Child Deaths in Custody - Learning lessons
In 2011-12, my office had the mournful responsibility of investigating three apparently self-inflicted deaths of children in Young Offender Institutions (YOls). These were the first such child deaths since 2007, although we have also investigated 38 self-inflicted deaths of young people aged 18 to 21.
The origin of this bulletin lies in the request for advice on any collective learning emerging from these investigations which was made to me by the Minister for Prisons and Rehabilitation, Jeremy Wright MP. This is absolutely appropriate: the death of three children in such circumstances is a tragedy that compels all those in authority to make every effort to ensure that lessons are learned as quickly as possible and effective steps taken to avoid any repetition. My investigations, individually and collectively, have a role to play in this - hence the production of this bulletin. It is anonymous and briefly draws out several recurring concerns that arise from the three investigations (which will be published after the inquests). The bulletin suggests the early learning that might help avoid such tragedies in future - and ensure children in custody are better safeguarded. I hope these lessons are learned. "Throughout this report we use 'young people' to refer to children and young adults eligible to be held in YOls (aged under 21 or sometimes 22). Nigel Newcomen CBE, Prisons and Probation Ombudsman
The three children were extremely vulnerable but, following the court decision to place them in custody, there does not seem to have been sufficiently detailed consideration given to the best placement to help manage their vulnerability. In two of the cases, when it became clear that the boys were struggling to cope with a normal YOI regime, they were not moved to specialist units within the YOls despite alternatives being available.
In two of the cases, the respective Youth Offending Teams (YOTs) had recommended placing the children in a Secure Training Centre (STC) rather than a YOL In both cases, the reason given was the child's vulnerability. Only one of the recommendations expanded on this to explain the nature of the vulnerability as learning difficulties, bullying and thoughts of suicide while he had been held in the YOI before sentencing.
Neither recommendation for an STC placement was acted upon. In one case, the recommendation was not communicated to the court so the Youth Justice Board (Y JB) were only able to allocate to a YOI (the law has since changed to allow flexibility). In the other case, the reasons why the placement recommendation, made by a YOT worker who knew the child, was rejected by the Y JB were not documented. An alternative to either an STC or YOI placement was the Keppel Unit (a specialist YOI unit reserved for the most vulnerable), but this was discounted.
Once at the YOls these two boys continued to show signs of extreme vulnerability including withdrawing from social contact and self-harm. Both of the YOls had units for vulnerable individuals and those experiencing some form of crisis. These units have fewer young people' and more staff to allow more intensive support. In both cases a move to the wing was considered and proposed. Both boys were initially resistant to a move and this in effect acted as a veto. As it was considered to be in the children's best interests to move, more active measures should have been taken to promote the benefits, such as arranging for them to visit the units and involving others such as families or the YOT workers in the discussions. Given their young age and the concerns of staff, including known bullying in one case, an enforced move may have even been appropriate. When one of the boys changed his mind about relocation there was no-one who had clear responsibility for acting on this and a transfer was not arranged.
There were different issues in the third case. The boy's vulnerability was less apparent, although his YOT worker had identified that he often hid his emotions. He also had some characteristics indicating increased risk, including it being his first time in custody and his status as a looked after child. However, he appeared to staff to be settling in to the YOI and socialising with the other young people. An incident where he behaved in an aggressive manner towards an officer led to him being moved from the induction wing earlier than usual and before he had completed his induction. Increased vulnerability due to his static risk factors and the change in circumstances do not seem to have been considered in this sudden move, nor the possibility of his challenging behaviour being a mask for acute distress.
2. Sharing Information:
Concerns had been identified about the risk of selfharm for all three boys before they entered the YOls. In particular, one boy was known to have selfharmed shortly before entering custody. Risk was recorded and reported in different ways for each child: a pre-sentence report, a 'Risk of Serious Harm' form completed after sentencing and a Person Escort Record (PER). Risk was also recorded inconsistently across the different records and forms for the same child - partly due to different approaches taken by different agencies, partly due to poor recording, and partly because of confusion about the relevance of certain information.
In one case, an incident of self-harm identified in police custody was shared with escort staff but not with the court custody staff. The court could have raised a self-harm warning form to notify the YOI on his arrival there, instead there was only an undated note in the PER about self-harm which was taken as a historical event and given little consideration. The YOT worker for another of the boys had recommended that an ACCT be opened upon arrival. This did not happen and there is no record that it was considered when he arrived at the YOI or why the recommendation was rejected.
3. Suicide and Self-Harm
3.1 Assessing risk: Two of the children were looked after children and the third had a statement of special educational needs. Two were in custody for the first time; the other had previously only spent a brief period in a YOI on remand. One had been sentenced for violence against his partner. These are known static risk factors for self-harm. The YOTs, the Y JB and the YOls were not consistent in their assessment or evaluation of the risk these children posed to themselves. Partly this resulted from communication difficulties and incomplete information being shared between these agencies and others (including the police, courts, and social services). However, there did not seem to be a shared understanding of vulnerability or how to assess it between these agencies.
