Is Speech Rec Wrecked?
In response to my fellow HIM Insider blogger Jeanne Johnston's post entitled "
Speech Wreck," I'd like to offer my own perspective on the use of speech recognition technology (SRT) in medical transcription. I have a fair amount of experience with back-end SRT, both as a working editor and in a supervisory capacity in a production environment. Over the past few years I've also managed to collect a lot of anecdotal and statistical information about how SRT impacts working transcriptionists.
I will tell you up front that I've long been a vocal advocate for fairness in compensation for SRT editors, as well as a vocal critic of the way SRT is often marketed. I believe there is plenty of objective evidence to prove that a good speech recognition platform will pay for itself on the basis of increased productivity alone without having to reduce the net compensation for editors. By that I mean that if SRT provides a 50% increase in productivity in the number of lines, characters, etc. that can be produced by a given number of editors in a given period of time, significant savings will be realized even if the payroll for editors remains the same. For the same money, the editors are producing 50% more work in the same amount of time. That's a win for the institution any way you look at it. For this reason, I get very cranky when I hear that SRT is being used as an excuse to reduce compensation for editors to the point where they make less money in a given period of time. But to be honest, some of the blame for that kind of scenario has to rest on the shoulders of the editors who put up with that kind of unfair treatment.
In 2006 I put together a small-scale survey (202 respondents) of MTs who had been working as speech recognition editors for at least six months, to find out how SR was impacting productivity and compensation. The results showed the following:
- - 28% of respondents found that editing SR took about the same amount of time it would have taken to keyboard manually, meaning no productivity gain overall.
- - 32% saw a 25% increase in productivity compared to manual keyboarding
- - 19% saw a 50% increase in productivity
- - 10% saw a 75% increase in productivity
- - 9% saw a 100% increase in productivity, i.e., double the amount of lines, characters, etc. in the same time period.
- - 2% saw a greater than 100% increase in productivity.
What this tells us is that the largest bloc of SR editors--a total of 51% of respondents--saw an average increase in productivity of between 25% and 50%. This confirms the anecdotal information I had collected via informal conversations with MTs working as SR editors in a variety of situations on a variety of SRT platforms.
In the same survey, I asked respondents how their rate of compensation had changed as a result of working as editors compared to traditional transcriptionists. Here are those results:
- - 31% received the same rate of compensation for editing as they had received for transcription
- - 18% received a rate that was 25% less for editing as compared to transcription
- - 9% received a rate that was 33% less
- - 27% received a rate that was 50% less, i.e., half of the rate they received for transcription
- - 9% received a rate that was 67% less
- - 6% received a rate that was 75% less
My interpretation of these figures is that while there was no doubt a statistically significant number of editors whose reduced rate of compensation combined with a relatively small increase in productivity had resulted in a decrease in their overall compensation, that group did not by any means represent a majority of SR editors.
A third question I asked respondents was how satisfied overall they were working as SR editors as compared to transcribing. The results I found quite fascinating:
- - 21% said they were very dissatisfied
- - 22% said they were somewhat dissatisfied. These two categories totaled about 43%
- - 31% said they were somewhat satisfied
- - 26% said they were very satisfied. These two categories totaled 57%
At least based on the results of this admittedly small sampling of editors, there was not an overwhelmingly negative attitude toward using SR, and in fact a solid majority was either somewhat or very satisfied with their overall experience with SR.
The overall picture I get from the survey results as well as the anecdotal information I've come across is that there is a wide spectrum in terms of the impact of SRT on productivity, compensation, and overall satisfaction among MTs working as SR editors. Consequently, I don't believe there is enough objective evidence to conclude that speech recognition has proven to be a widespread disaster for the MT working class. As with any scenario involving people, technology, and money, mileage is going to vary widely. In my experience, there are simply too many factors that can influence productivity, compensation, and overall satisfaction with speech recognition technology to draw hard and fast conclusions about the impact SRT is having on working MTs on the whole.
Having said that, as long as ANY working MTs or SR editors are being taken advantage of, there will always be a need for advocacy for fair compensation. Nothing I have written here is intended to minimize the culpability of employers who treat employees unfairly. But again, by far the most effective advocacy for fair compensation consists of being willing to say "No" to substandard compensation. Until MTs and editors are willing take that stand en masse, there will always be someone willing to squeeze one more drop of blood from the stone.