Quality Health Information – a Transcription Wake-up Call


Image via Wikipedia

“This is better quality than I get from <insert MTSO company name here>”

That is a comment from a doctor that I gave a demo to at RSNA this year. He dictated into our front-end speech recognition, using an untrained author profile, and was pleased with his results. He is a radiologist who works at a hospital that is telling him he should be using front-end speech recognition. He was resistant – why should he have to spend time editing when he could be reading studies? But – as he explained to me – if the documents he gets back from a medical transcriptionist require as many edits as he’d have to make using front-end speech recognition, why should he pay for transcription?

I heard a similar story from a customer yesterday. Some of this customer’s physicians were using front-end speech at one point, but quality of the information degraded because physicians weren’t correcting the errors in the speech recognized text. So they brought the MTs back. But as often happens when a service initially transitions to editing, MTs are missing errors and delivering poor-quality documents back to the physician. And again – if physicians are seeing errors in their transcription, then why not use front-end speech and correct their own errors?

This is a wake-up call for anyone in the transcription business. Quality and integrity of the health information that you deliver is your only defense against the push to get physicians to do their own documentation. If your quality is in question – your customers have no reason to keep you. That is harsh – but it is reality. You can talk about transcription’s role in meaningful use, and in allowing the physician to focus on the patient instead of on a keyboard until the cows come home. But if you don’t deliver quality – nothing else matters.

Many users react to MT quality problems by blaming the technology or by removing the physician from speech recognition. They want to know, “did this error come from the MT or from the speech recognition?”

Guess what the answer to that question is?  Errors delivered back to the physician are not the problem of the physician. They are not the result of using speech recognition. They are MT errors. Always.

If an MT typed a document from scratch, and delivered errors back to the physician, would you remove that physician from transcription?  Of course not – removing the physician has a direct impact on your revenue – and you wouldn’t be likely to keep your customer if you told them that your MTs don’t have the skills to produce quality documents for their physicians.

When faced with MT errors in transcription, you deal with the MT. An MT who is editing should be handled no differently than an MT who is transcribing from scratch.

When making the transition to speech recognition, or if you are facing quality complaints, ask yourselves the following questions:

  1. Have you assumed that speech recognition would allow you to use lesser-skilled, cheaper MTs?
  2. Have you assumed that speech recognition means that MTs should no longer have to send questions or blanks to QA?
  3. Are you doing QA on MTs as they transition to speech?
  4. Have the MTs been adequately trained to edit? Are they getting follow up?
  5. Did the MT training include attention to quality – or just attention to productivity?
  6. Did you cut MT line rates such that they have no choice but to rush through their work in order to keep their pay whole?
  7. Do you financially motivate MTs to produce work fast? Or to produce quality documentation?

Remember – an MT who has had sufficient training and can make edits quickly and efficiently is much more likely to pay attention to quality than one for whom every edit is money out of her pocket. The most proficient MT editors I’ve seen have also been the highest quality editors I’ve seen. Why? Because making edits is no threat to them – they can fly through documents containing multiple errors almost as quickly as they can through a near-perfect document.

Lastly, I am going to take this discussion back to the topic of compensation and management. The potential ROI for the use of speech recognition goes far beyond cuts in MT line rates. If you don’t manage to – and pay for – quality, then stick a fork in transcription folks. It’s done.

Need more on compensation and management?  See previous articles here and here and here and here.

Til Next Time,


16 Responses

  1. This is a cold hard truth that we have to come to terms with but few have the courage to speak about. I have believed for years that if we (as an industry) had been delivering on the quality that we advertise, more docs would be coming to our defense. Some doctors are pleased with our work and advocate for our continued presence, but until we clean up the industry and truly deliver on quality, there will be doctors that, like the one described above, believe that “free garbage” (ie FESR) is better than “paid garbage” (poor transcription).

    • Hi Laura: How are you? I like your line about “free garbage” being better than “paid garbage”…that is just how they look at it. I agree with you 100%.

      Thanks for commenting!


  2. “Errors delivered back to the physician are not the problem of the physician. They are not the result of using speech recognition. They are MT errors. Always.”

