Studies suggest that thoracic computed tomography (TCT) is superior to plain chest X-ray (CXR) in the detection of blunt chest injury. This study examined whether TCT provides additional information to routine CXR findings, whether the additional information results in a management change, and whether TCT is more useful in patients with particular mechanisms of injury. Level I trauma patients were prospectively placed into two groups. Control (CTL) group patients underwent TCT as a result of either clinical chest symptoms or abnormal CXR findings. The mechanism (MECH) group contained patients who had no thoracic signs and a negative CXR but experienced severe mechanisms of injury. TCT identified injuries not seen on CXR in 66 per cent of the CTL group and 39 per cent of the MECH group. Identification of these injuries resulted in a highly significant (P < 0.001) change in clinical management in 20 per cent of the CTL group and 5 per cent of the MECH group. TCT appears to be most helpful in the acute evaluation of trauma patients when roentgenographic evidence of chest injury exists and provides additional information impacting on the care of the patient 20 per cent of the time. In patients with severe mechanisms of injury and normal CXRs TCT expeditiously identifies occult chest injuries that require treatment in 5 per cent of this population.