Archived Talks

University of Pittsburgh Department of Biomedical Informatics Lecture Series

“AnyModal CDS - A Conversational Documentation System For Structured and Encoded Medical Reports”

Jürgen Fritsch, PhD
Chief Scientist
M*Modal

Friday, January 18, 2008
11:00 AM to 12:00 NOON
Parkvale Building (200 Meyran Avenue), Classroom M-184 (on the mezzanine level)

Abstract: I will present a conversational documentation system developed at M*Modal that allows physicians to document clinical encounters in the fastest and most natural way - using their voice to conversationally narrate their findings - yet generate structured clinical documents that represent the clinical facts of interested in encoded and validated form. One of the key characteristics of this system is a tight integration of speech recognition and natural language processing technologies. Rather than first transcribing a dictated medical report into text and then processing that text with a medical language processor, we perform both of these tasks in one single step, allowing for improved accuracy as well as for a single human reviewing step to validate the correctness of both the textual as well as the coded content. The system accepts and generates documents based on HL7’s Clinical Document Architecture (CDA) and specifically the Continuity of Care Document (CCD) standard for interoperability between clinical systems. Encoding of clinical facts such as medications, allergies, problems and procedures is based on standardized clinical vocabularies and ontologies such as SNOMED-CT, UMLS and ICD-9. After reviewing the system architecture I will focus on two specific aspects in detail: (1) Identification, interpretation and encoding of clinical facts directly from speech, and (2) implicit human validation of the correctness of coded medical information.

For more information: jxc3@pitt.edu or 412.647.7113