Yuhaniak Irwin Jeannie, Fernando Shawn, Schleyer Titus, Spallek Heiko
Center for Dental Informatics, Department of Biomedical Informatics, University of Pittsburgh, Pittsburgh, PA 15216, United States.
Stud Health Technol Inform. 2007;129(Pt 2):1127-31.
Speech recognition allows clinicians a hands-free option for interacting with computers, which is important for dentists who have difficulty using a keyboard and a mouse when working with patients. While roughly 13% of all general dentists with computers at chairside use speech recognition for data entry, 16% have tried and discontinued using this technology. In this study, researches explored the speech recognition features and functionality of four dental software applications. For each system, the documentation as well as the working program was evaluated to determine speech recognition capabilities. A comparison checklist was created to highlight each program's speech functionality. Next, after the development of charting scripts, feasibility user tests were conducted to determine if performance comparisons could be made across systems. While four systems were evaluated in the feature comparison, only two of the systems were reviewed during the feasibility user tests. Results show that current speech functionality, instead of being intuitive, is directly comparable to using a mouse. Further, systems require memorizing an enormous amount of specific terminology opposed to using natural language. User testing is a feasible way to measure the performance of speech recognition across systems and will be conducted in the near future. Overall, limited speech functionality reduces the ability of clinicians to interact directly with the computer during clinical care. This can hinder the benefits of electronic patient records and clinical decision support systems.
语音识别使临床医生在与计算机交互时无需手动操作,这对于在为患者治疗时难以使用键盘和鼠标的牙医来说非常重要。虽然大约13%在诊疗椅旁配备计算机的普通牙医使用语音识别进行数据录入,但16%的人曾尝试使用该技术但后来不再使用。在本研究中,研究人员探索了四种牙科软件应用程序的语音识别特性和功能。对于每个系统,评估了文档以及工作程序以确定语音识别能力。创建了一个比较清单以突出每个程序的语音功能。接下来,在制定图表脚本后,进行了可行性用户测试,以确定是否可以在不同系统之间进行性能比较。虽然在功能比较中评估了四个系统,但在可行性用户测试中只审查了其中两个系统。结果表明,当前的语音功能并不直观,而是与使用鼠标直接相当。此外,系统需要记住大量特定术语,而不是使用自然语言。用户测试是衡量不同系统语音识别性能的一种可行方法,并且将在不久的将来进行。总体而言,有限的语音功能降低了临床医生在临床护理期间直接与计算机交互的能力。这可能会阻碍电子病历和临床决策支持系统的益处。