Atul Gawande: To fix our broken healthcare system, start with primary care

The ex-CEO of Haven—Amazon, Berkshire Hathaway, and JP Morgan’s just-ended health venture—on what he’s learned as a surgeon, researcher, and entrepreneur.

Atul Gawande: To fix our broken healthcare system, start with primary care
[Photo: Haven Healthcare]

After three years, Haven, the much-anticipated health venture from Amazon, Berkshire Hathaway, and JP Morgan, is shutting down. Little information came from the company itself on the decision to end the business, merely stating that its founding companies would collaborate further in the future. When it disbands at the end of February, its staff will be divvied up among its three founding partners.


Haven was an attempt to exert some control over ever-rising healthcare costs while also ensuring quality. Dr. Atul Gawande, surgeon, researcher, and professor at Harvard Medical School, helmed the effort. Gawande was a surprising choice. As a professor of health policy and management at Harvard and founding executive director of Ariadne Labs, an incubator for health systems research, his healthcare credentials are formidable. He’s also written some compelling articles on the health industry for publications such as The New Yorker. However, as an academic and practicing surgeon, he’s something of an outsider in corporate America.

Few details about Haven emerged after it was announced to great fanfare in January 2018. The company has said only that it “explored a wide range of healthcare solutions, as well as piloted new ways to make primary care easier to access, insurance benefits simpler to understand and easier to use, and prescription drugs more affordable.” The project seems to have been beset with issues, though: Last August, Stat News reported that Haven had trouble retaining talent. The initiative has also been accused of lacking clear goals.

Another possible impediment is that the project seemingly faced direct competition from one of the companies that founded it. During the Haven years, Amazon made an independent push into healthcare. It bought online pharmacy PillPack and launched services such as a digital-first employee health offering called Amazon Care, a HIPAA-compliant data ecosystem for healthcare providers, and voice transcription for doctors.


In May 2020, Gawande stepped down as Haven’s CEO, though he continued to serve as chairman. He’s since been busy working at Brigham and Women’s hospital, teaching at Harvard, and chairing Ariadne Labs as well as a new COVID-19 testing initiative called CIC Health. He also serves on President-elect Joe Biden’s COVID-19 task force. Shortly before the news of Haven’s disbanding came out, he talked with Fast Company about his experience at Haven and the evolving business of healthcare in the U.S.

This interview has been edited for length and clarity.

Fast Company: You’ve had wide experience in the healthcare industry. What did you learn from being a CEO?


Atul Gawande: Before I came to Haven, I was the founder and executive director of a research center where we deployed healthcare solutions at a large scale. The critical thing I learned from effectively being the CEO of that kind of enterprise was all the usual things about getting your priorities ultra clear, being able to build your team around delivering on those priorities, and having the right kind of people for those tasks. And then really constantly working to keep everybody pulling in the same direction.

Being at Haven introduced for me two big things as CEO. One was understanding hiring, leading, and enabling the power of tech capabilities, and software engineers in particular. And then the second thing was the speed at which you could work to move things where you have some influence and control over the incentives as well.

When working from the outside, partnering with governments, partnering with the private sector, trying to introduce new methods in a non-nonprofit academic world, I couldn’t change the rules of the game unless the government was going to change the rules of the game. At Haven, you [had] the potential . . . of getting to set incentives for those you are working with, whether it’s how you design benefits or how you design the delivery of care.


I have a rule of thumb, which is that you have a variety of things that you need to pull together to make things happen to change healthcare. You need access to the patients, you need access to the clinicians, you need access to the incentives that guide the patients and the clinicians, and you need access to data. Nobody has all four. You’re always missing the financing and the capital or the access to the data or the access to the providers. So you’re constantly needing to pull partnerships together.

If you own all of [the pieces], you end up being in a narrow space where it’s hard to scale. You always end up having to partner either to get access to more providers and the patient populations or to get access to data. Your choices are you’re either small and have control of all four or to get to scale you have to learn to partner.

FC: During your time at Haven, what challenges were there in achieving better outcomes, better satisfaction, and a better cost efficiency?


AG: What I’ll say in general—across the last 20 years of doing this in a variety of different settings including now in lots of pandemic work—is that the greatest challenges are: you either have access to the patients, but you can’t change the providers and what they’re doing, or you have access to the providers and you can’t change the patient population and the incentives that are driving the ship.

FC: What is the most successful health project you’ve worked on?

One of the most exciting projects I’ve ever had was when [I] got to work with the government of Estonia and their national association of family physicians, who are the primary care clinicians of the country. This was a country where they have one and a half million people all in a single-payer system with a modernized electronic medical record system that is all accessible and tied into their Social Security system, their disability system, and everything else.


The one thing they didn’t have was capital, but you could go in and work with the clinicians to set up structures where you optimized for the big issues that they wanted to address and work on. [For example], men die 10 years younger than their women in the country. And they still have a hangover from the very low life expectancy of the former Soviet Republic. It’s alcohol, cardiac disease, diabetes, and addressing those issues in a systematic way, making sure people don’t fall through the cracks. The health system can be part of catching people and getting them back on track.

