103 Neither was conducted in the UK, and it is unclear to what extent they are applicable to UK because of differences in the configuration of emergency services in Germany and Switzerland. In addition, both studies are vulnerable to confounding and, therefore, better-quality study designs are needed to assess the clinical effectiveness of these interventions.
Cailhol et al. 102 found a large reduction in the risk of violent behaviour in people who had accessed the ED because of a suicide attempt [risk ratio (RR) 0.37, 95% CI 0.21 to 0.63]. However, an important limitation concerns the lack of clarity in the paper regarding the definition of violent behaviour.
Pajonk et al. 103 compared outcomes in two EDs in two similar cities: one provided access to trained emergency physicians and supporting health professionals, and the other did not. 103 The proportion of correctly identified and documented suicide attempts or suicide ideation was slightly higher in the city that had received the intervention (59.4%) compared with the city where this was not available (50%). However, these figures were still low. Medication use was higher in the control group (OR 1.49, 95% CI 1.04 to 2.14) but physicians who had received the educational intervention were more likely to prescribe medications specifically related to the psychiatric condition.
Other process interventions in the emergency department
Two further studies104,105 were also identified and it is difficult to conclude the extent to which these findings apply to the UK because of differences in service configuration. In addition, these studies are at high of risk of bias and, therefore, it is difficult to conclude anything from these findings.
Adams and Nielson104 found that a process improvement plan for people entering the ED within 30 days of psychiatric inpatient discharge reduced the readmission rate from 6.51% to 4.3% in the first 6 months of implementing the changes.
Strike et al. 105 investigated service users’ experience of the introduction of psychiatric assessment rooms. Many service users reported negative experiences including exacerbated symptoms, loneliness and seclusion.
Service-level interventions to improve urgent and emergency access to crisis care for people in contact with police
Table 3 summarises the studies included in the systematic review on improving urgent and emergency access to crisis care for people in contact with police. Six studies evaluated interventions involving police and their impact on access to crisis care. 32,107–111 Of these, three studies involved police officers receiving intensive mental health training (usually a 40-hour period), two studies examined the impact of triage provided by mental health professionals, and one study compared outcomes for different models of collaboration between police and mental health professionals being provided in different local areas. One study was conducted in the UK,32 three in the USA,108–110 one in Australia111 and one in Canada. 107
Author Country |
Study design | Interventions | Comparator | Outcomes |
---|---|---|---|---|
Edmondson and Cummins, 201432 UK n = 673 calls |
Mainly descriptive Follow-up period: 6 months |
Police officers received telephone triage from mental health professionals | Mainly none, but some before intervention data | Use of Mental Health Act detention, all other outcomes were descriptive |
Kisely et al., 2010107 Canada n = 4114 calls |
Quasi-experimental Follow-up period: 24 months |
Street triage crisis team formed in partnership between mental health services, municipal police and emergency health services | Area without access to the service | Police time on scene Engagement with outpatient treatment |
Steadman et al., 2000108 USA n ≈ 300 calls |
Quasi-experimental Follow-up period: 10 months |
Police officers received training (CIT model) Community service officers (trained social workers or other related fields) assist police officers in mental health emergencies Street triage crisis team including police and mental health professionals |
Interventions were compared with one another | Taken to treatment location, situation resolved on scene, referred to treatment, arrested |
Compton et al., 2014109 USA n = 1063 calls |
Quasi-experimental Follow-up period: 8 months |
Police officers received training (CIT model) | Police officers did not receive training | Level of force, referral or transport to mental health services, arrests |
El-Mallakh et al., 2008110 USA n = unclear |
Mainly descriptive | Police officer received training (CIT model) | Mainly none but some before intervention data | Inpatient referrals, all other outcomes were descriptive |
Herrington and Pope, 2014111 Australia n = 194 police officers |
Quasi-experimental Follow-up period: 24 months |
Police officers received training (MHIT model) | Police officers did not receive training | Level of force during Mental Health Act events Time at Mental Health Act events Dead time at Mental Health Act events |
Street triage or telephone triage by health- or social-care professionals in collaboration with police officers
Three studies were identified: one conducted in the UK,32 one in Canada107 and one in the USA. 108 The UK study was largely descriptive and provided very limited comparative data; therefore, little can be concluded on the clinical effectiveness of the intervention. Similarly, little could be concluded from the other two studies because of limitations in available outcome data or applicability of comparators.
Edmondson and Cummins32 investigated the impact of a RAID team in Oldham, England, providing telephone triage services. They found that between 2010 (before the intervention) and 2014 (implementation of intervention), the number of Section 136 detentions reduced from 71 to 43. They also found that 9.8% of calls to a telephone triage service resulted in a hospital admission (66 out of 673 calls).
Edmondson and Cummins32 reported that police officers felt that access to triage services from the RAID team improved communication, co-ordination of police–RAID work, timeliness of interventions and completion of assessments within satisfactory time scales. Some stated challenges to the use of RAID included unanswered calls, limited access to radio sets and mobile phones to use of RAID, and disagreements between police and health staff on appropriate designated place of safety. 32
A Canadian study provided more meaningful comparative data, but it is unclear the extent to which these findings can be applied to the UK. Kisely et al. 107 found that at the 12-month follow-up there were no differences between the street triage crisis team and controls for police time on scene of a suicide call. However, at the 24-month follow-up the street triage team [mean 136 minutes, standard deviation (SD) 136 minutes] was associated with a statistically significantly reduced police time on scene compared with the controls (mean 165 minutes, SD 165 minutes). In addition, after adjusting for confounders, people who had contact with the street triage crisis team were more likely to engage in outpatient treatment at the 12-month follow-up (β = 1.3; p < 0.001).
Although some comparative data were reported in a US-based study, the extent to which meaningful inferences can be drawn is unclear. Steadman et al. 108 found that the area focusing on police with mental health training was most likely to transport people to mental health services (75% of mental health-related calls, compared with 20% for the community service officers and 42% for the mobile crisis team). The situation was resolved on the scene most commonly by the community service officers (64% of mental health-related calls, compared with 17% for the street triage crisis team and 23% for the police officers with mental health training). Arrests were also highest for the team with the community service officers (13% of mental health-related calls, compared with 5% for the street triage crisis team and 2% for the police officers with mental health training).
Police officers receiving training in mental health
Three studies109–111 evaluated the benefits of police officers receiving mental health training; none were conducted in UK. No convincing evidence of benefit was found in any of these studies, and the extent to which the findings could be applied to the UK is also unclear.
Compton et al. 109 mostly found no differences in the level of force used by police with mental health training and those without, but police officers with mental health training were more likely than police officers without that training to describe their highest level of force as verbally engaging and negotiating with a person with mental health problems.
More than half of all encounters were resolved at the scene, with no differences found between trained and untrained police officers. However, police officers with mental health training were more likely to refer, or to transport, to a treatment facility (OR 1.70; p = 0.026) and less likely to arrest the individual (OR 0.47; p = 0.007).
El-Mallakh et al. 110 found that arrest rates were lower for calls taken by police officers with mental health training (2.1%) than the overall arrest rate (6.2%), a difference which was statistically significant (p < 0.01). In addition, referral rate to intensive psychiatric services fell from 53% before implementing the CIT training in 2001 to 26.8% in 2004.
Herrington and Pope111 reported no significant difference in the use of force before and after mental health intervention team (MHIT) training, or between front-line MHIT-trained and non-trained officers during Mental Health Act events. There was also no perceived improvement in the quality of relationship between police officers receiving training and health-care staff. However, police officers receiving mental health training reported spending less time dealing with Mental Health Act events than those who did not receive training (54.5 minutes compared with 99.5 minutes; p < 0.05). Police officers with mental health training also experienced less dead time waiting to hand over to health-care professionals (25.4 minutes compared with 54.8 minutes; p < 0.01). 111
Quality of treatment and care in crisis
Alternatives to inpatient treatment
We examined evidence for the effectiveness of a range of alternatives to inpatient treatment for both adults and young people. The findings are summarised in Table 4 and discussed in more detail in the text below.
Intervention | Evidence of clinical effectiveness | GRADE rating: quality of evidence |
---|---|---|
CRHTT for adults | Reduced risk of hospital admission and service user satisfaction compared with inpatient treatment | Low–very low |
Crisis houses for adults | Does not appear to be a difference in effectiveness compared with inpatient treatment | Low–very low |
Acute day hospital care for adults | Does not appear to be a difference in effectiveness compared with inpatient treatment | Low–very low |
Alternatives to inpatient treatment for children and young people | Inconclusive | Very low |
Crisis resolution and home treatment teams
Clinical effectiveness
The NCCMH36 identified six RCTs on CRHTTs published between 1964 and 2005. Evidence from the UK studies was largely consistent with those conducted in other countries and health systems.
There were large reductions in the probability of being admitted to hospital at 3 months, 6 months, 12 months (RR 0.4, 95% CI 0.31 to 0.51; GRADE rating, low; three studies, n = 400) and 24 months (RR 0.32, 95% CI 0.22 to 0.46; GRADE rating, low; one study, n = 118) for people receiving CRHTTs.
However, the probability of readmission was less conclusive at 12 months (RR 0.51, 95% CI 0.21 to 1.2; GRADE rating, very low; four studies, n = 601) and 24 months (RR 0.76, 95% CI 0.36 to 1.63; GRADE rating, very low; two studies, n = 306). It was also inconclusive whether or not CRHTTs reduced Mental Health Act admissions at 3 months (RR 0.65, 95% CI 0.31 to 1.35; GRADE rating, low; one study, n = 87).
There was also strong evidence of service user satisfaction with treatment at 6, 12 and 20 months [standardised mean difference (SMD) 1.21, 95% CI 0.85 to 1.58; GRADE rating, low; one study, n = 137].
Although there was positive evidence for clinical effectiveness, it should also be noted that the quality of evidence for all outcomes was rated low or very low, mainly because of the risk of bias in the included studies and the high level of heterogeneity.
Implementation studies (Crisis resolution team Optimisation and RElapse prevention study)
Crisis resolution team Optimisation and RElapse prevention (CORE) is a UK study currently being conducted at the time of this report whose aims include developing the evidence to optimise the functioning of CRTs.
As part of this study, a systematic review was conducted on how to implement CRTs that included studies examining effective components, service user and staff perceptions of effective elements of CRTs and recommendations by government and non-statutory organisations (Professor Sonia Johnson and Dr Brynmor Lloyd-Evans, personal communication).
Of five studies comparing different CRT models, one study found the presence of a psychiatrist was associated with reduced hospital admission for the CRT group. The other four studies did not find conclusive evidence for effective components of CRT.
They conducted a broader review of the clinical effectiveness of CRTs including 13 studies with similar conclusions that CRTs appeared more effective than treatment as usual in reducing hospital admissions. However, components of CRT were not well reported and there remained important gaps, even after contacting authors. There was a great deal of variety in terms of implementing the intervention between studies and it was unclear which components contributed to the clinical effectiveness and cost-effectiveness of CRTs.
A UK survey found that CRTs that offered a 24-hour service were more effective in reducing hospital admissions than those offering reduced hours, but secondary analysis of the data cast some doubt on the findings.
A review of qualitative studies was also conducted as part of the CORE study which suggested CRTs should provide: easy access, quick response, clarity and continuity concerning CRT treatment and aftercare, and care tailored to the needs of service users by staff with appropriate competencies (Professor Sonia Johnson and Dr Brynmor Lloyd-Evans, personal communication).
Recommendations by government agencies and non-statutory organisations included providing a 24/7 service, assessing all patients before hospital admission with the aim of home treatment where possible, multidisciplinary teams, staffing levels equivalent to 14 full-time staff per 150,000 population served and typical case loads of about 25 to 30 (Professor Sonia Johnson and Dr Brynmor Lloyd-Evans, personal communication).
The CORE study also includes a survey of CRT managers conducted in 2011/12. The findings of this survey are summarised in the May/June 2014 issue of Mental Health Today. 112
They found that only 40% of CRTs were providing a full 24/7 service, although 85% provide some 24/7 cover. Some teams were able to achieve 4-hour targets from referral to assessment, but for many CRTs it was routine practice to provide this the day after referral.
Crisis resolution team staff access to training, supervision and manualised resources to support psychosocial interventions was found to vary widely across the UK. In addition, a small minority had systems in place to monitor and develop staff performance.
Gatekeeping by CRTs also varied across the UK. Only 47% aimed to assess all service users in person before hospital admission, only 35% of CRTs had access to crisis houses and 22% had access to an acute day service.
Another important finding was the challenge to maintain the CRT focus. Many CRTs appeared to be working with a broader range of service users, with not all experiencing crisis. In addition, CRTs were often required to provide other services such as psychiatric liaison services in A&E departments.
This survey, along with interviews with 200 mental health staff, service users and carers resulted in the development of a 39-item CRT fidelity measure (see Appendix 2 or UCL113 for a copy of the fidelity scale). This was then followed up by a survey of 75 CRTs in 2013/14 to assess the extent to which services were consistent with the CORE model of good practice. The findings of this study are not yet published, but we were provided with a summary from the CORE research team. Although examples from CRTs of good fidelity were found for each item, few teams were putting together the whole package with no team meeting the criteria for good fidelity (mean score of 4 per item overall) (Professor Sonia Johnson and Dr Brynmor Lloyd-Evans, personal communication). Therefore, the key need for improvement among CRTs was greater fidelity across the range of items identified by the CORE model.
Cost-effectiveness
Two UK-based economic analyses were identified by the NCCMH. 35
One study compared CRHTTs with standard care based on a large RCT conducted in the UK (n = 260), which included people with psychosis, schizophrenia, bipolar disorder, personality disorders or depression. The authors considered NHS and criminal justice costs (although criminal justice only accounted for a small proportion of costs) based on a time horizon of 6 months. When including inpatient costs, CRHTTs were more cost-effective as they were associated with lower costs (by £2438). When excluding inpatient costs, CRHTTs were more expensive (by £768) but would be cost-effective if society was willing to pay £100 to avoid an extra inpatient day.
A further study conducted another analysis comparing CRHTTs with standard care, this time using data from a before-and-after study in the UK (n = 200) mainly including people with schizophrenia/schizoaffective disorder or bipolar disorder. The authors also included both NHS and criminal justice costs over a 6-month period.
The two studies concluded that CRHTTs resulted in cost savings of £1681 at 2001 prices. Although the clinical data were based on a before-and-after study, which maybe prone to risk of bias, there was extensive adjustment for baseline differences in the analysis.
Crisis houses
The NCCMH35 identified one RCT on crisis houses for people with psychosis, which was conducted in the USA. There appeared to be no additional benefit to crisis houses when compared with standard care (inpatient care) on hospital admission (RR 1.0, 95% CI 0.98 to 1.02; GRADE rating, low; one study, n = 185) or readmission (RR 0.9, 95% CI 0.76 to 1.05; GRADE rating, low; one study, n = 185). No data were available on Mental Health Act admission or service user satisfaction.
The NCCMH35 concluded that, although the evidence is currently inconclusive, crisis houses should be offered to service users as an alternative to standard care.
Lloyd-Evans et al. 66 examined the clinical effectiveness of crisis houses and other residential options in a broader range of populations. They identified 11 studies on crisis houses but only three of these were rated as moderate- or high-quality evidence. One study found some benefits in terms of quality of life, cost of index admission and homelessness at discharge. However, these differences were not sustained at the 2-month follow-up. One study found that crisis houses were less costly but associated with longer admission. Finally, another study found no differences between crisis houses and standard care.
Recent UK studies (randomised and non-randomised) have found similar results as above (and are also consistent with qualitative accounts of service user experience discussed in Chapter 1, Potential barriers to accessing support before crisis) including greater service user satisfaction, greater autonomy, reduced costs of admission, greater therapeutic alliances and no differences in service user and service utilisation outcomes compared with inpatient treatment. 114–117
Acute day hospital care
Clinical effectiveness
The NCCMH35 updated an existing Cochrane review, but did not find any additional trials. Ten RCTs, published between 1965 and 2007, were included in the review and compared acute day hospitals with routine inpatient care.
No evidence of differences was found between day hospital services and standard inpatient services in engaging participants. Studies conducted in the UK were largely consistent with those conducted in other health systems. For example, there was high-quality evidence of no differences between acute day hospitals and treatment as usual in terms of feasibility and engagement (RR 0.97, 95% CI 0.80 to 1.17; GRADE rating, high; one study, n = 1117). Although there were no differences between groups in total days in hospital, duration of hospital stay was shorter for those receiving inpatient care. The meta-analysis found that, overall, service users were more satisfied with day hospital care. However, when examining the Kallert-EU-2007 trial, which was judged to be more applicable to current UK practice, no differences in satisfaction were observed. 35
Cost-effectiveness
No studies were found on the cost-effectiveness of acute day hospital care. 35
Alternatives to inpatient treatment for children and young people
We selected the study by Shepperd et al. 65 as a comprehensive review of the evidence and did not consider any further systematic reviews. They identified seven RCTs on four distinct models of care: multisystemic therapy at home, specialist outpatient service, intensive home treatment and intensive home-based crisis intervention (‘homebuilders’ model for crisis intervention). There was very high conceptual heterogeneity and small sample sizes, and study quality was very low. The authors concluded that the quality of the evidence provided little guidance for the development of services.
Inpatient treatment
Clinical effectiveness and cost-effectiveness of inpatient treatment
Guidelines on psychosis, psychosis and substance misuse, depression, bipolar disorder and BPD all examined evidence for the clinical effectiveness and cost-effectiveness of inpatient treatment.
Studies comparing inpatient treatment with alternatives such as crisis houses and acute day hospitals are reviewed above in Alternatives to inpatient treatment. Most guidelines identified no other evidence on the clinical effectiveness and cost-effectiveness of inpatient treatment.
The NICE guideline Borderline Personality Disorder: Treatment and Management57 identified five studies on inpatient treatment; however, all the studies were conducted by the same institution in Finland, they lacked applicability to inpatient services in the UK and none used a control group. Therefore, it was not possible to conclude anything from these studies in terms of clinical effectiveness and cost-effectiveness of inpatient treatment in the UK. For all outcomes the GRADE rating of quality was very low.
Two studies identified in the Psychosis with Coexisting Substance Misuse: Assessment and Management in Adults and Young People NICE guideline58 showed some evidence for the benefits of inpatient treatment, but these studies were of low quality and difficult to interpret. Therefore, it was not possible to draw conclusions from these studies. The GRADE quality rating for all outcomes was very low.
Examining the clinical effectiveness of inpatient mental health treatment is challenging, as identifying an appropriate control group is extremely difficult, and this has led to a preponderance of largely descriptive studies in this area. An alternative method is to investigate what factors impact on service user outcomes, such as conflict and containment when receiving inpatient treatment; these are reviewed in Factors affecting conflict and containment in inpatient mental health settings.
Factors affecting conflict and containment in inpatient mental health settings
Bowers et al. have conducted a number of systematic reviews on conflict and containment in inpatient mental health settings. These include patient factors associated with psychiatric inpatient aggression118 and self-harm. 119
From these systematic reviews and other data, Bowers et al. 64 developed the Safewards model, which proposes six factors that influence conflict and containment in psychiatric wards: (1) staff team; (2) physical environment; (3) outside hospital; (4) patient community; (5) patient characteristics; and (6) regulatory framework.
The findings of their reviews are summarised in an overview of the evidence base on conflict and containment in inpatient mental health settings, which is grouped in terms of these six factors of the Safewards model. 64 Bowers et al. 64 acknowledged that the evidence base is very limited, consisting mainly of descriptive studies with few controlled trials available. Therefore, the data summarised below can be considered only preliminary.
One of the key factors influencing conflict and containment is structure. A cross-sectional study found that structure was a stronger predictor of conflict and containment than patient satisfaction. Similarly, a longitudinal study found that structure predicted subsequent conflict in wards. In addition, a participant observation with follow-up interviews of staff on four acute and two chronic wards found that violence was more common in wards with unclear staff functions and where there was unpredictability in activities/events and staff–patient interactions. 64 However, reviews of absconding and also substance use in inpatients did not find structure to be an important factor.
The physical environment was also found to be an important aspect of inpatient treatment. Bowers et al. 64 found that permanently locked wards were associated with reduced absconding. However, locked wards were also associated with greater aggression, self-harm and medication refusal. The complexity of ward layout was also an important factor. Suicides were more commonly attempted in private areas (such as bathrooms, bedrooms, toilets); therefore, greater availability of these areas was associated with higher rates of suicide. There was also evidence that removing ligature points in inpatient wards was associated with a reduction in suicides in England and Wales.
Outside hospital factors influenced risk of violence and self-harm. Bowers et al. 64 estimated that 3% of violent incidents were attributable to factors such as a lack of access to money and unresolved family problems. A relatively large study (n = 207) found that external factors accounted for 20% of incidents of self-harm.
