Voice Recognition Software
By Pat Bowen, CMT
After receiving several calls regarding voice recognition software (VRS) and its threat to our livelihood, I thought it would be a good time to address this topic once again. Personally, I’m not terribly concerned about VRS taking our jobs anytime soon. I have been fortunate to talk to many people about this, as well as see it demonstrated on several occasions. Here’s the scoop, as far as what I’ve seen and heard.
Voice recognition software can be added to any computer and is getting cheaper all the time. Basically, you talk into a microphone and your words appear on the computer screen. The user is allowed to enter text, correct errors, make revisions, add punctuation, etc. To "train" the VRS, the user will need to repeat 450 words three times each. After finishing the training, your voice files are built based on your unique speech pattern and are continuously adapting to your voice.
Sounds good so far, right? Well, here’s where the reassuring part comes in for me. Not long ago, I was channel surfing and stopped on one of the home shopping channels that was selling VRS. The sales rep from the company, who likely does these demonstrations every day and already had the VRS trained to her voice, demonstrated the software by writing (dictating) a short letter. Let me tell you … I was not impressed! I wish I had videotaped it so I could tell you exactly how it went, but I will try to summarize.
In the time it took the sales rep to finish the letter, I could have transcribed the same thing three times. The software consistently spit out the wrong words (some of them not even close!), e.g., address instead of addendum. She would then have to tell it to go back up and edit. The funniest part is, while she was bragging about her newly completed letter, a close-up of the three paragraphs showed at least three errors! Mind you, this is someone who is experienced (and selling) the program.
As medical transcriptionists, we know there are numerous sound-alikes in the medical language. Part of our job is to make sure the correct words are used in reports. Imagine how easy it would be for the computer to type "15 mg" instead of "50 mg," "Restoril" instead of "Zestril," "ilium," instead of "ileum."
Another part of our job is making sure the dictation is accurate. We make mistakes … and dictators make mistakes! We also monitor for consistency or palatability of dictation. An MT would definitely want to change, "The patient has had galactorrhea since the birth of her baby from both breasts," to "The patient has had galactorrhea from both breasts since the birth of her baby." How about those dictators that chronically spell things wrong? "Transcriptionist, that’s spelled P-R-O-V-O-C-H-O-L." "Excuse me, Doc, but it’s spelled Pravachol." If we are not there to catch these errors, who will? The next medical professional reading the report in the patient’s chart?
Remember, for most doctors, dictation is a chore they despise. They rush through it, talking as fast as they can, using all of the verbal shortcuts they can think of, and trusting us, the medical transcriptionists, to decipher the mess. For a physician to use VRS, he or she would have to speak very slowly and distinctly, spelling out many words, and correcting errors on the computer. How many doctors are likely to run out and purchase this program? Not many that I know. The biggest drawback, I see, is not being able to dictate an operative report while driving through a car wash, eating lunch, and playing classical music so loud the windows rattle.
VRS is useful to those who are no longer able to use their hands, e.g., MTs with CTS (carpal tunnel syndrome). However, VRS has created a new affliction … RSD (repetitive speech disorder). It is possible to seriously injure your voice through overuse/misuse just as quickly as your hands. No kidding!
Okay, say VRS is perfected to the point that it is easy to use. And, who knows? That could be next year. I believe we will still be able to use our language skills as "medical editors." Voice recognition software may replace the keyboard, but it will not replace the brains behind it.
References:
"Voice Recognition ... Say What?" by Bob Willard. MT Monthly, 12/94.
"Hands-Free Transcription?" by Doris Anderson, MT Monthly, 12/94.
"Making Voice Recognition Software Work for Us," by Barbara Grow, MT Monthly, 12/94.