NEW ORLEANS—In the face of economic uncertainty, health system executives and venture capitalists are setting a high bar for digital health and artificial intelligence investments.
That isn't to say that major healthcare organizations aren't placing bets. Digital health technologies for behavioral health, maternal health and patient engagement as well as AI and automation are all key priorities for many major health systems.
"Our organization continues to actively partner and to actively want to contract with innovative companies because we need innovation to pull us into the future and allow us to deliver a contemporary experience to our patients that they want," Christine Brocato, system vice president for strategic innovation at CommonSpirit Health, said during a session* at the American Telemedicine Association Nexus conference this week.
However, executives are being discerning about the technology companies they partner with and how their venture arms are making investments, health system executives said during ATA Nexus.
Rock Health researchers note that digital health startups and investors are navigating a market that is quickly shifting as the Trump administration reshapes budgets, leadership and oversight, which sets the stage for changes to reimbursement and innovation. Startups are facing a more selective investment climate, while challenging markets make it especially difficult for publicly traded companies to succeed, Rock Health noted in a recent report.
CommonSpirit Health is one of the largest not-for-profit health systems with 143 health systems and 2,000 sites of care. Brocato hears pitches every week, and her advice to startups and digital health innovators: Have a compelling data and return-on-investment story from an operational and financial perspective.
"I think a good data story is less about having big brand names of who you're working with or having a ton of pilots, but I think in this environment, it's more about finding three or four partners who are willing to work really deeply with you and develop high-quality data that proves that your product or service can really move operational and financial outcomes, improve metrics that matter. In this environment, sometimes it helps to go deep versus wide," Brocato said during a session* at ATA Nexus.
A flashy pitch with little substance won't get very far in this current environment, executives say.
"What is your evidence? Making sure you generate data that shows your product works and generates value to health systems. It is really important, and I think it gets often overlooked, especially in early days, but data generation and evidence generation is absolutely key to standing out," said Julia Monfrini Peev, partner at PACE Healthcare Capital, a venture capital firm that invests in digital health and health tech companies.
Digital health startups also should know their audience, she noted. "Understand the difference between a community hospital versus an academic versus a payvider and what are those different problems that we're experiencing, and how might you customize your pitch so that you are really tying your product to the key priorities that we are trying to solve," she said.
"As economics get tighter, people start to really start to be diligent with finances. You start to fund your key priorities. So make sure that you are connecting your product and service to those key priorities," she added.
Those key priorities vary across the provider market. Bo Wilkes, chief growth officer at Ballad Health, which operates hospitals that serve rural areas in Tennessee and Virginia, said he's interested in digital health technologies that support maternal health and behavioral health as well as oncology and pharmacy.
AI also is on the health system's radar, he noted.
"AI continues to be a buzzword—I think it's the flavor of the day right now—but it has significant opportunity," said Wilkes, who also serves as president of the health system's venture capital arm Ballad Ventures. "I think what really becomes interesting is in that startup space, not so much the AI, but what service are you providing? It's understanding the data elements there, that's what's really exciting."
MedStar Health, a not-for-profit health system that operates hospitals in Maryland and the Washington, D.C., area, has been focused on ambulatory care model redesign, inserting digital health technologies in its design to support remote workforces and other initiatives, said William Sheahan, senior vice president and chief innovation officer at MedStar Health.
"We're now focused on the inpatient space, acute care model redesign. We need to think about new technologies in patient rooms as well as similar things to ambulatory with remote workforce and how we add automation technology on top of it," Sheahan, who also serves as executive director of the MedStar Institute for Innovation.
The health system also is focused on deploying automation and AI for back-office processes and as it looks to "redesign the future of work," he added.
But technology is only one part of the strategy as health systems like MedStar Health look build the healthcare workforce of the future and evolve their operations.
"We often we focus on the technology as the transformation and the technology is enabling of the transformation, but the harder work is looking at yourself as an organization and doing the people and process change to deliver the value of the technology and implementation," Sheahan said.
Digital health startups and innovators need to have a "thoughtful playbook" on how their solution fits within that broader transformation, he noted.
"Often, what I look for in companies is companies that ask me questions instead of just tell me things, because I want them to situate their product within where we are and where our health system is and the individual hospitals or service lines within it. I think thoughtful dialog in the discovery phase, and often, what I'm actually listening for in that is how are they actually talking about how it's hard, not how it's easy," he said.
As a VC, Peev said she is looking for solutions to support women's health, including maternal health and fertility.
"I have an eye out for AI tools. But, for me, AI is really just a tool. What's most interesting is, what's the data set underneath? And how can we leverage the data set and the tool to then improve outcomes?" she said.
Fertility is one area of healthcare that is "ripe" for AI to harmonize patient data and improve outcomes, she noted.
"If we think about fertility, for instance, 75% of the IVF cycles turn into failure, and it's an emotional drain, it's a financial drain, and frankly, it's also taboo, because if we realize we put so many families through a process that results into so few positive outcomes, I think we would take a different approach," Peev said.
Ballad Health is looking to not only adopt new digital health solutions but also to find technology partners to co-develop with, Wilkes noted.