    This is not entirely true. I’ve been in this business 20 years, and I’ve encountered things in dictations that had to be heard to be believed: Doctors eating, talking to other people, trying to talk faster than they’re actually capable of talking, heavy accents, slurred speech – the list goes on. Sometimes I’ve heard all of these things in the same dictation. There’s also a learning curve with every new provider, to learn their style of speech and any quirks of their grammar.

    Most doctors are never taught how to dictate, and it shows. I’ve transcribed dictations for second-year residents that were clear and easy to understand, and I’ve also transcribed reports for doctors who have been in the field for years and who can’t put together a simple sentence. Some providers are speaking English as a second, third, or in some cases even a fourth or fifth language. While I admire people who can do this, it often makes transcribing their dictated reports a real challenge.

    • Hi Cynthia: Thanks for the comment! Yes, of course the doctors can sometimes dictate something that is completely incorrect and which you can’t possible be expected to correct for him. But I bet in your 20 years, you have placed many a blank in the document for the spots in his dictation that you couldn’t possibly be able to make out, and I bet you also were able to deduce what he meant to say, or at the very least were able to point out to someone that he dictated a medication or left/right discrepancy, or something like that. My point is that it is the MT’s responsibility to catch the errors in the draft text produced by speech recognition – just as you have done your best to produce a quality document, regardless of the physician’s particular challenges, in traditional transcription. Make sense?

      I think you make a good point though. Back when I was managing transcription, MTs were expected to send anything to QA that they couldn’t be sure about. Now I know many companies ding an MT for sending work to QA. And that also points back to the fact that management must bring the focus back to quality – or they’ll hasten their own demise.

      Thanks again!!

  3. Wow Lynn! What a fascinating article.

    You have touched on so many significant issues in this thought-provoking article. It is so relevant to the current state of transcription and its future direction and potential.

    I think you are absolutely right that the Quality and Integrity of health information delivered by MTSOs is more important than ever and that a solid QA program is absolutely imperative. I know that many MTSOs and MTs are feeling anxious and threatened, at this point, by the looming specters of both Speech Recognition and EHR/EMR technologies. They are perceived as Scylla and Charybdis – monsters that can destroy them – and that need to be carefully understood and navigated around. The winds of change are blowing hard and it is such a challenge to keep up with learning new platforms and tools that MTSOs are tempted to let their quality programs lapse, or to put off putting them in place, because they are too busy learning the new platforms etc. And yet, this is when a good quality program is most important. A good Quality Program is like a GPS – helps you to get your bearing – assess your current location – and then helps you to steer to a better course.

    A good Quality Program will support MTs through the transition from direct transcription to editing, or to entering health information into electronic medical records. It will provide the MTs with feedback and support, to help them to build up their proficiency and confidence more quickly, and help the MTSO to determine when they are proficient enough to no longer need daily feedback and support. A good QA program will help the MTSO to avoid being dashed on the rocks or sucked into the whirlpool of change.

    Last year, AHDI, MTIA (now CDIA) and AHIMA approved new QA Best Practices (available at http://www.ahdionline.org/Portals/0/downloads/QA_Best_Practices.pdf), to better address the changing transcription environment. This year, AHDI will soon be releasing a QA Best Practices Toolkit to help MTSOs to implement a QA program based on those new standards. I had the honor of participating in the committees that produced the standards and the toolkit, which inspired me to update my TQAudit software to support the new standards, as well as the original standards introduced by AAMT (now AHDI) in 2005. TQAudit is fully customizable so that MTSOs can also transition to adopting the new standards incrementally, and partially, and perhaps differently for different platforms or accounts, based on different TQAudit profiles.

    One of the most thought-provoking concepts in your article is:

    Errors delivered back to the physician are not the problem of the physician. They are not the result of using speech recognition. They are MT errors. Always.

    It brings into question so many issues:
    * who/what created the problem (or error)?
    * who/what should fix the problem (or error)?

    Ultimately, the transcriptionist (or MTSO) is uniquely positioned to:
    1. recognize that there is an error (problem)
    2. propose the best correction or fix (for the problem)
    3. bring it to the attention of the proper parties, depending on the situation.

    The new Best Practices document, mentioned above, focuses on patient safety and also stresses that:

    Integrity is achieved in healthcare documentation through a partnership between the author and the transcriptionist who transcribes and edits the report.

