Aneesh Chopra: Risk-Based Contracting Is Succeeding—Right Now

Oct. 17, 2023
Aneesh Chopra, one of the most respected industry leaders around the development of alternative payment models, shares his perspectives on this moment in APM development

Aneesh Chopra served as Assistant to the President and Chief Technology Officer in the Executive Office of the President in the White House of President Barack Obama, from May 2009 through February 2012, helping to develop policy around technology and to guide thought leadership around technology. He currently serves as cofounder and president of CareJourney, the Arlington, Va.-based firm that provides software solutions and consultative guidance to provider organizations involved in value-based, including risk-based, contracting. He spoke recently with Healthcare Innovation Editor-in-Chief Mark Hagland regarding this moment in the ongoing evolution of value-based care delivery in U.S. healthcare, and particularly around the subjects of alternative payment models (APMs) and accountable care organizations (ACOs), both those models and programs being developed and managed by the Centers for Medicare & Medicaid Services (CMS), via the Medicare Shared Savings Program (MSSP) and ACO REACH, and those being developed by private health plans, either through the Medicare Advantage program, or in some other form. Below are excerpts from that interview.

What is your sense of the APM landscape right now?

We are in a “Tale of Two Cities”: we have seen consistent and improved MSSP performance—let’s call that the Little Engine That Could—that has unequivocally delivered results to taxpayers and patients. ACO REACH will show that the possibilities can expand to capitated models and offer early evidence that we can do more, if we can deliver on these total-cost-of-care models. So I see MSSP and ACO REACH as building on each other, where one is demonstrating real-world evidence that we can accelerate the benefits of VBC through these models. While MSSP and REACH have certainly captured a sizable portion of the market in terms of provider participation rates; at the same time, a majority of providers are not enrolled.

So we are approaching a moment of convergence, where alternative payment models have created muscles for better primary care operations, including tighter referral patterns around high-value specialty care. And what I’m observing is that employer-sponsored care, probably the sector the least enrolled in value-based contracts until now, is ready to include those high-value principles into their operations.

“Value-based care the noun” means a provider signs up to contract with a payer; and there’s some complexity for the average employer-sponsored plan to administer. However, there are mechanisms for the employer-sponsored world to double down on providers enrolled in the federal programs. So the convergence is that value-based care the Verb may result in high-value steerage, patient nudges, fee-for-service bonuses that are not currently part of the traditional one-size-fits-all payment structures. So we may find the State of the Union of value-based care to look more positive and bullish when we ask how many employers, managed care organizations , MA [Medicare Advantage] plans, etc., incorporate high-value care and specialty routing into their operations. In that sense, the state of the union is strong, as the muscles are built.

Are you concerned about the relatively small numbers of organizations participating in the MSSP?

Actually, it’s a sizable number. If we could generate for you a list of what percentage of primary care doctors are treating patients in some kind of ACO model, it’s around 25 to 30 percent.

One advantage CMS has is that it has made a commitment to disclose, at the beneficiary level and provider level, how many people are enrolled in APM models. We have 2022 data, possibly 2023 data.

Is the pace really picking up and accelerating overall?

It’s an excellent question. When I show you the combined statistics of MSSP, REACH, and Primary Care First, the three anchor tenants in the building, MSSP has been flat year over year; however, if you add in the three together, it’s about 15.5 to 16 million lives… that should be 42 percent of FFS lives enrolled in those programs. That is not a small number. And REACH accounts for almost all the net new growth. It’s north of 3 million. And MSSP is at around 12 million.

And if you haven’t seen it, HCP LAN, the Health Care Payment Learning & Action Network, a stakeholder collaborative, puts out an annual State of the Union on APM adoption across commercial, Medicare Advantage, Medicare FFS, and Medicaid. And there’s been a shift: they categories four categories of risk contracts. And CMS deems categories 3B and all of 4 as the end state. And that number has grown. That’s a survey model. And we have actuals for CMS. And I’ll add that CMS late last week released the news that the MA Value-Based Insurance Design Model has gone from 1.5 million lives 4-5 years ago, to 12.4 million lives for next year. So if you look at the full picture, the Little Engine That Could, combined with these related models, shows the ball is moving down the field, even though the headline perception remains that’s a minority. The pieces are in place for it to take on a much more institutional role, because of those participating.

It's not your grandfather’s AI from eight months ago!

