KLAS: Hospitals should look for these four traits in AI imaging vendors

Success in artificial imaging work comes down to understanding how to pick and work with a vendor, the firm said.
By Mike Miliard
09:17 AM
Share
KLAS report

Artificial intelligence continues to make inroads in healthcare, particularly in imaging. As AI algorithms augment the brains of radiologists in large health systems and integrated delivery networks, many smaller providers are curious about how they might take advantage of the technology, too.

Of the 80-plus organizations KLAS surveyed for a new report, in fact, most are beta testing AI in limited imaging settings while investing in research and development. The firm also said some healthcare providers are sitting tight watching as the technology evolves and others are looking around to identify use cases and prove an ROI.

The report, titled “Artificial Intelligence in Imaging 2018: Early Users Speak Out,” also included some hype-tempering sentiments wherein anonymous users were quoted saying that some of the technologies are mostly marketing that lags behind expectations.”

[Also: AI adoption rate picking up and radiologists are paying attention to workflow, career impact]

“Over the last 20 years, KLAS has watched many emerging spaces gain momentum only for vendors to fail to deliver on the market’s expectations,” the authors wrote. “Where there is success, there are typically several key vendor attributes present. By looking to partner with vendors who exemplify these traits, providers can begin their imaging AI journey on the right foot.”

To that end, the authors recommend starting a set of clear expectations, including outcomes to be reached in a specific timeframe as well as the steps both the hospital and tech vendor must take to success.

Once those are in place, the firm suggested that customers strive for a strategic relationship with dedicated account management at a minimum.

The expectations and partnership should also lead to a strong focus on user training. Don’t be tempted to let this go for a lower price tag, either. Instead, pair users with trainers that understand each other and plan follow-up steps ahead of time.

Fourth is data governance, which can determine success or failure. Make governance a priority early-on, include as many relevant parties as possible, and institute best practices.

Perhaps unsurprisingly, IBM's Watson technology is top-of-mind for many making investment decisions, particularly with its initiatives such as the Watson Health Imaging Collaborative.

Hospitals and systems are also taking a close look at AI tools from Agfa, Carestream, GE Healthcare, McKesson, Philips, Sectra, Siemens and Zebra Medical Vision, according to the KLAS report.

Many respondents planning to put an imaging AI strategy in place said they're still more than a year away from adoption – and a bit more than one-third answered it will take at least three years to get their imaging algorithms up and running.

Twitter: @MikeMiliardHITN
Email the writer: mike.miliard@himssmedia.com