External validation of a five-variable clinical prediction rule for identifying children at very low risk for intra-abdominal injury after blunt abdominal trauma.

Abstract

A clinical prediction rule was previously developed by the Pediatric Surgery Research Collaborative (PedSRC) to identify patients at very low risk for intra-abdominal injury (IAI) and intra-abdominal injury receiving an acute intervention (IAI-I) who could safely avoid abdominal computed tomography (CT) scans after blunt abdominal trauma (BAT). Our objective was to externally validate the rule. The public-use dataset was obtained from the Pediatric Emergency Care Applied Research Network (PECARN) Intra-abdominal Injury Study. Patients 16 years of age and younger with chest x-ray, completed abdominal history and physical examination, aspartate aminotransferase (AST), and amylase or lipase collected within 6 hours of arrival were included. We excluded patients who presented greater than 6 hours after injury or missing any of the five clinical prediction variables from the PedSRC prediction rule. We included 2,435 patients from the PECARN dataset, with a mean age of 9.4 years. There were 235 patients with IAI (9.7%) and 60 patients with IAI-I (2.5%). The clinical prediction rule had a sensitivity of 97.5% for IAI and 100% for IAI-I. In patients with no abnormality in any of the five prediction rule variables, the rule had a negative predictive value of 99.3% for IAI and 100.0% for IAI-I. Of the "very low-risk" patients identified by the rule, 46.8% underwent abdominal CT imaging. A highly sensitive clinical prediction rule using history and abdominal physical examination, laboratory values, and chest x-ray was successfully validated using a large public-access dataset of pediatric BAT patients. Epidemiologic/prognostic study, level III; therapeutic care/management study, level IV.