Speech recognition moves into the mainstream
Speech recognition has reached a landmark. The technology has made the leap from early adopters and pioneers into the mainstream. “Speech has come of age,” sums Linda Reino, chief operating officer of Medquist, Inc.
Historically, speech overpromised and underperformed, says Nick van Terheyden, MD, director of business development for Philips Speech Recognition Systems. Those days are history. The current generation of solutions delivers, continues van Terheyden, with back-end systems bringing as much as a 30 to 40 percent increase in productivity. Turnaround times drop dramatically — as much as 90 percent — with speech, says Nuance Communications Senior Vice President Peter Durlach. And front end that completely bypass transcription and rely on the physician to edit reports can bring additional value. What’s more, the systems pay. According to MedQuist, some customers have realized ROI in as little as three to twelve months.
At RSNA 2005, vendors displayed desktop integration with RIS and PACS. This year, the newest systems from major vendors offer a high level of flexibility, enabling users to flip between front and back end recognition. Most systems don’t force physicians to self-edit, but many (even the skeptical) adopt self-editing when they realize its simplicity and efficiency. According to Nuance, up to 80 to 85 percent of PowerScribe users self-edit. What’s more, simple one-click and voice commands boost productivity and simplify the speech process. Both advances are sure to increase acceptance and adoption.
The current generation of goodies and demonstrated results are sure to tempt a wealth of new users from all avenues of healthcare, but speech is still young.
What’s on the speech horizon? Current systems like MedQuists’s SpeechQ provide XML output and send data to the clinical system in a standardized format. “This is the first step toward populating the clinical record [through speech],” notes Reino. Van Terheyden predicts the advent of a clinically driven workflow enabled by speech. The partnership between Nuance and decision support provider AMIRSYS demonstrate the additional clinical value speech can deliver. In the not too distant future, speech solutions will extract clinical data, provide feedback, facilitate knowledge exchange between the radiologist and clinician and integrate with the EHR, says van Terheyden.

AMIRSYS announced the integrations of STATdx Clinical Decision Support system with the Dictaphone PowerScribe for Radiology solution from Nuance and with Commissure’s RadWhere Suite. 
STATdx point-of-care, Clinical Decision Support system for imaging is designed to support the busy practicing radiologist by increasing speed, accuracy and diagnostic confidence in complex cases. STATdx also assists surgeons, neurologists, women’s health centers and emergency departments improve patient care. Now fully integrated with Dictaphone PowerScribe, STATdx streamlines workflow by reducing the time required to research and complete a difficult imaging analysis. Its digital format sets it apart from a textbook and ensures enterprise-wide access to on-demand reference tools at the point of care.
The agreement with Commissure fully integrates STATdx with the RadWhere Assisted Diagnosis module; a component of the RadWhere reporting suite that is driven by LEXIMER, a patented Natural Language Understanding algorithm. The integration reduces the time required to research clinical content by analyzing report content or accepting a spoken request for information at the time of interpretation and dictation.  

Crescendo announced a new front-end speech recognition module for its MD Center-XL dictation, speech recognition and electronic signature application.
MD Center-XL now automatically produces text from dictations right on screen. Voice commands allow physicians to navigate, edit, format, spell, play, fast forward, rewind, select and sign-off the report without using a mouse or keyboard. With the launch of this front-end module, MD Center-XL now allows physicians to oversee the entire documentation workflow, from dictation and correction to sign-off and final report distribution, using one single interface. The module allows healthcare facilities to leave their document creation options open: switching from back-end to front-end recognition may compensate for transcription resource shortage or periodic peaks of activity. A facility may decide that short reports can be reviewed by Authors in front-end mode, while more complex and detailed work can be routed to transcription for correction as a standard or on the fly.
Automatic template fill-in further accelerates the correction process. When a doctor creates a new dictation, the system creates the database record, recalls the template, populates the demographics and displays information on the screen in less than a second; the database record is then populated accordingly. This feature is particularly useful for “normal findings” and is configurable by work type.

