Every Patient Tells a Story (21 page)

On the other hand, the test was fast, convenient. It was easy to perform and carried no risk. One of the doctors who wrote to me about the test offered the following perspective: “Whether Adson’s maneuver is accurate or not hardly matters. The fact is that Duffy thought of the diagnosis—and if the maneuver promotes that, then it’s a good test.”

And yet if a particular exam is not reliable, how are doctors to judge the results they get? Can their findings be depended on? If the exam suggests the
presence of a specific diagnosis, will it pan out? If, instead, it suggests the patient doesn’t have the disease, can we rule it out?

We know how well many of the various technological tests work. For example, it’s been shown that an ultrasound is less reliable than a CT scan. And doctors can take that into account when they consider the test results—especially if the findings they get don’t support their own diagnostic hunches. But we don’t have that kind of data on many of the tests that make up the physical exam. And even for those for which we do have objective testing, the findings are often not taught. The result is that when we perform the physical exam we have no idea how much faith to put into what we find. That uncertainty can lead to the wrong diagnosis. Far more often it leads doctors to ignore or omit the exam and its findings and skip directly to a test that the physician can feel more confident about.

“The real problem,” says Dr. Steven McGee, who has collected and reviewed much of the research on the physical exam, “is that there is all this tradition handed down to us and our poor medical students try to learn all of it. Then they find out that some part of it doesn’t work and they throw the whole thing out. The truth is that there is a lot in the physical exam that turns out to be not terribly useful. But there are parts that are essential, even lifesaving.” McGee is part of a growing movement in research to assess the utility of various components of the physical exam.

The physical exam isn’t perfect, McGee told me, and we are all very much aware of that these days. “Our findings on physical exam feel like shades of gray while test results literally appear in black and white.” When we compare our own uncertainty with the confidence we feel when we look at a piece of paper—well, it’s no wonder we prefer tests. “But what you don’t see on that piece of paper and what we often forget is that these tests in which we have placed our confidence aren’t perfect either.” Take the chest X-ray. How reliable is that? One of the most basic findings we look for in a chest X-ray is the size of the heart—is it normal or is it large? A straightforward question and a chest X-ray should show that clearly enough. Having said that, if the same X-ray is read by more than one radiologist, how often will they agree about this simple finding?

Statisticians measure agreement using a tool called the kappa statistic. This takes into consideration the fact that sometimes with even random occurrences like flipping a coin, two people will agree or get the same answer merely by chance. To find real agreement rates you have to account for those that occur just by chance. So to use the example of two people flipping a coin, simple chance would have the coins both land on the same side about half the time. If the two coins were in agreement more often or less often, that would be their kappa statistic. You wouldn’t expect any more than 50 percent agreement and so the two coin tossers would be expected to have a kappa statistic of zero. On the other hand, if two individuals were looking at either a red card or a blue card and neither was color-blind, you would expect them to agree virtually all the time. Their kappa statistic would approach 100.

So how do radiologists do when determining if a heart is a normal size or larger? Their kappa statistic is 48. In other words, once chance agreement is taken into consideration, there’s a good chance the two radiologists will disagree at least some of the time. The same kind of disagreement occurs in other types of radiology—the problems with mammograms have been the most well described. Researchers calculated its kappa statistic as 47. Mammographers agreed with one another about 78 percent of the time. Pathology is another area of notorious disagreements.

Even laboratory testing is far from perfect.
Clostridium difficile
is a bacterium that causes severe diarrhea and requires treatment with antibiotics. Diagnosis is confirmed by detecting a toxin produced by the bacteria in the stool. When the test is positive, you can be certain that the patient has the disease. When the test is negative, however, it’s far from clear that the patient doesn’t have this infection. Studies show that up to one third of patients who have the infection will still have a negative test. Because it’s an important diagnosis to make, routine practice in the hospital is to repeat the test up to three times. Only when all three tests are negative can you be certain that the patient doesn’t have this potentially deadly infection.

