Background Head injuries are a common emergency department (ED) presentation. The National Institute for Health and Clinical Excellence (NICE) updated its guidance in September 2007 regarding imaging required for patients with head injuries.
Methods A two-centre observational ED study was carried out, examining imaging practice in adults and children with head injuries attending pre-guideline and post-guideline implementation. Guideline implementation occurred through a formal implementation programme at the teaching hospital, and informally at the district general hospital (DGH). Retrospective extraction took place of prospectively recorded data case records and radiology department imaging registers. Pre-implementation data were collected from Salford Royal Foundation NHS Trust (SRFT) from January and February 2008 and post-implementation data in May 2008. Post-implementation data was collated from Royal Bolton Hospital Foundation NHS Trust (RBFT) from September to November 2007. Compliance with NICE 2007 was the primary outcome assessed.
Results With the implementation of NICE 2007 guidelines at SRFT, a significant increase in compliance from 94.2% (92.9–95.5) to 98.8% (98.2–99.3) was observed for adults requiring head CTs, with an overall trend to improved clinical practice in the adult patient populations. However, a significant number of children (SRFT 68.7% and RBFT 77.1%) did not receive the indicated head CT scan following a head injury, after implementation of the guidelines.
Conclusions The SRFT implementation strategy employed was successful for adults, with the overall trend to increased clinical compliance post-guideline introduction. Evidence of a reluctance to adhere to the NICE recommendations for children indicated for CT head scan after a head injury was observed.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
From 2008 to 2009, a total of 272 485 patients were admitted to UK hospitals following a head injury.1 As such, traumatic brain injury remains a significant and expensive public health issue. It is therefore imperative that patients who will benefit from emergency treatment receive pertinent imaging that facilitates their expeditious diagnosis and treatment.
The National Institute for Health and Clinical Excellence (NICE) guidelines for the management of head and cervical spine (C-spine) injuries in patients presenting to the emergency department (ED) were updated in September 2007.2 They include revised head CT guidance for patients over/under 16 years (boxes 1 and 2).
NICE guidance: adult head CT
NICE: adult head CT
CT immediately, imaging and results analysed within 1 h of request:
Glasgow Coma Score (GCS)<13 on emergency department (ED) assessment
GCS<15 in ED 2 h post injury
Suspected open/depressed skull fracture
Sign of basal skull fracture (haemotympanum; ‘panda’ eyes; cerebrospinsal fluid leakage from ears/nose; Battle's sign)
Focal neurological deficit
>1 episode of vomiting
Any amnesia or loss of consciousness since the injury with coagulopathy (history of bleeding, clotting disorder, current treatment with warfarin).
CT should be carried out within 8 h of injury:
Amnesia of events > 30 min before impact
Any amnesia or loss of consciousness since the injury with either: age ≥65 years or dangerous mechanism of injury (pedestrian/cyclist struck by motor vehicle; occupant ejected from motor vehicle; fall from >1 m or five stairs).
NICE guidance: child head CT
NICE: Under 16s head CT
Witnessed loss of consciousness lasting >5 min
Amnesia (antegrade/retrograde) lasting >5 min
Three or more discrete episodes of vomiting
Clinical suspicion of non-accidental injury (NAI)
Post-traumatic seizure without a history of epilepsy
Age >1 year: Glasgow Coma Score (GCS) <14 on emergency department (ED) assessment
Age <1 year: GCS (paediatric) <15 on ED assessment
Suspicion of open or depressed skull injury or tense fontanelle
Any sign of basal skull fracture (as above)
Focal neurological deficit
Age <1 year: presence of bruise, swelling or laceration >5 cm on the head
Dangerous mechanism of injury (high-speed road traffic accident either as a pedestrian, cyclist or vehicle occupant; fall from >3 m; high-speed injury from a projectile/object).
