Radiology resident competency in diagnosing non-traumatic musculoskeletal conditions: A simulation-based assessment using WIDI SIM.

Abstract

Purpose: To evaluate radiology resident performance in diagnosing four non-traumatic musculoskeletal (MSK) pathologies that have historically yielded low scores on the Wisdom in Diagnostic Imaging Emergent/Critical Care Radiology Simulation (WIDI SIM).

Materials and methods: This multi-institutional, retrospective study analyzed WIDI SIM data collected from 2015 to 2021. A total of 351 radiology residents (R1-R4) interpreted 65 de-identified imaging cases, four of which focused on septic arthritis/osteomyelitis of the shoulder, septic arthritis/osteomyelitis of the hip, acetabular neoplasm, and Legg-Calve-Perthes disease. Each case was scored using a standardized 10-point rubric (0-2 = critical error, 3-6 = problematic omissions, 7-10 = effective report). Scores were further categorized into observational (missed findings) and interpretive (incorrect conclusion despite correct identification) errors. The Kruskal-Wallis test with Dunn's multiple comparisons was used to assess performance differences across postgraduate years.

Results: Among these four MSK pathologies, only hip osteomyelitis demonstrated a statistically significant difference across training levels (p = 0.0063), although no specific pairwise comparisons were significant. Average scores remained relatively low across all cases, with observational errors surpassing interpretive errors in frequency.

Conclusion: Radiology residents struggled to accurately diagnose non-traumatic MSK pathologies in a simulated on-call setting, predominantly due to missed imaging findings. Implementation of enhanced training strategies, such as targeted case review, high-yield simulations, and systematic visual search protocols, may improve MSK diagnostic competency and reduce the risk of clinically significant oversights.