Today we’re diving deeper into the HCA-Augmedix partnership announcement.

Like we covered last week, Ambient documentation has emerged as a real use-case for AI in healthcare. I wanted to learn more about what’s going on, so I reached out to Augmedix’s founder Ian Shakil.

I asked him some pointed questions about the partnership and space, and he was gracious enough to answer with a good amount of detail for Hospitalogy subscribers. I think you’ll enjoy his responses and I’ll mix in my thoughts as we move along.

• Join our Spaces today at 6pm CT to talk about everything AI in healthcare here: (Link to Spaces)

Key Takeaways on ambient documentation and the HCA-Augmedix partnership

• HCA and Augmedix are piloting ambient documentation in the acute care setting – the emergency room.
• Most use-cases for AI/documentation exist in the outpatient, ambulatory environment.
• HCA’s willingness to partner and invest in Augmedix signals that the space is credible.
• AI-enabled ambient documentation offers a triple whammy of increased efficiency, cost savings, and clinician burnout prevention.

Onto the Q&A!

What are the main value props you offer to a giant like HCA?

Ian: We have three main value propositions that are attractive to health systems. Those include:

• Higher quality documentation that is submitted faster.
• Clinician time savings, which can be repurposed for more patient throughput, clinician satisfaction, and/or other valuable tasks.
• Alleviated clinician burnout, leading to a better ability for health systems to attract and retain clinicians.

Blake’s Take: The value prop here lines up with what health systems are after when vetting a vendor. “Show me the money. Show me your ROI.”

Were there others involved in a bidding process? If not, how did the partnership manifest?

Ian: This was not a bidding process situation. We’ve been working with HCA for about a year. Before that, we were not materially active in the acute space. We were introduced to HCA by Google, through our Google partnership.

Blake’s Take: Ian’s comments here line up with my thoughts from last week’s piece on distribution. If you’re aligned with a major cloud player like Google, Microsoft (which likely has the biggest advantage here with Azure & Office), or AWS (but I’ve heard anecdotally 3M Modal isn’t as liked), this advantage probably gives you an edge to win partnerships as was the case here.

What gave you the edge over others in the space? How do you differentiate vs. a powerhouse like Nuance?

Ian: I suspect that HCA spoke with most of the ambient documentation players in the space, new and old, big and small. HCA chose to partner with us due to our willingness to pioneer in the acute care setting. Other large players in the space don’t materially operate in the acute care setting with ambient solutions; they just operate in the ambulatory space. We believe HCA appreciated our willingness to jump into more difficult and complex workflow areas with an innovation mindset.

Secondarily, HCA appreciated our transparency. Augmedix endeavors to avoid “black box” models and workflows. Our models also produce highly structured data feeds as output, above and beyond our core note output; this structured data approach adds a lot of value to HCA’s innovation and metric-keeping efforts.

HCA also invested $12M into Augmedix alongside Redmile, which indicates a long term, sustained partnership. Do you foresee HCA wanting to scale Augmedix across their entire portfolio? Will you eventually look to offer more than ambient documentation?

Ian: HCA indicates to me that they don’t engage in anything like this (and also take a financial stake) unless it has the potential to scale, enterprise wide. And as you know, scale for HCA is an entirely different beast when compared to scale at other health systems. That said, at this time we can only comment on the immediate plans to serve four HCA hospitals, two of which are live right now with two more coming at the end of the second quarter of this year.

Blake’s Take: the commentary here echoes what HCA CEO Sam Hazen told the audience at ViVE as well – when HCA considers vendors, its most pressing questions are…”can this scale across our hospital portfolio? Do they actually know what we need inside of our operations – are they immersed in the problems and pain points we face daily? Hazen also specifically called out Google & its cloud offering as a way to turn information in the cloud into actionable insights and align with partners. I guess that comment…makes a lot of sense now!

I saw that the pilot will roll out to two of HCA’s EDs in two hospitals, then expanding to two more. Since I imagine EDs in general are chaotic, what sorts of indicators will you and HCA be looking for in order to fine tune the product for the ED?

Ian: Augmedix was first to market; we invented ambient documentation (in the ambulatory setting). Now, as we deploy into the acute setting (the next frontier), it’s certainly more complex. We’re confident in our fine tuned technology and implementation techniques that will allow us to thrive in the acute space. We are looking at a number of metrics, including documentation quality, clinician adoption, clinician satisfaction.

Are you and HCA piloting a pure-AI, purely automated solution? If pure AI, are you worried about the cannibalization of your other products?

Ian: We are implementing Augmedix Live and simultaneously implementing and testing early versions of Augmedix Go in the HCA environment today.

Augmedix Live uses technology to produce medical documentation and point-of-care support in real time. A trained medical documentation specialist is in the loop with this product line. Augmedix Go features the full power of ambient AI to generate automated medical documentation that writes itself in real time, and made available in minutes for a clinician to review, edit and approve. Augmedix Go is our “Pure AI” solution.

We plan to GA (commercially launch) Augmedix Go for Acute by ~YE 2023. Thereafter, Augmedix Go is slated to be the preferred solution for future HCA adoption. The workflow learnings and machine learning data flywheel from Augmedix Live use in the emergency department (ED) are contributing to the optimization of the Augmedix Go solution for the ED. Our goal is to find the best Augmedix solution that fits the needs and goals for HCA’s clinicians.

Blake’s Take: Really insightful commentary here as well. I’m really interested to see how Augmedix Go and the other players (Nabla, Abridge, Ambience, etc.) continue to work on a fully automated solution. Since HCA seems to want to use Augmedix Go in the near future, I take that as a signal that the need for the human component of medical scribing will slowly fade. Automated solutions also offer higher gross margins (above its Notes margin mentioned below), and Augmedix will have its other products to help inform and train Go.

How will you and HCA think about physician adoption of ambient documentation? What are the major roadblocks there? Have any sort of incentives (financial or otherwise) worked to increase adoption in the past?

Ian: HCA wants to design Augmedix implementations that are highly standardized. With other health systems, we see a lot more note style and workflow variability from clinician to clinician. HCA wants implementations to be scale-ready and they want their enterprise-wide EHR data to be highly standardized. I think this is a very forward looking approach.

Which specialties, departments, and areas have you seen the greatest adoption of ambient documentation in general? Where does it work the best? Where are areas that still need improvement (if any)?

Ian: Prior to HCA, we’ve mostly operated in the ambulatory environment. Augmedix is flexible; we work with over 35 specialties across more than 30 EHR systems in a wide variety of care settings. Our 3 biggest specialties are primary care, oncology, and orthopedics.

One major success factor that we often consider is whether a health system has strong clinical leadership and is deeply involved in the implementation of the point of care technology projects. For example, leadership involvement for us is critical in the areas of physician selection, metrics, note standardization decisions, etc. We certainly have this at HCA.

As far as broader conversations around AI in healthcare, are you worried about Epic entering the generative AI space with Microsoft given that they have a close partnership and own Nuance?

Ian: I’m confident that, for the foreseeable future, Epic will be welcoming of multiple thriving ambient documentation solutions within their ecosystem. The same is true for traditional dictation transcription, where you see Nuance Dragon, 3M Fluency, and other solutions doing quite well on Epic. Indications are that Epic welcomes a similar dynamic in the newly emerging ambient documentation space.

Augmedix is quite successful in the Epic world today as we count numerous Epic clients as happy customers, such as Sutter, CHI, and TriHealth, to name a few.  Additionally, Augmedix is partnered with Google Cloud, and Epic and Google Cloud have a collaborative relationship; you probably recently saw the news regarding Hackensack.

What else should we know about AI in healthcare or Augmedix in the near future? What are you paying attention to, and what gets you excited?

Ian: I know there is a lot of attention on OpenAI GPT-4 right now, but I would keep your eyes on Google and their LLM roadmap. Frontier LLM technology is going to play a major role in accelerating our automation scores and increasing the number of settings amenable to Augmedix Go.

Augmedix is known for its ambient documentation solution – or our ability to do core notes. But if you think about it, we are building bidirectional pipes at scale into the offices of clinicians in most every specialty. That sets us up well to do so much more than just the medical note as we produce well-timed care gap nudges, real-time coding suggestions, and other forms of clinical support and the point of care. Stay tuned for more on this!

And lastly, the Augmedix team believes there is enormous value in the data running through our pipes. We see not only the super-structured EHR data (which we produce), but we also have granular access to the conversational graph, what’s said out loud, time-stamped to the EHR notes. There are many transformational clinical services that can be powered by this data set, and we plan to be on the vanguard of innovation here.

Thanks for reading and that’s all for this week! Subscribe to Hospitalogy here.

Next week I’ll be diving deep into the Kaiser Permanente and Geisinger deal announced yesterday. Make sure you’re subscribed to get it!

Here’s a sneak peek: