How Doctors Think (23 page)

Dr. Karen Delgado, the specialist in endocrinology and metabolism, is well recognized in her city for her lateral thinking, making diagnoses that require such creativity and imagination. When I asked her how she learned to think this way, she said she wasn't sure, but that when she was an intern she liked to play a mind game. When she admitted a patient with what seemed to be a clear and obvious diagnosis, she would stop and ask herself, What else could this be? Sometimes she was unable to come up with any other diagnosis. The obvious was almost certainly the answer. But on occasion she could rearrange the data in her mind to form another plausible picture, a different pattern that could also account for the patient's symptoms. If this proved to be the case, then she searched further. She was careful never to be satisfied from the outset. Often the search was fruitless, and the initial and obvious diagnosis was correct. But once in a while her deviation from vertical thinking, her breakout from the box where everything seemed to fit neatly, was critical in disproving that initial diagnosis—or in finding that lightning had struck more than once, that the patient had multiple problems requiring multiple diagnoses. This goes against the time-honored principle of Ockham's razor: go with a single cause if it appears to explain all the data.

Dr. Light said that during an exploratory arthroscopy in the OR, before the actual surgery, one of the other surgeons might have discovered what Dr. D found in his office, that the ligament was not functioning, that the joint between the scaphoid and lunate bones was distorted.

But, I pointed out, it doesn't inspire confidence when a doctor tells you, "I'll figure it out in the OR."

He agreed. Paradoxically, such confidence is bolstered, Light said, when a physician opens his mind to a patient and explains what he knows and what he doesn't know, what is firm about his findings and what is still unclear, which symptoms he can account for and which still demand explanation. Suppose Dr. C had stated things this way and explained that in the OR he would be able to assess the dynamic function of my wrist, be able to evaluate which joints might be deranged, instead of saying, "Leave it up to me." He would at least have shown me he was open to solving the puzzle rather than just throwing out a diagnosis, chondrocalcinosis, that didn't fit. Similarly, if Dr. B had explained, as Dr. Light did, that an MRI scan can overread a problem, showing changes in the hand that deviate from the normal but nonetheless should not be taken at face value, he would have inspired greater confidence and perhaps tacked away from deciding that three operations were necessary.

After several weeks of considering Dr. D's advice, and after getting a similar diagnosis from a hand surgeon who is a friend and member of my synagogue, I learned that a major orthopedic center in another city was beta-testing a new MRI that provided greater resolution of the bones, tendons, and ligaments in the hand. I was curious about the reach of the new technology and whether it might shed light on Dr. D's analysis. I went for the experimental scan. As Dr. D had predicted, the ligament between the scaphoid and lunate bones was frayed and lax. Moreover, tiny channels extended from the cysts. I inquired of friends at the orthopedic center and learned that a Dr. E, some thirty years the senior of Dr. D, had performed repairs of this type scores of times.

I met with Dr. E. He was formal and focused in his speech, saying that the arthroscopy and the surgery would be performed at one sitting. In addition, he used a new form of synthetic bone material for grafts, making it unnecessary to cut into my hip to harvest bone chips. Overall, the surgery was a success. After five months of rehabilitation, my wrist was about 80 percent of normal—not 100 percent. Struggling with a can opener set off swelling and a deep ache. "You have arthritis in the joint," Dr. E said in his no-nonsense manner. "Just be careful. There are limits."

This was another message that Dr. Light believes surgeons should communicate to their patients, especially in advance of an operation. "The perfect is the enemy of the good," Light said. "Nothing that you do in surgery is perfect. Everything is a compromise. Eighty percent of normal after surgery—well, that's pretty good," he said. Frankly, I had hoped for 100 percent, and like most patients I expected to be restored to pristine condition. More often than not, that is unrealistic. While you cannot predict a specific outcome for any particular patient, Light emphasized, you need to be candid and not paint too rosy a scenario.

