Speak To Me: Speech recognition with Nuance Healthcare’s Nick van Terheyden, MDby Stesha Doku on Apr 27, 2012 • 7:47 am
With the introduction of Apple’s Siri in the last year, free-form speech recognition has exploded in the mainstream. The technology around Siri, however, has been used in the medical field for some time for documentation with programs like well-known Dragon Medical™ by Nuance Healthcare. Nuance’s CMIO, Nick van Terheyden, MD was willing to speak with me about the advances Nuance has made in the domain of speech recognition, data mining, and documentation innovation in the healthcare space.
Dr. van Terheyden has been in the healthcare industry for > 25 years, working in imaging and other internet startups before coming to Nuance and finding his niche as a clinician advocate to adopting technology in standard practice. Dragon, now on its 10th medical version has been used commonly within healthcare enterprises, including in radiology reading rooms across the country for dictating reads of XRay/CT/MRI films and more. Over its time, it has been adapted to take into account accents across the world in addition to learning and increasing its vocabulary each time someone dictates.
The next innovations in this space are what one would expect: with the data Dragon is collecting, clinical language understanding is growing. With the use of such ontologies as SNOMED, ICD-10 and drug lists the language we use in medicine is better being understood by our machines. Van Terheyden recounts seeing IBM’s Watson take on two Jeopardy champions: “I thought, gosh, imagine all that unstructured data we have in healthcare — imagine pulling that into a Watson type toolset and then offering that as a resource to clinicians that sits in the background and applies that knowledge to your current patient.” The implications and capabilities of such a sophisticated machine to analyze and answer questions would be enormous.
There are many challenges in this domain, however. The amount of data that is being generated and mined through natural language processing for use in clinical decision making continues to grow. Not to forget, there is large variability in the willingness of physicians to adopt a new technology into their routine practice of medicine.
“The key to success with clinician engagement”, Van Terheyden comments, “is offering choice. If you force clinicians into a corner and require them to use something, it is a recipe for challenges in the short term and probably in the long term.” In his opinion, offering an option with a caveat to change one’s mind makes clinicians more willing to try out a new technology. When they realize it works well for them they are much faster to adopt. Sometimes being innovative in the specific domains in which physicians practice is also important. For example, Nuance was presented with the challenge of introducing speech recognition into pathology where the use makes sense but the challenge in working in a wet lab where the physician needs both hands is a barrier. To address this, foot petals were adapted and verbal phrases implemented to specifically control when the system would be “listening.”
In the next year, we can expect a lot from Nuance as their innovation in this space continues to add value to how physicians do their work efficiently intelligently. As a natural progression to speech, they are focusing on applying language understanding and analytics to healthcare to enable patient care and business improvements. It’ll be exciting to see what they come up with next.
Dragon Medical Practice Edition: http://www.nuance.com/for-
Consumer Version of Dragon Dictation for iPhone: http://itunes.apple.