3 Reasons Why Speech Recognition Software Sucks

This sucks

Haters Gonna Hate

Speech recognition seems to be a very polarizing subject.  You either love it, or you hate it.  And more often than not, stories are being told that you hate it.  (David Yeager, 02/10/12, For the Record)

1)  Dictator Must Enunciate
The truth is, many doctors do not take the time necessary to enunciate their dictation well enough for speech recognition to accurately interpret what they’re saying.  Even a simple sentence that’s dictated as, “She is not nauseated” comes out as, “She just got nauseated.”

2)  Doctor Spends Too Much Time Spent Editing
Doctors say they spend too much time editing the mistakes the speech recognition software made.  In a practice already pressed for time, do you really have time to edit your SR dictation?  When you do this, it essentially makes you a highly-paid document editor.  This is inefficient.  Isn’t the whole point of speech recognition to make the documentation process more efficient?  Yes, all dictated material needs to be reviewed for accuracy, even that transcribed by a human.  But because a human transcriptionist has a brain and therefore the ability to intuitively and intelligently interpret human speech, material transcribed by a human takes far less time to edit.

Even worse, speech recognition cannot recognize the subtle nuances in speech that could be cricital to a patient’s diagnosis and/or treatment. For instance, do you want to be checking all the time to make sure your program correct heard “dysphagia” instead of “dysphasia?” Or “abduct” instead of “adduct?” “Anuresis” instead of “enuresis”? “Ensure” (brand name) instead of “ensure” (the verb)? See what I mean? Kind of scary, huh?

3)  SR in Incapable of Thought, Recognition of Local Information
SR doesn’t recognize regional and local landmarks, names of local physicians, and local businesses.  Again, this is why it is far superior to have a human transcriptionist.  Even if your transcriptionist does not live in your area, he or she has innumerable resources at her fingertips to use to learn how “Dr. Loudermilk, neurologist,” spells his name.  When a human brain hears an unusual name (i.e., Loudermilk), it thinks, “Hmm, I don’t know how to spell that.  I need more information.  I’ll use my processing abilities.  I know Dr. Loudermilk practices in the Seattle area, and I know he’s a neurologist.  I’ll look in my resources to locate this person and then be able to transcribe the correct data.”  Speech recognition software with no reasoning ability, on the other hand, hears ‘Loudermilk” and probably types something like “louder milk” without giving it a thought.  (Vincent Desiderio, M.D., 11/13/12, Some Doctors Reluctant to go Digital)

Don’t be fooled by promises that speech recognition will make your life easier, cost less than a human transcriptionist, and eliminate mistakes.  It won’t.  In theory, it sounds great.  But in reality, it can be a nightmare for a busy physician.  Wouldn’t you rather put your patients’ medical records in the hands – and ears – of a human being capable of rational thinking and evaluative skills?

Call us now at 864-905-3559 to find out how simple, cost effective, and time effective it is to have us transcribe your dictation.  We can even import it right into your EMR.

Leave a Comment

Filed under Dictation and EMR

Leave a Reply

Your email address will not be published. Required fields are marked *