Investigation report: Non-accidental injuries in infants attending the emergency department

Due to the nature of the subject matter no specific incident was used to explore this area of care. Instead, the investigation analysed 10 serious incident reports (reports written by NHS trusts when a serious patient safety incident occurs) to identify the factors that contribute to non-accidental injuries not being diagnosed. These factors were grouped into themes, which informed the terms of reference for the investigation.

The investigation engaged with national stakeholders to sense-check the themes which had been developed to ensure that they were meaningful. The stakeholders also provided insights which helped to contextualise some of the themes.

The investigation also visited three acute trusts and interviewed medical and nursing staff to understand the frontline challenges associated with the identified themes and where change may mitigate this safety risk.

• There is no specific guidance for ED clinicians on the identification of suspected non-accidental injuries and what to do if they suspect an infant has a non-accidental injury.
• There may be barriers to routinely escalating cases of children with a potential non-accidental injury to paediatric (child specialist) and safeguarding teams.
• Delays in the availability of information about potential safeguarding concerns add to the pressures on ED staff when making decisions about infants with potential non-accidental injuries.
• There remain concerns about, and an inconsistent approach to, sharing safeguarding information between organisations.
• The Emergency Care Data Set (ECDS) gathers information about ED attendances and includes a field for when such attendances are related to safeguarding.
• The ECDS safeguarding information collected is not currently utilised within the NHS and there is minimal quality assurance in place to ensure that it is reliable.
• Risk factors for non-accidental injuries which do not meet the criteria to be included on the Child Protection – Information Sharing system (the electronic system designed for information sharing between the NHS and social services) are not included in a patient’s summary care record and may therefore remain unknown to clinicians.
• The investigation identified mechanisms which could enable safeguarding information that is not currently available to ED clinicians, to be made available through existing national and regional digital systems.
• Safeguarding teams are often located physically distant from EDs. This can create a barrier to communication and liaison with the team.

During discussions with trust staff, the investigation learned of actions that had been implemented locally to embed child safeguarding into the culture of EDs:

• Induction training: Benefits were seen where relevant staff had undergone a training session during which adult and child safeguarding was discussed, with a focus on how and when sharing of information between organisations is appropriate for safeguarding.
• Simulation training: The regular inclusion of a safeguarding element within simulation training has been seen to increase the confidence of staff to deal with safeguarding issues.
• Psychosocial multidisciplinary team (MDT) meetings: These meetings were held to discuss children (under 18 years) who had attended the ED and about whom there were safeguarding concerns. It was attended by various safeguarding stakeholders, including the police, violence reduction team and the probation service, as well as members of the local mental health team, ED clinicians and paediatricians. The MDT acted as a proactive step in sharing relevant multi-agency information to identify children who may be at risk of harm.

1.1.1 ‘Working together to safeguard children’ (Department for Education, 2018a) defines physical abuse to a child as ‘hitting, shaking, throwing, poisoning, burning or scalding, drowning, suffocating or otherwise causing physical harm’. The document states that ‘physical harm may also be caused when a parent or carer fabricates the symptoms of, or deliberately induces, illness in a child’.

1.1.2 This report uses the term ‘infant’ to describe children under 1 year of age. This age group is known to be at high risk of abuse (Davies et al, 2015).

1.1.3 There are various terms used by professionals to describe physical harm deliberately caused to a child. These include non-accidental injuries (National Society for the Prevention of Cruelty to Children, 2021) suspected physical abuse, child abuse and inflicted injury. The investigation uses the term non-accidental injury throughout this report for consistency.

1.1.4 This investigation focuses on infants who attend EDs which are consultant-led, available 24 hours a day and can treat any emergency patient. Some such EDs have a dedicated children’s area.

1.2.1 While abuse can happen to anyone regardless of their circumstances there are recognised factors which increase the risk of any type of child abuse or neglect. These include:

• domestic abuse
• parental mental health problems
• parental substance misuse (National Institute for Health and Care Excellence, 2019)
• a child being ‘looked after’ – that is, under the care of the local authority (Clever et al, 2011).

1.2.2 Children are most at risk if multiple risk factors are present (Clever et al, 2011).

1.3.1 Healthcare staff are ‘often the only eyes and ears in the community that can spot signs of abuse before young children go to school’ (National Society for the Prevention of Cruelty to Children, 2022).

1.3.2 In 2018 there were an estimated 21,489 child (under 18) attendances at EDs and minor injury units due to violence-related injury. Of these attendances, 1,781 were of children between 0 and 10 years old (Office for National Statistics, 2020a). Data from Great Ormond Street Hospital, Ofsted, and the government all show recent and significant increases in the number of suspected non-accidental injuries in children (National Society for the Prevention of Cruelty to Children, 2021).

1.4.1 Section 11 of the Children Act 2004 sets out the duty on the NHS and other organisations, such as the police and local authorities, to safeguard and promote the welfare of children. NHS employment contracts also place a safeguarding duty on all staff.

1.4.2 Additionally, the Children Act 2004 (as amended by the Children and Social Work Act 2017) identifies NHS clinical commissioning groups (CCGs) as one of the three ‘safeguarding partners’ for a local area. Integrated care boards have now taken on the statutory responsibilities of CCGs following the Health and Care Act 2022. The other two safeguarding partners are the local authority and police. Together these three agencies have a shared and equal duty to make arrangements to work together to safeguard and promote the welfare of children.

1.4.3 HM Government produced a guide on how these agencies should work together to fulfil these duties (Department for Education, 2018a). This document states that CCGs (and now integrated care boards) must have a designated doctor and a designated nurse who provide advice and support in relation to child safeguarding matters. Similarly, all providers of NHS-funded health services must have a named doctor and named nurse (and a named midwife if maternity services are provided) for safeguarding children. These individuals provide advice, promote safeguarding practices, and ensure that safeguarding training is in place.

1.5.1 The intercollegiate (multi-organisational) document, ‘Safeguarding children and young people: roles and competencies for healthcare staff’ (Royal College of Nursing, 2019) sets out a competency framework for child safeguarding training in the NHS. The standards set in this document are recognised by HM Government in ‘Working together to safeguard children’ (Department for Education, 2018a) as those to be taught in child safeguarding training.

1.5.2 The framework outlines five levels of competence (one being the lowest level and five the highest) and indicates the job roles that require the different levels of training. Level one, for example, must be completed by all staff working in healthcare services whereas level five must be completed by those in specialist roles, such as designated doctors and nurses (see 1.4.3).

1.5.3 The framework document states that an ‘annual appraisal is crucial to determine individuals’ attainment and maintenance of the required knowledge, skills and competence’. It also requires a mandatory 30-minute training session to be completed as part of general staff induction and refresher training to be undertaken a minimum of every 3 years.

1.5.4 Additionally, information to assist paediatricians (specialists in children’s health) with decisions about suspected non-accidental injury is set out in ‘Child protection companion’, published by the Royal College of Paediatrics and Child Health (RCPCH) (2017). The RCPCH also published ‘Facing the future: standards for children in emergency care settings’, a document developed by the Intercollegiate Committee for Standards for Children and Young People in Emergency Care Settings (Royal College of Paediatrics and Child Health, 2018). This document sets out standards of care for those providing urgent and emergency care to children, as well as guidance on the implementation and auditing of these standards.

1.5.5 The Royal College of Radiologists (RCR), in conjunction with the RCPCH, has produced guidance for clinicians, which states that ‘a senior clinician (usually a paediatrician) who suspects that a child has been subjected to physical abuse will need to request medical imaging to identify injury’ (Royal College of Radiologists, 2018). The investigation acknowledges that initial imaging may be requested by ED staff.

1.6.1 ‘Working together to safeguard children’ (Department for Education, 2018a) states that ‘Anyone who has concerns about a child’s welfare should make a referral to local authority children’s social care and should do so immediately if there is a concern that the child is suffering significant harm or is likely to do so’. Professional advisory documents echo this; for example, safeguarding guidance for doctors states ‘It is vital that all doctors have the confidence to act if they believe that a child or young person may be being abused or neglected’ (General Medical Council, 2020).

