Study finds high level of errors in speech recognition technology in ED EHR

- Print  | 

New research in the International Journal of Medical Informatics suggests there is a high prevalence of errors from speech recognition technology in emergency departments.

Researchers randomly sampled 100 notes dictated by attending emergency physicians who used speech recognition technology. The notes were collected from the ED's EHR. Two emergency physicians independently conducted an error analysis on the notes.

They found a total of 128 errors in the sampled notes, averaging 1.3 errors per note. Nearly 15 percent of those errors were deemed "critical," meaning they could lead to errors in patient care. Seventy-one percent of the notes had errors.

Researchers noted errors in annunciation were most prevalent (53.9 percent), followed by deletions of words (18 percent) and adding words (11.7 percent). Additionally, 10.9 percent of errors were "nonsense errors," homonyms and spelling errors.

More articles on health IT:

3 problems, solutions to EHR usability, patient safety
The role all executives play in cybersecurity
Global EHR solutions market to reach $30.3B by 2023: 4 key trends

© Copyright ASC COMMUNICATIONS 2019. Interested in LINKING to or REPRINTING this content? View our policies by clicking here.

To receive the latest hospital and health system business and legal news and analysis from Becker's Hospital Review, sign-up for the free Becker's Hospital Review E-weekly by clicking here.