Error in Radiology: An Evolutionary Inevitability?
Error in radiology reporting is a seemingly inevitable part of radiological practice, and a phenomenon that is ubiquitous, often without any respect for radiologist experience or expertise.
It is also a common phenomenon: although studies from the plain film era suggested an error rate of the order of 3-5%, more recent studies suggest that this rate is closer to 30% in the modern age of high-volume cross-sectional imaging – a rate of error that is almost inconceivable to non-radiologists.
The advent of digital image storage also multiplies the discovery of error in hindsight, and this is a trend that is likely to be further compounded by the increased volume of scanning and of stored images.
It is becoming apparent that a large proportion of errors in radiology reporting occur not because of mis-interpretation of the significance or otherwise of certain appearances, but because of a failure to perceive certain aspects on the scan. For example, it appears to be more often the case that mistakes are made because unsuspected, often second pathologies are not perceived.
It is also increasingly understood that as well as environmental factors, such as noise, interruptions and fatigue, cognitive biases play a very important part in error in radiology. These include framing bias (cognitive short-cuts introduced by knowledge of the referring clinical information or of previous reports) and ‘satisfaction of search’ (the premature termination of visual search following the identification of a primary pathology). Appreciation of the effect of such cognitive biases has been recently popularised, for example, by Daniel Kahnemann, in his book “Thinking Fast, Thinking Slow”, and by the demonstration of inattentional blindness in the “Invisible Gorilla” video on YouTube. All of these studies emphasise the innate human propensity to filter out extraneous visual information when a primary search activity is being undertaken. It may well be that our failure to perceive unexpected, non-searched for abnormalities is the result of basic human evolution.
back to all news