Medical Errors and the Way Doctors Think
By Michael Astion , MD, PhD
Jerome Groopman, MD, is professor of medicine at Harvard University, a columnist for the New Yorker, and acclaimed author of the best seller “How Doctor’s Think”. In his plenary lecture at the AACC Annual Meeting in July, Dr. Groopman talked about errors in thinking that can harm patients and how to overcome such errors.
Available data suggests that misdiagnoses occur in 15% or more of clinical cases, but overall there is very limited data on the frequency of misdiagnosis in medicine. According to Groopman, underlying these misdiagnoses are three common and related errors in thinking: availability; attribution; and anchoring.
Availability, also known as confirmation bias, refers to the bias caused by the experiences most readily recalled by a person. One example that Groopman provides in his writings is the case of an ER doctor who made many diagnoses of viral pneumonia over a short period of time during an epidemic in the community. One day, after making several of these diagnoses, a woman came in with some signs and symptoms of pneumonia; however, she also had findings that were not a good fit. For example, physical exam and chest X-ray were not consistent with viral pneumonia nor was her white blood cell count, which was within the reference range.
After the patient was admitted to the hospital for viral pneumonia, the physician on the hospital ward ordered additional tests and reconsidered the initial ER data, which included rapid breathing and acidemia. The hospital physician determined this was a straightforward case of aspirin toxicity. The ER doctor, on the other hand, had fallen into what Groopman calls a “thinking trap.” He was overly influenced by what he had recently seen—viral pneumonia— when there was actually a fair amount of data suggesting aspirin toxicity.
Attribution errors are based on prejudices about what is usually true.To illustrate this type of error, Groopman described the case of a very fit man in his 40s with chest pain. The patient was ruled out for myocardial infarction (MI) and discharged from the ER to home after having a negative EKG, one negative cardiac enzyme test, and a normal physical exam. The next day, however, the man was readmitted to the hospital with an MI. Here, the physician’s prejudice, which was based on the man’s rigorous physical appearance and the initial data, led him to conclude what would usually be true. He assumed the patient was not having an MI, but probably had indigestion or a strained a muscle.
Other examples of attribution include cases where a pheochromocytoma was missed in a middle-aged woman who was falsely diagnosed with symptoms of menopause and an alcoholic who was wrongly diagnosed with alcoholic cirrhosis when he had Wilson’s disease. In both these cases, the physicians attributed the patient’s presentation to common causes but failed to look deeper for less likely causes.
Anchoring is a thinking trap in which your thoughts get anchored to your first impression and then the search for a diagnosis is overly influenced by that impression. The anchor limits the ability to pursue alternative explanations and increases the likelihood that the first impression will be confirmed, even when it is wrong. Groopman said anchoring is related to attribution and availability since either of those thinking errors can lead to the wrong diagnosis, which acts as the anchor.
Groopman emphasized that thinking traps can be avoided by asking the following related questions about a patient’s presentation:
- What else can this be?
- Is there some other explanation for these data?
- Could two things be going on at once?
These questions can help clinicians get over the problem of “satisfaction of search” that closes the mind to alternative explanations for the patient’s condition. However, Groopman did not advocate using these questions to initiate mindless testing for every possibility. Rather, he proposes using these techniques to force the clinician to view the patient from another point of view before reaching a conclusion. This new viewpoint may lead to carefully-chosen, additional tests that are likely to be useful for reaching the correct diagnosis.
- Groopman J. What’s the Trouble? How Doctors Think. The New Yorker. January 29, 2007. Available online. Accessed August 25, 2009.
- Groopman J. How Doctors Think. Houghton Mifflin (Boston), 2007. 305 pp.
- JeromeGroopman.com, the official website of Dr. Jerome Groopman. Accessed August 28, 2009.