Speech Recognition Technology Does
Not Threaten Medical Transcription
Pat Forbis, CMT
Exhibit halls are interesting places, brimming with booths from which magazine advertisements and Web sites seem to come to life. Individuals step forward with bro-chures and pamphlets in hand, and with broad smiles designed to invite meeting participants to come a little closer. Always eager to describe their product or service in glowing terms, these carefully programmed company representatives seem to have the solution to just about any problem an industry can identify.
Certainly the exhibit hall at the recently held Medical Records Institute (MRI)-sponsored Toward an Electronic Patient Records (TEPR) meeting was no exception. Technology futurists were abundant, and of particular interest to medical transcriptionists (MTs) were those with speech recognition products and demonstrations.
Much to the amazement of the MTs, their inquiries were often met with resistance, hesitance or avoidance. “We’re surprised you’re here. You’ll be replaced soon, you know,” predicted one. And there was refusal from one company’s representatives to engage in conversation because “management instructed us not to talk with medical transcriptionists.” The message from some concluded firmly that, “medical transcriptionists will be replaced by computer-aided technology within five years, and we’re the ones who are going to do it.”
MTs spent time reassuring exhibitors that they were not only genuinely interested in their products, but that they were indeed purchasers of some of their products.
Are medical transcriptionists in agreement with these doom and gloom predictions? No.
The demand for medical transcriptionists has never been greater. It is impossible for the current transcription workforce to retire the ever-increasing amount of dictation being generated by health care providers. Dictation is no longer limited to physicians; nurses, physician assistants, therapists from a variety of disciplines, dieticians and social workers are among those now using dictation devices. And they rely on medical transcriptionists to complete the process. Transcriptionists would welcome technological assistance, speech recognition or otherwise.
The speech recognition technology that is available today does not consistently produce a document that is immediately useful, and there are those who believe that it never will. The use of speech recognition technology may be a satisfactory option for practitioners with limited vocabularies, but it is unlikely that it will ever replace the human ability to make sense from nonsense that so often occurs in the process of dictation, nor is it likely to be able to segregate hastily spoken messages from a din of noise emitting from a busy emergency department.
Companies that manufacture speech recognition equipment are proud of the quality of their products. They tell us that they produce documents of 80 percent to 95 percent quality. It appears that their definition of quality differs significantly from the quality expectation of a medical transcriptionist-produced document. While MTs are expected to differentiate sound-alike words, phrases, medications or lab values, errors made by speech recognition equipment are excused. The prediction is that the ability to reason will be left in the hands of the medical transcriptionists as they “edit” what the revolutionary technology cannot do…and that is to produce a consistently accurate patient care document.
These points and others were vigorously debated before a standing-room-only audience during the TEPR meeting by Brenda Hurley, CMT, of the American Association for Medical Transcription (AAMT) board of directors, and Joe Weber, chief executive officer (CEO) of Narratek. And it came as no surprise when Hurley was declared winner of the debate.
For those speech recognition researchers wanting to successfully create an electronic patient record that will meet the myriad needs of physicians, patients, payers, medical researchers and others who have the so-called “right to know,” here’s a word of advice: You will make great forward strides by involving those who are experts in speech recognition…medical transcriptionists. Instead of insulting or refusing to speak with MTs, draw them in, then listen carefully. There may be some valuable information to take home. And be assured that advances in speech recognition technology do not threaten medical transcriptionists. When the time comes, they will probably be its primary user. *
Pat Forbis is AAMT’s associate executive director for professional affairs.