This requires an uncommon degree of honesty—uncommon because it demands a certain deflation of the physician's ego. Here we have the contrast between what Selzer wrote about, the healthy ego necessary to putting a knife to another human being's flesh—a belief in one's own ability to make the right judgments in the OR and to work with dispatch—and the ego that imagines the scalpel as a magic wand that can restore a diseased body part perfectly. Such honesty is not rewarded in today's society. Patients shop for doctors; some doctors are keen to market themselves, knowing that it's easier to make the sale if they present their work as top-of-the-line, like a luxury automobile that flawlessly navigates the toughest curves and shifts gears seamlessly. But my banged-up hand was a 1952 Studebaker, and only so much could be done in the shop. It would never emerge as a brand-new Lexus.

"One of the intriguing things about hand surgery," Dr. Light said, is that "every patient comes in with his story, and you decode that story and then figure out not only what you can do for him, but what you can't.

"Early on in the life of a surgeon, the technical component is very important. When you are a resident in training, you say to yourself, 'I just did my first total hip replacement, skin to skin, and I feel great.' Oh, my gosh, what a sense of accomplishment! I remember the first time I reattached a finger, and I saw it pink up, and it was wondrous. As you get older, the sense of accomplishment becomes the patient who returns and tells you how much better he can function now. It's not the surgery per se but the person who is happy and pleased. So you don't want to leave people disappointed. To do that, you have to clarify the expectations up front. And then you have some people come in and demand a certain procedure, and you know in your heart that they are not going to be satisfied. Although it might be fine for someone else."

Dr. Light was revealing the kind of thinking that the most seasoned and expert doctors display: they think in sync with the patient. The patient should be helped to think in sync with the doctors, too.

In the case of my problem hand, did it all turn out pretty well because I am a doctor? Of course, I am at a great advantage being a physician and being married to one. But much of this three-year odyssey was guided by my having suffered earlier from a failed operation on my spine. Yes, my technical knowledge helped, but common sense was key. "There is nothing in biology or medicine that is so complicated that, if explained in clear and simple language, cannot be understood by any layperson. It's not quantum physics," Dr. Linda Lewis, my mentor at Columbia, once said on rounds.

There was an easily understood set of explanations for what was wrong with my hand. The cysts developed because of trauma, the wear and tear on the scaphoid and lunate bones from the banging on the computer, the karate chop in the swimming pool, the slamming elevator door. All this resulted in a breakdown of the matrix in these bones, and they filled with viscous fluid. The ligament probably was further damaged along the way by the extreme force I exerted while struggling with the juice bottle. Invented answers like "hyperreactive synovium" sound scientific and might initially impress a layman. Latin and Greek terms make up much of medical jargon and can take on unwarranted authority. But a layman speaking with another specialist, or checking in a medical textbook or on one of the better Internet sites, would soon discover that hyperreactive synovium is a unicorn, a mythical beast.

Dr. B was well intentioned, but not conservative enough. Sometimes less is more, and more can be too much. The compulsion to do everything, to address every abnormality even when those abnormalities are not particularly bothersome, reveals an irrationally idealized approach to practicing medicine. As Terry Light said, the perfect can be the enemy of the good.

Patients can help the doctor think by asking questions. If he mentions a possible complication from surgery, they can ask how often it happens. If he talks about pain and lingering discomfort from a procedure, they can ask how the pain compares with having a tooth pulled under Novocain, or some other unpleasant event. If he recommends a procedure, patients can ask why, what might be found, with what probability, and, importantly, how much difference it will make to find it. Some physicians will be uneasy, some even angry, when queried this way, because they may not have all the answers. Others will take the time and clearly respond to these simple, direct, reasonable questions. The kind of response illuminates how much the doctor really knows about your case, and how much still needs to be discovered.

Dr. D is a hero here. Not only did he think, and think independently, figuring out the genesis of my unusual problem, but he challenged today's high-tech god, the MRI scan. He was totally honest about his track record. He could have brushed aside my question and said something like, "I've done the operation successfully," which would have been true—in a single case. Terry Light had no doubt that with each passing year Dr. D will only get better, because a searching mind guides his hands.