1.6.2 Individual NHS trusts have developed internal policies and procedures for the escalation of safeguarding concerns.

1.7.1 NHS Digital developed the Child Protection – Information Sharing (CP-IS) electronic system, to enable information sharing between social care and the NHS.

1.7.2 Children are added to the system by social care if they have a child protection plan (that is, they are suffering, or likely to suffer, significant harm) or are looked after by (under the care of) the local authority. Pregnant women/pregnant people who have an unborn child protection plan will also be on this system (that is, where the child of a pregnant women/person will be on a child protection plan once born).

1.7.3 NHS staff can access the system and find out whether a child in their care is subject to such a safeguard. Social care will receive an alert when a child who is on the CP-IS system makes an unscheduled visit to a health or care setting.

1.7.4 The CP-IS system also allows social care and clinical staff to know if a child known to the system regularly makes unscheduled visits to a healthcare setting, such as an emergency department. The rationale for the system is that clinical staff:

‘… will be able to use this information as part of their overall clinical assessment, along with information about where and when children have previously been receiving urgent treatment. This will help them build up a better picture of what is happening in the child’s life so they can alert social services if they think something might be wrong.’
(Department of Health and Social Care, 2012)

1.7.5 This system is in place for all local authorities in England, has been endorsed by the Care Quality Commission and features in the NHS standard contract for service providers, which states:

‘The Provider must co-operate fully and liaise appropriately with third party providers of social care services as necessary for the effective operation of the Child Protection Information Sharing Project.’
(NHS England, 2022)

1.8.1 Expert opinion from the Royal College of General Practitioners in the co-authored ‘safeguarding children and young people’ (Royal College of General Practitioners and National Society for the Prevention of Cruelty to Children, 2014) highlighted possible barriers to suspecting non-accidental injuries. These were reiterated in national guidance (National Institute for Health and Care Excellence, 2019) and include:

• uncertainty about local procedures
• demands on time
• perceived difficulty associated with child protection issues
• concerns about confidentiality, consent, and data protection
• poor connectivity between the IT systems used by hospitals, midwives and health visitors (The Child Safeguarding Practice Review Panel, 2021).

1.9.1 There is a local safeguarding practice review system which seeks to identify learning from child protection incidents that result in serious harm or death. This high threshold means that many incidents of non-accidental injury, that do not fall into the categories of serious harm or death, are not investigated in this manner. They may be investigated at a local level by trusts, social services, primary care providers and others, often independently of one another and with no mechanism for routine sharing of learning.

2.1 Finn was 4 weeks old when he was brought into the emergency department (ED) by his mother and father due to sustaining a ‘bruise to the face’.

2.2 The healthcare assistant who was involved in caring for Finn was worried about how distressed Finn was and escalated this to the junior doctor and senior staff nurse on duty.

2.3 The doctor examined the 2 cm lump on the side of Finn’s head and spoke with Finn’s parents. The parents told the doctor that Finn’s head had dropped when he was being winded and had struck the father’s shoulder blade, causing the lump. The doctor described the parents as well dressed and as having appropriate interaction with their child. In addition, they considered that the parents had brought Finn into the ED promptly following the injury and noted that there was no social services involvement with the family.

2.4 The doctor considered the explanation of the injury to be acceptable. This explanation was, in fact, somewhat different to that given by the parents to the nurse looking after Finn. The nurse had been told that the father had been holding Finn out on his forearm when Finn’s head had fallen forwards, striking the father’s shoulder. The nurse documented that the father was unaffectionate towards Finn despite the baby’s evident distress.

2.5 An emergency nurse practitioner (ENP) working in another area of the ED happened to see Finn’s details on the electronic patient record system. They went to the area where he was being cared for in the ED to check whether non-accidental injury was being considered, given Finn’s age. The ENP briefly spoke with a senior nurse who explained that the doctor had examined Finn and spoken with the parents and that this was not considered to be a non-accidental injury. The ENP was reassured and returned to their area in the ED.

2.6 Finn’s temperature, heart and respiratory rates were all with the expected ranges. He was given paracetamol and was subsequently discharged home with his parents.

2.7 Details about Finn’s attendance were entered onto the electronic patient record. However, it was incorrectly recorded that he was a ‘mobile’ child (that is, he could crawl/walk) rather than ‘immobile’. This meant that the inbuilt alert to prompt consideration of non-accidental injury was not activated.

2.8 Four days later, Finn’s parents returned with him to the ED as he had blood coming from his nose and mouth. His parents explained that when they went to check on Finn sleeping, they had found him to have a blanket covering his face and had noticed the bleeding.

2.9 The safeguarding team were contacted and due to concerns about the injury further tests and scans were undertaken. These revealed multiple fractures, as well as other injuries.

2.10 The police and social services were alerted and came to the hospital. Finn was placed into foster care.

2.11 Jakob was 12 weeks old when he was brought to the ED by his mother. She told the doctor assessing Jakob that he was not moving his leg as normal and that he had caught it in the bars of his cot while rolling.

2.12 The nurse caring for Jakob noted from the electronic patient record that his mother was known to social services due to a history of domestic violence. However, there was no other information to indicate concern about Jakob’s welfare.

2.13 When Jakob was examined, no bruising or any other causes for concern were found. An X-ray of his leg was taken and was reported as normal – that is, no fractures were identified. Jakob was discharged home with his mother with a planned follow-up X-ray 2 weeks later. Information about Jakob’s attendance at the ED was shared with social services.

2.14 The following day, Jakob’s mother brought him back to the ED as she said she was very worried about his leg. A second X-ray was taken which was again reported as ‘normal’. Examination by an ED doctor did not find any swelling or tenderness of concern. Jakob was discharged home with a follow-up appointment for 3 days later.

2.15 At this follow-up appointment, no new concerns were identified, and it was agreed Jakob could be discharged with no follow-up appointments arranged.

2.16 A family nurse practitioner (a specially trained nurse who provides support through home visits) had been allocated to support Jakob’s mother in pregnancy and after his birth. They saw the information sent to the GP practice about Jakob’s visit to the ED. They were concerned about the injury and whether it might be non-accidental. The family nurse practitioner contacted the safeguarding team at the hospital who advised that the clinical team had not raised this as a concern. The safeguarding advisor agreed to contact the doctor involved and speak with relevant colleagues.

2.17 A meeting was held between the doctors who saw Jakob on his first and second ED visit and safeguarding colleagues. During this meeting, the doctor who saw Jakob on his first ED visit said that they had felt concerned about the possibility of non-accidental injury. However, they said as it was not clear that this was the cause, and as the X-ray was normal, they had not pursued their concern further by contacting the paediatric or safeguarding teams. During the meeting, and on further interrogation of records held, it was found that Jakob had a ‘child in need plan’ in place due to the history of domestic violence. All at the meeting agreed the possibility of non-accidental injury needed to be explored further.

2.18 Jakob’s mother was contacted and asked to bring him for a further X-ray. This X-ray showed a healing fracture of his left femur (thighbone) which, following expert review, was identified to be present on the X-ray taken at Jakob’s first ED visit (19 days earlier).

2.19 Jakob was admitted to a children’s ward while further tests and scans were undertaken.

2.20 Jakob was later discharged under an amended child in need plan which included regular visits from a social worker and his family nurse practitioner.

2.21 Amelia was 6 weeks old when she was taken to the ED by her father. She was bleeding from her nostril and mouth and her father said he had found her like this and had no idea of the cause. Amelia was examined by a paediatric doctor who found several marks, including a 2 cm bruise on her head, a graze on her chin, and swelling and redness to her nostril.

2.22 When asked by the doctor about the bruise, her father said he did not know what had happened but thought that Amelia’s brother, who was 18 months old, may have thrown a toy at her. Her father said he had no idea about the graze.

2.23 A consultant paediatrician was asked to see Amelia and they diagnosed a skin infection to her nostril. After further conversation with Amelia’s father, the consultant paediatrician was content that the bruise to her head was likely to be accidental.

