Letters, 5/15

Create PDF

Speech recognition

I found the article on speech recognition to be well reported and interesting (“Hear me now? Another audition for speech recognition,” March 2015). We implemented speech recognition at NorthShore University HealthSystem (as your writer reported) before I retired as director of the clinical pathology division. What William Watkin, MD, said in the article about our smooth implementation was so true. One of our pathologists was so computer-averse that he did not use email but he had no difficulty using VoiceOver. Another pathologist whose English is a second language and sometimes can be hard to understand even face to face had no problem using the system.

I am a touch typist, like Liza Jodry, MD, who said in the article it was just as productive for her to type than to use speech recognition. I initially thought the same, but I found that a small upfront investment in templates and corrections (as Voicebrook’s CEO recommended in the article) made my work more efficient.

It was interesting to read about the pathologists who had data to show that biopsy turnaround times improved with speech recognition. One of our quality assurance monitors had been biopsy TATs, of course, and I suspect, but don’t recall, that we must have had to start recording TAT in hours rather than days after speech recognition.

All of this is good for patients. But I still feel sad when jobs disappear. Of course, it was hard to find qualified transcriptionists; the stress and required skill are high, salaries woeful. I am a real fan of efficiency and automation. But I prefer, for example, not to use the self-checkouts at the supermarket. People need jobs.

Irene J. Check, PhD, D(ABMLI)
Fort Lauderdale, Fla.

Your article on speech recognition correctly distinguishes this method of transcription from that of voice recognition, though the terms are used interchangeably. Speech recognition can have a future as a computer transformation of intended dictation despite accents, dialects, and violation of Standard English; cost issues, shown by Iowa Methodist Medical Center’s study; or the adjustment problems some groups or individuals experience. Progress cannot be rolled back.

However, amid this inevitable progress, medical transcription as a specialty or transcriptionists as a part of the professional report-generation process should not be eliminated, though such progress tends to move in that direction. This is not unlike other trends in surgical pathology, such as the elimination of “the middleman” by automatic embedding and the grossing person in biopsy processing. Experienced medical transcriptionists, in addition to having the immediate task of transcribing reports, are gatekeepers against errors, discrepancies, and inconsistencies. I have always felt in my work at the grossing table that the transcriptionist had my back. Errors can and do occur in the hectic, assembly-line work of high-volume surgical pathology laboratories. Having transcriptionists review reports helps to minimize the chance of error. Pathologists ought to be relieved of the job of catching and fixing technical errors as much as possible.

As speech recognition continues to be brought into greater numbers of surgical pathology practices, the medical transcriptionist’s role will change to one of facilitator of the transcription workflow, separating template dictations from other types and having final control before delivery to the pathologist. To keep the pool of experienced medical transcriptionists, cost-conscious managers can manage not only wages but also other incentives, such as working from home, which is now fully attainable with modern cloud-computing technology.

Izak B. Dimenstein, MD, PhD HT(ASCP)
Grand Rapids, Mich.