An evidence synthesis of risk identification, assessment and management for young people using tier 4 inpatient child and adolescent mental health services

Article history

The research reported in this issue of the journal was funded by the HS&DR programme or one of its preceding programmes as project number 11/1024/08. The contractual start date was in February 2013. The final report began editorial review in May 2014 and was accepted for publication in October 2014. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HS&DR editors and production house have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the final report document. However, they do not accept liability for damages or losses arising from material published in this report.

Declared competing interests of authors

Dr Gemma Trainor is employed as a consultant nurse in child and adolescent mental health services.

Copyright statement

© Queen’s Printer and Controller of HMSO 2015. This work was produced by Hannigan 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.

Mental health in children and young people

One in 10 children and young people between the ages of 5 and 16 years living in Britain has a diagnosable mental health problem. 1 In England the total number affected is projected to increase by over 13% in the period to 2026. 2 In this context a priority for the NHS and its partner agencies is to make sure that the needs of each child are met in a tailored and timely way. To this end, child and adolescent mental health services (CAMHS) in England and Wales continue to be organised using a tiered approach. 3^,4 This is represented in Box 1.

Inpatient child and adolescent mental health services: a component within a complex system

With multiple groups of people and organisations located at different tiers, all interacting in mutual interdependence, CAMHS are an example of a complex system. 5–7 Within this system, as Box 1 shows, the most specialised services are available at tier 4 to young people with the greatest need. Those who use services at this level often have multiple disorders and difficulties, and, until relatively recently, tier 4 was largely synonymous with hospital care. 8 New service developments reflect the idea that care at this highest level should be provided in the least restrictive environment possible. Against this background a team funded by the National Institute for Health Research (NIHR) Service Delivery and Organisation programme systematically reviewed alternatives to hospital admission for children and young people and the evidence of their effectiveness, acceptability and cost. 9 This team described a number of alternatives to inpatient care in a typology of evaluated models, and from its mapping exercise reported a variety of services in use across the UK.

This evidence, plus evidence secured by the independent CAMHS review team3 and by the National CAMHS Support Service,8 points to a diversification at tier 4 which includes an expanded array of highly specialised and/or intensive community and out-of-hospital services. However, inpatient CAMHS units continue to play a major part in overall systems of mental health care, and, reflecting the terms of the NIHR Health Services and Delivery Research (HSDR) programme’s call under which this project was funded (NIHR HS&DR 11/1024: Innovations in secondary mental health services), it is hospital services that have exclusively been focused on here. Over the lifetime of this evidence synthesis, important reminders have appeared of the need for locally accessible, age-appropriate, hospital-based mental health care for children and young people. New data published in February 2014 following a joint British Broadcasting Corporation (BBC)/Community Care investigation pointed to the continuing problem of young people being admitted to adult wards and to wards many miles from their homes, with detrimental effects on individuals and on the maintenance of relationships with family and friends. 10 This, plus the fact that highly specialised, institutional health care also constitutes a substantial component of health service costs, makes this evidence review in the tier 4 hospital services field particularly timely.

At the turn of the new century it was estimated that over 2000 young people are admitted to English and Welsh CAMHS inpatient units each year, with the majority of specialist centres catering exclusively for those over 11 years old. 11 Variations in the characteristics of young people admitted are believed to exist, reflecting differences in the socioeconomic features of regions and differing levels of bed availability. Pressure on inpatient beds is considerable, and many who are referred for inpatient treatment are not accepted. 12 The admission of young people to adult mental health wards has been recognised as a problem for some time,13 with the recent BBC/Community Care investigation including vivid first-person testimony of the difficulties this causes. The increased use of adult wards in England has also happened in the face of active monitoring by the Care Quality Commission, which requires services to report each occasion on which an under-18-year-old is placed on a ward intended for adults for a continuous period of more than 48 hours. 14

It is notable that, until relatively recently, little was known of the actual interventions offered to young people admitted to mental health hospitals or the advantages of providing inpatient care. This general situation is changing, helped by the commissioning (including by the NIHR Service Delivery and Organisation programme, the predecessor to the body funding this current evidence synthesis) of investigations such as Tulloch et al. ’s15 into costs, outcomes and satisfaction in inpatient CAMHS. The wider evidence base has also been strengthened by new knowledge of systems and processes supporting young people making the transition from CAMHS to adult mental health services. 16 What has not been attempted hitherto is a synthesis of the evidence in the area of risk in the way that has been accomplished here.

Identifying, assessing and managing risk: the case for an evidence synthesis

Within the inter-relating CAMHS system, the identification, assessment and management of risk are key considerations for practitioners and managers working at the interface between community and hospital services. Work across this interface includes making decisions on the transfer of young people from one tier to another. Decisions made at this juncture, and risk-related decisions made in the within-hospital context, can have lasting consequences for young people, families and services. In preparing our bid for NIHR support we were aware of individual (and sometimes small-scale) investigations being conducted in the areas of risk identification, assessment and management in inpatient CAMHS. To the best of our knowledge, however, no previous studies have systematically brought together research and other evidence in the way that has been done here.

The word ‘risk’ in everyday mental health services is overwhelmingly used as a shorthand referring primarily to the possibilities of direct harm to self or others, or harm through self-neglect and physical deterioration. Risk-management interventions, including the admission of people to hospital as places of safety in which round-the-clock care and observation can be provided, are then introduced as means of minimising the likelihood of harm happening. Action in response to identified risks of this type is vitally important for young people. Beyond this, an ambition of this project has also been to search for evidence in the area of inpatient CAMHS risk where risk is understood in the broadest of senses, with the word itself having a number of different meanings. ‘Risky behaviour’ and ‘posing a risk’ are two, correlating closely with the dominant ways in which risk is thought of in mental health services but contrasting with the ideas of ‘risk factors’ and ‘being at risk’. 17 To this Coleman and Hagell17 add the idea of ‘risk reframing’, through which behaviours typically seen as risky might be reinterpreted as opportunities to develop resilience. In the mental health service context, this connects with the idea of ‘positive risk-taking’,18 used as a route to the promotion of individual responsibility and personal development.

As a project team we view risk as complex and multifaceted. Our original case for funding support included the idea that, in addition to the risks of harm to self and others and self-neglect, attention ought also to be given to the identification, assessment and management of other, less obvious, risks for young people using inpatient mental health services. In a context of bed scarcity and regional variations in patterns of provision, it is known, for example, that CAMHS clinicians describe the most significant reasons for hospital admission as a young person’s high risk of suicide, risk of physical deterioration owing to mental illness, need for round-the-clock observation and the presence of serious deliberate self-harming behaviour. 12 Pursuing the example of self-harm is illustrative of the thinking behind this study. Although good practice involves care and treatment in the least restrictive environment possible, for young people who seriously self-harm, safety can be difficult to achieve in the community, making this risk a common trigger for hospital referral. Anecdotal, practice-based evidence included in our original bid for support suggests that managers and professionals in hospitals also find self-harm difficult to manage because of problems of contagion. Contagion is the copying of harmful behaviours and, while important, is a risk that may not feature highly when hospital admission decisions are made. It is also a risk that may be underexamined in the research context, and about which little is therefore known.

Informing this project is also the idea that mental ill-health and hospital admission present risks to young people’s achievement of developmental milestones, psychological maturity, educational attainment, social integration with family and peers, and personal physical well-being. This overall perspective has informed the production of an evidence synthesis which embraces a broad view of risk, along with the idea that action to reduce the chances of one type of unwanted event happening potentially increases the chances of another occurring. For example, admitting a young person to hospital may represent a reasoned response to the risk of harm to self but is also an action potentially increasing the risk of other, undesired, events happening (including contagion, but also disruption to family and friendship networks, educational continuity and increased stigma). Just as risks are connected, so too are the people and services who might collectively respond to them. Addressing the full range of risks for young people admitted to mental health hospital is likely to draw on the efforts of workers located across the system: in health, social care and education services. The intention in this report has been to bring together the available evidence in ways that are helpful to this broad group of workers and to young people using services and their families.

Overarching research question and objectives

This project was funded under a NIHR HSDR programme call focusing on innovations in secondary mental health services in order to answer the overarching research question:

What is known about the identification, assessment and management of risk (where ‘risk’ is broadly conceived) in young people (aged 11–18 years) with complex mental health needs entering, using and exiting inpatient child and adolescent mental health services in the UK?

Objectives for the overall project were:

1. to summarise and appraise the evidence for the identification, assessment and management of risk for young people: as they make the transition into inpatient CAMHS; as they are cared for in inpatient CAMHS; as they make the transition from inpatient CAMHS to the community; and as they make the transition from inpatient CAMHS to adult mental health services

2. to identify and describe any underlying theoretical explanations for approaches used in the identification, assessment and management of risk

3. to understand the views and experiences of risk of young people (aged 11–18 years) with complex mental health needs using inpatient mental health services, and of those involved in the identification, assessment and management of risk in these settings

4. to synthesise the evidence for the identification, assessment and management of risk in young people (aged 11–18 years) with complex mental health needs entering, using and exiting inpatient services

5. to synthesise the evidence on the costs and cost-effectiveness to the NHS of different approaches to identifying, assessing and managing these risks

6. to identify the future priorities for commissioning, service development and research for young people (aged 11–18 years) with complex mental health needs entering, using and exiting tier 4 inpatient services.

A two-stage framework to evidence synthesis: the Evidence for Policy and Practice Information and Co-ordinating Centre approach

The project team was commissioned to meet these objectives using a specific approach to the identification, review and synthesis of the evidence. In recent years a variety of review approaches suitable for the health-care field have emerged,19 of which the EPPI-Centre (Evidence for Policy and Practice Information and Co-ordinating Centre) framework is one. This has developed over a period of two decades, led by researchers located primarily at the Institute of Education at the University of London and their affiliates. Documents outlining EPPI-Centre methods (and other materials) are freely available from http://eppi.ioe.ac.uk/cms/, where reviews informed by this framework are intended to have particular utility in informing policy and practice for public services. Recent published examples modelling varying degrees of detail and complexity completed by EPPI-Centre members and their collaborators include syntheses of the evidence for the socioeconomic value of nursing and midwifery,20 on commissioning in health, education and social care21 and on the effectiveness of interventions to strengthen national health service delivery in low- and lower middle-income countries. 22 Previous examples also exist of NIHR-funded projects using the broad EPPI-Centre approach. These include an HSDR programme project addressing self-care support for children and young people led by a member of the current research team (SP),23 and a Public Health Research Programme review into the effects of schools and school environment interventions on health. 24

Our use of the EPPI-Centre approach throughout the totality of this project is summarised in Figure 1.

FIGURE 1.

Describing the EPPI-Centre framework. 25

Like other frameworks guiding the review and synthesis of evidence, the EPPI-Centre approach stresses the importance of transparency and rigour. It also has a number of distinct, novel features. A first is an explicit acknowledgement that choice of research topic and specific focus (in the cases of both primary studies and syntheses of existing evidence) reflect particular sets of values and interests. 26 Research questions are not self-evident, but are constructed by particular people for particular purposes. In the case of the child and adolescent mental health field, the range of people with active interests in this complex system is wide indeed, potentially yielding a large variety of candidate priority research topics. Managers, practitioners from different backgrounds, commissioners, policy-makers, young people, families and others may have varying ideas on what kinds of research questions ought to be asked and what methods should be used to answer them. Often the authority to shape priorities remains limited to a relatively powerful few. These ideas underpin the observation that:

Involving representatives of all those who might have a vested interest in a particular systematic review helps to ensure that it is a relevant and useful piece of research.

