Alternative Input for Perfusion Management Devices: Voice Recognition for Data Input and the Effects on Charting and Perioperative Calculation Use

Abstract

Technology in healthcare has become increasingly prevalent and user friendly. In the last decade, advances in hands-free methods of data input have become more viable in a variety of medical professions. The aim of this study was to assess the advantages or disadvantages of hands-free charting through a voice-to-text app designed for perfusionists. Twelve clinical perfusion students using two different simulated bypass cases were recorded and assessed for the number of events noticed and charted, as well as the speed at which they accomplished these steps. Paper charts were compared with a custom app with voice-to-text charting capability. Data was analyzed using linear mixed models to detect differences in length of time until a chartable event was noticed, and how long after noticing an event it took to record the event. Timeliness of recording an event was made by assessing log-transformed time data. There was significantly more information recorded when charting on paper, while charting with voice-to-text resulted in significantly faster mean time from noticing an event to the recording of it. There was no significant difference between how many events were noticed and recorded. When using paper charting, a higher percentage of events that were missed were drug administration events, while voice charting had a higher percentage of missed events that were associated with cardioplegia delivery or bypass timing. With a decreased time interval between noticing an event and charting the event, speech-to-text for perfusion could be of benefit in situations where many events occur at once, such as emergency situations or highly active portions of bypass such as initiation and termination. While efforts were made to make the app as intuitive as possible, there is room for improvement.