Discrepancies between preliminary and final radiology reports: Impact of training level, imaging modality, and...

Discrepancies between preliminary and final radiology reports can lead to misdiagnoses and potential patient harm. These discrepancies are especially relevant in academic hospitals where radiology residents interpret imaging during off-hours. A common quality assurance practice involves preliminary interpretations by residents followed by attending radiologist review. Understanding the factors associated with clinically significant changes can enhance patient safety and radiology education.

We retrospectively reviewed 43,272 radiology reports initially interpreted by residents and later finalized by attendings. Variables included imaging modality, report time (weekday/weekend; hour of day), year of residency, and patient characteristics (age, ICU admission status). Clinically significant changes were defined as modifications that altered patient management.

In 1,162 cases (2.69%), clinically significant modifications were made. Discrepancy rates were higher in older patients (OR 1.006 per year, 95% CI 1.004-1.008, p < 0.001), ICU admissions (4.3%, OR 1.638, p = 0.008), during early evening hours 15:00-19:00 (OR 1.382, p = 0.048) and on weekends (OR 1.302, 95% CI 1.468-1.153, p < 0.001).). CT scans had higher discrepancy rates than ultrasound (US) studies (OR 2.02, 95% CI 1.57-2.59, p < 0.001). Report accuracy improved progressively with a decrease in discrepancy rate from 3.3% in PGY-1 to 1.6% in PGY-5.

Discrepancies between resident and attending radiology reports were associated with imaging modality, reporting time, patient complexity, and resident experience. These findings support the value of structured supervision and feedback, particularly in complex cases. Attending review of preliminary reports may serve as both a clinical safety mechanism and an educational tool within radiology training programs.