    And continues with:

    The author is responsible for clear, unambiguous, and complete dictation. The MT is responsible for preserving the author’s style and intended meaning with reports transcribed or edited in their entirety, including accurate demographics and appropriate distribution notations. Additionally, the MT brings integrity to the process through continuing education and commitment to the documentation process. Reporting errors as well as problematic practices that could potentially cause errors bring further integrity to the process.

    Thus, there might be times when the physician (author or dictator) really is the one who caused the problem, or perhaps due to a noisy dictation environment, the dictated words were just not at all audible.

    When Gary David reviewed my TQAudit software with me, last August, he brought up the need to document dictator issues in order to show transcription value add. Lynn, I recall in one of your presentations at an AHDI function, you brought up the point that even when the speech rec. engine is 100% true to the dictated verbiage, if the author misspoke, then the error of that poor dictation will remain in the document unless a transcriptionist/editor catches it. I think this is a very vital argument for keeping skilled medical language experts involved with the production of medical records, in whatever form them take. Doctors are human too, and yes, they sometimes do make mistakes. TQAudit features the ability to capture, track, and report on dictation errors as well as transcription errors. I am interested in seeing how they decide to use this information.

    I strongly encourage MTSOs to review the new QA Best Practices. They might or might not totally embrace everything in it, but I’m sure they will get some good ideas from it that will help them through these turbulent times. Like me, they may benefit from a fresh perspective and draw inspiration from it. The fact that it is approved by AHIMA as well as AHDI and MTIA is huge, I think. Most importantly, good quality transcription can save lives and improve health care outcomes.

    Thank you Lynn, for such an interesting and though-provoking post!

    • Thank you Christine! I hope this information helps – thank you for providing the links and suggestions!

  4. Thanks Lynn,

    Well, this is embarrassing … I had a mistake on my web site address that my name links to in the previous comment. It should be http://www.tyrrellsw.com with 2 r’s, but somehow I had 3 r’s in there. What do think Lynn, can I blame it on the technology (your web-site) or do think it was actually the author (me).

    Just goes to show how easy it is to make mistakes when doing your own documentation on line 🙂

    My humble apologies to anyone who tried to click my link above. I really am a bona-fide software developer who is very interested in supporting the transcription industry through these challenging times … even though I can’t type too well!

    • Oh we all make mistakes Christine! Thanks again for the information!


  5. What comes to my mind after reading this article is
    OUCH, ouch, ouch!

    I will definitely share this article with my staff, and other transcription managers in our system.

    • Oh no! I hope not too many Ouches!!! 🙂 Thanks so much for your comment!


  6. I am all for a quality document and do my best to produce them but when you have ESL dictators with multiple unintelligible words in a dictation and will be paid less (several cents per line less) because of sending this report to QA, what is an MT to do? We are being punished financially for trying to send in a correct document. Why is this happening?

    • Hello MT: I find this comment to be disturbing. I have heard that some companies are decreasing MT line rates for any documents sent to QA over a certain threshold. My question is – do they provide you the means to help you to submit these documents without QA? Do you have resources available, for example, to lend an ear to see if someone else can make out the unintelligible word? Is there any way a QA person can validate that the terms were indeed unintelligible, and that the MT was justified in sending the document to QA? I am going to talk about this in today’s blog.

      Thank you for commenting – and all the best to you.

      • Lynn, Thank you for your reply to my comment. I do have some resources from my company but not an extra ear to listen. We technically do not have QA people anymore but MTs that have taken a test and passed and now they are the “QA people” when one sends a report to QA. It doesn’t matter if all the blanks were left blank in every report sent to QA, what goes to QA is counted against you even if they know they have a problem dictator (ESL, too fast, cell phone user with interference, etc.). I am seriously considering another profession due to reduced pay for voice recognition, support for MT decreasing, no raises for years, company seeming to only care about quantity and not quality even though claiming otherwise, sending as much work as possible overseas, etc. I’ve been doing this for 10+ years and have had about as much as I can take. We need good people like you so keep up the good work and thank you so much! I just wish all the MT companies will hear you and take a look at what they are doing.

      • Hello MT: I sincerely hope this can be worked out with your employer – the industry sure can’t afford to lose skilled MTs.

        I’d be interested in hearing other people’s thoughts…any suggestions or comments?


  7. Excellent article – thank you! There are so many out there that really do not believe quality is #1. I will share this with many MTs.

    • Thanks for your comment Jennifer!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: