May 7, 2012

MTs Need to Get Off Inaccurate Pay Scale
For The Record
Vol. 24 No. 9 P. 29

I need to vent. It’s partly in response to the article by Elisabeth Altieri, BA, RHIT, in the March 26 issue of For The Record (“MTs Have Skills, so Let’s Use Them”) and partly in response to a comment that was recently made to me.

Addressing the article first, my impression is that Altieri is championing the medical transcriptionist (MT) cause, attempting to bolster our beleaguered egos, encouraging us to “get out there” and ... I don’t know ... do something. I’ve been a secretary, a physician assistant (PA) for 30 years, and an MT for 20. Yes, there are bosses who are honest enough to admit their secretaries are valuable. In fact, their secretaries are often referred to as administrative assistants, and their salaries are usually above secretary-level pay. If the secretary is a man, I doubt he is ever actually called a secretary, and I bet he doesn’t make standard secretarial pay.

In the transcription field, since almost all workers are women and most started out (at least in the old days) as secretaries, that’s as far as we got pay-wise. We have never risen above that level. Altieri makes a good point: MTs have been invisible for several reasons, and all the advances in technology have served to make us even more invisible. Out of sight, out of mind. I used to have one doctor who ended his dictations with “Thank you, operator.”

Now with speech recognition/voice recognition (SR/VR) programs, there is a whole other gestalt. There are issues on both the dictator and the MT/editor side. As a PA/MT/editor, I’m familiar with both sides.

Starting with the dictation side, the techniques for best practices are somewhat different when you’re dictating to an MT vs. dictating to an SR/VR engine for future editing by an MT. For example, in the past when I’ve dictated, I’ve spelled out the name of a doctor, a local town, or a medication that I’m not certain the MT will be familiar with (after all, the MT may live in India and not know how to spell Tamarac). With SR/VR, it’s pointless to spell out words—the software attempts to form words out of the letters you’re saying and comes up with a string of meaningless garbage. In addition, if you need to go back to add, delete, or change something you said four paragraphs prior, the SR/VR software isn’t going to do this for you. It just keeps rolling right along typing what you said, and you have a whole bunch of confusing nonsense smack in the middle of your report. A human being, on the other hand, will recognize that you’ve asked for the accommodation, find the section of the report to which you referred, make the correction, go back to where you left off, and pick up seamlessly. The people who are dictating don’t know if they’re dictating to a human or an SR/VR engine and wouldn’t know how to adjust their dictation styles even if they did.

On the transcription side, typing a document from scratch and editing an SR/VR-generated document are totally different skill sets. Good MTs get “in a zone” when they’re transcribing a report, making certain critical decisions and judgments. Did the doctor really mean left or did he mean right? Did the doctor say Zoloft or Zocor? Celebrex or Cerebyx? When the document is in front of you, although you don’t have to physically type an entire document, you are forced to decipher whether what is there is correct because the SR/VR software can’t. Then you need to make the corrections, which often involves deleting long strings of junk but may involve changing only one or two letters.

Aside from the obvious danger of wrong medications in the patient record, there are some errors that will escape the editor because they’re insidious. For example, the dosage is the same for two medications, so a flag won’t go up if you’re just skimming the document and not paying close attention because it “looks” correct. Standard doses of Flexeril and Dulcolax are 10 mg. Lamictal and Pyridium are both dosed at 100 mg. These medications do not sound or look alike, but they were incorrectly interpreted by SR/VR software. And were you alert to the fact that the patient in the report you’re working on is a man, but it says he’s on oxytocin, which is a drug intended to induce or assist labor in pregnant women? The SR/VR software “heard” that rather than doxazosin.

Which brings me to the comment that also spurred me to write this response: When I asked why I receive 10 cents per line to do transcription but only 4 cents per line to do SR/VR, I was told that it’s because I’m doing fewer keystrokes.

We have all these people—Altieri included—who tell us that we MTs/editors are valuable because we have numerous skill sets. We are so valuable that we can even move laterally to other jobs such as HIM. Altieri listed several MT skills but failed to mention typing. Why? Because computer programs can type! Typing is not where the value of an MT lies! And, by extension, typing is not how our pay should be decided. Typing more or typing less has nothing to do with the skills we bring to the table in our jobs.

The human who has to edit the garbage put out by the program requires more alertness, more attention to detail, more knowledge, more critical thinking skills. With all the substandard junk that is going into the SR/VR engines, MTs are being thrown under the bus. The accuracy of SR/VR is nowhere near what we were originally promised. In the beginning, we were told that unless the dictator’s output documents couldn’t achieve at least 80% to 90% accuracy, their dictations would not go to SR/VR. Now who knows what the accuracy rate is? I do just as much typing/correcting in SR/VR as I did in transcription but with more strain on my wrists because I’m doing less smooth typing and more jumping around, deleting, typing, scrolling, typing, cursoring, typing.

When I discussed this with somebody who is “high up” in the industry, she said that for us to keep up production and make money, SR/VR editors have to “eliminate the niceties,” by which she meant poor punctuation, misspellings, little words, word order, etc. In other words, only focus on the most egregious errors and don’t worry if the document isn’t a nice perfect package when you’re done. When I transcribe, my ID number is on the bottom of that report, but when I edit, there is no way to know it was me because all edited SR/VR reports at my company have the same generic ID number on them. So who would know?

I would. My work ethic and integrity are old school. I won’t compromise them for money, but I also can’t afford to work for 4 cents per line. How much longer can we maintain this? Unless and until the quality of SR/VR output improves and unless/until MTs and editors are paid for their skills and not their keystrokes, we are spiraling downward to doom. And I will leave the industry if things don’t change soon.

Sherry Roth, PA-C, DF-AAPA, CMT, AHDI-F
Coral Springs, Florida