@jeffmarc Hey Jeff..."WOW"...Your quote of the radiologist saying: "“I was using voice recognition to make that finding and it incorrectly wrote what I actually meant. There is no problem in that area." That is so scary. That tells you...all of us...that physicians using this voice recognition dictation technology, are not even "reviewing" the printed document for accuracy of what they said. They presume the technology is advanced enough to accurately place into printed form, that which they speak. With all due respect to physicians who speak English as their second language and often with a moderate to heavy accent, I don't even know how automated voice recognition works. Perhaps when they first start using it, there is a start up protocol where they must speak a prepared document into the system, so that the system can account for - recognize - any subtle to profound differences in the pronunciation of the English language by the user. That then forms the baseline for how the software will interpret everything they speak in the future...maybe??? I don't know...I have never used it.
These physicians must not be on Instagram and other social media sites that use voice recognition, because we have all seen written text scrolling at the bottom of an Instagram post where we "hear" what is being said (and in perfect English), but the printed words that are scrolling are completely wrong. Utterly scary. These physicians need to slow down and review/edit what they dictated. That is simple "quality control." As a former Director of Clinical and Anatomic Laboratory, it would be like testing patient blood samples without calibrating the instruments and running high, normal, and low quality control material to verify the accuracy of testing, and then just reporting to the doctor whatever the instrument measured. It could all be wrong! In the good old days, their medical transcriber who actually typed from their audio tape, would have enough knowledge to question what was spoken..."ask the physician for whom they were transcribing"...what they actually meant to say. This auto-dictation and artificial intelligence using voice recognition is clearly not developed well enough to be used in such critical applications, such as dictating critical pathology of cancer patients' biopsies and post-surgical tissue/slide examinations. SO, SO SCARY.