One of the biggest complaints from doctors trading in their paper files for electronic health records is that typing in patient info takes too much time.

For some doctors, an e-health record might mean 3 to 5 minutes of extra work per patient. In a busy practice or hospital, that time can add up quickly.

But some doctors are avoiding that hurdle by speaking commands as well as patient notes into their e-health record systems. Voice recognition functionality is not only helping to drive utilization of e-health record systems among doctors, it's also helping to dramatically reduce the cost of medical transcription services.

Dr. Jose Gude is the physician champion for an e-health project being rolled out at Swedish Medical Center, a three-hospital health care system in Seattle. Swedish's rollout includes e-health software from Epic Systems and Dragon NaturallySpeaking Medical Version 9.1 voice-recognition software from Nuance Communications. The systems are in a pilot program, but together the software already is saving time.

The e-medical record software has drop-down templates from which physicians can enter information about patients, such as during exams and in providing data about medical history. However, the voice-recognition software provides the e-records an added boost by allowing doctors to add details that would otherwise be excluded from the data, or have to be typed in.

"Doctors aren't good typists," said Gude. "I'm a hospitalist who works at night and does a heavy load of admissions," he said. After about five months of using the voice-recognition software with the e-medical records software, "I couldn't live without it," he said.

Patient medical histories "can be hard to template because they can include complex problems" that can't be detailed through drop-down lists featured in the e-health record system, he said. Speaking those details via a USB microphone into the system at a workstation at nursing stations saves time while ensuring "a clear clinical picture of a patient," he said.

At the University of Rochester Medical Center, DragonNaturally Speaking 9.1 is part of a rollout of new TouchWorks e-health record software from Allscripts. So far, about 100 doctors are using "this hybrid of templates and voice recognition" software, said Dr. David Krusch, chief medical information officer at University of Rochester Medical Center, which includes several primary care and specialty physician practices. Doctors are speaking into the system details about patients that add to the info entered by doctors via templates. "It's taking off like wildfire," he said.

Some doctor practices already are reaping tremendous savings in traditional transcription service fees. Rather than dictate notes that need to be transcribed by third parties, the voice-recognition software can add the information directly to patient's e-health records, where doctors can read through them immediately for editing or checking for errors.

Among the University of Rochester Medical Center doctors saving money and time is a six-physican primary care practice that had been spending $20,000 to $30,000 monthly in transcription fees, said Krusch. Since using the Allscripts software with Dragon voice recognition over the last six months, "that's down to zero now," he said. Eventually, the software will be rolled out so that more than 700 doctors, faculty, researchers, and others at the medical center will have access to it. The goal is to have at least half of them using the software, an achievable objective, Krusch said, considering the appeal of voice recognition, he said.

As a standalone product, Dragon NaturallySpeaking Medical is priced at about $1,200 per user and can be used with EMR packages purchased separately, said Peter Durlach, Nuance VP of product strategy and marketing for Dictaphone Healthcare Solutions division.

Also, some e-health records vendors, including Allscripts, offer versions of their software that have Dragon NaturallySpeaking Medical voice recognition functionality already embedded. That's the case for Allscripts' TouchWorks EMR, which has a voice recognition version that "costs a few hundred dollars" more per license, said Laurie McGraw, Allscripts president of clinical solutions group.

"It's no question that [voice recognition] is helping to drive physician adoption" of EMRs, said McGraw. "It's not the reason why EMR is getting adopted, but it's one more feature that adds ease of use," she said. Plus, "the savings are enormous" when voice recognition functionality reduces the need for transcription services, she said.

The University of Rochester Medical Center's Krusch said that, long term, voice recognition in EMRs will likely become a functionality most doctors expect in their software, rather than an extra perk.