Friday Reflection 48: Linguistics, Diagnosis, and Medical Error

FH is a 66-year-old woman who comes in for an urgent visit because she has been feeling woozy for two days. She is very anxious, almost distraught, because she thinks these symptoms are the same as the ones that her sister had before she died of a hemorrhagic stroke.

A few years ago, a team building exercise was proposed at a meeting I was attending. To say I hate team building exercises is a gross understatement.¹ I usually run for the door when these are suggested. On this day, I was too slow. For the exercise, I sat back-to-back with a partner who looked at a picture projected onto a screen. I could not see the picture. He described the image, and I had to draw what he described. After 5 minutes, I shared my drawing, and we discussed what worked and what didn’t.

Recently, I was at the Art Institute of Chicago, one of my favorite places on Earth, preparing to help lead a group of medical students around the museum. Our guide described a similar exercise while looking at a painting of a woman in mourning. Because my mind was on medicine, it struck me how similar this exercise is to what I do in clinic.

All diagnostic inquiries start with a patient experiencing a symptom. The symptom is a kind of platonic truth. What can make the search for an accurate diagnosis difficult is that a doctor seldom really has access to this truth. The doctor does not see or feel the symptom. Instead, the patient is asked to translate a sensation into language. Sometimes, the patient’s linguistic abilities are inadequate for describing the symptoms. Sometimes, our language itself is not up to the task.

Often there are issues working against the patient accurately describing his or her symptoms. The patient is anxious, in pain, exaggerating or minimizing symptoms, being rushed, or distracted.

No one can say if a patient is poorly describing his or her symptoms; that would be like telling someone that their description of red is incorrect.

FH describes her symptoms as wooziness. The doctor seeing her, Dr. S, not having a differential diagnosis for wooziness, asks her, “What do you mean woozy. FH says, “I feel floaty, foggy, out of it, off kilter.” FH is already getting a little exasperated. She is worried she might be having a fatal stroke.

To make a diagnosis, a doctor must characterize the concern, translating the patient’s words into a symptom with an established differential diagnosis and an associated diagnostic approach. This is where many diagnostic errors occur. This might happen if the doctor is not listening. But it also might happen if the doctor mischaracterizes what the patient is feeling because of how the patient reports the symptom. When that happens, the doctor begins evaluating a symptom that is not actually present.

The approach to the dizzy patient should begin with the doctor asking, “What do you mean dizzy?” and then just sitting quietly while the patient describes the dizziness. This question is supposed to force the patient to characterize the dizziness as vertigo, orthostasis, disequilibrium, or non-specific dizziness.² When Dr. S asked, “What do you mean by woozy?” she had decided that woozy meant dizzy and proceeded as if FH had complained of dizziness.

The clinical interchange has just started and already the patient has translated her symptom into language and Dr. S has translated that into a medically useful symptom.

After hearing wooziness described as “floaty, foggy, out of it, off kilter,” Dr. S. had had it with open ended questions. “When you feel woozy, does it feel like the room is spinning? Or does it feel like you are going to faint, you know like when your vision grays out? Or do you feel off balance, kind of drunk.”

FH answered, “Yes.”

At this point, we have a patient who is terribly worried about her condition and a doctor who is likely reconsidering her decision to come to work today.

In my experience, this juncture is not uncommon. A patient is having symptoms that need to be addressed. The way these symptoms are being presented linguistically is not leading the doctor to a familiar, workable symptom. Dr. S has tried to shoehorn woozy into the diagnostic rubric for dizzy and, not surprisingly, has gotten nowhere.

OK, tell me exactly what you were doing when you first got woozy?” asks Dr. S.

“I had just woken up. I rolled from my left side to my right to grab my phone to check the time and then I just about lost it. I mean really lost it. I was woozy AND nauseated.”

Dr. S. got really lucky. Although her interpretation of woozy as dizzy failed in her first two questions, she stuck with it with one more question. She hit on a suggestive answer, something that sounds like benign, paroxysmal, positional vertigo, BPPV. She performs the Dix Hallpike Maneuver and FH screams out. She has the most striking rotatory nystagmus Dr. S has ever seen.

“Are you feeling the wooziness?”

“Yes, this is exactly the sensation.”

At this point, the symptom has become a visible, objective sign.

What to take from all this? We always need to remember that reported symptoms are translations, one step removed from what is bringing a patient in. Unless you are lucky enough to be a dermatologist, when you can actually look at the problem, seeds for medical errors are sown as soon as a patient describes, translates, his or her symptom. The less specific the symptom, the more likely it is that the doctor will proceed down the wrong path. Acute onset pain at the base of the great toe might be reported as aching, burning, or searing, but you’re likely to end up thinking about gout.

Fatigue, on the other hand, might be describing tired, or weak, or sleepy, or short of breath. The differential diagnoses for those four translations probably includes every known diagnosis.

1. I also hate breakout sessions and role plays.

2. Dr. Martin Samuels, who delivered the best dizziness talk I ever heard, described non-specific dizziness as “dizzy dammit” -- as in “I am just dizzy, dammit, stop asking me these stupid questions.”

Photo Credit: The Art Institute of Chicago