There's more than meets the eye to the case of the dancing gorilla

Tools

As we report in FierceMedicalImaging this week, recent research out of Harvard University determined that a large majority of radiologists, when asked to look for nodules on a series of five CT lung scans, failed to notice an image of a dancing gorilla that had been added to one of the five.

Some have suggested that this study raises the "disturbing possibility" that the phenomenon known as "inattentional blindness"--focusing so intensely on one task that one misses distractions in the field of vision--could have "life-threatening" implications for patients. "[T]here is no way around the main finding," psychological science reporter Wray Herbert writes in a recent post. "[Eighty-three] percent of these highly trained physicians missed what might have been a life-threatening anomaly."

Actually, though, there is a way around it. Once we get beyond the fact that a dancing gorilla is hardly a life-threatening anomaly, there are several mitigating factors at work here.

Let's assume, for example, that each radiologist actually saw the gorilla instead of missing it. Wouldn't this just suggest that the radiologists weren't focusing on the task at hand--finding cancerous nodules? Instead of highlighting the fact that 83 percent of radiologists missed the gorilla, couldn't we applaud them for successfully ignoring important, though unusual, information while looking for cancer?

There's also the way in which the gorilla was presented on the scan: it's black, like the lung, so it's more inconspicuous. Yet lung nodules appear white on scans, which means a radiologist tasked specifically with identifying a nodule will focus on looking for bright spots

Then, as Ken Keller, M.D., medical director of the radiology department at Minot, N.D.-based Trinity Health points out in a commentary published last week in Diagnostic Imaging, the fact that the radiologists in this study were asked to look for just one thing (cancerous nodules) to the exclusion of anything else (such as a dancing gorilla) is atypical radiological practice.

"Although we may know a specific piece of clinical history or know that we are on a 'metastatic disease hunt,' we are trained to systematically review all areas of all images for any pathology that might be present," Keller said. "This is the art of being a radiologist and what separates us from other physicians. It's what keeps our miss rate so low relative to non-radiologists."

The study was fun. It was interesting. But the results of this study shouldn't worry anyone about the ability of radiologists to read and interpret studies. - Mike  @FierceHealthIT