The patient community was also found to impact on conflict and containment. A meta-analysis of the violence literature found that 25% of violent incidents occurred after patient–to-patient interactions including physical contact, intrusion into personal space and reactions to sexual approach.
A variety of patient characteristics were associated with conflict and containment. Younger age was associated with a greater risk across a range of outcomes, including violence, substance use, absconding and self-harm. Males were more likely to engage in violence (in acute but not forensic wards), substance use and absconding. Having a schizophrenia diagnosis was associated with greater violence, absconding, self-harm and suicide attempts. Depression was mainly associated with self-harm and suicide attempts. Formal detention was associated with most conflict/containment items such as violence, absconding, coerced medication and manual/mechanical restraint.
Although detention rates vary a great deal between countries, there are insufficient data to examine the influence of the regulatory framework on conflict and containment in inpatient wards. However, studies of national policy have found evidence of a reduction in seclusion use. Similarly, national policies on pro re nata medication (medication when required) have influenced behaviour. A study of European Union member states found that obligatory inclusion of a legal representative lowered compulsory admissions. 120
Bowers et al. 64 concluded that their Safewards model provides a foundation for a number of different interventions to seek to reduce conflict and containment.
Bowers et al. 64 have recently completed a single-blind cluster RCT based on their Safewards model. At the time of the review being conducted the trial results had not yet been published, but some data have been reported in conference presentations. The full study results have since been published in the International Journal of Nursing Studies. 121
They included 15 hospitals and 31 wards that were randomised to receive either the Safewards-based intervention or a control well-being intervention (information provision on diet, physical activity and health promotion). There was a statistically significant reduction in conflict (–14.6%, 95% CI –5.4% to –23.5%) and containment (–23.6%, 95% CI –5.8% to –35.2%) in the intervention wards compared with the control wards.
Promoting recovery/preventing future crises
For the review of all other stages of the mental health crisis care pathway we have focused on service and team models. However, as the focus of the Crisis Concordat for this stage of the care pathway is focused on individual-level interventions we have broadened the scope of the review for this area to include both service models and individual-level interventions recommended by NICE. The findings are summarised in Table 5 and discussed in more detail in Access and engagement for black and minority ethnic communities, Service-level interventions primarily aiming to improve mental health outcomes, Individual-level relapse prevention interventions, Joint crisis plans and Strengths-based approaches to recovery.
Intervention | Evidence of clinical effectiveness | GRADE rating: quality of evidence |
---|---|---|
Service-level interventions for treating mental health conditions | ||
EISs | Hospitalisation, PANSS total (i.e. positive and negative symptoms), employment, engagement with services, global state | Moderate to low |
ICM | Length of stay in hospital, retention in care and service user satisfaction However, effectiveness largely not replicated in UK studies |
High to very low |
Collaborative care for depression in people with chronic physical health problems | Depression symptoms, response, remission but judged to be of clinical importance | High to moderate |
Individual-level relapse prevention interventions | ||
CBT for people with psychosis | Reduction in readmissions, relapse rates and symptom-related outcomes | Moderate to high |
Family interventions for people with psychosis | Reduction in readmissions, relapse rates and symptom-related outcomes | Moderate to high |
Antipsychotics for people with psychosis | Reduction in relapse rates and overall treatment failure | Moderate to high |
Psychological interventions for people with bipolar disorder | Reduction in readmissions, relapse and other outcomes but was not possible to identify particular interventions more likely to be associated with benefit | Low to moderate |
Lithium for people with bipolar disorder | Reduction in relapse rates and overall treatment failure | Low to moderate |
DBT for people with BPD | Reduction in readmissions, self-harm and other outcomes | Low to moderate |
MBT for people with BPD | Reduction in hospitalisation, self-harm and other outcomes | Low to moderate |
Psychological interventions for self-harm | Reduction in repetition of self-harm, hopelessness, depression but was not possible to identify particular interventions more likely to be associated with benefit | Low to moderate |
CBT for depression | Reduction in relapse and depression symptoms | Low to moderate |
Antidepressants for depression | Reduction in relapse and depression symptoms | Low to moderate |
Strengths-based individual-level interventions to promote recovery | ||
Supported employment | Obtaining any occupation, obtaining competitive employment | Moderate to very low |
Peer support | Inconclusive for most outcomes | Low to very low |
Self-management | Positive and negative psychosis symptoms, hospitalisation | High to very low |
Crisis planning | ||
JCPs | Some evidence of benefit in pilot studies, but challenges identified when trying to implement more widely in routine practice | Low |
Access and engagement for black and minority ethnic communities
The NCCMH29 identified two reviews of the published67 and grey literature68 on improving pathways to care for BME communities. These reviews were not rated by the NCCMH. Owing to limitations in reporting of the review, we provided only overall GRADE ratings rather than for each outcome. There was substantial inconsistency of results for most outcomes, there were issues of applicability, particularly for the review of published studies,67 and there was also substantial uncertainty due to lack of reporting. Therefore, we rated the overall evidence as very low quality. However, it is acknowledged that there are substantial challenges in conducting research in this area.
Sass et al. 67 identified six studies on this topic. There was one RCT conducted in the UK that included 70 women of Indian origin with a common mental health disorder and provided an educational leaflet in English, Hindi and Punjabi to promote use of primary care. Although this did not change service use patterns it did result in a reduction in psychological distress among participants.
‘Ethnic matching’ demonstrated an impact on care pathways (two of three studies): improved use of outpatient care for Latinos and East Asians, increased use of continuing care, reduced use of crisis teams and shorter hospitalisations. However, two of the three studies were not able to report a change in pathways to care for African Americans.
Moffat et al. 68 identified eight relevant studies from the grey literature all conducted in the UK. They found three key overlapping themes of collaboration, facilitation of referral between services and improved access through outreach.
In terms of collaboration, they found pathway impacts arising from partnerships between NHS services and culturally specific community groups in three of seven studies. Collaboration and partnerships with local employment and education services also resulted in client completion of work placement scheme and attendance at college (two studies).
Collaboration between services aimed to facilitate routes through care. In addition, some services acted as a bridge between services and sectors to provide culture-specific advocacy through networking with a variety of agencies, and campaigns to raise awareness about the service. Services provided information on other available services and how to access them, but 44% of clients refused referral to suggested services (two studies).
Outreach services, including home treatment, home assessment, specialist clinics in hospitals, GP units or community settings, and delivering health education campaigns or leaflets (eight studies), showed that pathway impacts exist only for home assessment and treatment. Benefits included ensuring accurate assessment and identifying appropriate services for Chinese clients (one study).
Moffat et al. 68 noted there were important limitations in the data, including that most documents did not adopt a research framework or incorporate systematic collection of data. Therefore, it is difficult to draw conclusions from the data.
Service-level interventions primarily aiming to improve mental health outcomes
Interface between primary and secondary care
No robust evidence was identified on the interface between primary and secondary care for promoting recovery-preventing crisis. 35
Early intervention services
Clinical effectiveness
The NCCMH35 identified four RCTs on EISs. All were conducted in the UK or Europe and the length of follow-up ranged from 1 to 2 years.
They defined EISs as providing at least two of the following functions: early identification and therapeutic engagement of people with a first episode of psychosis; provision of evidence-based pharmacological and psychosocial interventions; and educating the wider community to reduce barriers to early engagement with treatment. 35
Early intervention services were more effective than standard care in reducing hospitalisation (RR 0.88, 95% CI 0.79 to 0.98; GRADE rating, moderate; three studies, n = 733), number of admissions (SMD –0.46, 95% CI –0.8 to –0.12; GRADE rating, moderate; one study, n = 136) and number of bed-days (SMD –0.18, 95% CI –0.33 to –0.03; GRADE rating, moderate; two studies, n = 683).
There was also a reduction in total symptoms based on the Positive And Negative Syndrome Scale (PANSS) (SMD –0.52, 95% CI –0.92 to –0.11; GRADE rating: low; one study, n = 99), positive symptoms based on the PANSS or Scale for the Assessment of Positive Symptoms (SMD –0.21, 95% CI –0.39 to –0.03; GRADE rating, low; two studies, n = 468) and negative symptoms based on the PANSS or Scale for the Assessment of Positive Symptoms (SMD –0.39, 95% CI –0.57 to –0.2; GRADE rating, low; two studies, n = 468).
In addition, there was evidence favouring EISs in terms of greater likelihood of being in work or paid employment (RR 0.72, 95% CI 0.54 to 0.97; GRADE rating, moderate) based on one trial of 436 participants.
Cost-effectiveness
The NCCMH35 identified six cost-effectiveness studies (two conducted in UK, two in Italy, one in Denmark and one in Australia). We will focus on the findings from the UK studies as these are most directly applicable to the review.
One study based their analyses on a RCT of 144 participants and used a time horizon of 18 months. The study estimated NHS costs and criminal justice costs. There were no statistically significant differences between EISs and standard care, whether or not criminal justice costs were included. The analyses concluded that EISs provide better outcomes at no extra cost and, therefore, are a cost-effective option. Another study by the same authors conducted a model-based cost analysis comparing EISs with standard care. Costs were reported for years 1 and 3, and were based on a NHS and Personal Social Services (PSS) perspective. It was found that EISs resulted in cost savings of £4972 and £14,248 in years 1 and 3, respectively (at 2006 prices). In addition, probabilistic sensitivity analysis indicated that there was a far greater likelihood of cost savings associated with EISs and the results were fairly robust.
Intensive case management
Clinical effectiveness
The NCCMH35 selected an existing Cochrane review which included 38 RCTs. They also conducted further subgroup analysis including only UK trials.
Intensive case management (ICM) was defined as a package of care using the assertive community treatment (ACT) model, assertive outreach model, the case management model, or reporting a caseload of up to 20 people. 35
Based on the international literature, ICM was found to be effective in reducing length of stay in hospital, and improving retention in care and service user satisfaction. There were less consistent data on benefits to quality of life. The quality of evidence based on GRADE criteria ranged from high to very low. There was moderate- to high-quality evidence on service user satisfaction, but for most other outcomes there was low- to very low-quality evidence.
However, when limiting the meta-analysis to only UK trials (k = 8) there was no evidence of benefit on reducing average number of days hospitalised, or on quality of life. However, ICM was associated with a greater probability of remaining in contact with service users.
Cost-effectiveness
The review of economic studies found four eligible analyses. 35 We focused on the two studies conducted in the UK as these are most directly applicable to our review.
One study conducted a cost minimisation analysis comparing ICM with standard care. No differences were found in clinical effectiveness and differences in costs were not statistically significant. It was concluded that ICM was not cost-effective, as there was no evidence of improvement in clinical effectiveness and no reduction in costs.
Another study compared cost-effectiveness of ICM with standard care Community Mental Health Teams (CMHTs). ICM was associated with a statistically significant increase in participant satisfaction and an increase in costs of £4031 that was not statistically significant. Therefore, the authors concluded that ICM was cost-effective. However, the NCCMH35 noted that the use of satisfaction scores as the outcome may make the analyses difficult to interpret.
Community Mental Health Teams
Clinical effectiveness
The NCCMH35 identified three RCTs conducted in the UK between 1992 and 1998. They defined CMHT as providing multidisciplinary community-based team care.
None of the three trials showed a statistically significant benefit compared with either standard hospital treatment or traditional psychiatric services. However, the NCCMH35 noted that these trials were unlikely to reflect the diversity of community mental health care in the UK at the present time, as many have assimilated practices used by more recent models of care, such as ACT, outreach services and ICM. The evidence was rated as very low quality.
Cost-effectiveness
One study, discussed in more detail in Early intervention services, found that CMHT was less effective than EIS although not differing (although potentially higher) in costs.
Psychosis and substance misuse integrated service models
Clinical effectiveness
The NCCMH58 identified a Cochrane review on integrated service models. Integrated service models were defined as those that unify services at the provider level, rather than requiring service users to negotiate separate mental health and substance misuse treatment programmes.
Although this Cochrane review has been subsequently updated,38 the authors found no further evidence on this model; therefore, we will focus on the data reviewed in the psychosis and substance misuse guideline.
Four trials were included in the review,58 two comparing an integrated service model (ACT/dual disorders treatment) with standard care, and two comparing integrated ACT with non-integrated ACT. The evidence was inconclusive (GRADE ratings ranged from moderate to low) whether or not an integrated approach was effective for people with psychosis and substance misuse.
In addition, three observational studies examining the clinical effectiveness of integrated service models were included. One study found that the integrated group was associated with a reduced number of days in hospital compared with parallel treatment. Another study found that, compared with standard treatment, integrated treatment was associated with fewer days in an institution, more days in stable housing and greater progress in substance misuse recovery. A further study successively enrolled four groups into a day hospital of a dual diagnosis treatment programme. All groups made sequential improvements when receiving treatment. Although the observational studies suggested some benefit, methodological limitations of the studies and lack of applicability to the UK meant that no firm conclusion could be drawn.
Cost-effectiveness
No UK cost-effectiveness analyses were identified in the NCCMH58 review.
Staffed accommodation for people with psychosis and substance misuse
Clinical effectiveness
The NCCMH58 defined included interventions as any staffed accommodation or supported housing for people with a diagnosis of psychosis and coexisting substance misuse that may include an element of specific treatment for the substance misuse.
There was one trial on staffed accommodation (n = 132) that provided an integrated mental health and substance misuse treatment programme. It was inconclusive whether or not this intervention was more effective than standard care (GRADE rating was low-quality evidence for all outcomes).
In addition, five non-randomised studies were identified in the review. 58 One study found that long-term residential programmes were more effective than short-term residential programmes in engaging service users in treatment and maintaining abstinence from substance use after discharge. However, no differences were found in hospitalisation and incarceration.
One study compared an integrated psychosis and substance misuse model with a therapeutic community approach to substance misuse treatment in people who were homeless. The more restrictive approach was associated with greater dropout, but there were few differences in outcomes between groups.
One study compared psychosocial rehabilitation with a modified therapeutic community approach for homeless people with psychosis and substance misuse. Those undergoing 2 years of psychosocial rehabilitation showed higher abstinence and treatment retention rates, and improved mental health, compared with those treated with a modified therapeutic community approach.
Another study compared a community residential treatment programme with a therapeutic community programme for people with psychosis and substance misuse. Benefits were reported for the therapeutic community approach on abstinence, psychological symptoms and level of functioning at the 12-month follow-up.
Finally, one study found that a medium-intensity therapeutic community was more effective than a low-intensity therapeutic community and treatment as usual, in terms of treatment retention. The medium-intensity group improved on more outcomes than the low-intensity or treatment-as-usual groups.
Although, most of the non-randomised studies found that longer duration residential programmes were more clinically effective than shorter duration, on substance use outcomes at follow-up, there were substantial methodological limitations in these studies. Therefore, it was difficult to conclude whether or not these interventions were clinically effective.
Cost-effectiveness
No UK cost-effectiveness analyses were identified in the NCCMH review. 58
Mood clinics for bipolar disorder
We identified a recent NICE guideline on bipolar disorder in adults, children and young people. 59 This was judged to include comprehensive and reliable reviews of the literature on promoting recovery; therefore, we did not seek to review further sources of data.
Clinical effectiveness
The NCCMH59 identified one RCT (n = 158) on mood clinics defined as an outpatient clinic providing evidence-based psychosocial and pharmacological interventions. This was conducted in Denmark and GRADE ratings ranged from low- to very low-quality evidence. This was effective in reducing the risk of hospitalisation (RR 0.66, 95% CI 0.46 to 0.95; GRADE rating, low; one study), but not the number of relapses (RR 1.10, 95% CI 0.85 to 1.42; GRADE rating, very low). It was concluded that services providing co-ordinated evidence-based psychological and pharmacological interventions were likely to be beneficial for people with bipolar disorder. 59
Cost-effectiveness
No UK cost-effectiveness analyses were identified in the NCCMH59 review.
Collaborative care for bipolar disorder
Clinical effectiveness
The NCCMH59 identified one trial on collaborative care for bipolar disorder. This was judged to be a comprehensive and reliable review on the evidence for this intervention.
Collaborative care was defined according to NCCMH:60
-
the provision of case management, supervised and supported by a senior mental health professional
-
development of a close collaboration between primary and secondary care services
-
provision of a range of evidence-based psychosocial and pharmacological interventions and medication management
-
provision of long-term co-ordination of care and follow-up.
There was evidence of effectiveness for reducing depression (SMD –0.56, 95% CI –1.06 to –0.07; GRADE rating, very low; one study, n = 66), but this was rated as very low quality.
For all other outcomes it was inconclusive whether or not collaborative care was more clinically effective than treatment as usual based on low- to very low-quality evidence.
Cost-effectiveness
No UK cost-effectiveness analyses were identified in the NCCMH59 review.
Service-level interventions for depression
We identified a recent NICE guideline on depression in adults and depression in people with a chronic physical health problems. 60 This was judged to include comprehensive and reliable reviews of the literature on promoting recovery. We also supplemented these data with a more recent Cochrane review of collaborative care. 63
Clinical effectiveness
Stepped-care models are commonly used in the treatment of common mental health problems in the UK. For example, NICE guidance on depression and other common mental health conditions assume a stepped-care model. Similarly, a report on two demonstration sites of the IAPT evaluation found good evidence for increased patient follows through the system. 60 However, there is currently limited evidence from direct studies on common mental health problems on the effectiveness of stepped care. However, there is a trial in the UK currently under way on stepped care for depression (led by Jacqueline Hill at the University of Exeter, Exeter, UK).
Fifty RCTs were included in the NCCMH60 review of collaborative care, which was limited to people without a comorbid physical condition. The definition of collaborative care was the same as that provided above for the treatment of people with bipolar disorder.
There were statistically significant differences between collaborative care and control on depression outcomes such as response (self-rated measures: RR 0.83, 95% CI 0.75 to 0.92; GRADE rating, high; seven studies, n = 1820) remission (RR 0.91, 95% CI 0.86 to 0.97; GRADE rating, high; three studies, n = 1480) and mean depression score (SMD –0.16, 95% CI –0.25 to –0.06; GRADE rating, high; 11 studies, n = 1876). The NCCMH60 concluded that these differences were of limited clinical importance.
Seventeen RCTs were included in the NCCMH60 review of collaborative care limited to people with a comorbid physical condition. Collaborative care was effective on depression outcomes including response (RR 0.79, 95% CI 0.73 to 0.85; GRADE rating, high; eight studies, n = 2652), remission (RR 0.81, 95% CI 0.73 to 0.90; GRADE rating, moderate; five studies, n = 2191) and mean depression score (SMD –0.31, 95% CI –0.40 to –0.22; GRADE rating, high; 10 studies, n = 1969). The NCCMH60 concluded that collaborative care in people with depression and a comorbid physical condition was likely to be clinically important.
A more recent Cochrane review63 on collaborative care for depression and anxiety with 79 included RCTs combined data from general adult and chronic physical health problem populations. The authors found similar magnitudes of benefit as the NCCMH60 and argued that sensitivity analyses excluding people with chronic physical health problems did not impact on effect estimates. However, a more detailed exploration of differences between these populations is needed to confirm their conclusion.
Cost-effectiveness
The NCCMH60 conducted a cost-effectiveness analysis of collaborative care for people with depression and a comorbid physical condition using a 15-month time horizon. They adopted a NHS and PSS perspective and the outcome measure was quality-adjusted life-years (QALYs). Collaborative care was found to be cost-effective as the incremental cost-effectiveness ratio (ICER) was below the threshold set by NICE (ICER of £4043 per QALY).
Individual-level relapse prevention interventions
Psychosis
Cognitive–behavioural therapy (CBT) and family intervention were recommended by NICE43 for preventing the risk of relapse and for promoting recovery. In addition, antipsychotic medication was also recommended.
Antipsychotic medication
Clinical effectiveness
The NCCMH35 identified 17 RCTs evaluating the clinical effectiveness of antipsychotics in people with schizophrenia that are in remission. All antipsychotics were found to reduce risk of relapse or overall treatment failure compared with placebo. There was insufficient evidence to determine whether or not there were any differences in effectiveness between antipsychotics.
An additional 18 RCTs were identified for people with psychosis who had not responded adequately to treatment. Clozapine appeared to be the most effective treatment for reducing relapse, but there were no clinically significant differences identified for persistent negative symptoms.
For people who did not respond adequately to clozapine, augmentation with a second antipsychotic may be more effective (based on six small RCTs).
There was insufficient evidence to recommend treatment with depot/long-acting injectable antipsychotic medication.
Cost-effectiveness
The NCCMH35 conducted a systematic review of cost-effectiveness studies for preventing relapse and promoting recovery. Most included studies were conducted outside the UK and, therefore, not applicable to the NHS health-care system. In addition, the two UK studies were associated with a high level of uncertainty and had important limitations.
Therefore, the NCCMH35 conducted a decision-analytic model from a NHS PSS perspective comparing olanzapine, amisulpride, zotepine, aripiprazole, paliperidone, risperidone and haloperidol. The model considered events such as relapse, discontinuation of treatment due to side effects and switching to another antipsychotic drug, discontinuation of treatment due to other reasons and moving to no treatment, as well as development of side effects such as weight gain, diabetes and mortality.