(p. 1)26

In any project informed by the EPPI-Centre approach a crucial task must therefore be the effective engagement of concerned people. Involvement in an EPPI-Centre synthesis is typically secured through the setting up of an advisory group or something similar, in which representatives of groups with clear interests and experiences relevant to a study’s broad focus and eventual findings participate. In reviews conducted in the health-care field, members of advisory groups are likely to be drawn from communities of practitioners and managers as well as from those with direct experience of using services. A representative stakeholder advisory group, in turn, assumes particular responsibilities at key moments in the lifetime of a project. One critical moment for a group’s involvement is the point at which major decisions need to be made on the direction of travel an evidence synthesis is about to take.

This idea that in-progress decisions have to be made on the future direction an evidence synthesis will take is a second hallmark of the EPPI-Centre framework. This is particularly evident through the EPPI-Centre’s commitment to the combining of an initial scoping and mapping phase with a second, more targeted, review in one or more negotiated priority areas. In EPPI-Centre reports a detailed account is often included of how an initial mapping of the territory (in our case, the territory relating to risk in inpatient CAMHS) has been presented to stakeholder participants drawn from across a system, and how this has been used as the basis for the considered selection of candidate subareas for subsequent, detailed, quality review and synthesis.

A third component of the EPPI-Centre framework relates to the sharing of information on a project’s methods and findings in ways that are sensitive and accessible to the needs of larger stakeholder groups. Representatives of these groups are offered opportunities to advise project members on the most suitable ways of disseminating new knowledge, through the use of varieties of media including briefing summaries, online materials and articles of varying lengths and complexity tailored for particular audiences. Again, in this report we show how we worked with stakeholders to develop strategies to share what we learned. We also include plans for further work in this area once this final report has been accepted for publication as an NIHR journals library monograph.

Structure of this report

This opening chapter has set the scene for this evidence synthesis in the area of risk for young people moving into, through and out of tier 4 inpatient child and adolescent mental health services. It has introduced the CAMHS system as a tiered one, in which hospitals fulfil important functions within a dynamic and inter-related network of services. The existence of risks, both obvious and less so, has been introduced and a case made for a project of this type. An overarching research question has been given, along with a series of objectives. The project team’s commitment to the EPPI-Centre approach to evidence synthesis has been stated, and an overview given of some of the key distinguishing characteristics of this framework. In the chapter immediately following, methods and findings are given of the project’s phase 1 scoping and mapping exercise. Chapter 3 describes the approach taken to working with stakeholders, and the process through which priority categories of risk were identified and carried forward into the second, in-depth, phase of the project. Chapter 4 addresses the concepts and methods used in the phase 2 review, beginning with a summary of the research questions and objectives guiding this segment of the project. Chapter 5 synthesises the evidence in the phase 2 priority risk areas. Chapter 6 discusses the study overall and its findings. It includes a single-page matrix summarising the project overall and outlines plans for dissemination and knowledge exchange informed by discussions with stakeholder collaborators. It also draws out the implications of this project for policy, services and practice and makes recommendations for future research.

Using the EPPI-Centre framework to map risk in inpatient child and adolescent mental health settings

The team’s use of the EPPI-Centre framework commenced with a scoping exercise (see Figure 2) to identify the broad contours of the research field relating to risk in inpatient child and adolescent mental health settings. The aim of this first phase was to identify and categorise research and other evidence at the intersection of the four key areas of young people, mental health, inpatients and risk (see Figure 3). During this phase no attempts were made to assess the quality of materials retrieved, the aim being instead to scope and map out the existing evidence at the meeting point of the areas lying at the heart of this project. In this initial searching, scoping and mapping, no attempts were made to determine which types of risk should be either included or excluded, and no attempts were made to identify underpinning theory or to identify costs or cost-effectiveness.

FIGURE 2.

Phase 1 of the EPPI-Centre framework in action.

FIGURE 3.

Phase 1 search.

As per the study’s commissioned protocol, searching for papers was conducted using the electronic databases MEDLINE and PsycINFO. For these two databases, controlled vocabulary and free-text terms which covered the four key areas above were combined. Details of the search strategies used are given in Appendix 1 (MEDLINE) and Appendix 2 (PsycINFO). This search was guided by a set of inclusion and exclusion criteria described immediately below and (in summary) by the four-arm strategy depicted in Figure 3. The end date for searching was March 2013.

Inclusion and exclusion criteria

To be included in this scoping exercise, a citation (which could be a research report, a review paper, a discussion piece, a published opinion, an editorial or something similar) had to:

1. Be written in the English language.

2. Focus on young people aged 11 to 18 years by:

   1. focusing exclusively on young people aged 11 to 18 years or

   2. focusing on a wider age group but including sufficient detail to enable the accurate identification of data relating to young people specifically or

   3. relating to a study sample where the mean age was between 11 and 18 years.

3. Focus on moving into (admission), through and/or out (discharge) of inpatient mental health services. As not all citations retrieved included unambiguous descriptions of service types, the decision was taken to consider ‘inpatient mental health services’ as any inpatient hospital services (and, in the case of US citations, residential treatment centres) staffed by mental health professionals.

4. Address risk identification and/or risk assessment and/or risk management. Reflecting the purpose of the review, risk was viewed in broad terms, with no a priori exclusions placed on citations because of the way risk was thought about and/or used by authors. Citations were therefore included which addressed the risks of (for example) harm to self or harm to others, but also the risks mental ill-health and hospital admission pose to young people’s physical, psychological, social and educational development.

All citations retrieved were downloaded into Endnote Web™ (2013, Thomson Reuters, New York, NY, USA) and duplicates removed. For the purpose of this scoping exercise, all citations to dissertations, theses or books were excluded. Each remaining citation retrieved was independently assessed for relevance to the review by two members of the study team, using the information provided in the title and abstract. Where any doubt existed, the full text was retrieved. In all cases, the full text was retrieved for all citations that at this stage appeared to meet the scoping review inclusion criteria.

Describing relevant papers

To achieve a high level of consistency, reviewers screened each retrieved citation for inclusion using a purposely designed form (see Appendix 3). Disagreement was resolved by discussion with a third reviewer. This phase 1 form included space for the extraction of key data from each citation: the country in which the study (in the case of research reports) had been carried out; if the citation focused on moving into (admission), through or out of (discharge) inpatient mental health services, or any combination thereof; the characteristics of the population studied (if relevant); the type of risk focused on; and, in the case of original research, the design used (e.g. cohort study, longitudinal study).

Evidence map

Figure 4 illustrates the flow of citations through the scoping exercise and shows how 124 individual items were eventually identified for inclusion. Details of the included papers are given in Appendix 4. Of these, the majority were primary research papers (115 articles describing 118 studies). The remaining papers (n = 9) described the development of guidelines (n = 1) or were a recommendation statement (n = 1), general discussions (n = 4), reviews (n = 2) and a letter (n = 1). In the following section we describe the characteristics of the 118 research studies according to the country in which projects were undertaken, the age of study participants, research design, the size of study populations and the year of publication.

FIGURE 4.

Flow of citations through phases of the scoping review.

Country where research was undertaken

Information on the country in which each research study had been carried out was obtained from the methods section of the retrieved full-text version of each paper, or where this information was not directly given from the country location of the corresponding author. The distribution of research by country in Figure 5 shows that by far the largest number of studies (over 60%) were conducted within the USA, with only 2.5% conducted within the UK. This preponderance of studies from the USA may reflect the limiting of the phase 1 scoping search to English-language citations only, this having a direct effect on the country of origin of papers retrieved.

FIGURE 5.

Country of origin of the 118 studies included in the scoping review.

Research design

Figure 6 summarises the research design used in the 118 research studies included in the phase 1 scoping exercise. Seventy-two of the studies used a cross-sectional design, of which 24 involved the collection of retrospective data from young people’s inpatient medical notes or charts.

FIGURE 6.

Research design used in the 118 research studies included in the scoping review.

Age of study participants

The mean age of the young people whose data were included in each of the 118 studies in the phase 1 scoping fell between 11 and 18 years. The age ranges of young people participating, or whose data were used, in individual studies are summarised in Table 1.

Size of study populations

The size of the populations included in each of the 118 phase 1 studies varied, with the largest number being 27,000 included in the single study by Chabra et al. 27 Figure 7 summarises this information.

FIGURE 7.

Size of the populations participating in phase 1 included research studies.

Year of publication

Data on the year of publication of the 115 papers reporting findings from the 118 studies included in phase 1 are given in Figure 8.

FIGURE 8.

Year of publication of phase 1 included papers.

Mapping exercise

As a way of presenting the evidence located and included in this phase 1 scoping, a number of broad, descriptive maps of the different ways in which risk has been identified, assessed and managed in the inpatient CAMHS area were produced. These are presented in this report as Figures 9–17. To group included papers thematically, an initial typology of risks was used, drawing on the work of Subotsky. 29 Subotsky uses four categories of risk: harm to self; harm to others; harm from others; and harm from the health-care system and its staff. In order to capture the full range of risks addressed across all 124 included papers, further categories were added by the project team as necessary, each reflecting the content of papers retrieved. Examples included the category of longer-term risks found at follow-up of young people previously admitted to inpatient CAMHS units, and risks of early disengagement from services. This full range of risks identified is reproduced in the figures immediately below, where the content of each map thematically summarises the main areas addressed by the papers included in each category.

FIGURE 9.

Descriptive map of ‘harm to self’.

Risk of harm to self (65 papers)28,30–93

Given what is known about the general usage of the term ‘risk’ in mental health services and the types of risk that attract most attention in policy and practice, it was not surprising to the project team to find that the category with the largest number of included phase 1 papers addressed the risks of harm to self (see Figure 9 for the associated descriptive map). Of the 65 papers included here, overwhelmingly most (n = 53)28,30–81 were concerned with the identification and/or the assessment and/or the management of suicide risk, followed by the risks to self which are associated with alcohol and drug abuse (n = 7). 82–88 Smaller numbers of included papers focused on general self-harm (n = 3)89–91 and the risks associated with human immunodeficiency virus (HIV) infection and acquired immunodeficiency syndrome (AIDS) (n = 2). 92,93 Noting the high volume of papers addressing risks associated with suicide, a decision was made to produce a separate descriptive map for this area alone (see Figure 10).

FIGURE 10.

Descriptive map of ‘suicide’.

Harm to others (20 papers)94–113

This category included a large number of papers addressing the risk of harm to others, with papers focusing on both aggression and violence (n = 18)94–111 and bullying (n = 2)112,113 (see Figure 11).

FIGURE 11.

Descriptive map of ‘harm to others’.

Longer-term risks found at follow-up (15 papers)114–128

This category was independently created by the project team to group together a series of papers focusing on the longer-term risks of being admitted to inpatient CAMHS units (see Figure 12). Included papers variously addressed the risks of readmission (n = 9),114–122 early death (n = 2),125,126 disrupted schooling (n = 2),127,128 homelessness (n = 1)123 and delinquency (n = 1). 124

FIGURE 12.

Descriptive map of ‘longer-term risks found at follow-up’.

Early disengagement from services (five papers)129–133

This category is depicted in Figure 13 and was developed by the project team to bring together papers concentrating on the risk of young people running away (n = 3),129–131 being discharged against medical advice (n = 1)132 and dropping out from treatment (n = 1). 133

FIGURE 13.

Descriptive map of ‘early disengagement from services’.

