Clinicians’ perceptions of the mental health needs of young people in alternative provision educational settings: An exploratory qualitative analysis

ABSTRACT It is estimated that 1 in 2 young people who are educated in Alternative Provision (AP) educational settings have social, emotional and mental health difficulties. However, the extent to which the mental health needs of these young people are currently being met by Child and Adolescent Mental Health Services (CAMHS) is less clear. In this qualitative study, semi-structured interviews were conducted with 15 CAMHS practitioners who had worked with young people educated in AP settings to explore secondary health care service engagement. We analysed the data using thematic analysis. Findings suggest that young people educated in AP settings have complex mental health needs for which they are currently not receiving adequate support, due to shortcomings within AP schools and numerous external barriers to care. Implications for clinicians and service commissioners include a need for further integration between CAMHS and education services, and better quality support across AP settings.


Introduction
Alternative Provision (AP) is an umbrella term which describes all education provided outside mainstream and special needs schools in the United Kingdom (Gill, Quilter-Pinner, and Swift 2017).AP schools primarily provide full-time education for pupils who have been permanently excluded from mainstream schooling, or who would not otherwise receive suitable education due to fixedterm suspension, illness or other reasons (HM Government 2022).The most common state-maintained AP schools are Pupil Referral Units (PRUs), but other settings include hospital schools, AP academies, AP free schools, and independent AP institutions (House of Commons Education Committee 2018; Mills and Thomson 2018).Although the Department for Education encourages eventual reintegration into mainstream schools (DfE 2013), permanently excluded pupils often remain in AP until they have completed secondary education (Gill, Quilter-Pinner, and Swift 2017).
Although there is a paucity of research into the characteristics of young people who are referred to AP for reasons of exclusion, illness or otherwise, an exploratory qualitative investigation by the Department for Education found that the most commonly cited reason for referral to AP was persistent disruptive behaviour, followed by mental health difficulties such as depression and self-harm (Mills and (Braun andClarke 2006, 2019).
2018).These were followed by other factors such as disengagement, non-attendance, violent behaviour and Attention Deficit Hyperactivity Disorder (ADHD).This finding is supported by UK-based epidemiological research which has identified significant associations between mainstream school exclusion and child psychiatric disorder, mental health and behavioural difficulties, social communication and language difficulties, and parental psychopathology (Ford et al. 2018;Paget et al. 2018;Parker et al. 2019).Boys are more likely to be excluded than girls (DfE 2022b).Individuals diagnosed with ADHD and disruptive behavioural disorders such as Conduct Disorder (CD) are the most likely to be excluded (Parker et al. 2016(Parker et al. , 2019)).Children with subclinical difficulties are also more likely to be excluded than children with no difficulties (Parker et al. 2019).
The Institute for Public Policy Research estimated that 1 in 2 excluded pupils in AP schools have recognised social, emotional and mental health difficulties, compared to 1 in 50 in mainstream schools (Gill, Quilter-Pinner, and Swift 2017).Students in AP settings often present with a range of complex vulnerabilities and are exposed to multiple risk factors for behavioural and emotional difficulties, such as poverty, developmental trauma, familial challenges and Special Educational Needs (SEN) (Apland et al. 2017;Cole 2015;Paget et al. 2018;Pirrie et al. 2011).
In addition to poor mental health, there is a high social and economic burden associated with exclusion.Young people who are excluded from mainstream schools are less likely to achieve an intermediate level qualification by age 19 (DfE 2011) and sustain engagement in further education, apprenticeship or employment after completing school (Madia et al. 2022;DfE 2022a).This is also the case for young people with poor mental health, as previous research has demonstrated that mental health difficulties are associated with lower educational attainment (Dickson et al. 2020;Lereya et al. 2019;Smith et al. 2021;Wickersham et al. 2021).Furthermore, excluded young people are more likely to have contact with the criminal justice system.A Ministry of Justice report found that over 60% of adult prisoners reported being temporarily excluded as a child, and over 40% reported being permanently excluded (Williams, Papadopoulou, and Booth 2012).Being permanently excluded from school is also associated with higher levels of antisocial behaviour (Hemphill et al. 2006;McCrystal, Percy, and Higgins 2007), substance abuse (McCrystal, Percy, and Higgins 2007) criminal involvement (McCrystal, Percy, and Higgins 2007;Wolf and Kupchik 2016), criminal victimisation (Wolf and Kupchik 2016) and imprisonment (Wolf and Kupchik 2016).
However, despite growing evidence that children with mental health difficulties are more likely to be excluded from school, and that both exclusion and mental illness are key risk factors for later negative outcomes, there is currently little research on the mental health needs of pupils in AP, and the extent to which these needs are being met (Parker et al. 2015;Whear et al. 2014).Approximately a quarter of referrals to Child and Adolescent Mental Health Services (CAMHS) are turned away due to high thresholds for access to treatment (Children's Commissioner 2022;Firth 2016aFirth , 2016b)).Further, national reviews have found that AP schools often have underdeveloped links with CAMHS (DCSF/DoH 2008; Firth 2016b; HM Government 2022).These findings suggests that young people in AP, despite being amongst the most disadvantaged in society, are at particular risk of receiving insufficient mental health care (Downs et al. 2017).
In this qualitative study, clinicians working in CAMHS were interviewed to explore the extent to which young people in AP are engaged in secondary health care services, in order to explore potential gaps in current provisions and to critically inform service developments in this area.The study had three primary research questions.Firstly, what kinds of mental health difficulties in AP settings appear particularly salient/problematic to clinicians?Secondly, do clinicians consider that CAMHS services as currently configured are able to address the mental health needs of young people in AP settings?Thirdly, how could the mental health needs of young people in AP settings be better met?