Probabilistic analyses identified a great deal of uncertainty in the results suggesting that no antipsychotic medication can be considered the most cost-effective intervention than other options.
Cognitive–behavioural therapy
Clinical effectiveness
The NCCMH35 found 11 RCTs that compared CBT with any control in participants during the promoting recovery phase. CBT was found to reduce readmission rates up to 18 months post follow-up (RR 0.76, 95% CI 0.61 to 0.94; five studies) and also to reduce the duration of hospitalisation (mean difference –8.26 days, 95% CI –15.51 to –1.01 days; five studies). CBT was also shown to reduce symptom severity [as measured by PANSS and the Brief Psychiatric Rating Scale (BPRS)] and depression outcomes.
Cost-effectiveness
The NCCMH35 also conducted an economic analysis showing that CBT was likely to be cost saving. The intervention costs were likely to be offset by a reduction in risk of future hospitalisation. The net cost of providing CBT was found to lie between –£2277 (overall net saving) and £557 per person with schizophrenia (for a mean duration of hospitalisation of 110.6 days) or between –£1017 and £751 per person (for a mean duration of hospitalisation of 69 days), using the 95% CIs of RRs of hospitalisation, as estimated in the guideline meta-analysis.
The limitation of this economic analysis is that it may have underestimated the potential cost savings of providing CBT for people with schizophrenia, as it did not take into account other potential reductions in health and social care resource use because of the intervention. Therefore, the NCCMH35 concluded that, taking into account these benefits, even a conservative assumption of the potential net cost of £751 per person was likely to be acceptable.
Family interventions
Clinical effectiveness
The NCCMH35 identified 32 RCTs of family intervention compared with any type of control. There was strong and consistent evidence for the clinical effectiveness of family intervention. For example, there was a large reduction in the risk of relapse (RR 0.55, 95% CI 0.45 to 0.67; 12 studies) up to 12 months following treatment. Family intervention also reduced hospital admission (RR 0.53, 95% CI 0.34 to 0.81; 10 studies).
Subgroup analyses suggested single-family intervention was more likely to be acceptable to service users and carers (as shown by data on leaving the study early). In addition, subgroup comparisons suggested that single-family interventions may have been more effective than multiple-family interventions in reducing hospital admission. However, these were based on indirect comparisons and would need to be established either by head-to-head trials or network meta-analyses.
Cost-effectiveness
The NCCMH35 also conducted economic analyses of family intervention for people with schizophrenia. They found that family intervention was likely to be cost saving because of the reduction in relapse rate offsetting intervention costs.
The net cost saving of providing family intervention ranged between £1195 and £3741 per person with schizophrenia, assuming a mean duration of hospitalisation of 110.6 days and the 95% CIs of RRs of relapse, as estimated in the guideline meta-analysis. When a mean length of hospital stay of 69 days was used, the net cost of providing family intervention was found to lie between –£1326 (overall net saving) and £263 per person with schizophrenia.
As above, this is likely to be an underestimate of the cost savings associated with family intervention, as relapse rates often reduce further after completion of the intervention. Therefore, the NCCMH36 concluded this was likely to be a cost-effective option.
Bipolar disorder
Pharmacological interventions
Clinical effectiveness
The NCCMH59 found 36 RCTs on long-term pharmacological management of bipolar disorder. Despite a relatively large number of included studies, the evidence is very heterogeneous, which precluded the use of meta-analysis on most occasions. Consistent with most reviews in the area, the report concluded that lithium is the most effective intervention in preventing relapse. However, the authors also point out that this treatment may not be effective in approximately 40% of people; therefore, alternative options are also required for these service users. For example, there is some evidence of the benefits of valproate alone or as an adjunct to lithium.
Evidence for antipsychotics, although largely positive (e.g. olanzapine and quetiapine), is mainly based on known responders’ discontinuation of acute treatment, which was not judged to provide a reliable test of the comparator agents.
All relevant pharmacological interventions were associated with serious side effects. For example, hyperparathyroidism and impaired renal function were associated with lithium, and weight gain was associated with olanzapine and valproate.
Cost-effectiveness
The NCCMH59 conducted a systematic review of cost-effectiveness studies on pharmacological interventions for promoting recovery in people with bipolar disorder. They found it was not possible to draw conclusions from the data because of serious study limitations, such as short time horizons in models, lack of consideration of side effects and potential conflicts of interest.
Psychological interventions
Clinical effectiveness
The NCCMH59 included 55 trials in their review of psychological interventions. The results of the meta-analyses suggested that psychological interventions may be associated with symptomatic improvement and reduced rates of relapse and hospitalisation. The majority of these trials were of moderate to low quality.
Individual psychological interventions reduced hospitalisation at end of treatment and follow-up, but the results lacked precision (RR 0.63, 95% CI 0.38 to 1.02; seven studies; quality, low). There were clear reductions in risk of relapse at end of treatment at follow-up (RR 0.74, 95% CI 0.63 to 0.87; seven studies; quality, moderate).
The evidence regarding the ability of group psychological interventions to reduce hospitalisation (RR 0.48, 95% CI 0.16 to 1.45; five studies; quality, very low) or relapse (RR 0.86, 95% CI 0.61 to 1.20; five studies; quality, very low) at both end of treatment and follow-up was inconclusive because of wide CIs.
Family psychoeducation was associated with reduction in hospitalisation (RR 0.05, 95% CI 0.00 to 0.83; three studies; quality, low) and relapse (RR 0.52, 95% CI 0.32 to 0.84; three studies; quality, low) at follow-up.
It was inconclusive whether or not family-focused therapy is effective in reducing relapse (RR 0.67, 95% CI 0.34 to 1.30; four studies; quality, very low), but appears to reduce hospitalisation (RR 0.24, 95% CI 0.08 to 0.74; four studies; quality, very low).
Cost-effectiveness
The NCCMH59 identified one economic study conducted in the UK that compared the cost-effectiveness of CBT with treatment as usual and adopted a NHS and PSS perspective over a time horizon of 12 months and 30 months. CBT was the dominant option, being more effective than treatment as usual at both 12 months and 30 months, although there were no differences in costs between the groups at both time points. Probabilistic analyses estimated a probability of 0.85 of CBT being cost-effective given a zero willingness to pay per additional day free from bipolar episodes at 12 months. There were only minor limitations identified with the study.
Borderline personality disorder
Psychological interventions
Clinical effectiveness
The NCCMH60 identified two multimodal psychological therapy programmes for people with BPD: dialectical behaviour therapy (DBT) and mentalisation-based therapy (MBT). A more recent Cochrane review77 provided an update of data on these interventions.
Eight trials of DBT, primarily conducted in outpatient settings, were identified. Large reductions in mean suicidality scores (SMD –1.26, 95% CI –2.24 to –0.29; one study, n = 20) and medium reduction in mean parasuicidality scores (SMD –0.54, 95% CI –0.92 to –0.16; three studies, n = 110) compared with control were observed. There was also evidence of reductions in inappropriate anger (SMD –0.83, 95% CI –1.43 to –0.22; two studies, n = 46), but evidence was inconclusive on total severity of BPD symptoms, impulsivity, interpersonal problems and dissociation/psychoticism.
No benefits were found for DBT compared with general management in one trial of 180 participants, and compared with community treatment by experts in one trial of 101 participants.
Two trials on MBT were identified: one with partial hospitalisation and the other in an outpatient setting. MBT with partial hospitalisation was associated with a large reduction in risk of suicidality (RR 0.08, 95% CI 0.01 to 0.58; one study, n = 38) and parasuicidality (RR 0.44, 95% CI 0.24 to 0.81; one study, n = 38) than outpatients receiving treatment as usual. Large reductions in interpersonal pathology (SMD –1.98, 95% CI –2.78 to –1.19; one study, n = 38) and depression (SMD –1.98, 95% CI –2.78 to –1.19; one study, n = 38) were also found, compared with controls. However, it was inconclusive whether or not MBT was associated with reductions in anxiety and general psychopathology.
Similar benefits were found in the larger MBT trial conducted in an outpatient setting. There were large reductions in suicidality (RR 0.11, 95% CI 0.03 to 0.46; one study, n = 134) and parasuicidality (RR 0.56, 95% CI 0.34 to 0.92; one study, n = 134) compared with treatment as usual. Large reductions in interpersonal pathology (SMD –0.95, 95% CI –1.30 to –0.59; one study, n = 134), small to medium reductions in depression (SMD –0.45, 95% CI –0.79 to –0.10; one study, n = 134) and medium reductions in general psychopathology (SMD –0.67, 95% CI –1.02 to –0.33; one study, n = 134) were also found compared with treatment as usual.
Cost-effectiveness
The NCCMH60 identified one cost-effectiveness analysis of both DBT and MBT based on a UK NHS perspective. They conducted four separate analyses on the cost-effectiveness of DBT using data from different RCTs. Three analyses found DBT to potentially be a cost-effective option, the other analysis found DBT to have a cost per QALY exceeding the NICE cost-effectiveness threshold. There was a high level of inconsistency across studies, most had small sample sizes and a high drop-out rate, as well as important limitation in the economic analyses; therefore, it is difficult to draw conclusions on the cost-effectiveness of DBT based on these data. 60
The NCCMH60 also found that MBT was potentially a cost-effective option, with an ICER below the NICE cost-effectiveness threshold. However, clinical data were based on only one small trial and there was great uncertainty regarding cost-effectiveness based on the probabilistic analyses. Therefore, as with DBT, it is difficult to draw conclusions on the cost-effectiveness of MBT.
Pharmacological interventions
Clinical effectiveness
The NCCMH57 identified 28 studies on pharmacological interventions for people with BPD. However, there were few studies for each individual drug and most were based on small sample sizes and broad variations in setting (inpatients, outpatients) and population (symptomatic volunteers, with and without comorbid axis I disorders). There was also variable reporting of outcome measures.
The NCCMH concluded that there was insufficient evidence for any pharmacological intervention being effective for the treatment of people with BPD. Therefore, they concluded that drug treatment should not be specifically used for the treatment of BPD. However, pharmacological treatments for comorbid conditions, such as depression, may be appropriate.
Cost-effectiveness
No evidence was identified on the cost-effectiveness of pharmacological interventions for people with BPD.
Self-harm
Psychological interventions
Clinical effectiveness
The NICE guidance56 recommends psychological interventions specifically structured for people who self-harm with the aim of reducing self-harm.
Psychological interventions were found to be more effective than treatment as usual in reducing repetition of self-harm (RR 0.76, 95% CI 0.61 to 0.96; GRADE rating, low; nine studies, n = 1323) and hopelessness (SMD –0.52, 95% CI –0.86 to –0.18; GRADE rating, moderate; three studies, n = 149). However, treatment modalities and settings differed considerably across trials, which resulted in high heterogeneity. Participants also differed between studies in terms of previous history of self-harm (ranging from 30% to 100%, where reported) and there was a lack of reporting of psychiatric diagnoses.
There was no convincing evidence for the clinical effectiveness of pharmacological interventions. When prescribing drugs for associated mental health conditions to people who self-harm it was recommended that the toxicity of the drug in overdose should be taken into account.
Cost-effectiveness
No evidence of UK studies on cost-effectiveness was identified.
Pharmacological interventions
Clinical effectiveness
There was no convincing evidence for the clinical effectiveness of pharmacological interventions. When prescribing drugs for associated mental health conditions to people who self-harm it was recommended that the toxicity of the drug in overdose should be taken into account.
Cost-effectiveness
No evidence of UK studies on cost-effectiveness was identified.
Depression
Pharmacological interventions
Clinical effectiveness
Based on 20 RCTs, the NCCMH60 concluded that long-term management with antidepressants was associated with reduction in risk of relapse for up to 2 years. However, no evidence was identified comparing the benefits of different durations of treatment. There was also no evidence to suggest that any particular antidepressant was associated with a greater reduction in risk of relapse.
In addition, NICE recommends the continuation of antidepressant medication or augmentation of medication as other options for preventing relapse in people at risk of relapse. 122
Cost-effectiveness
No evidence of UK studies on cost-effectiveness was identified.
Psychological interventions
Clinical effectiveness
Cognitive–behavioural therapy as an option for people at significant risk of relapse or with a history of recurrent depression is recommended by NICE. 122
The largest evidence base for CBT for treating depression is in comparison with antidepressants. The NCCMH60 found reduced relapse rates for CBT in these studies compared with antidepressants. There are also CBT interventions specifically designed to reduce relapse. Group mindfulness-based cognitive therapy (MBCT) was associated with the strongest evidence for reducing relapse. MBCT was more clinically effective than antidepressants, but with a lack of precision (RR 0.80, 95% CI 0.57 to 1.11; GRADE rating, low; one study n = 123), and more clinically effective than control treatments (RR 0.74, 95% CI 0.57 to 0.96; GRADE rating, moderate; one study n = 55) in people who had experienced three or more depressive episodes.
Cost-effectiveness
One relevant UK study was identified by the NCCMH. 59 This study compared MBCT with antidepressant medication based on data from a RCT. The time horizon was 15 months and it was based on a NHS and PSS perspective (although costs were calculated in US dollars). The authors found no statistically significant differences in costs between MBCT and antidepressants. The ICER was US$962 per relapse/recurrence prevented, which they argued justified provision of the intervention.
Psychosis
Joint crisis plans
Clinical effectiveness
Joint crisis plans (JCPs) are recommended by the NICE guideline on service user experience of mental health services. 55 This reflects positive evidence from an early RCT123 conducted in the UK and also qualitative evidence of UK service users’ experience.
However, a more recent larger-scale UK RCT in people with psychosis124 did not identify any benefits of JCPs. Thornicroft et al. 124 found that JCPs did not reduce risk of hospitalisation under the Mental Health Act5 (OR 0.90, 95% CI 0.58 to 1.39). This might reflect difficulties in fully implementing JCPs in routine clinical practice. 124
Cost-effectiveness
Flood et al. ,125 based on the data from the earlier Henderson et al. trial,123 conducted a cost-effectiveness analysis based on a NHS and PSS perspective and also took into account criminal justice costs. The authors found no statistically significant differences in costs and they concluded that there was a 78% probability that JCPs were more cost-effective than treatment as usual. However, Barrett et al.,126 based on the more recent larger UK trial,124 found that JCPs were not cost-effective.
Borderline personality disorder
Joint crisis plans
Clinical effectiveness
A recent pilot study,127 also conducted in the UK, of people with BPD found that the evidence that JCPs reduce self-harm is inconclusive (OR 1.90, 95% CI 0.53 to 6.5). Although consistent with the data from qualitative studies, there was a high level of acceptability, with 85% of service users stating that they would recommend the intervention.
Cost-effectiveness
Borschmann et al.,127 also based on UK data, found no statistically significant differences in costs between JCPs and treatment as usual. However, as there were no differences in clinical effectiveness, this would suggest JCPs are not a cost-effective option.
Strengths-based approaches to recovery
Vocational rehabilitation
The NCCMH29 identified three types of vocational rehabilitation:
-
prevocational training (receive training as preparation for seeking competitive employment)
-
supported employment (attempts to place service users in competitive employment immediately)
-
modifications of vocational rehabilitation programmes (either prevocational training or supported employment that includes techniques to increase motivation).
Clinical effectiveness
Thirty-eight RCTs were included in the NCCMH35 review of vocational rehabilitation in people with serious mental illness. Supported employment was found to be the most effective method for obtaining competitive employment and for any occupation (i.e. paid/unpaid or voluntary), but only 4 of the 18 trials were conducted in the UK.
There was strong evidence that supported employment was more clinically effective than pre-vocational training (RR 0.63, 95% CI 0.56 to 0.72; GRADE rating, moderate; 18 studies, n = 3627) in gaining competitive employment or any occupation (RR 0.70, 95% CI 0.56 to 0.87; GRADE rating, very low; seven studies, n = 1043).
Although no benefit was found in UK studies, this may reflect the fact that there were insufficient studies to reliably identify an effect.
Supported employment (standard or modified) was more clinically effective than non-vocational training (RR 0.46, 95% CI 0.25 to 0.85; GRADE rating, very low; three studies, n = 2277), but the evidence was rated as very low quality. High-quality evidence from one study involving 2055 participants found that supported employment was more clinically effective than non-vocational training on quality of life and occupational employment outcomes (e.g. obtaining an occupation, earnings or hours worked per week).
There was no evidence that prevocational training was more effective than non-vocational control for gaining competitive employment or increasing earnings, although there was very low-quality evidence to suggest that pre-vocational training was more effective for obtaining any occupation.
Evidence from one trial (n = 107) found that combined supported employment and prevocational training was more effective than supported employment alone in obtaining competitive employment (RR 0.23, 95% CI 0.13 to 0.39; GRADE rating, moderate).
Cost-effectiveness
Two cost-effectiveness analyses conducted in the UK were identified in the NCCMH35 review. However, both analyses were based on small studies with methodological limitations. Therefore, the NCCMH conducted economic modelling to assess the cost-effectiveness of a supported employment programme compared with treatment as usual based on the supported employment versus non-vocational training meta-analysis discussed above in Clinical effectiveness.
They adopted a NHS and PSS perspective and the outcome measure was QALYs. Supported employment programmes were found to be cost-effective compared with treatment as usual because of a low ICER (£5723 per QALY) that was below the threshold set by NICE.
Peer support
Clinical effectiveness
There have been over nine non-systematic narrative reviews of peer support and one systematic review published more than 10 years ago. Therefore, the NCCMH35 conducted a new systematic review on the clinical effectiveness of peer support. For further details on the methods and results see Lloyd-Evans et al. 69
The NCCMH35 distinguished three types of peer support: mutual support groups (where relationships are reciprocal in nature), peer support services (where support is provided by a peer support worker to one or more participants), and peer mental health services (where people who have used mental health services provide part or all of standard care provided by service).
The authors identified 16 RCTs on peer support [nine on peer support services (mean 43% had diagnosis of psychosis or schizophrenia), four on mutual support services (mean 42% had diagnosis of psychosis or schizophrenia) and three on peer mental health service providers (mean 68% had diagnosis of psychosis or schizophrenia)]; in all trials, the intervention groups were compared with a control group. The quality of evidence based on GRADE criteria ranged from low to very low.
Peer support services were associated with some positive effects on self-rated recovery (four trials) and functional disability (one trial), but no differences were found on empowerment or quality of life compared with control (two trials). Three studies found that people receiving peer support were more likely to have greater contact with services, but there was no conclusive evidence of benefits on hospitalisation or service user satisfaction. No follow-up data were available.
Three trials found evidence of benefit of mutual support groups in improving self-rated empowerment, quality of life and contact with services at the end of intervention. There was no evidence of benefits on hospitalisation, functional disability or service user satisfaction.
One trial on peer mental health service providers found that service users in the control group were more satisfied than those in the intervention group. It was unclear if receiving peer mental health services reduced risk of hospitalisation, and no data were reported on empowerment, recovery, functional disability or quality of life.
The NCCMH35 concluded that the benefits of peer-provided interventions were inconclusive in terms of both magnitude and direction of effect. Although there was some evidence of benefit, all outcome data were rated as low- or very low-quality evidence because of study design limitations and high heterogeneity across studies. These conclusions contrast with previous narrative reviews that have tended to emphasise positive findings from low-quality evidence. This might partly reflect that some of these reviews included a large number of non-randomised studies whose findings have yet to be replicated in more rigorous trials. 69
Cost-effectiveness
No UK cost-effectiveness analyses were identified by the NCCMH. 35 We identified one cost-effectiveness analysis based on a cluster RCT conducted in the UK. 128 This was judged to be a well-conducted study. There were no statistically significant differences between the peer support service and usual care. The incremental cost per unit of improvement on the Beck Hopelessness Scale was £12,555.
Self-management
Clinical effectiveness
The NCCMH35 included 25 RCTs in the review of self-management: 21 evaluated professional-led self-management and four evaluated peer-led self-management. A mean of 80% of participants in each study had a diagnosis of psychosis or schizophrenia. GRADE ratings ranged from very low to high.
Ten RCTs found evidence of benefit for self-management in reducing positive and negative symptoms of psychosis at the end of treatment, but no differences were found at follow-up. Benefits on total psychosis symptoms were inconclusive.
Five RCTs found that the risk of admission in the short term was lower in the self-management group, but there was no evidence of benefit at end of intervention or medium- or long-term follow-up.
There was inconclusive evidence for differences between the self-management and control groups on self-rated empowerment at end of intervention. However, there was evidence of benefit for self- and clinician-rated recovery, based on seven RCTs. No differences were found on functional disability at follow-up.
The NCCMH35 concluded that self-management appeared effective in reducing symptoms of psychosis and aiding recovery, but the evidence was less conclusive regarding the effect on the risk of hospitalisation.