2.24 No enquiries were made about social services involvement and when Amelia’s blood test results were found to be within expected ranges, she was discharged home.

2.25 Twenty-two days later, Amelia was referred to the paediatric team at the hospital by the family’s GP, who was concerned about her wellbeing. Amelia’s mother had made comments to the GP about not wanting her and that Amelia cried so much that the mother was worried she might harm her. As part of the referral, the GP mentioned that social services were involved with the family and that Amelia was known to the vulnerable babies team and had regular contact with her health visitor.

2.26 Amelia was admitted to the paediatric ward where, like the GP, nursing staff were concerned by comments Amelia’s mother made to them. For example, she had said that she had not wanted Amelia and that she had not bonded with her and found her ‘aggressive’. No injuries or marks were seen when the paediatricians examined Amelia. A discussion took place with social services, the outcome of which was that Amelia could be discharged home with her mother with the existing support in place.

2.27 Nursing staff on the paediatric ward were ‘surprised’ and ‘worried’ that Amelia had been discharged and shared their views with the trust’s safeguarding team. This led to a multi-agency strategy meeting being arranged.

2.28 In this meeting, information the clinicians were unaware of came to light. Specifically, that Amelia’s father had a history of domestic abuse and was currently denied access to his daughter from a previous relationship. The named doctor for safeguarding at the trust was on leave when the meeting took place, and it was agreed that they should review the case and make a decision about next steps on their return.

2.29 When the named doctor for safeguarding reviewed the case, they were concerned about the injuries Amelia had when she visited the ED and the comments made by Amelia’s mother on the ward. Another multi-agency strategy meeting was arranged. Here it was agreed that Amelia should be brought to the hospital for a child protection examination and other investigations.

2.30 Amelia’s mother brought her to the hospital and a full examination was carried out. This revealed multiple healing rib fractures. Clinicians involved in this attendance also considered that the injuries present at Amelia’s initial ED attendance were not sufficiently explained and were most likely non-accidental injuries.

2.31 Amelia and her brother were taken into emergency foster care.

3.1.1 In 9 of the 10 investigation reports analysed a key contributory factor identified by the local investigation was that the clinician(s) assessing the infant had not fully explored the way in which the injury occurred, or had accepted at face value an explanation that in hindsight did not fit with either the injury or the developmental ability of the infant. For example, Jakob (see 2.11) was 12 weeks old when he was seen by an ED doctor with a fractured femur. His mother said the injury was caused by Jakob rolling around in his cot and catching his leg in the bars. With the benefit of hindsight, it was recognised that an infant of this age would not have been able to roll in their cot.

3.1.2 Another example is that of an infant aged 5 weeks who was brought into the ED having ‘cut his top lip’ and was found to have a torn frenulum (tissue between the lip and the gum). His father said this happened during bottle feeding; this was accepted and the infant was discharged home. In hindsight, during the trust’s investigation it was agreed that such an injury could not have been sustained during bottle feeding without undue force having been used.

3.1.3 Alongside an acceptance of a family’s account of an injury, the reports found that clinicians gave attention to diagnosis and treatment rather than potential safeguarding concerns. Typical of this, one report described a ‘focus’ on ‘clinical management … rather than whether there was a non-accidental injury’. This focus was found to be particularly prevalent in junior doctors who had less experience of non-accidental injury.

3.1.4 In addition to not exploring the mechanism of injury sufficiently, the reports identified that vulnerabilities known to be linked to non-accidental injury (for example, a child having teenage parents) were not given sufficient weight in decision making.

3.1.5 While most of the reports did not include any information about why non-accidental injury was not recognised or explored, 3 of the 10 reports included some explanation of this. In one, it was documented that the clinician was reassured by the fact that the parent brought the child to the ED promptly following the injury. This was seen as appropriate care and responsible action, which was seen as evidence that the injury could not be non-accidental.

3.1.6 In another report, bias about personal attributes that might signify parenting ability appeared to have informed decision making. The doctor had written that the parents seemed ‘like nice parents and sensible’, contributing to the doctor’s opinion that the injury was not non-accidental.

3.1.7 The investigation found that in 3 of the 10 reports analysed, the staff involved had not completed the recommended safeguarding training. This meant that, in the majority of reports, a lack of training was not a contributory factor. While recognising the small number of reports reviewed, and that the content of training may vary, this finding reflects the fact that such administrative controls are recognised to be a weak mitigation to safety risks (The National Institute for Occupational Safety and Health, n.d.) and that stronger systemic barriers are needed.

3.1.8 While a lack of professional curiosity by ED staff was frequently identified as a contributory factor in the investigation reports, this curiosity was found in the community professionals with specialist roles such as family nurse practitioners and young parent nurses. In one investigation, the ED discharge summary shared with the family’s GP was reviewed by a family nurse practitioner who contacted the hospital for more information and to ask whether consideration had been given to the child’s injury being non-accidental. Similarly, in another report, on seeing the ED discharge summary, the young parent nurse contacted the safeguarding team at the hospital to enquire about any ongoing care plan for the infant and whether there had been full consideration and examination with the possibility of a non-accidental injury in mind.

3.1.9 The investigation was mindful that these practitioners have greater knowledge of the families they are involved with than ED staff. Therefore they have insight into the likelihood or not of non-accidental injury based on their understanding of the particular risk factors that exist within a family.

3.2.1 A common finding from the investigation’s analysis of the 10 reports was that the families of infants who had non-accidental injuries were known to, and had needed intervention from, social services. Staff in the ED would not know this from information systems unless the child was looked after by the local authority or on a child protection plan (see 1.2.1).

3.2.2 The investigation reviewed the serious incident reports to identify the reason for social services involvement. In all cases where social services were involved, this was due to the mother being vulnerable; for example, experiencing domestic violence from their partner, having mental ill-health or being a teenage parent. These are known contextual factors which can increase the risk of non-accidental injuries (National Institute for Health and Care Excellence, 2017; The Child Safeguarding Practice Review Panel, 2022).

3.2.3 The investigation’s analysis of the serious incident reports identified that of the 10 infants attending an ED with non-accidental injuries, 7 of the families were known to social services. However, the threshold for inclusion on the CP-IS system (see 1.7.2) was only met in one instance.

3.2.4 Of the 10 reports, 3 stated that the CP-IS system had been accessed to check for social services involvement with the family. In each of these cases, there was a history of domestic violence which is a known risk factor for non-accidental injuries (National Institute for Health and Care Excellence, 2019). The infants were not on the CP-IS system as they were not in the care of the local authority and did not have a child protection plan in place. This meant that checking the system served to reassure staff rather than raise concern. Two of these three families were, however, known to social services.

3.2.5 In four of the reports there is no comment about whether the CP-IS system was checked. In three of these four reports the families were known to social services. However, as the infants were not on child protection plans, or in the care the of the local authority, they did not reach the threshold to be on the CP-IS system.

3.2.6 Two of the reports state that the CP-IS system was not checked as part of decision making. No context was provided as to the reasons for this. In one of these reports, the infant had a child protection plan in place, so was on the CP-IS system. In the other report the infant was not on the CP-IS system but, in common with other cases in this sample, the family was known to social services. Therefore, a common finding from the investigation’s analysis of the 10 reports was that the families of infants with non-accidental injuries were known to, and had needed intervention from, social services. Staff in the ED would not necessarily know that a family was known to or had received intervention from social services.

3.2.7 In addition to accessing the CP-IS system, healthcare staff can contact social services to ask whether a family is known to them and if there is any information which may assist their clinical decision making. Such enquiries can be time consuming and rely on getting through to social services (which commonly operates between 09:00 and 17:00 hours, Monday to Friday, with on-call provision outside of these hours). There can also be delay in a suitable person being able to provide the necessary information in a very short timeframe. These are significant barriers for time-pressured healthcare staff with competing resources.

