On-call and in-demand: Managing interruptions in the reading room

Twitter icon
Facebook icon
LinkedIn icon
e-mail icon
Google icon
 - workflow interruption

It’s hard to get things done when you’re constantly being interrupted. This applies to any task, but is especially true for demanding tasks requiring focused concentration—like interpreting a medical image.

To some extent, though, interruptions come with the territory for radiologists. Other physicians want to review results, technologists may have questions about a protocol, or there could be any number of other reasons for radiologists to have to pull their attention away from their workstations.

How often do these interruptions occur? That was the question tackled by Akash Kansagra, MD, and colleagues at the University of California, San Francisco (UCSF). Kansagra is now a clinical fellow of endovascular surgical neuroradiology at the Mallinckrodt Institute of Radiology in St. Louis, but back at UCSF, he helped track the number of phone interruptions that typically bombarded radiology residents at the institution.

“Anecdotally, we knew that the on-call experience of the radiology resident is extraordinarily busy in a way that is often not appreciated by those outside the field and even by those inside the field before they have an opportunity to experience what it’s like to be the single on-call radiologist for an entire hospital,” Kansagra told Health Imaging.

Showing off their findings at last year’s RSNA conference in Chicago and publishing in the Journal of the American College of Radiology, Kansagra and colleagues demonstrated that over a 90-day period of data collection, on-call radiologists completed nearly 10,400 calls, with 55 percent being incoming calls.

Breaking the numbers down further, this averages out to 72 calls during a typical 12-hour overnight shift, and an average total time of 108 minutes spent on the phone (about 15 percent of a given shift).

Kansagra explained that, depending on the time of day, the interval between calls could range from 3 to 10 minutes. The more scans that needed to be seen, the more calls there were to answer. “When the hospital is busy, the scanners are busy,” he said. “When the hospital is busy, the phones are busy, too.”

A study requiring 5 minutes to read had a 37 percent chance of being interrupted, and this chance jumped to 59 percent for a 10-minute study.

But what is the impact of these interruptions? Hard to say, according to Kansagra. “There is a dearth of literature that connects interactions to negative outcomes. The reason for that is it’s hard to isolate the cause of bad outcomes.”

Some have tried to answer the question. Researchers from Indiana University Health, including Brad J. Balint, MD, and colleagues, recently published a paper in Academic Radiology where they argued that a call to a reading room within an hour of a resident generating a preliminary report increased the odds of an error by 12 percent.

However, the authors noted a number of limitations, including only tracking phone calls and not other interruptions, and having a low number of interpretation discrepancies as data points.

Keeping interruptions at bay

There’s still an open question of precisely how seriously interruptions can impact a radiologist’s workflow, but Kansagra says radiology should follow the lead of other industries that monitor interruptions in workflow. In aviation, for instance, pilots are mandated to have a “sterile cockpit” during critical phases of the flight, such as takeoff and landing.

“Even in the absence of a proven link between the phone calls and outcomes, we should still take action based on preliminary data alone,” he says. There is value to being available for a call or face-to-face consult with other physicians, but a balance must be struck.

Kansagra and his colleagues recently sought to expand on their work with interruptions, publishing a review featuring some strategies to decrease distractions. Among their recommendations were:

  • Asynchronous communication – Unexpected findings should be communicated to the treating physician, but it doesn’t have to be a speed bump in a radiologist’s workflow. This communication can be automated for important, but non-emergent findings, such as lung nodules. Examples include having functionality within dictation software to email or otherwise alert referring physicians of such findings with minimal interaction from the radiologist. “If the patient is not unstable…create some mechanism where [the referring physician] can leave a message for the radiologist and then, at a convenient time, the radiologists can return