Interpretation of chest radiographs of patients with thoracic trauma: influence of the experience of the interpreter.

Objective: To compare the quality of interpretation of chest radiographs (C Rs) by physicians of different levels of experience, with reference to data obtained from helicoidal computed tomography (HCT). Study design: Prospective observational study. Material: CRs of 50 thorax trauma patients as recent to HCTs as possible ob tained within the 48 h following admission to the intensive therapy unit. Method: CRs were analyzed according to a grid by observers included in one of the four groups: residents in anaesthesiology (n = 6), residents in radi ology (n = 3), senior anaesthesiologists (n = 5), and senior radiologists ( n = 3). The inter-observer agreements, the specificity and sensibility of e ach group with reference to HCT, and their global performances were assesse d. Results: Inter-observer agreements were poor and sensitivities low in compa rison to specificities. No group of observers performed better than another one. These data substantiate the low sensitivity of CRs in comparison to H CT and show that the quality of interpretation of CRs of thorax trauma pati ents is not influenced by physicians' experience. (C) 1999 Elsevier, Paris.