3.3.1 In 6 of the 10 reports analysed, no contact was made with the paediatric and/or safeguarding teams when the infant was first brought to the ED. (In some cases these teams were involved when the infant was later brought back to ED.) This was found to be a contributory factor in non-accidental injuries not being identified. The rationale for this lack of contact was predominantly that the parents’ explanation of the injuries were accepted as true.

3.3.2 Importantly, analysis from the reports revealed that when the paediatric or safeguarding teams were involved there was consideration of non-accidental injury. In 7 of the 10 reports, this led to non-accidental injury being identified.

3.3.3 The reports stated that the lack of contact with paediatric or safeguarding teams was not in line with local policies, which stated that they should be contacted when infants (under 1 year old) attend with an injury, whether or not parents’ accounts of an injury were credible. Some reports included the context surrounding these policies not being adhered to. For example, one report stated that the local policy was not readily available to staff through the trust intranet (local computer network); another that the ED doctor was not clear about the procedure for infants attending with injuries; and another that staff were unaware that there was a policy covering the attendance of infants with injuries.

3.3.4 There is a national standard which dictates that:

‘ICYP [infants, children and young people under the age of 18] at high risk of potential safeguarding presentations are reviewed by a senior (ST4+) paediatrician or PEM [paediatric emergency medicine] doctor (e.g. infants who are non-mobile presenting with injuries such as bruising, burns or fractures).’
(Royal College of Paediatrics and Child Health, 2018)

The investigation reports indicated that the trusts’ local policies reflected this standard. The investigation acknowledges that there are numerous national guidance documents and standards relevant to ED practice and it is a challenge for clinicians to be familiar with them all.

3.3.5 Despite national and local policy directing the involvement of paediatric and safeguarding teams when infants with certain injuries are seen in the ED, the evidence from the reports was that escalation only happens when clinicians have a suspicion that non-accidental injury may be a possibility. This suggests there may be barriers to routinely involving these teams which would benefit from further exploration.

4.1.1 Emergency department (ED) staff and national clinical leads told the investigation that diagnosing non-accidental injury in infants and children was extremely challenging. Specifically, the high workload, time pressure, sensitivity and potential difficulty of raising the issue, along with the competing demands and dynamic nature of the working environment, created barriers to the diagnosis of non-accidental injury.

4.1.2 Illustrating the difficulty of making the diagnosis in this environment, one consultant said that in the ED clinicians are required, on average, to make a significant decision every 10 minutes, but one potential non-accidental injury demanded, in effect, the clinician’s “diary to be cleared for the rest of the day due to the resource it requires”.

4.1.3 Reflecting on these pressures, one ED consultant said that they had, at times, discharged children about whom they had a low-level non-accidental injury concern, resulting in some unease about the discharge. They said that these cases were ones where they would have made more enquiries if there had been fewer time pressures and demands.

4.1.4 In addition, a senior ED nurse told the investigation that making a non-accidental injury diagnosis required a clinician to “slow down”, in effect putting usual timescales on hold. They explained this was because of the time it took to perform relevant investigations, make the necessary enquiries and have the conversations to inform a decision about whether it was safe to discharge the child home or not. The nurse said that this process inevitably had a knock-on effect on resources within the unit and the clinician’s capacity to focus on other clinical decisions.

4.1.5 One trust that was visited as part of the wider investigation had responded to the time and resource challenges by allocating a nurse on shift to be the gatherer of information when non-accidental injury was suspected. This nurse would liaise with the relevant teams and agencies and collate all the details obtained as well as providing direct care and speaking with the attending adult(s). The trust highlighted that this approach prevented the loss of information that can occur when multiple staff are involved.

4.1.6 Medical staff told the investigation that experience was one of the most important factors in identifying non-accidental injuries and that there were often junior doctors working within the ED who did not have this experience. The staff said that the nursing team, particularly if they had worked in the department for some time, were an important resource for identifying potential non-accidental injuries.

4.1.7 ‘Lack of professional curiosity’ was a phrase commonly used in missed non-accidental injury serious incident reports and one familiar to stakeholders. As clinicians and stakeholders pointed out, it is a term that is person-centred and can imply blame so is not helpful in a learning and system-focused investigation.

4.1.8 Evidence from interviews with ED staff identified the following elements associated with this ‘lack of professional curiosity’:

• accepting at face value the explanation of an injury to avoid the discomfort of probing and a potentially challenging conversation
• lack of knowledge of developmental milestones resulting in an acceptance of an unsuitable explanation for injuries (for example, a child aged 5 weeks sustaining a bruise while rolling)
• a focus on treating a child’s the injury without consideration of the wider risks associated with the presentation.

4.1.9 National guidance states that:

‘… an unsuitable explanation is one that is considered to be implausible, inadequate, or inconsistent with the child or young person’s presentation, normal activities, existing medical condition, age or developmental stage.’
(National Institute for Health and Care Excellence, 2019)

A clinician’s judgement of whether an explanation of an injury is ‘unsuitable’ will be subjective. In addition, an assessment of whether an explanation is consistent with a child’s developmental age relies upon a knowledge of milestones which the assessing clinician may or may not have.

4.1.10 In addition, clinicians told the investigation that diagnosing non-accidental injury requires them to be sceptical of information provided by parents. They explained that this approach was in contrast to all other scenarios where the clinicians should be guided by the information provided by parents. Furthermore, as well as being sceptical, diagnosing non-accidental injury required clinicians to view parents as potential harmers rather than protectors of their children. This unfamiliar mindset was described as further compounding the challenge of diagnosis.

4.1.11 Regional protocols and guidance documents have been designed to help mitigate the difficulty of diagnosing non-accidental injury in this challenging context and facilitate standardised practice. An example of this is ‘The regional multi-agency protocol for the management of actual or suspected bruising or other injury in infants who are not independently mobile’ (Hampshire Safeguarding Children Partnership, 2020).

4.1.12 The investigation found that these regional protocols related only to specific types of injury – for example bruising in a non-mobile baby or head injury. In addition, there was variability in the recommended response to, and suggested level of concern about, injuries such as bruising.

4.1.13 To illustrate the differing approaches of the regional protocols, wording from two such documents is included below:

• ‘… all infants with bruising who are not independently mobile must be referred in to children’s services and for a paediatric opinion’ (Surrey Safeguarding Children Partnership, n.d)
• ‘unexplained [emphasis added] bruising (or bruising without an acceptable explanation) in a child not independently mobile must always raise suspicion of maltreatment and should result in an immediate referral to Children’s Social Care Services and an urgent paediatric opinion’ (Greater Manchester Safeguarding, n.d.).

4.1.14 Highlighting the effect of this variability, one clinician said that their trust was on the border of three local authorities, which meant the process for a child with bruising differed depending on where the child lived. The Child Safeguarding Practice Review Panel, in its panel briefing ‘Bruising in non-mobile infants’ (2022b), recommended that there should be greater consistency between regional protocols on this topic. The evidence from this investigation further supports the Panel’s recommendation.

4.1.15 National guidance also allows for variability because of subjective clinical judgement. It states:

‘Suspect child maltreatment if there is bruising or petechiae (tiny red or purple spots) that are not caused by a medical condition (for example, a causative coagulation [blood clotting] disorder) and if the explanation for the bruising is unsuitable. Examples include:

• ... bruising in a child who is not independently mobile. ...’
  (National Institute for Health and Care Excellence, 2019)

4.1.16 ED consultants told the investigation that it would be helpful to have an agreed standardised approach to diagnosing non-accidental injury across all EDs. This view was echoed by national stakeholders. They thought a national document setting out the actions to be taken for potential non-accidental injury would lessen the decision-making burden on frontline clinicians and would be especially helpful for more junior ED staff.

4.1.17 In 2006 the Royal College of Paediatrics and Child Health (RCPCH) published the first iteration of the ‘Child protection companion’. This document provides detailed practical guidance on the different types of abuse, how to recognise the signs of such abuse, and the expectations of the clinician in relation to police proceedings and evidence. The document is published on the RCPCH website and accessible to its members (Royal College of Paediatrics and Child Health, 2017).

