DISCLAIMER: The views and opinions expressed are those of the interviewee(s) and are not necessarily those of McKinsey and Company.
Following is an edited transcript:
Pooja Kumar: Today, I’m joined by Dr. Jim Weinstein, who is the senior vice president for Microsoft Healthcare in Seattle and before that served as CEO of Dartmouth-Hitchcock Health System. We’ll be discussing his perspectives as a former hospital operator and current technology executive. Jim, thanks so much for joining us today.
Dr. Jim Weinstein: Great to be with you.
Pooja: Great. Jim, tell me about how you initially decided to go into medicine, your first career in this space.
Jim: I became sick as a child actually in high school and missed [almost] a year of school. And was quite fascinated because nobody could make a [clear] diagnosis but they had me in isolation and said I had some strange illness that they didn’t [clearly] define. And even today as a physician, I’m kind of curious [about] what it was. [During my senior year in High School], I worked [as an orderly] in a hospital, [with direct patient care,] every year until I was a junior in medical school. [That was a game changer.]
So, I just became fascinated with the hospital environment, and that there were technologies and people who actually thought deeply about helping somebody feel better. And I’d felt sick for so long as a child, that I thought, “This is a great opportunity.” So, that’s kind of how it started.
Pooja: Well and since then you’ve had so many careers, you know more than most people could dream of really, as a practicing spine surgeon, head of the Dartmouth Institute, a health services researcher, and a CEO. And now a [technology] executive.
Jim: People often ask me “How did you get to where you are? How did you do all that?” And with each journey, I always wondered, “Why me?” and, “How did I get here?” And I think what I’ve come to believe is that I’ve always been anxious to answer questions.
But what I really think is in spine surgery, I focused on cancer of the spine [and] tumors. And nobody had really described how you actually operate on those. They’re pretty rare, but pretty devastating if you have one. So, I spent a lot of time trying to figure that out, how could I actually take a tumor out and improve somebody’s life and longevity?
That was really exciting. [I] came out with a classification/treatment tool, that’s used today globally, how you actually do resect tumors of the spine. During my time at the University of Iowa, I also worked on pain models. Because all the patients I saw for back problems had back pain and I didn’t understand why. Even though I’d gone through medical school and residency, nobody actually explained to me why, the cause and effect, they have back pain.
So, we developed an animal model to study pain and I actually wrote the papers to describe why people actually have back pain and radicular pain. Then the question [becomes], “Why do you operate on them? What does that do?”
So, then we did the largest study ever done by the NIH [National Institutes of Health], at the time, on back surgery in the United States. We studied several thousands of patients across 14 states [and] 140 physician/surgeons. [We] came out with some pretty interesting studies that made the front page, above the fold, [of the] New York Times.
So, then I studied at Dartmouth and became director of the Dartmouth Institute (TDI), which studied healthcare for the United States. All of this just happened because I was asking, “Why?” And I probably asked too many times and became CEO. I’ve had a fascinating life before coming to Microsoft.
Pooja: If you think about the last two moves that you made from the Dartmouth Institute to [becoming] the CEO of a health system, and then from that to Microsoft, what has been the things that have been most surprising to you?
Jim: I think that there’s somewhat of a white space between healthcare and tech. And I think as somebody who’d studied health services research [and] the US healthcare system, I felt like we weren’t living the promise we wrote about, with the variation in our own system.
My mantra as CEO, for seven years, was to create a sustainable health system—not a healthcare system. And so, if you think about that, that’s a very divergent path for a healthcare system today.
And so, I did studies that said, “Less is more,” in spine surgery for example. And everything we studied in the [Dartmouth] Atlas showed tremendous variation, for many surgical procedures but [also] for heart failure, bypass surgery, [and] stents.
I break [“a sustainable health system”] down into three specific areas. One, “Is there effective care being provided by our system or our country?” When I mean effective care, things that are evidence-based like aspirin to prevent heart attacks.
[There’s also a] supply-sensitive [care]: Is it because we have more beds in Boston than we have in New Haven, that more healthcare is utilized and you’re more likely to die? But actually, you are more likely to die living in Boston than New Haven. And is that a supply-side issue?
So, again, “Why?” So, supply-sensitive care was important. [The last one] is preference-sensitive care. [Plus, a] notion that where two treatments [are] equal, a patient should have the choice.
I tried to develop a sustainable health system that embodied, “Let’s do effective care and let’s have a stream by which that occurs. Let’s not build more beds, even though that might make more money for us. Let’s not get into supply-sensitive care. Let’s start the first-in-the-world shared decision-making center, which we started at Dartmouth-Hitchcock, where patients are given tools to make decisions when [their] preferences matter.