Pre-sentence reports by the YOT included the factors that indicated the children were vulnerable. However, when considering risk of self-harm or suicide the YOls seem to have relied principally on what the children said and how they seemed to the staff at that time. While this is one aspect of a holistic assessment of risk, it apparently overrode consideration of other factors. The judgment of experienced staff and the child's own view are important but the assessment also needs to give sufficient weight to past behaviour and information contained in pre-sentence reports of those, such as the YOT workers, who are already familiar with the child. After all, the best predictor of future behaviour is past behaviour.
Following reception, patterns of behaviour that should have indicated heightened risk were not always recognised. Two of the boys increasingly isolated themselves, remaining in their cells during meals and association. The behaviour was known to staff and it is a known risk factor for self-harm but was not always well recorded. This self-isolation, and other incidents such as damaging the cell, were often considered separately rather than as part of a broader picture of the child's state of mind and as warning signs of distress.
3.2 Use of ACCTs: Assessment, Care in Custody, and Teamwork (ACCT) is the Prison Service process for monitoring and supporting those considered to be at risk of serious self-harm or suicide. Used principally with the adult estate, it is also the system used for children and young adults who are considered at risk. Two of the boys were being monitored under the ACCT system when they died. The investigations concluded that the ACCT process was insufficiently child-centred to deal with the complexities and special vulnerabilities of children.
One boy had reported thoughts of self-harm but repeatedly assured staff he had no intention of acting upon the thoughts as he would not want to cause his family any distress. Staff were concerned for him but it was not until after smashing his TV and causing cuts with the glass that an ACCT was opened. A recommendation from the YOT that an ACCT be opened on his arrival was not acted upon. In the second case, the ACCT was again not opened until after an act of self-harm when a member of staff noticed cuts, again despite previous concerns.
One of the boys consistently rejected the idea he was having thoughts about self-harm and was never subject to an ACCT. He displayed few signs of vulnerability while at the YOI, although he was a looked after child and it was his first time in custody. After an incident where the boy became extremely aggressive to staff and caused damage to his cell, his behaviour was treated solely as a discipline issue with little examination of the cause. This extreme behaviour should also have prompted consideration of his level of risk and an assessment of his mental health.
3.3 Conflict between care and discipline: In the two cases where ACCT was used, there was conflict between the ACCT processes and disciplinary procedures. All three children misbehaved - showing aggression to staff, damaging their cells, and blocking the doors or observation panels. Generally, even for the two children who were subject to ACCT monitoring, this was only approached as a disciplinary matter and not as an indication of increased risk or distress. A more holistic view of the children's risk and vulnerabilities that included poor behaviour was required.
During their time in custody, all three were subject to punishments arising from adjudications. One of the children subject to an ACCT also received a large number of warnings under the rewards and sanctions scheme and some sanctions were not legitimately applied. In most instances, managers conducting adjudications did not explore the full background and circumstances as they should have done. Despite their availability, none of the children had advocates to represent them at adjudications.
There were isolated examples of good practice with adjudications being stopped out of concern for the child and consideration of vulnerability influencing a rewards and sanctions review but overall there was little evidence of a more holistic approach. Instead, the approaches taken to discipline and to managing vulnerability appear to have operated entirely separately. This was particularly evident when adjudication punishments withdrew association and communal eating, while at the same time an objective in the child's ACCT was that he try to engage more with the other young people.
PSI 08/2012 which concerns care of young people states that the safeguarding children policy's 'essential purpose is to ensure that [other] policies are properly integrated in respect of safeguarding. This means moving away from a model of working principally with young people in "silos" and ensuring that work to safeguard children and young people is co-ordinated effectively across the core areas of the establishment'.
Unfortunately, in the three cases, the approach to safeguarding and behaviour management using ACCT, adjudications, and the rewards and sanctions system did not routinely operate as an integrated system. Care and discipline were not consistently co-ordinated and the formal adult-orientated adjudication system appeared an inappropriate way to manage vulnerable children.
3.4 Monitoring and reviewing AccTs: ACCT reviews are designed to monitor progress against the objectives of the care plan and determine appropriate action when the level of risk changes. The quality of the reviews varied. Often the child was consulted and there was an attempt to ensure the wing staff who had regular interaction with the child were present, but full multi-disciplinary reviews were rare. Greater involvement from the YOI healthcare teams and others involved in the child's care including the YOP, the family, and social services would have provided the perspectives necessary for a holistic assessment of the situation. In particular, the lack of senior management involvement in these reviews led to little high level information sharing about these vulnerable children, what should be done to help them and how this would be arranged.
