Tech Assessment for the Radiology Practice: Voice Recognition
The process of evaluating new technology has quickened its pace, but that does not make it any less challenging for a busy radiology practice, Ted Best says. Best is acting director of IT for Carolina Regional Radiology (CRR), Fayetteville, North Carolina, and is president of Xodus Technology Professionals. “The process flow hasn’t changed; what have changed are the tools that we use,” he notes. “One of the key factors in making these projects a success is communication, and using advanced tools such as group exchange servers, instant messaging, and video conferencing have helped us communicate more effictively and move the decisions forward at a faster pace.” He adds, however, that radiology practices can still succumb to indecision when evaluating new technologies if project best practices are not followed. Best helped shepherd CRR in its selection of a voice-recognition solution, in a project aimed at enabling the practice’s 21 radiologists to use a unified system, as opposed to the disparate voice-recognition systems in use by CRR’s hospital partners. “We desired a robust PACS that would give our radiologists the functionality to deliver patient care in the most efficient manner possible,” says Harry Ameredes, president of CRR. “Serving several different imaging centers and hospitals, we needed a PACS solution which would allow us to integrate multiple modalities from several different facilities into a single workflow. With regard to voice recognition, we desired an embedded platform over an integrated platform. An embedded platform is part of the PACS, not a stand-alone application; an integrated solution allows the radiologist to toggle patients with partial dictations open on each patient.” Naming the Team There are, Best notes, “so many different ways to keep these projects from being a success.” In a democratic governance structure like that of many radiology practices, he observes, “You find a lot of people who have ample opinions regarding a subject, but don’t really want the responsibility of making the final call.” Best recommends that practices choose both their evaluation teams and their final decision makers from the start. The evaluation team should include “anyone who uses the product or maintains the product,” Best says, along with “at least one person who completely understands it or has process-flow knowledge. On a voice-recognition project, you need someone who understands the reporting workflow from start to finish.” There should also be a handful of named decision makers. He adds, “When you don’t have named decision makers, things go out to committee and never get decided upon; everyone has input, but there has to be a decision maker (or two or three) who makes the final call.” In CRR’s case, this process was significantly streamlined because most of its radiologists had experience in working with several voice-recognition products—and by the fact that its PACS provider, Intelerad® Medical Systems, could offer the M*Modal voice-recognition solution as an embedded part of its InteleOne® PACS Multi-Method Reporting ModuleTM. “Intelerad has provided us a solution that combines all our disparate systems into a single workflow,” Ameredes notes. “We set up a test environment—a test workflow—and we liked what we saw. As a radiologist, it’s important to have an intuitive, robust, efficient workflow; the more quickly we can get a report out to the referring physician, the more value we will have in the patient-care process.” Trying the Product CRR’s input team included a handful of radiologists representing the practice’s various subspecialty areas. Ameredes says, “We took a sampling of the practice’s subspecialties.” Orthopedic radiology (the subspecialty of Ameredes) and mammography were two areas of the practice that embraced the solution early. Sheryl Jordan, MD, a breast imager with CRR, reports that today, the practice’s average work RVUs per breast-imaging radiologist are 85.17—a remarkably high figure, given the additional responsibilities of breast imagers, beyond reading. “It’s nothing short of remarkable, and that is solely because of the use of voice recognition and templates,” she says. CRR’s orthopedic radiologists have also seen benefits from the use of voice recognition and templates; Ameredes says, “We have built standardized templates which allow our radiologists to be more efficient, while providing the referring physician a consistent report across our practice—regardless of which particular radiologist is interpreting the study.” The practice is continuing to roll out the voice-recognition product, and the next step, Ameredes says, “is to take what we’ve done with orthopedic radiology and mammography and extend that across the practice. We’ve been planning to have Intelerad come back to train our radiologists, but we’ve already been able to get three or four radiologists up to speed without any official training.” On the IT side, Best says, teams should have an eye on the practice’s infrastructure. “Given what you want to do, do you have the workstations in place to handle it? Do you have the network capacity?” He adds, “Voice recognition is very sensitive to network capacity.” Post-implementation Results CRR implemented its embedded voice-recognition solution in 2011. Already, it has led to big gains in terms of patient and referring-physician satisfaction, the CRR radiologists note. “With our ability to work inside diagnostic templates, what we have is the ability to ensure that our diagnostic patients’ reports are already faxed to the physicians’ offices before the patients even leave the facility,” Jordan says. “We’ve certainly already spoken to the patient as well. The templates actually encourage you to spend more time talking to patients, and while that part slows you down, the templates put you ahead of the game.” Referrers are responding well, Ameredes says. “Not all radiologists report the same way, even within a practice, and that can make it hard for referring physicians to get answers for specific questions,” he notes. “Our orthopedic templates are set up as if an orthopedic surgeon were doing the arthroscopy—the first paragraph is the medial compartment, the second is the lateral, and so on; the referrers know where to find important information in each of our reports.” Looking ahead, Ameredes says, the added efficiency made possible by embedded voice-recognition positions CRR well for the future. “It hasn’t always been this way, but now, to stay relevant, radiologists need to get that report to the referring physician as soon as possible,” he says. “Volumes are always increasing, and like many radiologists, we’re keenly sensitive to efficient workflow that helps us take better care of patients.” Cat Vasko is editor of ImagingBiz.com and associate editor of Radiology Business Journal.