Hospital facilities for head injured children: UK national survey
- Correspondence to Rosie Houston, Centre for Maternal and Child Enquiries (CMACE), Chiltern Court, 188 Baker Street, London NW1 5SD, UK;
Contributors All authors had full access to all of the data in the study and can take responsibility for the integrity of the data and the accuracy of the data analysis.
- Accepted 4 March 2011
- Published Online First 23 May 2011
Aim To assess how child emergency department (ED) attendances are distributed between hospitals across England, Wales and Northern Ireland and how care is organised for children with a head injury.
Methods A cross-sectional survey was performed of the 245 hospitals in England Wales, Northern Ireland and the Crown Dependencies (Channel Islands and Isle of Man) which were eligible to participate in the enquiry from September 2009 to April 2010. The survey covered hospital details, departments and procedures, ED activity, imaging, admission and discharge procedures, referral and transfer, documentation, training and audit, information and advice, and non-accidental head injuries.
Results 64% of hospitals have an established pathway for management of head injured children. Not infrequently hospitals asserting designation as specialist trauma or specialist neurosurgical centres do not offer an intensive care service for children. 82% of child ED attendances are to hospitals that would not care for a critically ill child on-site. Hospitals that do offer such care are much more likely to have children's trained staff available in the ED. They are also more likely to have access to surgical support beyond neurosurgery.
Conclusion Given the extent of variation between hospitals in the facilities available for head injured children, further comparative studies into the standards of care delivered and outcomes (including a confidential enquiry) are indicated.
In England and Wales it is estimated that approximately 210 000 children attend hospital every year due to head injury.1–3 In 2006/2007, hospital episode statistics indicated that 34 483 children under 15 years were admitted to a hospital in England with a primary diagnosis of ‘injury to the head’ (ICD10 codes S00–S09). The annual rate of intensive care (IC) admission related to traumatic brain injury (TBI) ranges from 5.6 cases per 100 000 children in England and Wales to 7.3 per 100 000 per year in Northern Ireland.4
Although the majority of head injuries in children are minor,3 a small number have serious acute intracranial complications. TBI is the most common cause of morbidity, mortality, disability and lost years of productive life in children.1 In 2002, TBI caused 2% of deaths in those aged 0–14 years, and 30% of deaths due to external causes in 1–14-year-olds in England and Wales.5 6 The prognosis in severe TBI is poor: one study reported that the outcome of approximately 26% of patients admitted to hospital with Glasgow Coma Scale (GCS) <13 was either death or a persistent vegetative state, and that only 20% of survivors make a good recovery.7 These figures impact hugely on the individual and their family, and on the state for the long-term provision of care.
Prehospital care and emergency department (ED) professionals see a large number of children with head injury and have to identify the small number that will need further observation, investigation or intervention.3 While little can be done to alter the primary injury, effective optimal management to reduce secondary injury can significantly affect outcome.8 Previous studies reported large variations in practice for this group.9
This national survey formed part of the National Confidential Enquiry into Head Injury in Children by the Centre for Maternal and Child Enquiries. Its aim was to assess how child ED attendances are distributed between hospitals across England, Wales and Northern Ireland and how care is organised for children with a head injury. We also sought to identify and describe current practice in the acute management of head injury in children and assess variation between organisations in relation to national guidelines.3
A cross-sectional survey was performed of all the hospitals in England Wales, Northern Ireland and the Crown Dependencies (Channel Islands and Isle of Man) from September 2009 to April 2010. Information was collected at hospital rather than Trust level in order to give a better indication of the facilities available for a patient at the location where care is received. A local head injury enquiry coordinator was identified in each hospital to whom questionnaires were sent. However, it was emphasised that a multi-professional effort may be necessary to complete the questionnaire.
The survey instrument was developed following a review of the literature, including peer reviewed journals, surveys, standards and guidelines, and discussions with experienced professionals. The final questionnaire consisted of 10 sections covering: hospital details, departments and procedures, ED activity, imaging, admission and discharge procedures, referral and transfer, documentation, training and audit, information and advice, and non-accidental head injuries. A systematic approach to the design and implementation of the survey was adopted utilising key elements of Dillman's ‘total design method’. Careful design of the survey instrument and procedure minimised the burden on respondents and maximised their motivation to respond.10–13 Questions were designed to elicit objective/factual responses. The survey was piloted by nine coordinators from six regions. The final version was reviewed by the external advisory group for the project prior to national distribution.
