Healthy Conversations

Listening to Brad Bostic, founder, chairman, and CEO of HC1 -- and Daniel’s latest guest -- you can’t help but be excited about the future: “We've got this incredible opportunity that's once in multiple generations to advance the ball, and it's because you've got access to medical information that's digital unlike you've ever had, and you've got this access to compute, and you have a collection of really intelligent, committed people working on these different areas of innovation. You put all those things together and there could not be a better time to accelerate in healthcare.” He explains how the use of digital twins, and being able to be predictive sooner, can help.

Show Notes

Listening to Brad Bostic, founder, chairman, and CEO of HC1 -- and Daniel’s latest guest -- you can’t help but be excited about the future: “We've got this incredible opportunity that's once in multiple generations to advance the ball, and it's because you've got access to medical information that's digital unlike you've ever had, and you've got this access to compute, and you have a collection of really intelligent, committed people working on these different areas of innovation. You put all those things together and there could not be a better time to accelerate in healthcare.”  He explains how the use of digital twins, and being able to be predictive sooner, can help.

What is Healthy Conversations?

Healthy Conversations brings together leaders and innovators in health care to talk about the biggest issues facing patients and providers today. Every month, we explore new topics to help uncover the clinical insights and emerging technologies transforming health care in real time.

Brad Bostic:
The future is really all about empowering individuals to have a better understanding sooner of what they can be doing to improve their health, but having more of an operating system that directs that.

Dr. Daniel Kraft:
Welcome to Healthy Conversations. I'm Dr. Daniel Kraft. In Healthy Conversation, today with Brad Bostic, who's the founder, chairman and CEO of hc1.
Brad, you're a technologist, entrepreneur, data scientist. We're in a data-driven health care world, but we're still using fax machines to communicate and the CD-ROMs to transfer imaging data. Maybe let's dive into the pain points you're trying to solve for today, and then we'll arc to the future.

Brad Bostic:
As you say, there's increasingly now digital information, but it tends to be siloed and in many cases inaccessible.
What I saw with lab data was that it was the most effective way to tell a comprehensive story about what was going on with patients.
I was very fortunate to have a mentor early in my career, who was a very entrepreneurial surgeon. I was dating his daughter at the time. Now I've been married to her for 25 years.
So, he would share with me back in the late nineties, how he saw this need to move to a more value-based model and that we really were in a situation that was not sustainable financially if we didn't get that done.
I went to work for Ernst & Young, doing health care consulting and saw a lot of how things were working and how things weren't working.
Around that time, in the late nineties, my mom ended up getting sick with Stage 4 cancer. I saw the firsthand issues that resulted when health care was disconnected as it was.
That was what lit the fuse for me. From there I got into health care interoperability, looking to do a better job of bringing data together in common health records.
Fast-forward to today, where we're focusing on how you can transform health care by ensuring the right patient gets the right diagnosis through precision testing, and then ultimately on the right medication.
So, it's been quite a journey. Everything in health care I think is at a minimum of 10-year overnight success.

Dr. Daniel Kraft:
How do you try to solve it to make it more precise and personalized?

Brad Bostic:
If the data you're looking to leverage is not digitized consistently, you're in a losing proposition.
I zeroed in on, what can we do to be the best in the world at constructing these longitudinal patient profiles?
Now, setting out to do that, it sounded like it would be easier than it has been. A CBC generated in one lab is different than one in another lab. There are different analytes that are used, different instruments.
So, a big part of what hc1 has cracked the code on is using machine learning to harmonize and make sense of all of this lab diagnostic data at scale, and then zero in on where there are risk signals that say this individual patient is getting over tested, this is wasteful. Or more importantly, this patient has certain tests that should be done, that haven't been done.
In that example, physicians aren't looking for more computers to try to enter data into. Heck, when you're in medical school and you do six months worth of work on what lab testing is about and then you fast-forward 15, 20 years, there's no way to keep up with all of the latest, greatest, best and stay on top of what's going on with patients.
So, just bringing that component of, as you put it, the arc into the equation, is something that has been incredibly transformational where we're actually impacting lives every day for millions of people.