Risk factors influencing admission and length of stay (five papers)27,134–137

This category is depicted in Figure 14 and was created to bring together papers addressing factors influencing admission to, and length of stay in, inpatient CAMHS care. Factors identified included gender, ethnicity, general predisposing factors, and young people being adopted and living in disrupted family homes.

FIGURE 14.

Descriptive map of ‘risk factors influencing admission and length of stay’.

Risk of harm from the system (five papers)138–142

Directly drawn from Subotsky’s29 typology, this category includes the negative effects of treatment along with the risks of other adverse consequences of inpatient admission, such as loss of educational continuity, or exposure to risks, such as abuse by staff. Included in the mapping (see Figure 15) were papers dealing with the side effects of medication (n = 3),138–140 sexual abuse by staff (n = 1)141 and contagion (n = 1). 142

FIGURE 15.

Descriptive map of ‘risk of harm from the system’.

Responding to and managing risk (three papers)143–145

A small number of papers were brought together in this category, created by the team for the purposes of this mapping because they specifically addressed actions designed to manage or reduce risk (see Figure 16). These papers variously addressed improving educational attainment (n = 1),143 managing the environment to keep people safe (n = 1)145 and milieu-based behavioural management to reduce aggression (n = 1). 144

FIGURE 16.

Descriptive map of ‘responding to and managing risk’.

Predictors of restraint or seclusion (three papers)146–148

This set of three papers looked at predictors of restraint, seclusion or both (see Figure 17). Specifically, they sought to identify the factors that place a young person at increased risk of seclusion during his or her admission (n = 1),146 to establish whether or not a particular group of young people are at risk of restraint (n = 1)147 and to examine whether or not temperament characteristics (e.g. fear, anger) are risk factors for restraint and seclusion for young people in inpatient mental health hospital (n = 1). 148

FIGURE 17.

Descriptive map of ‘predictors of restraint or seclusion’.

Summary of phase 1 findings

Each of the 124 papers included in phase 1 of this project was identified following a search of two databases combining terms from four arms (mental health, young people, inpatients, risk). Each was reviewed by at least two members of the study team, who agreed that it met the project’s criteria for inclusion. Data (e.g. focus of paper, country of origin, type of study in the case of papers reporting original research) were extracted from each and summarised.

The central purpose of this first phase of our overall evidence synthesis was to scope and map out the territory in the area of risk for young people moving into, through and out of inpatient mental health services. This was achieved by grouping together included papers using categories developed by Subotsky,29 and by using new categories created specifically for this study by the project team. This theming of retrieved papers demonstrated that, overwhelmingly, it is clinical risks (particularly the risks of suicide, self-harm and harm to others) that have featured most prominently in published papers (or, at least, in the papers published in journals indexed in the two databases searched in this phase of the project).

This was not an unexpected finding. It suggests that the risks that dominate policy and practice in mental health services are the risks that occupy most researchers and writers in the field. With regard to the characteristics of the research in this area, few papers retrieved reported data from inpatient CAMHS risk studies conducted in the UK. Most papers reported findings from prospective or retrospective cross-sectional surveys designed to establish associations between the personal and clinical characteristics of young people using inpatient CAMHS and the chances of certain unwanted events (e.g. self-harm, violence, running away) happening. Very few papers were directed at establishing what might be done to respond to or manage risk.

Using the Evidence for Policy and Practice Information and Co-ordinating Centre framework to consult with stakeholders

The EPPI-Centre approach obliges researchers to pause at the point where the territory has been scoped and mapped out, and to consult with people with interests in the field. The purpose of this consultation is to share knowledge of what has been found and to invite views to be given on the priority areas to be taken forward into the in-depth phase of the review which follows. In this project, key groups to involve in the consultation process included young people with personal experience of using inpatient mental health services; family members; service managers; practitioners drawn from different occupational groups; and workers from non-statutory organisations campaigning for young people’s mental health and the development of services. Figure 18 illustrates the use of the EPPI-Centre framework in this phase of the project.

FIGURE 18.

Using the EPPI-Centre framework to consult and determine phase 2 priorities.

Stakeholder advisory group meeting: agreeing risk priorities for phase 2

In preparation for a first combined project team and stakeholder advisory group (SAG) meeting, a set of documents was prepared for invited members and posted out in advance. Included were an accessible overview of the project (see Appendix 6), a proposed set of terms of reference for SAG members (see Appendix 7) and an agenda (see Appendix 8). The meeting itself was convened in Cardiff on 24 April 2013, organised for the express purpose of generating candidate categories of risk to serve as priorities for the second, in-depth, phase of this review. The event was chaired by Dr Michael Coffey (Associate Professor, Swansea University), an academic with expertise in mental health services but not associated with the study in any other way.

Decisions on the full range of stakeholders to include in the SAG, and the identification of actual representative individuals to invite, were taken collectively by the project team, drawing on our understanding of the CAMHS system, our knowledge of specific inpatient units and our working relationships with managers, practitioners and others. In addition to members of the project team, participants (drawn from South Wales and Greater Manchester, reflecting the locations of project members) included Matthew Daniel, collaborating representative from YoungMinds (who had already completed the consultation exercise reproduced in Appendix 5); two young people with experiences of using CAMHS; a senior NHS CAMHS manager; a senior child and adolescent psychiatrist; a senior CAMHS therapist; and a senior nurse with inpatient CAMHS responsibilities. Invitations were also extended to individuals with backgrounds in child and adolescent mental health social work, clinical psychology and teaching in inpatient CAMHS settings.

The principles of nominal group technique were drawn on in planning the process designed to lead to the identification of priority risk categories for phase 2 of the project. This is an approach to group decision-making introduced in the early 1970s by Delbecq and Van de Ven,155 which places weight on all participants having an equal opportunity to express a view. Since its emergence it has been successfully used in the research context, including in a study funded by the NIHR Health Technology Assessment (HTA) programme involving a member of this current project team (ML) as a co-investigator. 156 In this current project, the use of an approach explicitly designed to avoid letting group decision-making become dominated by one or more vocal individuals appealed to the team, given our concern that young people, in particular, might find it difficult to contribute in the presence of professionals, managers and academics.

The SAG meeting opened with an overview of the project, along with a presentation of the broad themes arising from the phase 1 scoping and mapping exercise. To facilitate later discussion and priority setting, our maps remained on display throughout the duration of the event, each depicted on an individual sheet of A1 flipchart paper. A presentation from YoungMinds followed, centring on a summary of the outcomes of the consultation exercise previously conducted with young people.

Following a natural break in the meeting, participants were invited to record independently written responses to the question ‘What do you think the risks for young people are as they make the transition into, through and out of inpatient CAMHS?’

The idea of inviting written responses, without conferring, was used as a nominal group strategy to ensure that as a project team we had access to the full range of views available in the room in enduring form once the meeting had ended. With a member of the team (NE) acting as facilitator, all SAG and project team members were then invited to share their written responses individually. These were recorded on flipcharts using the words directly spoken by the participants, and, during this round, the views of the carer who had previously been consulted were also reported and recorded. In the facilitated discussion following, opportunities were taken to seek explanations or further details about any of the ideas that participants had produced that were not clear to all. In this segment of the meeting our aim was to generate as comprehensive a list as possible of the full range of risks associated with inpatient CAMHS, including (but not limited to) the risks which our phase 1 scoping and mapping had identified as occupying the interests of researchers.

By the end of this process, meeting participants had access to a range of materials and ideas presented and discussed in the earlier part of the day. These included the pre-prepared series of maps displaying the themed overview of existing research; the account of the YoungMinds pre-meeting consultation with young people; and a series of flipcharts carrying the words of individual stakeholders (and a carer, previously consulted) in response to the invitation for people to identify the risks facing young people as they make the transition into, through and out of inpatient CAMHS. In a second round of activity informed by nominal group principles, project team and stakeholder group members were then asked to rank, in writing, their personal priorities for the categories of risk to take forward into the second, in-depth, phase of the project. The preamble to this exercise invited people to draw as widely as they wished on the full range of risks as previously described and discussed.

This first combined project team/stakeholder meeting closed with the collection of lists of individually generated, ranked phase 2 priority categories of risk produced by each person taking part. To this collection of lists were added carer priorities previously identified and a composite list of priorities extracted from the YoungMinds consultation report. This exercise gave project team members a long list of candidate categories of risk for the in-depth segment of the project ordered by individual researcher and SAG member, and from the young people collectively consulted (see Appendix 9). From this collated list the top three ranked priorities were taken from each individual (and from the YoungMinds consultation collectively) to produce a new, shorter, list of 45 individual items. In order to group these items under broader categories, this list was independently coded by two members of the project team (DE and NE), who then met to discuss and agree on their categorisation. This joint categorisation was then checked and agreed on with the involvement of a third member of the team (BH).

This process led to the identification of eight ranked priority categories of risk, each having a count of the number of items from the total of 45 subsumed within it (see Appendix 10). This information was included in a summary document for SAG members, which was circulated for final responses before the initiation of phase 2 of the review, and is also summarised in Table 2. Further details are given in Appendix 11.

The document for SAG members noted that the top risk category priorities for the in-depth segment of the review were all examples of ‘less obvious’ risks, and, as such, were unlike many of the more ‘clinical’ risks identified in the phase 1 mapping of the literature. The project team’s categorisation of these less obvious risks included a version of this summary information, along with plans for the initiation of the next phase of the synthesis overall:

• Dislocation is the word the project team has used to describe the top priority emerging for the second phase of this review. This category of risk includes the ideas of young people being removed from ‘normal’ life, of being ‘different’ and of experiencing ‘fragmentation’.

• Dislocation as a category of less obvious risk also captures the ideas of being stigmatised and discriminated against, and of young people losing their previous identities, social contacts and friendship groups. It includes isolation from, and within, families.

• Dislocation also includes the risks presented to young people’s educational, psychological and social development.

• Dislocation implies unhelpful loss, and contagion (the second category of risk prioritised for the next phase of the review) implies unhelpful gaining: the risk of being exposed to and of learning abnormal behaviour, and of new and unhealthy friendships.

The project team’s plan presented to SAG collaborators indicated the aim of taking both dislocation and contagion forward as the two, linked, priority risk categories for the second phase of this evidence synthesis. Team members were aware of an equal number of items having been brought together under the categories of ‘contagion’ and ‘harm from organisation’, but noted that, in consulting with a carer, the risks from contagion (learning new and harmful behaviours) were a particular priority. This informed the plan, also included in the report to collaborators, to take this area forward into the next study phase. The report also drew attention to the fact that the proposed phase 2 plans implied a move away from the more ‘clinical’ risks identified in the phase 1 mapping towards a consideration of some of the ‘less obvious’ risks, and in this context invited feedback on three questions:

• Have we reflected stakeholders’ priorities accurately?

• If we have reflected the priorities for the in-depth part of our review accurately, then we welcome ideas on where we need to go for evidence in these areas. If we are gathering information on how the risks of ‘dislocation’ and ‘contagion’ are identified, assessed and managed as young people move into, through and out of inpatient CAMHS, then where should we look and whom should we approach?

• What other words can you think of that reflect the ideas of ‘dislocation’ and ‘contagion’, which we might use to continue to search for evidence?

