Analysis of Report Addenda as a Novel Approach to Characterization and Quantification of Errors in Diagnostic Radiology

Abstract

Diagnostic errors are common in radiology and can have a significant negative impact on patient care. Identifying the types of errors that occur can improve our understanding of the root causes and suggest pathways for improvement. To date, studies have focused on errors occurring in difficult cases, which have higher error rates but are not representative of the errors occurring in hospitals on a daily basis. In most hospitals, radiologists attach addenda to reports when an error needs correction. Therefore, addenda are markers for errors and provide a more complete, non-biased picture that may be more relevant to improving outcomes. Using report addenda at a large university hospital we analyzed the types of errors in 1,195 cases and found that radiology studies at our hospital have an error rate of 0.9%. Our results demonstrate that in daily radiology practice, errors of poor communication occur most frequently (36%), followed by under-reading (23%), procedure-related (20%), insufficient history (15%), over-reading (5%), and poor technique (0.5%). When analyzed by modality, most errors occurred in interventional procedures, followed by PET, MRI, and CT. Errors of communication are often preventable and suggest a clear area for intervention. More broadly, our success using addenda to study clinical errors demonstrates the feasibility of this novel approach, which would be reproducible at virtually all institutions.