Participants
Inclusion criteria were CAMHS practitioners from any professional discipline with experience of working with young people who have been educated in AP.The final sample included 15 participants (11 females, 4 males).All participants were based within the Greater London region.Work experience in CAMHS ranged from 1 to 21 years (mean experience 6 years) and work experience with AP ranged from 2 to 10 years (mean AP experience 4.5 years).Participant characteristics are displayed in Tables 1 and 2.
Ethical approval was attained through King's College London Research Ethics Committee (LRS/ DP-20/21-22015).Participants were recruited purposively, through snowball and chain-referral sampling.A recruitment advertisement was sent to clinicians with links to CAMHS, who were asked to distribute the advertisement to eligible participants.The advertisement was also distributed via a university research circular.Participant recruitment was discontinued when saturation occurred, and interview data ceased to offer novel insights into the research questions.

Measures
The proposed topic guide explored participants' professional experiences with AP, their experiences with young people who attend AP schools, their opinions on whether young people in AP are receiving adequate support for their mental health needs, and suggestions for how these needs could be better met.At the end of the interview, participants were given the opportunity to give any other information they felt was relevant.Open-ended questions were used to obtain information of interest, yet avoid biased questioning that sought specific answers (Braun and Clarke 2012).The flexible nature of the interview questions placed the participant as the expert and minimised the effects of the power dynamic between the interviewer and interviewee (DeJonckheere and Vaughn 2019).
It is considered good qualitative practice to revise the topic guide during data collection to ensure that the analysis is not limited to issues the researchers initially anticipate are relevant (Ziebland and McPherson 2006).Therefore, after review of initial interviews, the topic guide was amended to include additional prompts to help to ensure that the research questions were explored as fully as possible.The final topic guide consisted of 10 questions and 18 additional prompts (see Appendix B).

Procedure
Eligible participants were given an information sheet and a consent form, which were reviewed and completed prior to the interview.Participants were informed that interviews were confidential, and of their right to withdraw from the study.Demographic information was collected at the start of the interview.All interviews were conducted via Microsoft Teams video call, between 29 th April and NaN Invalid Date .Calls were audio recorded lasting between 24 and 66 minutes (mean length 48 minutes).Interviews were transcribed verbatim and checked for accuracy alongside the audio recordings.Personal identifying information was removed from transcripts, and responses were identified via their assigned participant ID.Transcripts were stored securely in a password-protected online repository and audio recordings were deleted after analysis.

Theoretical framework
Data was interpreted with an inductive, 'data-driven' approach, to produce explicit semantic themes.A deductive approach was not selected, as this area of study is relatively under-researched and we were not seeking to confirm existing theories (Creswell et al. 2007).Given the exploratory nature of the study, it was felt that semantic 'surface-level' analysis would be more suitable than latent level analysis, which attempts to capture underlying ideas, patterns, and assumptions with a more conceptual approach to the data (Braun and Clarke 2006).
An essentialist perspective was considered appropriate for the aims of this study.In line with the semantic approach, an essentialist framework prioritises reporting participants' explicit experiences over identifying underlying assumptions and discourses, and assumes a unidirectional relationship between meaning, experience, and language (Braun and Clarke 2006).

