00:00:00:

00:00:14: Welcome to a new episode of the Strategy and Insider podcast.

00:00:18: In recent episodes, we have looked at how healthcare is being reshaped by innovation digital transformation New business models in shifting expectations across the ecosystem.

00:00:30: Today's guest has helped shape the healthcare technology agenda from both sides, actually.

00:00:35: From public and private.

00:00:37: He spent his career at the intersection of health care policy data interoperability And now AI always with a focus on how Public-private collaboration can improve outcomes for real people.

00:00:51: That said very warm welcome to Anish Chopra.

00:00:55: It is really great to have you today.

00:00:57: Thanks

00:00:57: for having me.

00:00:59: Anish is the chair of The Arcadia Institute, a research and policy hub that analyzes large-scale real world healthcare data with the aim to generate insights on market dynamics for health care leaders, policymakers.

00:01:15: Prior to that, he served as the first chief technology officer of United States under President Obama where he drove open data initiatives and pushed toward electronic health records.

00:01:29: Earlier in his career, he held public office himself a secretary.

00:01:38: also was a successful entrepreneur and executive in the private sector, co-founding data analytics and tech firm Hunch Analytics back in twenty thirteen.

00:01:48: And serving as president of healthcare analytics platform care journey which was acquired by his current employer Acadia Back in twenty twenty four.

00:01:58: initially he studied health policy at the Johns Hopkins University and holds a master's in public policy from Harvard which may explain why his work has always combined health care, technology and public purpose.

00:02:12: Anish is the author of Innovative State a book that argues our biggest public challenges can only be solved through better handshake between government and private sector.

00:02:25: with all this said it's really great honor to have you.

00:02:29: very warm welcome again.

00:02:30: What kind introduction?

00:02:31: It's an honour

00:02:33: Next to the remarkable roles you held in the public sector, I just mentioned.

00:02:37: You have actually been active in consulting as well at the start of your career.

00:02:42: after nearly a decade adrenowned DC based consulting and healthcare research firm called advisory board company or ever?

00:02:49: You became Virginia's secretary of technology.

00:02:53: was this an intentional kind of long planned move into public office Or did this come about by chance?

00:03:01: And what is it about public purpose specifically that has drawn you away from a possible longstanding Korean consulting or even the tech industry?

00:03:11: Well, thank you for the question.

00:03:13: It was purposeful.

00:03:15: I Was in college around The time then.

00:03:19: US President Bill Clinton was just taking office and had made healthcare A major pillar of his presidency.

00:03:28: As a young student, I wanted to get into public policy on the count of the scale of impact and thought that healthcare might be the appropriate area given its unique social goals but business model opportunities in the United States.

00:03:47: Our health care market is effectively a fifty-fifty public private partnership.

00:03:52: now some may argue uh... this isn't an area has failed.

00:03:56: I'm a number of metrics most notably around affordability and access.

00:04:00: on the other hand it's a vibrant and competitive market that is resulted in breakthrough innovations life sciences and care delivery models.

00:04:11: And i had grown to appreciate that public sector that curates, and incentivize is the right behavior from private sector.

00:04:20: that handoff between public and private sector would be a place where we can have the greatest amount of change.

00:04:26: so given my history.

00:04:28: I anchor in my career on trying to understand health care for all dimensions work at Morgan Stanley as an investment banking analyst in healthcare group right out of undergraduate, And then when I finished my graduate school, went to the advisory board company and in fact authored some of our initial research studies on how the internet would affect health care.

00:04:49: So this theme of technology public policy-the healthcare industry that triangle had been mine North Star?

00:04:58: And while i had been grateful for the opportunity to grow and thrive at The Advisory Board Company was able to achieve When Governor Tim Kaine took office, it was a beautiful opportunity to join him in his cabinet.

00:05:16: And so that transition was quite natural.

00:05:19: from the tenure of The Advantage Word Company and that theme of being in government out-of-government and then enabling public private collaboration—that's been consistent theme throughout my life.

00:05:31: I hope your listeners might view themselves as individuals who want an impact at scale but are motivated by entrepreneurial spirit of our global economy.

00:05:42: And what I really think stands out when looking at you is that rare combination of, well let's say optimism and pragmatism... That you bring to the topic health care transformation.

00:05:54: Obviously there was a huge transformation ongoing as we speak.

