Innovating in Community Health Systems with Steve West of Healthliant Ventures

This is part of our weekly executive insights series where Elion CEO Bobby Guelich speaks with healthcare leaders about their tech priorities and learnings. For more, become a member and sign up for our email here.

Name: Steve West

Role: Managing Director

Organization: Healthliant Ventures

What are your key tech-related priorities for the rest of this year and into 2025?

AI is a huge priority for us, not just in terms of productivity, but also physician, nurse, and patient satisfaction. Where we tend to look at AI a bit differently, though, is that we’re not just focused on the LLM-based use cases; we also look at it more holistically, particularly in terms of how it can drive efficiency and speed-to-value for the health system.

For example, in a hospital, physical space and clinician attention are some of the most valuable resources. Where AI shows the most promise for us is in reducing the complexities associated with managing space utilization and making it easier for clinicians to know where to direct their focus.

Right now we’re exploring some interesting partnerships in different service lines, like palliative care. The type of work that is done within palliative care is always additive on top of oncology or cardiology, but it can be hard to find the time or attention to be able to do it in a great way. Using AI-enabled services, you can provide this care in a much more comprehensive fashion. For us, that is not just a win from an efficiency standpoint, it’s a win from a patient care standpoint.

On that last point, what does it open up on the palliative care side specifically?

So instead of thinking about palliative care as an inpatient-only endeavor, AI might make it easier to coordinate logistics across home-based palliative care. For example, you can allow your nurses to operate in a hybrid model, where sometimes the patient is coming into the clinic and sometimes it’s a virtual visit. AI allows us to take advantage of the enormous quantities of data—such as that coming from remote patient monitoring devices—to better manage the logistics of palliative care with a better patient experience.

It sounds like a lot of this is care orchestration. How are you developing trust in these types of systems?

To this point, people have over-emphasized making AI the decision-maker. We believe that where AI adds the most value is by reducing the time it takes to provide a recommendation. Ultimately, it’s the clinician or coordinator who makes the call, though. You build up trust over time that the recommendations are good, but you always have a human in the loop.

Are there any specific AI-driven companies that you’re currently working with?

Some that we’ve announced partnerships with are:

Acuity Behavioral Health, which uses scoring from the nurses on the ground inside inpatient behavioral health settings to figure out how acute the ward is. It uses some combination of rules-based scoring and AI-based scoring to predict a need for more nursing staff. Then later if that ward lowers acuity, you can flex those nurses elsewhere.

Similarly, we’re working with an Australian company, Artrya, that is in the FDA approval process right now. They’re automatically identifying plaque within the heart and calcium scoring—just using standard CCTA images—and can get information back within 15 minutes. We’re not going to be solely trusting that AI score, but it’s a useful data point to be able to rule out an immediate cardiac event and we can send the patient home rather than keeping them a day or two waiting for results. That bed can go to another patient that might be having a more severe issue and lower our wait times.

Any controversial takes that you want to share?

I think most of the generative AI market is going to go to Epic and Cerner in the long run. I think that’s why I’m a lot more bullish on other sorts of AI, like deep learning types of AI, as opposed to the generative side of it. Epic is always going to have a cost advantage, because they have a strategic relationship with Microsoft, who has a strategic relationship with Open AI.

Another is that I think right now, there’s a ton of risk in providing hospital at home services. You’re starting to see potential data concerns with hospital at home information. We’ve had major hacking events at Change Healthcare and other health systems. What does your risk profile look like when you start having a ton of different devices at home with patients?

Given your role, I know you’re a big advocate for startups working with smaller regional health systems. Can you share your perspective on this?

I think that an overabundance of attention goes to the Mayo Clinics and the Cleveland Clinics of the world—and I’m not calling them out in particular, as those folks provide great care. But I think that as a startup, figuring out novel ways to work with smaller health systems becomes a really important thing.

Smaller health systems can go faster. If they’re in more rural parts of the country, as opposed to cities, they actually have access to better longitudinal data. You can see a patient over the course of 10, 20, or 30 years. At a Mayo or a Cleveland, everybody goes to them for a second opinion; they fly in, they get the work up, but it’s a moment in time. They don’t see all of the data that led up to that. Those are the types of things that a community health system can provide.

When you’re thinking about developing for health systems, don’t think about developing for the fewer massive systems or most sophisticated academic medical centers. Think about developing for the hundreds of smaller systems that are ultimately going to be your biggest champions. Those are going to be the folks that ultimately generate the ARR necessary to achieve liftoff.

I can imagine being a startup founder and thinking, “Hey, that sounds great. I want to help raise the level of care across the country, not just make the best better.” But are there any best practices for getting started and getting your foot in the door with these regional systems?

I think a lot of startups really try to go after the C-suite. They inundate CMIOs and CIOs with messages, and they think, “Oh, well, that’s the natural buyer of my product. It’s a technology product.”

If it’s a clinical product, you want to find a clinical champion. Then you also have to engage your administration team—the head of the service line—to be able to help build the business case for it, as opposed to just “this provides better patient care.” Recognize that a lot of CMIOs and CIOs are veto votes in most of these types of interactions, but they are willing to take a look at things if there is a) a strong champion and b) a broader consensus (that your champion has helped cultivate) on how the product would interact within the health system. If you’ve been able to engage a lot of different folks across a particular service line, you’ve got three or four folks that all understand what the product is, you’ve spent time in their institution, and you understand their particular challenges, you’ll have more success.

Tech folks want to think about it as a technology problem to solve. But it’s a human problem. And so if it’s a human problem, you go after the people that are dealing with the human side of it each and every day; build the consensus there, and then say “Well, maybe we look at a pilot.” But you have to have people willing to go to war for your product to be able to do that. You can’t just do that from the quality of your code or expect the CIO to naturally advocate for your solution amongst the dozens of priorities they have.

You’ve had some great advice for the startup side of the equation, any advice or feedback for the folks who sit on your side of the equation?

I think one major thing I would say is start being expansive in terms of the idea of how to solve a problem. People are very concerned about adopting too many point solutions, and they want to work within their existing platforms. But when you get deep down into it, a lot of current platform solutions (with some exceptions) are just aggregations of point solutions that have been acquired. You potentially have just as much risk if you expand your platform contract as you do if you go with a newer entrant that may solve the problem in a better way, give you better customer service, and give you better attention from the founder. So in that way, I say point solutions are not nearly as scary as people think. It may seem like there’s more work on the IT side, but you’re getting that same amount of work on the platform too.