Reflexive thematic analysis
Reflexive thematic analysis (Braun and Clarke 2019) was selected to analyse patterns within the data.This method was considered appropriate for the exploratory nature of this research, as it allows rich data to be meaningfully combined (Braun andClarke 2006, 2014).
The analysis adhered to the recursive six-phase approach outlined by Braun and Clarke (2006), which is displayed in Figure 1.Once data had been transcribed and anonymised, transcripts were read multiple times to aid familiarisation of content, whilst noting initial thoughts.Transcripts were created by the researchers, as the transcription process informs the early stages of analysis and fully immerses the researcher in the data (Braun and Clarke 2006;Ziebland and McPherson 2006).The transcripts were uploaded to NVivo, and relevant codes were created across the data set.
Once all data extracts were coded, the One Sheet of Paper Method (Ziebland and McPherson 2006) was used to collate codes into potential themes.This method involved transferring the coded extracts onto one 'sheet' of paper and organising them into potential themes and sub-themes.These were presented on an initial thematic map.Provisional themes were reviewed to ensure that they were representative of the entire data set and addressed research questions.Themes and subthemes were then named, defined and presented in a final thematic map.To ensure quality and credibility within the analysis, the data was independently coded by two researchers (SKS and RW) but discussed regularly within the research team.Themes were collaboratively constructed.This ensured that coherent and meaningful data patterns relevant to the research questions were generated (Vaismoradi, Turunen, and Bondas 2013).Connecting analytic narrative to data extracts was an iterative process, to ensure that data was represented accurately (Ritchie et al. 2013).Furthermore, an audit trail of developing themes was maintained for future researchers.

Reflexivity
The researcher's epistemological position and background can shape research outcomes and mask the participants' voice (Berger 2013).Emotional response to the research can also affect the construction of data, consequently shaping the research conclusions (Kacen and Chaitin 2006).Furthermore, the process of generating codes and themes in thematic analysis is inherently subjective (Braun and Clarke 2021).Accordingly, reflexivity was employed throughout to maintain highquality research by engaging in reflective discussions with the research team, as well as recording internal dialogue and critical self-evaluations in a reflective log.

Results
Four main themes were identified, which summarised participants' experiences of working with young people who attend AP educational settings and their professional views on the challenges in meeting these young people's mental health needs.Themes and sub-themes are presented in Table 3.

Complex vulnerabilities
Participants characterised young people in AP as having more complex mental health presentations than those in mainstream schools.Several participants noted that young people had often experienced complex trauma.They were frequently victims of child maltreatment, and their traumatic life experiences contributed to a diagnostically challenging presentation, as articulated below: In doing something like an ADHD assessment, I think it would be far more complex trying to unpick for the PRU children whether it was kind of ADHD or whether it was more [.Participants observed that many of the young people attending AP had been through the care system and did not have a stable home environment, with parents often experiencing mental illness themselves.Participants highlighted conditions of social inequality, and noted that the young people were often from non-white or migrant backgrounds, and came from low-income and single-parent households.Participants observed further that the young people generally had low educational attainment and came from families with a history of school exclusion.
In terms of diagnoses, participants described frequently working with neurodevelopmental disorders (NDDs) such as ADHD, Autism Spectrum Disorders and Learning Disability.Many of the young people participants worked with had been diagnosed with Conduct disorder, and there were reports of risky, impulsive and antisocial behaviour.Further, many of the young people had difficulties with depression and anxiety, struggled to regulate their emotions and had an early onset of substance misuse.However, participants expressed an opinion that these externalising behaviours were often manifestations of unmet mental health needs.Participant 7 viewed disruptive behaviour as a manifestation of undiagnosed NDD.
It's not uncommon [for there] to be psychosocial problems around the family.So that becomes a distractor [. ..] and there isn't enough expertise to detect that underpinning the child's behaviour -independent of the psychosocial stressor -there are the neurodevelopmental deficits.[Participant 7, Community CAMHS] Several participants viewed externalising behaviours as a manifestation of unidentified complex trauma.Participants acknowledged a connection between disruptive behaviour and early-life traumas, which can contribute to attachment difficulties, poor mental health, and emerging personality disorders.

Detrimental AP environment
Participants observed shortcomings in the AP environment that may exacerbate poor mental health.However, there was some variation in how participants experienced the AP setting and staff.These experiences appeared to be moderated by the type of AP the participant had worked with.

Inconsistency of practice across AP settings
Several participants felt that APs could be chaotic educational environments that may not facilitate positive behavioural change and educational engagement.
[. ..]Young people with behavioural difficulties and, you know, all sorts of problems are all put together.There are a lot of threats made; there's a lot of bad language.You know, there [aren't] really many people in the class to offer that kind of model of good behaviour.[Participant 5, Community CAMHS] However, Participant 12 distinguished between APs with an invalidating learning environment and 'therapeutic' APs: I think a lot of kids know the reputation of the Alternate Provisions in the area and like what's associated with them, and feeling like it's a fight for survival when you get there [. ..]So they don't attend and if they do it's with a lot of hesitation [. ..]Unless it is like one of those therapeutic schools and it feels like: 'Okay this is something I've really struggled with in mainstream education and I'm finally getting access to an opportunity where I'm going to be supported [. . .] to learn in a style that meets my needs'.[Participant 12, In general, it appeared that the practitioners who reported positive experiences tended to be those who worked embedded within an AP.Most of these practitioners experienced the AP as a supportive and contained learning environment.However, these participants acknowledged that this experience may differ elsewhere: With the [APs] that aren't attached to us, they still have some limitations on the number of staff they have, the amount of space.They tend to have really small buildings, where they don't have the luxury of having like a sensory room, or a place where the young person can go with a member of staff and try to, like, deescalate things.[Participant 14, Hospital School Provision] Participants perceived inconsistency across provisions as problematic, as they felt that it was not guaranteed that all young people will have access to the same level of resources and supportparticularly those who do not meet any diagnostic criteria that could facilitate access to a more therapeutic AP, such as a diagnosis of NDD.

