Français

Errors and matters of judgment


Even good doctors make mistakes

Pitfalls of looking back


Hindsight bias

Hindsight is 20/20. Knowing an undesirable and unexpected clinical outcome has occurred increases the belief that it was predictable, therefore preventable and related to carelessness or poor clinical care.

Opens in popup

Some medical conditions are difficult to diagnose. Looking back on previous assessments of the patient, the symptoms and signs that signal the final diagnosis stand out clearly and now may seem obvious. However, those assessing the patient earlier had the challenge of picking these signals out from the totality of all symptoms and signs experienced by the patient at that time.

We can try to avoid hindsight bias by considering how factors such as the natural evolution of a disease, additional clinical information, and improvements in the technology used for clinical testing can affect retrospective assessments.

Case: A 56-year-old woman with breast cancer
Mammogram image

Background

A breast nodule is seen on a screening mammogram and investigations prove cancer.

The radiologist reviews the patient's previous mammograms and notes microcalcifications and architectural distortion are present on a mammogram done two years earlier.

Outcome

Not all breast cancers can be identified on mammograms.

Some mammograms show very subtle or even no detectable signs of malignancy, for example, a mass, microcalcifications, architectural distortion, or an asymmetric or developing density.

Some malignancies are completely obscured by overlying and highly dense glandular and fibrous tissue.

Lessons learned

It is common for radiologists to compare the current mammogram with one or more prior studies.

After a diagnosis of cancer, changes in the breast architecture that reflect the development of the cancer can be more easily identified.

Knowing where to look, one may, in hindsight, identify subtle changes in breast architecture that, at the time, were below the threshold of detection of a competent radiologist.

Lessons learned continued

This is not an indication of negligence (professional civil liability in Québec) or a measure of poor performance.

The reality is that a disease must progress to a certain point to be recognizable within the sensitivity of the diagnostic imaging used.

Sometimes, having done your best to eliminate hindsight bias, you may believe a lesion seen on a previous study should have been "obvious."

An obvious missed lesion is generally considered to be one that most radiologists would have recognized on the previous study.

Previous
Next