Dr. Daniel Kraft:
You mentioned a key word for clinicians, which is workflow. Often, the workflow is hit the easy button. Hit the whole panel of tests or not even know what's best. So, either over or under test.
Can you give us an example of how in the workflow platforms you've built, you help clinicians choose the right test to move the needle?

Brad Bostic:
There are really two levels to that. One is just a system-wide lab stewardship level, where you've got chief medical officers that need to minimize clinical variation.
People who are in administration and health systems, they generally will be like, "Yeah. We're already doing a great job with how we're doing our testing." We say, "That's fine, but let us go ahead and plug our system, in though."
Invariably what happens is you'll start seeing that, for example, in inpatient settings, you've got repeat normal tests that are just on a standing order that are run every day.
Doesn't it seem like that would be pretty easy to eliminate? But it's kind of rampant. That also exposes patients to potential infection. That's unnecessary. There's a lot of negative to that.
Proactively, from a precision population health level, identifying those patients that aren't even coming in for a care episode who have needs like, hey, I've identified this person has an elevated A1C. They're pre-diabetic. If you just stay on top of that, you can manage it.
Unfortunately, a lot of those folks don't ever show up again until they land in the ER, having some kind of a crisis.
Both sides, the bigger picture lab stewardship analytics, driving the best practices in an objective way, as well as the clinical decision support that injects those kinds of signals into the EHR workflow, those are equally important and they kind of need each other.

Dr. Daniel Kraft:
Right. So it can be as simple as in my inpatient experience, CBC every other day and chem-20 every day, which may not be indicated, but it ends up being stuck in the record.
The other challenge is synthesizing all that information. Obviously, now we have the multiomic age.
Can you touch upon, how do you help the clinician and caregivers make sense of these new fragments of data that might come from very different sorts of lab tests?

Brad Bostic:
Rather than experimenting on humans with respect to what the tests are that you run or what the medications are that you try, if we can bring together the most comprehensive view of the individual, where it's truly their digital twin, the computer, the system, especially with all this cloud compute, can automatically run these different scenarios and zero in on where you have risk and get predictive well beyond what has been possible in the past.
If you look at historically what has happened with Alzheimer's, there hasn't been a treatment at all really. What that's resulted in is a lack of incentive to diagnose Alzheimer's, because what do we do now?
Well, there are some really promising Alzheimer's drugs emerging now, that are proving to do well in trials, but the key is you've got to understand the individual early enough to be predictive about the development of that cognitive decline, because these drugs can't reverse cognitive decline. They can dramatically slow it.
There isn't one single test that determines whether or not you have early onset Alzheimer's. So what we're doing is using machine learning models against tens of thousands of records of historically diagnosed Alzheimer's patients, to identify these signals that otherwise weren't noticed.
I can tell you, we've already identified some incredibly exciting breakthrough ways of seeing that earlier becoming predictive. Then you can get ahead of the curve.

Dr. Daniel Kraft:
Digital twin is certainly in the zeitgeist now. Is what you're saying, you can now parse that with your system to get that early indication at stage zero, that someone has a higher risk for cognitive decline and potentially put them on a statin for the brain, as an example?

Brad Bostic:
That's the goal. We're by no means to the finish line yet, but I'd say we're making progress. 15, 20 years ago, it would've been impossible to have this kind of horsepower in a cost-effective way.
When I think about the digital twin, it is funny. I was using that term a long, long time ago, but now as you said, it's in the zeitgeist. I think that's awesome.
We need to eliminate the model where the human is treated as a guinea pig. We should be beyond that at this point.
Why is it that when you order a teddy bear on Amazon, they treat it like it's life or death that they deliver it to you on time? They gather all this information about you and that experience, in order to support you better in the future as a customer.
But you go through a health care experience where it may be life or death, and you're treated like a number.
When we first started, we thought lab was a great way to focus. The feedback we'd get was, "Why would you focus on lab? It's only 3% of the health system's budget. It smells bad in the basement, where it's located."
I said, "Well, no. It's not about how much you spend on lab tests, it's about how effectively you leverage the gold that exists in that data."