Lack of specialised training
Participants' experiences with staff tended to be varied, with Participant 5 describing staff's mental health literacy as a 'lottery'.Although a few practitioners had positive experiences, it was generally felt that there is a need for more specialised training of staff to understand and support mental health difficulties.
These are really complex young people, and people need to know about how to support them and how to understand the behaviours [. . .]We should be really valuing these people that are doing this incredibly hard job and looking after our young people, and they haven't been offered training to understand that.[Participant 8,Community CAMHS] Participants noted a tendency for staff to focus on correcting young people's outward behaviour rather than trying to understand the needs underlying their behaviour.
There seems to be more focus on the behavioural difficulties -i.e. the conduct -and the intention of the conduct always seems to be sort of malicious -or that's how it's construed by the educational facilities.Rather than this child has a need, or this child has this traumatic background.[Participant 6, Community forensic-CAMHS] They have very meagre resources, and they [the staff] start just focusing on medication, medication, and more medication.And they start not wanting to look at the hard work that is needed in terms of understanding the child and undoing certain psychosocial issues that are complicated to do.[Participant 7, Community CAMHS] Further, many participants found that where difficulties were identified, AP staff were sometimes unreceptive to CAMHS advice.
The staff, I think, found it quite difficult to hear that: 'Well these behaviours might have to do with what this child is experiencing from the past, or what is going on for them outside the school.'Or their behaviour is an expression of a need rather than just a difficult behaviour [. ..]It would perhaps feel pushy for them in a way that, you know, actually they're thinking: 'I'm finding it really difficult, and I think this behaviour is unacceptable, and I just don't want it.'[Participant 9,Inpatient CAMHS unit] As highlighted above, participants acknowledged the challenges that teachers face and suggested that they may require therapeutic support themselves to prevent burnout and meet the demands of their job.
You're working with quite a large group of quite dysregulated children or young people, and I think over time, as skilled as you might be, it can be incredibly exhausting and if you're kind of working off adrenaline every day, it makes it really hard to use some of that training that you've got [. ..]I think again they would have benefited from space for their own personal resilience, where they could actually stop to have a space to think a bit more, rather than working off adrenaline for weeks at a time until they would finally have a half-term break.[Participant 13, CAMHS based within PRU]

Barriers to young people receiving support
Except for a minority who attend APs with an explicit therapeutic focus, participants felt that, overall, young people in AP were not receiving adequate support for their mental health needs.Several barriers to young people receiving CAMHS support were reported.

Rigidity of statutory services
Participants highlighted a clear need for flexibility when working with young people in AP.However, the rigidity which was perceived to be inherent to statutory services like CAMHS was seen to contribute to the problem of accessing mental health support.
Many participants experienced young people in AP as difficult to engage therapeutically and noted that they were often hesitant to access available services due to the stigma associated with poor mental health.It was also felt that young people may be wary of approaching services due to feeling failed by them in the past.
I think we were working with a group of young people who were, you know, more complicated to build relationships with [. ..]When working with young people in mainstream schools [. ..]I think you managed to build a relationship a lot quicker and you are able to, kind of, use the suggested guidelines around assessment and treatment [. ..]Whereas with the PRU's I found that it took a lot longer to build the relationship.So you would spend five to six sessions just getting to know the young person [. ..] before you even start the idea of assessment or treatment in a more formal way.[Participant 13, CAMHS based within PRU] For these reasons, it was suggested that young people in AP may benefit from a period of preparatory therapeutic engagement to better prepare them to successfully utilise mental health services.However, participants believed that CAMHS tends to close these 'hard-to-reach' cases before they have been given this opportunity: They may be not at the point where they're ready to engage and do active work, but they need some sort of support . . .And often CAMHS will go, 'Well, they're not ready to engage so we'll close'.When, actually, if there was a way of supporting them towards engagement -that might help things.[Participant 10, Community CAMHS] Stringent entry requirements were highlighted as a barrier to young people receiving help.For example, as expressed by Participant 6, CAMHS may be reluctant to engage young people who are actively using substances: I think that lots of young people are, you know on the cusp of coming into CAMHS but aren't able to access the support that they desperately need because they [. ..] are using cannabis.And CAMHS don't always take that perspective that, you know, if we provide them with something else -another way that they can manage their behaviours -then they are less likely to use [. ..]They can actually stop that vicious cycle, but it's almost as if while the child is using, they're denied that access.[Participant 6, Community forensic-CAMHS] As a result of this rigidity, participants emphasised that the current system tends to be reactive rather than proactive in addressing needs, as they felt that young people are often only identified by services once their mental health has deteriorated significantly.
Once a problem is developed and created massive issues and disruption, then a need will be, kind of, identified or responded to [. ..]And that's often feedback I get from young people and parents, especially within the Youth Offending Service: 'Why did it take until my kid got in trouble with the law for them to get access to all this?' [. ..]We'll see them on our doorstep [. ..] when they meet diagnostic threshold.But it's much too reactive and too late in the game.[Participant 12, Participants also drew attention to systemic differences in the ways that young people from racial minority backgrounds access CAMHS.Participants noted that these young people tend to be referred to CAMHS via emergency routes, for example when they have entered the criminal justice system or have been admitted to hospital, rather than at a stage where early intervention is possible.
Practitioners acknowledged that the CAMHS referral process takes a long time due to the volume of requests and the fragmented nature of services.The protracted process of CAMHS referrals is captured by Participant 12: We have really long waiting lists at CAMHS as well, so while they might get access -what is, access?Is getting on a waiting list access?And if by the time you're at the top of the waiting lists, are you actually going to access it then?Or have things gotten worse?[Participant 12, YOS-CAMHS]

