by Lisa Sanders
In the emergency room they had drawn blood to try to grow the infecting bacteria, but so far they had shown nothing. They would need to be repeated whenever the patient spiked a fever—the time when the infectious agent was most likely to be found. A chest X-ray also done in the ER was normal but Stoppard ordered a second one—Kowalski had a fever, an elevated white blood cell count, and a cough—sometimes pneumonia can take a while to show up on an X-ray. He ordered tests to look for an infection in the patient’s kidneys, his liver, his gallbladder. They revealed nothing.
On the other hand, Kowalski seemed to be getting better: he still had fevers every night, but they were 100°–101°—much lower than they had been at home or in the ER. And during the day, when the medical team made their rounds, Kowalski looked tired but said he felt okay—no headache, no body aches. Whatever he had, Stoppard was relieved to see that it was responding to the antibiotics.
Or so he’d thought until this afternoon, when the patient’s temperature spiked to 104°, and the doctor had found him weeping in the darkened room. “Tell me I’m not going to die,” he pleaded with the young doctor. “Please help me.” He covered his head with the sheet and his shoulders heaved like a child’s.
In that darkened hospital room, confronted with the patient weeping beneath his sheets, his wife white-faced with worry, Stoppard was overwhelmed. What if he couldn’t figure this out? The day before, Dr. Huebner had suggested that they send the patient to the big university hospital thirty miles away, but the resident had disagreed. He thought they’d find the answer. But right at that moment he was worried he had been wrong. To see this tough guy reduced to tears seemed a reproach of his skills, of his doctoring, of his judgment in keeping him here at this small community hospital far from the fellows and subspecialists at Yale.
Stoppard, now a nephrology fellow at the University of Pennsylvania, remembers that moment well. “I didn’t think he was going to die. But I couldn’t promise that. And I couldn’t lie to him, of course. But I wanted him to know that we were working as hard as we could to figure it out. And I felt pretty sure we would.”
He outlined the plan that he’d worked out with the infectious disease specialist brought onto the case. Infection was still the most likely cause of the fever, he told the patient and his wife; they just had to find it. A CT scan of the abdomen and pelvis and an MRI of the brain would show if there were infections hidden there. An ultrasound of his heart would help them look for unusual infections in the valves—infections that can take weeks to grow in cultures. None of these infections is common, Stoppard explained, but neither was a fever that didn’t respond to a week of antibiotics.
And, while infection is the most common cause of fever, he continued, it’s not the only one, by any means. Blood clots can produce fevers; so can some cancers. Finally, diseases of the connective tissues of the body—the joints, blood vessels, and muscles—can cause fevers and body aches. They would run some specialized blood tests to look for these diseases. Something was sure to show up, Stoppard assured his patient.
The cool air of the hallway hit his face when Stoppard finally left the room, and he realized he was almost as sweaty as the patient. He wrote the orders he’d told them about and waited for something to turn up.
But nothing did. The tests were done over the next two days as the fever kept its nightly schedule. The scans of the brain and body were normal—no clots, no infections, no other enlarged lymph nodes. The ultrasound of the heart was unremarkable. The blood cultures remained negative. Tests of his liver, which had been abnormal on his admission, remained abnormal, but hadn’t worsened. One test stood out: the sedimentation rate. This is a very old test, and looks at how fast red blood cells sink to the bottom of a tiny capillary tube—a reflection of the amount of inflammation in the body. In this patient it was dramatically elevated. However, the test isn’t very specific about what is causing the inflammation—it’s one of the reasons it’s not used very often. It could be an infection but it could also be cancer or one of the diseases of the connective tissues. They hadn’t found any evidence of a cancer, and blood tests for lupus and rheumatoid arthritis—the two most common connective tissue disorders—were normal.
The resident was uncertain what to do next. Huebner once again brought up the possibility of transferring the patient to Yale. Kowalski had been in the hospital for almost a week, and they were still in the dark. Stoppard discussed the case with his colleagues and older, wiser physicians. Most of the tests they suggested had already been done. Then he spoke with Dr. Alfred Berger. A youthful man with a broad Irish face and easy laugh, he was new to the faculty but had already become a favorite with the residents. After Stoppard went through the complicated story, Berger asked only one question: “Does the patient have a rash?” No, they hadn’t seen one, Stoppard answered. But why had he asked that? Berger smiled. It’s all about patterns, he told the resident. The triad of persistent fever, joint pain, and a rash is the classic presentation of adult onset Still’s disease, an unusual and poorly understood disease of the connective tissue.
Still’s was first described in children, and in pediatrics it’s now known as systemic juvenile rheumatoid arthritis. Young adults are the usual targets. There is no way to test for it. It is a diagnosis of exclusion—in other words, before you diagnose Still’s you have to rule out everything else it could be. “If that’s what it is, it’s a great diagnosis. It’s rare and it’s cool,” he exclaimed enthusiastically. “Plus you’ve definitely got to know this one for the boards” (the tests you have to take in order to get licensed), the young teacher added as an afterthought.
In Still’s, a rash is usually seen on the trunk and arms and is often only visible when the patient is febrile. Neither the patient nor his wife had said anything about a rash. Stoppard’s team was on call that night so they would be able to look for the rash once the fever appeared.