Dr. Daniel Kraft:
Let's go back a bit to that UI for the clinician. How do they actually glean the insights and take action based on them?

Brad Bostic:
What we've found is, while we're data centric and we're all about, what are those risk signals that can drive an action that makes health care better, we also are of a belief that you need to have humans engaging in order to work with other humans to actually implement the change.
If you specifically look at where there are these diagnostic gaps, it's not just about, how do you inject some kind of an alert in the EHR for a doctor or a nurse?
It's also just as much about, how do you alert the care managers that work with the doctors and nurses and work to do the outreach to engage patients in the right focused way, so that the higher risk individuals get the attention that they need and that they've got somebody really looking out for them?
There are pharmacists who have fantastic knowledge. They can do a lot more than just counting pills. If you power them up with insights and tools to zero in on those high-risk individuals and then you have those pharmacists actually engage with the physician, "Hey, here's a change you can make to the medication regimen that is going to result in this person feeling better and ultimately having fewer ER visits," that human engagement is the end point of the power that happens behind the curtain.
We're way overboard on expecting an EHR to solve all these problems. You need some way to document the internal clinical process and diagnostic data points digitally and code for things, so you can bill for it. If you don't do that, you can't keep the lights on.
But this layer of intelligence that constructs this digital twin and then identifies the signal, getting that signal in the hands of the person who could then advise whoever the prescriber is, that to me is where you start driving positive change that gets adopted versus just saying, "Hey, here's yet another computer you've got to look at."

Dr. Daniel Kraft:
Absolutely. I mean the challenge, as we've talked about, is a lot of this lab and other data is quite siloed. Do you have other examples, oncology, cardiac care, surgery, where you've seen the digital twin model start to really make an impact? Where might that be in the next five, 10 years?

Brad Bostic:
Cardiac care is another key area where you can really do a lot of good. It just kind of blows your mind, where you'll have an individual who's had a lot of stents placed.
Recently, we saw that with somebody who they were on their 11th stent. It was identified that the medication that they were on, they had a genetic mutation that made it so the anticoagulant that they were taking didn't work. It was literally like just giving them expensive urine.
We know based on allele frequencies how common that is, but unfortunately it's not a standard of care because of the way reimbursement works.
In this case, it was determined that this alternative medication was something their body could actually benefit from. That change was made and then voila, problem solved.
I like dealing in the here and now realities. There are all kinds of really amazing things that we can do in the future that would save the lives of people like my mom, but some of these just day-to-day issues, like that example, we just fix those and you're going to move the needle on cost by billions instantly.

Dr. Daniel Kraft:
Yeah. I love that example. I mean, I'm always harping about pharmacogenomics. Where do we go between that gap between what's getting reimbursed and the huge opportunity, as you mentioned, just in picking the right anticoagulant?

Brad Bostic:
Incentive drives behavior. I think the move to value-based care through the direct contracting models and through what will become ACO REACH here effective January 1, where the primary care doctors, the health care providers actually can earn significantly more by keeping costs in check, but doing it by delivering more of the right care versus preventing access to care.
I had an opportunity a couple of years ago to work with a value-based care health plan that was inside of a very large health system. We had some of our best data scientists run our models at hc1 against the testing and prescribing that was happening.
I sat down across the table from the CFO of this health system. She looked at me and said, "If we do this, what's it going to do to my fee-for-service revenue?"
You're not going to get pharmacogenetics when that's the mindset. But when you move over to more of a value-based model, where it's every single physician can make 50% more money if they deliver better care and they realize that one of the ways to do it is to make sure you don't mess up the Coumadin or whichever medication isn't going to work for the person, they'll do a hundred dollars, which is going to soon be $80, which will soon be $50 pharmacogenetic test.
So it's the unsung hero of health care because the diagnostic power that's there, if we can harvest that at scale, we can truly make a difference.
Now we're all of a sudden having these leaps and bounds every three months, where it's like, look what we can do now.

Dr. Daniel Kraft:
Yeah. Outcomes are going to be the new incomes. Are you guys looking at, how do you integrate now these massive datasets, so it's synthesizable, understandable and actionable?