Grundig Business Systems showcased DictaCordEx wireless dictation microphone.
The DictaCordEx connects with the PC through a USB port offering radiologists maximum flexibility and freedom when recording dictations. The DigtaCordEx features a slim ergonomic design and intuitive slide switch with an extra battery option. Users can move freely inside a 33 feet radius. A visual and audible warning is issued when a user moves out of range, while the dictation is simultaneously saved in the temporary memory for voice data. 

MedQuist, Inc. hosted a series of seminars based on Inland Imaging’s, (Spokane, Wash.) deployment of SpeechQ for Radiology.
Topics included:
  • How to prepare and set realistic expectations for radiologists when implementing speech recognition;
  • Recommendations for involving radiologists in the selection process for speech solutions; and
  • Integration of speech recognition software within radiology workflow strategies.
MedQuist added several of the enhancements to its next release of SpeechQ based on input from radiologists at Inland Imaging.  Upcoming enhancements related to feedback from Inland Imaging include: improved speed, improved scalability, tighter desktop integration with PACS, and templating. In addition, MedQuist anticipates a new backup module and context improvements.

Nuance Communications, Inc. highlighted its Dictaphone PowerScribe system and its PACS/RIS Extension Partner (PREP) program. The company also announced an integration agreement with Vocada.
More than 900 customers rely on PowerScribe speech recognition solutions, Nuance said. The technology saves healthcare organizations thousands of dollars per radiologist each year in reduced or eliminated manual transcription costs, automates the clinical documentation process and reduces the long turnaround time associated with the manual transcription of radiology interpretations, part of a $10 billion transcription industry in North America alone.
PowerScribe integrated digital dictation and speech recognition application brings substantial productivity gains to the medical transcription process by combining highly accurate speech recognition software and powerful physician self-editing tools to reduce report turnaround time in even the most demanding medical environments. In many cases, organizations are able to completely eliminate manual transcription services, the cost of which is a growing problem for the industry.
Through its PACS/RIS Extension Partner (PREP) program, Nuance has integrated Dictaphone PowerScribe Workstation with a wide range of RIS and PACS, including Cerner, FUJIFILM Medical Systems, GE Healthcare, McKesson Information Solutions and Siemens Medical Solutions. Today, more than 45 PACS/RIS vendors provide customers with the ability to streamline workflow by using PowerScribe’s desktop integration. The desktop integration allows radiology departments to eliminate redundant hardware, providing real-time speech recognition reporting within the PACS/RIS workstation. Medical report errors can be reduced when multiple applications share patient context, eliminating the need for paper workflow.
The agreement between Nuance and Vocada creates a simplified user interface for radiologists and pathologists to quickly and accurately communicate critical patient findings and report test results. Using Vocada’s Veriphy system with Dictaphone PowerScribe from Nuance eliminates delays and inaccuracies in the communication of critical results, improving patient safety and streamlining reporting. Vocada’s Veriphy will become a component in the Dictaphone PowerScribe dashboard and the two systems will share the same physician directories and database. Using the Veriphy system with Dictaphone PowerScribe, a reporting clinician creates a voice message directly within PowerScribe, which is then automatically passed to the Veriphy system. Instant notifications are then sent to the ordering clinician via pager, cell phone, fax, and e-mail. This automated process saves valuable time and ultimately improves patient safety by ensuring that critical test results are delivered quickly and accurately.

Philips Speech Recognition Systems showcased SpeechMagic as a thin-client solution within Citrix environments and ProScribe, a mobile display based on thin-client technology.
An enhancement to SpeechMagic enables adequate speech recognition in Citrix environments. The SpeechMagic Citrix Extension enables the centralization of IT administration for easy installation and maintenance. No files are stored locally, significantly improving the protection of patient data, Philips said.
ProScribe allows radiologists to access front-end speech recognition for dictation into forms and fields, thus ensuring fast and easy report creation and access at any place within the radiology department.