[In the tech world,] these are some of the smartest people I’ve ever worked with, but they don’t [necessarily] know healthcare. I think the cloud, [which allows] for access to things that you don’t have to have on premises, security issues, backup opportunities, is incredible. We shouldn’t negate that.
The problem is, health systems are afraid and don’t trust to take their patient data, PHI [protected health information], and put it in the cloud. And I think as a tech company, sometimes you don’t appreciate what that means because you haven’t had that onerous responsibility of taking care of tens of thousands of patients who are very sensitive to what happens to their information.
Pooja: You’ve talked a lot about incentives, which I think are a key issue in [healthcare] and something that, for example, ACOs [accountable care organizations] initially tried to set out to address, amongst other types of payment mechanisms. Tell me a bit about what you see as the most exciting things on the horizon in terms of trying to solve the incentives problem?
Jim: So, what I would suggest is we think about first of all, a community by community approach. You can define a community by ZIP codes, by economics, by school districts, [or] by YMCA locations. We as a society have not yet defined a [health] system that we want to support and how we’re going to support it.
We’ve taken a nation of very diverse communities and people and put one system in place—a federal system. We have several commercial systems, but they’re pretty much fee-for-service, even though [some might suggest] they’re “value-based care” and I would ask, “Value to whom?”
So, how do we do this? My belief, right or wrong, is that we should take a community, I don’t care if it’s Baltimore or Chicago or LA [and ask,] “What is the population of those cities? What are their needs?” Not, “What can we build and offer them and get paid for?” [Instead], “What are the healthcare needs of those communities and those specific populations?”
And then provide the resources to address those needs, not [the] reverse. [This would let us] probably get to [the] reality of the social economic issues as being the most important issues to address. And, “By the way, I do need a hospital, [but] X number of beds less than I have now.” And, “I [might] use Walmart or Walgreens for my primary care visits.” And, “I don’t need all these diagnostic MRIs and CT scans,” which all goes back to the productivity question.
Because every system’s designed to get the results it gets. And we’ve built a system that is not performing well, because we’ve not built it with what we need to do today. We built it in the ’60s based on trying to get insurance. We revisited that under President Obama to get insurance, but we never actually designed the system to get the results [people need].
I would love to have an experimental pilot community where we, Microsoft, could partner with a health system, with [the] auto industry, with an insurance company, a pharmacy, and say, “Now we’re going to take this model city, [which is] five plus million people, and we’re going to work together to redesign this to what they really need and do the financial model that will be very different because; the workforce needs will be different, and the distribution of care across the continuum will be different.” That’s a dream I have.
And the data is critical. Interoperability, which is big for Microsoft and for healthcare and [is being] pushed by the CMS [Centers for Medicare & Medicaid Services] at a federal level, is really important.
But you have to architect things to be interoperable and to share them. I think if we could get as I said the food chain and the health systems and the commercial markets that people visit every day, and put all that data, which then gets to big data and artificial intelligence and machine learning capabilities, then we could do something for society with greater probabilities using the data in the right ways.
Pooja: Well one way to think about tech and back to the issue you raised that there is not a tech solution to every problem: There’s also not a match for every tech platform or solution that exists out there today. Where do you see the exciting avenues for tech to go?
Jim: I think Microsoft is really interested in the efficiencies of the workforce. So, “How do you not just become efficient but effective?” I think several of the product lines that have been announced by Microsoft, things like EmpowerMD [and its partnership with Nuance] that use natural language processing is used to [improve] a physician-patient encounter and make it so the doctor doesn’t have to sit there typing with his back to the patient.
EmpowerMD [and Nuance] will help to revolutionize that [exchange between the physician and patient]. But I also will be cautious about how fast. Because don’t forget, as I well know as a physician, and as somebody who led a health system, doctors [can be] slow to adopt new things despite how wonderful they are. When we implemented the electronic health record, because we were early in this space at Dartmouth, you know. We went big bang. You know, inpatient, outpatient, everything. And it was hard. And I had to work in all the spaces because the electronic health record’s different in the emergency room, it’s different in the operating room, it’s different in the clinic, and it’s different in inpatient.
And, you know, everybody had to be trained, let alone try to do their work. And, you know, I think we’ve come a long way. And now it’s wonderful, the gains we’ve made in such a short time of getting information into an [electronic] system. The big issue now is, “How do you use it?” And, “How do we make it useful at point of care?”
Pooja: Are tech companies natural partners to healthcare systems and payers and other stakeholders in addressing productivity? Why or why not?
Jim: I think Microsoft wants to be a natural partner. And I think there’s different parts of the health system, as you understand. There’s the payer side, which is a very different business of, managing claims.
Or talking about benefits, that’s where bots and chat bots can be really helpful for patients to, [for example,] check their benefits. The bots can deal with those [tasks] quite effectively. And I think on the administrative side of a hospital [as well]. If a patient’s calling in to ask questions, you can probably set up a chat bot to do a lot of that effectively.