Lack of support within AP
Many participants felt that AP schools were often not sufficiently resourced by local authorities to facilitate CAMHS support for learners.Participant 9 conveyed frustration with 'per pupil' funding models: The local authority of the CAMHS service that I was working for decided to offer all schools [. ..]CAMHS provision for free [. ..]And it was done on the basis of numbers.So obviously, you know, the bigger the school, the more hours they were getting.So obviously a PRU has a very small number of students [. ..]So the number of hours of free support the PRU was receiving were smaller than some of the primary schools, or the mainstream primary schools, just because the number of pupils were smaller.But obviously that doesn't correspond with the level of need.[Participant 9, Inpatient CAMHS] A lack of 'in-reach' CAMHS could result in referrals being missed, as participants believed that teachers often lack the required knowledge to make a 'successful' referral.Participant 10 described referrals being passed back-and-forth due to inadequate referral information.
We will get referrals through from schools saying this young person is really angry, they need support from CAMHS, and CAMHS will go, 'well that's not really a mental health need' [. ..]So there may have been previous referrals into CAMHS, but the information that was provided just didn't indicate that there were mental health needs.And so it wasn't picked up.[Participant 10,Community CAMHS] It was suggested that missed referrals may contribute to mental health difficulties not being identified, and manifestations of these difficulties being signposted as 'disruptive behaviour' within APs.
I think there was a lot of missed mental health [. ..]The PRU I worked in had [. ..] some places for young people that were there because of behavioural difficulties.Then there were some that were there for Special Educational Needs [SEN].So those young people [on the SEN pathway] would often have come with some kind of diagnosis or EHCP [Education, Health and Care Plan], that quite specifically had goals for them [. ..]But outside of the SEN kind of things, for the young people that were there for the behavioural pathway, quite often there wasn't a diagnosis in place.[Participant 13,CAMHS based within PRU] The difficulty of not having an EHCP in place was raised by several participants and was perceived as delaying the young people's needs being met.

Systemic family challenges
Practitioners expressed that it was often difficult to gain parental consent to engage young people in CAMHS work: I think with mainstream schools, the young people would mostly come from the kind of families where they were more open to input, and they would give consent openly at the beginning.Whereas with the PRU, sometimes it would be the case that you wouldn't have consent from the family [. ..] and if that wasn't possible, you might not be able to do your work.[Participant 13,CAMHS based within PRU] Participants had observed that parents could be particularly wary where social services were involved, due to fears about the potential consequences.Further, where parental consent was gained, participants noted difficulties for families in supporting the young people's needs at home, often due to complex difficulties within the family itself.
Success and outcomes, I found when in the AP, were very dependent on the willingness of the parent or carers to engage and to put in any strategies at home, and when there's, sort of, real kind of obvious trauma [. ..] it can be really painful and difficult I think for parents to be able to accept what effect it had on their child, when you're talking about, quite often, a painful and difficult part of their lives as well.[Participant 5, Community CAMHS]

