Each year around 3.5 million children pass through emergency departments and other ambulatory care settings such as walk in centres.1 Some of these children may present with non-accidental injuries, or with non-intentional injuries or illnesses that have nevertheless occurred within an abusive or neglectful context for the child. Emergency departments may be the first point at which such children come into contact with professionals who are able to act for their protection. Child maltreatment encompasses a spectrum both in terms of types and severity of abuse experienced. It is recognised, however, that many children who go on to experience more serious forms of abuse, including the small numbers experiencing fatal maltreatment, will have had previous contact with health services.2 The government document, Working Together,3 4 places a responsibility on health professionals to identify and respond appropriately to suspected abuse and neglect. Distinguishing those children whose circumstances may require further investigation from the majority with genuine accidental injuries can be a difficult task that is compounded further by the stresses and constraints inherent in a busy department striving to meet national targets for care.

Although there is a large body of research on risk factors for child abuse5 and on primary prediction and prevention,6 7 there is less published research on early indicators of abuse8 or procedures for detecting and responding to abuse in emergency care settings. Some previous attempts have been made to combine these indicators into screening tools that can be used in accident and emergency (A&E) departments to improve the recognition of child abuse.9^–12 Some audits have shown that clear protocols with a simple flowchart, backed up by training and liaison, can lead to increased awareness, consideration and documentation of intentional injury9 10 12 and improved effective social service referral.13 However, other audits have failed to show any significant improvement.14 A recent national survey of A&E departments showed that over 90% of departments do have written protocols for child protection.15 However, our perception, based on discussions with colleagues in child health and emergency medicine, is that there remains a great deal of uncertainty around their use and the guidelines in use are not based on empirical evidence.

The aims of the research were to document the range of information available to guide current practice in the management of child protection in emergency departments; to identify models of good practice; and to suggest ways in which departments can improve and clarify guidance given to staff.

METHODS

Questionnaires (appendix 1) were sent out addressed to the lead nurse/consultant in a sample of 81 emergency departments in England, selected randomly from a list provided by the British Association of Emergency Medicine, to provide a 20% sample of all departments. Questionnaires were also sent to all 20 emergency departments in Northern Irelandⁱ. Respondents were asked to return an anonymous completed copy of the questionnaire, along with a copy of their protocol if available. The returned protocols were scrutinised by all three authors using a structured proforma (appendix 2) looking for emerging themes, consistent threads, examples of good practice and any discrepancies within the key areas outlined below. Nine randomly selected protocols were read by a second reviewer to check for internal consistency; these indicated that the three reviewers were identifying the same themes and conclusions from their reviews. All three researchers compared notes to highlight the key features in the protocols along with both good and bad points arising from them. From these reviews, a series of recommendations were developed by consensus among the authors.

RESULTS

A total of 74 questionnaires were returned, 17 from Northern Ireland and 57 from England. Forty-five questionnaires were from general emergency departments, six from children’s emergency departments and 20 from minor injuries units. Three questionnaires were incomplete and have been excluded from the analysis.

Procedures

Sixty-three of 71 departments (88.7%) had written protocols for child protection, although only 27 of these supplied a copy, some of which consisted simply of pages copied from the Area Child Protection Committee (ACPC) procedures. The protocols returned showed a large variation in size (from 1 to 93 pages), format and content, and this significantly affected the ease of accessing relevant information. The most helpful were those that focused more on direct management, with only small amounts of background information, presented in short, numbered paragraphs and with an index, subheadings or other system of pointers. Seven protocols incorporated a single page flowchart, most of which were helpful, although some were too complicated to be of practical use. The majority of protocols were intended for both medical and nursing staff and this was considered to be helpful. Very few included quality measures such as date, author or sources of information.

Roles and responsibilities

Few protocols clearly outlined the roles and responsibilities of different staff groups. A common theme in many was the hierarchical approach with nurses being expected to refer their concerns to doctors, junior doctors to consultants, and emergency staff to paediatricians. Very few emphasised that each individual has a responsibility for the welfare of children. Several protocols included contact details, most commonly for the named or designated professionals; paediatricians or community paediatricians; the child protection register and social work departments. Forty-nine of 71 departments (69%) were able to identify lead professionals for child protection, of which 12 listed a lead doctor, 23 a lead nurse, and 14 both. Fifty-five departments (77.5%) reported that there were opportunities for regular liaison on child protection matters, but these varied in both frequency and in the personnel involved. Forty-one departments provided details of frequency, 12 having at least weekly liaison, 24 having liaison less than once a month or only on an ad hoc basis, and the remainder between weekly and monthly. The most common professional to provide liaison was a nurse practitioner or specialist nurse (13/51, 25.5%).

