Is this the end? Machine learning and 2 other threats to radiology’s future

Radiology is one of the cornerstones of modern healthcare, but according to a new analysis published by the Journal of the American College of Radiology, machine learning could potentially end the specialty as we know it within the next decade.

Authors Katie Chockley and Ezekiel Emanuel, MD, PhD, of the Perelman School of Medicine at the University of Pennsylvania in Philadelphia, called out machine learning as “the true threat to radiology.”

“Trends favor machine analysis over human analysis,” Chockley and Emanuel wrote. “Billions of images for training machines already exist. Technological advances mean that imaging equipment can capture ever higher resolution images.”

Machine learning represents many improvements for radiology, the authors said. For example, the machines allow data “to speak for themselves,” which can lead to trends being uncovered that could have gone unnoticed otherwise. Also, the pixel-by-pixel focus of machines can pick up key predictors.

“A human reading a chest radiograph may be inclined to simply interpret the image, determining if it represents someone healthy or sick and, if sick, whether infection, fluid, tumor or another issue is present,” Chockley and Emanuel wrote. “On the other hand, a machine will treat each tiny pixel on the scan as an individual variable and will seek to organize those pixels into shapes and patterns and, from there, make a diagnosis.”

Computers can also quickly digest complex data sets, the authors added. While even the most trained radiologists will have cognitive limitations, no such issues exist with machine learning.

Chockley and Emanuel also detailed how machine learning is already being used in computer-aided detection (CAD), showing that machines can work “as well as or better than” veteran radiologists.

2 other significant threats to radiology

The analysis from Chockley and Emanuel broke down two other threats to the future of radiology: the deinstitutionalization of healthcare and payment reform.

The deinstitutionalization of healthcare, they explained, can be seen in the way more and more care is being provided outside of hospitals. Incentives from the Affordable Care Act and the growing influence of facilities such as minute clinics and urgent care centers has played a role in this trend.

“Because care is moving out of the hospital, many hospitals are operating at a loss or are closing altogether,” Chockley and Emanuel wrote. “At least eight hospitals have closed in 2016 so far, and 74 rural hospitals have closed since 2010.”

This represents a decrease in demand for imaging, the authors said, which means less demand for radiologists as well.

The other significant threat to radiology’s future, payment reform, can also result in less demand for radiologists. The authors wrote that physicians participating in certain bundled payments are working toward reducing imaging and visits to the emergency room.

“As bundled payments and other payment reforms remove incentives to order costly and unnecessary tests and treatments, the number of images ordered will decrease,” Chockley and Emanuel wrote. “This, in combination with the reduction in tests run due to the deinstitutionalization of care, will decrease the number of images for radiologists to read.”

Other content related to this topic can be read here, here, and here.