Cost-effectiveness
No cost-effectiveness analyses were identified by the NCCMH. 35
Other strengths-based approaches
Ibrahim et al. 70 conducted a systematic review of strengths-based approaches. Five trials were included in the meta-analysis (four RCTs and one quasi-experimental study). Of these, one trial compared strengths-based case management with ACT, three trials compared strengths-based case management with standard care and one trial compared strengths-based case management with traditional case management services. Our GRADE ratings for all outcomes judged the evidence to be of very low quality.
A further three trials (one RCT and two quasi-experimental studies) were included in the narrative synthesis. Two trials compared strengths-based case management with standard case management services and one trial compared a strengths-based case management with treatment as usual.
Ibrahim et al. 70 found that the evidence as to whether or not strengths-based approaches were effective for level of functioning and quality of life was inconclusive. However, they found that the control group was more effective than strengths-based case management on psychotic symptoms, although this was mainly due to one study that was an extreme outlier (SMD = 40).
Risk of bias was unclear for most studies and there was high heterogeneity for all outcomes, reflecting substantial differences in both the interventions and comparison groups. In addition, many studies were conducted in the 1980s and 1990s, and the extent to which such studies can be generalised to current UK services was unclear. It was therefore not possible to conclude if strengths-based approaches are more effective than control approaches.
Chapter 6 Discussion
Principal findings
Access to support before crisis point
There is evidence from studies of a range of disorders that telephone support and triage result in quick access and acceptable referral decisions, and there is minimal evidence of harm. However, at present there are very limited data in relation to the use of telephone support and triage for providing support to people before the point of mental health crisis.
In addition, studies that have assessed the benefits of training and supporting primary care and community-based staff have not identified any clearly beneficial models.
The NICE guidelines55 on access to support before crisis point overlap largely with the recommendations from the Crisis Concordat3 and the London Strategic Network1 commissioning guide on the importance of receiving care with a minimum of delay, the importance of quick referral (either through self-referral or building links between mental health services, primary care and third-sector organisations) and equality of access. However, these are largely based on expert consensus and require further research.
Urgent and emergency access to crisis care
This review identified 15 studies relevant to urgent and emergency access to crisis care. Overall, studies were at high risk of bias and GRADE ratings ranged from low- to very low-quality evidence. These limitations, particularly the lack of randomised and cluster randomised trials, should be considered when interpreting the results.
Nine studies were conducted on improving access to crisis care and service user outcomes in the ED. Most of these studies evaluated liaison psychiatry models in which a liaison psychiatry team, a psychiatrist, or psychiatric nurse was located in the ED providing assessment, triage and treatment. The RAID model (conducted in the UK) found large reductions in risk of readmission, which potentially could lead to large savings in bed-days (bed savings largely occurred on geriatric wards), although controlled trials are needed to replicate these findings. Psychiatric liaison models mostly (although a large UK study implementing a psychiatric nursing service in EDs was not found to be effective) led to reductions in waiting times and improved service user satisfaction.
Providing educational interventions to ED staff may improve outcomes, but this was based on only two studies and neither was conducted in the UK; therefore, the extent to which their findings are applicable is unclear. One study found a large reduction in violent behaviour for people accessing the ED following a suicide attempt. Another study reported that services which received educational interventions prescribed less medication overall, but were more likely to prescribe medication appropriate for the psychiatric condition they were treating. There is a need for rigorously designed randomised and cluster randomised trials conducted in the UK to inform whether or not this is a promising approach to improving mental health crisis care.
Six studies were conducted on improving access to crisis care and service user outcomes when police were called out to mental health-related calls; many of the studies were primarily descriptive, with limited comparative data. In addition, only one study was conducted in the UK, which was also primarily descriptive.
There were two main intervention models: model (1) evaluated the effect of health- or social-care staff providing street triage or telephone triage for police officers attending mental health-related calls and model (2) provided police officers training in dealing with mental health-related calls.
Findings from one UK study suggested that providing telephone triage from a liaison psychiatry team may have reduced the number of detentions under the Mental Health Act, but it was not possible to confirm this because of the limitations in how the data were collected. Over time, the use of a street triage crisis team appeared to reduce police time at the scene of suicide calls and also may have resulted in greater engagement in treatment for mental health service users, but these findings were based on limited evidence from non-UK studies.
Mental health training did not appear to substantially influence the level of force used in mental health-related calls, but police officers with mental health training were more likely to transport people to a health-care setting and were less likely to arrest people with mental health problems. There was also some evidence for greater efficiency for police officers with mental health training resulting in shorter reported durations of Mental Health Act events and less ‘dead time’ waiting to hand over individuals to health-care staff. All studies provided limited data and none was conducted in the UK.
Quality treatment and care in crisis
Crisis resolution and home treatment teams were found to be both clinically effective and cost-effective, with benefits including substantial reductions in probability of hospital admission and greater service user satisfaction. Studies conducted in the UK were largely consistent with those conducted in other health systems. However, the quality of evidence for all studies was rated low because of the small number of studies, high risk of bias in included studies, and high heterogeneity. Reviews of factors affecting the clinical effectiveness and cost-effectiveness of CRHTTs found a great deal of variability when implementing these interventions.
Although there is no evidence that crisis houses or acute day hospitals are more effective than inpatient treatment, they also do not appear to be associated with greater readmissions than inpatient care. In addition, crisis houses are often associated with greater satisfaction than standard services. Therefore, both crisis houses and acute day hospitals are recommended as options by NICE as alternatives to inpatient treatment.
In terms of conflict and containment in inpatient mental health services, the evidence is largely based on descriptive studies with few controlled trials. Bowers et al. 64 propose a Safewards model as a foundation for future research on inpatient treatment. They propose six factors that influence conflict and containment: (1) staff team; (2) physical environment; (3) outside hospital; (4) patient community; (5) patient characteristics; and (6) regulatory framework. An intervention based on this model has recently been tested in a cluster randomised trial, which found evidence of reductions in conflict and containment compared with controls.
Promoting recovery/preventing future crises
Promoting recovery and staying well covers a large and diverse literature. We have sought to review this literature primarily by drawing on systematic reviews conducted for the purpose of informing NICE mental health guidelines. We highlight effective interventions recommended by NICE below.
In terms of service models, we identified evidence of benefit for EISs for people with psychosis and other serious mental illnesses based largely on data from the UK. ICM was effective based on the international literature but there was limited evidence of clinical effectiveness based on the UK literature. Collaborative care was effective for depression.
There were also benefits found for the use of antidepressants for people with depression and antipsychotics for people with psychosis and lithium for people with bipolar disorder in preventing relapse and reducing the risk of hospitalisation.
There were a number of clinically effective and cost-effective individual-level psychosocial interventions. For people with psychosis, CBT and family intervention were associated with reduced hospitalisation and risk of relapse. For people with bipolar disorder, psychological interventions were generally associated with reduced risk of hospitalisation and relapse, although it was not possible to distinguish which types of psychological interventions were more likely to be clinically effective and cost-effective. This was similar for psychological interventions for people who self-harm: there was evidence of reduced risk of repetition of self-harm, but it was not possible to identify which types of psychological intervention were most likely to result in benefits for service users. For people with BPD, DBT and MBT both appeared clinically effective in reducing risk of self-harm, hospitalisation and a range of other outcomes, although DBT has a much larger evidence base. For people with depression, CBT was found to be associated with reduced relapse rate there is some indication that MBCT maybe a particularly helpful approach.
In terms of strengths-based interventions, self-management was found to be associated with benefits for a range of outcomes. Peer support maybe clinically effective and cost-effective, but further research is needed to confirm this. Supported employment was found to help people with psychosis and other serious mental illness gain employment, although the number of UK studies is still relatively sparse.
Joint crisis plans are recommended by NICE,55 but more recent trials have found that implementing these interventions in routine practice may be challenging.
Limitations and future research
A common limitation across all four major elements of the care pathway was a general lack of rigorous randomised and cluster randomised trials evaluating models of mental health crisis care. Further high-quality trials conducted in the UK would have a considerable impact on reducing uncertainty regarding what the most clinical effective and cost-effective models of care for the treatment of people experiencing mental health crisis are. Specific challenges for each of the four main elements of the care pathway are summarised in Access to support before crisis point, Urgent and emergency access to crisis care, Quality treatment and care in crisis and Promoting recovery/preventing future crises.
Access to support before crisis point
There were limited data on the clinical effectiveness and cost-effectiveness of most interventions for improving access to support before mental health crisis. However, this is an active area, with many current developments and service initiatives occurring in the UK. 2 This may provide important opportunities to evaluate the benefit of these interventions in providing access to support before mental health crisis and also stimulate further, more formal high-quality research activity that informs future service development.
Urgent and emergency access to crisis care
The included studies had a number of limitations. Many studies were primarily descriptive with limited comparative data. Most studies did not control for confounding either in design or in the analysis, which substantially limits the strengths of our conclusions. In addition, although there are some positive findings, often these were based on a single study or a small number of studies, or there was a complete lack of UK studies.
Therefore, further and better-quality research is needed for improving outcomes both in the ED and for police officers attending mental health-related calls. There is need for more comparative data obtained in UK studies that aim to minimise confounding, either through design (e.g. cluster randomised trials) or, where that is not possible, through extensive adjustment for potential confounders in the analyses.
Quality treatment and care in crisis
For most interventions we have found limited or inconclusive evidence. This largely reflects the difficulty of conducting research on the clinical effectiveness and cost-effectiveness of inpatient and alternatives to inpatient treatment, but also reflects the need for more high-quality evidence to inform mental health crisis treatment.
Crisis resolution and home treatment teams were found to be both clinically effective and cost-effective, with benefits including substantial reductions in probability of hospital admission and greater service user satisfaction. However, the quality of evidence was rated low because of the small number of studies, high risk of bias in included studies and high heterogeneity. Reviews of factors affecting clinical effectiveness of CRHTTs found a great deal of variability when implementing these interventions. Currently, the CORE study, funded by the NIHR and directed by Sonia Johnson at UCL, has designed a resource pack to aid adherence to principles of best practice based on a fidelity scale developed for the study. 113 They are also currently testing whether or not fidelity to the CORE model is associated with better outcomes based on an evaluation of 25 CRTs.
Promoting recovery/preventing future crises
Limitations of the review
The scope for this section of the review is extremely broad; therefore, inevitably, any approach to summarise such a large body of literature within the time scale and resources of this review will have limitations. We limited our review of individual-level interventions to those recommended by NICE. Although this encompasses much of the relevant literature, it is possible important approaches may not have been covered. For example, there is promising evidence on the benefits of personal budgets for promoting choice and control, quality of life and service use among people with mental health problems. 129
Limitations of the evidence base
Although there is a wide-ranging literature on improving symptoms, social functioning and personal recovery, there is still very little on preventing repeat use of acute care and relapse of mental health problems after discharge from crisis services. In addition, evidence on the clinical effectiveness and cost-effectiveness of interventions to reduce risk of self-harm and suicide is very limited. Further research is needed on the clinical effectiveness and cost-effectiveness of low-intensity psychosocial interventions and psychological interventions with problem-solving elements for people who self-harm. In addition, as harm reduction approaches are being used by some mental health services in the UK, observational research to assess the benefits and potential risks of such interventions is needed.
In addition, very few data on the clinical effectiveness and cost-effectiveness of models of care for older adults and people from LGBT communities were identified.
A further limitation is that most evidence was rated as being of low or very low quality. This reflects the need for better-quality trials. However, this also reflects the challenges of conducting complex interventions in this area. Given the diversity and complexity of the populations and intervention components, it is not surprising that high heterogeneity in estimates of clinical effectiveness is common, and attempts to minimise risk of bias are often difficult to implement. However, recent methodological advances in the conduct of complex interventions should be better taken into account in future mental health research.
There was positive evidence for the benefits of strengths-based interventions such as supported employment and self-management, and some positive evidence for peer support. This is a developing body of literature and it is likely future trials with relevance to the UK may substantially impact on our conclusions. For example, a promising pilot study entitled ‘Connecting People a social intervention for promoting recovery in people with mental health problems’ is currently being conducted in the UK (for further details on intervention development see Webber et al. 130).
Some large-scale UK studies of peer support are currently in progress, for example the ENRICH project, funded by the NIHR (ENRICH project led by a joint team from South West London and St George’s Mental Health NHS Trust and St George’s, University of London). This RCT will compare the effects on a peer worker intervention with usual care for service users about to be discharged from a psychiatric ward, on readmission rates, experience of discharge and cost of services.
Mind is currently developing a peer support programme to improve access to peer support through funding from Big Lottery. The project aims to deliver peer support to 6000 people and provide peer support for 2400 people. In addition, the project also aims to measure changes in participant outcomes, assess cost-effectiveness and conduct a process evaluation.
Chapter 7 Conclusions
Implications for practice
Access to support before crisis point
-
NICE recommends services should ensure that people at risk of mental health crisis receive care with minimum delay and quick referral (either through self-referral or building links between services) and that there is equality of access to such care.
Urgent and emergency access to crisis care
-
Although there is evidence of benefits for liaison psychiatry teams in improving waiting times and reducing readmission, this is largely based on uncontrolled studies and a lack of data from the UK.
Quality treatment and care in crisis
-
Crisis resolution teams are more clinically effective than inpatient care for a range of outcomes, although implementation of this model of care varies across the UK with few teams meeting all evidence-based criteria for good practice.
-
Crisis houses and acute day hospitals appear as clinically effective as inpatient treatment, but are associated with greater service user satisfaction.
Promoting recovery
-
Clinically effective service models include EISs for people with psychosis and other serious mental illnesses and collaborative care for depression (particularly for people with chronic physical health problems).
-
Clinically effective pharmacological interventions include antidepressants for people with depression, lithium for people with bipolar disorder and antipsychotics for people with psychosis.
-
Clinically effective individual-level strengths-based interventions include self-management and supported employment. There is also some evidence for benefit for peer support (but this needs further high-quality research to validate these findings).
-
Individual-level interventions with evidence of benefit include for people:
-
with psychosis: CBT, family interventions
-
with bipolar disorder: psychological interventions
-
who self-harm: psychological interventions
-
with BPD: DBT and MBT
-
with depression: CBT (particularly MBCT).
-
-
Crisis planning is currently recommended by NICE, although more recent research has raised questions regarding the clinical effectiveness of this intervention, therefore further research is needed on whether or not this is an effective approach to promote recovery.
Recommendations for research
Access to support before crisis point
-
Most current recommendations and service developments are based on expert opinion with limited research in this area. Rigorous evaluation of current service developments is needed to ensure evidence-based and effective support for service users.
Urgent and emergency access to crisis care
-
Potential benefits of liaison psychiatry teams are based on limited evidence; therefore, confirmation of the clinical effectiveness and cost-effectiveness of these models of care in high-quality trials (e.g. cluster randomised trials) is needed.
-
Data on clinical effectiveness and cost-effectiveness of mental health training of police officers, street triage and telephone triage to assist police officers with potentially mental health-related incidents is very limited and requires rigorous high-quality evaluation.
Quality treatment and care in crisis
-
Current work from the CORE study aims to improve implementation of good practice in CRTs and is an important component of improving quality treatment for people in crisis.
-
Further work is needed to examine the clinical effectiveness and cost-effectiveness of various aspects of inpatient care on service user outcomes.
Promoting recovery
-
Many of the key service models, to provide long-term management and treatment of mental health problems, lack a clear evidence base (e.g. CMHTs, ICM); therefore, further developments are needed.
-
There is a key need to develop models of care that reduce self-harm, suicide and relapse after discharge from crisis services and inpatient treatment.
-
Large-scale studies are currently under way to investigate the clinical effectiveness and cost-effectiveness of peer support, which is a key area of uncertainty.
-
Interventions on improving social networks and social capital are also important developments currently being evaluated in the UK.
-
Interventions to promote equality of access to mental health services for BME populations are needed.
Acknowledgements
We thank NHS England and the DH Policy Research programme for their role in the project being commissioned. In particular, we would like to thank Geraldine Strathdee (National Clinical Director for Mental Health, NHS England) and Alison Tingle (Research and Development Directorate, DH).
We also thank Jim Symington (Director, Symington-Tinto Health and Social Care Consultancy), Sophie Corlett (Director of External Relations, Mind), Tim Kendall (Director, NCCMH), Ian Watt (Professor of Primary Care, University of York) and Rachel Churchill (Reader in Psychiatric Epidemiology, University of Bristol) for their contributions as members of the advisory group. We also thank Dawn Craig (Principal Scientist, Newcastle University), Nerys Woolacott (Senior Research Fellow, CRD, University of York) and Gill Norman (Research Associate, University of Manchester) for early contributions to the review protocol; and Claire Khouja (Technical Editor, CRD, University of York) for editorial assistance. We finally thank the three anonymous reviewers and Sarah Yiannoullou (Managing Director, National Survivor User Network) for their comments on the draft report.
Contribution of authors
Fiona Paton (Research Fellow) was involved in all stages of the rapid synthesis from development of the protocol, through screening studies and data extraction to analysis and synthesis and production of the final report.
Kath Wright (Information Service Manager) conducted literature searches and contributed to the protocol, methods and appendices sections of the report.
Nigel Ayre (Community Development Worker) provided advice throughout the rapid synthesis process and commented on the draft report.
Ceri Dare (Service User Representative) provided advice throughout the rapid synthesis process and commented on the draft report.
Sonia Johnson (Professor of Social and Community Psychiatry) provided advice throughout the rapid synthesis process and commented on the draft report.
Brynmor Lloyd-Evans (Non-Clinical Lecturer in Mental Health and Social Care) provided advice throughout the rapid synthesis process and commented on the draft report.
Alan Simpson (Professor of Collaborative Mental Health Nursing) provided advice throughout the rapid synthesis process and commented on the draft report.
Martin Webber (Reader in Social Work) provided advice throughout the rapid synthesis process and commented on the draft report.
Nick Meader (Research Fellow) was involved in all stages of the rapid synthesis from development of the protocol, through screening studies and data extraction to analysis and synthesis and production of the final report. He took overall responsibility for the rapid synthesis.
Data sharing statement
All available data is included in the appendices of the report.
Disclaimers
This report presents independent research funded by the National Institute for Health Research (NIHR). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, NETSCC, the HTA programme or the Department of Health. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, NETSCC, the HTA programme or the Department of Health.
References
Appendix 1 Search strategies
Guidelines
NHS Evidence database
URL: www.evidence.nhs.uk.
Date range searched: January 1999 to 24 June 2014.
Date searched: 24 June 2014.
Search strategy
“mental health crisis”
Systematic reviews
Database of Abstracts of Reviews of Effects, NHS Economic Evaluation Database and the Health Technology Assessment database
URL: www.crd.york.ac.uk/CRDWeb/.
Date range searched: January 1999 to 25 June 2014.
Date searched: 25 June 2014.
Search strategy
Line # | Search line |
---|---|
1 | mental health crisis |
2 | mental health crises |
3 | mental health |
4 | mental* ill* |
5 | mental* disorder* |
6 | psychiatric |
7 | psychotic |
8 | schizophren* |
9 | bipolar |
10 | #3 OR #4 OR #5 OR #6 OR #7 OR #8 OR #9 |
11 | crisis or crises or emergenc* or acute |
12 | #10 AND #11 |
13 | #1 OR #2 OR #12 |
Cochrane Database for Systematic Reviews
URL: www.cochranelibrary.com/.
Date range searched: January 1999 to 25 June 2014.
Date searched: 25 June 2014.
Search strategy
-
mental near/4 (crisis or crises)
-
(mental* ill*) near/2 (crisis or crises or acute or emergenc*)
-
(mental* disorder*) near/2 (crisis or crises or acute or emergenc*)
-
psychiatric near/2 (crisis or crises or acute or emergenc*)
-
psychotic near/2 (crisis or crises or acute or emergenc*)
-
schizophren* near/2 (crisis or crises or acute or emergenc*)
-
bipolar near/2 (crisis or crises or acute or emergenc*)
-
#1 or #2 or #3 or #4 or #5 or #6 or #7
Primary studies on urgent and emergency access to crisis care
MEDLINE
Date range searched: January 1999 to 11 November 2014.
Date searched: 11 November 2014.
Search strategy
-
mental health crisis.ti,ab.
-
mental health crises.ti,ab.
-
(mental health emergency or mental health emergencies).ti,ab.
-
(psychiatric adj (crisis or crises or emergency or emergencies)).ti,ab.
-
1 or 2 or 3 or 4
-
Emergency medical Services/ or Emergency Services, Psychiatric/
-
Police/
-
Emergency medical Technicians/
-
Allied health Personnel/
-
(paramedic$ or police$ or fireman or fire officer$ or firemen or firefight$ or fire fighter$ or first responder$).ti,ab.
-
(ambulance adj (staff or personnel or officer$ or worker$)).ti,ab.
-
6 or 7 or 8 or 9 or 10 or 11
-
5 and 12
-
Mental disorder/
-
mental health.ti,ab.
