When Experience Helps—and When It Gets in the Way

Walking through an amusement park, I often pause at the easels of the sketch artists. With a blank sheet of paper, a handful of colored pencils or chalk, and perhaps ten seconds spent studying the stranger sitting before them, they can produce a recognizable likeness in five minutes. The subject smiles, pays the artist, and walks away with a portrait suitable for the refrigerator door.

I am not an artist, but I imagine the work of an accomplished portraitist is quite different. The goal is no longer simply to capture the most recognizable features of a face. The artist studies the subject more carefully—perhaps learning something about the person before putting brush to canvas. Colors are mixed, brushes selected, shadows considered. The finished portrait attempts to reveal something deeper than a likeness.

Over the course of my medical career, I watched physicians—both trainees and those long established in practice—approach a new patient in much the same two ways.

Sometimes we are sketch artists. At our best, however, we should aspire to be portraitists.

The Five-Minute Diagnosis

Consider a patient who comes to the clinic because of a persistent cough.

An experienced physician has seen hundreds, perhaps thousands, of patients with this complaint. That experience is enormously valuable. The physician already knows the common explanations and can quickly begin sorting among them.

“Did you have a respiratory infection before the cough began?”

“Do you ever wheeze or become short of breath with exercise?”

“Do you feel drainage in the back of your throat?”

A few affirmative answers may quickly suggest a plausible explanation. Perhaps the patient recently had a viral respiratory infection and has been left with unusually sensitive airways. A working diagnosis is established, an inhaler is prescribed, and the patient is reassured that the cough will probably improve with time.

The sketch is complete.

And it may be correct.

But is it a sufficiently accurate portrait of the patient's illness?

That is where the very experience that makes physicians efficient can sometimes get in their way.

When Experience Becomes a Shortcut

Experience teaches physicians to recognize patterns. That ability is indispensable. An experienced clinician can often recognize in minutes what once required much longer as a trainee.

But pattern recognition carries a risk. If a patient looks sufficiently like the last hundred patients with the same complaint, it becomes tempting to assume that the rest of the picture will look the same as well.

The physician begins with the common causes of cough and asks questions designed to distinguish among them. Once the answers seem to support one of those possibilities, the diagnostic search may begin to narrow.

In doing so, the physician has behaved much like our amusement-park sketch artist: identify the prominent features, put them on paper, and produce a recognizable picture as efficiently as possible.

In medicine, however, a recognizable picture is not always the complete picture.

This is where physicians can fall victim to a cognitive error known as premature closure.

Premature closure occurs when a physician becomes satisfied with a diagnosis before adequately exploring reasonable alternatives. The problem is not that the first diagnosis is necessarily foolish. Quite the opposite—it is often highly plausible. The danger comes when that plausibility discourages further inquiry. New information may then be interpreted to support the favored diagnosis, while details that do not fit receive less attention.

A familiar illness can thereby conceal an unfamiliar one.

One way to resist this tendency is to ask two questions before considering the diagnostic process complete:

Does everything I know about this patient fit my diagnosis?

And:

What else could explain what I am seeing?

The difficulty lies partly in those five words: everything I know about this patient.

The Questions We Don't Have Time to Ask

Learning enough about a patient to paint a full diagnostic portrait takes time, and time has become one of the scarcest commodities in clinical medicine.

A patient with a cough could generate dozens of potentially useful questions.

“Do you use a humidifier, or have you noticed mold or persistent dampness in your home?”

“Do you keep birds?”

“Does your work expose you to wood dust or other inhaled materials?”

“Do you have heartburn, or is the cough worse after meals or when you lie down?”

“Have you recently started a new medication?”

Each question opens another diagnostic door. Bird exposure might raise concern about hypersensitivity pneumonitis. Certain occupational exposures can cause airway or lung disease. Gastroesophageal reflux can contribute to chronic cough. ACE inhibitors are well known for producing cough in some patients.

Even the physical examination can contain surprises. Stimulation of the external auditory canal—including, occasionally, something as innocuous as hair or cerumen touching the canal—can trigger cough through the auricular branch of the vagus nerve, sometimes called Arnold's nerve.

No physician can pursue every conceivable possibility during every brief office encounter. Nor should they. Medicine requires judgment about which questions matter most.

But therein lies the dilemma: How does a physician remain efficient without allowing efficiency to become premature closure?

Perhaps the First Diagnosis Doesn't Have to Be the Last

One solution may be to change what we expect from the first visit.

The goal of an initial encounter need not always be to arrive at the diagnosis. Sometimes it should be to establish a provisional diagnosis—a hypothesis that is reasonable but remains open to revision.

The physician's first responsibility is to determine whether the patient is safe. Are there warning signs suggesting a serious illness? Is there a diagnosis that cannot safely wait? Does the patient require immediate testing, treatment, or referral?

If those concerns have been adequately addressed, the physician may reasonably tell the patient, in effect:

Here is what I think is most likely, but I am not yet certain.

That sentence changes the nature of the diagnosis. It becomes the beginning of an investigation rather than the end of one.

Let the Patient Help Paint the Portrait

Technology could make this approach even more useful.

Imagine that the patient with cough leaves the first visit with a structured questionnaire specifically designed for chronic cough. At home, without a physician watching the clock, the patient can think carefully about questions involving medications, occupational exposures, pets, environmental conditions, reflux symptoms, nasal symptoms, previous respiratory illnesses, timing of the cough, sputum, exercise, sleep, and other potentially relevant clues.

The purpose would not be to have a questionnaire make the diagnosis. It cannot replace a thoughtful physician.

Instead, it would gather information.

Patients often remember important details only after leaving the doctor's office. A structured questionnaire gives them another opportunity to think about their illness and provides the physician with information that there simply wasn't time to collect during the initial encounter.

The patient then returns for a planned diagnostic reassessment.

At that visit, the physician does more than ask, “Is the treatment working?”

The better questions are:

Does what I have learned since our first visit strengthen my original diagnosis?

Does anything weaken it?

Is there now another explanation I should consider?

That second encounter deliberately reopens the diagnostic process. In doing so, it creates a defense against premature closure.

From Sketch to Portrait

Questionnaires have obvious limitations. Patients can misunderstand questions. They may overlook important symptoms or attach significance to things that are medically unimportant. A checklist can never reproduce the give-and-take of an experienced physician taking a careful history.

And no questionnaire should delay the evaluation of a patient whose symptoms suggest a potentially serious illness.

But for a stable patient with a symptom that does not require an immediate definitive diagnosis, a two-stage approach deserves consideration: first, identify danger, develop a provisional diagnosis, and begin reasonable management; then gather additional structured information and deliberately reconsider the diagnosis.

Paradoxically, this may be more thoughtful than trying to squeeze an exhaustive history, examination, diagnosis, treatment plan, and patient education into a single hurried appointment.

Experience remains one of a physician's greatest assets. It allows us to recognize patterns, concentrate on what is most likely, and make decisions efficiently.

But experience should tell us where to begin—not necessarily where to stop.

A five-minute sketch may capture a remarkable likeness. Sometimes that is all that is needed.

But when someone's health depends upon getting the picture right, we should know when it is time to put down the sketching pencil, pick up the brush, and finish the portrait.