Trying to make that happen in the United States, whether you’re doing that from an employer point of view and insurer point of view—we don’t have those natural pieces put together. And that is the constant uphill battle. We have a fundamentally broken system by having a healthcare system built around where you work. It means when you stop working, you lose your healthcare, you lose your primary care. You don’t have the connections sustained for big parts of your life like it should be.

FC: And untethering healthcare from work is ultimately up to politics.


AG: Yeah. It’s all ultimately government and the ongoing national debate.

FC: What do you think about this confluence of payer-providers like Aetna and CVS? Do these kinds of systems have enough incentive to provide high-quality care, or does the ability to tightly control costs conflict with that?

AG: When they become this kind of vertical stack of insurance and the provider you have the potential for the right incentives to come together. You are never going to be free of needing regulation to ensure that quality is going to need to be there. The big challenge is that people are hopeful that there’s competition that would drive the CVS folks against the Optum folks, and you have therefore some choice and competition driving how these kinds of vertical structures work.


The reality of the country is that two-thirds of counties in the United States really don’t have more than one major provider available to them. And that is a problem for how you achieve what you’re going to achieve.

We don’t have a system for the fact that half of your life or more is going to be spent with at least one or more chronic illnesses.”

A century ago, we really finally drove childbirth from being one of the biggest killers for women and newborns to becoming a rare cause of death for women and a much, much lower likelihood of death for children. It was a combination of the government and the private sector. The private sector innovated around how you save the mothers’ lives and babies’ lives. But public information—a century now of [recorded] death rates of every hospital, every death of the baby and the mom—total transparency about where things go wrong [also played a role].

The Department of Public Health would potentially shut down a hospital if it was having bad results. It would hold people’s feet to the fire to maintain a minimum standard of care. You could get real competition between places over everything from the experience of care and whether they have a Jacuzzi or whether they have great clinicians. But you weren’t competing over whether your baby would die or the mom would die in one place versus another. The lesson I’ve [learned] from working on the nonprofit side, working on the private sector side, and having worked on the government side is: Given how much the big have gotten bigger, there will need to be much more data transparency about the actual services that are provided and what the outcomes are—whether it’s childbirth, primary care, cancer care, or surgery.


There is going to need to be close involvement from people in the public health system around whether it’s serving its minimum quality and appropriate levels of care. And the pandemic has made that eminently clear. We don’t have the basic information on whether testing is being done and where, who has access, who doesn’t have access, how well it’s being done, where the holes are, how much mental health has been damaged, how much our specialty care has been affected.

FC: Before the pandemic struck, there was a big trend towards consumerization of healthcare—the unbundling of healthcare services into tests you pick up at the store or apps that you buy for a specific ailment or subscription services. How does that fit into the overall healthcare ecosystem?

AG: I think it’s a natural consequence of weakened primary care. And it is extremely damaging to the long term health of people.


You have to be able to take some chances and move new things out into practice without putting people at risk of their safety.”

Primary care is a service where you should have ready contact for the majority of your medical needs at any time of day. It should be a team that you have contact with, who knows you well enough to know what your goals are and how you’re doing over time against those goals. And it then should be able to deploy the assets of the healthcare system, which are enormous, right? Whether it’s highly complex, specialty care, getting you a COVID test, or recognizing your goal is to get pregnant and getting you on the right path there. It should be a system that could enable all of the assets of the system to serve your needs and then follow along to make sure that when it doesn’t serve your needs, you’re in good hands.

That basic service is a fundamental building block of many systems around the world, but it is not of ours. We don’t have a system for the fact that half of your life or more is going to be spent with at least one or more chronic illnesses, whether it’s high blood pressure or a complex heart disease.

My son has had, since his birth, congenital heart defects. Having someone who can follow you over time and bring you the powers of what science is discovering can save your life years in advance. When you bounce around the system, you miss out on that opportunity. And now that we’re in a world of genomics, a world where we’re able to make changes that can affect your outcomes 40 to 50 years from now— it’s crazy that we are responding to it by severing the relationships people have over time.


When you’re young and don’t need much, you don’t value that until the day that you have a major issue and then don’t have anybody who you can call, who knows you.

FC: What do you ultimately think is the best economic system that we could have for healthcare?

AG: I think the way that Medicare Advantage is evolving, Medicare [Accountable Care Organizations], some of the capitated [fee per patient] Medicaid plans is the ideal way to go. And what is that? It means you have a primary care relationship. Dollars flow on a monthly basis to that team to meet a certain level of convenience and efficiency, to get the full range of your needs from medicines to preventive care, to be advocates for you. And if they don’t do a good job, then you take your payment from them and it goes to the next clinician and the insurers reward them for being able to deliver on a certain quality and manage the cost constraints. That has been immensely popular in Medicare with more and more people flowing into that part of the system. I think it would be of great interest if we were to expand in the under-65 population.