Coordinators were informed of the survey 1 month prior to the planned distribution. Collaboration was facilitated by email and telephone prompts and support over the next 5 weeks. Data from completed questionnaires were entered into PASW Statistics 18 for analysis. Unless otherwise stated, reported percentages and frequencies have been calculated excluding missing data. Comparisons were made using χ2. Where cell counts were <5, Yates' continuity correction was used.
Questionnaires were returned by 230 out of 245 eligible hospitals, corresponding to a 94% response rate and describing the services available for over 3 million child ED attendances. We present the results first by the descriptions that the hospitals chose for themselves but subsequently by the level of service they stated that they offer to head injured children.
Table 1 shows the response rate by type of hospital—self reported using mutually exclusive categories. The proportions of each type of hospital admitting children are listed in table 2. In addition to the designations in table 1, hospitals were able to assert their status as specialist neurosurgical centres, paediatric neurosurgical centres, trauma centres and/or paediatric trauma centres. These data are also presented in table 2.
Four of the 22 hospitals claiming specialist status as neurosurgical units would not provide IC for children in their hospital. In two of these a critically injured paediatric trauma patient requiring hospitalisation would not be cared for in their hospital, and in another it was noted that if the child required surgical care they would be transferred. One specialist neurosurgical unit in a district general hospital provides IC for children on the general (adult) intensive care unit (ICU).
Forty per cent (20/50) of hospitals claiming to be trauma centres would not care for a paediatric trauma patient in their hospital, and in 70% (35/50) a critically injured child requiring a paediatric ICU (PICU) would not be treated there.
In light of the many idiosyncrasies and variation over which facilities are provided by which type of hospital, we abandoned these titles and present an analysis based on the level of service offered for children.
In 205/230 hospitals a critically injured paediatric trauma patient requiring IC would be looked after in another hospital. Of the 25 hospitals that would receive these patients, 22 would care for them in a PICU and three in an adult ICU in their hospital. Three of the 25 do not have access to paediatric surgery and one of the hospitals with a PICU has no ED. Two of the hospitals providing IC cannot accept head injuries of all severities. Specifically the latter do not have on-site paediatric neurosurgery. In some cases these hospitals attempt to divert children directly from the referring hospital on a case-by-case basis (making the judgement that surgery is required on the basis of the history and remote access to the CT scan), but in others they end up accepting patients who require a further transfer.
Departments and procedures
A total of 210 hospitals have an ED; 32 of these do not admit children. Of the 178 hospitals with an ED which admit children, 154 (87%) would refer a child needing IC from their ED. These hospitals have different levels of resource when compared to hospitals with an ED that would provide critical care on-site (n=24, including 21 with PICUs and three in which the care would be delivered in an adult ICU). They are less likely to have available a consultant with training in paediatric emergency medicine (OR 2.67, 95% CI 1.08 to 6.61), a paediatrician with training in emergency medicine (OR 10.95, 95% CI 4.18 to 28.71) or a paediatric anaesthetist (OR 47.08, 95% CI 2.81 to 788.95). They are as likely to have at least one children's trained nurse on their staff but less likely to have cover for every shift (OR 9.23, 95% CI 2.88 to 29.57). They are as likely to have general paediatric surgery and ear, nose and throat surgery on-site, but less likely to have thoracic surgery (OR 21.27, 95% CI 7.27 to 672.23) or maxillo-facial surgery (OR 11.89, 95% CI 2.7 to 52.35) compared to hospitals providing ED and IC for children on-site.
Overall, 64% of hospitals have an established pathway for the management of head injured children in the ED; 58% could describe this as a protocol (written or otherwise).
Figure 1 is a map showing where children present to ED. Eighty-two per cent of child ED attendances are to departments in hospitals that would not care for a critically ill child on-site. The EDs in the 24 hospitals (with an ED) that would look after a critically ill child on-site see an average of 25 000 (SEM) children a year compared to an average of 15 000 (SEM) a year in the 186 hospitals that would send a critically ill child elsewhere. Figure 2 is a map showing centres that would admit a critically ill head-injured child.