00:05:57: Can you share with us where does this optimism come from?

00:06:01: How do you sustain it in healthcare systems?

00:06:04: As sometimes know It can move very slow at times.

00:06:08: Yeah, I as a child was regaled with tales by my father of an colleague in his from undergraduate studies and India.

00:06:18: By the name of Sam Petrota.

00:06:20: he Was an electrical engineer From a village who came to United States like many Of his peers including My Father right out of college And joined The telecommunications industry at A moment that sector's change from analog to digital infrastructure.

00:06:39: he was quite successful in his corporate career but saw a window of opportunity as an entrepreneur very early life.

00:06:47: To build and design digital switches.

00:06:51: that led do i technology startup which is later acquired made him quite successful, at a very young age, almost like in one of those penny-a-year salary type roles guiding then Prime Minister Rajiv Gandhi on how to wire up the country.

00:07:11: Now back to optimism—the country with three hundred million people at that time only had three hundred thousand phone lines and if you followed traditional models of governance trying to raise public expenditures necessary was going to be an unbearable financial burden on the country.

00:07:33: So you had this terrible choice where only them rich few could get access to telecommunications or, You'd have to massively underwrite a big project and those did not feel very optimistic.

00:07:45: pathways for success.

00:07:47: Sam invented if he will are innovated by looking at his move from analog digital And recruited about three hundred engineers almost akin to the NASA moon mission and built an indigenous low-cost modern infrastructure capable of wiring, The entire country.

00:08:08: And giving every single person in the country access To a phone line.

00:08:13: that breakthrough mindset solving problems through innovation Public-private collaboration once you learn that story is possible.

00:08:23: It's almost infectious and impossible to ignore as a vehicle for the future.

00:08:28: So I carry this story of Sam Petrota with me in every aspect of life because there's always away.

00:08:36: We just may not have been able to find it yet.

00:08:38: we haven't looked around The quarter, but he hasn't connected enough dots.

00:08:41: And so part of the optimism is to believe Naively in some cases, but also you know with some confidence and history now behind me that there is a way.

00:08:51: And finding that ways what I think will make our health care system ultimately thrive despite its current challenges.

00:08:58: Big thanks for sharing the story and this should actually make us confident we can move out healthcare systems more broadly.

00:09:07: one of your major roles from outside.

00:09:09: looking at probably the most rewarding or thrilling jobs that you had.

00:09:14: You were first CTO of United States and served under Obama administration, helping him drive his technology agenda.

00:09:23: President Obama is often described as a confident conciliator calm on pressure highly analytical someone who favoured consents building over confrontation.

00:09:36: How did you experience that first hand in your role, and which experiences did take away from leadership style ever since?

00:09:46: Well thank-you for the question.

00:09:48: Your assessment is spot on And maybe to set context let me give this story perhaps into three parts When President Obama was running for office.

00:09:59: The perception he may be an underdog at a time of primary.

00:10:04: But by tapping the infrastructure of our emerging social media and technology-powered economy, he was able to give voice to the grassroots.

00:10:15: And connect them in coordinate that energy towards a successful election.

00:10:20: so prior to becoming president had naturally already seen the power potential of harnessing all this emerging capability is campaigned.

00:10:32: So it didn't require my involvement.

00:10:35: It just basically made the case that if this worked in a campaign, we should apply the same logic and

00:10:40: governing.".

00:10:41: That was sort of chapter one—his natural inclination.

00:10:45: Chapter two…and maybe this is more about his governing style...he created before I took it.

00:10:53: so on day one of his administration despite economic crisis he set a memorandum on open and collaborative government, where he named a role chief technology officer.

00:11:05: And in Washington circles, He had done something quite unprecedented.

00:11:10: He had named that role an assistant to the president.

00:11:13: there are very few people have carried that title.

00:11:15: it means you report directly to the President.

00:11:18: It creates the conditions.

00:11:20: were his advisors.

00:11:22: all of them carry this Assistant to The President role collaborate.

00:11:26: To understand this sort of second chapter, you don't take an item to the president unless there is internal division within his advisory group.

00:11:37: It's

00:11:38: almost

00:11:39: challenging for the President but his day is filled with thirty minute difficult questions that need to be answered.

00:11:46: and if it was easy we reached consensus he would not waste time with the issue.

00:11:52: You wouldn't go there, okay got it?

