Pelvic CT Imaging in Blunt Abdominal Trauma

Abdominopelvic computed tomography (CTap) utilization rose significantly in blunt trauma patients over the last decade but failed to reduce mortality or missed injury rates. Primary Hypothesis: In stable, alert patients (GCS \> 14) without clinically evident fractures of the pelvis, hip or lumbar spine, CT abdomen (CTa) alone (diaphragm to iliac crest) identifies intra-abdominal injury (IAI) with an accuracy that is not inferior to routine CTap (diaphragm to greater trochanter) with a clinically relevant reduction in radiation exposure. Primary Aim: Compare the accuracy of CTa alone versus CTap to detect IAI in two age groups: ages 3-17 years and 18-60 years. Blunt trauma patients requiring CTap will be enrolled. Data obtained prior to CT imaging include demographics, injury mechanism, exam and pelvic radiograph findings and FAST results from the trauma evaluation. The original CTap will be digitally reformatted to create matched pairs of de-identified CTa and CTap studies. A board certified study radiologist, blinded to the original CT and clinical outcome, will interpret the CTa studies in injured patients. McNemar's chi-square test will be used to evaluate the null hypothesis for injuries in matched pairs assuming no difference for uninjured patients. Te the test characteristics of the CTa versus CTap will be determined. The reference standard will include initial radiology reports, operative reports, and 7-day medical record review. If the upper limit of the 95% confidence interval for the difference in the performance (accuracy) of CTap and CTa alone is less than 3%, the conclusion will be that CTa alone is not inferior to CTap to diagnose IAI. Secondary Aims: The mean effective doses of radiation will be calculated and compared with reductions up to 50% expected. The study will determine if physicians' pretest probability accurately identifies clinically significant pelvis, hip and lumbar spine fractures (CTp indications) in two age groups: ages 3-17 years and 18-60 years. Using a gestalt pretest probability of ≤ 2% as "negative for injury", and a pretest probability \> 2% as "positive for injury", the test characteristics of physician estimation will be determined. The test characteristics and interobserver agreement (Cohen's kappa statistic) of exam findings expected to predict fractures of the pelvis, hip and lumbar spine will be reported separately.