Objectives: To externally validate the Israeli Decision Algorithm for Identifying traumatic brain injury (TBI) in Children (IDITBIC), and to compare its performance as a screening tool for clinically-important TBI (ciTBI), with that of the previously validated PECARN rule.
Methods: We performed accuracy assessments of IDITBIC and PECARN rule on a large multicenter cohort of children used in the Australasian Pediatric Head Injury Rules Study. Diagnostic accuracy for detection of ciTBI was calculated.
Results: The comparison cohort included 18,913 children; 160 (0.85%) had ciTBI, 24 (0.13%) needed neurosurgery and 251 (1.3%) had TBI on CT. Applying IDITBIC would have missed 11 ciTBI patients. All the missed cases were classified as such based on the admission to hospital >2d, none of them died, needed neurosurgery, or was intubated. In children aged <2 years, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of IDITBIC and PECARN rule were [95.2%, 79.5%, 3.8%, and 99.9%] and [100.0%, 59.1%, 2.0%, and 100.0%]; respectively. In children ≥2 years, sensitivity, specificity, PPV and NPV of IDITBIC and PECARN rule were [92.4%, 75.3%, 3.1%, and 99.9%] and [99.2%, 52.9%, 1.7%, and 100.0%]; respectively. Area under the receiver operating characteristic curve values of IDITBIC and PECARN rule for children aged <2 years, and children aged ≥2 years were 0.88 and 0.77 and 0.83 and 0.72; respectively.
Conclusions: IDITBIC demonstrated high accuracy in predicting patients at risk for ciTBI. As a screening tool for ciTBI, PECARN rule showed superiority over IDITBIC but had a lower discrimination performance.
