Radiologists Read Scans Better if They Have Patient's Photo



By Anne Harding

TUESDAY, Dec. 2, 2008 ( — Radiologists who read imaging exams—such as magnetic resonance imaging (MRI) and CT scans—may do a better job if they see a picture of the face that goes with the diagnostic test, according to research presented this week at the Radiology Society of North America's annual meeting in Chicago.

Radiologists in the study said they read CT scans more meticulously and felt more empathy when they saw a patient's face, although it's not clear if the photo actually improved their accuracy in interpreting the test results.

Yehonatan N. Turner, MD, a radiology resident at Shaare Zedek Medical Center in Jerusalem, Israel, came up with the idea of adding photos to patients' files after reading dozens of CT scans as a medical resident. "I noticed that I know the patient's liver and spleen better than I know him," he says. "I thought maybe attaching the patient's photograph to the file may make the scan unique and important."

Dr. Turner says he was also drawing on the twentieth-century philosopher Emmanuel Levinas's idea that seeing another person's face instills a sense of responsibility for that person.

In the study, Dr. Turner and his colleagues photographed the faces of 267 patients (with their consent) two minutes before they underwent a CT scan. Fifteen radiologists interpreted the CT exams with patients' photos alongside, and then completed a questionnaire. Three months later, 30 of the CT scans that included an incidental finding (an unexpected finding that may or may not have health implications) were shown to the radiologists again, without the patient photos.

Study participants said that seeing the photograph didn't increase the amount of time it took them to interpret the scan, but they did say they interpreted it more meticulously. In fact, in 80% of the scans reviewed without photos, the radiologists didn't report the incidental findings that they had seen when they originally viewed the images.

"I think it's a really fun study to have done. I think it's very imaginative," says Etta D. Pisano, MD, a Kenan professor of radiology and biomedical engineering at the University of North Carolina School of Medicine in Chapel Hill. She's also the director of the UNC Biomedical Research Imaging Center. "It sort of makes intuitive sense to me."

But the implications are limited for the time being, she adds, given that the researchers didn't test the accuracy of the radiologists' readings with and without the photographs. "Anything that you do, even taking a photograph of a patient, costs money and time," Dr. Pisano says. "If radiologists' performance could improve for something important, then it might be worth implementing. Something like incidental findings isn't worth changing our system of providing care."

There's also the possibility that seeing what a patient looks like could have a negative impact on the radiologists' ability to interpret images accurately, notes Robert Smith, PhD, director of cancer screening at the American Cancer Society in Atlanta. "The physician could bring stereotypes and other information to bear that may diminish the accuracy, a perfectly human thing to do, by the way."

He points to a study in which radiologists who learned about a patient's family history of breast cancer spotted more tumors than if they didn't have this information. But they also came up with more false-positives.

"It's hard at face value to imagine that having the patient's photograph could be harmful," Smith says. "Typically, radiologists may work in isolation. They may actually find that this helps them to connect with the patient and helps them think more holistically about the image they're reviewing."

The radiologists in Dr. Turner's study seemed to like the idea; all of them recommended adopting it for routine practice.