Twenty years ago, technologists predicted that medical transcription would be replaced by speech recognition in 3 to 5 years. Five years later, they uttered the same prediction. And that's the way it's been ever since-a rolling 3- to 5-year window.

MedSpeak, IBM ViaVoice, and Dragon Systems Naturally Speaking are the major packages. They are pretty good now, reaching up to 95-98% accuracy at normal speaking speeds for a good dictator, but that 2% error rate kills it! In practice, physicians are slamming out dictations way too fast for the system to work. Internist Jeffrey Clode, MD, of Spokane, Wash., remembers a colleague at his 20-physician practice who told his computer that a female patient had come in after choking on peanuts. Except on screen, the software didn't get the dictation quite right. Instead of "peanuts," the software entered...well, something more anatomical. "It was funnier than hell because our nurse practitioner found the error and was just in hysterics," Dr. Clode said. "You have to watch out [and proof carefully] when you use it."

Doctors not only have to spend considerable time learning the software but also must "adapt to the technology," says Joe Marion,  an executive director with Superior Consultant Co., Southfield, Mich. "If they are not willing to do that and [think] that the computer is going to figure them out, then they will struggle." Speech recognition is "not for everybody," he says. "One of my partners who tried it would yell into the microphone, 'Damn it! I said ...' Well, if you're kind of rigid like that, don't want do to proofing and are the type of individual who can't stand any errors, it's not going to work."

Also, physicians use "meta-language" like "Start with a normal standard CT Chest and add that there is a 2cm nodule in the lateral segment of the middle lobe." and a medical transcriptionist will produce the whole report properly. Or, "go back to where I said left subclavian and change it to right" and the medical transcriptionist figures it out.

The transcriptionist does a lot of "back-office" work in addition to just turning speech into text. Sometimes the charge codes are wrong. They fix all that, and format and deliver the work. They call the physician or flag the report if they don't understand a word, or if a transcription gets cut off.

Speech recognition technology is increasingly being used to compensate for the shortage of MTs. However, it is impossible for this technology, with all of its limitations, to completely eliminate the need for medical transcriptionists. Even at its best, machine-translated text contains errors that need to be corrected by professionals with language skills and an understanding of the health record. MTs continue to be the best qualified to discern the nuances of human speech. Technology, such as background speech recognition, will cause MTs to be transformed - they will become the editors and quality assurance people. Transcriptionists have a huge amount of knowledge, such as medical terminology and an understanding of the institution and physicians they're transcribing for. They will always be in demand. Medical transcriptionists will remain the best qualified to discern the nuances of human speech - indeed, the ultimate speech processors.