IBM MED SPEAK / RADIOLOGY

For nearly a century, radiology reporting has been an inefficient and time-consuming process. Until now, that is, with the introduction of IBM MedSpeak/Radiology(TM), a high-accuracy speech recognition solution that enables radiologists to create their own diagnostic text reports -- quickly, easily and almost instantaneously -- by simply speaking aloud. With little or no special training of the system, most North American English speakers can use MedSpeak/Radiology to dramatically reduce the time and cost of radiology report generation.

Simple and natural
Unlike most "discrete" speech recognition systems, which require the user to speak slowly and insert distinct pauses between words, MedSpeak/Radiology lets you speak continuously and at a natural pace with the text appearing instantaneously on the screen.

The product of more than twenty years of research by IBM, the Speech Company, MedSpeak/Radiology has an average accuracy rate in beta tests of over 95 percent for most North American native English speakers in the United States and Canada. It includes a special radiology vocabulary of more than 25,000 words and a language model that understands the common word patterns used in radiology reporting. (You can also add your own words, such as proper names and abbreviations.)

MedSpeak/Radiology's statistical language model allows it to distinguish easily between words and symbols that sound alike, such as "to," "two" and "too," or "colon" used as punctuation and "colon" as it refers to a section of the large intestine. Moreover, the system's sophisticated adaptive capability permits it to learn your unique dictation style as it is being used.

Virtually all the functions necessary to manage the radiology report dictation process are packaged into a complete application that includes a voice-activated graphical user interface that lets you use a combination of voice commands, mouse or keyboard to control the system's functions. The software comes packaged with a high-quality, noise-canceling microphone in either a hand-held or headset model.

Finally, dictation can be performed virtually eyes-free -- without looking at the screen -- so you can concentrate on reading and interpreting diagnostic images.

Greater control,
enhanced flexibility
With MedSpeak/Radiology, you can edit, sign and store your own reports instantly, or pass the draft to a colleague or transcriptionist for further editing. The system saves both the original recording and the transcribed text of any preliminary, unsigned version of the report, so a reviewer or transcriptionist can play back selected portions of the report, focusing only on those areas, if any, that need to be changed. The confidentiality of patient reports may be increased because of reduced reliance on outside transcription services.

MedSpeak/Radiology gives you greater control over the entire reporting process by enhancing accuracy and reducing the average time to complete a radiology report from hours or days to just minutes. In research studies conducted by the Massachusetts General Hospital, radiologists using MedSpeak Radiology virtually eliminated the administrative portion of the radiology dictation cycle when compared with their existing transcription system.

And the new automated process is more flexible than ever. Completed reports can be stored and retrieved right on the system via MedSpeak/Radiology's own powerful database manager or linked to patient records on other hospital or radiology information systems.

Integration made easy
Whether you are a single practitioner or a member of a large practice with centralized reporting requirements, MedSpeak/Radiology's client/server architecture offers the right solution for you. In a standalone environment, each MedSpeak/Radiology workstation can perform all the functions -- including permanent report storage and retrieval -- of a complete dictation system on a single machine.

For larger practices, multiple MedSpeak/Radiology workstations can be connected via a local area network (LAN) to a centralized server. In this configuration, any radiologist with an approved password can use any of the MedSpeak/Radiology workstations with consistent recognition accuracy and access. Preliminary reports can be centrally stored for editing by other radiologists or transcriptionists, if desired. The system tracks original authorship and the status of each report, a plus in teaching hospitals, where several levels of approval are required).

The bottom line? MedSpeak/Radiology enables you to improve your service to patients and referring physicians while reducing costs.

The following advanced features are part of the complete MedSpeak/Radiology solution:

• Open RIS/HIS Interfaces. Healthcare providers with a Radiology or Hospital Information System (RIS or HIS) that generates radiology orders may want to create an electronic interface to download them directly to MedSpeak/Radiology and upload the finalized radiology reports to be stored directly on the RIS/HIS. Several standardized interfaces are available for an additional charge. For providers with other systems, MedSpeak/Radiology also contains a documented HL7-compliant application programming interface, with standard format and standard protocols. This means you can send the information wherever you want it to go with a minimum of programming.
• Simplified Order Entry. For those users without a RIS/HIS interface, MedSpeak/Radiology includes a simple order entry application module that displays the order in the header of the MedSpeak/Radiology report dictation screen.
• Bar Code Interface. MedSpeak/Radiology includes a standard bar code interface that users can modify to suit their own bar code environment.
• Pre-Stored Reports. The individual radiologist or institution can create sets of canned text reports (for a "normal chest" diagnosis, for example) and store them for later recall with a single command word or acronym. When displayed, the pre-stored report can be further customized or simply signed as a final report.
• System Security. MedSpeak/Radiology's user ID and password protection at sign-on and password reentry can be used for electronic signature of reports to ensure the security of patient information.

The configuration will vary depending upon whether MedSpeak/Radiology is installed in a client or server mode. However, a typical MedSpeak/Radiology client system will generally require:

• Windows NT 3.51 operating system
• Pentium/Pro 200 MHz uniprocessor IBM PC or IBM-compatible system
• 32 MB of available memory (RAM) and 256 KB L2 cache (includes Windows NT 3.51 operating system requirements; additional memory may be required for other applications)
• At least 1 GB of disk space (accommodates operating system and MedSpeak/Radiology software operating environment; additional disk space may be required for other applications)
• CD-ROM drive (quad-speed or faster)
• 3.5" high-density diskette drive
• Standard SVGA, 1024 x 768 resolution, 256-color personal computer monitor
• Creative Labs SoundBlasterx 16 (or 32) Sound Card
• LAN Card (only if LAN attached)
• A high-quality, noise-canceling microphone with cable (included in the MedSpeak/Radiology package)
• Optional external loudspeaker (for voice playback in correction mode) and cables compatible with the audio subsystem
• Optional bar code reader

All in the family
IBM offers the most advanced selection of speech products in the world. In addition to MedSpeak/Radiology, IBM's proven VoiceType(TM) family of products includes speech recognition offerings for OS/2(TM), Microsoft Windows(TM) 3.x and Windows 95 on the 486 and lower-end Pentium platforms. Clinicians and physicians other than radiologists who want to use speech dictation to record patient notes, will find the IBM's VoiceType family well-suited to their needs. Reflecting more than two decades of IBM research and development -- and over 50 patents in speech recognition -- IBM VoiceType speech recognition technology has enabled business people and professionals at every level -- and across a wide spectrum of industries -- to enjoy freedom of speech and its rewards.

"We have been following speech recognition technology at the Massachusetts General Hospital for over 10 years, mostly with great disappointment, until our experience with the IBM MedSpeak system. From my standpoint of having responsibility for generating more than 400,000 radiology reports each year, I see enormous potential for speech recognition technology to be a valuable and cost-effective tool for radiologists."

Dr. James H. Thrall, Radiologist-in-Chief, Massachusetts General Hospital

"For those familiar with older voice recognition systems, MedSpeak represents an important leap forward. The system allows users to speak completely naturally and accounts for differences in speech patterns, all with a very high accuracy rate. This facilitates reporting and eases the radiologist's workflow."

Dr. Lawrence Schwartz, Department of Radiology, Memorial Sloan-Kettering Cancer Center.

MedSpeak/Radiology is priced at $19,495 for a turnkey 1 workstation system, hardware, software, installation, training and maintenance included.