Identification of children at risk

Overall there was a lack of clarity about the purpose and appropriate use of indicators of concern. Twenty-five of 71 (35.2%) departments reported that they used a checklist to highlight concerns. Only nine respondents reported the factors they would use to highlight concerns (table 1). The median number of items listed was 7 (range 5–14). A large number of the protocols returned included long lists of signs and symptoms of abuse (up to 68 separate factors in one protocol), with no guidance on how to use these. Those felt to be of more practical value included a short (up to 8 items) checklist, with clear guidelines on what to do if concerns were identified. Some indicators were very specific (for example, “children under 1 year with a fracture, burn or scald”), while others were more descriptive (for example, “the parents’ behaviour gives rise to concern”). One included recognised risk factors (for example, teenage parents, low income, or prematurity) as well as signs and symptoms, but gave no indication on how these were to be interpreted. The majority of departments (47/71, 66.2%) checked the child protection register if there were suspicions for a particular child, and 38 specified reasons for doing so (table 1). Sixteen departments (22.5%) checked the register for all children and eight (11.3%) did not check it at all. Twenty-five departments checked a database or list kept in the department, while 30 used a telephone check, and nine a combination of these. In the protocols supplied, there was often some confusion about how and when to check the child protection register, or on what to do following a positive result.

Referral process

This was the strongest element of most protocols, though surprisingly totally lacking from four and unclear in a further five. Those protocols supplied by walk in centres or minor injuries units limited their referral pathway to sending the child to the main emergency department. A few emphasised the importance of backing up referrals in writing, but few included any subsequent check of management. Some protocols incorporated safeguards—for example, action to take if a child is taken from the department; informing the locality manager or child protection nurse of all attendances; or information on contacting the social worker or police urgently if the child is perceived to be at immediate risk. A number of protocols specified different pathways according to the level of concern (for example, child in need/possible child abuse/definite child abuse) either in the body of the protocol or in a flowchart.

Subsequent management

The most common aspects of subsequent management incorporated in the protocols were guidelines on documentation, details on informing primary care, and action around discharge. However, most protocols provided very little detail beyond initial recognition and referral. Very few included any mention of dealing with the child’s presenting complaint, or assessing their medical needs first. One protocol inappropriately commented that nurses should not undertake treatment of any injuries. None gave any details on management in cases requiring admission and very few on closing the case if concerns are not verified.

Interagency working

This did not always feature highly and was not always clear. Some protocols made no mention of social services at all. Some gave helpful advice on early liaison with social services, the importance of joint planning and ongoing information sharing. One protocol repeatedly emphasised that staff should not share information with other agencies until this had been approved by senior managers or professionals. Some of the protocols provided helpful details on how to prepare a report for a case conference or statements for the police.

Information sharing, consent and confidentiality

No protocols clearly set out guidelines on information sharing, very few gave any details on consent, and none mentioned confidentiality. The few that did give details limited this to reminding staff of the importance of explaining concerns and action to the accompanying adult, but also did not give guidance on when it is not appropriate to do so. One protocol specifically stated that “The parents should not usually be told that [child abuse] is the reason for admission”. One protocol gave helpful guidance on what to do if consent to examination is not given.

Evidence of a child centred approach

There were few protocols that retained an emphasis on the welfare of the child and none that gave any specific guidelines on listening to the child. The inclusion of a clear statement early on in a few of the protocols did, however, convey a child centred approach. These included statements emphasising that “the welfare of the child is paramount” or that “children are vulnerable”. Other helpful points were consideration of the environment, providing a quiet private room with toys available. Only two mentioned the importance of abuse in disabled children, and none referred to cultural, ethnic or gender issues.

Other features

Some protocols had information on the management of specific issues, including alleged sexual abuse and the taking of forensic samples; possible fabricated or induced illness (Munchausen syndrome by proxy); unaccompanied children; domestic violence; and parental mental illness. These were most helpful when they were included in separate, and therefore identifiable, headed sections.

DISCUSSION

Maximising the recognition of children at risk and optimising any subsequent response within emergency department settings is a priority that can be enhanced by promoting recognition and development of clear, standardised, accessible procedures. Such procedures should accord with both national and local guidelines,1 4 16 allowing for consistency in standards, while remaining locally relevant. This representative survey of procedures in emergency departments has enabled identification of good practice around the country, provides pointers towards developing robust local protocols, and suggests some quality standards for the development of such protocols (table 2). These quality standards have been developed by consensus among the authors and reflect our opinions based on the review of supplied procedures. We have not been able to identify any clear evidence base to support these recommendations, and further research is needed to evaluate their validity and usefulness.

