| Document
|
Author Rachel Magennis |
| Abstract Radiology is prone to cognitive bias. High workload, interruptions, pattern-recognition fatigue, and time pressures create situations where anchoring, satisfaction of search, and premature closure can occur. Radiology Events and Learning Meetings (REALMs) provide a forum for sharing discrepancies, but their value depends on psychological safety. Without a blame-free environment, clinicians cannot fully explore how bias influences decisions. This work examines how human-factors-informed REALM adaptations can improve bias awareness, strengthen reflective decision-making, and enhance patient safety. Approach We reviewed REALM processes across a multi-site radiology department using a human-factors framework: * Identifying bias-prone moments (overnight reporting, complex cases, mismatch between clinical request and findings). * Observing barriers to open discussion, including hierarchy, fear of judgement, and limited confidence among staff unfamiliar with structured feedback. * Introducing parallel forums (Reporting Radiographer REALM, Resident REALM) to create psychologically safe peer-level spaces. * Embedding bias-spotting prompts to normalise discussion of cognitive shortcuts. * Collecting feedback on whether these adaptations encouraged open discussion of errors and decision-making. Findings Smaller peer-level REALMs made it easier to discuss bias explicitly. Participants reported increased readiness to reflect on anchoring, automation bias, and satisfaction of search. These forums built confidence, enabling richer discussion in the main REALM and clearer identification of system factors affecting errors. Key Takeaways * Reflection on errors must include bias awareness. * Psychological safety is essential. * Parallel REALMs are optional but beneficial for large or multi-site departments. * Bias vocabulary normalises discussion of shared cognitive tendencies. * Human-factors-informed REALMs support safer systems and more resilient decision-making. |