4.1.18 The RCPCH also published ‘Facing the future: standards for children in emergency care settings’ (Royal College of Paediatrics and Child Health, 2018), which details the standards of care for children and young people in the ED. This was produced in conjunction with others including the Royal College of Emergency Medicine (RCEM). This guidance recognises that ‘Emergency Departments may be the first point at which children who have been subjected to abuse or neglect come into contact with professionals who are able to act for their protection’. The guidance makes clear that systems and processes should be in place to help protect children but does not specify the detail of these.

4.1.19 The investigation spoke with representatives from the RCEM about the potential benefit of ED-specific guidance on diagnosing non-accidental injury. They explained that the RCPCH has led the guidance in relation to non-accidental injury and said that ED doctors are guided by this and National Institute for Health and Care Excellence (NICE) recommendations. However, the relevant RCPCH guidance (the ‘Child protection companion’), which provides the practical details to support identification and diagnosis of non-accidental injury, is not published on the RCEM website; a subscription to the RCPCH is required to access it. It was also noted that the RCPCH guidance was very detailed and therefore long, as well as directed specifically at the role of paediatricians. The investigation spoke to members of the wider Paediatric Emergency Medicine Professional Advisory Group who felt that specific guidance produced by the RCEM would help ED clinicians by “demystifying” this area of practice and standardising the approach to such presentations.

4.2.1 The ‘Child protection companion’ published by the RCPCH (2017) states that injuries should be ‘assessed in the context of the medical and social history, developmental stage, explanation given, full examination and relevant investigations’. In addition, national guidance about non-accidental injury (including National Institute for Health and Care Excellence, 2017) highlights the significance of social service involvement with families, a point echoed by staff interviewed. Given that clinicians in the ED will not know the child’s family circumstances and whether there is any social service involvement the information must be sought from other sources.

4.2.2 The adult who brings the infant to the ED will be the primary source of information about the injury and the child’s medical and social history. Staff explained that as part of the information gathering at triage, the adult will be asked if there is any social services involvement with the family.

4.2.3 Evidence from the serious incident reports analysed by the investigation identified that the reliability of the response given to this question was variable. Similarly, staff interviewed said that some parents were forthcoming about social services involvement, but others were not. Staff said that they corroborated the answer they were given through other sources – such as contacting social services or accessing the Child Protection – Information Sharing System (CP-IS).

4.2.4 The CP-IS system (see section 1.7) provides limited information. An infant or child may have multiple risk factors for non-accidental injury, and social services may be concerned about them. However, if they have not reached the threshold for a child protection plan, or they are not looked after by the local authority, the child will not appear on the system.

4.2.5 Analysis of the serious incident reports by the investigation evidenced the significance of social services involvement. In 7 of the 10 reports, the family had social services input although the level of concern was not sufficient to have generated a child protection plan.

4.2.6 Staff interviewed said that there was a risk of taking false reassurance that an infant or child was not at risk if their name was not on the CP-IS system.

4.2.7 In addition to the limitations in information, the investigation was told that the system can sometimes be out of date. For example, one staff member said that when there is an unborn child protection plan in place, it can take time for this to be associated with the child once born. This means that an infant may not show as being on a child protection plan even though they are.

4.2.8 There was recognition that the CP-IS system is “one part of the jigsaw”, and that clinicians needed “as many bits of information as possible” to be sufficiently informed to make a diagnosis of non-accidental injury. Available information should be considered alongside the clinical signs and symptoms.

4.2.9 Staff told the investigation that conversations with social workers, together with any information they have about an infant and their family, were extremely helpful and influential in informing the diagnosis of non-accidental injury. ED consultants interviewed told the investigation that there were similarities in social worker and ED roles. Both required fine judgements about degrees of risk and decisions that had potentially significant consequences in relation to people’s welfare.

4.2.10 EDs do not have access to information held by social services. The information is usually obtained by telephoning the relevant local social services team. The investigation found variability in the time taken (both in and out of office hours) to speak to a social worker to discuss an infant and request information, and in the time taken to receive the requested information.

4.2.11 The investigation found that in hours (09:00 to 17:00 hours, Monday to Friday) the ED clinician would contact the social services team for the area the infant lived in. Details of the infant would be provided, and if known to the team the clinician would be told to await contact and information from the allocated social worker for that infant.

4.2.12 The three EDs visited said that it could take several hours to receive the relevant information back. This created problems as it meant either keeping the infant in the ED for longer than was clinically required, admitting the infant to a ward, or discharging the infant despite the risk. These options were described as creating different problems. Keeping the infant in the ED necessitated ongoing, difficult conversations with the attending adult(s). Admitting the infant to a ward also meant difficult conversations and required a bed being available on a suitable ward and potentially occupying a bed unnecessarily. Discharging the infant before receiving a response from social services meant taking a risk that may ultimately result in further non-accidental injury. ED clinicians commented that ideally systems would be in place to allow easy access to relevant information and conversations which would be of benefit to both services in protecting the infant’s welfare.

4.2.13 The investigation was told that out of hours there would be an on-call social worker to telephone. They would take the infant’s details and the ED clinician would then await a call back from the on-call social worker after they had accessed the necessary information systems. ED clinicians said that this call back took variable amounts of time depending upon the competing demands on the social worker. As with the situation in hours, this meant taking one of the options described above.

4.2.14 The safeguarding teams at two trusts visited as part of the investigation had been granted access to a limited amount of information on the local social services information system. The system did not cover the trusts’ total catchment area, but the teams said it had still proved very useful for ED and other staff as it enabled them to quickly provide information to support safeguarding decisions. However, as access to these systems had only been granted to the safeguarding teams, the information was not available outside of the teams’ typical Monday to Friday office hours.

4.2.15 In addition to the time it took to get information and (in hours) have a conversation with a social worker, staff told the investigation that there were differing degrees of openness to information sharing depending on the individual social worker and particular social services team. Some teams were described as “very open”, providing a clear picture of the family circumstances, while others were described as “very reluctant” to give information. The perception of ED staff was that this reluctance was due to concerns about breaching confidentiality and information governance. For example, one consultant described a scenario where they had contacted the social worker, who confirmed that the family was known to them but could not provide any details due to data protection. The clinician explained to the investigation that the misunderstanding about confidentiality and data protection in safeguarding situations was very unhelpful and resulted in clinicians having to make decisions about whether an infant could be safely discharged without the information that was available to help them in that decision.

4.2.16 This clinician’s comments about the importance of sharing relevant information in the interests of safeguarding are reflected in serious case reviews where this has been found to have resulted in harm. The importance of protecting children is acknowledged in national guidance on safeguarding which states ‘Fears about sharing information must not be allowed to stand in the way of the need to promote the welfare and protect the safety of children’ (Department for Education, 2018b).

4.2.17 The investigation spoke with the Data and Analytics team for System Transformation at NHS England about ED access to information held by safeguarding partners such as social services. The data team explained how information held by different organisations and services could be made accessible through “collaboration frameworks” – that is, interfaces which allow controlled access to relevant information. One of the projects currently in progress is focused on building a collaborative framework between the information systems used by GPs and EDs to mutually support clinical decision making.

4.2.18 Information about patients attending EDs and the treatment they receive is entered into a national Emergency Care Data Set (ECDS) using standardised descriptors with associated codes.

4.2.19 Within the ECDS, there is also the ability to enter information relating to safeguarding. The investigation spoke with the ECDS team at NHS England. They explained that the safeguarding element of the ECDS had been developed in recognition of the fact that one of the “key features of high-profile cases where non-accidental injury had been missed was due to multiple providers being involved and the reluctance of clinicians to make a formal diagnosis of non-accidental injury”. To mitigate this risk, the safeguarding data includes the ability to record ‘concern’, as well as when a child is considered to be ‘at risk’ of abuse or where abuse is ‘suspected’. The team said that this was to allow clinicians to highlight those cases where there was a “low-level” concern that was not sufficient to warrant intervention.

