Diagnostic error remains to be a major setback in improving the quality of radiology and overall interpretation of imaging examinations. What consistently induces diagnostic error is a radiologist's cognitive bias, according to a recent article published online March 15 in the American Journal of Roentgenology, as approximately 75 percent of medical malpractice claims against radiologists are related to diagnostic error.  

Researchers Jason Itri, MD, PhD, from Wake Forest Baptist Medical Center in North Carolina and Sohil Patel, MD, at the department of radiology and medical imaging at the University of Virginia Health System assert that radiologists rely on heuristic principles, or mental shortcuts, to reduce complex tasks into simplify judgmental operations in medical imaging. However, the researchers wrote, although heuristics based on experience and assumption can be helpful for interpretation purposes, they can also produce cognitive bias and significant error.  

"An understanding of the causes of cognitive biases can lead to the development of education content and systematic improvements that mitigate errors and improve the quality of care provided by radiologists," the researchers wrote.  

Cognitive Bias and Heuristics  

According to Itri and Patel, there are two types of cognitive biases that exist in radiology: perceptual errors and interpretive errors. Accounting for 60 to 80 percent of errors in imaging, perceptual errors occur when an abnormality is identified in a diagnostic imaging but remains unseen by the radiologist. Interpretive errors make up the remaining 20 to 40 percent of imaging errors when an abnormality is identified on an image, but its diagnostic meaning is incorrectly interpreted.  

Radiologists may take advantage of heuristics to minimize delay, cost and uncertainty in clinical decision making by "assessing probabilities and predicting values into simpler judgmental operations, the researchers explained. However, they also may also be the source to diagnostic error.  

"Awareness of the spectrum of cognitive biases is an important step toward a comprehensive strategy to learn from diagnostic errors and ultimately improve patient care," according to the researchers.  

Cognitive Biases Contributing to Imaging Error  

Itri and Patel discuss ten types of cognitive bias that contribute to diagnostic error in medical imaging, which include: 

• Availability bias – diagnostic assessments influenced by past experiences  
• Alliterative bias – when one radiologist's diagnostic judgement influences another radiologist  
• Anchoring bias – settling on a diagnostic hypothesis based on previously gathered information  
• Framing bias – a tendency to be influenced by how a question is asked or a problem is presented  
• Attribution bias - attributing characteristics to a certain image, patient, or case because of the class it belongs to  
• Blind spot bias – failing to recognize certain characteristics in an image, potentially leading to misdiagnosis  
• Regret bias – overestimating the likelihood of a disease because of the unwillingly acknowledging an adverse outcome from a failure to previously diagnose that disease  
• Satisfaction of search – stopping a visual search after identifying an abnormality that can explain the patient's symptoms and the radiologists then already satisfied with a determined diagnosis  
• Hindsight bias – overestimating the predictability of an event after the event is known  
• Scout neglect bias – neglecting scout images on a cross sectional imaging study or exam   

A compartmental strategy that requires identifying errors, analyzing errors for systematic causes and biases, using what is learned from errors to develop educational content and systematic improvements and developing a culture which promotes reporting and learning from errors may reduce the effect of cognitive biases to learn from diagnostic errors, the researchers wrote.  

"Implementing an educational curriculum for trainees, radiologists and staff as a foundational element of this strategy, because awareness is arguably one of the most effective told we have to reduce errors," the researchers wrote. "Awareness that errors occur and are often the result of systematic causes (rather than person failures) is a requisite to having a robust and effective reporting system to identify errors."