-
(mental$ adj disorder$).ti,ab.
-
(mental$ adj ill$).ti,ab.
-
Schizophrenia/
-
schizophren$.ti,ab.
-
exp Personality Disorders/
-
14 or 15 or 16 or 17 or 18 or 19 or 20
-
12 and 21
-
crisis intervention/
-
12 and 23
-
13 or 22 or 24
-
limit 25 to yr="1990 -Current"
-
(editorial or comment or letter).pt.
-
26 not 27
Appendix 2 Crisis Resolution Team Fidelity Checklist (CORE study)
Reproduced with permission from Dr Brynmor Lloyd-Evans, UCL, London, 2015, personal communication.
CORE Crisis Resolution Team Fidelity Scale Version 2
Developed by Brynmor Lloyd-Evans, Sonia Johnson and the CORE Research Group. [The CORE CRT Fidelity Scale constitutes independent research funded by the NIHR under its Programme Grants for Applied Research programme (Reference Number: RP-PG-0109–10078). The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the DH.] www.ucl.ac.uk/core-study.
Enquiries to Brynmor Lloyd Evans –Division of Psychiatry, University College London.
The CORE Programme is managed by Camden and Islington NHS Foundation Trust and University College London.
Please do not use or adapt without permission from the developers.
CORE Crisis Resolution Team Fidelity Scale
# | Item | Evidence | Scoring criteria | Item definitions and scoring guidance | Score |
---|---|---|---|---|---|
1 | The CRT responds quickly to new referrals | P, M |
4: 5 criteria are met 3: 4 criteria are met 2: 3 criteria are met 1: 2 or fewer criteria are met |
|
|
M, S, O (P) | |||||
M, S, O (P) | |||||
M, S, O (P) | |||||
M, S, O (P) | |||||
M, S, O (P) | |||||
2 | The CRT is easily accessible to all eligible referrers | P, M, S, O | Scoring criteria:
4: 3 criteria are met 3: 2 criteria are met 2: 1 criterion is met 1: no criteria are met |
This item only assesses how easily accessible the CRT is to eligible referrers. [The range of eligible referrers is assessed in item 3]
|
|
SU, FF, M, S | |||||
M, S, O | |||||
P | |||||
3 | The CRT accepts referrals from all sources | P, M, S, SU, FF, O | Scoring criteria: accepts referrals from:
4: 6 criteria are met 3: 4 criteria are met 2: At least 3 criteria are met 1: 2 or fewer criteria are met |
|
|
4 | The CRT will consider working with anyone who would otherwise be admitted to adult acute psychiatric hospital | P, M, S, O | Scoring criteria: will work with the following in circumstances where they would otherwise be admitted to an acute mental health ward:
4: 4 criteria are met 3: 3 criteria are met 2: 2 criteria are met 1: One or no criteria are met |
|
|
5 | The CRT provides a 24 hour, seven day a week service | M, S, SU, FF, O | 5: The CRT visits service users at home, when needed, 24 hours a day, seven days a week 4: The CRT visits service users at home at least 14 hours per day, and can provide telephone support and assessments at health service premises at other times 3: The CRT visits service users at home at least 12 hours a day, and can provide telephone support and assessments at health service premises at other times 2: The CRT visits service users at home less than 12 hours per day but can provide telephone support and assessments at health service premises at other times 1: The CRT does not provide telephone support and assessments on health service premises 24 hours a day |
|
|
6 | The CRT has a fully implemented ‘gatekeeping’ role, assessing all patients before admission to acute psychiatric wards and deciding whether they are suitable for home treatment | M, S, O (P) | Scoring criteria:
4: 5 criteria are met 3: 4 criteria are met 2: 3 criteria are met 1: 2 or fewer criteria are met |
|
|
P, M, S, O | |||||
7 | The CRT facilitates early discharge from hospital | M, S, O (P) | Scoring criteria:
4: 5 criteria are met 3: 4 criteria are met 2: 3 criteria are met 1: 2 or fewer criteria are met |
|
|
M, S, O (P) | |||||
M, S, O (P) | |||||
P, M, S | |||||
M, S, O(W), SU, FF | |||||
M, S, O(W), SU, FF | |||||
8 | The CRT provides explanation and direction to other services for service users, carers and referrers regarding referrals which are not accepted | M, S | Scoring criteria:
4: 3 criteria are met 3: 2 criteria are met 2: 1 criterion is met 1: no criteria are met |
|
|
M, S, O | |||||
M, S, O | |||||
M, S, O (P) | |||||
9 | The CRT responds to requests for help from service users and carers whom the CRT is currently supporting | P, M, S, SU, FF | Scoring criteria: Service users and carers are all given a contact phone number for the CRT
4: 3 criteria are met 3: 2 criteria are met 2: 1 criterion is met 1: no criteria are met |
|
|
M, S, SU, FF (P) | |||||
M, S, SU, FF (P) | |||||
M, S or C, SU, FF | |||||
10 | The CRT is a distinct service which only provides crisis assessment and brief home treatment | M, S (P) | Scoring criteria:
4: 4 criteria are met 3: 3 criteria are met 2: 2 criterion is met 1: 1 or no criteria are met |
|
|
P, M, S, C | |||||
P, M, S, C | |||||
M, S, O, C | |||||
M, S, O, C | |||||
11 | The CRT assertively engages and comprehensively assesses all service users accepted for CRT support | C | Scoring criteria:
4: Assessments documenting at least 8 domains are completed with at least 90% of service users and the CRT uses assertive engagement strategies 3: Assessments documenting at least 8 domains are completed with at least 80% of service users 2: Assessments documenting 7 domains are completed with at least 80% of service users 1: Assessments documenting at least 7 domains are completed with fewer than 80% of service users |
|
|
M, S, O | |||||
12 | The CRT provides clear information to service users and families about treatment plans and visits | M, S, SU, FF | Scoring criteria:
4: 5 criteria met 3: 4 criteria are met 2: 2 or 3 criteria are met 1: 1 or no criteria are met |
|
|
C, SU, FF | |||||
SU, FF | |||||
SU, FF, S | |||||
SU, FF, S | |||||
SU, FF, S | |||||
13 | The CRT closely involves and works with families and wider social networks in supporting service users | C (SU, FF) | Scoring criteria:
5: All 5 criteria are met 4: 4 criteria are met 3: 3 criteria are met 2: 2 criteria are met 1: 1 or no criteria are met |
|
|
C (SU, FF) | |||||
C (SU, FF) | |||||
C (SU, FF) | |||||
SU, FF, S | |||||
14 | The CRT assesses carers’ needs and offers carers emotional and practical support | C, FF | Scoring criteria:
5: All criteria are met 4: 3 criteria are met 3: 2 criteria are met 2: 1 criterion is met 1: No criteria are met |
|
|
C, FF | |||||
C, FF | |||||
S (P) | |||||
15 | The CRT reviews, prescribes and delivers medication for all service users when needed | M, S (P) | Scoring criteria:
5: All criteria are fully met 4: 3 criteria are fully met and one partially met 3: 3 criteria are fully met and one unmet OR two are fully met and two partially met 2: 2 criteria are fully met and at least one unmet OR all 4 are partially met 1: Only one criterion is fully met OR < 4 are partially met |
|
|
M, S (P, C) | |||||
C, SU, FF, M, S | |||||
P, S | |||||
16 | The CRT promotes service users’ and carers’ understanding of illness and medication and addresses concerns about medication | S, P | Scoring criteria:
4: 5 criteria are met 3: 4 criteria are met 2: At least 2 criteria are met 1: 1 or 0 criteria are met |
|
|
C (SU, FF, S) | |||||
C, SU, FF, S | |||||
C, SU, FF, S | |||||
C | |||||
C, M, S | |||||
17 | The CRT provides psychological interventions | P, M | Scoring criteria:
4: 4 criteria are met 3: 3 criteria are met 2: 2 criteria are met 1: One or no criteria are met |
|
|
S, C | |||||
C, S | |||||
M, S | |||||
M, S | |||||
18 | The CRT considers and addresses service users’ physical health needs | C (M, S, SU) | Scoring criteria:
4: Three criteria are met 3: Two criteria are met 2: One criterion is met 1: No criteria are met |
|
|
C, SU, FF, M, S | |||||
C (M, S, SU) | |||||
M, S | |||||
19 | The CRT helps service users with social and practical problems | C, M, S, SU, FF | Scoring criteria: The CRT helps service users with the following life domains:
4: 2 life domains are fully supported and 1 is partially supported 3: 2 life domains are fully supported and one is not supported; or all 3 domains are partially supported 2: One life domain is not supported and at least one other is only partially supported 1: Two or three life domains are not supported |
|
|
20 | The CRT provides individualised care | C (SU, FF) | Scoring criteria:
4: At least one criterion is fully met and the others are partially met 3: All criteria are partially met OR two criteria are fully met and one is unmet 2: Two criteria are partially met and one is unmet OR one criterion is fully met and two are unmet 1: one or no criteria are partially met; others are unmet |
|
|
SU, FF (C) | |||||
C | |||||
21 | CRT staff visits are long enough to discuss service users’ and families’ concerns | P | Scoring criteria:
4: Criteria C and D are met, but one of criteria A or B is not met 3: Criteria C and D are met but criteria A and B are not met 2: Criterion C OR criterion D is not met 1: Neither criterion C nor D is met |
|
|
P, C (M) | |||||
P (SU, FF, M, S) | |||||
S, SU, FF | |||||
22 | The CRT prioritises good therapeutic relationships between staff and service users and carers | M | Scoring criteria
4: Criterion D is fully met and 2 other criteria are met 3: Criterion D is fully met but fewer than two other criteria are met 2: Criterion D is partially met 1: Criterion D is unmet |
|
|
M, S | |||||
P, M | |||||
SU, FF (C, S, M, O) | |||||
23 | The CRT offers service users choice regarding location, timing and types of support | P (M, S, SU, C, O) | Scoring criteria:
4: Criterion A and 2 other criteria are met 3: Criterion A and 1 other criterion are met 2: Criterion A and no other criteria are met 1: Criterion A is not met |
|
|
SU (C) | |||||
SU, FF, S | |||||
SU, FF, S | |||||
24 | The CRT helps plan service users’ and service responses to future crises | C | Scoring criteria:
4: Three criteria are met 3: Two criteria are met 2: One criterion is met 1: No criteria are met |
|
|
C | |||||
C, P, S, SU | |||||
C, SU, FF, M | |||||
25 | The CRT plans aftercare with all service users | C (S, O, SU, FF) | Scoring criteria:
4: Criterion A and 4 other criteria are met 3: 4 criteria are met 2: At least 2 criteria are met 1: No or one criteria are met |
|
|
C (S, O, SU, FF) | |||||
C (S, O, SU, FF) | |||||
C (S, O, SU, FF) | |||||
S, SU, FF | |||||
S, SU, FF | |||||
26 | The CRT prioritises acceptability to service users in how CRT care is ended | C, S, SU, FF | Scoring criteria:
4: 4 criteria are met 3: 3 criteria are met 2: 2 criteria are met 1: 1 or no criteria are met |
|
|
C, S, SU, FF | |||||
C, S, SU, FF | |||||
S, SU, FF (C) | |||||
S, SU, FF (C) | |||||
27 | The CRT has adequate staffing levels | P, (M) | Scoring criteria:
4: Criterion A is fully met and two other criteria are met 3: Criterion A is fully met and one other criteria is met OR Criterion A is partially met and all other criteria are met 2: Criterion A is partially met and no more than two other criteria are met 1: Criterion A is not met |
|
|
M, S, SU, FF | |||||
M, S, O | |||||
P (M, S) | |||||
28 | The CRT has a psychiatrist or psychiatrists in the CRT team, with adequate staffing levels | P, M | Scoring criteria:
4: 2 criteria are fully met and one is partially met 3: 2 criteria are fully met and one is not met OR one criteria is fully met and two are partially met 2: Only one criterion is fully met and at least one is unmet 1: No criteria are fully met |
|
|
P, M | |||||
M, S, O | |||||
29 | The CRT is a full multi-disciplinary staff team | P, M | Scoring criteria: CRT staffing includes dedicated time from: i) nurses; ii) occupational therapists; iii) clinical or counselling psychologists; iv) social workers; v) psychiatrists; vi) service user-employees; vii) other support staff without professional mental health qualifications; viii) pharmacists ix) Approved Mental Health Professionals or equivalent; x) non-medical prescribers; xi) family therapist; xii) accredited cognitive behavioural therapist 5: The CRT team includes 8 or more of the listed staff groups 4: The CRT team includes 7 of the listed staff groups 3: The CRT team includes 6 of the listed staff groups: 2: The CRT team includes 5 of the listed staff groups 1: The CRT team includes 4 or fewer of the listed staff groups |
|
|
30 | The CRT provides a thorough induction programme for new staff and ongoing training and supervision in core competencies for CRT staff | M, S (P) | Scoring criteria:
4: 4 criteria are met 3: 3 criteria are met 2: 2criterion is met 1:1 or no criteria are met |
|
|
M, S, P | |||||
M, S (P) | |||||
P | |||||
P | |||||
31 | The CRT has comprehensive risk assessment and risk management procedures, including procedures for safeguarding children and vulnerable adults living with CRT service users | C | Scoring criteria:
4: 5 criteria are met 3: 4 criteria are met 2: 3 criteria are met 1: Two or fewer criteria are met |
Reviewers please double-check you are shown all relevant risk documents (e.g. if risk updates or contingency plans are stored separately from an initial risk assessment) |
|
C | |||||
C | |||||
P | |||||
M, S | |||||
M, S | |||||
32 | The CRT has systems to ensure the safety of CRT staff members | P, S, M | Scoring criteria:
4: Criteria A and B and two other criteria are met 3: Criteria A and B are met and fewer than 2 other criteria are met 2: Either criterion A or criterion B is not met 1: Criterion A and criterion B are both not met |
|
|
M, S | |||||
P | |||||
M, S | |||||
P, S, M | |||||
33 | The CRT has effective record keeping and communication procedures to promote teamwork and information sharing between CRT staff | M, S | Scoring criteria:
4: 4 criteria are met 3: 3 criteria are met 2: 2 criteria are met 1: 1 or no criteria are met |
|
|
M, S | |||||
M, S | |||||
M, S (C) | |||||
M, S | |||||
34 | The CRT works effectively with other community services | C (S, O, SU) | Scoring criteria:
4: Criterion A and E and two other criteria are met 3: At least 3 criteria are met 2: 2 criteria are met 1: One or no criteria are met |
|
|
C, S, M, O | |||||
M, S O | |||||
M, S O | |||||
M, S O | |||||
35 | The CRT takes account of equality and diversity in all aspects of service provision | M, S (C, SU, FF) | Scoring criteria:
4: Criteria A and E and 2 other criteria are met 3: Criterion A and E are met but other criteria are met 2: One of criteria A and E is met 1: Neither criterion A nor criterion E are met |
|
|
P, M | |||||
M, S | |||||
P, M, S | |||||
SU, FF (C, M, S) | |||||
36 | The CRT has systems to provide consistency of staff and support to a service user during a period of CRT care | M, S | Scoring criteria:
4: Criterion D and two other criteria are met 3: Criterion D and one other criterion are met 2: Criterion D is met but no other criteria are met 1: No criteria are met |
|
|
C, SU, FF | |||||
M, S | |||||
S, SU, FF | |||||
37 | The CRT can access a range of crisis services to help provide an alternative to hospital admission for service users experiencing mental health crisis | M, S | Scoring criteria
4: Three criteria are met 3: Two criteria are met 2: One criterion is met 1: No criteria are met |
|
|
38 | The CRT provides frequent visits to service users | C | Scoring criteria:
4: 4 criteria are met 3: 3 criteria are met 2: 2 criteria are met 1: 1 or no criteria are met |
|
|
C, SU, FF | |||||
C | |||||
C | |||||
P | |||||
39 | The CRT mostly assesses and supports service users in their home | C (SU, FF) | Scoring criteria: Where the service user has not actively expressed a preference to meet elsewhere: 5: At least 80% of CRT contacts with service users take place in the service user’s home or current place of residence 4: At least 70% of CRT contacts with service users take place in the service user’s home or current place of residence 3: At least 60% of CRT contacts with service users take place in the service user’s home or current place of residence 2: At least 40% of CRT contacts with service users take place in the service user’s home or current place of residence 1: Less than 40% of CRT contacts with service users take place in the service user’s home or current place of residence |
|
CORE CRT Fidelity Scale – Guidance for Reviewers
CRT fidelity reviews will consist of the following elements: review of 10 consecutive service users’ records; review of paperwork (policies, protocols and other documents relevant to the team); interview with CRT manager; interview with members of the CRT staff team; interviews with CRT service users; interviews with CRT carers; interviews with managers or staff from other services. An example of the schedule for a review is provided in the document Preparing for your CRT Fidelity Review.
When conducting a review, reviewers should use the following the interview schedules and checklists of information:
-
CRT Manager interview schedule
-
CRT staff interview schedule
-
CRT service user interview schedule
-
CRT carer interview schedule
-
Managers from other acute services interview schedule
-
Managers from community services interviews schedule
-
CRT case note review checklist
-
CRT paperwork review checklist
Please use evidence collected with these schedules and refer to the CORE CRT fidelity scale criteria and scoring guidance when scoring a CORE CRT fidelity review. Reviewers should confer and come to an agreed score for each item following the review.
General Scoring Guidance
In addition to the item-specific scoring guidance included within the CORE CRT fidelity scale, reviewers should follow the general scoring guidance below:
-
Work down the scale for each item when scoring – i.e. does the CRT meet the criteria to score 5; if not does it meet the criteria for 4 . . .?
-
All source agreement need not mean unanimous reports that a criterion is met. If there is a clear prevailing view among a respondent group (e.g. service users, or CRT staff) that a criterion is met, then this may constitute agreement even though a single respondent may have a different view.
-
For scoring based on percentages of service users receiving an intervention, combine information from case note reviews and service user and family interviews where appropriate. For example, if 7 out of 10 case notes but all service user and family respondents reported that something was done, an 80% target could be scored as met.
-
If scoring guidance is not explicit, then prioritise direct evidence in deciding whether criteria are met. For example, rely heavily on evidence from case notes and reports from service users about whether interventions are provided, prioritising these over written policies or the CRT manager’s report.
-
Additional evidence may be considered in scoring items where available and relevant. For example, good quality audit data for a CRT may provide stronger evidence than a review of 10 sets of case notes in scoring an item.
