What are doctors thinking when you’re talking in the exam room?

Ever wonder what your doctor is thinking while taking your history? If we’re doing it right, we’re looking at you instead of a computer. We’re making appropriate eye contact while displaying welcoming body language. And we’re letting you tell your story with as few interruptions as possible. Clearly we are listening intently, but did you ever wonder what’s going through our minds while you’re speaking?

I’ve been thinking about this lately in the context of teaching medical students about history-taking. They’re being taught all the right questions to ask and how to ask them (body language, open-ended, etc.) but it often seems like they don’t know what they’re supposed to be listening for, or how to elicit the kind of information that will let them make a diagnosis (which is, of course, the necessary prerequisite for appropriate treatment).

So here’s an analogy to try and help both patients and learners better understand what’s going on inside the doctor’s head.

Imagine that someone is telling you a story. There are lots of different stories it could be, but the person has no idea which one it is. In fact, as far as he’s concerned, no one has ever heard the story before because it’s the first time he’s telling it. But he’s telling it to you because he’s hoping you’ll recognize the story, and tell him how it’s going to end — or, more importantly, intervene to change the ending, if it happens to be a story that ends badly.

What are your prerequisites? First, you need to know a lot of stories. Because you can’t help the storyteller if you’ve never heard of his story. (Even if you’ve never actually heard it, you need to have heard of it.) Next, you need to know what elements are intrinsic parts of the story, and which ones are minor details that can vary from one telling to another without materially affecting the essence of the story. There are also significant clues you can take from who’s telling you the story. Younger women, older men, little kids each seem to tell their own specific stories. Timing also matters. Some stories are told more often within several months of giving birth, or within a week of major surgery for example.

Say the storyteller begins, “Once upon a time …”

You wouldn’t immediately jump in and start asking a whole bunch of specific questions like:

  • Was there a little girl?
  • Was she in a forest?
  • Was she in a castle?
  • Was she under a spell?

No, you’d let the storyteller go on as long as possible on his own.

“Once upon a time, there was a little girl in a blue riding hood, whose mother sent her off to her grandmother’s house with a basket of goodies.”

If you’re sufficiently familiar with enough stories, you should immediately recognize that the color of the girl’s clothing may not be as important as getting her to turn around and not get any deeper into that forest until you’ve had a chance to do a scan and a Wolf-ectomy. Or at least make sure she’s got a hunter escort.

Even though the stories are basically the same, each person tells them differently. Think about how many different ways you’ve heard stories that are basically Romeo and Juliet.

As the storyteller is relating his tale, you start thinking about the various possible stories he could be telling. Once you start to recognize the story (or even what kind of story: Is this a short one where nothing much happens of consequence, or one where things don’t end well for the protagonist?), you begin to listen to specific details from one story or another. As you hear them, they lead you down the path towards figuring out the correct story. If they’re not forthcoming, you try and ask open-ended type questions to elicit them. Things like, “What kind of place was the princess living in?” as opposed to, “Was she in a forest? Was she in a town? Was there a castle?” (Analogous to asking, “What does eating do to the pain?” instead of, “Is the pain relieved by eating or does eating make it worse?” Open-ended instead of yes-no.)

Once you’ve heard a hundred people tell you the story of their gallbladder (“I’ve been getting this pain in my stomach, up high on the right, for a while now, but it’s getting worse after I eat, especially fatty foods. It nauseates me, and sometimes I throw up,”) or the story of their appendix, or the story of their hemorrhoids, or the story of the flu, or the story of their ulcer or their pneumonia or their MS or their UTI … you get the idea — you start to recognize it within a few sentences. You want to keep listening, however, because this just might be the person telling you the story of their brain tumor instead of the story of their migraine headaches. The best way to figure it out for sure is to keep listening, which means keep the storyteller talking.

There’s lots of overlap between stories. There are many conditions, like upper and lower respiratory infections, viral syndromes, gastroenteritis, pinkeye, and many other conditions which will all get better with time. Assuming you’re very confident that the story is one of the many short, simple ones with lots of overlapping features where nothing really bad happens, recommending things like rest and plenty of fluids will often result in the storyteller living happily ever after. But your patient is counting on you to correctly recognize which story they’re telling, to know how it progresses — including how it ends — and how to intervene to change a bad ending to a better one.

If you decide too soon that you know what story it is, you may stop listening for clues to other stories and miss them altogether. Then, as the story progresses (or as the disease progresses along its natural history) you may be making things worse instead of better because you’re not getting the right story. Many times, the most important clue is recognizing that there’s a significant detail that just doesn’t fit. The girl in the blue riding hood is being imprisoned in a tower with no doors and happens to have really long hair. An adult with a sore throat that began with a runny nose two weeks ago is now getting much worse; Rapunzel does not have strep.

There are two basic cognitive strategies to figuring out the story. At first, when you don’t know many stories or what you can do to modify the scary ones, you do something called “hypothesis testing.” You take the first story that pops into your head and start asking all kinds of questions specifically about that story. Once you decide that’s not it, you take the next most likely one and start specifically seeking information to confirm or deny it. And so on. Thinking about it in terms of trying to figure out which fairy tale someone is telling you helps you appreciate the inefficiency of this approach.

The strategy used by experienced clinicians is called “pattern recognition.” That’s what we mean when we tell learners to “listen to the patient,” and “the patient will tell you what’s wrong.” As we listen, we hear familiar riffs: chest pain that worsens with exertion steers us toward the heart; pain with nausea after eating pushes us towards the gall bladder. We have to keep the patient talking, but we have to know how to ferret out key details that will guide our questions when we finally get around to asking them. You just need the confidence that you will recognize the story the patient is telling you.

How do you learn these stories? Aside from hearing them from lots of different patients, of course, you can read them in medical textbooks, where they are called things like “presentation” and “natural history.” Except in this case, you know what the story is (because the title is on top of the page), as opposed to learning how to figure out which one the patient’s telling you.

These stories have also been called illness scripts. Most of the academic work on them has been done with groups of medical students in small group settings. But I believe there would be value in writing them down in transcript form, just as a hypothetical patient might relay the story, with a sidebar containing the contemporaneous text of what’s going through my head as I listen to the patient. Symptom-based as opposed to diagnosis-based, I think it would be useful to entry-level medical, nursing, and allied health care students to help make the jump to pattern recognition from hypothesis testing.

Any takers for a collaboration?

Lucy Hornstein is a family physician who blogs at Musings of a Dinosaur, and is the author of Declarations of a Dinosaur: 10 Laws I’ve Learned as a Family Doctor.

View 7 Comments >

Most Popular

✓ Join 150,000+ subscribers
✓ Get KevinMD's most popular stories