Radiology at tipping point with limitations of RVUs and the growing shortage of radiologists

She spoke with Radiology Business in an interview where she outlined some of these challenges.

Is the RVU still a good measure?

Chen shed light on the limitations of RVUs that are the basis for paying radiologists. With annual reductions in Medicare payment, there are concerns that radiologists are not being fairly compensated for the time they spend reading some exams. Same for hours dedicated to speaking with referring physicians, preparing for tumor boards, or other duties not directly related to reading. These concerns have been a contributing factor to burnout and the growing shortage of radiologists.

"It's important for people to understand that RVUs are a way that we value work that we do. But it may not take into account the nuances of a particular practice, and may disincentivize certain areas that you may need to address," Chen said. "X-rays are not really valued very highly, but someone has to read them. And so you need someone to read a hundred plain films a day and not just focus on the MRI list. And so it's important to take those into consideration because it's easy to disincentivize things by focusing just on one number."

RVUs may not consider the full scope of a radiologist's duties, she noted. The concern is that the emphasis on a numerical value overshadows other critical responsibilities, leading to a potential imbalance in the types of exams addressed.

There need to be ways to account for other nonclinical work, she said. For an academic setting, how do you take into account research and teaching, which are important parts of the mission?  

How many radiology exam reads in a day is too many?

As patient volumes rise and the number of available physicians falls, radiologists are expected to read an increasing number of exams. Chen wonders: What is a safe number of exams a radiologist can reasonably read in a day? Oftentimes institutions use data to measure how well radiologists are performing, and that benchmark always moves up amid increasing caseloads. 

"At some point, we're going to reach a point where we have to limit the number of hours that are safe to read scans. What is the number of studies that a radiologist can read in a day?" Chen said. 

There are 3 open positions for each new radiology resident

Chen said there are currently three open positions for every radiology resident graduating, posing a challenge for practices aiming to cover the workload adequately. With the pandemic significantly speeding up clinician burnout rates across healthcare, radiologists left clinical work in higher than expected numbers, bringing the shortage to the forefront.

"Our practices are trying to cover all of the work to be done, but it becomes a challenge," Chen said.

AI may help augment the radiology workforce

The use of artificial intelligence emerged as a potential solution to enhance efficiency and alleviate workload pressure. Chen acknowledged the presence of numerous AI vendors at RSNA, emphasizing the importance of AI in workflow orchestration, triaging studies, and improving communication between clinicians. However, concerns were raised about deploying AI without careful consideration of its impact on radiologists' workload and the need for seamless integration into existing workflows.

"I think there's still a lot of hype in AI. A lot of people are talking about it, but when you survey practices, only about 50% of practices are actually using AI," Chen said. 

Most appear to be leveraging AI for work list efficiency and workflow orchestration to get the right exam to the right radiologist. These systems include AI to flag studies that appear to have critical or timely findings, bringing them to the top of the queue. 

Some AI technology is receiving reimbursement, such as auto detection of stroke. Viz.AI is a good example of this technology, where stroke algorithms help the acute care team to communicate quickly via text messaging and rapid sharing of patient data. 

"It's more about the orchestration of communication across providers rather than actually some sort of AI technology that's doing some magical diagnosis," she said.

Another important area of AI, Chen said, is dictation software that enables automation of some of the impressions in the radiology report, facilitating actionable findings, and helping to decrease errors and improve communication between radiologists and other clinicians.