Nachos Neurosurgery

Jon Ford Healthcare Leave a Comment

About the Author

Jon Ford

VP Engineering

When we deployed the first version of SayIt almost ten years ago, we delivered a solution that was limited when compared to the SayIt of 2018. We were first to the web and the product was good. But, looking back, the tools at our disposal seem primitive when compared to those available today. The state of the art dictation was speaker dependent. New users invested 10-20 minutes reading to the system, teaching it how they spoke. The idea of speaker independent dictation was still just that - an idea. Specialty topics were limited and very narrow, requiring users to speak within a what would now be considered a very narrow band of acceptable phraseology. Users received good results if they followed these rules and spoke in an “average” way. But, who wants to be average?

If you follow our blog here at nVoq, or read other tech news, you are aware of how advances in machine learning have transformed speech recognition. Speaker independence and increasing topic breadth help us immediately deliver accurate transcriptions to almost everyone speaking about almost anything in their medical specialty. But, “almost everyone” and “almost anything” are not our ultimate goal. We want SayIt to work for everyone, regardless of what they are saying, or even where they are saying it.

The engineers at nVoq have been working to steadily increase coverage and create a system that is more forgiving in how it listens to speech. Our last release greatly increased accuracy and reduces errors for accented speakers and common disfluencies such as truncated words, hurried speech, etc. And, if a speaker drifts off topic, we are much more likely to transcribe that correctly as well. We will continue to push out from the mean, reaching further afield to create a solution that works for everyone. But, what about working everywhere?

Medical professionals are constantly on the go. nVoq provides iOS and Android wireless microphone applications that enable this mobility, creating a consistent dictation experience as clinicians move about doing their work. But, going from room to room, building to building, and clinic to clinic typically still involves an IT managed PC or Virtual Desktop. Even at home on a Mac or PC, the computing environment is stable and SayIt provides a great dictation experience. But, the fastest growing segment within the healthcare industry is home healthcare. Still required to provide accurate and timely documentation, home healthcare professionals are sent to the field with a tablet running mobile EHR software. The iOS or Android built-in dictation solutions are not HIPAA compliant and the home healthcare provider is forced to “thumb out” a report on a less-than-ergonomic virtual keyboard, unless that tablet is equipped with nVoq’s mobile dictation for Android. With wireless data access, a simple installation, and no programming, home healthcare providers can be dictating patient encounter notes on-site with the same fast and accurate dictation Mac and PC SayIt users enjoy. Now, everyone can dictate notes about all things medical from any location.

Harried schedules and shrinking staff sizes have led to not only working in different locations, but working on the move. We make phone calls from our cars and over meals – sometimes simultaneously. Rescheduling that appointment while taking a bite of your hamburger at a red light doesn’t seem so strange any more. In fact, some medical dictations are submitted this way. Dictating from a cell phone in a car, while walking down a hospital hallway, or even while eating a sandwich is common for medical professionals who work with transcriptionists. The Onion, a satirical news website, recently published an article titled “New Speech Recognition Software Factors In User’s Mouth Always Being Full”. This is funny because there is some truth to it. Perhaps in addition to our standard offering, we will need new specialty topic such as “Pizza Pathology” for those late-night work sessions dictating notes while eating delivery or “Nachos Neurosurgery” for those grueling surgery sessions that demand some sort of refueling.

While we may not be able to extract text from speech mixed with crunching chips, we are making great progress on adjusting to changing backgrounds, speech patterns, and all kinds of other inconsistencies. Even in the case of traditional transcription, an accurate first draft of this type of dictation makes the human transcriptionists more productive by transforming their work from simply converting speech into text to where they truly add value – creating great medical documentation from the speech provided by the physician.

The last ten years at nVoq have been devoted to helping our customers and our partners’ customers get their jobs done. Through speech and automation we have continued to broaden the applicability of our desktop productivity tools. Whether real-time, backend, stationary, or mobile, nVoq can help you produce accurate medical documentation fast. Just remember what your mom taught you years ago – don’t talk with your mouth full.

Leave a Reply