The integration of a continuous-speech-recognition system with the QMR diagnostic program

Area of Science:

• Medical Informatics
• Speech Recognition Technology
• Clinical Decision Support Systems

Background:

• Physician data entry for clinical decision support systems like Quick Medical Reference (QMR) can be time-consuming.
• Existing systems often lack efficient methods for incorporating spoken observations from physical examinations.

Purpose of the Study:

• To develop and evaluate a continuous-speech interface for the Quick Medical Reference (QMR) system.
• To address the challenges of using natural language medical terminology in speech recognition.

Main Methods:

• Designed a continuous-speech recognition interface for QMR.
• Developed a method for matching spoken medical findings to standardized QMR terms.
• Utilized a semantic representation of findings to improve recognition accuracy and minimize misrecognition effects.

Main Results:

• Successfully created an interface allowing physicians to input physical examination findings via continuous speech.
• Demonstrated a method for accurately mapping natural language spoken terms to QMR database entries.
• The semantic approach effectively handled variations in spoken medical terminology.

Conclusions:

• Continuous-speech interfaces can significantly enhance data input for medical reference systems.
• The proposed method offers a robust solution for recognizing and integrating spoken medical observations.
• This technology has the potential to streamline clinical workflows and improve the usability of electronic health records.