A structured reporting system supported by speech recognition for electronic medical records

概要

論文の詳細を見る
Background: Structured reporting provides a great advantage in the secondary use of entered information. Although speech recognition is increasingly popular as a man-machine interface, the results of recognition are often provided by free-text reports. Data in prose form are not efficient for secondary usage.Objectives: To develop a system for judging whether terms entered by speech recognition fit to the preset structured terminology. When the terms fit to the structure, they are stored in a database; otherwise, the system displays an alert to prompt correction of the term. Methods: The system was designed for use on personal computers. The structured terminology is based on the “Minimal Standard Terminology Ver. 2 Japanese version", which was developed for gastroenterological endoscopy reporting. We compared two kinds of reporting methods, data entry through the speech recognition structured reporting system (SRSR) and through conventional handwriting. Results: The average entering time by SRSR was 40Descriptionhorter than that by handwriting. Six of 168 words were recognized incorrectly; and were recognized correctly within second retry. Conclusions: Data entry time, including additional free-text entries, by keyboard was shorter than that by handwriting. The SRSR system also has the potential advantage of hands-free data entry during endoscopy procedures.
The Chiba Medical Society,千葉医学会の論文
2009-02-01