Need for a therapeutic approach
Many participants felt that the goals of AP should not be purely educational, and that young people's therapeutic and social care needs should be prioritised.Participants highlighted the relationship between good mental health and having a positive educational experience.They emphasised that young people are less able to succeed educationally when their mental health needs have not been met.I would say that alternative provisions should be set up and run based on a therapeutic model.And I think that the education side should come after building that secure base [. ..]Kids can't do any of those kinds of higher functions until the basics are in place, and I think that a lot of PRUs would do better working on that first and building the relationships, and then introducing, you know, academic stuff a little bit later on [. ..]Yes, of course, the main aim is to educate, but being reasonable, you're not going to be able to do that until somebody is in a safe enough space to learn.[Participant 11,CAMHS based within PRU] Likewise, participants expressed that it is difficult to support young people's mental health needs when they do not have a stable educational setting.
Having a stable and, kind of, responsive and understanding educational placement is so, so important.I think that mediates good outcomes for young people [. ..]It's often really difficult [. ..] when it's not set up to meet their needs.Then it's often really difficult to effect change in mental health, because the environment just isn't conducive to it.[Participant 8,Community CAMHS] Further, many practitioners discussed how therapy can play a complimentary role to education by improving young people's functional abilities, self-esteem, concentration and emotional regulation, and in some cases can assist with re-integration back to mainstream schools.Participant 8 suggested offering therapy to all young people in AP, rather than only those who are referred to CAMHS: My team's been looking at offering DBT [Dialectical Behaviour Therapy] skills programmes in schools [. ..]So intervention for all the young people to kind of think about ways to manage emotions [. . .]We know that most young people in those sectors will have difficulties with those things.So we can offer something to everyone.

[Participant 8, Community CAMHS]
A number of participants felt that schools should take a more holistic, trauma-informed approach on an individual and organisational level.
The trauma-informed approach, you know [. ..] it targets the whole organism, and therefore you need different people with different specialisms in order to do that, and you need to be able to organise your school day to accommodate that, and you need, you know, people to work with families and parents as well.[Participant 5, Community CAMHS] Participants expressed a need for partnership working to tackle the complex difficulties young people in AP face.In general, it was felt that there is a lack of integration between CAMHS, education and social services.
I always emphasise that children with complex needs [. ..] -the boundaries between social care, health, and education are blurred.So we need to work together and you can't just contact a doctor and say, 'Can you please give medication for this child' [. ..]Because unless you approach from the education point of view and also from the social care point of view, we will not get anywhere.[Participant 7,Community CAMHS] Participants noted challenges in inter-agency working related to efficient communication and information sharing -often due to different understandings of the same behaviour.
If something's happened, like, at school or at the alternative provision and that hasn't been shared with methat can really influence [. ..] how I'm understanding the risk [. ..]Often there's really different perspectives from different disciplines, which is a strength in that, you know, you're getting lots of perspectives, but I think sometimes it can be difficult when responding to behaviours in different ways.[Participant 8,Community CAMHS] In terms of service integration, practitioners who worked with an AP on-site tended to describe higher levels of integration between CAMHS and education, more multidisciplinary team working and better shared understanding.Below, Participant 2 describes her experience of working with the AP at an inpatient CAMHS unit: We were very integrated, and the education team were, you know, part of our multidisciplinary team.We would feedback quite a lot of our assessments which would help them, you know, structure their sessions or their, say, their risk assessments [. ..]We were really lucky because we had that culture of -you do your therapy and you go to education.[Participant 2, Inpatient CAMHS] All participants expressed a belief that there is a need for CAMHS to have an embedded role in AP schools.Many participants felt it would be helpful to have CAMHS on-site, and highlighted multiple benefits to further partnership.These included early identification and intervention in mental health needs, improving staff's understanding of mental health, easier communication and informationsharing, greater ability to improve therapeutic engagement, and increased access to support for young people, families and staff.Participant 7 expressed a sense of urgency for a regular CAMHS role: We need to be in the school [. ..]The battleground of the brain is in the school, so that's where we need to be.And I hope that eventually this will happen because we're just, you know -trying to grasp on thin air.[Participant 7, Community CAMHS] However, two participants who had worked within an AP setting felt that in order for the alliance between CAMHS and APs to be effective, more thought needed to be put into building shared expectations and understandings.
For a lot of head teachers and senior leaders, you know, the idea of having mental health support for their children is: 'Yes, great -makes sense' but if you actually dig into it and ask more . . .I don't know that they actually have a good understanding of what the purpose of it is, and what the rationale is, and how it can be most helpful.[Participant 9,Inpatient CAMHS unit] Sometimes being in the school made it really quite difficult as a practitioner . . .I think there was always a pressure that the consultation would turn into direct work, which wasn't useful or indicated at that time.So I do think that it is important to have CAMHS support and to have CAMHS input, but I think there needs to be a lot of thinking about how its offered, thinking about also managing your resources, as well as managing expectations, really, about what realistically can be done.[Participant 13, CAMHS based within PRU]