00:11:54: So given that these two aspects of the fact he understood the power technology and governs with his advisors who could collaborate hopefully eighty twenty rule ninety ten rule reach consensus without bothering him only really take is time.

00:12:10: when they were complex issues needed resolution here in this point to highlight the uniqueness president ran To the complexity.

00:12:22: so as a matter of just process we would write our arguments almost like illegal brief and president will be prepared.

00:12:31: And then when you gather the advisers and maybe half dozen times an issue i cared about kind had to get his buy in because of division within the advisors.

00:12:42: he was methodical what is amazing.

00:12:45: despite his perhaps not formal training it in the tech innovation space, he knew exactly the right questions to ask in the room, to get at the exact crux of the issue.

00:12:57: that would help him make the right decision.

00:12:59: Of course I supported every decision you made and obviously for the most part we could see from his history um... He leaned towards embracing technology data innovation when given the choice And that open mindset really did serve the country well i think became a precedent.

00:13:17: even future administrations comment about him being a thoughtful conciliator, it's driven by his desire to have all voices communicate to him and engage in collaborative dialogue so that the decision can be made.

00:13:34: That is informed at best of the group's ability.

00:13:37: And surely something to learn from also for own leadership style?

00:13:42: Thanks for sharing those very personal moments you shared with President Obama.

00:13:57: Now moving more a bit into healthcare and pharma industry specific topics.

00:14:02: Compared to other sectors, health care has not yet seen the same productivity gains from digitization despite really enormous investments in Health IT.

00:14:14: What do you think digitalisation has so often added complexity instead of reducing it?

00:14:22: And what has changed for that to finally shift in the end.

00:14:27: Let me make this as plain.

00:14:47: that is fragmented, competitive and largely incentivized through the fee for service system to generate more units of service.

00:15:04: Technology in the context of the fee-for-service world maximizes inflation per unit.

00:15:11: cost go up.

00:15:13: while there may be some operational savings revenue growth exceeds costs savings as the driver.

00:15:19: And what we found ourselves was a high tech act injection of capital into the technology sector, when the demand signal for its highest and best use was overwhelmingly informed by the Ifever Service Incentives.

00:15:35: But the Affordable Care Act shifted those incentives towards more of a coordinated model.

00:15:42: but the demands signal for that functionality has still played out almost fifteen years later as in our system.

00:15:53: so you have a dynamic where we injected the necessary infrastructure.

00:16:00: We didn't optimize last mile incentive problem and so he kind of has to put all these pieces together.

00:16:09: one need to have the right incentives for this is two.

00:16:13: you need to have a more connected system, so whether your independent primary care doctor or integrated health systems.

00:16:20: You can communicate and share information about patient that prevent the next avoidable healthcare visit or admission.

00:16:29: And you need to have an aggregate kind of performance tracking model so that we can shine light on where were falling short and when performing well.

00:16:37: That framework, opening up data for performance tracking interoperability and aligned incentives powered by these emerging technologies is only now coming into balance in a much more confident enter the next wave of reform and a more eighty-twenty style ROI than in the first wave.

00:17:00: And actually, what you're just referencing is also recurring theme when your work that healthcare data is still fragmented across providers specialists imaging centers pharmacies payers who name them right?

00:17:14: From your perspective, what are the most important barriers still standing for us to truly convene on a connected interoperable ecosystem where different data sets can flow?

00:17:26: Can find its way to interconnect and allow meaningful analysis on top of that.

00:17:32: So I'm going provide more nuanced answer because my goodness if there was magic wand we were one integrated nationalized healthcare system IT developed in a manner that facilitated this sort of data sharing and enabling insights to the parts of health care system.

00:17:51: That could take action, an advance if someone's health deteriorating.

00:17:55: so you know ideal world is single organizational unit driving towards his future.

00:18:00: I think thats not happening in United States for sure but lets assume its like straw man.

00:18:06: what enterprise might look do?

00:18:08: they were fully integrated aligned toward productivity goal.

00:18:13: We also acknowledge that in a fragmented system, what you need to do is sort of mandate if you will certain functionality.

00:18:21: So the idea would be you yourself may not be motivated to turn on the computer and look for insights on ways to help care for the folks who are not coming into.

00:18:30: see you that day.

00:18:31: That data must be made available To others that might be motivated.