Chapter 8

The Eye of the Beholder

P
RIMARY CARE PHYSICIANS
regularly look to doctors like Dennis Orwig to confirm the health of certain patients and identify disease in others. But despite his importance in the diagnosis and treatment of patients, very few ever meet him. In fact, Dr. Orwig spends most of his day alone, sitting in the dark. If there were a window in his office, he would look out on some of the most stunning landscape in the United States. Marin General Hospital, north of San Francisco, where Orwig works, has a commanding view of Mount Tamalpais, the undulating rise named by the Miwok Indians for its resemblance to a sleeping maiden. Eucalyptus trees surround the hospital, and soft breezes filter through their branches. But Orwig is purposely cloistered from these surroundings because he is a radiologist. Nothing should distract him from the three monitors at his workstation. On a typical day, he reads the digital images projected on the monitors: chest x-rays that show the heart, lungs, ribs, and clavicles; mammograms that unmask growths, some benign, others malignant; CT and MRI scans that reveal the architecture of organs, blood vessels, and bones.

"Radiology is a discipline broken down into two processes: the process of perception and the process of cognition," he said. This means that first a radiologist must make an observation; second, he needs to analyze what he perceived, what it means, the possible explanations for the finding. This dual process is repeated second by second, minute by minute, hour after hour during his working day. Like primary care physicians, he risks missing something significant in the blur: a change in contour of a tissue or a variation in density of an organ that he needs to notice. And as with Dr. Victoria McEvoy and others, the sheer volume has become daunting. A decade ago, a radiologist like Orwig in private practice might evaluate from twelve to fifteen thousand cases a year. By one estimate, the workload currently reaches from sixteen to twenty-five thousand cases. Some cases generate only a few images, but others involve hundreds or thousands. For example, a patient in the emergency room with fever and a cough may have a chest x-ray. This study is done with the individual in two positions: one with the chest pressed against the plate and the second with the chest pressed at the side. Thus, there are two images that a radiologist examines in this case. But a CT scan of the abdomen, which is often ordered in the ER when there is a suspicion of appendicitis, generates many hundreds of images, and the radiologist has to select the key ones to analyze from this multitude.

For that reason, radiologists are expected to look at and analyze images very quickly. In fact, conclusions from first impressions, or "gestalt," are supposed to be the mark of good training, much as "shooting from the hip" is prized among ER doctors. But Orwig takes issue with this celebrated form of thinking. "I was trained at the University of California, San Francisco," Orwig told me. "It is ranked as one of the top programs in the country. But I believe there was a deficiency in the training, and I know from colleagues that the flaw was not only there, but at many other centers." As novices, they were taught to systematically inspect each anatomic component on the x-ray. But the aim was to develop sufficient expertise so that they could abandon the deliberate deconstruction of pictures and see at a glance what is abnormal. "Somehow, over time, you were supposed to 'get it' when looking at an image." The stated reason for fostering this mode of perception and cognition was that it fit with the large number of images that a radiologist would view each working day. And, indeed, many doctors in this specialty rely heavily on first impression—gestalt—rapidly distinguishing normal from abnormal, drawing conclusions within seconds of viewing an image. But Orwig soon realized that while this often succeeded, many radiologists, including seasoned ones, missed important findings. His concern about gestalt comes not only from his own experience in practice but from studies in the medical literature.

Dr. E. James Potchen at Michigan State University in East Lansing has studied performance in reading chest x-rays. More than one hundred certified radiologists were assessed. These studies at Michigan State used a series of sixty chest x-rays that included duplicates of some of the films. When the radiologists were asked, "Is the film normal?" they disagreed among themselves an average of 20 percent of the time. This is called "interobserver variability." When a single radiologist reread on a later day the same sixty films, he contradicted his earlier analysis from 5 to 10 percent of the time. This is called "intraobserver variability."