Discussion
To the best of our knowledge, the current study is first exploration of CAMHS clinicians' perceptions of the mental health needs of young people in AP settings.We sought to answer three research questions.Firstly, what mental health difficulties appear particularly salient/problematic to clinicians in AP settings?Our findings confirm existing evidence highlighting that pupils in AP have more complex mental health difficulties than those in mainstream schools and are exposed to multiple vulnerabilities.Secondly, do clinicians consider that CAMHS services as currently configured are able to address the mental health needs of young people in AP settings?Participants felt that students in AP settings lack access to sufficient support for their mental health due to numerous barriers.The type of AP setting appeared to influence the amount of support CAMHS could offer, highlighting inconsistencies in the quality of APs and adequacy of staff training.Thirdly, how could the mental health needs of young people in AP settings be better met?Our results indicate the need for APs to incorporate a more therapeutic, trauma-informed approach to education, with CAMHS forming an integral part of AP.
Participant's descriptions of 'complex vulnerabilities' within AP are consistent with findings that excluded young people are characterised by poor mental health and complex family problems, increasing the risk of negative life outcomes (Gill, Quilter-Pinner, and Swift 2017;Paget et al. 2018).Participants noted that diagnoses including NDD and CD, together with lifetime histories of complex trauma, were common amongst youth in AP.Associated difficulties such as aggression, impulsivity, rule-breaking and poor emotional regulation can present challenges for mainstream schools and cause difficulties for pupils to cope with academic demands, which increases the risk of exclusion (Ford et al. 2018;Parker et al. 2016).Mental health difficulties are often not formally recognised before exclusion, and many excluded young people fall below formal diagnostic thresholds which secure access to CAMHS services or specialist AP enrolment (Gill, Quilter-Pinner, and Swift 2017;Parker et al. 2019).Our results also suggest that young people with behavioural difficulties often do not receive specialist support unless they have a clear comorbidity, such as an NDD or a diagnosed severe mental illness.In many cases, they are only identified by services once their mental health has deteriorated significantly.Metaanalysis demonstrate that the onset of mental illness occurs before the age of 14 for one third of individuals, and before the age of 18 in almost half (Solmi et al. 2021).Mental health intervention is most impactful in lessening poor outcomes when targeted at the time of onset (Correll et al. 2018).A review of CAMHS transformation plans by Firth (2016b) suggests that eligibility thresholds based exclusively on diagnosis may prevent individuals with sub-threshold difficulties, behavioural problems and complex trauma histories from receiving early assessment and intervention.
Our findings support evidence that fragmented services reduce accessibility to CAMHS for those enrolled in AP.Mental health care is spread across separate organisations, creating gaps in provision and confusing referral processes (CQC 2017).These difficulties are exacerbated by inadequate funding, resulting in staff shortages and longer waiting times (Firth 2016b;Pettitt 2003).Firth (2016b) further notes that the fragmented commissioning process results in young people being offered different services depending on their address -a phenomenon termed the 'postcode lottery' (Gill, Quilter-Pinner, and Swift 2017).When a therapeutically orientated AP is unavailable, students with mental health difficulties are often sent to PRUs without a needs-led plan for their pathway (DCSF/DoH 2008).As observed in our results, mental health needs in PRUs may be missed or viewed simply as behavioural difficulties.Young people with psychological needs require support that is sensitive to their developmental stage, therefore being placed in a PRU may delay timely intervention and result in missed opportunities for improving long-term outcomes of mental illness (Fusar-Poli 2019; HM Government 2022).
Both qualitative research and our findings highlight inconsistency in the quality of APs with regard to their suitability for young people with mental health needs (Apland et al. 2017;McCluskey, Riddell, and Weedon 2014).Differences in funding provided by local authorities may lead to APs receiving insufficient resources and training (HM Government 2022; Taylor 2012).Subsequently, teachers hired often do not have Qualified Teacher Status (Gill, Quilter-Pinner, and Swift 2017) and lack guidance in identifying and supporting mental health needs (DCSF/DoH 2008; Firth 2016b; Lowry et al. 2022).AP staff are often expected to manage complex needs with inadequate support (Ofsted 2022;Walter, Gouze, and Lim 2006), potentially explaining reports in our study of AP teachers experiencing burnout.These factors may contribute to a deleterious educational environment and the deterioration of existing mental health difficulties for students.
The present study also showed the difficulties CAMHS experience in engaging youth in AP.Firth's (2016b) review of CAMHS transformation plans highlighted that young people with the highest level of need are generally the most difficult for external services to engage.As suggested in our study, CAMHS may not be flexible enough to offer support to these 'hard-to-reach' young people.Prior research has highlighted both stigma and negative perceptions of statutory services as barriers to young people accessing mental health services, resulting in difficulties becoming more complex and difficult to address due to lack of treatment (CQC 2017;Gould et al. 2009;Holland et al. 2020).Further, racial minorities represent a large majority of the AP population and are underrepresented in CAMHS, yet over-represented in adult inpatient services (Bradley 2009;Malek and Joughin 2004) suggesting that early intervention is less likely.Prior research has overlooked differences in the way that racially diverse groups in APs access CAMHS and represents an important area for future research.
Our findings suggest that the remit of AP needs to expand to meet young people's therapeutic needs alongside their educational requirements.There is emerging evidence that mental health services embedded within schools can positively impact both mental health and educational attainment, particularly when interventions are targeted towards students with existing emotional and behavioural difficulties, and those at-risk of mental illness (Fazel et al. 2014).An Ofsted (2010) review noted that therapy which supports young people's emotional wellbeing was more effective when integrated with educational provision and considered the young person's family and social context.However, as highlighted in our findings, there may be challenges in implementing this strategy within the AP context due to budgetary constraints.Equally, the optimal therapeutic approach for externalising behaviours in later adolescence remains unclear (Fonagy et al. 2018).
As suggested in the present findings, differences in priorities and training can often result in inefficient joint-working between different agencies involved in young people's care (Firth 2016b;Vostanis et al. 2012).Existing evidence points to lack of integration between CAMHS and education providers, noting that where CAMHS interventions exist in AP, they are often an add-on service rather than an integral part of the institution (Cole 2015).However, a randomised controlled trial by Panayiotopoulos and Kerfoot (2007) found that a multi-disciplinary intervention for pupils excluded from primary school due to disruptive behaviour significantly reduced the number of future excluded days and emotional and behavioural difficulties compared with routine care.Further, in a recent report by Ofsted (2022), staff employed in APs for primary-age pupils reported that working with CAMHS helped them better understand pupils' needs and triggers for difficult behaviour.These findings suggest that a collaborative approach within APs is potentially key to intervening in the mental health of excluded youth and improving future outcomes.
In the present study, CAMHS clinicians who worked on-site with AP schools reported higher levels of integration and better multidisciplinary working.They also evaluated the AP staff as more supportive and better trained to understand mental health.Our results suggest that future policy should focus on securing regular CAMHS roles within AP schools.This role could assist in reducing inconsistencies across provisions, promoting multi-disciplinary working, increasing access to hardto-reach groups, providing family interventions, facilitating staff training regarding mental health awareness and managing challenging behaviour.There is a significant financial burden in neglecting the negative outcomes associated with exclusion, with an estimated cost of £370,000 per excluded young person in lost taxation, benefits, healthcare and criminal justice costs (Gill, Quilter-Pinner, and Swift 2017).Given the established links between poor mental health and negative life outcomes in each of these domains, future policy should prioritise improving the interface between education provision and health care, by developing an integrative AP curriculum that targets mental health and behaviour alongside educational attainment.