The 178 hospitals with an ED and which admit children for secondary care all have access to CT at all hours, as do the hospitals which provide critical care for head injured children. Ninety per cent of hospitals have a written protocol stating which patients should have a cranial CT. Nevertheless compliance with NICE guidelines (stated intended management) for the use of CT was incomplete (table 3). The differences between hospitals that would provide IC on-site if required and those that would transfer a child for such care were not significant at the 95% level.
Neurosurgical referral almost universally requires remote review of the cranial CT; however, while only 89/154 (58%) had an established protocol for teleradiology, 138/154 (90%) had an ‘effective image transfer facility available’ (undefined).
Admission and discharge procedures
Thirty-two hospitals reported themselves to have an ED but claimed not to admit children for inpatient care. These hospitals have a mean of 41 000 ED attendances per year (SEM 6200), 9000 of which are children (SEM 1400). Twenty-three of these hospitals (72%) have a bypass protocol to minimise the risk of a severely injured child attending via the ambulance service. Overall, 79% (181/230) of hospitals reported that they admit children for secondary or tertiary care, three of which do not have an ED. The composition of the teams most likely to provide this care is heterogeneous across all sites, with no specific difference in pattern (other than neurosurgical involvement) between hospitals that provide critical care on-site and those that do not.
Hospitals described two patterns of clinical practice based on the use (and indication for) cranial CT. These could be summarised as ‘scan and discharge (if CT normal)’ versus ‘admit and observe’. The choice between these treatment plans is affected by the ease of access to CT (which was not graded in the survey), the likelihood of cooperation by the patient (related to age and access to anaesthetic support) and the availability of short term observation units (59% of hospitals).
Retrieval and transfer
Eighty-seven per cent of hospitals with an ED transfer critically ill children for definitive care. All have access to CT. Twenty-five hospitals have critical care facilities for children (22 PICU, 3 adult ICU), 23 of which include some access to neurosurgical care. Fifteen claimed to be specialist paediatric neurosurgical units and 13 to be paediatric trauma centres. However, a further 38 hospitals (all of which would transfer a child requiring critical care) considered themselves eligible to receive a child from another hospital. This was usually district general hospitals supporting minor injury units.
Seven hospitals with an ICU for children on-site have a designated consultant responsible for retrieval. Fifty-two per cent of units providing IC had written agreements regarding transfer arrangements with at least one referring unit. Similarly, 50% of referring units had written agreements regarding transfer. Only 7% of referring hospitals have a named consultant associated with the policy. For hospitals that would not provide care for critically injured children, written agreements with ambulance providers about the transfer arrangements were less common (24%).
Seventy-eight per cent of hospitals meeting head injured children in their ED or providing inpatient care for them report the use of a standard proforma for their record keeping. Sixty-three per cent of referring hospitals reported having a separate proforma for children.
Training and audit
ED equipped hospitals that provide inpatient care for children but which transfer critically ill children out have similar levels of advanced paediatric life support (APLS) trained staff in the ED to those which provide IC for children. However, the proportion of the 32 hospitals which do not admit children (but which see 32×9000 child attendances a year) having comprehensive cover from APLS trained staff is significantly smaller (p<0.0001, OR 13.2, 95% CI 5.44 to 31.95).
Only 65% of all the hospitals responding to the survey had a training programme for staff on the assessment and recording of observations in children, and only 54% of these included key risk factors in head injury and indications for urgent CT. Fifteen per cent of the programmes did not include calculation of the GCS in children, and 11% did not include any teaching of neuro-observations in children. Twenty-five per cent of hospitals had conducted an audit in relation to head injury in the 12 months prior to survey completion.
Information and advice
Fifty-eight per cent of hospitals have a clinical lead with designated responsibility for implementing the NICE guideline. Fifteen per cent have a board or display area with details of guidance and support organisations, as recommended in the NICE guideline.3
Hospitals without critical care facilities for children are marginally more likely to provide written advice on discharge (OR 0.07, 95% CI 0.006 to 0.84) but the proportion is high in both instances. Information leaflets rarely include details about the long-term impact on families (15%) or public health literature and non-medical sources of advice (9%).