4.2.20 The ECDS is collected by NHS Digital (now part of NHS England), which is responsible for anonymising the data, and sending the information contained within it to the relevant teams or bodies responsible for acting on the data. For example, information relating to clinical care and treatment and the associated timeframes for this is sent to the team within NHS England looking at performance times within EDs. However, the investigation was informed by NHS Digital that the safeguarding data contained within ECDS is not currently sent to any team or body for review and action. The ECDS team said this was a missed opportunity. They noted that there was over 5 years of data relating to safeguarding which has the potential to identify trends and inform care in the same way as other data collected by the NHS.

4.2.21 As an example of how such data can be used, the ECDS team referred to ECDS data that was shared with community safety partnerships (including police and local authorities) to help tackle violence. This data includes details of injuries resulting from violence, the time the patient was seen and the location where the injury occurred. This information is used to help inform interventions such as restricting alcohol licencing and targeted policing in particular areas.

4.2.22 The investigation spoke with the Deputy Director for Safeguarding at NHS England. They said there were limitations in the ECDS data captured; for example, there is no information about family circumstances, and the field as it currently stands would not allow clinicians to categorise the concern sufficiently for the data output to be useful. The ECDS team said that most fields within the ECDS had been gradually revised over time and this had not taken place for the safeguarding field. Both the Deputy Director for Safeguarding and ECDS team said that a group made up of interested persons could review the field to ensure that they worked well from all perspectives.

4.2.23 The Data and Analytics team for System Transformation at NHS England said that as the safeguarding data is not currently used, it is not part of any data quality or validation process so may well be poor and not necessarily representative. They noted that these limitations would likely impact on the initial usefulness of this data if it were to be used. This point was echoed by the Deputy Director for Safeguarding who said that there is an issue of completeness of data input affecting quality.

4.2.24 The ECDS team at NHS England agreed that data quality was an issue and that if the safeguarding data was incorporated into the top tier data quality matrix indicators (like fields such as chief complaint, acuity and diagnosis) and into the data quality dashboard then this would mean trusts had to report on it and would therefore improve the data quality. They also said that clarity was needed about who had overall ownership and responsibility for providing clinical and information governance and assurance. In addition, there needs to be agreement about who the data should be sent to for actioning. Given their role, integrated care boards may well be central in determining local actions to be taken based on the data.

4.2.25 GPs, and other professionals such as health visitors and midwives working at a GP practice, are an important source of information about infants and their family circumstances. The investigation spoke with a representative from the Royal College of General Practitioners (RCGP) who said that GPs hold lots of information about “family context”.

4.2.26 Frequent visits to EDs are a recognised risk factor for non-accidental injury (National Institute for Health and Care Excellence, 2019). Reflecting this, ‘Facing the Future: standards for children in emergency care settings’ (Royal College of Paediatrics and Child Health, 2018) recommends that ‘Systems are in place to identify children and young people who attend frequently’. Currently, unless an infant is on the CP-IS system, there is not a system in place to identify infants who attend a number of different hospital trusts.

4.2.27 GPs receive information whenever one of their patients (adult or child) visits an ED or other urgent care setting. Information about an infant’s attendance at an ED will be on that hospital trust’s patient administration system. However, the attendance will not appear on the patent administration system at any other hospital trust as such systems are individual to the trust and functionally independent from those of other trusts. Thus, if an infant attends the ED at four different hospital trusts, the ED clinicians at each will only be aware of the one attendance they are involved with, unless informed otherwise by the adult(s) present. The exception to this is infants on the CP-IS system. In this situation, clinicians can see up to 25 previous visits to any ED or other unplanned care setting. Thus, a GP may be the only health professional in possession of information on infants who have attended multiple different healthcare settings but are not on the CP-IS because they are not on a child protection plan or a looked-after child.

4.2.28 Staff commented on this weakness in current systems and the risk it posed. They said that when non-accidental injury was being considered as a diagnosis, staff mitigated the risk by telephoning neighbouring EDs to see if the infant had attended. If so, they would ask for details of the injuries and explanations provided. As one ED nurse highlighted, this mitigation is not a reliable system as only obvious neighbouring EDs were contacted, and a safer system would be one where information relevant to the diagnosis of non-accidental injury was accessible to ED staff.

4.2.29 However, evidence from interviews identified that ED clinicians do not seek information from GP surgeries. While aware that the data held within GP surgeries may be a rich source of information to inform a diagnosis of non-accidental injury, ED clinicians highlighted the difficulty of getting through to GP surgeries and, once through, of speaking with a GP. In addition, they highlighted the issue of working hours and the fact that infants with a possible non-accidental injury may well be brought to the ED outside of GP surgery opening hours. It appeared that these challenges meant ED clinicians did not attempt to contact GPs as a source of information. The RCGP representative said that in their experience as a GP, contact from an ED requesting information in relation to a safeguarding concern was rare. Thus, an important source of information about an infant and their surrounding situation was lost to ED clinicians to support a diagnosis of non-accidental injury.

4.2.30 Some information held by GPs is available to ED clinicians in the form of the summary care record. The summary care record system was created to reduce the risk of prescribing errors by enabling access to current medications, allergies and bad reactions to medications, and basic details of the patient for identification purposes (name, date of birth, address and NHS number). It can be viewed through clinical information systems or through a specific application and can only be accessed by those providing direct care and who have the appropriate level of access.

4.2.31 The summary care record was developed to allow for ‘additional information’ such as significant medical history or long-term conditions to be included with the patient’s explicit consent. In response to the COVID-19 pandemic the NHS included ‘additional information’ by default on a patient’s summary care record, including COVID-19 related information. Since then, a decision has been made to continue the inclusion of additional information in the summary care record.

4.2.32 The investigation spoke with the Live Services Team at NHS Digital, who have summary care records within their portfolio. They explained that the ‘spine’ information system which currently allows access to the summary care record is soon to be replaced by the National Care Records Service (NCRS), which is currently being piloted. The NCRS will still enable access to patients’ summary care records, but also allows the viewing of health care plan records. They added that the NCRS system is more powerful than its predecessor and potentially could support further safeguarding information nationally. For example, the national record locator, which is part of the NCRS, is designed to hold patients’ mental health crisis plans, which can then be viewed by out-of-hours services and the ambulance service.

4.2.33 A representative from the Live Services Team explained that the summary care record is populated from GP records using an inclusion dataset of over 3,000 items. Within these items there already exist some which may highlight potential safeguarding concerns, such as ‘social worker involved’.

4.2.34 However, they explained that they purposefully restrict any codes directly related to child safeguarding from the summary care record inclusion dataset. This is to avoid clinicians using the summary care record instead of the CP-IS system to look for safeguarding information. The reason given was that if a child was on the CP-IS system, and the CP-IS information was not accessed because the information was seen in the summary care record, then there would be no trigger to notify the local authority of the contact. Whereas, if the CP-IS was accessed, the local authority would be made aware of the contact.

4.2.35 The robustness of data items relies upon the coding used in the GP system from which the data is gathered. Such data may include information which comes from the ECDS and is included in a discharge summary sent to the GP following a patient being seen in the ED.

4.2.36 The representative from the Royal College of General Practitioners (RCGP) said that discharge information from EDs can be lacking in detail and may not mention whether safeguarding has been considered when a child has attended with an injury. This means that any ED concerns about a potential non-accidental injury may not transfer into a child’s primary care record.

4.2.37 The RCGP representative explained that there was also a lack of consistency in how discharge summaries from EDs were managed in GP practices. In some practices the summaries were reviewed by a doctor, in some they were reviewed by an administrator and pertinent information transferred into the patient’s record, and in others they were simply filed if there were no actions for the GP. What was done with these summaries depended on the individual practice and how much information was detailed in the summaries, so there were challenges present from both ED and GPs.

4.2.38 The investigation spoke with the Deputy Director for Safeguarding at NHS England who said that the recording of safeguarding information was a complex area with potential legal implications.

