The CHIP criteria appears to be a valid tool for reducing unnecessary head CT scans in patients with mild head injury

EBM Focus - Volume 13, Issue 33

Read the full EBM Focus and earn CME credit.

Reference: BMJ. 2018 Aug 24;362:k3527 (level 2 [mid-level] evidence)

Several clinical decision tools are available for evaluation of patients with minor head injury, but two of the most widely used tools were only validated in patients with loss of consciousness (LOC) or memory loss. The CT in head injury patients (CHIP) decision rule was developed to estimate risk of intracranial trauma in those patients with head injury in the absence of LOC or memory loss, but to date has lacked external validation. To that point, a multicenter, prospective cohort study compared the four commonly used clinical decision tools CHIP, New Orleans Criteria (NOC), Canadian CT Head Rule (CCHR), and National Institute for Health and Care Excellence (NICE) guideline in patients aged 16 years and older with blunt head trauma in the past 24 hours and Glasgow coma scale (GCS) of 13-15, with or without LOC or post-traumatic amnesia (BMJ. 2018 Aug 24). The researchers compared the test characteristics of these four clinical decision tools for the outcomes of traumatic finding on CT scan and potential neurosurgical lesion.

Of the 4,557 patients included in this validation cohort, 82% underwent CT scan, 8.4% had a traumatic finding on CT, 1.6% had potential neurosurgical lesion and 0.4% underwent neurosurgical intervention for head injury within 30 days after the injury. In general, the tools that were the most sensitive for identifying any traumatic intracranial injury on CT or potential neurosurgical lesion were also the least specific, and vice versa. The NOC and CHIP criteria had the highest sensitivity (98.8% and 94.1% respectively), but low specificity (4.4%b and 21.6%) while the NICE guideline was the most specific (60.9%) but had sensitivity of only 72.5%. For the CCHR, the sensitivity was 80.3% and specificity was 44.2%. The authors estimated 96% of patients in the validation study would need CT with NOC, 80% with CHIP criteria, 58% with CCHR and 42% with NICE guidelines. They concluded the CHIP criteria best balanced the benefit of identifying findings on CT with the harms of both false negative and false positives.

This study was limited by the significant number of participants who did not have a CT and the need to impute missing data. It is worth noting that both CCHR and NICE criteria have had higher sensitivity in previously tested cohorts. Finally, given the very low rate of lesions that would warrant neurosurgical intervention, all 4 decision rules had negative predictive values > 99% meaning any of them can be used with reasonable confidence.

Focus Point: For adults and older adolescents with head injury but without LOC or memory loss, the CHIP criteria appears to be a valid decision tool that balances risk of missing a clinically important finding with the need to limit the overuse of head CT.

For more information, see the topic Decision rules for computed tomography in adults with head injury in DynaMed Plus. DynaMed users click here.