This report originally appeared in the February 2011 issue of DOTmed Business News

The difference between what’s said and what’s heard can be huge. Sometimes the speaker can be blamed for miscommunication, other times the listener is at fault. For speech communication software, it’s a different situation. The speaker is still available, but on the listener side there’s only the manufacturer to blame. Fortunately, the software has come a long way since its inception. It’s a good thing too, since some manufacturers believe speech recognition will provide big benefits to health care providers in the future.

“Connecting the voice to the data in an EHR is the key to achieving meaningful use,” says Christopher Spring, vice president of product management with Medquist. “Some hospital executives don’t even want to show new technology, which bridges the gap between their current dictation methods and the newer EHR templates to their providers because they are worried that it will discourage them from using the EHR.”

As people start to better understand the challenges and technological capabilities of clinical documentation related to EHRs, they will begin to understand that they don’t have to be mutually exclusive.

“New speech recognition platforms encourage physician use of EHRs by eliminating the frustrations inherent in the pull-down menus of the past,” Spring says.

Keith Belton, senior director of product marketing for Nuance, says that the biggest buzz in the industry is the application of speech recognition to address the federal requirements regarding electronic health records.

“The physicians didn’t go to medical school to type. In order to get the data into the EHR efficiently and completely, speech recognition is really the way to enable that process,” he says.

Pros are many
Advancements in speech recognition are coming fast. Speech recognition technology is well on its way to becoming one of the most widely adopted technologies in health care settings because it can significantly reduce documentation time and can boost both the availability and accuracy of patient records.

“We see this as a growing trend in the industry,” says Don Fallati, senior vice president, marketing for speech recognition vendor M*Modal. “It tends to be a richer documentation than traditional methods found in EHRs through pull-down menus. Physicians appear to favor speech as a [documentation method].”

Time is money
The oft-stated maxim of time equaling money holds true in the realm of physician documentation software, with speech recognition serving as a real time saver.

“Speech recognition gives time back to physicians that they would otherwise spend on more manual documentation,” Belton says. “Increased productivity leads to cost savings, with more time to focus on patient care.”

Once physicians learn how to easily dictate straight into the EHR and have essential information automatically extracted while at the same time meeting meaningful use criteria, adoption could be swift.

The biggest challenge, Spring says, is balancing the need to populate EHRs with complete and accurate information while giving providers documentation methods that do not limit their productivity and ability to focus on patient care.

Ease of use
Installing an EHR system in a hospital is the easy part. The performance of an EHR system—improving “ease of use” for providers and gathering meaningful data—is much more complex.

“Every day, doctors tell us that dictation is their preferred method of documentation because it adds color and meaning to the patient record,” says Spring. “When it comes to complete and accurate patient information, their voice is incredibly valuable. It’s important for us to develop products that not only convert their voice to useful information, but support a provider’s workflow so that it’s easy for them to create high-quality documentation, and also work seamlessly with the EHR.”

Technology is moving at lightning speed, especially in the mobile health area. With the ease of Web-based solutions that are affordable, user-friendly and improve productivity, physician receptivity is increasing. “Couple this with the 16,000 annual medical school graduates who are encouraged from day one to use technology in the classroom, as well as during their clinical training, and physician attitudes toward technology are no longer a barrier to progression, but rather the vehicle for it,” says Ruthann Russo, managing director and a documentation specialist with Navigant Consulting.

Early speech recognition software was known for problems centered around its inability to extract appropriate information from unstructured text. Through the advent of natural language processing, great strides have been made to automate the dictation process through speech recognition. This allows for needed patient information to be extracted from freely dictated text directly into the appropriate EHR data fields in an accurate and efficient manner.

“Creating software that’s more user-friendly is accomplished by understanding exactly how it’s going to be used in real life. You have to build tools that make the process simple and easy for the end users,” Spring says. “If an older physician is used to dictating, platforms that incorporate speech recognition allow them to continue to dictate, while the technology transforms their voice into useful information for the EHR.”

Front-end speech recognition employs static or handheld recording devices where physicians see their words on a screen as they speak. So instead of using transcriptionists to proof or edit the physician note, the physician can immediately see errors and correct mistakes using speech-activated commands.

Backend speech recognition software utilizes transcriptionists to oversee physicians’ notes. Using a backend speech system, a physician dictates notes into a static microphone or headset.

“The biggest impact speech recognition has made is improving the productivity of the clinician who is using it,” Spring says. “It has given the end user the ability to document faster and more accurately. Documents that are completed and signed more quickly can go to the billing department faster. This is a huge benefit for a hospital’s revenue cycle.”