What we’ve ended up with, says McGee, is a culture where test results have too much credibility and the good parts of the physical get too little.
Neither is good for the patient. And we forget that for many diseases the diagnostic standard is still the physical examination: there is no test better than the physical exam to diagnose Parkinson’s disease or Lou Gehrig’s disease. Same with many dermatologic diseases. We need to weed out the useless components of the exam. Stop teaching those parts, says McGee. The rest can play an important role in diagnosis. We lose our skills, McGee suggests, at our patients’ peril.

David Sackett, a Canadian physician considered the father of evidence-based medicine, has been one of the strongest advocates of a more evidence-based approach to the physical exam. In the 1990s he started working with the
Journal of the American Medical Association
to develop a series of articles called the Rational Clinical Exam. Each article in the series asks a question: does this patient have (some disease)? The article reviews the parts of the history and the exam and then provides the doctor with a measure of the test’s accuracy and precision. The first article focused on ascites—fluid in the abdominal cavity. In the intervening years the series has looked at everything from asthma to appendicitis. It’s been enormously successful, devotedly read and cited by physicians long frustrated by the vagaries of the physical exam.

For example, the gold-standard physical exam to find ascites, I was taught, was the puddle sign. In this exam, you ask the sick patient to get on his hands and knees, as if he were playing horsie with a child. Theoretically the free-flowing ascitic fluid in the abdomen would collect at the lowest part of the belly—the part hanging down. By striking that with your finger you would hear a dull sound if there was fluid there, a tympanic sound if there was only bowel there. It turns out that this embarrassing and uncomfortable test isn’t very useful. What was shown to be a more effective test was to check for fluid when the patient was lying on his back. The patient puts his hand on the middle of his abdomen, holding the subcutaneous fat in place, and the doctor taps sharply on one side of the abdomen while feeling the other side. If there’s fluid in the abdomen, you’ll feel it slosh against the inner wall of the abdomen. If there’s only abdominal fat, you will feel no movement.

I went to hear Steven McGee speak at a meeting of the American College of Physicians. The large room was filled to capacity. After the introduction, he walked up to the stage, a small man, trim and owlish, with horn-rimmed glasses hiding his eyes. He spoke in a quiet baritone about his own approach to making the physical exam worth doing again. Sometimes, the exam will give you all you need to make a diagnosis. Sometimes, he said, it will tell you what the patient doesn’t have. You just have to know which parts you can depend on. “Who uses Tinel’s test when you’re seeing a patient with hand numbness and tingling?” he asked the audience. Hands appeared across the room. Bad news, he told us. Not a good test. Asking the patient to show you where the symptoms occur on the hand is a better test. Those with carpal tunnel are most likely to point to the thumb and first two fingers. Finding decreased sensation on the thumb and first two fingers is a fast and simple technique that may help you make that diagnosis.

His goal, he told his audience, is to help doctors examine patients more confidently and accurately. “Once versed in evidence-based physical diagnosis, clinicians can then settle many important questions at the time and place where they first arise—at the patient’s bedside.”

When his talk was over I overheard snatches of conversation as the audience left the hall to go to their next lecture. There was excitement, hope, and passionate discussions of the accuracy and validity of favored physical exam tests. As I walked through the double doors into the crowded hallway, I fell behind a group of young doctors and overheard their brief conversation on the talk. One tall, dark-haired young doctor nudged his friend with an elbow and said simply, “As if.” Then laughed. I didn’t see his face, but the meaning was clear: as if this research could change a fait accompli, the death of the physical exam. The others laughed with him. Another in the group said, “Like I’m not going to get the test.” It was an abrupt reminder of the conservative nature of doctors. Changing this new status quo would be a challenge.