On the other hand, I think the notion that tech is going to save you time, I’m just not sure yet. Because I haven’t seen it. And so, I think we have to be really careful what we promise and not overpromise.
I did some work for the federal government on inequities. You know we have more people in this country now on Medicaid than we do on Medicare. That’s going to grow further. People talk about equality. And I say, “Equality’s interesting, but if I give everybody a size eight gym shoe, that’s not equitable.”
And so, the reason I’m bringing this up is technology’s interesting, but we have to be careful about how it misses some of the largest parts of the population. We have to be sensitive to solutions for all, not just some. I hope to influence that more, because it’s going to become a bigger problem in the world.
Pooja: [We often] talk about technology as being this silver bullet and it’s going to solve everything—cancer, diabetes. What problems that are productivity-related in healthcare do you think that’s not something that tech will solve, whether it’s because it’s primarily an operational problem, or primarily a cultural problem, or something else entirely?
Jim: You have to have the right culture to move change forward. That’s hard. I think most leaders who try to do it appreciate that comment. I’ve been privileged, as you said, to be in a lot of positions, and lead a lot of organizations, and [drive] a lot of change.
But I’ve had a lot of pushback sometimes too. They don’t always tell you as the leader because people are [oftentimes] afraid to talk to the leader directly. But believe me, I heard it. But all of us need to do better. We tend to want to have a solution for everything. And there is not a solution for everything, especially in healthcare, and quite frankly in almost everything.
I don’t think technology will be successful and artificial intelligence in the way people might perceive. Everything’s not an artificial intelligence problem. And, I think there’s risk in artificial intelligence. People can corrupt networks and you might not perceive it. What if my model didn’t account for you, as an African-American female, but I’ve used that model broadly? We could end up overdiagnosing or underdiagnosing in exponentially greater numbers than we are now. I like the idea of validation of models.
[It’s important to be] honest about what’s possible and be excited about that future, but we have to also deal with today’s inefficiencies.
Pooja: Absolutely. I’d love to push a bit more on the automation piece, especially in light of the waste that we know is inherent in administrative tasks in the healthcare system. Do you see tech actually replacing automated tasks that people do in a big way in healthcare?
Jim: Yeah. But, we have to start training people differently for a different workforce future. What is a future radiologist’s work going to be? If a chest X-ray can be read better by a machine, a plain chest, I don’t need to have someone sit there all day and read them.
But do we need to do some quality checks, on certain numbers, to make sure that there’s an agreement? Or do we use that information in a quality way to improve the training of future radiologists? It’s just an example. But we’re not designed as a system, as a country to do that. I think many things could be done in an automated way.
If you think about consolidation, to me those are the things—supply chain, revenue management, pharmacy—that could be consolidated across many organizations with a centralized set of services [while] still maintaining privacy with your own container of information, but why not centralize all that? Why does every hospital need an IT department and a revenue management department and a supply chain department, a pharmacy department? They don’t.
With a $20 trillion deficit and a growing healthcare economy we can’t sustain, productivity becomes really important, amongst many things.
Pooja: I’d love to ask, if you take providers and payers in turn, what would you say to leaders who want to think total transformation for their systems in terms of a productivity journey partnered with tech companies, such as Microsoft, or other types of players? What advice would you have for them? What would you say they should be thinking about?
Jim: The inequities in our society are growing. There’s a paucity of solutions that are solving the problems for those most in need. Yes, we have drugs that are very exciting [but] costly, but they are not a solution for a society this large.
[These] partnerships with payers and providers [and] pharma and policymakers are really important right now. I’m hopeful Microsoft will lead much of that. Because to your point about productivity and efficiency, it’s in those businesses, too, which impact the provider organizations as well.
My latest role is around strategy and thinking about these issues and data analytics, because I think they’re the foundation of the solution set that you’re asking about. Those payers, which have tremendous power, including the federal government, the VA [US Department of Veterans Affairs], need to be critical partners in all the things we’re talking about.
Pooja: It is so true. I think we, as a society, have created this very fragmented and quilted healthcare system to try to cover who we can, where we can, and to the level that is tenable at any given point in time. But without really stepping back and thinking about how to do this in a better way. I think as we looked at the issue of productivity, and the cost of care keeps rising, there really does [surface] an equity question, people are falling through the cracks all the time, or not.
Jim: And it’s going to get greater. And we judge things the wrong way. We need big ideas to share, to make everybody’s boat rise and move the whole curve to the right and stop regressing to the mean.
Pooja: Jim, thank you so much for taking the time to speak with us today. It was a real honor.
Jim: Thank you.
For more on specific opportunities to better control healthcare spending, see “The productivity imperative for healthcare delivery in the United States.”