Recent developments in artificial intelligence (AI) for radiology appear more likely to add to the workload of radiologists than reduce it, Dutch researchers reported in an article published online on 29 June in Insights into Imaging.
After reviewing a random sample of 440 medical imaging research studies published in 2019, Dr. Thomas Kwee of University Medical Center Groningen and Dr. Robert Kwee of Zuyderland Medical Center in the Netherlands found that over 86% of AI-focused studies were associated with a higher workload for radiologists.
This was largely due to the need for longer postprocessing and interpretation times, according to their analysis.
These AI studies were associated with an increased workload (p < 0.001) in both the academic tertiary care setting (odds ratio, 10.64) and nonacademic general teaching hospital setting (odds ratio, 10.45). The findings were statistically significant.
Although AI has been regarded as a tool for decreasing the workload of radiologists, the study shows that most current AI applications in medical imaging may have the opposite effect, according to the researchers. That's because these algorithms require additional postprocessing and interpretation time, rather than being seamlessly integrated in the workflow and taking over radiologist tasks, they said.
"Therefore, we believe that there is currently no scientific basis for policy makers to use AI as a reason to refrain from expanding the radiology workforce or to cut reimbursements for imaging procedures," they wrote. "To reverse the (looming) shortage of radiologists, it may be necessary to enroll more residents into radiology training programs, train and employ radiologist assistants, and/or increase financial resources to employ (new) radiology staff."
Overall, the researchers found that approximately two-thirds of the medical imaging research studies they reviewed could directly contribute to patient care. A little less than half of these studies would increase radiologist workload, however.
"Recently published medical imaging studies often add value to radiological patient care," they wrote. "However, they likely increase the overall workload of diagnostic radiologists, and this particularly applies to AI studies."