4.2.39 However, the Deputy Director could see the benefit of greater access to the safeguarding information and suggested that data from all healthcare providers was important in this context, including maternity information systems and data from health visitors.

4.2.40 The investigation acknowledges that the use of the summary care record to provide access to all safeguarding information held by healthcare providers is just one option and there may be other information systems which could be explored to facilitate such access.

4.2.41 Additionally, some local information systems may help to inform clinicians about potential safeguarding concerns. For example, the Live Services Team at NHS Digital said that shared care records, previously known as local data sharing records, may be used to identify multiple attendances at urgent care settings in a region and could therefore assist in identifying known safeguarding risks as mentioned in 4.2.27.

4.3.1 The investigation spoke at length to clinicians as part of the wider investigation about the safeguarding support available. They spoke of the systems that supported safeguarding, the associated challenges, and the adaptations being used to provide such support. These are explored below.

4.3.2 All NHS staff have a safeguarding responsibility, both from a statutory and contractual perspective. As mentioned in section 1.5, there are different levels of training required for staff depending upon their contact with patients and responsibilities. All trusts are required to have a named doctor and named nurse, and a named midwife if providing maternity services. The different levels of training were developed because different roles were seen to require different levels of safeguarding knowledge. This means that while all staff will have some safeguarding knowledge, there are named individuals who are required to have a greater understanding of safeguarding, and staff will call upon those individuals when they need advice.

4.3.3 The majority of staff spoken with as part of the wider investigation said that safeguarding teams were a very useful resource, providing advice and guidance on this sensitive area. Some staff said that safeguarding teams may also lead difficult conversations with parents to explain the process when there are safeguarding concerns. The investigation was told that part of the importance of this team was them having the power and authority to provide challenge where appropriate, as well as an independent view on the presentation.

4.3.4 Safeguarding teams in trusts typically work Monday to Friday during office hours, with no on-call arrangements available outside of these hours. The investigation heard from staff members, across different organisations, that potential non-accidental injuries in infants would often be seen outside of usual office hours. This was reflected in the serious incident reports reviewed; five of the ten reports were about infants seen in ED outside of Monday to Friday office hours.

4.3.5 Staff described managing potential non-accidental injury or other safeguarding concerns outside of these hours as “tricky” as the support provided by the safeguarding team was not available.

4.3.6 The investigation visited one trust which had set up an on-call safeguarding system so that clinicians could seek advice day and night.

4.3.7 Staff said the benefit of this system was that in a scenario where they were “not quite sure what to do” they “could pick up the phone any time of the day or night and speak to someone with safeguarding experience”.

4.3.8 It was explained that this enabled issues to be dealt with in “real time”, avoiding the need to recall someone to the hospital the following morning who had been discharged. This system worked because the trust was formed of multiple hospitals which meant it was possible to have a number of advisors on the rota providing this support.

4.3.9 Staff explained that this service could similarly be shared across hospitals in a locality through a service level agreement. They said there was scope to develop such provisions though integrated care systems or by linking up with community services which may be able to assist with service provision.

4.3.10 The investigation spoke with the Chief Social Worker for Children and Families. They said that it was time to rethink how local safeguarding partners respond to child protection enquiries more generally and particularly outside of Monday to Friday office hours. They said that while there is always an emergency duty response available, this was very often limited. In cases of suspected non-accidental injury, this sometimes left families stuck on hospital wards, with hospital staff having to manage very tense situations, overnight and even over whole weekends. The Chief Social Worker said this raised important questions about the ethics of this practice, with families receiving limited information about their rights and the consequences for them if they did not comply.

4.3.11 The independent review of children’s social care and the National Child Safeguarding Practice Review into the murders of Arthur Labinjo-Hughes and Star Hobson, both published in May 2022, (The Child Safeguarding Practice Review Panel, 2022a; MacAlister, 2022) highlighted the need to strengthen the multiagency arrangements currently in place to safeguard children. In Response to these reports the Department for Education has published an implementation strategy and consultation on children’s social care reforms (Department for Education, 2023).

4.3.12 The implementation strategy has a focus on child protection and the development of a national children’s social care framework. This includes the development of a new child protection response from a dedicated group of multi-agency professionals from the police, health and local authorities. These practitioners would work side by side as an integrated team and lead on specific child protection functions including facilitating ‘timely decision-making’.

4.3.13 The operating model for this response will be tested though ‘pathfinders’ in a limited number of areas to establish how the changes can be implemented most effectively. As part of this, the pathfinders will ‘explore how other professionals from across agencies can access timely, expert child protection advice through this model’.

4.3.14 There is a clear overlap between the need highlighted in this report for expert safeguarding advice out of hours and the ambitions outlined in the Department for Education’s recently published implementation strategy. However, the strategy does not detail the working hours of the expert multi-professional team nor the timeframe within which advice will be provided for those seeking it. The evidence from this investigation is that advice needs to be in real time or within a very short time frame (a matter of hours) to be of most value to ED clinicians. This is because, as the Chief Social Worker highlighted, for a potential diagnosis of non-accidental injury, the infant may be kept in hospital (with potentially challenging conversations ongoing with parents) while waiting for safeguarding advice.

4.3.15 It was noted that the majority of safeguarding teams worked in offices based outside of treatment areas.

4.3.16 Clinical staff told the investigation that they felt it would be “more helpful to have the safeguarding team close by”. They said being able to “put their head around the door” and ask a question would feel less formal than calling a team located elsewhere in the hospital.

4.3.17 This view was mirrored by the safeguarding teams interviewed who said being physically present created a familiarity which would enable easier conversations.

4.3.18 Clinical staff and safeguarding teams acknowledged the limited space in clinical areas which prohibited non-clinical staff being located there. Some safeguarding teams had mitigated this barrier by attending the ED daily to have a presence within the department. This also enabled staff to understand the challenges within the ED. Some safeguarding teams said that having a presence within the ED made their ability to appropriately challenge decision making easier.

4.3.19 During stakeholder consultation on the draft report, feedback was received echoing the value of safeguarding teams being located physically near to the ED and the importance of this being highlighted to trusts.

4.4.1 During discussions with staff at various trusts it became clear that some trusts had implemented actions locally designed to embed child safeguarding into the culture of EDs. Some of these initiatives are detailed below as trusts may find this information useful for their own local arrangements.

4.4.2 One example of integration of safeguarding into an ED’s culture was the regular inclusion of a safeguarding element within simulation training. It was explained that this aimed to make clinicians think not just about the medical situation they were faced with but also the wider social picture. Studies on the use of simulation training for safeguarding scenarios have found that it increased the confidence of staff to deal with such issues (Hall et al, 2015).

4.4.3 The investigation acknowledges that guidance and training are not strong interventions to effect safety improvements, as they are people-focused rather than system-focused. However, training can add value by helping to embed safeguarding into the culture of a trust or team.

4.4.4 One trust visited had developed a psychosocial multidisciplinary team (MDT) meeting which was attended by various safeguarding stakeholders, including the police, violence reduction team and the probation service, as well as members of the local mental health team, ED clinicians and paediatricians. In the meeting they discussed children (under 18 years) who had attended ED and about whom they had safeguarding concerns. Prior to the meeting they went through all of the attendances which met the criteria for discussion and removed those which required a social services or health visitor notification, for example where a stairgate was not secured which had led to an injury but where there were not significant safeguarding concerns. These are actioned by the safeguarding team. The MDT then discussed children who have attended about whom they may have emerging concerns, as well as cases where a child was not the patient, but their parent had come to the ED because of, for example a mental health issue, drugs or alcohol, or an injury sustained through violence. The MDT was therefore acting as a proactive step in identifying children who may be at risk of harm.