ACCT reviews are meant to determine what further action is necessary, so involvement from all those caring for the child is important to ensure a consistent approach. One ACCT was closed and then reopened very quickly on a number of occasions. The decision depended a lot on the way the child presented and the ACCT was closed despite key objectives in the care plan not having been achieved. This is contrary to ACCT guidelines and, as the objectives are intended to address the factors leading to heightened risk, this is concerning. The ACCT care plans tended to be focused on addressing specific behaviours such as refusing food or selfisolation, rather than focusing on the reasons behind such behaviours.
3.5 Responding to risk: The response to risk was not always consistent or sufficient. It was not until the children self-harmed that ACCTs were opened, despite existing concerns. Even with the ACCT in place, the choice and numbers of observations of the vulnerable children were sometimes not proportionate to risk. In one case the child smashing his cell furniture did not lead to an ACCT review, despite his unusually aggressive and unpredictable behaviour. In another case not only was the child subject to an ACCT but on the evening he fatally self-harmed there were specific and serious concerns raised by the staff on duty. The number of observations carried out each hour was increased but the possibility of constant supervision was never considered. The concerns were raised in the evening when a multi-disciplinary review may have been difficult but the decision about observations was taken by a member of staff who was unfamiliar with the situation.
Ultimately, ACCT is used mainly with adult prisoners but few adaptations had been made to reflect the needs of children. In particular, there was little discussion at ACCT reviews that challenging behaviour from children could be a sign of distress
and that children might be less able to communicate this distress. One boy subject to an ACCT repeatedly blocked his cell observation panel which obstructed monitoring, but alternative approaches that might have been feasible for a child - such as leaving his door open - were not considered.
4. Mental Health
4.1 Access to mental health care: All three children entered custody with previously identified mental health concerns. Two had diagnoses of ADHD, and were prescribed medication, although of the two only one was in regular contact with the Child and Adolescent Mental Health Services (CAMHS). In the third case, the concerns were less well defined: there was a known history of cannabis use and the YOT worker was worried about his ability to come to terms with the recent death of a family member. The YOT worker also noted how adept he was at concealing his emotions.
The two children with ADHD were referred for mental health screening and had appointments with psychiatrists after their arrival, although this was focused mostly on their ADHD medication rather than wider mental health issues. One was assigned a named mental health nurse. Both were seen by mental health staff but often presented more positively in these reviews than at other times in the YOL In both cases mental health staff noted that the children scored their moods much lower than staff expected. While there seems to have been regular monitoring, acts of self-harm, missed medication and concerns from wing staff do not seem to have led to an escalation in the mental health support provided (except for closer monitoring through ACCT).
The other child was not referred to mental health screening after reception. The assessment relied mostly on self-reporting by the child who denied feeling miserable or experiencing powerful memories of traumatic events. The nurse identified 'no concerns' and so no referral for a full assessment was made. He was seen by the Young Persons' Substance Misuse Service but, even following an aggressive outburst, was not referred for mental health assessment.
Problems with the mental health screening tool do not seem to be unique to children. We have investigated the deaths of two 18 year olds at a YOI (in 2010 and 2011) who both scored a 0 (no concerns) on reception screening despite having recent, documented, mental health problems.
4.2 Medication: Two of the boys had prescribed medication for ADHD. Both missed or refused to take their medication frequently in the period before their deaths. A doctor spoke to one boy about the need to take his medication regularly but there was no systematic approach or escalation following missed medication. Although it is ultimately up to the individual to take medication, the reasons why it was being refused should have been explored - whether by healthcare or another trusted adult - and the situation more closely monitored, particularly as they were children.
Refusal to engage with treatment can be a sign of increased risk and, in one case, there was also a suggestion that the need to leave the wing to collect the medication marked the boy out as a target for bullying. (ADHD medication is controlled and must be dispensed centrally.)
5.1 Tackling bullying: Bullying is a widely recognised problem in YOls and known to increase the risk of self-harm. It has been a factor in almost all PPO investigations of self-inflicted deaths in YOls of both children and young adults. Two of the investigations in this report uncovered evidence of bullying and found that staff were aware of, or suspected, bullying but there was a lack of a robust response.
One child reported this bullying repeatedly to a variety of staff. However, he gave no names and without names staff said they were unable to act, even though it was clear who he meant. Evidence from CCTV suggested that even when staff witnessed harassing behaviour from other young people it was not adequately challenged.
In the other case, bullying was not reported and might have occurred only the evening before the boy took his own life. However, bullying by shouting out of cell windows was a recognised problem at the establishment and indeed across YOIs. The YOI had produced a new policy to tackle the problem but the investigation found it was not being implemented robustly. Shouting from windows was not regularly recorded in the Shout Out Log and staff said that it was difficult to challenge when the large number of young people made it hard to identify either the protagonists or the target.