One stakeholder participant from our April 2013 meeting responded to the summary and phase 2 initiation document by advising the team to be mindful of the dangers of ‘demonising tier 4 admissions’, noting that hospitals are a crucial and very scarce resource playing an important part in overall systems of care. Our response as a team, then and now, is that we concur: inpatient admission for young people with mental health difficulties is indeed sometimes necessary, as it is here that round-the-clock specialist care is provided to people with the greatest need. What the consultation exercise indicated, however, was an appetite among project members and collaborating stakeholders for a phase 2 review of some of the risks less often considered for young people passing into, through and out of inpatient care.

Summary of consultation with stakeholders

Following EPPI-Centre principles, a series of broad, phase 1 maps of the literature on risks for young people admitted to mental health hospitals were taken to a stakeholder advisory group with a view to determining the focus for the second, in-depth, phase of this study. Project team members engaged with the charity YoungMinds, which conducted a consultation on the team’s behalf. A series of maps arising from the phase 1 scoping exercise were devised and presented to people taking part in a key stakeholder advisory group meeting convened in April 2013. Collaborators were then asked to identify what they thought the risks to young people moving into, through and out of mental health hospital were. Responses were displayed alongside the risks occupying earlier teams of researchers identified from the scoping exercise. Informed by nominal group techniques, all team members and collaborators were invited individually to rank the categories of risk they saw as priorities for the in-depth phase of the project. These priorities were grouped together under a series of broader categories, with a large number of risks brought together under the umbrella category ‘dislocation’. The team determined to make this a first priority risk category for phase 2 alongside a second category, ‘contagion’.

Introduction and conceptual framework

During the phase 1 scoping of the literature using two databases, a preponderance of research and other outputs was found that concentrated on the clinical risks to young people admitted to mental health hospitals, and particularly in relation to the risks of suicide and self-harm. Typically, it is in response to these risks that young people are admitted to tier 4 inpatient settings, as it is in hospitals that round-the-clock specialist care and treatment in conditions of safety can be provided. The risks of suicide, self-harm, physical deterioration and harm to others are vitally important ones: for young people, their families, friends, practitioners, managers and the wider society. Nonetheless, the steer from young people contributing to the YoungMinds consultation exercise conducted on the project team’s behalf, and from stakeholders participating in the end-of-scoping priority-setting meeting, was to conduct a search and synthesis of the evidence across a series of other, less obvious, risks.

This part of the project was again informed by the idea that inpatient CAMHS units exist within a larger, complex and inter-related system of people and organisations. Young people themselves are part of this system, as are their families and friends, the schools they attend and the community-based health services they may use. Decisions to admit young people to mental health inpatient units are often made in the context of high levels of clinical risk, and have important consequences: for the individuals admitted, for families and friends and others immediately surrounding, and for those providing direct services. Hospital admission also has consequences at the system level, at which services are organised and delivered: where teams and workers in inpatient and community health and social care services, in schools and elsewhere have responsibilities to manage young people’s transitions into, through and out of hospital in as seamless and integrated a manner as possible. In this context a consideration of the broader, less obvious, risks of ‘dislocation’ and ‘contagion’ for young people and those around them is consistent with a systems perspective and it is these largely non-clinical risks that form the focus of this part of the project.

Having identified dislocation and contagion as the categories of risk to prioritise, the task of searching comprehensively for research and other evidence in these areas in the context of young people passing into, through and out of inpatient mental health hospital proved challenging. This segment of the report sets out the research questions guiding phase 2. It also addresses the approach used in the interrogation of databases, and the sequential strategy developed for sifting through citations and making decisions on including and excluding materials identified. Reflecting EPPI-Centre commitments to seeking out evidence of the widest variety (rather than research outputs alone), a description is also given of the steps taken to locate and secure grey literature, clinical case reports and reports of local practice initiatives on how inpatient CAMHS units identify, assess and manage the categories of ‘less obvious’ risk focused on in phase 2. A detailed account is then given of the materials included in this in-depth component of the project and how decisions were made on the quality of these.

Phase 2 research question and objectives

The research question for phase 2 was:

What is known about the identification, assessment and management of dislocation and contagion in young people (aged 11–18 years) with complex mental health needs entering, using and exiting tier 4 inpatient services in the UK?

Objectives for the in-depth evidence synthesis were:

1. to identify and appraise the evidence for the identification, assessment and management of the risks of dislocation and contagion for young people using inpatient CAMHS

2. to identify and describe any underlying theoretical explanations for approaches used in the identification, assessment and management of the risks of dislocation and contagion

3. to understand the views and experiences of dislocation and contagion of young people (aged 11–18 years) with complex mental health needs using inpatient mental health services, and of those involved in the identification, assessment and management of these risks in these settings

4. to synthesise the evidence for the identification, assessment and management of dislocation and contagion in young people (aged 11–18 years) with complex mental health needs moving through inpatient services

5. where possible, to synthesise the evidence on the costs and cost-effectiveness to the NHS of different approaches to identifying, assessing and managing the risks of dislocation and contagion for young people using inpatient mental health services.

Review design

As with the first phase of the project, the design of this second segment was informed by the approach developed by the EPPI-Centre, as Figure 19 illustrates.

FIGURE 19.

Phase 2 of the EPPI-Centre framework in action.

Review methods

As per the approach described in the project protocol, once we had agreed the focus for this second phase of the project, work commenced with electronic searches of specified databases. In the case of citations finally included, reference lists were searched to identify additional materials for possible inclusion (back-chaining). Searches were conducted for UK-only grey literature, and the project team published (and widely circulated) a call for evidence. This requested examples of local service responses to the less obvious risks and was sent to all NHS and non-NHS inpatient CAMHS units in the UK and was additionally distributed via online discussion and mailing lists.

Types of participants

To be included, research and other materials needed to be written in the English language and to focus on young people aged 11–18 years by:

1. focusing exclusively on young people aged 11–18 years or

2. focusing on a wider age group but including sufficient detail to enable the accurate identification of data relating to young people specifically or

3. relating to a study sample where the mean age was between 11 years and 18 years.

Types of intervention and phenomena of interest

All citations were considered which addressed ‘dislocation’ and ‘contagion’ for young people moving into, through and out of inpatient mental health services.

Context

All citations were considered where care was provided in inpatient mental health services. Following the decision made in phase 1, ‘inpatient mental health services’ was defined in this phase of the study as any hospital setting staffed by mental health professionals. In the case of US studies retrieved, the decision was again taken to include citations centring on residential treatment centres (RTCs) or residential treatment programmes. These facilities are comparable to hospitals, with the American Academy of Child and Adolescent Psychiatry defining them as providing:

intensive help for youth with serious emotional and behaviour problems. While receiving residential treatment, children temporarily live outside of their homes and in a facility where they can be supervised and monitored by trained staff.

(p. 1)157

Types of evidence

This part of the review considered a variety of types of evidence: published and grey literature original research; review, discussion and opinion papers; reports of local practice initiatives or service developments; clinical case studies; and policy and guidance documents. All evidence relating to outcomes, views and experiences, costs and cost-effectiveness, policies, and service and practice responses in the areas of ‘dislocation’ and ‘contagion’ for young people (11–18 years) with complex mental health needs using inpatient mental health services was sought. Materials published in the English language since the introduction of the tiered system in CAMHS in 1995 were sought.

Exclusion criteria

Excluded materials were those not in the English language; addressing inpatient mental health services for adults over 18 years; centring on inpatient mental health services for children under the age of 11 years; relating to any community-based mental health services (e.g. outpatient, day care or wraparound); addressing juvenile justice services.

Search strategy

Preliminary scoping searches of MEDLINE and PsycINFO performed for phase 1 of the project (see Chapter 2 and Appendices 1 and 2) highlighted a number of potential challenges to be considered when developing the in-depth searches required for phase 2, principally surrounding the use of the term ‘risk’. The outputs from phase 1 had predominantly focused on the clinical risks to young people in inpatient mental health settings, such as the risk of suicide and self-harm. Comparatively little evidence was found focusing on the less obvious risks such as the risk to young people admitted to mental health hospital of negative peer group influences or the loss of educational opportunities. The priority risk categories, to be taken forward to the second, in-depth, phase of the study and decided in consultation with the stakeholder advisory group, were all examples of these less obvious risks. An examination of the papers known to the investigators from phase 1 discussing these less obvious risks confirmed the challenge ahead, with very few using the term ‘risk’ and the language used within each varying considerably.

Led by one member of the project team (EG), an initial attempt was made to try to define and describe the two priority areas of ‘dislocation’ and ‘contagion’, with the intention of directly including these in one arm of the phase 2 search strategy. Potential keywords reflecting these areas were generated, drawing on contributions from the project team and from members of the stakeholder advisory group. Examples of candidate keywords connected to the priority risk areas are given in Table 3.

However, testing these terms in pilot searches produced an unmanageable volume of citations. Following consultation with colleagues, including Mala Mann [an information specialist associated with the Systematic Review Network (SysNet) at Cardiff University], the decision was made to remove this arm of the search entirely. This decision produced a highly sensitive final search strategy for phase 2 comprising three arms: (1) young people, (2) mental health and (3) inpatient. This was recognised as the best approach to capturing all relevant material, leaving project investigators the task of reviewing all search results and manually identifying citations that also addressed the priority risks of dislocation and contagion, from which any papers addressing costs and cost-effectiveness could also be located. Searches were developed using a combination of controlled vocabulary [e.g. medical subject headings (MeSHs) in MEDLINE] and text words, and, before proceeding, this strategy was thoroughly tested to ensure that it retrieved all of the core papers identified as relevant during the phase 1 scoping. Information on search strategies used across all databases is provided in Appendix 12. The 17 individual databases searched were:

1. American Economic Association Database (EconLit)

2. Applied Social Sciences Index and Abstracts (ASSIA)

3. British Nursing Index (BNI)

4. Cochrane Library (including the Cochrane Database of Systematic Reviews, Database of Abstracts of Reviews of Effects, Cochrane Methodology Register, HTA and NHS Economic Evaluation Database)

5. Cumulative Index to Nursing & Allied Health Literature

6. Education Resources Information Center (ERIC)

7. Excerpta Medica (EMBASE)

8. Health Management Information Consortium (HMIC)

9. Medical Literature Analysis and Retrieval System Online (MEDLINE)

10. PsycINFO

11. Scopus

12. Social Care Online

13. Social Services Abstracts

14. Sociological Abstracts

15. System for Information on Grey Literature in Europe (OpenGrey)

16. Turning Research into Practice (TRIP) Plus

17. Web of Science.

To ensure that the project remained manageable in the time frame given, a number of limits were also applied to contain the number of results. Text word searching was used for titles and abstracts only, and, as per the inclusion criteria given in Types of evidence above, the search was limited to English-language publications from 1995 to September 2013. When numbers of results were particularly large, the option of restricting the search to ‘human’ was also utilised.

As described in our Review methods section above and in Call for evidence below, a range of snowballing techniques were also used to increase the sensitivity of the search. References from citations located and included in phase 2 were searched for additional studies (back-chaining), and the Journal of Child and Adolescent Psychiatric Nursing, European Child and Adolescent Psychiatry and Psychiatric Services were hand-searched for the previous 2 years to ensure that any relevant papers that may not have been indexed in the major databases were located. Searches for relevant grey literature were conducted using directories of conference proceedings, HMIC, OpenGrey and Index to Theses. A search was also made using the website of the NIHR, including for information on studies funded by the HSDR programme and its predecessors.