Brad Bostic:
This is a multivariate equation. You do have certain standard, straightforward indicators of things like blood pressure or bad cholesterol, or this specific diagnostic result relates to this specific need for a medication therapy.
But then what increasingly happens and where most of the cost goes in health care is where you end up with that individual who's on 8, 9, 10, 12 medications that have been prescribed by four or five different physicians, who have no good way to keep all of that in sync. They literally are making the situation worse and worse and worse.
The financial industry is a good point of reference. You can pull together all of your financial relationships in one dashboard and make it really easy for your financial advisor to see that.
It's more black and white. You're losing money. You're making money. You're treading water. Nobody has enough time in a seven-minute office visit to go understand all that.
So, I think if the system could pre-qualify who is highest risk and predictive of that risk based on this multivariate equation, surface it as an alert, have it be multisystem...
It's never going to be one EHR. The answer isn't, Epic runs the world, it just isn't. Then, these alerts have to be smart enough to get you to focus on the things that are really going to move the needle.

Dr. Daniel Kraft:
One thing I've been thinking about, I've developed a little platform in telemedicine, where you could take all the data about a patient and maybe 3D print the personalized polypill with their aspirin, statin, beta blocker, Synthroid, Vitamin D, things they might be taking every day or might even need to tweak.
How do you see narrowing that gap between data to insight to action? Because again, alerts go only so far.

Brad Bostic:
I love that, by the way, what you just described. There's no chip we can put in everybody's brains when they're born, to do right by themselves on eating perfectly and exercising all the time. That stuff's hard.
The future is really all about empowering individuals to have a better understanding sooner of what they can be doing to improve their health, but having more of an operating system that directs that.
We see this happening with CVS and all these organizations that have local presence. They're becoming more of the quarterback on, hey, where do you need to go to get care?
Then the specialists though engaging and looking at adherence more holistically, I think about involuntary non-adherence is where you take the pills as prescribed and they can't benefit you. Maybe even they hurt you. Fixing things like that can do a lot to bring us into the future.

Dr. Daniel Kraft:
Right. You essentially want to take... whether it's the Amazonification or Uberification, take something complex with lots of moving parts and have an easy button and have a user interface that can even be delightful and easy for a five-year-old to a 55-year-old to use.

Brad Bostic:
Right.

Dr. Daniel Kraft:
So just help me understand, how does hc1 work today? Who's your customer? What's your most common use case?

Brad Bostic:
So hc1 set out to bring together the provider and the prescribing entities in one virtuous cycle.
First of all, people weren't getting diagnosed effectively because they weren't getting tested with the best possible diagnostic tests.
Then, once they were getting diagnosed, they weren't getting the best medication therapies. In some cases, they were actually patients who needed life-saving medications that their physician didn't even know existed.
So, our customers really are on the provider side and on the pharma side. We work with laboratories that are independent commercial labs, doing a lot of specialized testing.
Then, we also work with health systems that need this insight into how to ensure the right patient's getting the right test, so that they're diagnosed effectively.
Once somebody's diagnosed, it empowers the health care provider with on-demand insight into approved medications, that can help potentially save the lives of patients who are diagnosed with rare disease.

Dr. Daniel Kraft:
So you're really an insights engine. While we talk a lot about precision medicine today, we're really still in an era of relative imprecision globally, for about the top 10 grossing on-patent drugs sold in the United States over the last few years.
They're only really effective for about one in four to one in 24 of the patients who takes them, whether that's an antidepressant or a statin. So, it seems like we have a little ways to go.

Brad Bostic:
Normally, when people say precision medicine, the person hearing that phrase thinks oncology. But in reality, 95% of people who are patients thankfully don't have cancer. But they have some other kind of a health condition, and diagnostic lab testing really spans all of it.
You can actually sit and talk about the future we all want to be part of, but sometimes we overlook that the present is so full of just base hits.
If you just get somebody on a medication they can metabolize, you can actually save them a hell of a lot of lifestyle and life expectancy and cost to the system.
One of the things I've learned over time is you need to have a vision, but then it needs to be broken down into steps.
We just talk about right patient, right test, right prescription. You have to figure out how to get a car on the road before you can figure out how to do low earth orbit.
Just like with the genome, where it costs a billion dollars to get to the first one, I think that this digital twin concept, it's a similar acceleration we're seeing.
We're really excited about where hc1 is positioned to help fulfill that promise, not only in the States but globally.

