Now you see me: A pragmatic cohort study comparing first and final radiological diagnoses in the emergency department

Abstract

Objectives To (1) compare timely but preliminary and definitive but delayed radiological reports in a large urban level 1 trauma centre, (2) assess the clinical significance of their differences and (3) identify clinical predictors of such differences. Design, setting and participants We performed a retrospective record review for all 2914 patients who presented to our university affiliated emergency department (ED) during a 6-week period. In those that underwent radiological imaging, we compared the patients' discharge letter from the ED to the definitive radiological report. All identified discrepancies were assessed regarding their clinical significance by trained raters, independent and in duplicate. A binary logistic regression was performed to calculate the likelihood of discrepancies based on readily available clinical data. Results 1522 patients had radiographic examinations performed. Rater agreement on the clinical significance of identified discrepancies was substantial (kappa=0.86). We found an overall discrepancy rate of 20.35% of which about one-third (7.48% overall) are clinically relevant. A logistic regression identified patients' age, the imaging modality and the anatomic region under investigation to be predictive of future discrepancies. Conclusions Discrepancies between radiological diagnoses in the ED are frequent and readily available clinical factors predict their likelihood. Emergency physicians should reconsider their discharge diagnosis especially in older patients undergoing CT scans of more than one anatomic region.