Searching for policy and guidance

As per the study protocol, searches were made of organisational websites to locate and include important contextual material (such as policy drivers, standards and cross-UK priorities for inpatient CAMHS), along with new material to include as direct evidence. Lists of current and archived websites to search were generated by project team members drawing on their understanding of the field. Those searched were:

• CAMHS Outcomes Research Consortium

• Care Services Improvement Partnership

• Centre for Mental Health

• Child and Maternal Health Observatory

• Children and Young People’s Mental Health Coalition

• Department of Health, England

• Joint Commissioning Panel for Mental Health

• Mental Health Foundation

• Mental Welfare Commission

• National CAMHS Support Service

• National Institute for Health and Care Excellence

• National Mental Health Development Unit

• Northern Ireland Executive

• Quality Network for Inpatient CAMHS

• Royal College of Psychiatrists

• Scottish Executive

• Social Care Institute for Excellence

• Solutions for Public Health

• Welsh Government

• YoungMinds.

Call for evidence

A single-side call for evidence was produced (see Appendix 13). This introduced the study and invited recipients to forward evidence relating to the identification, assessment and management of the categories of less obvious risks focused on in this phase of the review. This call was sent in the second half of 2013 to all NHS and non-NHS inpatient CAMHS units identified within the UK. In January 2014 the same call for evidence was electronically circulated via the Child and Adolescent Mental Health Service National Academic Mailing List Service discussion list (see www.jiscmail.ac.uk/cgi-bin/webadmin?A0=CAMHS). Contact was made with the Royal College of Psychiatrists’ Quality Network for Inpatient CAMHS (QNIC), and through the group’s Deputy Programme Manager the call for evidence was forwarded to all members. The team was unable to also circulate this call to members of the Royal College of Nursing’s Children and Young People forums (as had initially been intended and written into the study protocol) as no online forum was found to exist. Instead, snowball sampling was used to continue the circulation of the study among members of the nursing profession. Contact was made with the Royal College of Nursing’s Mental Health Advisor (who has UK-wide responsibilities), and then with a former National Nursing Officer in England with mental health responsibilities, for onward distribution via their lists. The call was also circulated to all members of Mental Health Nurse Academics UK (MHNAUK, an organisation with representatives of over 60 universities; see http://mhnauk.swan.ac.uk/main.htm), and to members of the All Wales Senior Nurse Advisory Group for Mental Health.

Screening citations: a staged approach

The reference management software package EndNote (Version X7.1, Thomson Reuters, New York, NY, USA) was used to manage all citations found through the team’s database searching. First, duplicates were removed. Because of the large volume of retrieved citations remaining, a staged approach to screening was then conducted. In an initial round, the title of each citation was read by two members of the review team to determine whether or not the item fell at the intersection of ‘young people’, ‘mental health’ and ‘inpatient services’. For citations included at this stage, abstracts where available were read by two members of the team to determine whether or not items fell at the intersection of ‘young people’, ‘mental health’, ‘inpatient services’ and, now, the additional category of ‘risk’, broadly defined. Abstracts were then screened a final time to determine whether or not citations were concerned with the ‘less obvious risks’ forming the focus of this phase of the overall review. Where any doubt existed, full texts of articles were retrieved.

The full texts of all citations included at this stage were screened independently by two reviewers using a screening tool to aid this process (see Appendix 14). Multiple articles by the same authors reporting findings from the same study were linked together to help inform decisions on which studies to include. At all stages of this screening process, any citations that reviewers were unsure of were taken through to the next stage to allow a further round of consideration. Disagreement on the inclusion of citations at any of the stages based on topic relevance was resolved by discussion with a third member of the review team. Studies identified from the back-chaining of reference lists were assessed for relevance based on publication titles, and then on the retrieval of full texts. Following the retrieval of back-chained full text publications, citations were managed in the same way as citations retrieved via database searching.

Policies and guidance, reports of practice initiatives, clinical case studies and other non-research materials found in the grey literature were also read by two members of the project team and considered against the topic inclusion criteria, with disagreements resolved as above. Initial judgements, using the criteria developed by the National Institute for Health and Clinical Excellence (now the National Institute for Health and Care Excellence) in 2006,158 were:

• category A (accepted): material judged to be of sufficient quality to be included in the next stage of the selection process

• category B (background): material judged to provide background or contextual information

• category R (rejected): material judged to be of no use for the purposes of this review.

Figure 20 shows the flow of citations through each stage of this overall process. Despite interest being shown in the project from individual respondents (some of whom forwarded published papers to the team), no local policies, practice initiatives or service developments were secured in response to our call for evidence (see Appendix 13). Searching of websites for policy and guidance found a total number of 20 documents assessed as relevant to the review.

FIGURE 20.

Flow of citations through the phase 2 in-depth review.

Data extraction

Each of the 40 citations and 20 policy and guidance documents was allocated to a member of the research team. Data from each were extracted and entered into a series of electronic tables. Table format followed guidance issued by the Centre for Reviews and Dissemination,159 with choice of table reflecting the type of material considered (see Appendices 16–21). Data extraction was independently checked for accuracy and completeness by a second researcher (DE), with any disagreements noted and resolved by consensus. A record of corrections was kept. Where multiple publications from the same study were identified, data were extracted and reported as a single study.

Quality assessment

As per the study protocol, each included study that passed the initial topic inclusion screening process was read and assessed independently for quality by two team members using one of a number of agreed appraisal checklists. For studies using an experimental design, the Effective Public Health Practice Project (EPHPP) tool was used. 160 This is a generic instrument which has been used to evaluate a variety of studies using an intervention design and has been deemed acceptable for use within systematic reviews. 161 Study quality is assessed using six criteria:

1. selection bias: the extent to which study participants were representative of the target population

2. study design

3. control of confounders

4. blinding: whether or not outcome assessors, intervention providers and participants were aware of the research question

5. data collection methods

6. withdrawals and dropouts.

An overall assessment of the quality of each study is also calculated, based on ratings of ‘weak’ through ‘moderate’ to ‘strong’ across each of the six criteria. Further study quality criteria are also included in this tool (including intervention integrity and statistical analysis), but these do not affect the overall assessment.

For studies using a survey design we used the checklist designed by Rees et al. ,162 and for qualitative studies the appropriate checklist available from the Critical Appraisal Skills Programme (CASP). 163 In order to calculate an overall assessment for both surveys and qualitative studies, appraisals were then considered alongside these criteria adapted from Kirk et al. :164

1. clear statement of the aims of the study

2. adequate description of the context for the study

3. clear specification of research design and its appropriateness for the research aims

4. reporting of clear details of the sample and method of recruitment/sampling

5. clear description of data collection

6. clear description data analysis provided

7. attempts made to establish rigour of data analysis

8. discussion of ethical issues/approval details

9. inclusion of sufficient original data to support interpretations and conclusions.

Based on the assessments of studies across all of these nine criteria, reviewers judged each study in terms of overall quality as ‘low’ (achieving quality ratings of 1–5), ‘medium’ (achieving quality ratings of 6–7) and ‘high’ (achieving quality ratings of 8–9). Resolution of discrepancies in the independent assessments of quality within each pair of reviewers was resolved through discussion, and, in a small number of cases, a third member of the team was invited to help achieve consensus and to arbitrate. No studies were excluded based on quality assessment ratings alone, although quality was used to inform the interpretation of findings.

Economic analysis

The plan for the economic analysis was sensitive to the type and quantity of evidence found. Where evidence of effectiveness in retrieved papers was accompanied by robust costing and/or cost-effectiveness analyses, our intention was, led by the health economist member of the project team (ML), to use the Drummond checklist165 to assess the quality of this. If little or no such economic evidence was found, attempts were to be made to estimate the cost of the most effective alternative approaches using resource data from the published/grey sources together with the most appropriate unit costs (e.g. Curtis;166 NHS Reference Costs, available at www.gov.uk/government/collections/nhs-reference-costs). Initial plans to add costs and economic terms to the search strategy to identify documents including an economic component were modified in light of the development of the three-arm search strategy already described above.

Data synthesis

Following the searching, sifting and appraising of the evidence, included materials (see Appendices 16–21) were brought together in a series of individual syntheses. Each of these focused on a priority risk category: a type of dislocation or contagion. Researchers informed by the EPPI-Centre approach recognise that a number of different synthesis strategies exist, with decisions on which to adopt reflecting the types of evidence located, theoretical orientations and overall review purpose. 26 Here, a configurative synthesis approach167 was used, a style suited to projects in which research and other materials from different contexts need to be meaningfully and coherently combined in narrative form.

Findings extracted from included citations (see Appendices 16–21) were inspected by members of the project team, and codes were applied to denote the category (or categories) of less obvious risk that each addressed. Available codes were created deductively, to reflect the categories and concepts arising from the joint project team and stakeholder consultation:

• Dislocation: Normal Life

• Dislocation: Identity

• Dislocation: Stigma

• Dislocation: Friends

• Dislocation: Families

• Dislocation: Education

• Dislocation: Psychological Development

• Dislocation: Social Development

• Contagion.

The contents of policy and guidance documents (see Appendix 21) were inspected by members of the project team working individually, and information was coded as above. For each category of risk for which materials had been found, findings and extracts from relevant policies and guidance documents were grouped together. This grouping resulted in the identification of six types of dislocation risk (Normal Life; Identity; Stigma; Friends; Families; Education) and the risk of contagion. No material was coded as either Dislocation: Psychological Development or Dislocation: Social.

Led by the project manager (DE), as per the study protocol, packages of category-specific materials were distributed to review team members. Initially working independently, each reviewer inspected the detailed findings of each included study, clinical case report or practice initiative (see Appendices 16–20) and the relevant contents of policy and guidance (see Appendix 21). Subcategories were created in inductive fashion as reviewers developed familiarity with the materials assembled, an EPPI-Centre approach described by Gough et al. 168 This is consistent with the configurative synthesis approach used here, is recommended for reviews including mixtures of research evidence and policy,169 and is suited where a wide variety of study designs are being considered within tight time frames. 170 Each category-specific narrative synthesis summarising findings from included studies, juxtaposed with the relevant policies and guidelines, was then forwarded to two project members (DE and BH) for review and editing.

The strength of synthesised findings for stream 1 (intervention studies) was assessed using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach,171 where certainty of evidence is reported as being high, moderate or low/very low. Confidence in synthesised qualitative and survey findings was assessed using the Confidence in the Evidence from Reviews of Qualitative research (CERQual) tool, which uses a similar approach to GRADE. 172 The original CerQual approach was designed for qualitative findings and we used the same process but included findings from surveys in the assessment of confidence. Confidence in findings is described as high, moderate or low.

Results

Study characteristics

Summary of phase 2 concepts and methods

A research question and associated objectives were devised to guide phase 2 of the project, centring on the search for evidence relating to the identification, assessment and management of the risks of dislocation and contagion for young people using inpatient mental health services. A comprehensive search strategy was devised to secure evidence of a variety of types. Multiple databases were searched for published and grey material, using combinations of words. The reference lists of included papers were scrutinised for possible additional citations and key journals were hand-searched. Government websites and the websites of relevant organisations were searched, and a call for evidence was prepared and sent to all NHS and non-NHS inpatient CAMHS units identified in the UK. This same call for evidence was circulated via discussion lists and directly to members of organisations with interests in the CAMHS field, and to targeted individuals for onwards distribution via their networks.

Evidence identified was screened through a number of sequential filters. Forty citations (reporting on 38 research studies, reports of local initiatives or clinical case reports) were finally included, alongside 20 policy and guidance documents. Data from the 40 citations were extracted, and quality was assessed using appropriately selected appraisal checklists. The characteristics of included studies were summarised, in preparation for the production of a series of individual syntheses, each focusing on a priority risk category: a type of dislocation, or contagion. These syntheses are reproduced in Chapter 5.