I thought again of my sister-in-law, Joanie, who’d offered to teach me on her own cancer. The gesture suggested she had far more confidence in the diagnostic potential of the physical exam than just about anyone in that
lecture hall. Would she care if these skills were just allowed to die? Would she even notice? Can simply updating our armamentarium of physical exam techniques—eliminating those that don’t work, buffing up those that do—be enough to reanimate the corpus of the physical exam? If not, what else might be needed?

CHAPTER SEVEN
The Heart of the Matter

I
leaned forward in my seat and pressed the cheap plastic earpieces of the stethoscope deeper into my ears. I could hear the normal double knock of the heart at work, but there was another sound there too—one I didn’t recognize. It was a quiet scratchy noise—regular, rhythmic, driving—like a percussionist thrumming out a rhythm on a washboard.

At the business end of the stethoscope I wore about my neck, the end that I would normally place on the patient’s chest, the silver-dollar-sized disc was missing. In its place was a small black box made of cheap plastic, about the size of a pack of cigarettes. It was a lightweight radio receiver and the sounds I heard through the earpieces were being broadcast to me.

What is that noise? I should know this
.

I sat among a dozen or so other doctors listening intently, trying to identify the cause of these abnormal sounds. All of us, medical school graduates, several years of specialty training and practice under our belts, were here at a class taught at the American College of Physicians conference, to relearn one of the fundamentals of the physical—the examination of the heart. I glanced at the woman next to me; her casually curly gray hair framed a brow wrinkled with concentration. She caught my look and smiled sheepishly. Clearly she too was stumped. A younger guy with oversized glasses stared intently at the floor.

“Who can tell me anything about what we’re hearing?” asked Dr. Vivian Obeso, the course leader. She scanned the faces of the dozen or so doctors who sat before her, on the other side of a life-sized mannequin of a young man. His chest was exposed, a sheet covered the rest of him, and his plastic legs were amputated mid-thigh. The missing end of our stethoscopes rested on the upper left side of the mannequin’s chest, a couple of inches below the clavicle, demonstrating where the sound we heard would be coming from, had this plastic dummy been a living patient. The tiny class sat silent. Despite the age and years of experience of most of the doctors, there was an awkward pause as we hesitated to answer—it was a moment straight out of sixth grade. I knew from my own years of teaching medical residents that it’s often hard to tell what that silence means. Is the question too hard? Or too easy? Both provoke the same uneasy hush. I still hadn’t recognized the heart sound and suspected that was true of the others as well.

“All right. Don’t tell me what you think it is—we’ll get to that. Just describe the sound.” Obeso tried again. “First, when does it occur? Is it systolic or diastolic?”

A normal heartbeat has two sounds separated by a very short period of what is usually silence—these two beats and the pause between them is known as systole (from the Greek word s
ystole
, which means contraction, so named by William Harvey when he first described the circular motion of the blood through the body in the seventeenth century). These are the noises made as the heart squeezes the blood into the lungs (the right side of the heart does this job) and into the general circulation (done by the left side of the heart). This double knock, onomatopoeia’d as lub-dup, is followed by another pause, this one often longer than the first. During the pause between lub-dups blood pours back into the heart, refilling each side before the next contraction. This longer pause is called diastole (from the Greek for drawing apart, because the heart enlarges as it relaxes and fills with blood). Because the activities in these two phases are so different, heart sounds are usually identified by where in this cycle they occur.

“Who can tell me? Systolic or diastolic?” The woman next to me looked up. “It’s both,” she offered quietly.

“Right. Did everybody hear that? There is both a systolic and a diastolic component.”

I listened again. Indeed, the staticky sound came between the lub and dup, but then reappeared in the time between beats.

The teacher continued: “The patient is a young man who comes to the emergency room complaining of chest pain. This is his heart exam. Can you describe the sound?” A young man in the front row looked up. “It’s scratchy,” he said.

“Exactly right.” Obeso nodded. “So what is this? There are three components to this sound. You don’t always hear all three, but even just two of them will allow you to make this diagnosis.”