Head injuries in children are a common presentation in acute care settings. The major uncertainty about management of these injuries is whether the child should undergo cranial CT. Most head injuries are mild and do not require neurosurgical management. However, a small proportion of patients might present as having mild injuries but have clinically significant intracranial injuries. Although CT provides definitive and rapid diagnosis to confirm or exclude intracranial injuries, there is concern about radiation-induced cancer, particularly in young patients.1, 2, 3 Furthermore, CT scanners are resource-intensive and sedation might be required for the scan.4, 5 Reports of large increases in CT rates and wide variability in its use for paediatric head injuries are also of concern.6, 7, 8, 9
Clinical decision rules have been developed to identify children at high risk of intracranial injuries, aiming to assist clinicians to minimise CT scans while still identifying all relevant injuries.10, 11 Three clinical decision rules derived in large multicentre studies with high methodological quality are: the prediction rule for the identification of children at very low risk of clinically important traumatic brain injury (TBI), developed by the Pediatric Emergency Care Applied Research Network (PECARN; USA);9 the Canadian Assessment of Tomography for Childhood Head Injury (CATCH) rule;8 and the Children's Head Injury Algorithm for the Prediction of Important Clinical Events (CHALICE; UK).12 Unfortunately, a direct comparison of the three rules is not possible because they addressed different questions (who to CT vs who not to CT), targeted different age groups and injury severities, and used different outcomes (table 1).10 Despite having undergone only limited external validation,13, 14, 15, 16 these rules are widely used or recommended: the American Academy of Pediatrics suggests that PECARN criteria should be used to determine whether imaging is indicated,17 elements of CATCH are in the Canadian Paediatric Society position statement,18 and CHALICE has been incorporated into UK guidance.19 In some countries, such as Australia and New Zealand, no clinical decision rules predominate.20
Research in context
Evidence before this study
We searched MEDLINE, Embase, and the Cochrane Library for reports published from Jan 1, 2006 (the publication year of the CHALICE rule) until June 1, 2016, with the following search terms (with acronyms, synonyms, and closely related words): “craniocerebral trauma”, “tomograph, xray computed”, “decision support techniques”, “newborn, infant, child, adolescent, paediatric”, and “Pediatric Emergency Care Applied Research Network, PECARN, clinically-important brain injury, Canadian Assessment of Tomography for Childhood Head Injury, CATCH, Children's Head Injury Algorithm for the Prediction of Important Clinical Events, CHALICE”. We did not apply any study design or language restrictions. We identified further studies by examining the reference lists of all included articles and searching relevant websites. We reviewed titles or abstracts for relevance, and assessed original reports and reviews related to PECARN, CATCH, and CHALICE head injury rules. We did not find any external validation studies (not including derivation sites or derivation authors) of the PECARN, CATCH, and CHALICE rules or comparative analysis of the rules in large multicentre samples.
Added value of this study
To our knowledge, this study is the first large, appropriately powered, multicentre study to externally validate the PECARN, CATCH, and CHALICE clinical decision rules. Although all rules had high performance accuracy, the PECARN rules did not miss a single patient requiring neurosurgery.
Implications of all the available evidence
The externally validated performance accuracies of the injury rules are an important starting point for clinicians considering the introduction of one of the rules. Although a number of factors apart from rule accuracy need to be considered as well, PECARN seems to miss the fewest patients.
For clinicians, hospitals, or national bodies contemplating implementation of one of these rules, confirmation and comparison of their accuracy in an appropriately powered external validation study is essential. Two single-centre comparative validation studies have been done, but their results are difficult to translate to practice; one had very wide confidence intervals affecting the interpretation of sensitivities14 and the other had a very low underlying CT rate.16
We designed a multicentre external validation study of these three clinical decision rules for childhood head injuries, aiming to establish their diagnostic accuracy outside their derivation setting and investigate the clinical decision rules' performance in a clinically homogeneous cohort of children with mild head injuries—the population that creates the greatest dilemma for clinicians.