Intended for healthcare professionals

Editorials

Managing minor head injury in children

BMJ 2006; 333 doi: https://doi.org/10.1136/bmj.38954.516933.DE (Published 31 August 2006) Cite this as: BMJ 2006;333:455
  1. Harvey Marcovitch, associate editor, BMJ (h.marcovitch{at}btinternet.com)
  1. BMJ Editorial, Tavistock Square, London WC1H 9JR

    Early discharge after computed tomography is safe and cost effective

    In the United Kingdom some 500 000 children attend accident and emergency departments after a head injury and about 50 000 are admitted (nearly 10% of all hospital admissions). The death rate is 5.3 per 100 000 children.1 Despite various guidelines on the management of head injury, clinical practice varies. Two papers in this week's BMJ shed new light on the management of mild head injury in children admitted to emergency departments.23 They compare immediate computed tomography for triage with inpatient observation in mild head injury—920 of the 2602 enrolled patients were aged 6-15. No child died or needed surgical intervention. Early discharge after immediate computed tomography was no worse in terms of recovery at three months, later complications, and patient satisfaction; it was also less costly than inpatient observation.

    Here is a scenario familiar to emergency department doctors, paediatricians, and general practitioners. A child suffers apparently mild head trauma with brief loss of consciousness. When seen, he or she has normal neurological findings and a Glasgow coma scale score of 15. The child makes an uninterrupted recovery but spends 24 hours in hospital, with worried parents at the bedside and nurses performing hourly neurological observations of variable quality. Fortunately a less common scenario is that, in the hour or so between the decision to admit the child and actual admission to the ward, he or she develops neurological signs from an unsuspected subdural haematoma and is left with permanent sequelae, despite transfer to a neurosurgical centre. How then, do we differentiate between children who need urgent imaging and referral to a neurosurgical service and those who need observation only?

    A literature review designed to inform a clinical practice guideline for the American Academy of Pediatrics found 108 relevant studies before 1999, but it concluded that there was not sufficient scientific basis for evidence based recommendations on key management issues.4 A subsequent meta-analysis of studies investigating clinical variables that might predict serious intracranial injury in children showed, unsurprisingly, that predictors include reduced level of consciousness, focal neurological signs, and skull fracture. Neither headache (difficult or impossible to recognise in children who have not yet started to speak) nor vomiting increased the risk.5

    Skull radiography was used for many decades to identify children at greatest risk. However, although skull fractures increase the risk of intracranial injury fourfold, the presence of a fracture is of limited value as a diagnostic test—with a calculated sensitivity of 59% and specificity of 88%.5 A retrospective audit of two cohorts of children presenting to an emergency department in a UK paediatric teaching hospital (before and after restricting the use of skull x rays) concluded that abolishing the use of x rays, except in infants under 12 months, did not increase the admission rate or the rate of missed intracranial injury and reduced overall radiation dosage, despite doubling the proportion of children who receive a computed tomography scan from 1.0% to 2.1%.6 The authors' protocol permitted skull x rays for the youngest infants if non-accidental injury was suspected or evidence of head trauma existed.7

    In June 2003, the National Institute for Clinical Excellence (NICE) published clinical guidelines for managing head injury; these guidelines were mainly based on the Canadian computed tomography head rule and the New Orleans criteria.89 Both these sets of criteria are 100% sensitive for the need for neurosurgical intervention, and the Canadian rules have higher specificity for important clinical outcomes (50.6% v 12.7%) so are likely to result in lower imaging rates.10

    These guidelines advocate a strategy based on early computed tomography scanning for patients deemed clinically to be at high risk and a major reduction in the use of skull x rays. Audits in the UK, conducted since the guidelines were adopted, have shown an increase in the incidence of performing computed tomography head scans: from 1-2% to 7-9% of attendees of all ages in hospitals following NICE guidelines1112 and from 14% to 20% at a hospital following the Canadian rules.13

    The practicability of applying any of these guidelines will depend to some extent on the availability of staff and imaging facilities and the cash to pay for them in a clinical situation more commonly met out of hours. One question that arises is whether extra expenditure on scanning might be balanced by savings on hospital admission—unlikely in circumstances where even patients with a normal scan might still be admitted.

    Clinicians may have little difficulty in predicting the response of their radiological colleagues to a request to image all patients presenting with minor head injury, however much they point out the savings of about €200 (£135; $256) per patient calculated by the Swedish group.3 To convince radiologists, further studies are needed to define a highest risk group who could be scanned on arrival then discharged (or x rayed if under 12 months) and a lowest risk group who could be sent home from an observation unit without being scanned. Two subjective issues would need to be dealt with by qualitative research: the degree to which patients are reassured by (unnecessary) imaging thereby preventing re-attendance, and the degree to which doctors' (possibly irrational) fears of medicolegal consequences of not imaging can be assuaged.

    Footnotes

    • Competing interests None declared.

      Research pp 465, 469

    References

    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    11. 11.
    12. 12.
    13. 13.
    View Abstract