Last month, the New York City-based Peterson Health Technology Institute (PHTI), which describes itself on its website as “a $50M initiative that provides independent evaluations of innovative digital health technologies to improve health and lower costs,” published a new report, entitled “Adoption of Artificial Intelligence in Healthcare Delivery Systems: Early Applications and Impacts.”
Among other findings, the report’s authors noted that, “In part, health systems are open to adopting AI-enabled administrative solutions because of the hype surrounding AI in general, but also because these solutions are being positioned to address major and seemingly intractable health system challenges, from financial pressures to workforce shortages. Representatives from health systems in the Taskforce” organized by the Institute cited as reasons investigating and investing in artificial intelligence, the need to create financial sustainability for their organizations; the desire to address ongoing staffing shortages; the need to address worker burnout, especially among frontline clinicians; and ongoing administrative burdens, especially around clinical documentation.
Per all that, the reported noted that “Healthcare costs are rising faster than inflation, wages, and the overall economy, and administrative complexity alone causes an estimated $250 billion in wasteful spending.1 Concurrently, the clinical workforce has seen limited productivity gains over the past few decades and is suffering from high levels of burnout following the COVID pandemic, increasing administrative burdens, and a lack of autonomy. This has motivated health systems to embrace new technologies — particularly those focused on administrative tasks — that promise to increase productivity and reduce the paperwork burden on clinicians, with low perceived risk to patient outcomes.”
Shortly after the publication of the report, Prabhjot Singh, M.D., senior adviser for strategic initiatives, at the Peterson Health Technology Institute, spoke with Healthcare Innovation Editor-in-Chief Mark Hagland regarding the AI-focused report and PHTI’s broader aims around technology diffusion. Below are excerpts from that interview.
I’d like to begin by asking about your background personally?
I trained as a physician scientist in New York City; I trained as a physician, and practiced in primary care, both in academics and entrepreneurial settings. My wife and I built a provider group during the COVID-19 pandemic. We ended up selling it to Oak Street Health, which became part of CVS; she went on to found a company called Diverge Health.
But for the past ten years, since the Center launched in 2015, I’ve been affiliated with Peterson, focusing on how economics meets healthcare delivery. I grew up in Kenya, and I’ve spent most of my career focused on low- and middle-income countries and rural areas. There are a lot of tradeoffs we shouldn’t have to make. What’s really working, and what’s worth it? The Peterson Center is really focused on big conditions and policy and incentives and financial infrastructure and markets, and how does all that make care more affordable and effective? And over the past three or four years, we focused on the Peterson Health Technology Institute. And Carolyn Pearson (?) joined us to be the leader of both.
When technology enters, you see affordability and access improve, but that’s not been the case in HC. It’s led to higher prices and hasn’t led to improved access as one might think. And we starting talking to the purchasers and payers of HC, and they said, there’s no one out there calling balls and strikes; there’s no independent entity saying, these things are worth it or not, and why, without a vested interest. So two years ago, we launched the Peterson Health Technology Institute, to really look at those issues; started with diabetes, musculoskeletal care, hypertension, and are doing work on anxiety and depression now, core areas that impact spending and care delivery.
But we’re also looking at AI-enabled solutions that are impacting how doctors and patients interact and are also impacting call centers, for example. So we do both formal assessments of mature technologies, and we also do early views: how are things really going, and what would it take for technology adoption to lead to a more affordable, higher-performing sector? I feel privileged as someone who’s had frontline primary care experience, to work alongside people like Caroline with her background in consulting and policy, and colleagues like Meg Barron, who really knows the sector, and David Sill. That team at PHTI, you get a really strong, multidisciplinary view of how technology is shaping healthcare, looking at the policy issues, and at the highest level, the Peterson Institute, which focuses on the economy overall.
How big is the Institute altogether?
It’s a self-funded, non-profit philanthropy. There are about a dozen-plus people at the Technology Institute and another dozen-plus at the Center. And we work with an extensive group of other organizations. We provide grants and also contract with them to develop work—like our health economics and research partners, that help us with these assessment.
How would you summarize the main findings in the AI report that you’ve just released?
The obvious point is simply that the speed of adoption of particularly ambient scribes, is among the fastest in the history of the sector. And that in itself is really notable: they speak to frontline needs for things that could help clinicians like myself, and they’re addressing a pain point. And they combine that with excellent business acumen. That might be stating the obvious, but it’s actually remarkable.
Point number two is that we’re incredibly early with these technologies: they’re in the pilot and testing phases, they’re still trying to figure out what they’ll be when they grow up. So the insights are very heterogeneous. There’s a huge range of experiences; it’s early days for these emerging technologies that want to become mature platforms, but there are huge questions around them. That’s why we were so pleased to work with CFOs and COOs at health systems, to engage in that dialogue at this early stage; our hope is that it will help move these things along into the future.
And for the first time, you’ve got executives and people thinking about investments, thinking very much about the day-to-day workflow of clinicians and patients. And they are keyed into these microsteps. You’ve heard the frontline saying, we’re experiencing burnout, patients are having bad experiences. So this work is shifting attention to the frontlines. And as a result, a lot of surprises are taking place: why don’t frontline clinicians talk to the revenue cycle people? That’s not how organizations are structured. So shifts are taking place in attention inside patient care organizations. But adoption gets complicated and slows and organizations need to digest information and get comfortable with it, and adapt.
And so this task force, we’re having a really interesting dialogue, where people want to move faster, but at the same time, are realizing that these are organizational issues as much as technological ones. So the report reflects the speed of development, the earliness of the development, and that we need to take all of this seriously, measure what works and what price it’s worth it at, and the sector will be stronger for that focus.
How should people think about this?
This is a really thoughtful discussion that’s taking place now in the industry. People are thinking in advance about risks and benefits, and about why adoption fails. Adoption has failed in the past because people don’t know their users or user needs. And you’re seeing people being much more savvy now. So it’s an incredibly important moment for healthcare; it’s been a twenty-year journey, but then again, everything in healthcare is a twenty-year journey. But it’s really important to understand in terms of day-to-day operations, what are you trying to improve? And how much resource are you putting into an effort? We so often hear from the innovators and the enthusiasts. But a lot of the folks who have to make the key decisions around whether to lean into AI versus basic patient care services, they need to understand that their investment in AI will be just part of moving the organization forward. It’s not a PR strategy; it’s how care should be delivered and how we should be driving results. And to do that, we need to make sure there’s dialogue between those making the decisions and those on the cutting edge. We see our role at Peterson in terms of helping to speed up that process. This technology is interesting and clearly transformative; its applications are very early in their development.
So we say, why don’t you have the early conversations with the executives and the leaders who will have to make the hard calls. And instead of waiting six or seven years in an innovation cycle to have those conversations, we hope that we can move those conversations up, by a matter of years, to have them early. We want to have these honest, thoughtful, brass-tacks conversations early. Those are the discussions I’m excited about. And even the leading health systems adopting this technology are well-resourced health systems.
There’s a lot of talk about responsible AI, but there also needs to be talk about responsible purchasing of AI solutions. For smaller, rural health systems, etc., it really matters how you spend those dollars, because you’re not one of the “big kids.” So I loved that those involved in this conversation saw this as a public service. And I really appreciated the service orientation that folks had, to get this right; and that’s the overarching spirit around AI, folks want to get this right.
