Researchers Use Eye-Tracking Technology to Assess Wrong-Patient Errors


R&E researcher examines the value of pairing photographs with radiographs in detecting wrong-patient errors.

New RSNA-funded research shows that patient photographs paired with radiographs may reduce wrong-patient errors and improve image assessment without increasing interpretation time.

With a 2015 Canon U.S.A./RSNA Research Medical Student Grant, Alex Chung, MD, and colleagues used eye-tracking technology to assess the visual attention of radiologists examining radiographs with and without paired photographs.

Despite protocols that require ensuring patient identity by checking two unique identifiers whenever a procedure is done, radiographs taken in emergency departments (EDs) or intensive care unit settings are at a higher risk of wrong-patient identification errors because patients often cannot provide the unique identifier information.

While previous research has demonstrated that a paired photograph of the patient taken simultaneously at the time of the radiograph increases the detection rate of wrong-patient errors, there is some concern that the photos are a distraction or may increase interpretation time.

“Our specific aim was to incorporate eye-tracking technology and objectively quantify the degree of distraction posed by photographs,” said Dr. Chung, a transitional year resident at Emory University School of Medicine, Atlanta.

Eye-Tracking Technology Yields Results

The study comprised 10 radiologists (six male, four female) from the University of Arizona specializing in a variety of areas including general, abdominal, cardiothoracic and pediatric radiology. Dr. Chung and colleagues obtained patient data (radiographs and photographs) at Emory University and conducted the eye tracking observer study at the University of Arizona.

The subjects reviewed 21 portable chest radiographs in two phases. First, the images were provided without patient photographs and the radiologists were asked to note the placement of various tubes and lines. To prevent recall bias, Dr. Chung and his team allowed at least three weeks to pass before performing the next review. In the second phase, the images were paired with photographs and subjects were asked to perform the same task noting placement of lines and tubes.

Eye-tracking technology measured how long the subjects focused on various areas of the image and the total time spent looking at each case. The technology also noted the distraction rate, or number of times the subjects’ eyes scanned off the radiograph either to view image labels, or in the second phase, to view the photograph.

In both phases, the images were presented on an LCD display, ambient room lighting levels were controlled, the average distance between the observer and the screen was 35 cm, and the eye-tracking equipment sampled eye positions every one-sixtieth of a second with an accuracy of 0.4 degrees and a precision of 0.34 degrees.

The findings from both studies showed that overall time spent viewing the cases did not increase with the addition of the photograph.

“Radiologists compensated by integrating the examination of the photo into their search by decreasing, somewhat, their time on the x-ray image,” said Elizabeth Anne Krupinski, PhD, professor and vice chair for research in the Department of Radiology and Imaging Science at Emory University, who supervised Dr. Chung’s research.

After each phase, the subjects completed a survey collecting demographic information and answering questions about how they acquire patient information, their opinions about the photographs and which body areas they would like included in patient photographs.

Following the first study, all subjects indicated that they would be significantly likely to contact the referring provider if they detected a critical result. After the second study, the subjects were asked to indicate how much more likely they were to call the referring provider if an important finding was detected with the photographs present than without. Seven providers reported no difference, two reported slightly more, and one reported significantly more when photographs were present.

“This research shows that having photographs may help communicative and empathic dimensions of the interpretation process as well,” Dr. Chung said.

“This study was important, because it was the first of its kind to assess the impact of providing patient photographs during image interpretation on the way radiologists search images for findings,” Dr. Krupinski said.

Results May Personalize the Reading Experience

The research findings also show promise for the patient-radiologist relationship, said Srini Tridandapandi, PhD, MD, MBA, a radiologist at Emory University who helped develop the technology to add photographs to radiographs.

“With widespread adoption of PACS over the last couple of decades, we have ‘lost the patient’ and such photographs may help bring the patient back to the radiologist,” said Dr. Tridandapandi, who also served as Dr. Chung’s mentor.

Dr. Chung agreed, noting that this preliminary study provides a firm rationale for conducting a clinical study. “In the future, a study that explores whether the diagnostic accuracy of the reports is affected by the presence of photographs is warranted,” he said.

Dr. Chung credits the RSNA Research Medical Student Grant program with helping him develop grant writing skills and providing support for securing protected time at his institution. He plans to conduct additional research as he pursues a career in academics.

Source:http://www.rsna.org

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s