Voice recognition: ready for prime time?

t h a p c t o t d r a t t oice recognition systems have een advocated for radiology reorting systems as a means of imroving health care by reducing imging reporting turnaround time nd cutting costs. Several serious isues still need to be addressed, inluding decreased radiologist prouctivity, decreased report quality, egative effects on resident educaion, and cost shifting from inexensive medical transcriptionists to xpensive physicians, before such ystems should be universally instiuted. Voice recognition systems are ow being advocated for the dictaion of imaging reports. The appeal f such systems seems to be the intantaneous turnaround time of ictated reports, with essentially eal-time availability to an unlimted number of clinicians (who have lectronic access to reports), and reuced personnel issues, including he quality and availability of medcal transcriptionists and direct or ndirect transcriptionist costs. Venors marketing such systems emhasize deceased turnaround time “only 5 months after going live ith [proprietary name of voice ecognition system], 95% of our eports were transcribed and aproved in less than 1 hour”) and ecreased transcription expenses “with [proprietary name of voice ecognition system], our . . . overll costs were reduced approxiately 87% . . . and we saved over 1 million in transcription serices”) [1]. I believe that at present, voice ecognition systems are a false conomy whose benefits are exagerated and whose costs in terms of octor time, resident education, t