Appendix 3 Reviews data extraction and quality assessment
Review characteristics
Study | Pathway component | Objectives | Number of studies | Number of patients | Countries/health-care systems included |
---|---|---|---|---|---|
Alvarez-Jimenez et al., 201171 | Promoting recovery | To determine the clinical effectiveness of pharmacological and non-pharmacological interventions to prevent relapse in patients with FEP | Nine | 1227 | European countries |
Anderson et al., 201046 | Promoting recovery | To assess the clinical effectiveness of pathways to care for patients with FEP and the association with DUP | 30 | 2851 | Australia, Canada, the People's Republic of China, France, Germany, Iran, Ireland, Malaysia, Trinidad and Tobago, Singapore, South Africa, Switzerland, the UK and the USA |
Babalola et al., 201472 | Crisis treatment (short vs. long inpatient stay) | To evaluate the effect of short stay/brief admission hospital care with long stay/standard inpatient care in people with serious mental illness | Six | 2030 | USA and the UK |
Bird et al., 201073 | Promoting recovery | To assess the clinical effectiveness of EISs, CBT and family intervention in early psychosis | Four | 800 | UK, unclear |
Blank et al., 201290 | Access to support before crisis point (teletriage – health) | To determine the appropriateness of, and compliance with, telephone triage decisions | Two RCTs, 52 observational studies | Not reported | New Zealand, Australia, USA, European countries and Canada |
Campbell and Kisely, 200974 | Promoting recovery (advanced treatment directives) | To assess the effects of advance treatment directives for people with severe mental illness | Two | 321 | UK, unclear |
Choo et al., 2012131 | Access (computer technology in EDs) | To evaluate the use of computer technologies in addressing health behaviours in EDs | Four (two RCTs, two cross-sectional observational studies) | 13,076 | USA |
Reilly et al., 201375 | Promoting recovery (collaborative care – severe mental illness) | To assess the clinical effectiveness of collaborative care approaches in comparison with standard care for people with severe mental illness who are living in the community | One | 306 | US Veterans Administration Medical Centre |
Compton et al., 200876 | Access to support before crisis point (training to recognise mental illness) | To explore the research on CIT programmes | 12 | Unclear (but did include at least 452 police officers; 485 patients) | USA |
Dieterich et al., 201077 | Promoting recovery (ICM – severe mental illness) | To assess the effects of ICM in people with severe mental illness and evaluate whether or not the effect of ICM on hospitalisation depends on its fidelity to the ACT model and on the setting | 38 (12 multicentres) | 7328 | USA, Canada, European countries and Australia |
Duncan et al., 2010132 | Promoting recovery (shared decision-making) | To assess the effectiveness of provider-, consumer- or carer-directed shared decision-making interventions for people with mental health conditions | Two cluster RCTs | 518 | Germany |
Hamm et al., 2010133 | Crisis treatment (crisis intervention in children and young people – ED) | To evaluate the clinical effectiveness of ED-based management interventions for mental health presentations | 12 observational studies (prospective cohort, controlled before-after, uncontrolled before-after, interrupted time series designs) | At least 2253 patients; 80 triage nurses; 1991 psychologists; 1966 psychiatrists | USA, Canada, Luxembourg, Australia and Belgium |
Hickling et al., 200786 | Crisis treatment (general medical vs. specialist psychiatric units – acute psychosis) | To compare the clinical effectiveness of open medical wards vs. conventional psychiatric units on outcomes in people with acute psychosis | None | Not applicable | Not applicable |
Hubbeling and Bertram, 201278 | Crisis treatment (CRTs – severe mental illness) | To assess the evidence on whether or not CRTs in the UK and elsewhere reduce the number of hospital beds | One RCT, 15 naturalistic studies | Where reported, 4262 CRT and 2,886,015 residents | UK, USA, Australia, Luxembourg and France |
Huibers et al., 201191 | Access to support before crisis point (teletriage – out of hours care) | To assess the evidence on safety of telephone triage in out-of-hours primary care | 34 (13 observational studies, 11 simulated patient studies) | 7042 unselected/high risk patients | UK (n = 8) and USA (n = 12), unclear |
Ismail et al., 201392 | Access to support before crisis point (primary care service intervention to reduce A&E presentations) | To evaluate primary care service interventions to reduce inappropriate A&E attendances | 34 (six systematic reviews; 13 before-and-after studies or interrupted time series; seven cross-sectional studies; six non-comparative case studies; one cohort; and one non-RCT) | Not reported | Europe (11 UK) and Australia; health system settings ranged from tax-funded or social insurance systems to private insurance systems |
Kidd et al., 201479 | Crisis treatment (inpatient recovery based care – severe mental illness) | To assess the effects of recovery-oriented practices on hospital wards | 27 (four controlled designs; remaining studies were descriptive or uncontrolled) | Not reported | Not reported |
Lloyd-Evans et al., 200966 | Crisis treatment (residential alternatives to hospital admission – severe mental illness) | To assess the clinical effectiveness and cost-effectiveness of and satisfaction with residential alternatives to standard acute inpatient mental health services | 27 [seven RCTs, 15 non-randomised studies (seven prospective, six retrospective, two unspecified), five interrupted time series] | At least 9705 | USA, Canada, UK (six studies) and Switzerland |
Maitre et al., 2013134 | Promoting recovery (advance directives – chronic psychosis) | To review the evidence on psychiatric advance directives for people with chronic psychotic disorders | 36 | Not reported | Not reported |
Marshall et al., 201180 | Crisis treatment (day hospital – severe mental illness) | To compare the effects of day hospital vs. inpatient care for people with acute psychiatric disorders | 10 RCTs [type I trials excluded patients considered ineligible for day hospital treatment (e.g. people too violent or under compulsion); type II trials randomised everyone presenting for admission but admitted to inpatient ward any people allocated to day hospital who were too unwell for immediate day hospital treatment] | Number of patients 2685 | UK, USA, Germany, Poland the Slovak Republic, the Czech Republic and the Netherlands |
Morriss et al., 201387 | Access to support before crisis point (training to recognise recurrence of schizophrenia) | To compare the clinical effectiveness of early warning signs interventions plus treatment as usual involving and not involving a psychological therapy | 34 RCTs (including two cluster RCTs) | 3554 | USA, Canada, European countries, East Asian countries and Australia |
Murphy et al., 2012135 | Crisis treatment (crisis intervention – severe mental illness) | To compare the clinical effectiveness of crisis intervention models vs. standard care in people with serious mental illness experiencing an acute episode | Eight | 1144 | Australia, Canada, the USA and the UK |
Ramos-Rios et al., 201293 | Access to support before crisis point (teletriage – geriatric psychiatric patients) | To assess the evidence on telepsychiatry programmes, particularly psychogeriatric care via videoconferencing in nursing homes for the elderly | 10 studies: case study, pilot study, programme description, cost-effectiveness analysis | 258 | USA, the Republic of Korea and Hong Kong |
Robinson et al., 201181 | Promoting recovery (suicide prevention in young people); crisis treatment (crisis treatment for suicide attempts in young people); promoting recovery (promoting recovery in young people) | To assess the effects of interventions for adolescents and young adults presenting to a clinical setting with suicidal behaviours | 15 | Unclear; sample sizes ranged from 22 to 448 patients | USA, unclear |
Sass et al., 200967 and Moffat et al., 200968 | Promoting recovery (pathways to care – equality of access) | To evaluate research on initiatives to increase, improve or enhance access to care or pathways to mental health care for BME groups | Six studies: one RCT, two quasi-experimental design, two observational studies; and one retrospective analysis. Eight studies; in-house project report, external evaluation, semistructured interviews, case history, book chapters | Range 70–5983 | USA, Australia and the UK (one study); the UK (eight studies) |
Shek et al., 200982 | Crisis treatment (day hospital – schizophrenia) | To assess the effects of day hospital care vs. continuing outpatient care for people with schizophrenia and other similar severe mental illness | Four (published between 1966 and 1986) | 351 | USA |
Shepperd et al., 200965 | Crisis treatment (alternatives to inpatient care for children and young people) | To assess the clinical effectiveness, acceptability and cost of mental health services that provide an alternative to inpatient care for children and young people | Seven RCTs | 783 | USA, Germany and the UK (one RCT) |
Skalli and Nicole, 201183 | Promoting recovery (FEP services) | To assess the efficacy and efficiency of specialised FEP services and identify patients who benefit the most from such programmes | Five | 1633 | Scandinavian countries |
Stoffers et al., 201284 | Promoting recovery (psychotherapies – borderline disorder) | To assess the effects of psychological interventions for BPD | 28 | 1804 (ranging between 16 and 180 patients) | European countries (including two RCTs in the UK), USA, Canada and one study in New Zealand |
Thomas and Rickwood, 201385 | Crisis treatment (acute and subacute residential services – severe mental illness) | To assess the clinical effectiveness of acute and subacute residential services in providing transitional care to support people with mental illness | 26 [11 RCTs, one mixed methods (cross-sectional, qualitative), one longitudinal, two prospective, eight repeated measures, one cross-sectional, and two qualitative] | 5161 | USA, the UK (10 studies) and Norway |
Professor Sonia Johnson and Dr Brynmor Lloyd-Evans, personal communication | Crisis treatment (crisis intervention) | To assess the evidence regarding best practice for organisation and service delivery in CRTs | 63 (24 qualitative; 24 quantitative; and 15 guidelines) | Approximately 16,921 | Australia, USA, Canada, the UK, Norway and Germany |
Grading of Recommendations Assessment, Development and Evaluation profiles
Urgent and emergency access to crisis care
Psychiatric liaison teams compared with standard care for people presenting with mental health crisis
-
Patient or population: patients with for people presenting with mental health crisis.
-
Settings: ED.
-
Intervention: psychiatric liaison teams.
-
Comparison: standard care.
Outcome | Before implementation | Psychiatric liaison teams | Relative effect (95% CI) | Number of participants (studies) | Quality of the evidence (GRADE) | Comments |
---|---|---|---|---|---|---|
Readmission | Not estimable | 6613 (two studies) | ⊕⊝⊝⊝ very lowa | |||
Waiting times | Not estimable | 6901 (three studies) | ⊕⊝⊝⊝ very lowb,c |
Educational intervention compared with standard care for urgent and emergency access to crisis care
-
Patient or population: patients with urgent and emergency access to crisis care.
-
Settings: ED.
-
Intervention: educational intervention.
-
Comparison: standard care.
Outcome | Standard care | Educational intervention | Relative effect (95% CI) | Number of participants (studies) | Quality of the evidence (GRADE) | Comments |
---|---|---|---|---|---|---|
Violence | Not estimable | 479 (one study) | ⊕⊝⊝⊝ very lowa,b | |||
Medication use | Not estimable | 646 (one study) | ⊕⊝⊝⊝ very lowa,b |
Street/telephone triage compared with standard care for people in mental health crisis
-
Patient or population: patients with people in mental health crisis.
-
Settings: Mental Health Act-related events.
-
Intervention: street/telephone triage.
-
Comparison: standard care.
Outcome | Standard care | Street/telephone triage | Relative effect (95% CI) | Number of participants (studies) | Quality of the evidence (GRADE) | Comments |
---|---|---|---|---|---|---|
Mental Health Act detention | Not estimable | 673 (one study) | ⊕⊝⊝⊝ very lowa,b | |||
Police time on call | Not estimable | 4114 (one study) | ⊕⊝⊝⊝ very lowb |
Police officer training compared with standard care for people in mental health crisis
-
Patient or population: patients with people in mental health crisis.
-
Settings: Mental Health Act-related events.
-
Intervention: police officer training.
-
Comparison: standard care.
Outcome | Standard care | Police officer training | Relative effect (95% CI) | Number of participants (studies) | Quality of the evidence (GRADE) | Comments |
---|---|---|---|---|---|---|
Level of force used | Not estimable | 1257 (two studies) | ⊕⊝⊝⊝ very lowa | |||
Duration of Mental Health Act event | Not estimable | 194 (one study) | ⊕⊝⊝⊝ very lowb | |||
Arrest rate | Not estimable | 1063 (two studies) | ⊕⊝⊝⊝ very lowa,c |
Quality of treatment and care in crisis
Alternatives to inpatient treatment compared with standard care for young people with mental health problems
-
Patient or population: patients with young people with mental health problems.
-
Intervention: alternatives to inpatient treatment.
-
Comparison: standard care.
Outcome | Standard care | Alternatives to inpatient treatment | Relative effect (95% CI) | Number of participants (studies) | Quality of the evidence (GRADE) | Comments |
---|---|---|---|---|---|---|
Mental health symptoms | 560 (seven studies) | ⊕⊝⊝⊝ very lowa,b,c | ||||
General psychological functioning | 560 (seven studies) | ⊕⊝⊝⊝ very lowa,b,c |
Promoting recovery
Strengths-based service models compared with control for promoting recovery in people with mental health problems
-
Patient or population: people with mental health problems.
-
Settings: any.
-
Intervention: strengths-based service models.
-
Comparison: control.
Outcome | Strengths-based service models | Number of participants (studies) | Quality of the evidence (GRADE) | Comments |
---|---|---|---|---|
Level of functioning | The mean level of functioning in the intervention groups was 0.03 SDs higher (0.37 lower to 0.43 higher) | 314 (four studies) | ⊕⊝⊝⊝ very lowa,b,c | |
Mental health symptoms | The mean mental health symptoms in the intervention groups was 0.04 SDs higher (0.88 lower to 0.97 higher) | 150 (three studies) | ⊕⊝⊝⊝ very lowa,b,c | |
Quality of life | The mean quality of life in the intervention groups was 24 SDs lower (76.94 lower to 28.93 higher) | 105 (two studies) | ⊕⊝⊝⊝ very lowa,b,c |
Appendix 4 Primary studies data extraction and quality assessment
Study characteristics
Primary studies: study characteristics
Study | Study design and brief description of methods | Objectives | Number of patients | Number of centres/teams | Country/health-care system | Primary outcome | Time points |
---|---|---|---|---|---|---|---|
Improving access to crisis care in the ED | |||||||
Adams and Nielson, 2012104 | Pre-/post-process plan implementation Data collected on revisit rates to the psychiatric ED within 30 days of discharge (March 2009–March 2011) |
To provide evidence on the implementation of a proactive process improvement plan to decrease mean psychiatric emergency revisit rates | NR (medical centre has 532 inpatient beds, 252 located at one site, 66 of which are for inpatient psychiatry) | One hospital, two locations | USA | ED revisit rates | Follow-up: 10 months |
Cailhol et al., 2007102 | Data collected prospectively before (5 months) and after (5 months) introduction of a staff educational crisis intervention Patients assessed using an ad hoc questionnaire, completed by clinicians |
To investigate the impact of a staff educational crisis intervention on violent behaviour of patients admitted to EDs following drug suicidal attempt | Of 4791 patients consulting psychiatric ED, 478 patients (10%) had attempted drug suicide | One | Switzerland | Violent behaviour | 10 months |
McDonough et al., 200499 | Before-and-after data collected using a specifically designed EMTaCS database and data obtained from the Australasian Triage Scale (July 2001–July 2002) | To evaluate the impact of the EMTaCS on the ED | 604 patients (803 contacts) | One | Australia | Average nightly contact rate; mean time of contact; average time lag between arrival at ED and contact with EMTaCS Patient outcomes (e.g. hospitalisation, referrals to other services) |
12 months |
Pajonk et al., 2008103 | Prospective comparison of two urban settings with comparable sociodemographic profiles over 12 months (starting July 2003). Calls recorded and diagnoses collected (there may be multiple reasons for call) | To compare the frequency, diagnostic reliability and treatment of psychiatric emergencies in two different emergency medical systems | 5477 emergency situations: psychiatric emergencies (n = 646; 11.8%) | Two cities with a two-tiered emergency medical system – ambulance staffed with emergency medical technicians – defibrillation as the first tier The second tier comprises vehicles staffed with an emergency medical defibrillator and an emergency physician (mostly anaesthesiologists with additional training in pre-hospital emergency medicine) |
Germany | Diagnoses and treatment of psychiatric emergencies (defined as acute onset or exacerbation of a mental illness that may lead to danger to the life and health of the patient or others) | 12 months |
Sinclair et al., 2006100 | Crossover design study introducing a 3-month psychiatric nurse intervention into two busy UK A&E departments Standardised assessments completed for each patient, retrospective case note review, patient and staff surveys |
To assess the impact of a dedicated A&E psychiatric nurse service on several outcomes relevant to patients and clinicians | Patients presenting to A&E with mental health problems and 527 general A&E patients; 34 members of staff | Two A&E departments | UK | A&E attendance; referrals; repeat attendances | 12 months |
Strike et al., 2008105 | Exploratory qualitative design using semistructured interviews and iterative, thematic analysis A purposeful maximum variation sampling strategy was used to recruit participants between January 2004 and October 2004 Patient interviews (lasting from 45 minutes to 2 hours) were conducted after patients were discharged from the ED or after admission to the psychiatric ward Staff were recruited through posting flyers in high-traffic staff locations, by referral and word of mouth. Staff interviews lasted 30–90 minutes |
To explore the impact of, and attitudes towards, the ED psychiatric assessment rooms from the perspectives of patients and medical staff | Recruited after discharge from the ED (n = 18) Recruited after admission to the psychiatric ward (n = 7) Six nurses; five doctors; six non-medical ED staff members (i.e. crisis team workers and security officers) |
One large urban general hospital | Canada | Patient and medical staff experiences of psychiatric safe rooms | Approximately 10 months |
Tadros et al., 201397 | Evaluation of data already held by the City Hospital, Birmingham. RAID group referred to and managed by the RAID team (December 2009–July 2010) and the RAID-influence group not referred to RAID but had a discharge date between December 2009 and July 2010 In addition, a retrospective control group (pre-RAID) including patients admitted between December 2008 and July 2009 |
To evaluate whether or not an integrated rapid response mental health service improves access to psychiatric assessments and services, in addition to reducing costs of health service provision in acute hospitals | RAID group: 886 RAID-influence group: 2654 Pre-RAID control group: NR |
One hospital | England | Length of hospital stay and readmission rates (including length of survival in the community after discharge); cost savings | 6 months |
Wand et al., 2011101 | Mixed-methods study assessing service user perceptions of a MHNP service for people presenting to the ED | To assess service user satisfaction and experience of the MHNP service | 23 | One | Australia | Service user satisfaction and perceptions of the service | Study conducted between June 2008 and March 2010 |
Woo et al., 200798 | Retrospective cohort study comparing psychiatric emergency service model with psychiatric consultation model Participants were randomly selected from 3532 evaluations |
To compare the psychiatric emergency service model with the consultation model | 200 | One medical centre | USA | Admission to inpatient unit; timeliness of medical clearance; timeliness of psychiatric evaluation; mental state examination; urine toxicology; pregnancy testing; vital signs; emergency medication given; seclusion and restraint; elopement; follow-up care provided; readmission within 30 days | 1 year |
Police involvement in improving access to crisis care | |||||||
Compton et al., 2014109 | Data collected from March to November 2010 | To assess the effects of CIT training on levels of force and dispositions | 180 police officers (91 with CIT training); 1063 encounters | Six departments | USA | Level of force; disposition (resolution at the scene, referral or transport to services, or arrest) | NR |
Edmondson and Cummins, 201432 | Telephone/triage pilot project; modified mixed-method approach including qualitative data collection and thematic analysis (including semistructured individual and group interviews with police officers on the use of the pilot scheme), quantitative methods (December 2013–May 2014) and case study approaches to evaluation. Greater Manchester Police officers able to contact a 24-hour telephone number for professional advice and assistance from the NHS trust’s psychiatric liaison service, RAID team. RAID staff were experienced trained mental health workers (working with hospital colleagues) who supported people with mental health and/or alcohol problems | To assess the impact of the Oldham Phone Triage/RAID pilot project in relation to:
|
673 incidents used for quantitative analysis | One | England | Use of pilot service; use of Section136; hospitalisation; length of time officers spent with person; eventual outcome; police officer experiences with pilot, the strengths and deficits of the pilot, whether or not it had value and achieved its aims, identifying areas for future development | NR |
El-Mallakh et al., 2008110 | CIT programme instituted in 2001, began operation in January 2002 and expanded to include the county in 2003 Descriptive statistics presented and a test of significance of proportional data was employed |
