This is a voice enabled reporting system for Radiology and Emergency Medicine. Speech is converted to text, which is then processed by an NLP system (MedLEE, from Columbia University). The output of the NLP system is represented in XML. The XML document is then processed further to find codes (such as BIRADS or ICD9 codes).
We'd also like to map it to whatever PRA comes up with.
Carol Friedman representing MedLEE Project, Columbia University Medical Informatics Dept.
Project has been underway for between 1 and 2 years. It is currently considered to be an alpha implementation.
No specific problems solved were listed.
Insights: Speech recognition problems have, so far, swamped all other issues as far as users are concerned.
Obstacles: Lack of Standards for XML Clinical Documents
Contact Name: David Rosenthal
Email Address: drosenthal@lhs.com