You say tomato, I say tomahto
Doctors come from all over the world and the number of accents that can be heard in hospitals is extensive. Speech recognition software has to be up to the task.

“There are 46 sounds in the English language and what speech recognition does is understand and keep examples of all of the sounds that you make,” says Jason Kolinoski, senior vice president of Medquist. “Let’s say you say the word ‘cardiomegaly.’ This is a combination of sounds number 1, number 4, number 12, and so on. The computer keeps a scientific imprint on what they think that word is for you. It also keeps what we call a language model for every user because you might say it a different way and the combination of sounds may be different.”

So the trick isn’t to ditch the accent, it’s to say something the same way every time.

Nuance software comes with about 80 different medical sub-special categories and it doesn’t matter what sort of accent someone has.

“About 400,000 to 450,000 physicians today use some sort of speech recognition, with roughly half the physicians in our country having adopted this in the last four to five years,” Belton says. “If you are not using speech recognition, you’re sort of behind the curve in completion of records and time savings and reimbursement.”

Experts say that over time, speech recognition software can help create a dramatic return on investment if the systems are used correctly.

“There’s also a patient care aspect to it because from a production perspective, when you use a manual process, regardless of how good the process is, there is time involved,” Fallati says. “Speech can literally bring those reports to chart-ready in minutes.”

ChartWise Medical Systems offers software that helps doctors properly code and use phrases that will ensure optimized reimbursements.

“Doctors are being evaluated and graded on how well they participate in these programs and unfortunately, some hospitals are penalizing their full-time doctors for patient charts where they didn’t get as precise as they needed to be in their documentation,” says Jon Elion, a cardiologist and founder of ChartWise Medical Systems. “I hate the penalty idea and some are offering incentives for doing the most complete job and in that situation a [physician] would be advised to use our software.”

For instance, if a doctor used the phrase “unstable angina” it knows to translate that to “intermediate coronary syndrome” and the doctor and hospital will still be credited with the procedure they did, even though it may not have been initially said in the dictation.

Training time
Speech recognition software generally requires little training—it can take as little as 30 minutes to train the software and a couple of hours to train users. This increases its acceptance among clinicians.

“Speech recognition software does not require a major retraining of physicians since most are already using or have used some type of dictation,” Fallati says. “Speech recognition software will, in most cases, integrate into electronic medical record systems, should the practice decide to upgrade in the future. It’s also been shown that speech recognition makes a practice function in ‘real time,’ since speech recognition files can be dropped to paper if needed and faxed the same day as the patient is seen.”

Most companies recommend 2-3 hours per physician of training on the software and the key to success is follow-up.

On the go
The movement toward more mobile offerings designed for physician interface and documentation is the trend among speech recognition vendors.

Belton says Nuance has approximately 35,000 physicians in the United States who are using iPhones to look up something in their medical records.

“They simply speak the medication or the disease treatment plan into their phone and it instantly comes back with medical reference information,” Belton says. “The iPhone is clearly exploding in the health care space.”

M*Modal has also extended its speech recognition product to the iPhone and is seeing a rise in physicians using this.

The final word
Belton sees two scenarios unfolding for physicians and EHRs. Either physicians will continue to use traditional dictation methods, using the EHR for only certain functions, or they will adopt speech recognition technology, basically ending the costs associated with transcription. Those costs could run as much as $20,000 a year.

A continual challenge is whether dictation—the physician’s voice, which documents the patient’s story—is an essential component of meaningful use with EHRs.

“Because dictation remains the preferred method of documentation for physicians, there is an ongoing struggle between balancing the need for EHRs and the effect EHRs have on the physician’s time,” says Kolinoski. “Personal experience shows that provider frustration levels are rising over EHRs and the time it takes to use them — typically three to four times longer than dictating. Soon, the issue will come to a head. We must find a way that makes this technology work for providers and encourages them to continue to use the EHR.”

Speech recognition technology is a great option for medical practices for many different reasons. Whether it’s being used alone, as a stepping stone to full EMR, or in conjunction with an EMR system, practices will be pleased with the cost savings, ease of transition and workflow improvements.

“The future is very much focused on how to best support the entire clinical documentation lifecycle from voice capture to final document and well beyond,” Spring says. “Ensuring patient information is structured to support EHRs, coding, billing, revenue cycle integrity, analytics, regulatory reporting and all of the processes that follow the creation of clinical documentation is the key.”