To determine the impact of implementing a police CIT on police calls that involved people with mental illness | NR (number of CIT calls: 2269 in 2002; 2112 in 2004; 3068 in 2005; and 3237 in 2006) | NR | USA | Number of CIT calls; use of force; police and citizen injuries; special weapons and tactics callouts; hostage negotiation team callouts; arrest rates | 2002–4 |
Herrington and Pope, 2014111 | Multiphased mixed-methods pilot study initiated in New South Wales (January 2008–December 2009) Mental Health Act events recorded using computer operational policing system (3 months in 2007, 2008 and 2009) to represent comparison months before and after introduction of the intervention MHIT trained officers completed questionnaires at three time points (pre training, 2 months and 18 months post training). Non-MHIT trained participants recruited as a convenience sample and completed questionnaires Surveys completed by police and people with mental illness involved in a (systematic) sample of 171 Mental Health Act events recorded at four evaluations sites (October 2008–February 2009) Semistructured interviews undertaken by stakeholders and interpreted using thematic analysis Two focus groups conducted with people with mental illness Observational data from training, and four 12-hour shifts with police in four evaluation sites (192 hours) |
To assess a police initiated response to addressing incidents resulting from police interactions with people with mental illness | New South Wales Police Force: Members of MHIT project team (n = 9) Front-line police receiving MHIT training (n = 111) Front-line officers not receiving MHIT training (n = 74) New South Wales health participants: 5 Non-government organisations: 12 People with mental illness: 17 Ambulance service of New South Wales declined to participate in the research Mental Health Act events: n = 471 |
Nine sites | Australia | Mental Health Act events, including de-escalation, use of force and engagement with stakeholders | 2 months and 18 months after training |
Kisely et al., 2010107 | Mixed-methods design: controlled before-and-after comparison; qualitative assessments (interviews and focus groups) on views of service, using purposive sampling Service users and families or friends recruited through community and mental health settings, and advertisements Intervention implemented in June 2006 |
To assess the impact of an integrated mobile crisis team | NR | NR | Canada | Changes in availability and accessibility by contacts with crisis team and mental health calls to the police; efficiency (call-to-door time for visits); service user and staff satisfaction with the service | 1 and 2 years post implementation |
Steadman et al., 2000108 | Comparative cross-site descriptive design (October 1996–August 1997) | To compare three different models of police responses to calls for ‘emotionally disturbed persons’ | Approximately 100 police dispatch calls at each site, and an additional 100 calls from each site in which a specialised response occurred | Three sites | USA | Police responses and dispositions (taken to treatment location; situation resolved on scene; referred to treatment; arrested) | 10 months |
Primary studies: intervention details
Study | Intervention characteristics | Comparator characteristics |
---|---|---|
Improving access to crisis care in the ED | ||
Adams and Nielson, 2012104 | Description of intervention Proactive process improvement plan for psychiatric patients entering the ED within 30 days of an inpatient discharge: incorporating inpatient unit’s discharge instruction into the ED’s plan of care, involving the ED’s community-based on-call psychiatrists, standardise communication between ED nurses and on-call psychiatrists, include communication with outpatient care providers, utilise decisional support from the ED’s informatics system Setting Community-based not-for-profit psychiatric ED; one with a level 2 trauma service and a dedicated psychiatric area Implementation Psychiatric patients cleared by the ED’s triage process. Patients in the psychiatric emergency area remain under care, seen by a nurse who then contacts on-call psychiatrist. Patients tracked using a stand-alone ED information system and interfaces with the attached medical centre Fidelity to model NR |
Description of comparator N/A Setting N/A Implementation N/A Fidelity to model N/A |
Cailhol et al., 2007102 | Description of intervention Specific staff educational crisis intervention (educational programme for clinicians, nurses and security staff focusing on restraint and violent behaviour (early screening of potential violent behaviour); emphasis on regular communication between relevant staff; medical presence during all security interventions for restraint; debriefing after each intervention Setting Psychiatric ED Implementation Psychiatric emergency team (six residents and six psychiatrists), security and nurses Fidelity to model NR |
Description of comparator N/A Setting N/A Implementation N/A Fidelity to model N/A |
McDonough et al., 200499 | Description of intervention Night (21.00 to 07.30) EMTaCS (Thursday–Sunday for 3 months, then 7 nights per week) Setting ED Implementation Mental health Directorate in collaboration with ED Fidelity to model NR |
Description of comparator N/A Setting N/A Implementation N/A Fidelity to model N/A |
Pajonk et al., 2008103 | Description of intervention City 1: emergency physicians received standardised 3-hour educational course about symptoms, diagnoses, crisis intervention and pharmacological treatment of psychiatric emergencies. Health professionals provide daily briefings and regular supervision throughout the study Setting ED Implementation Emergency physicians Fidelity to model NR |
Description of comparator City 2: not informed about the study and did not receive educational meetings, briefings or supervision Setting Hospital ED Implementation Emergency physicians Fidelity to model NR |
Sinclair et al., 2006100 | Description of intervention Dedicated A&E psychiatric nurse service; 130 hours per week from 08.00 to 12.00. Clinical supervision provided by A&E nurse managers and, where necessary, a consultant psychiatrist Setting A&E department 1 Implementation Four experienced community psychiatric nurses Fidelity to model NR |
Description of comparator Dedicated A&E psychiatric nurse service; 130 hours per week from 08.00 to 12.00. Clinical supervision provided by A&E nurse managers and, where necessary, a consultant psychiatrist Setting A&E department 2 Implementation Same four experienced community psychiatric nurses Fidelity to model NR |
Strike et al., 2008105 | Description of intervention Psychiatric assessment (safe) rooms in EDs Setting Hospital ED Implementation Medical staff Fidelity to model NR |
Description of comparator N/A Setting N/A Implementation N/A Fidelity to model N/A |
Tadros et al., 201397 | Description of intervention Liaison psychiatry service delivering rapid response (within 1 hour), 24-hours, 7 days a week, age-inclusive service and comprehensive range of mental health specialities, including old age, working age, post-natal mental health and substance misuse Setting City Hospital, Birmingham Implementation The RAID model provides a multiprofessional team who provide formal and informal training to acute hospital staff Fidelity to model NR |
Description of comparator RAID-influence group (all emergency admissions with a mental health diagnosis not referred to RAID) Pre-RAID retrospective control group (all emergency admissions with a mental health diagnosis admitted before introduction of RAID) Setting City Hospital, Birmingham Implementation Hospital staff (not specified) Fidelity to model NR |
Wand et al., 2011101 | Description of intervention Mental health nurse practitioner outpatient service Setting ED Implementation Mental health nurse practitioner Fidelity to model NR |
Description of comparator N/A Setting N/A Implementation N/A Fidelity to model N/A |
Woo et al., 200798 | Description of intervention Psychiatric emergency service: provides immediate psychiatric assessment to minimise harm from inappropriate decisions for people in mental health crisis. Also provides a therapeutic environment for people in crisis to receive psychiatric, medical and social support Setting University of California Los Angeles-Kern Medical Centre Implementation NR Fidelity to model NR |
Description of comparator Consultation model: consultant psychiatrist providing assessment and treatment for mental health crises in the ED Setting University of California Los Angeles – Kern Medical Centre – before psychiatric emergency service had been implemented Implementation NR Fidelity to model NR |
Police involvement in improving access to crisis care | ||
Compton et al., 2014109 | Description of intervention Police-based CIT model (40-hour training) Setting NR Implementation Police officers Fidelity to model NR |
Description of comparator Police officers who had not received CIT training Setting NR Implementation Police officers Fidelity to model NR |
Edmondson et al., 201432 | Description of intervention RAID mental health phone triage pilot; available 24 hours a day, 7 days a week for use by officers at the scene of an incident. Enables speedy access to professional mental health expertise and advice; sharing and exchange of information; agreement on appropriate course of action Setting On scene Implementation Police officers Fidelity to model NR |
Description of comparator N/A Setting N/A Implementation N/A Fidelity to model N/A |
El-Mallakh et al., 2008110 | Description of intervention Police CIT programme Setting NR Implementation Police and special weapons and tactics officers Fidelity to model NR |
Description of comparator N/A Setting N/A Implementation N/A Fidelity to model N/A |
Herrington and Pope, 2014111 | Description of intervention MHIT; central project team to deliver training Setting NR Implementation Police officers Fidelity to model NR |
Description of comparator Sites neighbouring the pilot programme sites who interact with the same hospital Setting NR Implementation NR Fidelity to model NR |
Kisely et al., 2010107 | Description of intervention Integrated mobile crisis team formed in partnership between mental health services, municipal police and emergency health services. Offering short-term crisis management, including mobile interventions, 24-hour telephone support Setting On scene Implementation Mobile interventions attended by a plain-clothes police officer and a mental health professional Fidelity to model NR |
Description of comparator Area similar to intervention area but without access to such a service Setting NR Implementation NR Fidelity to model NR |
Steadman et al., 2000108 | Description of intervention Birmingham programme; police-based specialised mental health response. Community service officers (civilian police employees with professional training in social work or related fields) assist police officers in mental health emergencies by providing crisis intervention and some follow-up assistance. Community service officers participate in a 6-week classroom and field training programme; available Monday–Friday from 08.00 to 22.00, 24-hour coverage rotating on-call duty during weekends, holidays, and off-shift hours Setting On scene Implementation Community service officers Fidelity to model NR |
Description of comparator Memphis programme (police-based specialised police response): CIT with specially trained officers (40 hours of training) to transport individuals they suspect of having mental illness to psychiatric emergency service. Covering four overlapping shifts, providing 24-hour service. Knoxville programme (mental-health-based specialised mental health response): mobile crisis unit responding to calls in the community, handling telephone calls and referrals from prison. 24-hour coverage Setting On scene Implementation Police officers Fidelity to model NR |
Primary studies: participant characteristics
Study | Mean age | Mental health diagnosis | Comorbidities | Sex | Ethnicity | Other relevant characteristics |
---|---|---|---|---|---|---|
Improving access to crisis care in the ED | ||||||
Adams and Nielson, 2012104 | NR | NR | NR | NR | NR | NR |
Cailhol et al., 2007102 | NR | Patients attempting drug suicide and fulfilling the Diagnostic and Statistical Manual of Mental Disorders-Fourth Edition criteria for BPD, psychotic disorder, or major depression, life crisis situations, and alcohol or drug intoxication | NR | NR | NR | NR |
McDonough et al., 200499 | 34 years (range 15–90 years) | Drug and alcohol use, mood, disorder, psychosis and anxiety | NR | Male: 328/604 (54.3%) | NR | Reasons for contact:
|
Pajonk et al., 2008103 |
|
Severely impaired patients: city 1 (102/433, 23.4%); city 2 (68/210, 32.2%); OR 1.55 (95% CI 1.08 to 2.23) Emergency physician diagnosis alcohol intoxication: city 1 (143/433, 32.9%); city 2 (69/210, 32.7%); OR 0.99 (95% CI 0.70 to 1.41) State of agitation or violence: city 1 (131/433, 30.1%); city 2 (47/210, 32.7%); OR 0.66 (95% CI 0.45 to 0.98) Suicide attempt: city 1 (43/433, 9.9%); city 2 (31/210, 14.7%); OR 1.57 (95% CI 0.96 to 2.58) Psychosis: city 1 (33/433, 7.6%); city 2 (17/210, 8.1%); OR 1.07 (95% CI 0.58 to 1.96) Depression: city 1 (26/433, 6%); city 2 (9/210, 4.3%); OR 0.7 (95% CI 0.32 to 1.52) Leading psychiatric diagnoses after assessment Dementia/personality disorder: city 1 (1/433, 0.2%); city 2 (1/210, 0.5%); OR 2.07 (95% CI 0.13 to 33.20) Drug/substance misuse: city 1 (215/433, 49.4%); city 2 (109/210, 51.7%); OR 1.09 (95% CI 0.79 to 1.52) Schizophrenia/psychotic disorder: city 1 (31/433, 7.1%); city 2 (14/210, 6.6%); OR 0.93 (95% CI 0.48 to 1.78) Bipolar/mood disorders: city 1 (20/433, 4.6%); city 2 (7/210, 3.3%); OR 0.71 (95% CI 0.30 to 1.71) Anxiety/stress disorders: city 1 (116/433, 26.7%); city 2 (42/210, 19.9%); OR 0.68 (95% CI 0.46 to 1.02) |
NR | Male: city 1 (232/433, 54%); city 2 (118/210,56%) | NR | NR |
Sinclair et al., 2006100 | NR | NR (approximately half of the 411 patients seen by a psychiatric nurse had a psychiatric history). 92% (n = 37) of reviewed patient management plans formulated by the psychiatric nurses were judged appropriate and of good quality | NR | NR | NR | NR |
Strike et al., 2008105 | 20–29 years: 8/25 (32%) 30–39 years: 11/25 (44%) 40+ years: 6/25 (24%) |
Males were suicidal (acute and chronic), had a history of suicidality and current substance use disorder Axis 1 disorders
|
Alcohol abuse and dependent: 22/24 (92%) Substance abuse and dependent: 16/24 (67%) |
Only males were eligible for inclusion in the study | NR | Some men visited ED during a period of intense suicidal ideation or following a suicide attempt. Other men were assessed as suicidal during triage, and some men went to ED as a way to see a psychiatrist |
Tadros et al., 201397 | Patients referred from A&E: 36.4 years (range 16–64 years) Patients admitted because of self-harm: 34.5 years (range 16–64 years) The mean age of referrals from all wards was 65.7 years (range 16–64 years) |
Category F, ICD-10 diagnosis, including psychosis, schizophrenia, self-harm | NR | Male:
|
NR | NR |
Wand et al., 2011101 | 32 years (range 18–44 years) | NR | NR | 16 females and 7 males | NR | NR |
Woo et al., 200798 | 38 years | Most common diagnoses were schizophrenia (n = 19), bipolar disorder (n = 19) and depression (n = 17) | NR | 53 males and 47 females in both groups | NR | All participants were involuntarily admitted |
Police involvement in improving access to crisis care | ||||||
Compton et al., 2014109 | Mean age of officers: 36.6 years | NR | NR | Male: 135/180 (75%) | Caucasian: 93/180 (52%) African American: 75/180 (42%) Hispanic: 8/180 (4%) Native American or Pacific Islander: 2/180 (1%) Asian: 2/180 (1%) |
NR |
Edmondson and Cummins, 201432 | Of 40 cases resulting in use of Section 136: 36+ years: 12/40 (30%) 26–35 years: 16/40 (40%) 18–25 years: 7/40 (17.5%) Under 18 years: 5/40 (12.5%) |
Features of mental health-related calls Alcohol and/or drug use: 153/673 (23%) calls 67/673 (9.95%) calls identified no obvious mental health issues Self-harm; agitated, erratic and/or unsettled behaviour; paranoia or psychosis |
Alcohol and/or drug use; present in approximately 10/40 (25%) of Section 136 cases | Male:24/40 (60%) | NR | Prior contact with health and social care services Mental Health Care Programme approach: 5/40 (12.5%) Secondary health or social care services: 30/40 (75%) |
El-Mallakh et al., 2008110 | NR | NR | NR | NR | NR | NR |
Herrington and Pope, 2014111 | NR | NR | NR | NR | NR | NR |
Kisely et al., 2010107 | NR | NR | NR | NR | NR | NR |
Steadman et al., 2000108 | NR | NR | NR | NR | NR | NR |
Study outcomes
Primary studies: patient-related outcomes
Study | Self-harm | Violence | Service user experience |
---|---|---|---|
Improving access to crisis care in the ED | |||
Adams and Nielson, 2012104 | NR | NR | NR |
Cailhol et al., 2007102 | NR | 5 months prior to intervention (n = 254):
|
NR |
McDonough et al., 200499 | 257/604 (43%); 171 patients self-harmed as a result of situational crisis rather than an underlying psychopathology | NR | NR |
Pajonk et al., 2008103 | Correct identification and documentation of suicide attempts or suicide idea was low: city 1, 59.4%; city 2, 50% Certain suicide attempt: city 1, 52/106 (12%); city 2, 38/62, (29.3%); OR 1.62 (95% CI 1.03 to 2.55) Probable suicide attempt: city 1, 16//106, (3.7%); city 2, 10/62, (4.7%); OR 1.30 (95% CI 0.58 to 2.92) Suicidal ideation: city 1, 38/106 (8.7%); city 2, 14/62 (6.6%); OR 0.74 (95% CI 0.39 to 1.40) |
NR | NR |
Sinclair et al., 2006100 | NR | NR | Patient satisfaction: 31% (n = 156) of patients with mental health problems and 52% (n = 272) of general A&E patients completed a questionnaire. Levels of satisfaction were high for all patients with non-significant differences between intervention and non-intervention periods The most important concern for patients with mental health problems related to lack of empathy from A&E staff. For general A&E patients the most important issues were physical environment and waiting times |
Strike et al., 2008105 | All men reported prior ED admissions related to suicidal ideation or attempts. Despite repeated visits, the ED was frequently viewed as a last resort for care | NR | Many patients presented at the ED because they were alone, scared, and had few other social supports; being assigned to a psychiatric assessment room felt like a punishment. One patient stated:It was very jail-like and reminded me of prisonThe rooms might not have made the patients feel safe at all, rather than calming the patients, being both separated from other patients and observed was believed to escalate some symptoms, especially for patients with paranoia or depression:I spoke with uh one of the workers there, staff, and I explain. I, I, I believe I told him ‘you know what, I just cannot stay in this room by myself here.’ I feel like claustrophobic, scared, empty, nothingPatientThe length of time spent alone was also noted to contribute to negative experiences:I had to sit around and wait, and wait and wait and wait. If I have any disgruntlement about uh the emergency services it’s waiting, but unfortunately that’s the way it isPatientOften patients and staff spoke about the rooms as analogous to seclusion rooms used in the psychiatric inpatient setting particularly for violent patients. Furthermore, patients questioned the use of the rooms for non-violent patients by stating. . . I was put into a room with a bed with walls. It was pretty plain and I think they use it to lock up irate mental patients that they receive . . .Patient |
Tadros et al., 201397 | Referral to RAID team: 32% | NR | NR |
Wand et al., 2011101 | NR | NR | Most highly regarded aspects impacting on their experience of the service:
Service was perceived as easily accessible and free:I didn’t feel I had to wait around and ponder on whether it was worth it. It was dealt with really so fast and so positively that it was very effectiveLocating the service in the ED made it a lot easier and more likely to happenSubtheme: continuity of care Benefits of follow-up from the same person after presenting to ED:I just felt like I wasn’t really forgotten about . . . I felt like after being in an emergency, I was kind of taken care of in terms of going to see the . . . nurse practitioner after thatSubtheme: need for extended service Some participants suggested need for further staffing and service to be available outside business hours:If you are working a job, it is sometimes hard to get that time off during the day, and you don’t actually want to tell your employer that you’re going to see somebody because of a certain situation . . .Service user satisfaction (survey data) (n = 51): 74.5% (38/51 availability) and 68.6% (35/51 accessibility) strongly agreed the service was convenient. 70.6% (36/51) felt listened to and understood, 60.8% (31/51) that they had received useful information and health education, 74.5% (38/51) felt they had received support and encouragement, and 68.6% (35/51) were satisfied with overall standard of care provided |
Woo et al., 200798 | NR | NR | NR |
Police involvement in improving access to crisis care | |||
Compton et al., 2014109 | NR | NR | NR |
Edmondson et al., 201432 | 220/673 (32.7%) calls involved self-harm, threatened suicide or overdose (usually owing to personal crisis such as relationship breakdown or financial problems) | NR | NR |
El-Mallakh et al., 2008110 | NR | Patient injuries: 2002: 7/2269 (0.31%) 2004: 2/2112 (0.09%) 2006: 4/755 (0.53%) 2002 vs. 2004, p < 0.01; 2002 vs. 2006, non-significant; 2004 vs. 2006, non-significant Officer injuries: 2002, three officers (0.13%); 2004, 0; 2006, two officers (0.26%) |
NR |
Herrington and Pope, 2014111 | 37 self-inflicted injury | NR | Several people with mental illness expressed a feeling that police were fearful of them, and that this fear inadvertently escalated events:The way I see it is their assumption is you’ve gone psychotic or suicidal, I’m dangerous and they’re afraid of me. They’re afraid of what I’ll do. What I really need is for someone to say ‘you’re okay. You’re safe with us. We’re gonna help you’. But what they say is ‘come on hurry up, get in line, get in the car’, ‘Duck your head’. They’re aggressive towards me because I feel that they’re afraid of what I’m gonna do to them . . .Person with mental illness, non-MHIT area |
Kisely et al., 2010107 | 1534/3123 (49%) were for suicidal ideation | NR | Themes from focus groups for service users, family or friends, and mental health staff/police officers reflected the main themes identified by quantitative data. Service users found having someone to talk to, obtaining advice and support, and facilitating referral helpful:So I just felt completely isolated except for the ability to be able to talk to somebody from Mobile CrisisOne service userAvailability and accessibility: positive that services offered help to everyone, not just people with a fixed address; greater accessibility on the telephone; better understanding of mental health issues by police officers, and improved partnering on-scene:I think it is an essential service and I would love to see it expand even furtherOne key informant interviewee |
