You have arrived at your doctor’s surgery or the consultant’s clinic. You have the story of your symptoms roughly ready in your head: for example where the pain is, when it began, how it has changed over time. You will likely have some suspicion of what might be causing it and perhaps some subconscious fear of a specific serious illness.
The consultation starts. Ideally the doctor will not interrupt before you have finished so that you get to outline a framework of the reasons you have sought medical help. Then you wait for the ball to be lobbed back over the net at you, usually in the form of a prolonged rally of questions and answers as the physician tries to make sense of your unique story.
What exactly is going through your doctor’s head at this point? At one level she is formulating a number of possible diagnoses, and asking questions, the answers to which help to provisionally rank the diagnostic slate. By adding your age, sex, past medical history and other elements into the equation, the diagnostic algorithm begins to take shape.
At this point a good physician will run her thoughts by you, inviting further clarification of her summary. She will then rank the most likely diagnoses and outline some investigation or treatment options. Cue a further exchange of views before a way forward is agreed by both doctor and patient.
You may well ask, why, as an opening gambit in this modern technological age, doctors don't simply order a wide battery of tests and scans? Dr Siddartha Mukherjee, in his book The Laws of Medicine: field notes from an uncertain science, makes the crucial point that "a test can only be interpreted sanely in the context of prior probabilities".
This apparent paradox is based on the fact that every test in medicine has a false positive and a false negative rate. In a false positive, a test is positive even when the patient does not have the disease. With a false negative test, a patient tests negative but actually has the abnormality being tested for. Without getting bogged down in statistics, essentially a doctor increases the likelihood of a test being accurate when she uses this “prior knowledge” to decide when and what type of test to order in each case.
In the words of a group of professors at Stanford Medical School, “ technology drives diagnosis, but it often merely substitutes our fears of uncertainty with delusions of certainty”.
Without wishing to promote a return to medical paternalism, the prior knowledge approach to diagnosis is something that “old school” medics tend to do very well. But the siren call of new technology is always there, with an understandable attraction for the neophyte. Rather than pitting the accuracy of bedside evaluation against that of technology, physicians who judiciously combine both are more likely to achieve better outcomes for their patients.
Jerome Groopman argues that the frequency of medical mistakes, as well as their severity, can be reduced by understanding how a doctor thinks. "Cogent medical judgments meld first impressions . . . with deliberate analysis. This requires time, perhaps the rarest commodity in a healthcare system that clocks appointments to minutes."
Can we, as patients, help reduce flawed reasoning by our doctors? In his book, How Doctors Think, Groopman suggests three questions we might ask when an initial diagnosis hasn't worked out:
“What else could it be?” allows a doctor to acknowledge the uncertainty inherent in medicine. “Is there anything that doesn’t fit?” will prompt your doctor to pause and encourage her to think laterally. “Is it possible I have more than one problem?” helps challenge the conventional medical school teaching that a doctor should seek one cause for a patient’s many complaints.
Doctors, for their part, would do well to be mindful of Mukherjee’s first law of medicine: “a strong intuition is much more powerful than a weak test.”