Speech Recognition Makes a Statement in Health Care

Is data entry the hurdle blocking electronic medical records from wide adoption?

Speech recognition could be the weight needed to push broad adoption of electronic medical records in health care.

Nuance, an ISV that makes the popular Dragon speech recognition software, is betting data entry is the primary stumbling block to EMR adoption. The company is building tools and ties to EMRs, hoping doctors and technicians will see it as a time saver and more efficient way to practice medicine.

"EMRs can slow physicians down," said Keith Belton, senior product marketing manager for Nuance. While health care industry analysts estimate that about 30 percent of physicians are currently using EMRs, Belton said using speech recognition technology in conjunction could speed adoption, since it could eliminate the need for transcription or for manual data entry into EMRs.

Dr. Eric Fishman, a former orthopedic surgeon who is now a distributor of speech recognition technology, has used speech recognition software in his practice since 1994. Dr. Fishman said he initially used the technology to produce medical records in WordPerfect and then Microsoft Word, and until he stopped seeing patients in June 2007, was using Nuance's Dragon NaturallySpeaking Medical speech recognition software to dictate information into EMR software.

Dr. Fishman said that while the EMR software he used included a lot of "templated" data such as age, date, gender and other information that didn't change much from patient to patient and could be selected using a mouse click or a drop down menu, inputting specific patient complaints unique to them as well as his treatment plan required inputting multiple pages of text and was time consuming.

Using the EMR software in conjunction with NaturallySpeaking Medical allowed Fishman to generate records much more quickly and increase the time he could spend talking to and treating patients. Using integrated "voice macros" sped up the process, he said. The software can be programmed to insert multiple paragraphs of text into a record when he spoke a specific combination of words, he said.

Dr. Fishman said that, for example, he could pre-program the speech recognition software to insert two or three paragraphs of text to describe a specific treatment he had prescribed to a patient. Without the speech recognition software, he'd have to input the text manually, often repeating the same paragraphs for multiple patients who he had prescribed the same treatment.

Currently, Belton said, Dragon NaturallySpeaking Medical is being used by 50,000 to 60,000 physicians, including 7,000 nurses and doctors within the U.S. Veterans Administration hospital system, which integrates the technology with its EMR software.

The vendor said its speech recognition software is compatible with EMR software from major EMR providers Eclipsys, Misys, GE Healthcare, McKesson, NextGen and Allscripts. While Nuance offers speech recognition for a variety of markets including government and telecommunications, Belton said that the health care market accounts for about 25 percent to 30 percent of the company's business.