"We're trying to find those partners that we can pull into our environment to make sure that we're focusing on the right things, the quality, the experience and not only the experience for patients, but the experience for our caregivers. We've got to make sure that that our solutions and services are built within our community and not adding to our workflow," he said. "As a healthcare corporate venture investor, we have the unique opportunity to not only invest, but then also kind of partner as an initial use case."
There is a place for point solutions, or technologies that address a narrow clinical or operational need, but there needs to be consideration for how the technology "plugs into the larger ecosystem of technology and could potentially solve adjacent problems," Sheahan noted.
"From a health system perspective, I think we do have to take a rational approach to thinking about platform, and are there solutions that cross multiple parts of our business that can deliver 80% or 90% of what individual point solutions could do, and have not as much technical debt?" he said.
Health systems and investors are also developing their strategies around AI technologies as the sector evolves at lightning speed.
"AI is a tool. It's a calculator. It's a powerful calculator and it can do amazing things, but it's still just a calculator. We try not to invest in the tools. We try to invest in the better business models that leverage those tools," Peev noted.
There are big opportunities to use computational AI for predictive algorithms to improve workflow and aid clinical decision-making, Sheahan noted.
The emergence of large language models also unlocks opportunities to improve healthcare operations. "AI has the opportunity to really add far more automation into all parts of our business and democratize data and make it accessible in a way for every associate in a healthcare organization and every patient who is served by healthcare in the U.S. That's a challenging problem, though, and I think the maturity of our interoperability data structures in healthcare that's going to get in the way of deploying that largely. But, I think advancements happening now in regulation and between health systems and payers should unlock some of that value."
AI startups also need to understand how different health systems and healthcare organizations are deploying AI and the resources they have to support the technology. Not every hospital has the infrastructure and expertise to be on the leading edge of AI.
Ballad Health is going to be a "fast follower" when it comes to clinical AI for diagnostics, Wilkes noted.
"We don't have this large research infrastructure. We're a $2.5 billion healthcare system in rural East Tennessee and southwest Virginia. What we're really focused on is leveraging AI from a workforce perspective. The clinical aspects will occur at other institutions. They'll validate those models, and as opportunities present themselves for us then to pull it into our system, we'll then take that advantage," he said.
"I think a lot of community health systems will take that approach where we should be walking into that space slower, we should be crawling," Brocato noted. "We should be understanding the ethics of it all before we start to apply it live. And I think that's the right thing to do. Certainly, on the back end, all the administrative aspects of reaping efficiencies in AI is kind of a no-brainer."
Patient engagement is an area that's ripe for AI tools to modernize manual processes and revamp patients' experience with healthcare, she noted. AI technology that can save patients time by helping them bypass the clinic waiting room or get them seen by a clinician faster in the emergency room can actually build an individual's trust in the health system.
"Patients fall through the cracks all the time because, despite best intentions, someone forgot to call. There's a piece of paper that didn't get sent or no one followed up. AI can be the glue that really holds that together and ensures a consistent experience throughout the entire journey for a patient," she said.
The health system applied an AI tool in its primary care physician offices to streamline patient check-in and prep them for appointments. "We found that patients really love that. I think the true north for this AI journey is that it brought trust in us because we are respecting their time," Brocato said.
Another AI tool used in the emergency department notifies patients and their loved ones about the wait time and keeps them informed about next steps. "Changing the communication through this AI tool dramatically increased patient satisfaction," she noted.
With all the excitement around AI, though, health systems are trying to find the best approach to build guardrails and effective AI governance strategies.
"As we get more solutions into market and into our individual systems, I think that governance becomes really challenging. We need to focus on it so that we can understand how AI is both creating outcomes, but also where it is creating risks," Sheahan said.
The MedStar Health National Center for Human Factors in Healthcare recently published a study that found human-in-the-loop models, in which clinicians or physicians maintain oversight of the AI outputs, weren't effective at catching AI-generated errors. The stimulation study assessed whether primary care physicians could identify and correct errors in AI-generated draft responses to patient portal messages. All but one physician "sent" at least one fictitious response to a patient that contained an error, the study, published in NPJ Digital Medicine, found.
"That's really problematic," Sheahan said during the session, referring to the study. "If we're saying that AI is safe because there are clinicians that are going to be watching the output of AI, we have to validate that that's true, and where we're inserting algorithms into complex clinical care, we really need to make sure that we understand that breadcrumb of safety. We don't want to harm more people, because we're inserting non-transparent AI into the system."
Looking at the overall digital health investment outlook, there was optimism going into 2025 that this would be a strong year for mergers and acquisitions and exits. Digital physical therapy player Hinge Health's IPO plans earlier this year seemed like a good sign.
That enthusiasm has been tempered amid fears of a potential recession, Peev noted. In early April, Business Insider reported that Hinge Health was considering delaying its IPO given the shaky stock market.
"Unfortunately, we're all in waiting mode in terms of what the exits will look like," she said. "It's a choppy environment to land a plane."
"In an environment where the recession is looming, though, not everything is negative for healthcare. Healthcare tends to be recession-proof. It's not absolutely perfect. I know we all work in healthcare, and we do fear the tariffs and we do fear changes in insurance coverage for our members. But in general, it's a good sector to be in."
*The session on digital health investment at ATA Nexus was moderated by the author.