Background: A quarter of all patients presenting to emergency departments are children. Although there are several large, well-conducted studies on adults enabling accurate selection of patients with head injury at high risk for computed tomography scanning, no such study has derived a rule for children.
Aim: To conduct a prospective multicentre diagnostic cohort study to provide a rule for selection of high-risk children with head injury for computed tomography scanning.
Design: All children presenting to the emergency departments of 10 hospitals in the northwest of England with any severity of head injury were recruited. A tailor-made proforma was used to collect data on around 40 clinical variables for each child. These variables were defined from a literature review, and a pilot study was conducted before the children’s head injury algorithm for the prediction of important clinical events (CHALICE) study. All children who had a clinically significant head injury (death, need for neurosurgical intervention or abnormality on a computed tomography scan) were identified. Recursive partitioning was used to create a highly sensitive rule for the prediction of significant intracranial pathology.
Results: 22 772 children were recruited over 2½ years. 65% of these were boys and 56% were <5 years old. 281 children showed an abnormality on the computed tomography scan, 137 had a neurosurgical operation and 15 died. The CHALICE rule was derived with a sensitivity of 98% (95% confidence interval (CI) 96% to 100%) and a specificity of 87% (95% CI 86% to 87%) for the prediction of clinically significant head injury, and requires a computed tomography scan rate of 14%.
Conclusion: A highly sensitive clinical decision rule is derived for the identification of children who should undergo computed tomography scanning after head injury. This rule has the potential to improve and standardise the care of children presenting with head injuries. Validation of this rule in new cohorts of patients should now be undertaken.
- CHALICE, children’s head injury algorithm for the prediction of important clinical events
- LOC, loss of consciousness
- RCS, Royal College of Surgeons
- SXR, skull radiograph
Statistics from Altmetric.com
Funding: This study was funded by the Enid Linder Foundation Research Fellowship from the Royal College of Surgeons of England, The Child Brain Injury Trust and The Dickinson Trust. None of these funding bodies had any role in the conduct of the study or in the preparation of the manuscript.
Competing interests: None.
This paper conforms to the STARD guidelines for reporting of diagnostic cohort studies.
Contributors: JD was the lead researcher, and led the study and prepared the manuscript. KM-J conceived the project, obtained funding, supervised the research, and oversaw and supervised its conduct. JPD created the proforma and supervised data collation. J-PL carried out the analysis for interobserver agreement. FL and JB organised the multicentre participation and supervised its conduct. All authors had access to all study data and participated in manuscript preparation and checking.
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.