4.4.5 It was explained that there had a been a lot of work involved in developing this aspect (where an adult patient’s circumstances may have implications for a child’s welfare), including having those working in the adult ED routinely asking patients about their family circumstances, particularly those whose reasons for attendance were violence, drug or alcohol related. However, staff told the investigation that such questions were important for all admissions. They cited the example where an elderly woman may visit the ED following a fall and, having been asked about her family circumstances, reveals that they are the primary carer for a young teenage granddaughter whose mother left some years before. There may be no formal arrangement in place and no social worker involved, but by asking about the family circumstances staff may identify children in the household who may be at risk. As a way of highlighting the importance of this consideration, adult ED nursing staff would rotate into the paediatric ED.

4.4.6 A member of the RCEM Paediatric Emergency Medicine Professional Advisory Group told the investigation that not identifying vulnerable children whose parent(s) have attended for reasons relating to mental health, violence, drugs or alcohol was a significant risk and one which is of particular concern to the Group.

4.4.7 As stated in 4.2.16, the sharing of information between organisations has long been a recognised issue in relation to safeguarding. Recent national reports, such as the national review into the murders of Arthur Labinjo-Hughes and Star Hobson (The Child Safeguarding Practice Review Panel, 2022a) indicate that information-sharing issues remain. The investigation discussed this with clinicians and national leads who said that apprehension about sharing information outside of an organisation still exists. One clinician noted that when they had attended induction training at a trust there had been a 30-minute session during which adult and child safeguarding was discussed, as well as when information can be shared. This was followed by an hour-long session emphasising the importance of information governance and the risk of sharing information.

• There is no specific guidance for ED clinicians on the identification of suspected non-accidental injuries and what to do if they suspect an infant has a non-accidental injury.
• There may be barriers to routinely escalating cases of children with a potential non-accidental injury to paediatric (child specialist) and safeguarding teams.
• Delays in the availability of information about potential safeguarding concerns add to the pressures on ED staff when making decisions about infants with potential non-accidental injuries.
• There remain concerns about, and an inconsistent approach to, sharing safeguarding information between organisations.
• The Emergency Care Data Set (ECDS) gathers information about ED attendances and includes a field for when such attendances are related to safeguarding.
• The ECDS safeguarding information collected is not currently utilised within the NHS and there is minimal quality assurance in place to ensure that it is reliable.
• Risk factors for non-accidental injuries which do not meet the criteria to be included on the Child Protection – Information Sharing system (the electronic system designed for information sharing between the NHS and social services) are not included in a patient’s summary care record and may therefore remain unknown to clinicians.
• The investigation identified mechanisms which could enable safeguarding information that is not currently available to ED clinicians, to be made available through existing national and regional digital systems.
• Safeguarding teams are often located physically distant from EDs. This can create a barrier to communication and liaison with the team.

During discussions with staff at various trusts it became clear that actions designed to embed child safeguarding into the culture of the emergency departments had been implemented locally:

• Induction training: benefits were seen where relevant staff had undergone a training session during which adult and child safeguarding was discussed, with a focus on how and when sharing of information between organisations is appropriate for safeguarding.
• Simulation training: the regular inclusion of a safeguarding element within simulation training has been seen to increase the confidence of staff to deal with safeguarding issues.
• Psycho-social multidisciplinary team meetings: A meeting to discuss children (under 18 years) who had attended ED about whom they had safeguarding concerns that was attended by various safeguarding stakeholders, including the police, violence reduction, the probation service as well as members of the local mental health team, ED clinicians and paediatricians. The MDT acted as a proactive step in sharing relevant multiagency information to identify children who may become at risk of harm.

The HSIB Chief Investigator authorised a national investigation based on an assessment of the risk against HSIB’s patient safety risk criteria:

Outcome impact – what was, or is, the impact of the safety issue on people and services across the healthcare system?

As well as the potential for serious injury or even death, non-accidental injury can have long term effects on a child’s health. A link has been suggested between ‘non-sexual child maltreatment’ and drug abuse, suicide attempts and a range of mental disorders (Norman et al, 2012). Non-accidental injury is thought to have a negative effect on a child’s developing brain, which can lead to impaired cognitive (mental) function (National Scientific Council on the Developing Child, 2014).

In addition to the harm caused to the child, it is likely that staff involved in any such cases could suffer emotionally and could be considered ‘second victims’ (Second Victim Support, n.d.).

Systemic risk – how widespread and how common a safety issue is this across the healthcare system?

The investigation could not find figures for attendances at emergency departments (EDs) of infants under 1 year of age with violence-related injuries. However, the Violence Research Group reported that in 2018 there were 1,781 attendances of children aged between 0 and 10 years for such injuries (Sivarajasingam et al, 2019). The investigation acknowledges that this category will encompass many injuries unrelated to abuse and that the subset relevant to this investigation will be much smaller.

In 2020 the Office for National Statistics published a report of collated data on child abuse which showed that 4,170 children in England were the subject of a child protection plan due to experience or risk of physical abuse (Office for National Statistics, 2020b).

Of the 91 incidents of missed or delayed opportunities to recognise non-accidental injury and other forms of abuse or neglect, 63 involved infants aged 1-year-old or younger. Of these, 33 (52%) of the missed or delayed opportunities happened in an emergency department (ED) or other acute setting (e.g. a children’s ward).

Due to a number of limitations, it is likely there will be more instances of missed diagnosis of non-accidental injury than reported here. Even when instances are reported, there are multiple ways to record these incidents meaning it is hard to understand the full scale of this issue. For example, other StEIS categories, such as ‘Diagnostic incident’ are also known to contain relevant incidents. However, it was expected that the category ‘Abuse/alleged abuse of a child patient by a third party’ would be used most often in these circumstances and so this was the focus of the StEIS search.

Learning potential – what is the potential for an HSIB investigation to lead to positive changes and improvements to patient safety across the healthcare system?

Missed non-accidental injury in infants in the ED is a well-recognised but persistent safety risk. This suggests there are complexities associated with diagnosing such injuries that need to be understood and acknowledged.

An HSIB national safety investigation can provide insight into persistent safety risks and make recommendations that stimulate change. In addition, they provide an opportunity to share learning from stakeholders and/or healthcare providers who have made beneficial improvements to positively influence processes and practices across organisations.

Thematic analysis of serious incidents reports: incidents were identified through StEIS and the relevant trusts were contacted and completed reports requested.

Other sources of evidence gathered by the investigation:

• analysis of 10 serious incident investigation reports from different trusts across England
• review of literature relevant to the safety risk – specifically, literature regarding the responsibilities of healthcare staff and organisations, training in child safeguarding, risk factors for non-accidental injury in infants and information sharing
• observational visits to three acute trusts
• semi-structured interviews with ED clinicians (foundation trainees to consultants)
• semi-structured interviews with ED nursing staff
• semi-structured interviews with trust safeguarding teams
• semi-structured interviews with paediatricians
• focus group of ED clinicians.

The investigation undertook a thematic analysis of 10 serious incident reports to identify factors that may contribute to non-accidental injuries in infants being missed in the ED. The information within the reports was taken as it appeared, with no challenge to the detail of the incidents. The reports were jointly reviewed by two investigators and inductive coding (where codes are derived from the data rather than the data being categorised into predetermined codes) was undertaken to give an objective, unbiased consideration of the data contained. When aggregated, this coding revealed contributory factors to non-accidental injury not being suspected and themes were developed from these.

The investigation then engaged with stakeholders. This enabled sense-checking of the themes which had been developed to ensure that they were meaningful, and provided insights which helped to contextualise some of the themes. These themes informed the terms of reference for the investigation.

The investigation used the Systems Engineering Initiative for Patient Safety (SEIPS) model of incident analysis (Carayon et al, 2016) when exploring the context of decision making for the wider national picture. This tool was used as a guide during site visits for evidence collection and in the analysis of the data gathered. SEIPS provides a human factors framework for understanding the work system (that is, the external environment, organisation, internal environment, tools and technology, tasks, and persons), work processes (including physical, cognitive and social/behavioural aspects) and the relationship between these and the resulting outcomes in healthcare.

The investigation findings were shared with the stakeholders and subject matter advisors. This enabled checking for factual accuracy and overall sense-checking. The stakeholders contributed to the development of the safety recommendations based on the evidence gathered.