Factors Influencing Dictation Errors in Radiology Resident Reports: A Retrospective, Cross-sectional Analysis at a...

Radiology reports serve as the primary communication tool between radiologists and referring clinicians. Dictation errors-textual inaccuracies in finalized reports-can compromise clarity and accuracy. This study evaluates the frequency and nature of dictation errors in resident-generated reports and explores associations with training level, gender, and time of year.

This retrospective review examined 600 chest radiograph reports from a large academic medical center in the United States between January 2023 and December 2023. The sample included 500 resident-generated reports (dictated by residents and finalized by attendings) and 100 attending-only reports (dictated and finalized by attendings alone). Dictation errors-defined as inaccuracies introduced during dictation, editing, or proofreading-were categorized as nonsense, omission, redundancy, or internal inconsistency and assessed for clinical significance. Two PGY-2 radiology residents performed manual review, and a board-eligible cardiothoracic radiologist adjudicated all flagged errors.

Dictation errors were more frequent in resident-generated reports (14.6% vs 3.0%; p<.0001). Reports dictated by male residents had borderline significantly higher error rates than reports dictated by female residents (16.6% vs 9.7%, p =.0495), a difference which was attenuated when accounting for report word count (OR = 1.56, p =.168). The frequency of errors increased by approximately 42% for each additional year of residency (p =.0016) or 29% for each additional year when corrected for word count (p =.0345), with the highest error rates in reports dictated by senior residents (20.6%).

Dictation errors in finalized reports may arise from factors related to resident dictation, attending review, or shared workflow pressures. Differences by postgraduate year and gender may reflect resident performance, attending oversight, or both. Further study is needed to clarify these relationships. Targeted interventions, including ongoing supervision, feedback, and use of automated error detection tools, may help mitigate error risk, and improve documentation quality in radiology training programs.