If you read this site, then you know that my favorite topic right now is probably ambient clinical voice. This shouldn’t come as a surprise since if ambient clinical voice is effective, then it could solve so many of the EHR documentation features that put such a burden on physicians in healthcare. The idea of an AI system listening to a clinical visit and automatically documenting the patient visit including meeting all reimbursement documentation requirements and other government requirements is just almost too tantalizing to believe. Although, it seems like telehealth is going to provide us the opportunity to see if the technology is finally there to make EHR auto documentation a reality and we’ll soon find out if the price is still too high for ambient clinical voice and we need a bit more time (or competition?) for that price to come down.
If you want to better understand how this technology works and where we’re at in the development of this technology, then you’ll love the interview below with Raghav Mani, Product Manager for Healthcare AI at NVIDIA. While I’m sure many people have seen the name NVIDIA when dealing with the video drivers on their computers (especially gaming computers), their realization that GPUs could be used for so much more than video has positioned them squarely in the healthcare AI space as well. So, Raghav Mani is well positioned to share where we’re at on the journey to EHR auto documentation. Plus, it doesn’t hurt that he previously worked at Epic.
Along with talking about automating EHR charting for doctors, we also talk about what this technology could mean for patients. We all know how little a patient actually retains during a relatively high stress appointment with a doctor. There’s something really beautiful about providing patients a recording of the visit that’s been processed through an engine that can take them directly to the point of the visit they’re searching for or even process the audio into text in a format that is useful to the patient. These types of features are going to be possible too.
Plus, I ask Mani to share the results of some recently published research that looked at Entity Recognition and Concept Mapping from conversations. That’s a lot to chew on I know, but it’s basically how to effectively train an NLP (Natural Language Processing) engine that will allow an organization to build out features like EHR documentation automation and converting a visit into a useful document for patients. If you’re not familiar with this topic, Mani does one of the best jobs explaining some of the history of AI and NLP and where we’re at today. Then, he shares how NVIDIA researchers are contributing to that progression.
Much like the video card in your computer, most people won’t know NVIDIA’s doing the hard work of NLP and AI in the background. However, their NVIDIA Jarvis platform is available to developers that want to roll out these types of features on top of their existing application. I love that their latest research is based on a large set of clinical terminology (PubMed) that can be leveraged by startups who don’t have a large set of data to train a medical NLP engine. Plus, if you’re a vendor with a lot of healthcare data, you can leverage that data to make the NVIDIA Jarvis platform even more effective for your users.
To learn a lot more, check out our interview with Raghav Mani from NVIDIA below:
Learn more about NVIDIA: https://www.nvidia.com/en-us/industries/healthcare-life-sciences/
Listen and subscribe to the Healthcare IT Today Interviews Podcast to hear all the latest insights from experts in healthcare IT.
And for an exclusive look at our top stories, subscribe to our newsletter.