Readability of the radiologic report

Abstract

To objectively assess the linguistic complexity of radiographic reports, the authors retrospectively performed readability analysis on 10,361 reports dictated by seven attending radiologists in 1989. A commercial writing evaluation program was used to determine the Flesch-Kincaid readability index of each report. Four radiographic modalities were studied: general radiography, mammography, ultrasound, and magnetic resonance imaging (MRI). Analysis of the reports demonstrated that the average readability index was consistently higher for mammography and MRI when compared with ultrasound and general radiology. Reports were categorized by radiologist and modality, and a two-way analysis of variance was performed that showed a significant difference between radiologist and modality (F = 18.6, P less than .001). To assess the relevance of this observation, 40 chest radiographic reports, randomly selected from the reports in the sample, were sent to family practitioners. The practitioners were asked to evaluate each report for clarity and the certainty of the radiologists about their findings. There was a strong negative correlation between readability and average clarity (r = -0.63, P less than .001) and readability and average certainty (r = -0.58, P less than .001). Thus, the results of this study suggest that there are differences in the linguistic complexity of radiographic reports, as measured by the readability index, among individual radiologists and different modalities. The survey data indicate that reports containing lengthy, complex sentences are viewed negatively by family practitioners and may impede effective communication between radiologists and clinicians.