Rule No. 18: If you can’t figure out the diagnosis in two minutes, you won’t figure it out for at least two days.

The culmination of the art of the practice of Medicine is to arrive at the correct diagnosis with the least amount of data. There comes a point in each medical career where the best physicians develop a feel, a sixth sense, for where the disease really lies. Once you know how and where to look, you don’t really need a lot of “signs” to successfully track it down. That’s not to say that there are not diagnoses that are bizarre and rare. My impression has been that such diagnoses take a lot of time to nail down. So, by and large, most diagnoses will just jump right into your head. When they don’t, there’s a temptation to start calling specialists to help divine what is going on. I choose consultations carefully. I want physicians who will complement my strengths, who’ll think differently from me. Also it’s been said: “The more doctors, the sadder the prognosis,” so keep the consultants few.

I once had a fellow faculty member in the Psychology Department who had performed detailed studies about how quickly physicians arrived at clinical diagnoses. He told me his original research hypothesis was that successful physicians sifted through data, as if they were running along a computer algorithm, logically narrowing the data down into a final diagnosis--precisely how computer software attempts to arrive at a diagnosis is simulation exercises.

Contrary to everything he initially expected, he found that master clinicians leap intuitively, virtually instantaneously, to a provisional diagnosis. They then see what bits of data will support it—not eliminate it. Only when there are data points that are too significant, or too many data points of lesser value stacking up against the first guess, will the doctor go to a new diagnosis and subject it rapidly to the same assessment.

At first glance, the differences between these two “thought processes” may seem to be splitting hairs but it isn’t. A computer starts with all the data available in order to be able to narrow down to all the possible diagnoses. It conserves multiple diagnoses until the data eliminates them, one by one. The master clinician jumps to a single diagnosis and then picks data that will support it. This sounds to me to be precisely how the internal process works. So if, as the doctor, your brain does not jump real quickly to a provisional diagnosis, you’re probably not going to narrow anything down until you and your co-workers have spent a lot of time collecting more data.