You’ve nailed it. The world changed November 30 with the release of ChatGPT. Why did it change? We’ve had AI winters and AI springs. And to those who’ve historically seen AI as better math for predictive analytics, there’s been hype and frustration over the years. However, generative AI shifted the landscape from better predictive models in general, towards productivity-enhancing tools for everyone in every industry, across the board.

So while I’ve been monitoring and supportive of more R&D of AI, more real-world testing of AI in the tradition prediction space, I’ve gone all in on harnessing the possibilities of true productivity growth in healthcare, because of generative AI. When we initially invested in the HITECH Act, a big goal was that the derivative benefits of digitization would be the same levels of productivity growth as what Alan Greenspan noted in the 1990s when explaining why the American productivity was so… it was computerization.

However, a decade after the HITECH Act, it’s really hard to see evidence of healthcare productivity improvement. So we’ve had this computer investment productivity paradox nagging at us ever since the HITECH Act was passed. And what excites me about generative AI is that I see a direct line between investments, regulations, and the actual productivity growth that we need in HC across the board. So I am extraordinarily bullish on AI and transformer models, with the most obvious manifestation of transformer models being LLMs. It’s a productivity question, not a technological one.

What about the potential dangers?

I have much angst on the upside as I have on the downside. Let me be really clear: as with any technological advancement—the advancement of the Internet has brought great societal good, but also social media disinformation and cybersecurity risks. So I am not a Pollyanna. I’m focused on opening up while locking down. And so the goal is to maximize the benefits while mitigating the harms. And where I’m spending my personal time on the issue of AI is trying to strike the right balance between industry guidelines—call them regulations, rules of the world—and the critical need to foster more experimentation, more shots on goal, more investment—and the key is the need for enforceable codes of conduct that allow this nascent industry and the large universe of HC IT purchasers to ensure we are elevating the bar on what these systems have to do, while opening up the models for bias, for security risks, for other unintended consequences, so that we can proceed with caution but comfort. It won’t be a single vendor convincing a single organization to take advantage of these possibilities, but instead, it will have to be a multistakeholder, collaborative effort.

What should the c-suite leaders of health system leaders be thinking broadly, in that area?

As a base principle, are you subject to interoperability regulations? If yes, then the question to the board is a fairly basic one: will we embrace this opportunity, or will the folks who have access to the data because of interoperability, do it for us? Let’s take the example of Getty Images, with its proprietary library of mazing images. You can’t use those photos. But last month, Getty announced that it was building its own LLM. Now there will be a publicly available tool with access to the Internet, and there may be a private version that can be bought. So those with IP rights to healthcare data might organize information. That could be a product growth opportunity for those with intellectual rights to data that nobody else has. That’s Path 1. Path 2 is the challenge/opportunity of interoperability: if I won’t provide these services, will someone with access to data, do it to me? So I believe that because of the CURES Act [the 21st Century CURES Act, passed by Congress and signed into law by President Barack Obama on Dec. 23, 2016], where by luck, all these LLM capabilities are coming online while we’re simplifying access rights to consumers and doctors, that should unleash a very competitive marketplace that I hope will provide for consumer and societal benefit, or at a minimum, corporate benefit. And in my opinion, there isn’t a boardroom in HC that hasn’t already had a board meeting where AI hasn’t been discussed. The limitation is a mixture of comfort and caution on account of us not having a clear regulatory roadmap. So the clearer we can get that regulatory roadmap, the more likely we’ll unlock possibilities; there’s just too much potential upside.

Sponsored Recommendations

State of the Market: Transforming Healthcare; Strategies for Building a Resilient and Adaptive Workforce

The U.S. healthcare system is facing critical challenges, including workforce shortages, high turnover, and regulatory pressures. This guide highlights the vital role of technology...

How AI-Native Locating Intelligence Revolutionizes the RTLS market

Discover how leveraging an RTLS solution with artificial intelligence as the location engine can increase efficiency, improve safety, and elevate care without the compromises ...

Harnessing the True Power of Cultural, Clinical and Operational Data

Optimize healthcare performance by combining clinical, operational, and cultural insights. A deeper understanding of team factors improves care and resource management.

How Digital Co-Pilots for patients help navigate care journeys to lower costs, increase profits, and improve patient outcomes

Discover how digital care journey platforms act as 'co-pilots' for patients, improving outcomes and reducing costs, while boosting profitability and patient satisfaction in this...