Introduction

This chapter brings together findings from the series of separate syntheses of the evidence, each addressing one category of less obvious risk. Where findings were reported across two publications, data were extracted as follows: the study by Haynes180 is reported in greater detail in the thesis by Painter,181 and all data are extracted from the longer document; the YoungMinds report authored by Svanberg and Street is published across two outputs, the first reporting methods198 and the second reporting findings. 199 All data are extracted from the second document only. In the case of the study by Clemens, two publications177,178 reporting different findings from the larger study are included, and data have been extracted from each separately. In reporting statistical information from the included studies, as much detail has been included (e.g. p-values and effect sizes) as was given by authors of the original papers.

Table 5 summarises the categories and subcategories of less obvious risk for which evidence has been brought together in this chapter.

Dislocation: Normal Life

Five of the included studies report findings related to the risk to young people in inpatient mental health hospital of dislocation from normal life. 176,181–183,199 Three of the policy and guidance documents address this area. 201–203

Two subcategories were created within Dislocation: Normal Life. These were (1) everyday life and interactions in hospital and (2) missing out on life outside and transition home.

Dislocation: Identity

Three of the included studies report findings related to the risk to young people in inpatient mental health hospital of experiencing identity dislocation. 179,181,183 No policy documents were found addressing this area.

Two subcategories were created within Dislocation: Identity. These were (1) mental health problems as identity-changing and (2) responding to threats to identity.

Dislocation: Friends

Ten of the included studies report findings related to the risks around Dislocation: Friends. 175–177,179,181–184,188,199 One of the included clinical case reports addressed this category, giving an account of a 13-year-old boy’s journey through hospital, including a consideration of peer group identity. 221

Two subcategories were created within Dislocation: Friends. These were (1) relationships with young people outside hospital and (2) relationships with young people in hospital. Policy and guidance documents addressing these subcategories are included in each relevant section below.

Dislocation: Stigma

Six of the included studies report findings related to the risk of stigma for young people admitted to inpatient mental health hospitals,176,181–183,190,195 and 11 policy and guidance documents address this area. 8,201,204,206–213 These documents recognise stigma as an issue for users of mental health services201,203,206–211 and that services should be provided in a non-stigmatising environment. 8,209,213

Three of the included papers provided descriptions of the term ‘stigma’. Terminology used and reported included ‘general devaluation’,190,195 ‘disrespect’, ‘emotional insult’, ‘underestimation by others’,190 ‘rejection’, ‘exclusion’, ‘people looking down on you when they find you are in treatment’195 and ‘being treated differently to other patients’. 183 Studies included also address both the apprehension of stigma among young people recently discharged195 and experiences of actual or enacted stigma. 190

Two sub-categories were created within Dislocation: Stigma. These were (1) young people’s experiences during admission and (2) young people’s experiences after discharge.

Dislocation: Education

Seventeen of the included studies report findings related to the risk to young people in inpatient mental health hospital of dislocation from education,11,15,127,128,174–178,181,183,185,186,189,193,199,200 as did 15 of the policy and guidance documents. 93,201–207,210,213–218 One of the included clinical case reports addressed this category, giving an account of a 13-year-old boy’s journey through hospital including a consideration of school provision,222 and one of the included practice initiatives addressed this category. 221

Four subcategories were created within Dislocation: Education. These were (1) education provision and facilities, (2) the quality of inpatient education, (3) academic progress and (4) reintegrating with school after discharge. Policy and guidance documents addressing these subcategories are included in each relevant section below, along with the single practice initiative.

Dislocation: Families

Seventeen of the included studies report findings related to the risk to young people in inpatient mental health hospital of dislocation from families. 11,15,173,175,176,179,183,184,186–188,191,193,194,196,199,200 Five of the included clinical case reports221–225 and one of the included practice initiatives addressed this category. 220 Jones221 gives an account of a 13-year-old boy’s journey through hospital, including a consideration of family functioning; Wilkinson225 describes the principles and practice of involving family members in an inpatient unit; and LePage222 describes the impact of early family participation in inpatient assessment and treatment of youth diagnosed with a first episode psychosis and includes the case of a 15-year-old boy. Nejtek et al. 223 present a case study of a 15-year-old boy and focus on issues around maternal non-compliance, and Olson224 discusses engagement using a parent/adolescent activity group and provides a case description of two 16-year-old boys and a 17-year-old girl.

Three subcategories were created within Dislocation: Families. These were (1) impact on family relationships, (2) family involvement and (3) maintaining contact with families. Policy and guidance documents addressing these subcategories are included in each relevant section below, along with the single practice initiative.

Contagion

Seven of the included studies report findings related to the risk of contagion for young people in inpatient mental health hospital. 142,176,179,183,192,197,199 No policy or guidance documents addressed this area.

Within the literature examined, two aspects of unhelpful behaviour acquired through the process of contagion were identified: deliberate self-harming behaviour in the forms of ‘self-mutilation by skin cutting’ or ‘blood-letting’, and suicidal behaviour. 197 Deliberate self-harming contagion was defined as two or more acts of deliberate self-harming behaviour that involved two or more young people and occurred on the same day or consecutive days. Suicide contagion was described by King et al. 142 as when one completed suicide triggers a cluster of subsequent suicidal behaviours among the contacts of the person who has died. One study192 set out to examine the spontaneous occurrence of self-harm among inpatients without a history of self-harm, but did not define the concept of behavioural contagion. Two studies addressed, briefly, the sharing of unhelpful eating behaviours. 179,183

Two subcategories were created within Contagion. These were (1) experiences of contagion and (2) evidence of contagion.

Economic analysis

None of the studies reported above included an economic analysis or an economic evaluation of alternative ways of identifying, assessing and managing less obvious risks for young people using tier 4 inpatient CAMHS. This is not surprising considering that many studies used qualitative designs and focused mainly on the identification phase of the less obvious risks prioritised by the SAG. The data derived from these studies could not be used to inform an economic modelling exercise of likely NHS costs and analysis of cost-effectiveness. The health economist member of the project team (ML) contacted internationally leading health economists with expertise in the mental health field for specialist advice, and from Professor Martin Knapp (The London School of Economics and Political Science) two additional papers reporting on the economic impact of childhood mental disorder on public services were identified. 226,227 In the study reported in the first,227 of the 445 children with psychiatric disorder included, only three had been admitted to an inpatient psychiatric unit, meaning this paper could not be used to inform our economic analysis. The second227 used data from the same epidemiological survey and could not be used either.

The documents included in this review and summarised in Appendix 21 suggest, however, that the less obvious risks identified in this report exist and are recognised by the Department of Health. This additional economics paper includes the observation that health services in the UK (as in many other countries) are under tremendous financial constraint, dictating that NHS resources be allocated efficiently. 226 In the context of mental health care and services, study designs traditionally thought of as gold standard (i.e. randomised controlled studies) are not always possible and the use of alternative designs should be considered. 228 Modelling exercises are one way of generating evidence in these circumstances, but they rely on good-quality primary data of a type not available to the project team in this synthesis.

Conclusion

This chapter has presented seven individual, contextualised, syntheses of the evidence in the series of priority risk areas focused on in phase 2 of the review. In the chapter following, these are brought together in a metasynthesis, and, in the light of discussions held at a second SAG meeting, plans are outlined for the sharing of this project’s findings and promoting knowledge exchange. Strengths and weaknesses of this project are identified, and implications for future services, practice and research drawn out.

Introduction

Findings from phases 1 and 2 are first brought together in this closing chapter in a matrix, which summarises what has been learned in accessible, single-page style. Key findings are then reviewed and set in a systems context. The project’s original overarching research question and objectives are returned to and reviewed, and, in the light of discussions held at a second SAG meeting, plans are outlined for the sharing of the project’s findings and the promotion of knowledge exchange. Strengths and limitations of the project are considered, implications for services and practice drawn out and recommendations for future research made.

Summary: what has been learned?

This project has been completed at a time when promoting the health and well-being of children and young people is an absolute priority. 229 The matrix in Figure 21 summarises findings from both the phase 1 scoping review and the phase 2 priority area syntheses. A headline is that, although this review has focused on a series of ‘less obvious’ risks which are important to people with stakes in the child and adolescent mental health system, little evidence to guide services and practice in these areas exists.

FIGURE 21.

Using the EPPI-Centre framework to summarise and synthesise all findings.

A first, and obvious, observation is to draw a contrast between the types of risk identified in the phase 1 scoping and the types of risk addressed in the in-depth phase 2 review and synthesis. It is clinical risks, relatively narrowly defined, which occupied the majority of researchers whose outputs were located in the initial segment of this project. As a project team we recognise the importance of research that improves the identification, assessment and management of clinical risks (including the risks of self-harm, suicide and harm to others) in the case of young people with mental health difficulties. As we acknowledged in our original proposal for funding using the NICAPS study11 in support, these are also precisely the types of risk that most often bring young people to the attention of hospital services. Inpatient child and adolescent mental health hospitals exist because it is here that high levels of skilled, continuous care can be provided to young people with the greatest need whose difficulties challenge the capacity of community-based workers. 230 Noting all of this, we are nonetheless also struck by the appetite of the stakeholder representatives we worked with for the prioritisation of a largely different set of risks to take forward into phase 2. Inpatient CAMHS play a vital, but changing, part in complex systems of care (for a discussion, see McDougall et al. 231) and the identification of a series of ‘less obvious’ risks for this project’s in-depth review points to an awareness that inpatient admission can have wider, and long-lasting, consequences about which more needs to be known.

In the areas of risks to normal life and identity policy and guidance was sparse but did recognise that young people undergoing treatment within inpatient settings should be able to lead as normal a life as possible. Views and experiences were reported in rich detail and young people and health-care professionals described boredom, stringent ward rules and routines, and a lack of opportunity for everyday interactions (CerQual – high). Feeling separated from life outside and the subsequent difficulties experienced on returning home were identified as pressing issues by some young people and health-care professionals (CerQual – high). Young people with eating disorders talked about mental health problems eroding their identities (CerQual – moderate), along with the experience of not being treated as individuals (CerQual – low). For other young people it was a struggle to manage threats to the sense of self during admission and treatment (CerQual – low). There were no intervention studies found that focused on the testing of actions to mitigate the risks to normal life or to identity.

In the case of risks associated with friendships and peer relations, policy and guidance are limited to making recommendations on inpatient units having space for visitors. The evidence included in this segment of the project pointed to the difficulties (and ambivalence) young people can experience in maintaining home friendships at a distance (CerQual – high) and in reconnecting with their friends after discharge (CerQual – high). In some cases, connections with friends were significantly associated with levels of postdischarge depression and suicidal ideation (CerQual – low). No intervention studies were found investigating actions to help young people in hospital maintain good relations with their peers at home. Evidence was found pointing to young people’s positive views of being with others in a similar position during hospital care and treatment, in terms of mutual support and companionship (CerQual – high). Young people also spoke of the negative aspects of living with other young people with mental health difficulties (CerQual – high). Some parents were found to be concerned about their children’s sharing of living space with other vulnerable people and at least some young people expressed ambivalence (and even fear) in their relationships with other inpatients (CerQual – low). No intervention studies were found investigating actions to promote positive peer relations among young people who were inpatients.

