Speech Technology Solution for Dental Healthcare System

A significant part of dental healthcare professionals’ work time goes into documenting and reporting patient related information. Documenting during a medical procedure can often be cumbersome or even impossible, especially for alone working clinicians. Conventional input devices also often cause ergonomic and hygiene problems in medical environment and thus, many clinical tasks require the working input of two healthcare professionals. The target of this thesis was to study modern speech technology and develop a speech technology solution that can be used in dental healthcare and patient information management context. Speech technology solution is intended to ease the workload of dental healthcare professionals and add automation into the documentation process. The primary application target of the speech solution is the task of dental check-up. During the dental check-up clinician goes through all patient’s teeth and documents the dental status observations from each tooth into the patient information system. For clinicians who work alone, this task is especially cumbersome and time consuming. With the developed speech solution clinicians are able to document and add teeth related status observations into the patient information system via speech commands. The speech technology solution was developed using Microsoft Speech Platform technology, the tools it provides and agile user-centered-design principles. In order to evaluate speech solution’s performance and viability for the task it was intended, a usability testing was conducted. In the usability testing, a group of test participants used the speech solution to perform tasks in its intended end-use context. In documenting the dental status observations with the speech solution, test participants achieved an aver-age success score of 96 percent. Solution’s speech recognition accuracy ranged from 87 to 100 percent between the test participants. The research and speech recognition solution developed as result of this thesis, works as a pilot project for enhancing patient information system with speech recognition and speech synthesis. The built speech recognition solution was unveiled and demonstrated publicly to end users at Finnish Dental Congress and Exhibition in November 2018