As artificial intelligence reshapes radiology reporting and workflow, many health systems see an opportunity to modernize legacy infrastructure and reduce administrative burden for radiologists.
Yale New Haven Health System is deploying Rad AI solutions across its imaging network, spanning more than 16 outpatient imaging centers and five hospital campuses. Connecticut-based YNHHS operates five hospitals, including a nationally ranked academic medical center as its flagship hospital, several specialty networks and Northeast Medical Group.
YNHHS is affiliated with Yale University and Yale Medicine, the clinical practice of the Yale School of Medicine and the largest academic multi-specialty practice in New England.
As its name implies, Rad AI provides AI-powered radiology workflow solutions. The company's cloud-native reporting platform uses its generative AI models trained specifically for radiology and healthcare, along with a large proprietary radiology report dataset, to deliver solutions that save physicians time, reduce burnout and improve patient care, according to the company.
The radiology reporting market is seeing major shifts. Microsoft is pulling the plug on its PowerScribe 360 software, clearing the field for new players in the market.
Originally developed by Nuance, PowerScribe 360 has become the dominant reporting solution for radiologists, commanding 75% to 80% of the U.S. radiology reporting and speech recognition market at its peak. Microsoft, which acquired Nuance in 2022, announced in March that it was sunsetting PowerScribe 360 with maintenance and support for the software ending August 31, 2026. Full support will cease in 2027, Microsoft said. PowerScribe users will need to convert to the latest version of PowerScribe One, a newer cloud-based reporting solution available on a subscription basis only.
Given the forced migration, health systems and radiology practices are evaluating infrastructure and long-term technology strategy, and many are shopping the market. There's a growing list of health tech startups offering AI-powered solutions for radiology, including Jacobian, Aidoc, Sirona Medical, Rad AI, Viz.ai and Harrison.ai, to name a few. Companies are also adding capabilities as NewVue launched native AI-driven reporting to its platform.
Following the Microsoft announcement, Rad AI saw a 4x spike in inbound inquiries from health systems and a 72% year-over-year surge in new sales deals, the company reported. Rad AI's inbound forms saw a 233% jump in health systems explicitly saying they want out of the Microsoft ecosystem rather than upgrading to PowerScribe One, according to the company.
"We've had to expand our implementation teams significantly, and we're anticipating demand so that we're ready for it at the moment that we need to have people that are already trained and ready to go," said Elizabeth Bergey, M.D., Rad AI's chief clinical officer and one of several radiologists on staff at the company.
The transition to PowerScribe One comes with change management for every healthcare institution, Melissa Davis, M.D., vice chair for imaging informatics, radiology and biomedical imaging and associate professor of radiology and biomedical imaging at YNHHS, told Fierce Healthcare.
"It's a completely different product. That allows us to actually have room to take a step back and say, 'Is this the product that we want, or do we want to explore something different?' I think a lot of health systems are having that conversation right now, given where technology is in radiology, and, if we can project out the next five to 10 or 15 years, what's going to be the tool that is going to allow us to grow most efficiently in that direction?" Davis said.
YNHHS chose to look for a new radiology reporting partner, seeking a more open and flexible platform capable of integrating across its existing clinical and diagnostic systems, according to health system executives.
"We established our relationship with [Rad AI] a couple of years ago with the intention of developing a relationship with a partner that we can grow with over time, especially given the changing landscape of technology, especially in the radiology space," Davis noted.
"From a product standpoint, we're really interested in the company's Rad AI Impressions and their reporting tool, because we all know that reporting is changing. With PowerScribe 360 sunsetting, there's a big opening for what reporting looks like in radiology at this point, and we wanted to be on the forefront of that conversation," Davis told Fierce Healthcare.
Rad AI can integrate with existing systems, and its gen AI technology can reduce documentation burden, automate repetitive tasks and improve workflow efficiency, according to YNHHS executives.
Rad AI was founded by radiologist Jeff Chang, M.D., and serial entrepreneur, Doktor Gurson. They launched the company in 2018 with the aim of building gen AI tools for radiologists to help streamline workflows.
The company says its solutions are designed for radiologists by radiologists, and Rad AI asserts its gen AI-based solutions can save radiologists more than 60 minutes per shift and cut dictation time by nearly half. Rad AI's solutions include Impressions, which automatically generates report impressions from dictated findings; Reporting, a solution that uses advanced machine learning algorithms and gen AI to quickly create accurate reports; and Continuity, which closes the loop on follow-up recommendations for significant incidental findings in radiology reports.
The company claims it works with more than 200 customers across hospitals, health systems and radiology groups in the U.S, accounting for nearly 50% of all U.S. medical imaging. Rad AI works with nine out of 10 of the largest radiology groups in the country.
"We have probably the longest experience with AI and have been around and stable for a long time now, expanding relatively rapidly as the demand has increased for the products that we're producing. But we're still small enough to be a boutique provider and provide that kind of hands-on service," Bergey said.
YNHHS has been deploying Rad AI's solutions in a phased roll-out. In one week of implementation, YNHHS saw a 12% increase in efficiency, according to the health system.
YNHHS and Rad AI also plan to co-develop specialized AI tools pairing the vendor's ability to automate the documentation process with Yale's clinical research expertise, according to organization executives.
"As an academic medical center, our priority is always to advance the quality of patient care. In evaluating the future of our radiology infrastructure, we realized that standard software vendors couldn't keep pace with our evolving needs; we required a true co-development partner, capable of building alongside us," Christopher Whitlow, M.D., Ph.D., YNHHS Radiologist-in-Chief and Chair, Department of Radiology and Biomedical Imaging at Yale School of Medicine, said in a statement.
Improving operational workflows in radiology is an area that YNHHS is interested in exploring with Rad AI, Davis said.
"We're always excited to have a development partner because our radiologists are really eager to be a leading voice within radiology when it comes to what tools our radiologists are going to see," she said. "We're looking at operational workflows to say, 'I have X number of radiologists and I have many more films to read. What is the appropriate workflow for our radiologists on a day-to-day basis, so that we get the work done, but our radiologists feel like they've done good work, and they're not burned out on the other side of that, and we still have capacity for training our residents as well?' There's all these operational workflow questions and we believe that there is a solution that can be targeted towards the radiologic practice that we don't see in the market today."