Montage–Nuance integration synergizes analytics capabilities

William Boonn, MD, had barely begun his career as a radiologist when his department colleagues started approaching him and fellow IT-savvy radiologist Woojin Kim, MD, with questions about analytics and data-mining. And why not? The two physicians had been interested in informatics and healthcare IT since the beginning of medical school.

As it turned out, many within their radiology department needed access to reliable data on a daily basis. Various department stakeholders were looking for data to successfully streamline operations and efficiency, improve quality and patient safety, ensure Joint Commission compliance and manage medico-legal risk, as well as advance the educational and research missions of the institution.

“These queries would get bottlenecked by a workflow that required contacting a data analyst to mine the RIS with obscure SQL or Crystal Report tools,” Boonn recalls. “In today’s world of Siri and Google, when you need to access information, you should be able to just go and query it yourself.”

That was the propelling vision and in early 2010 Boonn and Kim, together with radiologists Curtis Langlotz, MD, PhD, and Rajan Agarwal, MD, MBA, launched Montage Healthcare Solutions, Inc. Today, Boonn is the company’s CEO and Kim its Director of Innovation.

By now, Montage has refined its core functionality beyond simple search and productivity metrics.  Leveraging their clinical domain expertise in radiology, the founders have developed sophisticated natural language processing algorithms and workflow tools to enable extraction of key concepts from unstructured radiology reports to drive improved quality and performance. 

This powerful capability enables sophisticated analysis that “allows us to deliver a much richer, fuller set of quality metrics,” says Boonn, who remains an adjunct assistant professor of radiology at the Perelman School of Medicine at the University of Pennsylvania while remaining actively involved with the RSNA, ACR and Society of Imaging Informatics in Medicine (SIIM). “These can then be used to not only improve diagnostic confidence and patient safety but also drive revenue and improve quality and outcomes.”

When Nuance met Montage

Given these and other advances innovated by Montage, it made perfect sense when, in November 2010, the Philadelphia-based company began partnering with Nuance Communications, Burlington, MA, to integrate their analytics tools with Nuance’s industry-leading speech recognition platform. Users of Nuance’s PowerScribe 360 are the beneficiaries of this synergistic integration.

Not least among the benefits are proprietary data-mining techniques that pave the way to targeted quality, business and marketing analytics. Business performance can be measured, clinical decision support enhanced—and practice leaders, clinicians, researchers, and educators can search the radiology information system (RIS) and electronic medical record (EMR) for all sorts of dynamic data.

“As U.S. healthcare moves from volume-based to value-based reimbursement models, radiology practices and departments are looking to make sure they are contributing to providing optimal patient care with minimal expense,” explains Don Naugler, Director of Product Management, Diagnostics, Nuance. As Montage is integrated within the PowerScribe 360 platform, “we are now able to have report-level metrics really come to life. The facility can look for existing strengths or weaknesses, have those borne out by the sheer numbers of the metrics, and, even more important, apply those metrics to an individual, a group of individuals or a practice.”

Meanwhile, certain analytics are available even as reports are being dictated. Naugler points to client sites that have driven tightly directed initiatives from the ability to search and analyze reports in real time.

“If an imaging examination is performed on an inpatient and a follow-up biopsy or secondary procedure is recommended, you can have a very focused effort to make sure that it is done in an optimal time and thereby reduce length of stay,” Naugler says. “This is an absolute cost saver for a hospital.  Similarly, you can also target lost revenue as a result of missed follow-ups.”

To this Boonn adds that the Montage–PowerScribe integration allows a radiologist to perform searches and queries directly within the workflow.

“If I’m looking at a study of a patient who has been in the hospital for several weeks and has dozens of prior imaging studies, and I come across an unusual finding, I want to quickly find out if that finding has been previously described,” Boonn says. “Is this new? If it has been seen previously, I want to find that prior examination out of the dozens of priors to use as a comparison. By being able to do that, I can save a significant amount of time in rendering my interpretation.”

Dynamic differentiator

Boonn stresses that, in the fading days of volume-based reimbursement, basic metrics on turnaround times, RVUs and referral patterns sufficed. However, as radiology is increasingly challenged to demonstrate value, metrics that do a “deeper dive” are not luxuries but necessities. That is because it is vital to compete against other practices on quality rather than strictly on price.

“The problem with competing on price in healthcare is that it’s really a race to the bottom,” he says. “You become commoditized and undifferentiated.”

In many cases, practices still using simple analytics tools cannot produce the kinds of metrics hospitals, ACOs and payers are increasingly demanding. “You need to be able to show your follow-up rates, your compliance rates with Joint Commission regulations, your clinical outcomes,” Boonn says. “These are metrics that can only be obtained by understanding and extracting data from the report itself.”

He cites a leading radiology practice in Denver that has differentiated itself by competing on quality rather than on price.  This focus on quality has not only delivered better patient care; the practice has doubled its business in six years while maintaining margins.

“When you meet with the C-suite of a hospital, you don’t want to go in talking about anecdotes and feelings,” Boonn says. “You want to be able to give them hard data, highlight demonstrable improvement and where you stand compared to your competitors.”

Blended platforms, unified aims  

Naugler says that, historically, radiology has tended to view analytics as “a third of the equation” for showing quality. The first third was meeting guidelines as set by Medicare, the ACR and others. The second was understanding and demonstrating how one’s practice was doing with regards to follow-up, compliance, error rates and other real-time concerns. Within PowerScribe 360, there are functions to point out at the time of report creation where missteps/omissions occur.

Then, there is the third piece—the retrospective. “We told you what you are supposed to do, and we gave you some visual feedback to make sure you did it,” Naugler says. “Did you do it?”

Going forward, he says, look for Nuance and Montage to further blend their platforms in such a way that report-level analytics capability becomes “much more integral” to daily workflows.

“We want to make sure that there is actionable reporting on the metrics themselves, so you actually have the ability to track them and follow up on them,” Naugler says. “True follow-up tracking and communicating is one of the first things we are working towards perfecting. It’s a true hot-button issue out there. People are looking to see how it can be done.”

Beyond that, he says, “We are communicating those report-level analytics back out through some of the other PowerScribe notification platforms, including the critical results platform. We want to really reach out with that information proactively.”  Finally, the opportunity to leverage the PowerShare network and enable cross-enterprise data analysis is the true holy grail.

Reflecting on how far Montage has come since its early beginnings, Boonn expresses similar excitement about the future. “We’re really at the tip of the iceberg for using analytics in healthcare,” he says. “There are tremendous opportunities for us to use data to optimally deliver personalized care to patients within the context of population health management.  We’re just getting started.”