Non-accidental head injuries and safeguarding
Child protection training and procedures are asserted to be well delivered; 95% and 91% of hospitals, respectively, had a named nurse and named consultant for child protection/safeguarding children. The majority of hospitals (82%) have a dedicated and integrated team consisting of named nurse, paediatric liaison advisor and child protection coordinator. A safeguarding forum exists in 78% of hospitals. Protocols and guidelines for the assessment/management and referral of children with a suspected non-accidental injury (NAI) are also widely available (97%); in 78% of hospitals this guideline has been reviewed at least once since 2007. However, although initial mandatory training in identifying potential signs of NAI is provided in 96% of hospitals and annual safeguarding training in over 90%, a standard assessment tool to help identify cases is only available in half of them.
Only 71% of hospitals with direct access to CT reported that children who required imaging of the head or cervical spine would be assessed by a clinician experienced in detection of NAI.
In this survey, hospitals were allowed to choose their designation in categories which (for the purposes of the survey) were mutually exclusive; however this led to semantic confusion. For example, some children's hospitals coexist within larger hospitals and there was little consistency surrounding the services available in ‘trauma centres’. To circumvent the idiosyncrasies of the pattern of service provision, we divided the data by the level of service that each hospital would provide to a head injured child.
Twenty years ago, 42% of deaths of head injured children involved avoidable factors in hospital care.14 The centralisation of paediatric IC in the late 1990s was justified by evidence of better outcomes,15 some of which were specific for head injured children.16 It has led to a difference in practice compared to adults. Head injured adults are moved to specialist centres if neurosurgical intervention is required, the necessary triage being enabled by electronic transfer of CT images. In paediatric practice, the decision about transfer is made on the basis of the need for IC rather than neurosurgical intervention. It is therefore much more likely that a head injured child will need to be transferred between institutions and even more pertinent in this group to consider primary transfer to specialist centres from the scene.
The survey was not designed to compare standards of care or outcomes. However it can be used to help model workload, particularly in relation to patient transfer. If head injured children present to the ED in the same pattern as the totality of child presentations, then 82% of the workload is borne by departments with limited resource. This will not be problematic for minor trauma (the vast majority), but those requiring IC and/or neurosurgical intervention then have to undergo further transfer, resulting in critical delay before definitive care with anticipated poorer outcome.17 Eighty-seven per cent of hospitals (accounting for 82% of the workload) would transfer these children out. Put another way, of the 578 serious head injuries in children (aged 0–14 years) in a year in England, Wales and Northern Ireland (‘serious’ defined by need for IC admission),4 only 104 present to hospitals capable of providing them with critical care. The remaining 474 need urgent transfer to another institution. A total of 117 of the 57817 require emergency surgery, of whom only about 20 will have coincidentally presented to an institution that can care for them through to IC, and only 16 to a hospital that can also provide the necessary surgery. The remaining 101 are among those who require emergency transfer.
We found striking differences in staffing which reflected the level of service. Although they have the benefit of greater access to support from paediatric IC and paediatric anaesthetists, major trauma is less frequently seen in paediatric EDs (children are only one-quarter of the population and the incidence is low in the lower age groups). Studies suggesting advantages to the centralisation of paediatric trauma usually include the advantage of proximity to paediatric IC.18 19 Since 35 of the 50 hospitals (70%) claiming ‘trauma centre’ status in this survey did not provide IC for children, it is likely they may not experience the improved outcomes. By contrast, even urgent paediatric neurosurgical procedures do not automatically require postoperative IC and hence services are not always co-located.
The self-report survey was dependent on a reliable response. Effort was made to ensure comparability of the data from respondents through piloting the survey and providing guidance, yet there were still some ambiguous responses. Respondents may have misinterpreted some questions and subsequently provided incorrect responses, or responses that could be misinterpreted by the researcher.
Given the extent of variation between hospitals in the facilities available for head injured children, further comparative studies into the standards of care delivered and outcomes (including a confidential enquiry) are indicated.
We would like to thank all of the respondents and their colleagues for taking the time to complete the survey.
Funding This work was undertaken by the Centre for Maternal and Child Enquiries (CMACE) as part of the Child Health Enquiry under the Confidential Enquiry into Maternal and Child Health (CEMACH) programme. The CEMACH programme of work is funded by the National Patient Safety Agency. Additional contributors to the Child Health Enquiry include the Department of Health, Social Services and Public Safety of Northern Ireland, the Isle of Man and the States of Jersey and Guernsey. The views expressed in this publication are those of the authors and not those of the funding bodies.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.