interpretation of computed tomography of the head: emergency physicians versus radiologists

conclusions both eps and er residents either missed or falsely called a significant number of pathologies in their interpretations. the interpretations of eps and er residents were more sensitive and more specific, respectively. these findings revealed the need for increased training time in head ct reading for residents and the necessity of attending continuing medical education workshops for emergency physicians. results out of 544 ct scans, em physicians had 35 false negatives and 53 false positives compared with radiologist’s interpretations (p < 0.0001). em residents had 74 false negatives and 12 false positives compared with radiologist’s interpretations (p < 0.0001). background many patients are brought to crowded emergency departments (ed) of hospitals every day for evaluation of head injuries, headaches, neurologic deficits etc. ct scan of the head is the most common diagnostic measure used to search for pathologies. in many eds the initial interpretation of images are performed by emergency physicians (ep). since most decisions are made based on the initial interpretation of the images by emergency physicians and not the radiologists, it is necessary to assess the accuracy of interpretations made by the former group. objectives the objective of this study was to compare the findings reported in the interpretation of head cts by emergency physicians and compare to radiologists (the gold standard). materials and methods this was a prospective cross sectional study conducted from march to may 2009 in a teaching hospital in tehran, iran. all non-contrast head cts obtained during the study period were copied on dvds and sent separately to a radiologist, 6 emergency medicine (em) attending physicians and 14 senior em residents for interpretation. clinical information pertaining to each patient was also sent with each ct. the radiologist’s interpretation was considered as the gold standard and reference for comparison. data from em physicians and residents were compared with the reference as well as with each other and statistical analysis was performed using spss 18.5.