What mechanism justifies abdominal evaluation in motor vehicle crashes?

Background: Significant resources are spent on assessment of the abdomen in motor vehicle crash (MVC) occupants based solely upon crash mechanism. Most have no clinically significant injuries. We sought to define crash and occupant characteristics that would be associated with a low probability of clinically significant abdominal injury. Methods: Drivers in MVCs were analyzed from the National Automotive Sampling System from 1993 to 2001. The probability of abdominal injury with an Abbreviated Injury Score > 1 was measured over a range of crash velocities controlling for other covariates using multivariate logistic regression. A receiver operating curve was used to assess the model. Results: Age, restraint use, net change in velocity, ejection, rollover, vehicle type, other passenger mortality, and other injuries were included in the model yielding an area under the receiver operating curve of 0.948. The probability of abdominal injury increased significantly at velocities >20 km/h and reached 5.0% at approximately 30 km/h. Age >75 years old or the presence of head, leg, or chest injuries significantly increased the risk of abdominal injury, even at low velocities. Absence of injury to the head, spine, legs, or chest indicated that the abdomen was unlikely to be injured regardless of crash velocity. Conclusion: Absence of injury to the head, chest, spine, and legs indicates an extremely low probability of abdominal injury, even at high velocities. Abdominal evaluation in this group of patients for reason of mechanism alone is therefore unnecessary.