The Evolution of Speech Recognition - Cover

The Evolution of Speech Recognition

Remember when doctors recorded their notes on cassette tapes? They hired transcriptionists to listen to the tapes and turn them into documentation that went into patient records. Then came back-end speech recognition wherein physicians could dictate with software, creating a voice file to send to a transcriptionist. Of course, doctors loved the fact that they could share the burden of documentation with their transcriptionists, but many hospitals and clinics found that method to be costly in resource and time. It became popular to place the responsibility on the doctor to transcribe their own voice files—hence front-end speech recognition. 

These days, speech recognition vendors have found a way to assist the work of both doctors and administrations by automatically converting speech patterns into words on a screen. Though the accuracy of the technology is not perfect, it has vastly improved over the years. Now less effort (and money) is spent on transcriptionists or doctors checking the software’s work. 

The speech recognition market is also transitioning to the cloud. Thick-client users have to install an instance of the product on each device they use. That means that each instance of the software has to be trained to the provider’s accent and speech patterns. For doctors that work in multiple locations, that can be frustrating. The cloud allows users to access the same unique version of the product wherever they are.

On the list of healthcare providers’ concerns, physician burnout is at the top. Speech vendors and other new entrants are striving answer the call. We are seeing capabilities used that align with the consumer experience much like Amazon Alexa and Google Home do in our houses. Adding concierge voice capabilities to the exam room is a step forward. However, products such as Nuance’s DAX and 3M’s own ambient speech functionality are promising providers to capture conversations between patients and clinicians to create the needed documentation in close to real time. KLAS will soon be speaking with customers that have deployed this technology to capture their experience and the resulting outcomes and report to the market.

A Focus on Delivery

Our early KLAS reports about speech recognition software were all about functionality. Differentiation was largely in the accuracy of the different engines. Today, the focus is on delivery. Customers want to know who can be a real partner. They want to know who leads the most effective implementations and how many of their peers have adopted the product. Customers also want to know whether a speech recognition vendor will partner with their organization’s EMR vendor and whether the product can scale effectively.

In the Speech Recognition: Front-End EMR report, we answered the call of providers. Our focus is on delivery. We also focused on vendors that were known for integrating with major EMRs: Dolbey, 3M, and Nuance.

Market Metamorphosis

The last year has been exciting for speech recognition vendors. In February 2019, 3M completed their acquisition of MModal. We are closely watching how that acquisition has impacted customer satisfaction. Also, Dolbey recently released their new cloud product, Fusion Narrate. Nuance also launched their Dragon Ambient Experience solution (DAX). We were eager to know how the products were faring with early adopters.

We have discovered that all the products that we measured are strong contenders. Customers scored the products very high. Fusion Narrate is providing a great experience for their early adopters, which are mostly small hospitals and clinics. It has been more than a year since the acquisition of MModal, and customer satisfaction has held steady thanks to the strong, high-touch relationship that has been maintained through this period of change. Customers haven’t seen the expected technology benefits, but they don’t feel that their experience over the last year has gotten worse. Will the strong relationship continue? Will 3M make changes that affect customers? Will 3M technology begin to benefit the MModal customer base? The next six to nine months will be key to answering those questions.

Nuance is seen as an innovator with strong relationships with EMR vendors and is working on building implementation and adoption strategies to deliver their new technology more effectively.

Innovation and Integration

Two strong predictors of customer success and satisfaction are how vendors perform in supporting the integration and the delivery of new technology. Rarely do vendors succeed long term if these are not relative strengths. The market is no longer just about innovation; it’s about the delivery of that innovation, and the health system has made that focus of new technology vs integration


A good partnership with a speech recognition vendor is key in implementing effective integration and provider adoption. And while some vendors may be more innovative, the delivery of that new technology is what drives higher provider satisfaction.

Speech Recognition solutions are continuing to evolve. KLAS will continue to report on these solutions to address increasingly complex healthcare challenges.

     Photo Credit: Adobe Stock, WrightStudio