They got that chance in just a couple of hours. Late that afternoon Stoppard got a call from the medical student on the team. “The rash, the rash—he’s got it!” she shouted excitedly. The medical student had told the patient and his wife to be on the lookout for a rash that afternoon. When she’d come to the room to check on him, Kowalski gave her a big smile, then said gruffly, “Hey, Doc, wanna see a nice ass?” He’d turned and dropped his pants to show a rash across his backside.
The resident hurried to the room. The rash was made up of painless, slightly raised, irregular patches of an unusual shade of pink; in textbooks, it’s often described as salmon-colored. The patient was started on prednisone, the usual treatment for Still’s, and his response was a near instantaneous confirmation of the diagnosis. When he was given the first dose of medicine, his fever was 102.7° and the rash glowed. One hour later, both had completely disappeared.
The next morning the patient was up and dressed when Stoppard brought the team on rounds. His hair was combed, his mustache was waxed, and the car keys were on the bedside table. He was, he told them as soon as they walked in the door, ready to go home now. The fatigue, the muscle pains, the sore throat were completely gone. They wanted to keep him one more day—just to be sure—but the patient wouldn’t hear of it. “Aren’t you sick of me yet? Because I sure as hell am sick of you guys.” Reluctantly, he agreed to stay until late afternoon, when the fever usually started, and when it didn’t arrive he and his wife went home.
Why wasn’t the rash obvious until that evening? Was that the first day he’d had it? In reviewing the chart, I saw that the attending had noted a rash several days earlier. At the time Huebner had attributed it to a simple skin infection and no one else had commented on it. And when asked, none of the team members recalled seeing the rash at all. It was outside their set of expectations. They simply didn’t see it. Knowing what to look for makes it far more likely that you will find it.
The patient took prednisone for six months. He followed up with a rheumatologist who was familiar with the disease. She warned him that the disease recurs. It
’s been a couple of years and the disease reappears occasionally. “I like the house cold when I sleep—always have—but when I wake up and my pillow is sweaty I know the Still’s is on the warpath,” says Kowalski. “But I don’t let it slow me down.” He takes a week’s worth of prednisone and again the symptoms disappear as quickly and mysteriously as they had that first time. He has to take it easy for a day or two, but knowing the diagnosis, understanding the course of the disease and what to expect, allows him to tolerate the symptoms with equanimity. The fear, the not knowing that made the fever so intolerable in the hospital, is gone. What’s left is just the discomfort. “I never even heard of that disease before I got it,” Kowalski told me, then added: “To tell you the truth, I don’t think my doctor had either.”
It’s a truism in medicine that difficult diagnoses are most likely to be made by the most or least experienced doctors. The most senior have a broad set of experiences that allows them to consider many different possibilities. Because they are open to a wide variety of observations, fewer pertinent findings are filtered out. What about the novice? They have no expectations and there is some evidence that this lack of preset experience-based biases allows them to look more carefully at the entire picture.
Dr. Marvin Chun points to an experiment conducted by his lab a couple of years ago. Participants in the study were shown two pictures; they were identical except that in one picture, a single element had been changed. The participants were shown one of two pairs of pictures. In one set the object that had been changed was central to the image. This picture showed a large room in which three people dressed in laboratory clothing are standing before a background of complicated machinery. In the first image, two support arms located just behind the people are painted bright yellow; in the next image that’s changed—they’re orange.
The second couplet featured a group of hot air balloons in the shape of farm animals. Well above them at the top of the photograph hovers a large hot air balloon with a clown’s face painted on the side. A large, vibrant pink dot is visible on the clown’s cheek. In the distance a second hot air balloon can be seen. A bright pink scarf flutters from the surface of the balloon. In the second image of that series, the hot pink spot and scarf disappear.
The researchers’ hypothesis was that most viewers would notice the changes in the image of the laboratory immediately because the object that changes color was located right behind the people at the center of the image. The change in the second set of pictures, they thought, would be harder to see because the change was peripheral. They were right. Subjects needed much more time to identify the peripheral change. Experience has taught us that important information in a photograph is principally found in the center, so that’s where we look first.
The researchers took the experiment one step further. What if the picture defied our usual expectations? Would that change how quickly we were able to find the difference in the two images? To answer this question researchers showed another group of subjects the same pictures with one difference: this time the pictures were upside-down. In this experiment the subjects would have no experience in the new inverted world, no experience-based biases, and in this setting, Chun hypothesized, the change that was peripheral to the action would be just as obvious to the viewer as the change that was central to the content of the picture. In fact, that was the case. With the upside-down pictures it took about the same amount of time to identify either of the changes.
So the novice has no expectations, the expert has many expectations. Both states facilitate close observation. Where does that leave doctors (like me) in the middle—after our neophyte days but while still on the road to expertise?
This is an area of great interest to Chun and other researchers, and it’s a hot topic in error reduction research. “I don’t know that we’ve found much that’s useful yet,” admits Chun. “I think the most important thing we’ve learned is that the control of this is primarily in the brain of the viewer.” He believes that drivers—and even doctors—need to be taught to direct their attention more broadly. When we focus too narrowly we will certainly miss something. “It’s all there for the taking. We just have to learn how to see it.”
