Interesting – if limited in prospective use – retrospective comparison of the New Orleans, Canadian, and NEXUS II instruments for risk stratifying adults suffering minor head trauma. Busy, urban children’s ED went through 8 years of data to find over 6000 patients with minor head trauma.
Unfortunately, they only looked at the 2,101 that received a head CT, so we lose a huge chunk of our population to “clinical judgement” that could have profoundly affected the specificity of the rules and perhaps had small effects on their sensitivity.
Full of interesting tidbits – 25% of their study population was under 2 years old, but 41% of their injuries were detected in the under 2 population. Sensitivity and specificity essentially rose and fell with the percentage of the cohort scanned – the New Orleans rule would have scanned 89% of their cohort…that had an incidence of 4.4% of intracranial injury. That made the sensitivity 96%, but the specificity 11% – and I hate to think what the specificity would have been if the other 4000 patients had been included. The Canadian Rule scanned the least, missed the most at 65% sensitivity, but achieved a 36% specificity.
But the real question is – what’s the point? The PECARN criteria get you up to ~96% sensitivity with a specificity of 53-58%. Kids aren’t small adults – especially infants, and especially in trauma. Don’t apply adult criteria in kids.