There were concerns raised by staff and family about possible bullying in the third case, but the investigation found no evidence of this.
5.2 Personal Officer Scheme: The Personal Officer Scheme at one of the YOls had been criticised several times by prison inspectors. The PPO investigations in the other YOls also noted that it took too long for a personal officer to be assigned or introduced to the child. The personal officer is supposed to be the child's main point of contact and should try to establish a relationship of trust with the child. In the case of two children identified as particularly vulnerable, it was unfortunate that more priority was not placed on ensuring they had effective personal officer support.
This is not to say that efforts of staff to offer support were not commendable. Examples included staff informally taking on the role of personal officer, special effort to ensure a child was able to make a phone call and visits to one child by a particular staff member whom he trusted. However, busy YOls can struggle to ensure a consistent and reliable staff presence which allows for the building of trusting relationships and a supportive environment. Given the vulnerability of the three boys, and the difficulty at least two had in trusting adults, such relationships might have been more effectively developed with staff in another setting with better staff resources, such as a Secure Training Centre or the Keppel Unit at Wetherby YOL
5.3 Training: The investigations raised concerns about the appropriateness of some of the adult-orientated processes in YOls, such as the health screening tool, the adjudication process, and the operation of rewards and sanctions schemes. At the same time staff would benefit if the training they received was more childfocused; helping them to recognise behaviour that indicates heightened risk, increase understanding of the issues faced by vulnerable or looked after children, and supporting a more robust approach to bullying.
One boy disclosed past sexual abuse to a member of staff (the first time he had spoken of this at the YOI). While there was no criticism of the officer involved who appeared to have handled this to the best of their ability, it is a concern that prison officers who work with (often vulnerable) young people are not given training about how to handle such discussions.
6. External Support
Continuing involvement of those who had been supporting the children before they entered custody was also problematic. YOTs, the family, and social services were not routinely included in care planning as a result of the two ACCTs. Such input might also have been useful if the refusal of medication had been more comprehensively followed up and when, in two cases, transfers to wings in the YOls for vulnerable children were discussed. One child who had been in care was not assigned a social worker for most of his time in custody. He was seen by several duty social workers but this meant continuity of his care and planning for the future was further complicated.
Contact with family was important in all three cases, although for different reasons. One boy had repeatedly tried to telephone his mother without reaching her. An incident of aggressive behaviour seems to have been at least partly a result of distress at lack of contact on Mothers' Day. This does not seem to have been recognised by the YOI.
Anxiety about his family, and particularly his mother's health, was regularly spoken of by another boy - contact with his family was sporadic as the boy often spent all of his phone credit the day he received it. He was known to rely heavily on his family for support but, despite an open ACCT, little thought was given about how to allow him to better access that support, such as through enhanced telephone contact. It is notable that some STCs permit young people to have telephones in their rooms, which would have given these young people much greater flexibility about when they were able to access support from friends and family.
In the third case, contact with family was especially problematic and had to be arranged through the Local Authority. The YOI did try to help; facilitating visits and calls with his foster parents, and efforts were made to arrange contact with his grandmother through the Local Authority.
7. Lessons to be learned
Many of the issues raised by the three recent deaths are not unique. The impact of bullying, weakness of reception assessments of vulnerability and mental health, weakness of personal officer schemes, and problems with AccTs have been identified in our past investigations of child deaths between 2004 and 2007. However, by drawing common themes together in this report there may be scope to learn lessons and avoid such tragedies in future. These include:
Lesson 1 - Busy YOls struggle to give the individual attention necessary to care for the most vulnerable. Accordingly, allocation to an STc or specialist unit within YOls, such as the Keppel Unit at Wetherby, needs to be considered and pursued.
Lesson 2 - There are problems sharing information and a lack of shared understanding of vulnerability which can hinder co-ordinated care of the children across agencies.
Lesson 3 - Assessments of vulnerability and risk of self-harm did not adequately weigh static risk factors against presentation or fully take into account the complex ways children can show emotional distress.
Lesson 4 - AccT processes were insufficiently child-centred, and the involvement of senior managers, families and outside agencies in care planning was too limited.
Lesson 5 - Managing risk, treating mental health, and the wider operation of YOI processes, particularly disciplinary procedures, need to be better integrated to ensure children are treated holistically and consistently.
Lesson 6 -YOls need to ensure a more robust response to bullying and that reports of bullying are acted upon.
Lesson 7 - Personal officers offer an important point of contact and support. They should be assigned quickly upon reception and regular contact with the child or young person fully documented.
Lesson 8 - Sources of external support, including but not limited to families, can be extremely important. Enhanced access to this support at times of crisis and for those at particular risk of selfharm should be facilitated wherever possible.