The three principal modifications between the 2007 and 2003 NICE head injury updates are: the expedition of times to adult CT head scanning; the advocation of CT for c-spine imaging; and the implementation of head injury imaging guidance validated for use in children.2–5
Looking at each of these in turn, adults requiring head CTs are, with the 2007 guidance, now stratified into two groups: those requiring a scan immediately or within 8 h.2 The provision of NICE C-spine imaging guidance is a new development, building on the Canadian C-spine rules used in current clinical practice, and advocates imaging of the entire c-spine with CT (rather than radiographs, or selected CT imaging, eg that of C1 and 2 in intubated patients) once certain criteria are met.2 4 Previous CT imaging guidance in children was adapted from adult-derived guidance. The highly sensitive children's head imaging guidance now incorporated in the new NICE guidance was derived for the paediatric population via the CHALICE study in 2006.5 Consequently, all of these changes necessitate 24 h access to CT scanning and interpretation.
The objectives of this study were to establish whether clinical practice followed the standards set by the NICE guidance, for all ED patients presenting with a head injury, before and after the new 2007 guidelines were introduced with and without a formal implementation strategy. The implications of resource utilisation in the implementation of evidence-based clinical practice were also examined.
This study comprised a retrospective review of contemporaneous, anonymised patient records. This was conducted before and after implementation of the NICE 2007 guidelines in the ED of a large, urban, university teaching hospital with tertiary neurosciences services–Salford Royal Foundation NHS Trust (SRFT). Post-implementation data were collected from a busy district general hospital, Royal Bolton Hospital Foundation NHS Trust (RBFT).
SRFT ED had already established good compliance with the departmental policy following the introduction of the NICE 2003 adult head injury guidance.3 6 Although updated guidance was published in September 2007, it was not implemented in the department until the last fortnight in April 2008. This change in practice occurred post agreement with the radiology department and involved an ED consultant-led publicity programme. Information was conveyed through the use of teaching sessions and posters displayed in all departmental areas and propagated via a staff dissemination programme. RBFT implemented the updated NICE guidance in September 2007 without a formal implementation strategy.
The pre-implementation data collection consisted of 2 months, and the post-implementation 1 month, of data of all patients attending the SRFT ED with a head injury. Post-implementation data from RBFT were collected from an 8-week period from September to November 2007. Patients with relevant presentations were identified using a comprehensive search strategy in the ED audit databases. Mechanisms of injury suggestive of possible head injuries were also examined, including patients presenting the following: facial injuries; assault and post motor vehicle collision (MVC). To ensure that no relevant patients were omitted from the data collection, radiology records of imaging procedures performed during these timeframes were also reviewed.
The data collected were examined for demographic information and clinical compliance with the new head imaging guidance; additionally, the use of skull x-rays, (a practice that had a diminished role in the 2003 NICE update and is only advocated in the 2007 guidance as part of a skeletal survey if non-accidental injury (NAI) is suspected in children) was recorded.
The primary outcome measure was clinical compliance with the imaging selection guidelines. Secondary outcomes looked at resource use, ie number of CT scans (head), radiographs (skull) and admission rates.
Baseline characteristics for adults and children with head injuries attending EDs during the two study periods are shown in table 1. Eleven-thousand nine-hundred and eighty-eight patients attended SRFT in study period 1, of whom 347 (2.9%) were adults and 136 (1.1%) children presenting with head injuries. During study period 2, at SRFT 409 (6.3%) adults and 177 (2.7%) children attended with head injuries of the 6535 patients. Of the 17 823 patients to attend RBFT during the post-implementation study period, 353 (1.9%) were adults and 369 (2.1%) were children with head injuries. As is typical of patients presenting with head injuries, a male predominance was seen within the adults and children across the study periods and departments.
Falls were the most common mechanism of injury in children presenting to both study centres. In adults, the most common mechanism of injury resulted from assault in period 1 at SRFT, and was attributed to falls during period 2 at SRFT and BRFT.
The number of adults sustaining an intracranial abnormality on CT increased from seven (2.0%) to 14 (3.4%) across the study periods at SRFT and decreased from one (0.7%) to zero in children. At RBFT, seven (1.9%) adults and two (0.5%) children were found to have an intracranial abnormality during their post-implementation period. Mortality for adults and children remained static in periods 1 and 2 at SRFT. This outcome was not recorded in the RBFT data.