Managing the risks of stigma and discrimination is a high priority for policy-makers. Young people felt that stigmatising experiences can occur as a result of being admitted, as well as during their inpatient stay (CerQual – moderate) and at discharge (CerQual – low). Being with similar young people can also lead to feelings of acceptance, in contrast with the experience of being rejected in the community (CerQual – low). No intervention studies were found evaluating actions to mitigate the risks of stigma or discrimination to young people admitted to mental health hospital.

Hospital admission poses risks to young people’s schooling. Health-care professionals, parents and young people all recognise the importance of educational provision with appropriate facilities for young people in inpatient CAMHS (CerQual – high), which is also identified as a policy and guidance priority. Smaller class sizes utilising a multiclass format with specialist teaching have been shown in a study involving young people in a RTC in the USA (GRADE – low) to be effective in increasing the amount of work young people are able to produce while in hospital. In the UK, education is provided as standard across inpatient units, but in a majority of hospitals only core National Curriculum subjects are taught (CerQual – high). Improving quality and maintaining good communication and co-ordination across hospitals and schools feature prominently in policy. Within units in the UK, varying teacher/student ratios are found in NHS and non-NHS units (CerQual – high), and good (but not universally so) relations between parents and teachers have been reported (CerQual – low).

One of the disadvantages of inpatient care recognised in policy and guidance is the effects of admission on family life. Training inpatient staff working with young people and their families through the use of role plays or mindfulness did not have a significant impact on the family-friendliness of the admission process (GRADE – low). While on an inpatient unit, young people often feel homesickness (CerQual – high) and experience a range of negative feelings (CerQual – moderate). Associations between family connectedness and postdischarge depression and suicidal ideation have been reported (CerQual – low). Some family members need additional support during their children’s admission (CerQual – low).

Partnership with families during inpatient care is strongly recommended in policy and guidance. Young people whose parents do get involved make significant improvements across a range of treatment and postdischarge outcomes (CerQual – low) but health professionals report that a number of obstacles exist to enable this to take place (CerQual – low).

Whether or not families are fully involved in a young person’s care, the evidence suggests that units should have procedures on visiting and that flexible arrangements should be made for family contact. A particular risk of family dislocation is reported in instances where young people are admitted to hospitals located far from home, in terms of keeping in touch and cost (CerQual – high). For some, the quality of care at inpatient units is considered to be more important than the distance from the hospital to the family home (CerQual – moderate). Some young people also appreciated being away from the home environment (CerQual – low).

The risks of young people in hospital learning harmful behaviours was a priority area for phase 2 of this project, but no policy or guidance was found addressing this. Health professionals and parents have concerns about young people acquiring unhelpful, destructive behaviours while they are inpatients (CerQual – moderate). Young people with eating disorders very quickly copy the behaviour of those around them with the same condition (CerQual – moderate). There is mixed evidence of recorded contagion in inpatient mental health facilities for young people (CerQual – low), with no fixed definition of what constitutes ‘contagion’. No evidence was located investigating actions to mitigate the risks of contagion in inpatient settings.

None of the studies reported above included an economic analysis or an economic evaluation of alternative ways of identifying, assessing and managing less obvious risks for young people using inpatient CAMHS. This body of evidence suggests that this area of research is still in its embryonic stage, where hypothesis generating and exploratory work needs to be carried out to inform the design of large longitudinal studies. Economic analyses would not normally be carried out at this stage, as the size of the problem still needs to be quantified and alternative models of addressing the problem have to be formulated.

Systems and theory

Across the materials brought together in this report, the limited evidence of approaches explicitly used to identify, assess and manage the less obvious risks of dislocation and contagion challenged any attempt to locate and describe any underlying theory informing action in these areas. More generally, underpinning this project has been a recurring concern to set findings in a complex systems framework,5–7 and where possible to develop new theory associated with our narrative synthesis. Our contribution to theory particularly lies in our development and deployment of ‘dislocation’ as a broad class of risk, within which we identified a number of identifiable subdivisions. This concept building guided our phase 2 searches and our subsequent narrative syntheses. It allowed us to organise meaningfully our assembled reports of participant views and experiences, descriptions of services at national or local levels, descriptions and/or evaluations of local initiatives, and quantitative investigations (prospective, retrospective or cross-sectional) into aspects of young people and their care and/or outcomes.

Our commitment to existing theory extends to a reiteration of the idea that inpatient mental health units for children and young people must be thought of as being just one part of a larger, interacting network of services. 231 This system includes community CAMHS, but also, for example, schools. It was in synthesising the evidence surrounding the risks hospital admission poses to education that the clearest appreciation was found (by researchers, policy-makers and standard-setters) of the importance of working across health and education organisational boundaries during young people’s journeys through the system. This kind of thinking also draws attention to the highly specialised, geographically dispersed character of inpatient CAMHS units within the larger systems of which they are a part. Admission to hospital may be clinically necessary and helpful, but the distances units may be located from young people’s homes risk dislocation from family and friends, as well as from school.

A systems approach of this type can also be applied to the more microscopic understanding of young people’s experiences, and of those surrounding them. The evidence synthesised across all the phase 2 priority risk areas reveals the extent to which hospital admission can be disruptive and have major consequences for young people viewed in the context of family and friendship networks, and implications for identity and peer and community acceptance.

Reviewing the overarching research question and project objectives

This project was commissioned to summarise and appraise the evidence in the area of risk for young people moving into, through and out of inpatient mental health care. This has been achieved using a two-phase approach, beginning with a scoping of the literature using selected databases followed by a comprehensive search and in-depth review for evidence in a number of priority areas determined in collaboration with stakeholder representatives.

Specific objectives for the project are stated in Chapter 1 above, and these are reviewed here against the progress made in meeting each. The first objective, relating to summarising the evidence, has been addressed across both phases of the study, but the objective of appraising the evidence has been confined to the second (in-depth) phase. The attempt to identify underlying theory (objective 2) was also confined to phase 2 and was challenged by the limited evidence of approaches explicitly used to identify, assess and manage the risks of dislocation and contagion. Objective 3, relating to views and experiences, has been addressed in the appraisal and synthesis of evidence contained in phase 2 (and particularly the qualitative evidence located and included). Objective 4 has been comprehensively met in phase 2, and in Chapter 5 specifically. The synthesis of costs and cost-effectiveness (objective 5), where the evidence located supports this, has been confined to phase 2, where the limited literature located meant that no economic analysis proved possible. Implications and future priorities (objective 6) have been identified in the concluding section of this chapter, below.

Promoting knowledge mobilisation

As per this project’s commissioned protocol, a second stakeholder advisory group was held with the explicit purpose of sharing progress and discussing strategies for the dissemination of the review’s findings. This was convened in Cardiff on 31 March 2014, and all members of the SAG were again invited. Dr Michael Coffey, from Swansea University, once more acted as independent chair.

A key message arising from this meeting was the value of using a range of available media, both traditional and non-traditional, to promote the reading and uptake of summary and detailed findings. The project protocol refers to the production of accessible summaries, and the team’s intention is that these be produced (drawing on, for example, the single-page matrix reproduced as Figure 21) following peer review and acceptance of this report for the NIHR Journals Library. Care will be exercised to avoid duplicate or redundant publication, and advice will be taken where necessary, but the plan remains to prepare findings and process papers for high-impact journals (with at least one paper published in gold open access format) and in print and online publications targeting practitioners, managers, and young people and their parents. Abstracts have been submitted to conferences, and the first presentation at a national-level event has been made.

Strengths and limitations

This project has succeeded in bringing together and narratively synthesising the evidence in an area which has been largely overlooked in everyday practice and services: the identification, assessment and management of the less obvious risks for young people moving into, through and out of the inpatient component of the inter-relating CAMHS system. Our findings in no way undermine the importance of attending to the more obvious, clinical, risks: but they do alert people with interests in the CAMHS field to the importance of a broader constellation of risks associated with hospital admission. Recognising the importance of these less obvious risks as priorities for the second phase of the review was made possible only because of a close collaboration (in EPPI-Centre style) between members of the project team and stakeholders at a key decision-making event following an initial scoping of the literature. The CAMHS arena is populated by people with a variety of backgrounds and expertise, not all of whom have routine opportunities to shape the direction of policy, services or research. In this context we consider it a strength that this study was informed to the degree it was by the views, experiences and recommendations of people using and working in the system.

The importance to stakeholders of the identification, assessment and management of the less obvious risks to young people moving into, through and out of mental health hospital contrasts with the limited quantity and quality of completed research capable of informing policy, services and practice in these areas. Having first found an under-recognised area for our evidence synthesis, we now find ourselves unable to make comprehensive suggestions for improvement. This recognises, in particular, that very little evidence was found to support actions designed to mitigate the less obvious risks to young people admitted. We return to this theme in our section Implications for health care and recommendations for research below.

This project also has limitations. The limiting of all searches to English-language publications is one, with obvious implications for the non-retrieval of potentially relevant materials from teams possibly working in the non-English-speaking parts of the world. Another relates specifically to phase 1, where the search strategy used included the term ‘risk’. As we later found in phase 2, papers reporting on the less obvious risks do not necessarily use the language of ‘risk’ at all, and a different search strategy (across different or additional databases) might have yielded a series of phase 1 maps of a different character. This leads to a further potential limitation, which relates to our use of umbrella terms and concepts for phase 2. We developed the idea of ‘dislocation’ as an overarching construct to bring together a variety of risks. This helped us to connect different types of evidence across a number of distinct but relatable areas and gave us a robust framework for both searching and synthesising. However, we recognise that the idea of ‘dislocation’ is not a self-evident one. Our processes for searching and sifting materials were scrupulous, but we recognise that degrees of interpretation were used in the identification of evidence judged to make a contribution to knowledge in this broad, constructed, area.

Two final limitations remain. The first relates to the disappointing lack of information retrieved from NHS and non-NHS inpatient CAMHS units, and from others with interests and expertise in the field, in response to our widely circulated call for evidence. Enthusiastic responses to the focus of the study were received, and published papers were sometimes sent to us, but no written local service responses about the less obvious risks were sent. We return to this in Implications for health care and recommendations for research below. Second, the paucity of evidence on alternative approaches to identifying, assessing and managing the less obvious risks of dislocation and contagion challenged our aspiration of conducting a robust health economics analysis.

Implications for health care and recommendations for research

At the risk of overstating, we reiterate that nothing in this evidence synthesis should be interpreted as undermining the value of hospital services for young people with the greatest level of mental health need. However, the trend observed in Chapter 1 towards the diversification of tier 4 services to include specialist, intensive care provided in the community is one that we applaud. We therefore suggest that:

• One way of mitigating the less obvious risks associated with hospital admission is to admit only when all other options have been exhausted, for periods as short as possible and close to home.

The evidence synthesised in this review has implications at the point that judgements are being made on admitting young people to the inpatient component of the CAMHS system. It suggests that:

• As routine, explicit attention be paid to the assessment of the risks of dislocation and to the proactive planning of care which mitigates these.

The evidence suggests that, for young people in hospital:

• Services should balance structured therapeutic activity with space for normalising, everyday interactions.

This review also suggests that:

• Services should be able to help young people in hospital keep in touch with both family and friends, including in the case of admission to units located a long way from home.

The evidence found and included also underlines that:

• It is important to help young people continue with their education during periods spent in hospital, and proactive care planning includes preparation for school reintegration.

Other evidence found suggests that:

• Behaviour by staff (e.g. at mealtimes) can reveal the character of young people’s specific mental health difficulties and this can be experienced as stigmatising.