Steadman et al., 2000108 | NR | NR | NR |
Primary studies: service user outcomes
Study | Hospital admissions | Use of force/restraint | Waiting times | Mental health outcomes |
---|---|---|---|---|
Improving access to crisis care in the ED | ||||
Adams and Nielson, 2012104 | Mean psychiatric ED 30-day revisit rates for patients not admitted, 12 months prior to process plan: 6.51% (range 1.83–9.53%) At 6 months post implementation of process plan: 4.3% (range at 10 months, 3.53–5.56%) |
NR | NR | NR |
Cailhol et al., 2007102 | NR | NR | NR | NR |
McDonough et al., 200499 | Patient outcomes (803 contacts with ED): Admission for observation: 177/803 (22%) Mental health unit admission: 169/803 (21%) Medical admission: 24/803 (3%) |
NR | Average time lag between arrival at ED and contact with EMTaCS: Baseline (235 minutes) 12 months (36 minutes); 95% reduction in ‘seen by’ time |
NR |
Pajonk et al., 2008103 | NR | NR | NR | NR |
Sinclair et al., 2006100 | Referral rates per 100 A&E mental health attendees (n = 6097) admitted to general hospital ward: Pre-intervention: hospital 1 (n = 3957), 39.9%; hospital 2 (n = 2140), 48.2% Intervention: hospital 1, 38.8%; hospital 2, 41.7% Post-intervention: hospital 1, 43.8%; hospital 2, 44.9% Repeat attendances to A&E by patients with mental health problem: non-significant between intervention and non-intervention periods of both A&E departments |
NR | Mean waiting times from arrival until assessment and treatment [calculated for 4364/6097 (72%) of attendances at A&E] Pre-intervention Hospital 1: 1–3 months, n = 700; 65.9 minutes (SD 53.4 minutes), 95% CI 62.0 to 70.0 minutes Hospital 2: 1–6 months, n = 672; 50.5 minutes (SD 48.3 minutes), 95% CI 46.9 to 54.2 minutes Intervention Hospital 1: 4–6 months, n = 807; 62.9 minutes (SD 52.3 minutes), 95% CI 59.3 to 66.5 minutes Hospital 2: 7–9 months, n = 331; 46.0 minutes (SD 48.9 minutes), 95% CI 40.7 to 51.2 minutes Post-intervention Hospital 1: 7–12 months, n = 1514; 74.1 minutes (SD 54.7), 95% CI 71.3 to 76.8 minutes Hospital 2: 10–12 months, n = 340; 48.8 minutes (SD 46.3), 95% CI 43.8 to 53.7 minutes Total Hospital 1: n = 3021, 69.2 minutes (SD 54.0 minutes), 95% CI 67.3 to 71.1 minutes Hospital 2: n = 1343, 49.0 minutes (SD 48.0 minutes), 95% CI 46.4 to 51.5 minutes Both hospitals Pre-intervention: n = 1372, 58.4 minutes (SD 51.5 minutes), 95% CI 55.7 to 61.1 minutes Intervention: n = 1138, 58.0 minutes (SD 51.9 minutes), 95% CI 55.0 to 61.0 minutes Post-intervention: n = 1854, 69.4 minutes (SD 54.1 minutes), 95% CI 67.0 to 71.9 minutes Total hospitals 1 and 2 n = 4364, 63.0 minutes (SD 53.0 minutes), 95% CI 61.4 to 64.6 minutes Although there were few significant differences between pre and post intervention in the unadjusted analyses, adjusted analyses found that A&E attendance during the intervention period and seeing a psychiatric nurse were associated with shorter waiting times |
NR |
Strike et al., 2008105 | NR | Many staff believed that some patients deliberately misbehaved to ensure they would be restrained and assigned to the psychiatric room believing that they enjoyed the attention they received:He ends up in restraints every time he’s here but he likes that; he likes the attentionED nurseSome staff acknowledged that disruptive patients were likely to be seen sooner than calmer patients. Some patients admitted to using disruptive behaviour (e.g. yelling, threatening) to avoid an overly-long waiting time and ensure that a physician would see them more quickly:She [the crisis nurse] came from the Crisis Centre Unit, so she was coming in every little while, and she would say ‘you have to wait, you have to wait; that’s all there is to it’ and she was very good to me. However, I don’t think she did enough to speed up the process and it wasn’t until I went into a temper tantrum that the process was speeded up like thatPatient | NR | NR |
Tadros et al., 201397 | Mean length of hospital stay: RAID: 9.4 days Pre-RAID: 10.3 days 79 cases (p = 0.31); 0.9 days saved per patient (797 days saved over 8 months) RAID-influence group: 5.2 days Pre-RAID: 8.4 days 359 cases (p < 0.001); 3.2 days saved per patient (8493 days saved over 8 months). Total savings over 8 months, 9290 bed-days in city hospital; savings over 12 months, 13,935 bed-days, equating to 38 beds per day Discharge effect: 75% of patients were discharged within 1 week and 65% were discharged within 3 days of seeing RAID independent of their length of stay |
NR | NR | Referral to RAID team: Depression: 18% Alcohol-related problems: 13% Psychosis: 9% Anxiety: 4% Drug misuse: 4% Others (not specified): 2% Increases in mental health diagnosis at city hospital: Eating disorders: 46% Other anxiety disorders: 44% Dementia: 22% Mental and behavioural disorders due to alcohol: 21% Depression and related disorders: 17% Schizophrenia: 8% |
Wand et al., 2011101 | NR | NR | NR | NR |
Woo et al., 200798 | Admission to inpatient unit (n = 100 each group) consultation model: 47%, PES model: 52%; p = 0.48 Readmission within 30 days (n = 100 each group) consultation model: 28%, PES model: 24%; p = 0.52 |
Seclusion and restraint (n = 100 each group) consultation model: 15%, PES model: 6%; p = 0.05 | Timeliness of psychiatric evaluation (n = 100 each group) consultation model: mean 639 minutes, SD 508 minutes PES model: mean 330 minutes, SD 291 minutes; p < 0.01 Timeliness of medical clearance (n = 100 each group) consultation model: mean 328 minutes, SD 333 minutes PES model: mean 377 minutes, SD 317 minutes; p = 0.29 |
NR |
Police involvement in improving access to crisis care | ||||
Edmondson and Cummins, 201432 | 66/673 (9.8%) calls Outcomes of Section 136: converted to ‘informal’ (voluntary) admission to hospital, 17/40 (42.5%) Converted to compulsory detention to hospital lasting up to 28 days: 4/40 (10%) Converted to compulsory detention to hospital lasting up to 6 months: 1/40 (2.5%) Section 136 completed (discharge with follow-up from specialist services): 15/40 (37.5%) Section 136 completed (discharged home with primary care follow up or no further action): 3/40 (7.5%) |
NR | NR | Section 136 considered: 78/673 (12%) calls Section 136: used 43; 40 analysed, of which police officers contacted RAID on 31 (77.5%) occasions. Of the three incidents where RAID was not contacted, two already had mental health professionals on scene, one call to RAID went unanswered Section 136 detention (by calendar year):
Reasons for use of Section 136:
|
El-Mallakh et al., 2008110 | NR | Use of force:
|
NR | NR |
Herrington and Pope, 2014111 | NR | Coercive force: 58/471 Mental Health Act events; no significant difference in coercive force between months, and no difference between trained mental health officers and non-trained officers Injury resulting from police actions, n = 4 Transportation of people with mental illness and use of force Use of force: transport by police (n = 16); ambulance (n = 2); and other means (n = 0) No use of force: transport by police (n = 55); ambulance (n = 18); and other means (n = 2) |
NR | NR |
Kisely et al., 2010107 | NR | NR | Police time on scene Pre-intervention: intervention (n = 798) mean 185 minutes (SD 185 minutes) vs. control (n = 258) mean 132 minutes (SD 130 minutes) 1 year post-intervention: intervention (n = 1058) mean 161 minutes (SD 160 minutes) vs. control (n = 349) mean 164 minutes (SD 165 minutes) 2 years post-intervention: intervention (n = 1267) mean 136 minutes (SD 136 minutes) vs. control (n = 384) mean 165 minutes (SD 165 minutes); intervention vs. control (p < 0.001) Call-to-door time 1 year post-intervention: intervention (n = 172) mean 73 minutes (SD 81 minutes) 2 years post-intervention: intervention (n = 552) mean 36 minutes (SD 39 minutes) |
Global Assessment of Functioning scores post-intervention [318/1093 (32%) mobile visits]: mean score 52 (SD 17.7), representing moderate symptoms/difficulty in social, occupational, or school functioning |
Steadman et al., 2000108 | NR | NR | NR | NR |
Primary studies: other outcomes
Study | Other relevant outcomes |
---|---|
Improving access to crisis care in the ED | |
Adams and Nielson, 2012104 | NR |
Cailhol et al., 2007102 | NR |
McDonough et al., 200499 | Average nightly contact rate: baseline (2.9 patients); 12 months (3.75 patients) Mean time of contact: 1 hour 40 minutes Patient outcomes (803 contacts with ED); review and management by other ED staff: 128/803 (16%); discharged home: 128/803 (16%) Discharged with social worker follow-up: 16/803 (2%) Discharged with community mental health nurse follow-up: 88/803 (11%) Discharged to inpatient psychiatric day programme: 8/803 (1%) Discharged to police: 8/803 (1%) Left without plan: 44/803 (6%) Referral to other mental health services: 8/803 (1%) |
Pajonk et al., 2008103 | Treatment transport to hospital: city 1, 355/433 (83.1%); city 2, 188/210 (92.6%); OR 2.54 (95% CI 1.42 to 4.56) Any medication: city 1, 278/433 (63.9%); city 2, 153/210 (72.5%); OR 1.49 (95% CI 1.04 to 2.14) Non-pharmacological crisis intervention: city 1, 71/433 (16.3%); city 2, 18/210 (8.5%); OR 0.48 (95% CI 0.28 to 0.83) |
Sinclair et al., 2006100 | Number of patients assessed and psychiatric nurse time utilised. Approximately one-third (411) of patients with mental health problems presenting to A&E while a psychiatric nurse was on duty were referred to the psychiatric nurse team, with just over 90% of these patients assessed Referral rates per 100 A&E mental health attendees (n = 6097) Discharged to GP care Pre-intervention [hospital 1 (n = 3957), 36.0%; hospital 2 (n = 2140), 28.5%] Intervention (hospital 1, 36.4%; hospital 2, 37.1%) Post-intervention (hospital 1, 35.5%; hospital 2, 35.4%) Self-discharged Pre-intervention (hospital 1, 11.8%; hospital 2, 10.1%) Intervention (hospital 1, 12.1%; hospital 2, 7.8%) Post-intervention (hospital 1, 13.4%; hospital 2, 9.2%) Transferred to psychiatric hospital Pre-intervention (hospital 1, 5.7%; hospital 2, 8.9%) Intervention (hospital 1, 5.6%; hospital 2, 7.6%) Post-intervention (hospital 1, 3.7%; hospital 2, 7.2%) Referred to outpatient clinic Pre-intervention (hospital 1, 4.0%; hospital 2, 3.0%) Intervention (hospital 1, 4.0%; hospital 2, 2.0%) Post-intervention (hospital 1, 2.7%; hospital 2, 2.3%) Other Pre-intervention (hospital 1, 0.8%; hospital 2, 1.1%) Intervention (hospital 1, 1.8%; hospital 2, 2.2%) Post-intervention (hospital 1, 0.6%; hospital 2, 0.5%) Not recorded Pre-intervention (hospital 1, 1.8%; hospital 2, 0.2%) Intervention (hospital 1, 1.3%; hospital 2, 1.6%) Post-intervention (hospital 1, 0.3%; hospital 2, 0.5%) |
Strike et al., 2008105 | In general, comments about the psychiatric assessment room were negative and focused on both the physical attributes and location and the emotional responses it evoked. The psychiatric assessment rooms were designed for patients requiring a quiet environment with close supervision. However, staff commented that the ED setting was not appropriate for patients in need of a calm environment Staff also noted that the rooms were intentionally stark and lacking in furnishings, decoration, or freestanding equipment in order not to impede the work of the staff and to reduce the injury to self, staff, or other patients. Patients also commented negatively on the physical aspects of the room |
Tadros et al., 201397 | Response times (RAID assessments): A&E referrals: 113 of 124 (91%) per month assessed within 1 hour (average time 24 minutes) Ward referrals: 788 of 886 (89%) assessed within 24 hours (average time 16 hours) Patient survival rate RAID group vs. pre-RAID group and RAID-influence group: p < 0.0001 RAID-influence group vs. pre-RAID: p = 0.21 Readmissions RAID influence vs. RAID group: HR 2.80, 95% CI 2.68 to 2.92 Pre-RAID vs. RAID group: HR 2.45, 95% CI 2.33 to 2.57 |
Wand et al., 2011101 | NR |
Woo et al., 200798 | For all outcomes n = 100 for each group Mental State Exam conducted Consultation model: 49% PES model: 95%; p < 0.01 Emergency medication given Consultation model: 74% PES model: 53%; p = 0.01 Elopement Consultation model: 13% PES model: 5%; p = 0.05 Follow-up care provided Consultation model: 58% PES model: 69%; p = 0.11 |
Police involvement in improving access to crisis care | |
Compton et al., 2014109 | Disposition Situations resolved: 47% CIT officers/48% non-CIT officers; OR 0.96, non-significant Referrals: 13% CIT officers/24% non-CIT officers; OR 1.70, significant (p < 0.05) Arrests: 13% CIT officers/24% non-CIT officers; OR 0.47, significant (p < 0.05) Disposition by condition Arrests: Mental illness only (40%): CIT trained officers (n = 16)/non-CIT trained officers (n = 23); OR 0.34 Mental illness and drug or alcohol problem (10%): CIT trained officers (n = 10)/non-CIT trained officers (n = 10); OR 1.09 Drug or alcohol problem only (34%): CIT trained officers (n = 34)/non-CIT trained officers (n = 81); OR 0.68 Disability or nothing indicated (16%): CIT trained officers (n = 7)/non-CIT trained officers (n = 18); OR 0.26 Situations resolved (number of encounters): Mental illness only (40%): situations resolved (number of encounters) – CIT trained officers (n = 111)/non-CIT trained officers (n = 94); OR 0.96 Mental illness and drug or alcohol use (10%): situations resolved – CIT trained officers (n = 19)/non-CIT trained officers (n = 25); OR 0.60 Drug or alcohol problem only (34%): CIT trained officers (n = 53)/non-CIT officers (n = 106); OR 0.73 Disability or nothing indicated (16%): CIT trained officers (n = 62)/non-CIT trained officers (n = 35); OR 1.90 Referrals: Mental illness only (40%): CIT trained officers (n = 104)/non-CIT trained officers (n = 74); OR 1.54 Mental illness and drug or alcohol use (10%): CIT trained officers (22)/non-CIT trained officers (n = 18); OR 1.52 Drug or alcohol problem only (34%): CIT trained officers (n = 45)/non-CIT officers (n = 43); OR 2.16 Disability or nothing indicated (16%); CIT trained officers (n = 34)/non-CIT trained officers (n = 21); OR 1.06 |
Edmondson and Cummins, 201432 | Use of RAID: 217/673 (32%) calls Length of time spent with individuals: resolved within 1 hour – 669/673 (99.4%) Of the 40 Section 136 cases, 30/40 (75%) resolved within 1 hour; 5/40 (12.5%) delays of over 4 hours Write up using the national decision-making model: 282/673 (42%) calls. However, this may be misleading as officers following this model may not be formally identifying that they are doing so Police officers strongly indicated that access to RAID improved communication, co-ordination of police – RAID work, timeliness of interventions and completion of assessments within satisfactory time scales Challenges to the use of RAID included unanswered calls, access to radio sets and mobile phones to use RAID; disagreements between police and health staff on appropriate designated Place of Safety |
El-Mallakh et al., 2008110 | Prison mental health unit occupancy remained unchanged: 1123 pre CIT in 2001; 1244 post CIT in 2002; 1165 in 2003; 1027 in 2004; and 1347 in 2005 Intense psychiatric service referrals significantly decreased: 595 (53%) pre CIT in 2001; 275 (26.8%) in 2004; p < 0.01; and 320 (23.8%) in 2005; p < 0.01 Arrests rates: CIT arrest rates (2.1%) vs. overall arrest rate (6.2%); p < 0.01 |
Herrington and Pope, 2014111 | Interactions between police and people with mental illness: Most of the 42 trained officers believed the training changed their approach to mental health-related events; half believed the changes were considerable, particularly relating to empathy. Self-reported behaviour change (39 police officers, 18 months post-training): Better understand relations with other (health) agencies: n = 5 More calm approach: n = 2 Take more time/patience: n = 6 Understand/empathise more: n = 16 More aware of own behaviour when interacting/confidence: n = 4 Better planned approach/use of other options: n = 6 Confidence: pre training vs. 2 months post training (p < 0.001); pre training vs. 18 months post training (p < 0.001); 2 months vs. 18 months post training (p = not significant) De-escalation: 344 events; no differences found between trained and non-trained officers in terms receiving schedule 1 (mental health-related) events, in conveying people with mental health problems to health-care services, or in average duration of Schedule 1 events. However, self-reported data suggested trained officers perceived a reduction in average duration of events from 106 to 56 minutes |
Kisely et al., 2010107 | Availability and accessibility: 12 months pre-intervention – crisis team responded to 464 people; 2793 contacts Year 1 of intervention: 1414 people, 5911 contacts Year 2 of intervention: 1666 people, 7558 contacts Outpatient attendance: pre-intervention (8380 visits) vs. year 1 of intervention (8523 visits). After adjusting for confounders (e.g. mental health service use), people who had contacted the services (n = 295) showed greater engagement than controls as measured by outpatient contacts (p < 0.001) |
Steadman et al., 2000108 | Specialised response on the scene Birmingham, 28/100 (28%); Knoxville, 40/100 (40%); Memphis, 92/97 (95%); total, 160/297 (54%); p < 0.001 for the difference between programmes. Knoxville had lengthy response time, which posed a significant barrier to use of the service by police. In Birmingham there were only six community service officers for a police force of 921, severely restricting the availability of the officers with special training Disposition Taken to treatment location: Birmingham, 20%; Knoxville, 42%; Memphis, 75%; total, 46% Situation resolved on the scene: Birmingham, 64%; Knoxville, 17%; Memphis, 23%; total, 35% Referred to treatment: Birmingham, 3%; Knoxville, 36%; Memphis, 0%; total, 13% Arrested: Birmingham, 13%; Knoxville, 5%; Memphis, 2%; total, 7%; p < 0.001 for the difference in dispositions between programmes |
Study quality
Primary study quality assessment (comparator studies)
First author (year) | Adequate allocation concealment | Random sequence generation | Blinding of participants | Blinding of personnel | Blinding of outcome assessors | Incomplete outcome data | Selective outcome reporting | Other sources of bias |
---|---|---|---|---|---|---|---|---|
Improving access to crisis care in the ED | ||||||||
Pajonk et al., 2008103 | No | No | Unclear | No | No | Yes | Unclear | Unclear |
Sinclair et al., 2006100 | No | No | Unclear | No | No | Yes | Unclear | Unclear |
Tadros et al., 201397 | No | No | No | No | Unclear | Unclear | Yes | Yes (selection bias) |
Woo et al., 200798 | No (not a RCT) | No (not a RCT but aimed to match on age, sex, ethnicity, and primary diagnosis) | Unclear | Unclear | Unclear | None | Unclear | Unclear |
Police involvement in improving access to crisis care | ||||||||
Compton et al., 2014109 | No | No | No | No | No | Yes | Unclear | Unclear |
Herrington and Pope, 2014111 | No | No | No | No | Unclear | Unclear | Unclear | Unclear |
Kisely et al., 2010107 | No | No | No | No | No | Yes | Unclear | Unclear |
Steadman et al., 2000108 | No | No | No | No | Unclear | Unclear | Yes | Unclear |
Primary study quality assessment (no comparator)
First author (year) | Study sample representative of population | Loss to follow-up unrelated to key characteristics | Prognostic factor adequately measured in study participants | Outcome of interest adequately measured | Important potential confounders are appropriately accounted for | Statistical analysis is appropriate for the design of the study |
---|---|---|---|---|---|---|
Improving access to crisis care in the ED | ||||||
Adams and Nielson, 2012104 | Yes | Unclear | Unclear | Yes | Unclear | Unclear |
Cailhol et al., 2007102 | Yes | Unclear | Unclear | Unclear | Unclear | Yes |
McDonough et al., 200499 | Yes | Unclear | Yes | Yes | Unclear | Yes |
Police involvement in improving access to crisis care | ||||||
Edmondson and Cummins, 201432 | Yes | Unclear | Unclear | Unclear | No | Yes |
El-Mallakh et al., 2008110 | Yes | Unclear | Unclear | Unclear | Unclear | Yes |
Primary study quality assessment (qualitative)
Study | Clear statement of aims | Qualitative methodology appropriate | Research design appropriate | Recruitment strategy appropriate | Data collected in a way that addressed the issue | Relationship between researcher and participant adequately considered | Ethical issues addressed | Data analysis sufficiently rigorous | Clear statement of findings | Value of the research |
---|---|---|---|---|---|---|---|---|---|---|
Improving access to crisis care in the ED | ||||||||||
Strike et al., 2008105 | Yes | Yes | Yes | Yes | Yes | Unclear | Yes | Yes | Yes | The data are valuable in terms of highlighting experiences in a specific service user population |
Wand et al., 2011101 | Yes | Yes | Yes | Unclear | Yes | Unclear | Yes | Yes | Yes | It has some value in terms of service user satisfaction and perceptions |
List of abbreviations
- A&E
- accident and emergency
- ACT
- assertive community treatment
- BME
- black and minority ethnic
- BPD
- borderline personality disorder
- CBT
- cognitive–behavioural therapy
- CI
- confidence interval
- CIT
- crisis intervention team
- CMHT
- Community Mental Health Team
- CORE
- Crisis resolution team Optimisation and RElapse prevention
- CQC
- Care Quality Commission
- CRD
- Centre for Reviews and Dissemination
- CRHTT
- crisis resolution and home treatment team
- CRT
- crisis resolution team
- DARE
- Database of Abstracts of Reviews of Effects
- DBT
- dialectical behaviour therapy
- DH
- Department of Health
- DUP
- duration of untreated psychosis
- ED
- emergency department
- EIS
- early intervention service
- FEP
- first-episode psychosis
- GP
- general practitioner
- GRADE
- Grading of Recommendations Assessment, Development and Evaluation
- HR
- hazard ratio
- HTA
- Health Technology Assessment
- IAPT
- Improving Access to Psychological Therapies
- ICER
- incremental cost-effectiveness ratio
- ICM
- intensive case management
- IPS
- individual placement and support
- JCP
- joint crisis plan
- LGBT
- lesbian, gay, bisexual and transgender
- MBCT
- mindfulness-based cognitive therapy
- MBT
- mentalisation-based therapy
- MHIT
- mental health intervention team
- NCCMH
- National Collaborating Centre for Mental Health
- NHS EED
- NHS Economic Evaluation Database
- NICE
- National Institute for Health and Care Excellence
- NIHR
- National Institute for Health Research
- OR
- odds ratio
- PANSS
- Positive And Negative Syndrome Scale
- PSS
- Personal Social Services
- QALY
- quality-adjusted life-year
- RAID
- Rapid, Assessment, Interface and Discharge
- RCT
- randomised controlled trial
- RR
- risk ratio
- SD
- standard deviation
- SMD
- standardised mean difference
- UCL
- University College London