Voice Recognition

Twitter icon
Facebook icon
LinkedIn icon
e-mail icon
Google icon
Speech recognition and voice dictation software vendors embraced the term 'real-time' at this year's Radiological Society of North America show in Chicago, highlighting systems that efficiently and accurately convert voice to text and improve report turaround time and increase productivity. The technology has been raising some eyebrows, and show goers were curious to see just how vendors have been addressing some of the technology's past problems. Vendors demonstrated flexible systems that permit dictation via telephone, PC, internet, handheld recorder and PDA, as well as showcased systems - integrated with numerous report generating tools - that radiologists can launch instantly while reviewing PACS images.


Commissure introduced at RSNA its RadWhere Suite and RadPort Suite - two robust software systems for clinicians and radiologists that increase efficiency and communication.

RadWhere Suite improves the means by which reports are created, decisions are made and communications completed, the company says. Its real-time speech reporting component uses patented natural language understanding algorithms to speed the reporting process that reduces how much the radiologist must say to complete a report. Real-time auto structuring enables automated knowledge access for mining clinical data as the radiologist speaks, allowing for intelligent access to relevant data in the diagnostic decision support process. RadWhere Suite also is capable of analyzing radiology reports and coding them with the proper ICD9 codes, eliminating the need for manual coding. RadWhere's enterprise teleradiology component solves the problems associated with servicing multiple sites and using multiple applications by connecting to all sites and providing a single workstation in which to access studies. 

RadPort Suite focuses on assisting referring clinicians with the task of ordering radiology exams and accessing multi-media results for their patients. RadPort's Decision Support component adds a comprehensive and continuously updated knowledge base that provides real-time decision support information to clinicians at the time of ordering by suggesting the most appropriate exam based on indications and history. RadPort also has the ability to allow the referring physician to easily check the status of an exam and review the report after the exam has been performed, solving the problems associated with results distribution. RadPort's Practice Builder component allows a radiology administrator or manager to examine ordering patterns of referring clinicians for utilization management and quality assurance purposes.


Crescendo showcased the MD Center-XL with Speech Processing, a new author correction module for its speech recognition software powered by SpeechMagic from Philips.

According to Crescendo, users can now dictate, review and correct recognized text returned from the speech recognition server, perform synchronized playback and sign off dictations and reports from one single application. The application is best suited for environments such as radiology, pathology and ER, where reports are relatively short and experienced medical transcriptionists hard to find, Crescendo says.


Dictaphone Corp. at RSNA debuted the PowerMicII microphone that combines dictation and review functions with Microsoft Windows mouse capabilities to support physicians who are dictating and self-correcting reports generated with speech recognition technology.

The PowerMicII is designed for healthcare providers using Dictaphone's Enterprise workstation and PowerScribe workstation speech recognition systems. The new microphone buttons include: a fingertip track-point button to control the onscreen cursor; a dedicated tab backward button to complement the tab forward functionality on the current microphone, voice-driven templates and form fill-ins; a transcribe button to recognize speech or send dictation to a transcriptionist automatically, when physician self-completion will not be used; and a dedicated "enter/select" button.

Dictaphone says the microphone's functions are adjustable through the Windows control panel. Microphone users also get audible feedback upon successful barcode scan when using the microphone's optional integrated scanner.

The PowerMicII will begin shipping with new Enterprise workstation and PowerScribe orders later this year, Dictaphone says.


MedQuist highlighted SpeechQ for Radiology, a speech recognition system designed specifically for radiology medical reporting. Radiologists use the product's front-end speech recognition capability to dictate, edit and electronically sign a report in one session. SpeechQ for Radiology can be used as a traditional digital dictation system. Editing the recognized text automatically updates each physician's speech recognition profile, improving speech recognition accuracy.


Nuance Communications, formerly ScanSoft Inc., used RSNA to demonstrate Dragon NaturallySpeaking Medical 8, the newest version of the company's flagship speech recognition system designed for healthcare organizations.

Converting speech into text at up to 160 words per minute, Dragon NatruallySpeaking Medical allows users to create patient notes, reports, letters and other documents in real-time, all by voice. A new roaming user interface allows doctors, nurses and other medical personnel to dictate on-the-go using a variety of input methods, including PCs, tablets and digital voice recorders.

According to Nuance, the system enables Microsoft Windows applications, including most PC and web-based EMR systems, as well as Microsoft Office, Corel WordPerfect and Lotus Notes. Version 8 has an expanded set of specialty vocabularies from four to 14: cardiology; emergency; family practice; gastroenterology; general medical; mental health; neurology; ObGyn; oncology; orthopedic; pathology; pediatrics; radiology and surgery. The product provides a core vocabulary of 300,000 words and thousands more in its specialty vocabulary sets, and also makes it easy to add additional words to existing vocabularies, or to create new custom vocabulary sets.

Nuance also showcased the Dragon MT Workflow System, a scalable, HIPAA compliant, web-based platform for end-to-end transcription processing. The Dragon MT Transcription Server manages initial capture of physician dictation through transcription, QA, e-Signature and document distribution, including network fax or printer, e-mail clients, and HL7 interface to any HIS system. The Dragon MT Speech Server manages front-end and back-end speech recognition, including automatic identification of physicians that are good candidates for speech recognition.


Philips Speech Recognition Systems showcased its voice recognition technology and announced the release of an enhancement to SpeechMagic that enables speech recognition in Citrix environments.  
Philips' SpeechMagic, which supports up to 23 languages, integrates Intelligent Speech Interpretation (ISI) technology that is able to interpret what the speaker means rather than just successfully recognizing words. High-quality medical reports can be produced with a minimum of human intervention, resulting in efficiency gains and cost savings in medical documentation.

SpeechMagic for Citrix enables adequate speech recognition from Citrix servers, which are widely used in healthcare institutions. By adding bi-directional audio capabilities, Citrix enabled the digital recordings to be uploaded and Philips developed a real-time speech recognition channel. The channel improves the usability of dictation hardware, such as the Philips SpeechMike and allows for the deployment of the full range of speech recognition features within a Citrix environment. Users can dictate simultaneously anywhere within the Citrix network and either delegate the dictation to a transcriptionist or correct it themselves, Philips said.


Provox at RSNA announced the availability of VoxSDK, which the company calls "the most advanced software development kit that enables healthcare system vendors to streamline embedding speech recognition driven medical documentation capabilities in to PACS, RIS and EMRs."
           
VoxSDK combines speech recognition, word processing, command and control, device management, text-to-speech, voice model management, and SmartMacros in one development kit. This new SDK reduces the complexity of adding enterprise-wide speech recognition to any PACS, RIS or EMR as an integral part of a vendor's system. SDK is the result of eight-plus years of development expertise that thus reduces the time to deploy speech recognition for use in more 33 medical specialties using diverse documentation work flows, the company says.
           
The VoxSDK enables other developers to utilize Provox's "rules driven" macro engine to intelligently build radiology and other specialty specific templates. An advanced word processor is integrated with the speech engine to manage healthcare text and numerical formatting interactively with the speech recognition process. Voice model management enables the speech engine to "learn" as edits are entered by providers and/or editors at one or more locations on multiple workstations.


WinScribe highlighted its digital dictation workflow system at RSNA. WinScribe is an integrated system for dictation, transcription and dictation workflow management. The technology provides anytime document creation, with flexibility for telephone, PC, internet, handheld recorder and modern PDA dictation.
   
According to the company, users of WinScribe software can dictate anywhere, anytime and have their dictation automatically routed to the relevant transcription staff, wherever they are located. The system also supports the HL7 standard so that it can integrate into hospital system environments, WinScribe says.