Home

Cover Story

Table of Contents

E-Newsletter

Article Archive

Editorial Calendar

Datebook

Writers' Guidelines

Orgs/Links

Opinion Polls

Reprints

Forum

Search

For other articles and previous issues click here.

May 24, 2004

New Voice-Controlled Integrated Reporting Tool
Vol. 5 No. 11 p. 28

Provox Technologies’ VoxReports speech-recognition–based reporting system enables radiologists to quickly dictate and create comprehensive interpretations without changing the reporting process. At the same time, VoxReports appeases cost-conscious rad administrators by using technology to reduce transcription costs.

The reporting tool works by linking speech-recognition technology with a “rules engine” to generate a report. It automates patient demographics, diagnostic information, gender- and age-related information, referring physician information, and normal clinical reviews.

It also gives radiologists the option of self-editing their reports or sending them to a transcriptionist, or “editor,” for review. (Steve Rothschild, vice president of marketing for Provox, notes that Provox terms transcriptionists “editors” because with VoxReports, transcriptionists no longer have to type entire reports; rather, they edit as needed.)

The PACS/RIS Equation
VoxReports is part of the VoxEnterprise suite of software, which allows multiple departments and more than 33 medical specialties to create and manage medical documents throughout the enterprise. The VoxExchange server handles all system administration, user profiles, and speech files and can be the repository for all documents and audio files.

Ideal for radiologists who work in several offices and at home, VoxReports can be used from multiple locations through the use of a headset or handheld microphone.

Each installation varies depending on whether the site is a radiology group, imaging center, or hospital department and what type of equipment the site has. Most installations are done remotely, with system integrations and interfaces scheduled beforehand. Radiologist training takes anywhere from two to four hours. “The speech engine usually doesn’t require much training. We will augment it depending on the proficiency of the radiologist [specifically, for radiologists with different accents],” says Rothschild. “Usually, we spend more time on how to proficiently create reports. We also spend time on customizing macros, although we usually just charge a service fee to build custom macros for them.”

The reporting tool integrates with an institution’s PACS or RIS application to display study information via predefined and customer-specific templates. Using the Provox VoxBroker Integration Engine, data are passed between the RIS and VoxReports to automate the document creation process.

VoxReports also integrates with PACS and RIS to allow the PACS or RIS work list to drive the report creation workflow. “With an integrated VoxReports solution, the PACS or RIS work list is used to identify the image being interpreted. With a single command or button press, VoxReports activates to allow dictation to begin on the selected order,” says Rothschild. “Radiologists can focus on interpreting the image and not be concerned about the dictated report until they want to review it.”

Rothschild says Provox is currently working with PACS and RIS vendors such as Fuji, IDX, Brit Systems, and AMICAS. “Our systems integration team works with each PACS and RIS vendor to ensure an optimal and reliable interface or integration is provided to our clients,” he says.

Dictation Shortcuts
With VoxReports’ SmartMacros feature, rules-based macros (voice command short-cuts created for commonly used phrases or norms) are established that interpret PACS and RIS data to select the appropriate form, fill in standard text, or conditionally convert data into the appropriate format. For example, a “normal” macro can be created that checks the age, gender, and type of exam and builds a normal report based on each specific patient.

According to Rothschild, VoxReports is the only voice-based reporting system to provide an interface that allows reports to be created and sent for proofreading without the user having to look at the computer screen. The operations are speech command-driven and have audio cues that verbally confirm a completed command.

User Feedback
Though the technology was publicly launched at RSNA 2003, it has been in use at the Austin Radiological Association (ARA) in Texas since August 2003. Since implementing the technology, the ARA centralized its outpatient facilities and integrated its PACS and RIS to VoxReports. ARA physicians can now self-edit or defer reports, and reports are processed within minutes—meeting their turnaround time of a maximum of four hours.

“Thus far, Provox has exceeded our expectations of transcriptionist turnaround time and productivity,” says Doyle Rabe, ARA CEO. “Radiologists are making an easy transition and their accuracy is improving with the editors’ corrections. Provox is meeting our needs, and we are confident that we will continue to be impressed and pleased with the product.”

No Self-Editing Necessary
One of the reasons why the ARA chose VoxReports is because ARA radiologists wanted the option of self-editing some reports and deferring others. “Provox was the only voice dictation system we evaluated that allows editors [transcriptionists] to correct the radiologists’ voice models,” says Rabe. “With every report, the radiologist’s accuracy is improving without having a negative impact on his or her productivity.”

Editors are automatically assigned reports to proofread. Through an editing module, corrections can be made using foot control and synchronized voice and text audio playback. Once a report is edited, the radiologist is notified to review and sign the report. With power sign-off, radiologists can electronically sign multiple reports with one command. Rothschild notes that radiologists can turn to Provox for their John Hancock. “If they use us for an electronic signature, we will automatically sign the report, put it on their own letterhead, and store the document,” he says. “We can also automatically e-mail, fax, or print it. We interface to the RIS, so we know who to send documents to. In other cases, the radiologist will send the report back to the RIS for an electronic signature and delivery. It all depends on the customer.”

VoxReports provides a Web tool to monitor and manage both radiologist dictation performance and editor corrections per minute, allowing system administrators to set, meet, and exceed performance goals. “What we’re finding is that it usually takes the editors 30% of the time it would take to type the entire report,” says Rothschild.

VoxReports allows radiologists to document interpretations faster than any other product on the market, according to Rothschild. “Combining speech-recognition technology with a rules engine gives us the unique advantage of being able to continue to make the engine more intelligent to further automate radiology workflow,” he says.

For more information, visit www.provox.com.

— Amy L. Conver is the assistant editor of Radiology Today.

Subscribe to Radiology Today Magazine!