For any child attending an emergency department, the first responsibility of the staff is to attend to that child’s needs. Medical treatment, including treatment of injuries and administration of analgesia, should not be delayed because of concerns about possible child abuse or neglect.

Where concerns about possible abuse or neglect are identified, the lead agency for investigation is social services.4 There should be clear lines of referral and responsibility in place, emphasising that all staff members have a responsibility to protect children. Consideration should be given to different responses according to the level of concern including clear procedures for responding to situations where a child, other family member, staff member or member of the public is potentially at immediate risk of harm. There should be clear procedures to be followed in the case of admission to a ward, which include full handover and transfer of responsibilities; and also for children who are discharged from the department, including arrangements for follow up and for notifying the primary care team and other relevant professionals. For minor injuries units and other small units, procedures need to be in place to facilitate good liaison between the small unit and any larger referral unit providing support and secondary care. While the larger units may follow through on concerns, the referring unit must take responsibility for ensuring that information is passed on and received, and that concerns are acted on. This may require safeguards to ensure that a child has attended the main department, or for contacting social services directly in cases of concern.

There is currently no scientific evidence to support the use of screening tools or checklists in identifying children at risk of abuse or neglect and no evidence from this or other studies to inform which procedures are effective at identifying children at risk. However, some early indicators are perceived by professionals to be suggestive (though not diagnostic) of abuse or neglect.17 If a checklist is incorporated, it should be simple and specific, and staff should be reminded that they serve as an aide memoire and not as a screening tool. Staff should also be reminded of the importance of listening to the child.18 This may involve both verbal and non-verbal forms of communication, and consideration should be given to the age, developmental stage and ethnicity of the child, along with any disability.

The presence of a child’s name on the child protection register, or any other database, should not be used as an indicator of risk, nor should it be used as the sole basis for decision making in relation to possible concern, but is a relevant additional piece of information and an important part of understanding the full context of the child’s presentation. Each child should receive an evaluation of the risks and concerns identified through the presentation, and any action should be based on such an evaluation. Nevertheless, for any child in whom there are pre-existing concerns, an emergency department attendance is a significant event (this may be positive, indicating appropriate care on the part of the parents/carers). In order to facilitate the flow of information to those professionals working with the child and family, all emergency department attendances should be notified to the child’s primary care team. Recent government guidance recommends that local authorities should no longer hold separate child protection registers.4 Children about whom there are concerns will still be subject to child protection conferences at which a child protection plan will be put in place, but their names will no longer be held on a separate register. Emergency departments will therefore need to work with their Local Safeguarding Children Boards (LSCBs, formerly ACPCs) and named professionals to put in place appropriate systems for accessing and sharing information.

Effective child protection requires joint working of professionals from different agencies. Information sharing is essential in order to fully evaluate and appropriately respond to possible concerns. Staff should be encouraged to share concerns with other agencies, to respond to requests for information in relation to child protection, and to question other professionals where there are differences in opinion.16 18 19 Advising staff not to share information with other agencies is inappropriate and goes against national guidance in relation to child protection.16 Concerns about possible child abuse or neglect should normally be shared openly with the parents or carers unless to do so might further increase the risk to the child, or could compromise any criminal investigation. Consent to examination or any investigations should normally be obtained from a person with parental responsibility and from the child, providing he/she is competent to give such consent.16 20

CONCLUSIONS

This representative survey of child protection procedures in emergency departments has confirmed that most departments have operational protocols, but has highlighted a wide variation in the quality and appropriateness of those protocols. Some examples of good practice have been identified and the process of evaluating local protocols has informed some recommendations on good practice. In his report on the death of Victoria Climbie, Lord Laming emphasised that “the investigation and management of a case of possible deliberate harm to a child must be approached in the same systematic and rigorous manner as would be appropriate to the investigation and management of any other potentially fatal disease”.18 Lord Laming made numerous recommendations on the appropriate investigation and management of suspected child abuse, much of which is being implemented within the health service.1 21 In order to ensure that children are given the highest level of protection from harm, emergency departments need to have in place robust and workable procedures that are in keeping with national recommendations, responsive to local needs, and accessible to all staff. The results of this survey should assist emergency departments in developing child protection guidelines to give their staff clear direction when faced with concerns about a child.