FC: What do you think about companies that have no experience in healthcare getting into healthcare? Because we’re seeing that quite a bit.

AG: It’s not uncommon that they come in saying, “It’s crazy that this thing doesn’t work. We know how to do service,” and so on. And they’re often right that they’ve got some real skills to bring. But healthcare is such a different production model for innovation. One of my board members at Ariadne Labs was David Ebersman, who was the chief financial officer in new business development for Genentech, a drug company, and then Facebook. And we often talked about how at Genentech, [when] you get to work on a new innovation it takes 10 to 13 years and you get one chance of iteration with a clinical trial. And at Facebook, you get to throw a bunch of 20-somethings into a room, solve a problem on the internet, throw it out in eight weeks to a hundred thousand people, and see what happens.

The model in doing healthcare innovation is exactly in the middle of that. You can’t just come up with a solution and throw it out to a hundred thousand people and say, “Hey it’s okay if 90% of the people have a terrible birth experience and it makes their kid worse.” You have a lot of learning before doing, and iterating before you can deploy it. And yet you can’t be so slow that it takes you 10 years, like with a drug. You have to be able to take some chances and move new things out into practice without putting people at risk of their safety.

It’s also not as tightly regulated to [make] change [in other industries]. So when people jump out, they imagine that they can innovate the way that they innovated in the world that they came from, but the innovation cycle is extremely different. We’ve only had about a decade of really bringing tech to changing delivery in healthcare. We have 30, 40 years of doing that in retail. We haven’t even learned how to make innovation organizations that really know how to do this well in healthcare.

FC: Let’s talk about retail. You see companies like Walmart, which on the one hand has this huge benefit of incredible scale that could help fix the rural health problem. On the other hand, it has a less-than-great track record of how it treats its employees.

AG: Places like a Walmart or an Amazon or others who are in the services side have paid a huge amount of attention to the experiences of the person who comes through to make it a better experience for them. A lot of discipline around cost and a recognition of how important it is to drive for scale. Those are the things we do not bring in healthcare. They need to marry that with a deep understanding of the complexities in healthcare and building relationships that are not just about the momentary transaction, but the reality that people need someone who will stay with them for years of their life. You can create incredible outcomes for people when you enable that relationship.

The national chains of the world haven’t yet embraced that notion. [CVS] MinuteClinics solve that acute problem, but they don’t do what really great systems around the world do, which is help people on their journey, to access all of the powers in the system to optimize a better outcome per dollar. It’s very hard to find examples of organizations that have really been committed to doing that.

FC: Healthcare has come under a lot of pressure this year because of COVID-19, and healthcare institutions are not doing well. But there were problems before the pandemic. I’m curious if you think this year will spark change or if we’ll just sort of weather through it and keep on keeping on?

AG: I have a complicated answer. Part of it is trying to figure out how much of the healthcare system has been permanently damaged and how much is just like a train being taken off of the track. And then when COVID has gone, it’s going to be a little creaky, maybe a little rusty. But you put it back on the track [and] it’s going to sail ahead.

Primary care, in a pandemic in particular, is where you can really look after a population of people.

Very clearly, fee-for-service [the model where a patient pays a fee for an individual medical service] is not made for a crisis like this. And so the systems of health that have been most damaged are the ones that haven’t had the reserves to buffer against what’s happened. What are those parts of the system? Number one has been rural hospitals and practices, because they’re typically small and have not had a lot of cash on hand. And you’ve seen a lot of closures in those communities that were already underserved.

The second group has been primary care. It’s been always underfunded relative to specialists like me as a surgeon. And yet primary care, in a pandemic in particular, is where you can really look after a population of people, make sure they don’t fall through the cracks for anything from vaccination to their medical issues. That fragile system has been damaged as well.

One of the consequences is that in the [business of] healthcare part of the system, we have seen the big get bigger. Distressed practices and hospitals have been bought up by the bigger systems. As the big get bigger, they can demand higher fees, but the direction we really should be going is towards paying for better population health in the system.

The really interesting thing that has emerged is some inklings of a better system. Massachusetts and North Carolina, for example, have adopted the Blue Cross Blue Shield insurers. And those two states adopted a rescue package for primary care where they committed to paying them to sustain their incomes per patient at the 2019 level as a flat fee instead of a fee for service. They hold them to requirements to contact the patients, be in touch with their needs, and maintain a certain level of quality. If they actually control costs as well as improve the quality—and that’s to a large extent what’s been happening in Medicare Advantage and in Medicare ACO—that move away from fee for service towards paying for the full range of population services from prevention to acute care to chronic care management is the forward direction. I expected that to be a shift that would happen even more rapidly [in the U.S.], and I think it’s not yet clear that that will happen that rapidly.


About the author

Ruth Reader is a writer for Fast Company. She covers the intersection of health and technology.