Читать книгу The Invisible Gorilla: And Other Ways Our Intuition Deceives Us - Christopher Chabris - Страница 14

Even within their field of specialty, experts are not immune to inattentional blindness or the illusion of attention. Radiologists are medical specialists responsible for reading x-rays, CT scans, MRIs, and other images in order to detect and diagnose tumors and other abnormalities. Radiologists perform this visual detection task under controlled conditions every day of their careers. In the United States, their training involves four years of medical school, followed by up to five years in residency at a teaching hospital. Those who specialize in specific body systems spend another year or two in fellowship training. In total, they often have more than ten years of post-undergraduate training, followed by on-the-job experience in studying dozens of films each day. Despite their extensive training, radiologists can still miss subtle problems when they “read” medical images.

Consider a recent case described by Frank Zwemer and his colleagues at the University of Rochester School of Medicine.⁴⁶ An ambulance brought a woman in her forties to the emergency room with severe vaginal bleeding. Doctors attempted to insert an intravenous line in a peripheral vein, but failed, so they instead inserted a central line via a catheter in the femoral vein, the largest vein in the groin. Getting the line in correctly requires also inserting a guidewire, which is removed once the line is in place.

The line was introduced successfully, but due to an oversight, the physician neglected to remove the guidewire.⁴⁷ To address her blood loss, the patient was given transfusions, but she then developed difficulty breathing due to pulmonary edema (a swelling or fluid buildup in the lungs). She was intubated for respiratory support, and a chest x-ray was taken to confirm the diagnosis and make sure that the breathing tube was placed correctly. The ER doctor and the attending radiologist agreed on the diagnosis, but neither of them noticed the guidewire. The patient went next to the intensive care unit for several days of treatment, and after she improved she went to a standard unit. There she developed shortness of breath, which was caused by pulmonary embolism—a blood clot in her lung. During this time she received two more x-rays, as well as an echocardiogram and a CT scan. Only on the fifth day of her stay in the hospital did a physician happen to notice and remove the guidewire while performing a procedure to correct the pulmonary embolism. The patient then made a full recovery. (It was determined later that the guidewire probably didn’t cause the embolism because it was constructed of so-called nonthrombogenic material specifically intended not to promote blood clotting.)

When the various medical images were examined afterward, the guidewire was clearly visible on all three x-rays and on the CT, but none of the many doctors on the case noticed it. Their failure to see the anomalous guidewire illustrates yet again the dangers of inattentional blindness. The radiologists and other physicians who reviewed the chest images looked at them carefully, but they did not see the guidewire because they did not expect to see it.

Radiologists have a tremendously difficult task. They often review a large number of images at a time, typically looking for a specific problem—a broken bone, a tumor, and so on. They can’t take in everything in the image, so they focus their attention on the critical aspects of the image, just as the subjects in the gorilla study focused on counting the passes of one team of players. Due to the limits of attention, radiologists are unlikely to notice aspects of the image that are unexpected, like the presence of a guidewire. But people assume that radiologists should notice any problem in a medical image regardless of whether it is expected; any failure to do so must therefore be the result of the doctor’s negligence. Radiologists are regularly sued for missing small tumors or other problems.⁴⁸ These lawsuits are often based on the illusion of attention—people assume that radiologists will notice anything anomalous in an image, when in reality they, like the rest of us, tend to see best what they are looking for in the image. If you tell radiologists to find the guidewire in a chest x-ray, they will expect to see one and will notice it. But if you tell them to find a pulmonary embolism, they may not notice the guidewire. (It’s also possible that when searching for the guidewire, they will miss more pulmonary embolisms.) An unexpected tumor that was missed during the original reading might seem obvious in hindsight.

Unfortunately, people often confuse what is easily noticed when it is expected with what should be noticed when it is unexpected. Moreover, the procedures frequently used in hospitals when reviewing radiographs are affected by the illusion of attention; doctors themselves also assume that they will notice unexpected problems in an image, even when they are looking for something else. To reduce the effects of inattentional blindness, one can deliberately reexamine the same images with an eye toward the unexpected. When participants in our studies know that something unexpected might happen, they consistently see the gorilla—the unexpected has become the target of focused attention. Devoting attention to the unexpected is not a cure-all, however. We have limited attention resources, and devoting some attention to unexpected events means that we have less attention available for our primary task. It would be imprudent to ask radiologists to take time and resources away from detecting the expected problem in an x-ray (“Doctor, can you confirm that this patient has a pulmonary embolism so that we can begin treatment?”) to focus instead on things that are unlikely to be there (“Doctor, can you tell us whether we left anything behind in this patient’s body?”). A more effective strategy would be for a second radiologist, unfamiliar with the case and the tentative diagnosis, to examine the images and to look for secondary problems that might not have been noticed the first time through.

So it turns out that even experts with a decade of training in their medical specialty can miss unexpected objects in their domain of expertise. Although radiologists are better able than laypeople to detect unusual aspects of radiographs, they suffer from the same limits on attention as everyone else. Their expertise lies not in greater attention, but in more precise expectations formed by their experience and training in perceiving the important features of the images. Experience guides them to look for common problems rather than rare anomalies, and in most cases, that strategy is wise.