Speech recognition software meant to streamline clinical documentation within an EHR increases the risk of documentation errors, including those with the potential to cause patient harm, according to a study in the Journal of the American Medical Informatics Association.
The researchers — led by Tobias Hodgson, a PhD candidate at North Ryde, Australia-based Macquarie University's Centre for Health Informatics — had 35 emergency department clinicians complete clinical documentation tasks in an EHR. The clinicians used either speech recognition assisted clinical documentation or standard keyboard and mouse clinical documentation.
The clinicians' mean task completion times were 18.11 percent slower when using speech recognition software compared to using a standard keyboard and mouse. The researchers also observed more errors when clinicians used speech recognition software, including those with the potential to cause clinical harm.
"Use of [speech recognition] to drive interactive clinical documentation in the EHR requires careful evaluation," the study authors concluded, noting some errors may have been caused by poor EHR integration. "Current generation implementations may require significant development before they are safe and effective."