Dr. Daniel Kraft:
Are you able to start to crowdsource the insights from thousands of patients on your platform, so you can better inform the suggestions?

Brad Bostic:
It's using that composite view of lots of different patients to identify these signals that end up having a high probability of predicting certain outcomes as that early warning system.
But on the other side, the data they're fed needs to encompass the input from the experts who can curate and make the final call.

Dr. Daniel Kraft:
Right. You can upskill everybody across the care continuum. Speaking about the continuum, a lot of criticism to some degree of AI and let's say genomic datasets, is that they traditionally have come from Caucasian Europeans.
How do you see the issue about data equity feeding the insights that these AI engines then generate?

Brad Bostic:
In particular, as it relates to medications, there's a heavy need to make sure that we've got the broader diversity of datasets to guide these pathways for people, based on truly their unique qualities.
You've got the data challenge as well, where I do see the progressive leaders in our US health care system leaning into the fact that you can't use HIPAA and protecting people's privacy as a way to completely eliminate the possibility of actually delivering better health care.
All the things we're doing on these models, it's based on de-identified data. But when you step over to even the EU, where every country has its own unique privacy angle on data, we're going to have to find ways to strike that balance.
I'm all for privacy, but ultimately I'm all for a healthier planet and enabling people to live the best lives that they can live.
In the next era, what can we possibly do? The only thing is generate longer life expectancy, higher quality of life. This has to get solved.
If you look at medical devices, that's another area that we're in through one of our Health Cloud Capital portfolio companies. There is a massive surge of regulation around medical devices that is actually resulting in about 30% of the devices being taken off the market in certain parts of the world. It's kind of whack-a-mole. There's always something to swing at.

Dr. Daniel Kraft:
Oftentimes, the regulators are not necessarily recognizing our exponential aid. I've seen many examples of where patient data didn't flow. The patient may have died with their privacy intact, but they might rather be alive and thriving if their data was shared more readily.

Brad Bostic:
I kind of harken back to the Diamandis book Abundance. The subtitle is, The Future is Much Brighter Than You Think.
We've got this incredible opportunity that's like once in multiple generations, to advance the ball. It's because you've got access to medical information that's digital, unlike you've ever had.
You've got this access to compute. You have a collection of really intelligent, committed people working on these different areas of innovation.
You put all those things together and there could not be a better time to accelerate in health care.
If you just read headlines, you might be despondent. But if you actually are in the mix, in the front lines, seeing how things are changing for the better, I have no doubt that within our lifetimes we will have new predictive models that allow us to get ahead of the curve and cut way back on people developing Type II diabetes.

Dr. Daniel Kraft:
Yeah. It's a very exciting time. Certainly, Covid has been a bit of a catalyst to bring us into this health age. The trick is to translate that into real action.
It's also, the future is already here. It's just not evenly distributed. I've run a conference for 10 years, called Exponential Medicine. Now it's called NextMed Health, where the whole theme is, how do you bring people together to see what is now and what's near and what's next? Because often we don't need to reinvent the wheel.
Any other closing thoughts you'd want to share with clinicians out there, who are often struggling to deal with an overload of laboratory and other data?

Brad Bostic:
First of all, thank you to everybody who is in the care delivery environment. It's not an easy job. All this change, while exciting, also I think is accompanied by a fair amount of stress and inefficiency as we make the change.
I think it's a fool's errand to attempt to just perpetuate fee-for-service as a vehicle to generate economics as a individual who's delivering care.
I don't think you should consider value-based care and reducing care cost as something that can only be done by limiting access to care. That's the negative side of it, potentially.

Dr. Daniel Kraft:
Well, Brad Bostic, thank you for joining us in Healthy Conversations and to you and your team at hc1 for hopefully helping move that needle to continuously improve health care for all.