Strengths and limitations
This study has several strengths.The use of semi-structured interviews permitted collection of rich data which enhanced the understanding of an under-researched topic.Furthermore, CAMHS practitioners from multiple disciplines were interviewed, capturing a range of perspectives.The researchers were independent of CAMHS, minimising social desirability bias and encouraging open responses.
The study also has limitations.Firstly, most participants were based in London, where there are bigger budgets for CAMHS and a higher prevalence of social, emotional and mental health needs (Firth 2016a).Future research could explore whether the experiences of clinicians in other regions of the U.K. converge with the findings of this study.Secondly, the mean years of participants' experience working with young people in AP was 4.5 years, with just over half of participants having 3 or less years of experience in the field.Researchers did not, however, observe significant thematic differences in the data provided by participants with the highest and lowest levels of experience.Thirdly, the use of purposive sampling may have resulted in participants with strong positions on the AP educational system.However, this exploratory research aimed to capture novel insights into how current provision of mental health services for young people in AP could be improved, and such potentially 'strong' positions were therefore to be welcomed.

Conclusion
Although young people in AP present with complex difficulties and have a high level of need, their mental health is not always adequately supported by current statutory provision.Implications for clinicians and service commissioners include a need for further integration between CAMHS and education services, and better quality support across AP settings.A proactive stance in addressing mental health difficulties has the potential to improve outcomes for young people.
. .] complex trauma or dysregulation [. ..]You would always have to do that [. ..] extra thinking about past experiences.It would be hard to determine, to be honest, what exactly the difficulties were.[Participant 13, CAMHS based within PRU] I think there's a lot of complex trauma, and I think because of the overlap [in] presentation with ADHD, hyperkinetic Conduct [Disorder] [. ..] emotional regulation, hypervigilance, restlessness and externalising, avoidance, anger [. ..]I do think there's a lot of under identified or under-diagnosed trauma and some kind of need in that regard [. ..] that's masked within these externalising behaviours.[Participant 12, Youth Offending Services (YOS)-CAMHS] Thomson Figure 1.Recursive phases of reflexive thematic analysis

Table 1 .
Participant characteristics and demographics.

Table 2 .
Participant ID and role within child and adolescent mental health services.