Software or Service?

November 15, 2001

New technology offers an alternative to old-fashioned transcription


In your practice's never-ending quest to reduce expenses while maintaining or improving quality, one area that's worth a look (or a listen) is the method of transcription you use. Like so many other physician support systems, the options run the gamut, from the old-fashioned, tried-and-true to the newest and most-improved. There is no single solution for everyone, but depending on your practice - size, specialty, required turnaround time, and computer capabilities - targeting transcription tools to address your needs can save time, improve accuracy, and trim thousands off your bottom line.

While many offices are still using in-house transcriptionists or sending out dictaphone tapes, there are two technology-based transcription options on the market vying for your business. One is a voice recognition software system and the other is a transcription service with a modern twist. Users of both claim a high accuracy rate and a smooth transition from their former method, and most vendors offer "try before you buy" demonstrations or sample software.

Voice e-volution

Voice recognition has, in fact, been around for a number of years, but vendors and users say it continues to evolve. Here's how it works: Clinicians speak into a microphone plugged into a computer, and their words appear instantly on their computer screen as digital documentation. Each user has a voice file, which is customized to the way the person speaks.

Greg Hackney, president and CEO of MD Productivity, a voice recognition system vendor based in Austin, Texas, says there have been two major improvements in voice recognition technology in the last five or six years: hardware processing speed has become fast enough to run the programs, and continuous speech feature has improved - users don't have to pause as they did before.

Most voice recognition systems use specialized macros - automatic insertions of blocks of text that appear on the screen when a certain patient condition or disease is dictated. In other words, large amounts of supporting documentation will be added to the patient's file when just a few common words are spoken. These systems also provide immediate editing features, such as commands to change typefaces or to add color to type for emphasis.

"Voice activation is here to stay," says Jim Zielinski, whose practice uses an MD Productivity product. "This is a way to cut costs without cutting the quality of care. The days of transcription as we know them are getting shorter."

Zielinski, who works in a small surgical practice in Mt. Pleasant, Iowa, as a certified registered nurse, spent about one hour to "train" the system to recognize his voice, and about a week to customize the software for his practice. The $18,000 fee the practice used to spend annually on traditional transcription has since been reduced significantly. Plus, Zielinski says, his office now sees more patients and completes their charts by the end of each day.

"If you can talk in a clear way, like a TV announcer, it works well," he says. Zielinski generally completes charts for existing patients in a minute, and new patients take about a minute and a half. He stresses that a "good computer" is a must - "the fastest one you can get and as much memory as you can buy. It's a very powerful program, and it requires a lot of computer to run it," he says.

Capturing details and saving time are the reasons John Levinson, a Boston cardiologist, uses voice recognition software. "I want to be able to convey the flavor of what the patient told me, and it's important for the consulting doctor to get notes done quickly. I want to be the cardiologist that the internist knew they would hear from that day. If you send to a transcriptionist there's a cost and a delay," says Levinson who uses a Lernout & Hauspie product.

The human touch

Some physicians, who may be reluctant to reduce the process to just themselves and their computer - or who prefer to know a real person is on the other end of the transcription process - may find a technology-based transcription service is a good way to go, and a few steps ahead of the dictaphones of the past.

Once they have contracted with a service, clinicians dial into the system from any phone (or a digital hand-held device), provide a physician and patient identification number, dictate their notes, then send the information to a centralized storage server. Next, a person on the other end transcribes the notes, and the document becomes available to the physician through e-mail, Internet, fax, or hard copy. (In some cases the voice is transcribed electronically and the document is sent to quality control experts to check accuracy.) Physicians have access to the documents within several hours or several days, depending on the contract.

"Our doctors are more comfortable knowing human contact is involved," says Sandy Kece, transcription manager of Delaware Cardiovascular Associates in Wilmington, which uses Scribes Online. She notes that it can be hard to convince some physicians to switch from methods they've used for years. "It was difficult enough for our doctors to know that the Internet would be involved."

Adds Hackney, whose company has developed several transcription products, "For doctors who don't want to sit in front of a computer [a technology-based transcription service] is the way to go. As far as [physicians] are concerned it's what they've been doing for 20 years."


The payoff has been significant for Kece, whose practice has 14 cardiologists in eight offices. Her practice used to employ five in-house transcriptionists. "The service costs about half the price, plus you don't have to pay vacation or medical benefits," she says of the method they've used for about a year now.

Users of these transcription services may find that they have to speak fairly slowly, and the editing process to correct inaccuracies can be time consuming. Still, says Kece, "The transcription service works well for this practice because many of the physicians have heavy accents," says Kece. "You have to train the voice recognition programs, and it would have been too much time up-front for these doctors to do that."

Something old, something new

Considering that both transcription services and voice recognition software vendors emphasize that different options will work best depending upon office variables, the answer may even be to combine methods, as some physicians do.

Allen Rothpearl, a Long Island radiologist, says the way to go is using the new and the old options. "I use both because my transcriptionist is saturated," he says. Rothpearl, who was a programmer and systems analyst prior to attending medical school, adds, "My number one choice for transcription is in-house. They pick up things a service wouldn't. If you're in-house you know how the practice operates - they pick up misspellings of a patient or doctor's name. And my transcriptionist can say to me, 'did you mean to say this or did you mean to say that?'"

It seems that for some practices, despite continuing advances in technology, the human touch remains as important as ever. "If I have Lorraine downstairs, who has worked for me for 15 years and does everything for me and will make my report exactly how I want it to be, why should I do anything different?" says Geeta Sankappanavar, CEO of Intellirecords.

Rothpearl adds: "As computers get more powerful, ultimately I think voice recognition will be the choice. But I don't think it will ever really replace a live transcriptionist. Machines are machines, but with humans you get that personal touch."

Karen Gatzke can be reached at editor@physicianspractice.com.

This article originally appeared in the November/December 2001 issue of Physicians Practice.