Project BriefInformation Infrastructure For Healthcare (September 1995)Intelligent Spoken Medical RecordsDevelop prototype multimedia technology for electronic patient records that allows users to input and retrieve patient data using natural, unconstrained language, and that uses intelligent information processing methods and event-tracking techniques to summarize relevant information and highlight important changes. Sponsor: Berdy Medical Systems365 West Passaic StreetRochelle Park, NJ 07662
Berdy Medical Systems proposes an innovative, technically challenging approach to building a spoken-language interface for an electronic patient record system. Harnessing information technology to improve the handling of patient records -particularly to gather, correlate, and analyze patient data from a variety of normally isolated records -- is generally acknowledged as an important goal for significantly reducing healthcare costs while improving patient care. To be useful, however, such a system must give healthcare professionals a convenient, accessible user interface. Several research groups are studying multimedia, spoken-language systems that allow the user to enter data and request information in natural language. Such a system must be fast and accurate to be useful, but to achieve this, existing language-recognition systems have had to make compromises: they are "tuned" to be dependent on a particular speaker, or have small vocabulary, or use a very stylized and rigid syntax. Berdy Medical proposes to avoid these limitations with a dynamic system that uses the current context of the dialog to limit the search for solutions, allowing the physician to use natural, unconstrained language without artificial limitations. Another potential problem with such systems is, paradoxically, that they can provide the physician with too much information to effectively assimilate and use. The software R&D proposed by Berdy Medical could result in new, intelligent information processing and machine learning techniques to distill summaries of information from multiple sources, avoiding redundancies or irrelevant information. The proposed system also will be capable of detecting and highlighting trends in data over time, as an aid to improved diagnoses. The Berdy system will develop in-depth capabilities in two specialties, anesthesia-pain management and oncology, and more general capabilities for other specialties.
|
|
ATP website comments: webmaster-atp@nist.gov
Privacy Statement / Security Notice • NIST Disclaimer • NIST Information Quality Standards NIST is an agency of the U.S. Commerce Department |