STRUCTURE AND CONTENT OF 400 CT REPORTS IN FOUR TEACHING HOSPITALS USING A NEW, WINDOWS-BASED SOFTWARE TOOL

Abstract

To test a new software tool developed for analysis of radiology reports and to compare CT reports from four different teaching hospitals. Four hundred CT reports were randomly collected from four Finnish university hospitals. A Windows 3.1-based software tool was developed to make a comparative analysis of the information content of CT reports. The structure of the reports was partly analyzed manually. The new software tool greatly facilitated semiquantitative analysis of the information content of residents' and radiologists' reports. There were some local differences in the length and structure of the reports, the choice of vocabulary and the number of differential diagnoses given, and large differences in the use of an "impression" section. Thorough description of focal lesions was included in less than 50% of the reports from each of the four institutions. The variation in well-structured CT reports indicates considerable geographic differences in radiology reporting, which may reflect the long-term influences of a few teachers. Rational application of the communication standards should improve the quality of written radiology reports. Such standards should be emphasized, particularly in teaching hospitals.