After a morning spent with patients at his office in Montefiore, Dr. Stanley Wainapel settled back in his chair and loosened his tie. “People often ask me how I’m able to do the physical exam without my vision. Actually, for me, that’s the easiest part. If you go to hear a heart murmur, what’s the first thing you do? You close your eyes. You don’t want anything else to interfere with your ability to hear. And if you thought you just felt a liver edge you close your eyes to feel it.” I found myself nodding in agreement—once again forgetting he couldn’t see me.
He paused, thoughtfully adjusting his glasses. One of the lenses suddenly burst out of the frame. I jumped up and offered to help him find it but before I’d even finished my sentence he’d found the lens and put lens and glasses into his breast pocket.
“My patients learn pretty quickly that I don’t see much, but here’s the strange part.” He leaned forward and looked me squarely in the face. “They still bring in their MRIs; they still want me to look at their X-rays. Why do they do that? They know I can’t see them.” I considered this paradox—why bring pictures to a man you know is blind? “They don’t want me to see them. They don’t care about that,” Wainapel explains. “They want me to help them see what’s going on. They want me to help them understand. That’s really my job. Same as any doctor.”
CHAPTER SIX
The Healing Touch
The healing power of touch has long been part of Western culture. The prophet Elisha was said to have brought the dead to life with a simple touch. Jesus laid his hands on a leper and he was cured. His disciples were also granted this power to heal. Christian saints often performed miracles of healing by touch. And since Western monarchies were granted their power by divine right, many kings claimed this power as well. Until the eighteenth century, a single touch from the monarchs of England, Germany, or France was thought to be able to cure scrofula, a chronic infection of the skin—a therapy just as effective but far less painful than other available cures.
The use of the touch in medical diagnosis has had a spottier history. Hippocrates relished and eagerly employed the data provided by the senses. He wrote, “It is the business of the physician to know in the first place things … most important, most easily known, which are to be perceived by the sight, touch, hearing, the nose and the tongue.” And texture, temperature, and contour were often provided in the description of patients and their diseases in his works. That approach to medicine was followed only intermittently until the Renaissance, and not until the Enlightenment was it fully re-embraced by physicians who sought to use the concrete data provided by the body to make medicine a true science in an age of scientific achievement. Ultimately it is the same quest for the precision and accuracy of a true science that has practically destroyed the physical exam. The doctor’s touch seems primitive and uncertain when compared to what we can find out through the marvels of technology.
That’s the perception, but is it true? There’s mounting evidence that the hand of the doctor provides information that can’t be gained from the cool eye cast by its technological replacements. Take, for example, the issue of screening for breast cancer. What can an exam pick up that can’t be discerned by a machine? The machines in question—mammography, ultrasound, magnetic imaging—play a powerful role in the detection of breast cancers. But so does touch. Most breast cancers—well over 70 percent—are detected by women who feel a lump in their breast. Mammograms account for another 20 percent—clearly an important tool in the detection of this common disease. Yet studies suggest that the breast exams performed by a physician account for another 5 percent of breast cancers detected—given the number of breast cancers in this country, that comes out to ten thousand cancers picked up on exam every year, making touch a surprisingly powerful tool as well.
The a
ssessment of abdominal pain—one of the most common and problematic emergency room complaints—is another example where the physical exam may work better than even the best technology. Every year over three million patients come to an ER somewhere in the country complaining of pain in the belly. A quarter million of those patients end up in an operating room, having their appendix taken out. Most of the time, it’s a good call—the surgeon will remove a diseased organ. But on average 20 percent of those who take that trip to the OR will have what the surgeons call a negative appendix—that is, an appendix that is completely normal. For women the rate of unnecessary appendectomies can be twice that, up to 45 percent in some studies. And these statistics have been unchanged for decades.
For many years this was considered an acceptable rate. Overall it was clear that early intervention was the safest way to deal with this potentially fatal disease and that the benefit of rushing patients with suspected appendicitis to the OR outweighed the potential harm of the unnecessary surgery.
Twenty years ago, Alfredo Alvarado, an emergency room physician in Florida, developed a method of separating patients who may have appendicitis from those whose pain probably comes from something else. Using the Alvarado score, doctors can then identify those with likely appendicitis, who can be taken directly to the operating room, from those with possible appendicitis, who should be watched. Alvarado considered three components in developing his score: three symptoms—nausea, anorexia, and abdominal pain that migrates to the right lower quadrant; three physical exam findings—fever, tenderness over the right lower quadrant, and the presence of what is known as “rebound tenderness,” where the sudden release of pressure on the abdomen is more painful than the pressure itself; and a single test showing the number and type of white cells in the blood. Each factor present carries one or two points for a maximum score of 10. Those patients with scores of 7 or more probably have appendicitis and can go straight to the operating room. No further testing is needed. Those with scores of 4 and under probably don’t have appendicitis at all and should be evaluated for other causes of abdominal pain. In studies, this system has been shown to reduce the rate of unnecessary appendectomies to less than 5 percent.