A Conversation With William Shrank, M.D., on Artificial Intelligence, Value-Based Care and Wasteful Healthcare Spending

This transcript has been edited for clarity and length.

We’ve been talking to several editorial advisory board members for this series. We begin with a warmup question: When you started to think about healthcare today, what was on your mind? Which problems or puzzles were in the forefront of your thinking?

I think, for me, it’s the sort of orientation that I always circle back to … just how personal everyone’s path to their own best health is. And we spend a lot of time thinking, at a high level, about how to optimize a health system, how to make sure patients have access to therapies and treatments
and providers.

But the critical sort of feature to anyone’s health really has to do with what motivates them — what they want to be able to do that day, that week, that month, that year, with their parents, their family, their children. And the barriers to health often aren’t the issues that the healthcare system [is] focused on. There’s social context. There’s cultural context. There’s a whole host of things that impact one’s ability to achieve their best health.

Shrank

Trying to make sure we have a health system and a way of delivering care that really meets people where they are and understands what their motivations are, I think, is the job of a really effective healthcare system. That’s a place that I continue to return to every morning when I wake up. I think about what role I could play and what role we can play as we’re trying to continually improve on the health system.

You’ve made a switch in your career from working for payer organizations to working for a very well-known venture capital firm. Could you fill us in on what you’re doing with Andreessen Horowitz and how you’re usingyour experience in healthcare, as you help decide what sort of companies to invest in?

We’re at a really unique time where the use of better technology can and will revolutionize how patients get care. And if we do it right, it will improve, I believe, the humanity of how people get care. I would love to see an environment where better technology — the use of AI [artificial intelligence], the use of data analytics software — makes it a lot easier for providers to focus on the right things, to spend their time not documenting and not addressing all the administrative complexities of a very fragmented health system, but rather talking to patients and understanding what they need and addressing their needs.

I think Andreessen Horowitz is sort of at this right intersection of trying to understand where the greatest opportunities to invest in technology will be, but also within an ecosystem where they’ve overindexed on our commitment to value-based care. There’s a great opportunity to leverage technology to support the transformation [to value-based care] and that intersection of better use of technology in a value-based care world, I think, is where we all want to be. It’s just an honor to be able to hang out with entrepreneurs who don’t necessarily have a huge amount of experience working with payers or with providers but who are deeply passionate and committed to solving a problem or a set of problems that can meaningfully impact the lives of the patients served or the providers that leverage the technology.

Is there any common thread to the problems in healthcare for which AI is particularly well suited?

There’s a huge range, from companies that really focus on simplifying a process, all the way to companies that deliver full-stack care for patients with cancer or deliver care for patients who have behavioral health problems. All the way along that continuum you can surmise the incredibly important role that AI can play.

It can be something that just makes the work more efficient — something as simple as a company that is an electronic scribe. If I’m a primary care physician and I’m seeing a patient, and if that note is being written in a thoughtful way and I don’t have to write it, that’s going to be useful, not just for me but for other providers that have to see that patient. [It might] distill the next steps into some recommendations. That all can make my life a lot easier and allow me to be much more focused on the patient in front of me and the conversation we’re having.

It could be something that just really simplifies the billing process or the prior authorization process. It could be something like a company like Memora Health, a company that simplifies the care management process. For a payer or risk-bearing provider, it could be a company like Pearl Health, which offers a rich data analytics platform to help risk-bearing providers proactively manage the health needs of the populations they serve.

It could be a company like Thyme Care that supports value-based care by offering a sort of a technology-enhanced wraparound service around those who are receiving cancer care. Or a company called Waymark that provide community health worker services to Medicaid patients to improve primary care. There’s a huge, broad range.

The way some at Andreessen Horowitz describe it is that if we were to blow up the entire healthcare system, our hope is to have invested in a bunch of things that if you piece them together you would have a better health system than you did in the first place. I think that parts of that are true — that there’s an opportunity to kind of rethink the infrastructures so that we take better care of people.

I don’t want to be the Luddite in the conversation. But do you worry that AI, rather than opening opportunities for a more personal sort of care, will routinize it, that providers will start flying on autopilot, checking boxes because their work is now regimented by an algorithm or template?

The move to technology has created some impersonal features of documentation, especially since there’s so much typing. Without better technology, this would continue. I think with better technology, it can make it easier for providers. I think that transition is always bound to be clunky. Can we get rid of some of the inefficiencies in the process? I’m sure we can.

The second point you made is, do you get to this place where things are so routinized, and you’re so dependent on an AI copilot, that you stop paying attention yourself? I’m much more hopeful that doctors and providers care about the people that they’re taking care of, and when given the opportunity to spend more time talking to them and more time trying to understand their problems, it will lead to better outcomes and better experiences.

You were the lead author of an often-cited paper published in JAMA in 2019 that argued that 25% of healthcare spending was wasteful. Has anything you’ve read, seen, heard or experienced since changed your mind about that 25% waste figure?

I think there’s massive variation across environments and systems in terms of where the reservoirs of waste are biggest. And I think they are very much related to how providers are being paid.

There’s much more administrative complexity in a setting where payers and providers are not aligned, and the provider is constantly asking for prior authorization for any procedure. And that should be meaningfully diminished in a value-based environment where the provider is at risk and the payer turns off prior authorization.

In that paper we also looked at the evidence that’s been published to date of interventions to address the six categories of wastes that we reported on. And it turns out, there’s a considerable amount of evidence, particularly in the areas of improving care.

I’m very hopeful that as we move toward more global capitation and more environment where providers are fully incentivized to drive better care at lower cost — and where the relationship between the provider and the payer is not confrontational; in fact, it’s perfectly aligned — that we will get to a much less wasteful and much more productive healthcare system [where] dollars are really being spent on the highest-value services.