Head CT scans
The frequency of adult patients with head/neck injuries receiving head CT scans decreased significantly after implementation at SRFT, down from 75 (21.6%) to 45 (11.0%). There was no change in the frequency of CT scans for children, see table 1.
A significant reduction of un-indicated head CT scans for adult patients occurred after implementation, down from 11 (3.2%) to four (0.9%). The number of patients not receiving indicated head CT scans also reduced from nine (2.6%) to one (0.2%). At RBFT, after implementation, the number of adult patients not receiving indicated scans was higher than SRFT at 15 (4.3%) and the number receiving un-indicated head CT scans similar to SRFT at one (0.3%).
A significant increase was observed in the proportion of adult patients on whom appropriate clinical decisions were made regarding appropriate imaging, from 327 (94.2%) to 404 (98.8%), see table 2.
No head CT scans were performed without indication in children before or after implementation at either of the study centres. Compliance remained static with 127 (93.4%) children before implementation and 166 (93.8%) children after implementation receiving appropriate clinical decisions at SRFT; however, the percentage of indicated and not performed head CT scans remained high.
After implementation, 16 (9.0%) children at SRFT and 35 (9.2%) children at RBFT with head/neck injury were indicated for head CT scans. Of these indicated head CT scans, only five (31.3%) and eight (22.9%) scans, respectively, were performed. In the SRFT pre-implementation period, when the 2003 NICE guidelines were being utilised, of the 15 indicated child CT head scans, six (40.0%) were performed, see table 2.
No single clinical sign predominantly indicated that head CT scans were performed on paediatric patients; however, children with more than one pertinent clinical sign were more likely to receive a scan. Three separate episodes of vomiting was the clinical sign most frequently present in those children not receiving an indicated head CT scan, see table 3.
Skull x-rays (SXR) have a diminished role in the 2003 NICE update and are only advocated in the 2007 guidance as part of a skeletal survey if NAI is suspected in children. Clinical compliance with the NICE 2003 guidance was already high at SRFT, only three (2.2%) children with head injury were indicated for SXR and all were performed. After implementation of the 2007 guidelines, one (0.2%) adult received an un-indicated SXR and no paediatric patients received an indicated or un-indicated SXR. At RBFT, where a less formal introduction of the NICE guidelines was implemented, four (1.1%) adults and three (0.8%) children received un-indicated SXRs.
Admissions and discharge
Significant reductions in admissions were seen for all patients presenting after implementation of the updated 2007 guidance at SRFT. A total of 89 (25.7%) adults were admitted with head injuries before implementation and 60 (14.9%) adults after implementation. Child admissions were also down from 11 (8.1%) to six (3.4%). At RBFT, 45 (12.7%) adults and 24 (6.5%) children were admitted with head injuries.
Of the nine paediatric patients indicated for head CT scans that were not performed, five (55.6%) were subsequently discharged without observation before implementation at SRFT and after implementation eight (72.7%) of the eleven children. At RBFT, 15 (55.6%) of the 27 children were discharged without observation, after not receiving an indicated head CT scan.
The results of this study can be subdivided into those looking at compliance, the primary outcome, and those representing resource utilisation, eg secondary measures such as imaging and admission rates.
Looking at the rates of guideline-compliant imaging for adults, significant changes were seen regarding CT head usage at SRFT. The percentage of adults who correctly received no imaging significantly increased from 75.8% to 88.8%, and there were significant reductions in those patients receiving un-indicated scans. Overall compliance in this population increased, significantly, from 94.2% to 98.8% at SRFT. At RBFT, where a less formal implementation of the guidelines was implemented, 88.1% of adults correctly received no imaging with a compliance rate of 95.5%. Thus, the implementation strategy with formal staff training sessions increased compliance with NICE 2007.
Although adult-indicated CT head scanning was consistently performed, this was not mirrored in the paediatric population. Only five of the indicated 16 CT head scans were performed on children after implementation at SRFT and eight of the 35 at RBFT.
It had been anticipated prior to data evaluation that a negligible number of skull radiographs were being performed, given their lack of value in current clinical practice. The present results support this assumption with over 99% compliance before and after implementation.
