AHRQ Study Estimates Frequency of Errors When Speech Recognition Technology Is Used To Enter Patient Data
Critical errors were found in 15 percent of data entered into patients’ electronic health records by physicians using computerized speech recognition technology, according to a pilot study funded by AHRQ. The study evaluated front-end speech recognition technology that allows dictation and editing in an electronic record’s text field. The study examined speech recognition errors based on 100 patient notes by attending emergency department physicians in Boston from January to June 2012. Findings showed that there were 128 errors, or 1.3 errors per note, and that of the 71 percent of notes that contained errors, 15 percent contained one or more critical errors that could potentially lead to miscommunication affecting patient care. Annunciation errors were most common, followed by deletions and added words. Study findings represent the first estimates of speech recognition errors in dictated emergency department notes, researchers said. The study, “Incidence of Speech Recognition Errors in the Emergency Department,” andabstract were published in the International Journal of Medical Informatics.Incidence of speech recognition errors in the emergency department. - PubMed - NCBI
Int J Med Inform. 2016 Sep;93:70-3. doi: 10.1016/j.ijmedinf.2016.05.005. Epub 2016 May 26.
Incidence of speech recognition errors in the emergency department.
Copyright © 2016 Elsevier Ireland Ltd. All rights reserved.
Emergency medicine; Patient safety; Speech recognition