From the evidence located and synthesised in this review:

• There is no clear support for the implementation of any specific menu of interventions to manage the less obvious risks (as many studies were isolated and of variable quality, sometimes conducted in different parts of the world, where health systems are different from those in the UK, limiting their direct applicability, and knowledge has not accumulated).

What this review therefore does is lay bare areas of real importance for young people’s health-care experiences, and for their longer-term flourishing, about which little evidence to guide services and practice exists. We therefore recommend that:

• Research is needed into the risk of inpatient contagion, where the current evidence on the existence of this is equivocal.

• High-quality studies are also needed to inform action to identify, and mitigate, the risks of dislocation associated with hospital admission; to assess the costs to the NHS of different approaches; and to place service commissioning and provision on a more evidence-based footing.

To this we add the suggestion that:

• Studies should equally address the assessment, identification and management of the less obvious risks across all tiers and components of the CAMHS system (including newer, specialised community services such as those for young people with eating disorders or with challenging behaviours), reflecting the extent to which services are entirely interlinked.

We therefore see that:

• There is a place for a programme of research which begins with the careful exploration of current services and practice with a view to identifying those features found locally to be helpful.

We do not interpret the absence of written policies and service specifications from local CAMHS units in response to our call for evidence as signifying an absence of actions routinely taken to offset the less obvious risks. Describing what happens and what helps to mitigate these, including actions which are not committed to paper in the form of formal policy or standards, would represent a reasoned first step towards isolating and potentially testing specific interventions.

Contributions of authors

Dr Ben Hannigan (Reader in Mental Health Nursing) screened, selected, appraised and synthesised materials, and wrote for and edited the final report.

Mrs Deborah Edwards (Research Fellow) screened, selected, appraised and synthesised materials, and wrote for and edited the final report.

Dr Nicola Evans (Lecturer in Mental Health Nursing) screened, selected, appraised and synthesised materials, and wrote for the final report.

Mrs Elizabeth Gillen (Subject Librarian, Healthcare Sciences) led the development of search strategies, screened, selected, appraised and synthesised materials, and wrote for the final report.

Dr Mirella Longo (Senior Research Fellow in Health Economics) led the health economics analysis, screened, selected, appraised and synthesised materials, and wrote for the final report.

Professor Steven Pryjmachuk (Professor of Mental Health Nursing Education) screened, selected, appraised and synthesised materials, and wrote for the final report.

Dr Gemma Trainor (Consultant Nurse in Child and Adolescent Mental Health Services) screened, selected and appraised materials, and commented on the final report.

The project team acknowledges the help of:

Dr Michael Coffey (Associate Professor, College of Human and Health Sciences, Swansea University) for independently chairing the meetings of the project team and stakeholder advisory group.

Mr Matthew Daniel (Training and Consultancy Manager, YoungMinds) for conducting the series of consultations with young people to inform the identification of priorities for the second phase of the project.

Professor Sally Holland (Professor of Social Work, Cardiff School of Social Sciences) for sharing her experiences and expertise in involving young people in research.

Professor Martin Knapp (Professor of Social Policy, The London School of Economics and Political Science) for his help identifying papers reporting the economic aspects of child and adolescent mental health needs and services.

Professor Lesley Lowes (Florence Nightingale Foundation Chair of Clinical Nursing Research, School of Healthcare Sciences, Cardiff University) for sharing her experiences and expertise in involving young people in research.

Ms Mala Mann (Information Specialist/Systematic Reviewer, Cardiff University) for her help refining the phase 2 search strategy.

All members of the stakeholder advisory group, who helped identify priorities for the second phase of the project, informed plans for dissemination and knowledge mobilisation and generated ideas for future research.

Funding

This project is funded by the National Institute for Health Research Health Services and Delivery Research programme (project number 11/1024/08). The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the HSDR programme, the NIHR, the NHS or the Department of Health.

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 HS&DR 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 HS&DR programme or the Department of Health.

Project aim

The overall aim of this project is to synthesise the international research evidence and the UK grey literature relating to the identification, assessment and management of ‘risk’ in young people (aged 11–18) with complex mental health needs using inpatient mental health services, and to establish what is known about the costs and cost-effectiveness to the NHS of different approaches.

Objectives

Specifically, the project is examining and synthesising the evidence for, and the costs to the NHS of, approaches to risk used as young people:

• make the transition into inpatient CAMHS;

• are cared for in inpatient CAMHS;

• make the transition from inpatient CAMHS to the community;

• make the transition from inpatient CAMHS to adult mental health services.

Acknowledgement and disclaimer

This project is funded by the National Institute for Health Research Health Services and Delivery Research Programme (project number 11/1024/08).

The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the HS&DR Programme, NIHR, NHS or the Department of Health.

Background to the project

One in 10 children and young people between the ages of 5 years and 16 years has a diagnosable mental health problem.

NHS services are provided using a tiered approach.

The most specialised child and adolescent mental health services (CAMHS) are available at tier 4 to young people with complex and/or severe needs.

Tier 4 services have diversified, but inpatient CAMHS units continue to have an important place.

Often ‘risk’ is used as a shorthand to refer to the possibilities of direct harm to self or others, or harm through self-neglect.

These are important areas but, in this study, we are also interested in other, less-recognised, risks (e.g. to the achievement of developmental milestones, psychological maturity, educational attainment, social integration with family and peers, and personal physical well-being).

Methods

We are conducting an evidence synthesis guided by the Evidence for Policy and Practice Information and Co-ordinating Centre (EPPI-Centre) framework.

This approach is particularly recommended for reviews where the findings are destined for practical use by policy-makers, managers and other decision-makers.

The EPPI-Centre approach is two-stage, transparent, rigorous and sensitive to stakeholders.

Applied Social Sciences Index and Abstracts (via Proquest)

URL: http://search.proquest.com/assia/advanced

Date range searched: 1995 to September 2013.

Date search: 12 September 2013.

Records retrieved: 663.

1. su.Exact(“adolescents”)

2. su.Exact(depressed adolescents”)

3. (adolescen* or teen* or “young adult*” or “young people” or “young person” or youth* or juvenile*).ti, ab.

4. 1 or 2 or 3

5. su.Exact(“mental health”)

6. su.Exact(“mental health care”)

7. (mental* or psychiatr* or “tier 4”).ti, ab.

8. 5 or 6 or 7

9. 4 and 8

10. (CAMHS or “child and adolescent mental health service*).ti, ab.

11. 9 or 10

12. su.Exact(“inpatients”)

13. su.Exact(“hospital discharge”)

14. su.Exact(“admission”)

15. (inpatient* or admission* or admitt* or discharge* or transition* or transfer*).ti, ab.

16. 12 or 13 or 14 or 15

17. 11 and 16

18. 17 Limited to 1995-2013

* The BNI and ERIC were also searched using Proquest using a very similar/slightly adapted search to the one above.

Cochrane Library: to include Cochrane Database of Systematic Reviews, Database of Abstracts of Reviews of Effects, Cochrane Methodology Register, Health Technology Assessment and Economic Evaluation Database

URL: www.cochranelibrary.com/

Date range searched: 1995 to September 2013.

Date of search: 16 September 2013.

Records retrieved: 949.

1. MeSH descriptor: adolescent

2. MeSH descriptor: young adult

3. (adolescen* or teen* or “young adult*” or “young person” or “young people” or youth* or juvenile*).ti, ab, kw.

4. 1 or 2 or 3

5. MeSH descriptor: mental health

6. MeSH descriptor: mental health services

7. MeSH descriptor: mental disorders

8. (mental* or psychiatr*or “tier 4”).ti, ab, kw.

9. 5 or 6 or 7 or 8

10. 4 and 9

11. MeSH descriptor: Adolescent Psychiatry

12. (CAMHS or “child and adolescent mental health services”).ti, ab, kw.

13. 11 or 12

14. 10 or 13

15. MeSH descriptor: inpatients

16. MeSH descriptor: patient admission

17. MeSH descriptor: patient discharge

18. MeSH descriptor: transition to adult care

19. (inpatient* or admission* or admitt* or discharge* or transition* or transfer*).ti, ab.kw.

20. 15 or 16 or 17 or 18 or 19

21. 14 and 20

22. MeSH descriptor: hospitals, psychiatric

23. 4 and 22

24. 21 or 23

Cumulative Index to Nursing & Allied Health Literature (via EBSCOhost)

URL: www.ebscohost.com/nursing/products/cinahl-databases/cinahl-complete

Date range searched: 1995 to September 2013.

Date of search: 13 September 2013.

Records retrieved: 1148.

1. (MH “Adolescence”)

2. (MH “Young Adult”)

3. (adolescen* or teen* or “young adult*” or “young people” or “young person” or youth* or juvenile*).ti, ab.

4. 1 or 2 or 3

5. (MH “mental health”)

6. (MH “mental health services”)

7. (MH “psychiatric patients”)

8. (mental* or psychiatr* or “tier 4”).ti, ab.

9. 5 or 6 or 7 or 8

10. 4 and 9

11. (MH “Adolescent Psychiatry”)

12. (CAMHS or “child and adolescent mental health service*”).ti, ab.

13. 11 or 12

14. 10 or 13

15. (MH “inpatients”)

16. (MH “patient admission”)

17. (MH “patient discharge”)

18. (MH “transfer, discharge”)

19. (transition* or transfer*) N5 “adult care”.ti, ab.

20. 15 or 16 or 17 or 18 or 19

21. 14 and 20

22. (MH “Hospitals, Psychiatric”)

23. 4 and 22

24. 21 or 23

25. 24 limited to English language 1995–2013

EMBASE (via Ovid)

URL: www.ovid.com/site/catalog/databases/903.jsp

Date range searched: 1995 to September 2013.

Date of search: 13 September 2013.

Records retrieved: 3617.

1. exp Adolescent/

2. (adolescen* or teen* or young adult* or young people or young person or youth* or juvenile*).tw.

3. 1 or 2

4. exp Mental Health/

5. exp Mental Health Service/

6. (mental* or psychiatr* or tier 4).tw.

7. 4 or 5 or 6

8. 3 and 7

9. (CAMHS or “child and adolescent health service*”).tw.

10. 8 or 9

11. exp Hospital Patient/

12. exp Hospital Admission/

13. exp Hospital Discharge

14. exp Transition to adult care/

15. inpatient*.tw.

16. 11 or 12 or 13 or 14 or 15

17. 10 and 16

18. exp Mental Hospital

19. 3 and 18

20. 17 or 19

21. Limit 20 to (human and English Language and yr=1995-current)

Health Management Information Consortium

URL: www.ovid.com/site/catalog/databases/99.jsp

Date range searched: 1995 to September 2013.

Date of search: 16 September 2013.

Records retrieved: 1053.

1. exp Young People/

2. (adolescen* or teen* or young adult* or young people or young person or youth* or juvenile*).tw.

3. 1 or 2

4. exp Mental Health/

5. exp Mental Health Services/

6. (mental* or psychiatr* or tier 4).tw.

7. 5 or 6 or 7

8. 3 and 7

9. (CAMHS or “child and adolescent health service*”).tw.

10. 8 or 9

11. exp In Patients/

12. exp Patient Admission/

13. exp Patient Discharge/

14. (transition* or transfer*) adj5 “adult care”.tw.

15. 11 or 12 or 13 or 14

16. 10 and 15

17. exp Mental Health Hospitals/

18. exp Mental Health Units/

19. exp Psychiatric Units/

20. 17 or 18 or 19

21. 3 and 20

22. 16 or 21

23. Limit 22 to (English language and yr=“1995-Current”)