The design of a reconstructive hand surgery text database based on a speech recognition system in Iran

摘要

Speech recognition system (SR) is a tool to create comprehensive, accurate and legible information in the patient's medical records. The present study is a part of a project to create a specialty text database for SR software in the field of reconstructive hand surgery. This project is considered a fundamental work for practical application of the system to enter the healthcare field in Iran. An interventional, quasi-experimental study and fundamental work to create surgical text database in the existing Persian software of "Nevisa". 1863 reconstructive hand surgical descriptions during a 3-month period (March 20, 2012 through June 20, 2012) in 15-Khordad Subspecialty Hospital were used to create database. The statistical population of the study consisted of all patients admitted with hand disorders within this time interval. About 108 dictated voices of physicians were collected after surgery. After the type and required analysis, the reconstructive hand surgery text database was created and tested in a real operating room. In this paper, a reconstructive hand surgery text database was created for Persian SR software. The database contains capital vocabulary and more than 1200 words for use in the hand surgical unit in specialty hospitals and clinics. The record of findings and results of surgery in the patients' medical records is very important. The lack of computerized data recording is one of the reasons for failure and illegibility of medical records in health centers. To solve these problems, software was used for data entry in the medical records and creating comprehensive electronic records. The designed software was tested in the operating room. The obtained accuracy demonstrated the potential for practical applicability of the developed software in Iran.