Visual hindsight bias for abnormal mammograms in radiologists

Area of Science:

• Radiology
• Cognitive Psychology
• Visual Perception

Background:

• Hindsight bias, the tendency to believe one could have predicted an outcome after it occurred, impacts decision-making.
• Evidence suggests hindsight bias extends to visual perception, where prior knowledge alters how images are seen.
• This bias is relevant in interpreting radiological images, potentially affecting diagnostic accuracy.

Purpose of the Study:

• To investigate visual hindsight bias in expert radiologists interpreting mammograms.
• To determine if prior knowledge of an abnormality enhances visual perception of mammographic images.
• To differentiate visual hindsight bias from decision-level bias in radiology.

Main Methods:

• Expert radiologists reviewed unilateral abnormal mammograms.
• Confidence ratings for abnormalities were collected on a 6-point scale.
• The random image structure evolution method was used to isolate visual biases.

Main Results:

• Radiologists were more accurate in identifying abnormalities when initially shown clear images compared to degraded ones.
• This indicates prior visual experience with an abnormality enhances perception.
• The findings suggest a visual hindsight bias in radiologists' image interpretation.

Conclusions:

• Expert radiologists demonstrate both decision-level and visual hindsight bias.
• Visual hindsight bias influences how radiologists perceive abnormalities in mammograms.
• These findings have implications for medical malpractice and negligence lawsuits.