Journey to Imaging 3.0: A road paved with information technology

As an early adopter of speech recognition technology in 2000, Richard H. Wiggins, III, MD, CIIP, FSIIM, witnessed a stunning reduction in turnaround times (TAT) at the University of Utah Health Care, Salt Lake City.

The organization grew, imaging volumes proliferated, TATs plummeted, radiologist RVUs soared and inpatient length-of-stays declined—a classic Imaging 2.0 success story propelled by speech recognition, PACS, and RIS.

Squeezing every last ounce of efficiency out of the systems that have smoothed the transition from an analog to digital radiology workflow has been a quest of Wiggins, professor, department of radiology, University of Utah Health Sciences Center. “We want to get every bit of efficiency we can out of those systems,” he says.

For the next leg in radiology’s journey, the transition to value-based imaging or what the American College of Radiology (ACR) calls Imaging 3.0^TM, Wiggins anticipates a movement to structured reporting to better mine the massive amounts of information the reports contain. “We expect this move will allow us to search reports better and do better analytics,” he shares, with the potential to improve patient care quality. 

The University of Utah health system has taken initial steps toward that goal in its use of templates, macros and pick lists available in PowerScribe 360 from Nuance, Burlington, Mass. The preference for these different speech methodologies varies by radiologist, but having a system that allows for their use to different degrees facilitates acceptance and increases utilization by imagers both inside and outside of the department of radiology, Wiggins says.

Taking a leap

Encompassing four separate hospitals and ten outpatient clinics scattered along the Salt Lake Valley, the University of Utah Health Care sites are all connected, completely filmless and receive more than one million clinical visits annually.

“We have a huge referral base, about 10% to 15% of the geographic continental United States,” Wiggins says. “There is not a lot between Denver and Seattle besides us for academic sites, so we get a lot of the rural referrals from the surrounding five states.”

With the adoption of speech recognition in 2001, TATs dropped from 40 hours to under 10 hours. A longtime user of Nuance speech recognition technology, University of Utah Health Care upgraded to PowerScribe 360 in 2012, and this time TATs dropped more than 50%.

“Our ER reports had about a 2-hour turnaround time in 2007, and that dropped down to under an hour in 2012 with PowerScribe 360,” he reports.

TATs for plain film were reduced by more than 50%, declining from an hour and 40 minutes to just 40 minutes in 2012. CT emergency department (ED) cases also declined more than 50%, from 2.5 hours to an hour; and MRI went from 16 hours down to only an hour and a half.

Wiggins attributes the results in part to a better quality, faster speech recognition engine and improvements to the software that recognizes and digitizes the speaker’s voice.

Another factor was a cultural change that emphasized improved communications and frequent meetings with ED physicians, surgeons, and referring physicians to assess when reports are received and where they are going. “It’s a constant effort,” Wiggins says.

Radiologist workflow

Implementing the new reporting platform to support—rather than detract—from the organization’s workflow was a priority for Wiggins. He also was unwavering on a single mouse and keyboard for each PACS workstation with integration of RIS and PowerScribe 360 to maximize “eyes-on-images” time.

As a result, University of Utah Health Care installed a bidirectional workflow that streamlines the reporting process for attending radiologists, whether they are reading studies in an attending workflow, or utilizing the resident workflow to sign studies read by residents.

When reading in production mode, the attending chooses a study in PACS from an unread list populated by the RIS, and the study opens in PACS, which directs PowerScribe 360 to open the accession number and automatically load the appropriate macro or template, based on the body and modality, in preparation for dictation.

When reviewing and signing studies read by residents, the attending radiologist looks at PowerScribe 360 and chooses a report dictated in his name, which in turn opens the report in SR and directs PACS to open the correlating images.

Wiggins explains: “Our site is such that our radiologists may all day long be going back and forth between an attending workflow and a resident workflow. That bidirectional workflow between PACS and PowerScribe 360 is huge for us.”

Wiggins sees the movement to structured reporting and the ability to accept DICOM structured reporting as an imperative future step. Continuously evolving, the DICOM standard has added DICOM Structured Reporting (DICOM SR), which will allow for improved communication with referring providers, improved imaging workflow and streamlined reimbursement.

The next leg of the journey

As radiologists like Wiggins set their sights on better management and mining of report data, the company behind PowerScribe 360, Nuance Communications, is laying the groundwork to support radiologists in their next steps toward that goal.

“Radiologists were engaged with speech recognition technology from the early days, and helped evolve the technology to a solution that integrates seamlessly into the radiology workflow,” says Christy Clark Murfitt, director, diagnostic solutions marketing, Nuance Healthcare. “PowerScribe played a critical role in making the radiology department the most efficient in the healthcare enterprise.”

The software has evolved to offer more than increasingly accurate voice recognition to streamline productivity, Murfitt says. Templates, or “pick lists”, that align to specific procedure codes, add efficiency, and a framework for quality documentation that meets regulatory and referring physician requirements.

“We know that in the era of value-based payment models, the focus will be on volume and value,” Murfitt says.  “Quality has to be an imperative; but technology solutions have to deliver quality while maintaining the current efficiencies that have come to be expected in the enterprise.”

Radiology is in the information business, Murfitt reminds. Not only are the customers asking for structured reports, but structured reports that facilitate data mining.

“To stay relevant in the Imaging 3.0 era, we need to ensure that our clients are empowered to aggregate all necessary data from an actionable report and have the ability to data mine to prove quality in the value-based care environment,” she says.

It’s personal

In some ways, Wiggins’ quest to wring maximum efficiency from PowerScribe 360 is personal: His father, Richard Sr., produced much of the original work in digital signal processing and speech synthesis that resulted in the Texas Instruments© digital speech synthesis chip within the handheld educational toy, Speak and Spell, in the ’70s, paving the way for speech synthesis and speech recognition software in the modern imaging world.

The role of speech recognition and other imaging informatics technology solutions in the historic transition of radiology from an analog to digital is indisputable. As these systems continue to evolve to support the quality needs, data mining and regulatory requirements of the current healthcare environment, they will be equally important in moving radiologists into the Imaging 3.0 era.

At University of Utah Health Care, Wiggins envisions the next step as adoption of the quality assurance modules in PowerScribe 360 that leverage natural language processing to aid radiologists in avoiding laterality (right–left) and gender errors. 

As in the transition from analog to digital radiology, the evolution of imaging from version 2.0 to 3.0 will require changes in culture, practice patterns and priorities to meet the demands of value-based healthcare.

“The big change in thinking about our modern imaging informatics world is that we are not looking at film, we are looking at data,” Wiggins notes. “How we handle data is very different than how we used to handle hard copy film, and that change was not easy for everyone.”