Using speech-to-text (STT) technology in nursing health communication
Background: Communication failures at change of nursing shift (handover) or in patient documentation may account for nearly 50% of all adverse events or patient incidents. Aim: We aim to minimise communication breakdowns using STT procedures. Method: Diverse voice-recording (VR) technologies including MP3 players, voice-recorders, mobile phones and IPAD (combined with omnidirectional and noise-cancelling lapel microphones) will be assessed for word recognition accuracy, technical quality and usability. Assessments will be conducted under various conditions including use by different clinicians (diverse speaking styles/accents/voices), single vs. many people, and in diverse settings. Settings will be varied by gradually increasing noise levels and introducing different noise types (e.g., ambient vs. intrusive, interruptions by people, and clinical factors). Outcomes: The accuracy, functionality and feasibility of STT procedures for deriving patient information will be defined. Implications: Use of STT technology at end-of-shift nursing handover may provide the first significant change in decades in how nurses communicate patient information. Applications: STT can be used by all health professionals if effective.
Keywords: computerized patient records; health information technology; information storage and retrieval; natural language processing; nursing records; speech recognition software