On examining the secondary outcome measures to look at resource utilisation, significant reductions in rates of imaging were seen after implementation for adult head CTs and child skull radiographs at SRFT.
Looking at the limitations of this study, small numbers of adult and child patients attended within the study periods with head injuries. Data were collected retrospectively from clinical records. This made documentation and interpretation errors possible within the data collection. From the baseline study population characteristics, listed in table 1, it can be deduced that a greater proportion of less severely head-injured patients attended during the post-implementation phase, thus necessitating fewer admissions and a reduced imaging rate.
This study has highlighted the decision-making of when to use CT in the paediatric management of head injuries. Browning et al presented guidelines for the radiological imaging in head injuries of children,10 which are in keeping with the findings presented here. Although the clinical features of reduced Glasgow Coma Score, focal neurological signs or signs of a base of skull fracture are powerful indicators of clinical course of serious head injury and result in an immediate head CT, the softer signs of loss of consciousness, amnesia, headache, vomiting and lethargy present a dilemma to the physician.
A recent cohort study suggests that clinician experience and parental preference should be taken into account in the CT decision-making of this intermediate risk group.11 Their algorithm, is ‘assistive rather than directive’ for determining the need to perform a CT scan on those children presenting with softer signs. However, Dunning et al suggest that, due to the small number of head injuries with intracranial abnormality, clinical experience of such cases will be very limited and therefore should not be relied on for a clinical diagnosis.11 The difficulty clinicians face in determining if such children should receive a CT head scan was also highlighted in the conclusions of a recent systematic review of clinical prediction rules.12
The present results indicate that clinicians appear to be less likely to perform the indicated CT head scan on a child presenting with these softer signs. When balancing the issues regarding sedation to obtain compliance and/or perceived radiation risks associated with CT in the young versus the benefits of ruling out intracranial abnormality, it was found that 68.7% (SRFT) and 77.1% (RBFT) of children did not receive the indicated head CT scan and these realistic findings of physician practices are more in keeping with the guidelines of Brown et al.10 However, as the number of intracranial abnormalities in children is small (ranging from zero to 0.7% in this study), clinicians are likely to be ‘correct’ if they decide not to CT a child with a head injury with an indication for scan based on ‘softer’ signs. However, this approach will miss the rare event of a child presenting with softer signs and having a yet undiagnosed serious intracranial bleed. The NICE guidance is a safety net that includes these children, yet, in reality, clinicians appear on balance to prefer not to CT, and in some cases not even observe children presenting with these ‘softer’ signs.
It has been suggested that admission is a poor diagnostic test for accessing if neurosurgical intervention is needed.13 This presents an additional dilemma for the clinician, having decided not to perform an indicated CT head they then have to question the value in keeping the child in for observation. In the study group of children of intermediate risk for whom CT scans were indicated and not performed, more than half were then discharged without a period of observation–the clinicians judging in these cases that the presenting clinical features of the head injury did not warrant admission. Following up the eleven cases of ‘discharged without observation’ at SRFT, one child returned to SRFT ED. This child was transferred to a paediatric hospital for observation and a subsequent CT scan was normal. The ten other children did not return to SRFT.
From the data analysed in this study, it can be concluded that an overall trend towards increased clinical compliance occurred for adults requiring head CTs and regarding SXRs for adults and children following guideline introduction, conferring some measure of success for a formal ED implementation strategy where 24/7 access to head CT is available.
The development of paediatric guidelines highlights an important topical issue within today's clinical practice, as to how and when they should be implemented.
An obstacle to implementing the head injury guidance published by NICE in 2003 is the 24 h access to radiology. This necessary service change was further reinforced by the updated 2007 NICE guidance. Consequently, the adaptation and evolution of radiology service provision is essential, necessitating a review of the current specialty training programme and allowing accessible, available 24 h facilities to develop.
As recently advocated by Lord Darzi, in these times of improved research capabilities, studies such as this are important and indicate to what extent and in what ways the implementation of all clinically pertinent emergent guidelines can be achieved.14
Funding This research was funded in part by a Foulkes Foundation Fellowship for A Yates.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.