00:18:36: And so In the other extreme you could see a world of mandates where information collected at every touch point is aggregatable and shareable to organizations that are incentivized.

00:18:49: To put that information towards highest investors, we've made some progress there.

00:18:53: but one of the challenges Is that the technical requirements on how to navigate privacy insecurity?

00:19:01: And existing let's just call it managed access to information versus fully open access to anyone anywhere.

00:19:08: that capability requires a bit more technical standardization, API standards the role of AI safety standards all range of gaps on the IT side.

00:19:20: That are the last mile gaps that need to be resolved.

00:19:24: so in world these two extreme examples we had one buyer driving what adoption use case within one enterprise relatively straightforward model.

00:19:33: if we have more organizations aligned that way great but we don't.

00:19:37: versus this fragmented with mandates on sharing, not so great in that we haven't quite closed the loop of what exactly that individual has to do.

00:19:48: To make that information available everyone without violating privacy and security?

00:19:52: We're left with a third optimistic path which is that we recruit early adopters to demonstrate what can be done and achieve some form evidence industry adoption or technical approach that allows me to pull my personal medical records from all the connected doctors in the country without the risk, then it'd be given to a nefarious cybersecurity hacker who might wish to do the same but for harm.

00:20:21: And that early adopter cohort has the ability of more rapidly testing the technical standards leading away and if that works than we can scale results through that regulatory apparatus.

00:20:36: And I think, that's the motion we are in over the last five years and it is paying

00:20:40: dividends.".

00:20:41: You've recently also argued how you get paid actually determines what tech you're using?

00:20:49: As well as seeing a fundamental shift increasingly from fee-for service towards more paying for better outcomes.

00:20:58: How does that change?

00:20:59: which technologies actually stick and make a difference?

00:21:02: And are there any tools, finally starting to earn their keep.

00:21:07: Well yeah I mean here i'm bit personally biased.

00:21:09: of course That's the bedrock in our work at Arcadia.

00:21:12: but just before I mean, sort of unabashedly celebrate our own firm's role.

00:21:19: I want to highlight a bit of the dynamic.

00:21:22: The typical physician might have say two thousand patients in their panel.

00:21:25: Let's just put that as an order of magnitude for primary care.

00:21:29: and let's say on a given day twenty Of those patients are coming to see the doctor.

00:21:33: i know it's probably less It's probably undercount relative To the reality of life.

00:21:37: but let's just for this purpose Technology with the purpose of making those twenty visits optimized, making sure the physician knows exactly what's happening to that patient.

00:21:50: What needs an effective visit.

00:22:15: So let's call that demand signal fee for service.

00:22:19: on the other hand, there are nineteen hundred and eighty patients that are not coming to see you today

00:22:26: or

00:22:27: they didn't feel sick enough to pick up the phone in call for an appointment.

00:22:32: but their health may be quietly deteriorating.

00:22:34: And if you could catch the right subset of those nineteen hundred twenty patients at a time before Reaches a true deterioration where they have to go the emergency room or need more acute intervention, then we might be able keep people healthier for longer especially those that suffer from chronic illnesses.

00:22:56: So technology look for signal on the nineteen and eighty.

00:23:00: there are not visiting you that particular day but maybe touching someone else in health care system may have been different specialist have ordered a new med or failed to pick up their existing medications, whatever the signal may be.

00:23:15: You would put a premium on collecting signal On behalf of the patients for whom you feel accountable and that that will Put much higher premium on pulling data in That's generated from outside your individual localized electronic health record And sift through all Of that information To find the signal as what some group of patients should call instead of them waiting for them to call you.

00:23:42: So, we can see the job be done in a fee-for-value model has lot more do with staying proactive and engaged with individuals while they personally might feel healthy or not sick enough want on your physician visit versus one where you are waiting that appointment making sure it is handled effectively and efficiently.

00:24:06: That is the world, that's a big divide in the technical demand signal from healthcare.

00:24:11: And I... The more we shift towards that longitudinal population-based signal of the better were going to be at all things you know and literature vast majority of heart attacks we could prevent if just managed blood pressure control effectively in the ambulatory setting.

00:24:30: We've known this for twenty plus years yet haven't made as much progress on that quality goal as we should have and more like throughout the health care economy.

00:24:42: And on a similar note, but also in a way very differently in Europe.

00:24:48: We're trying to keep those patients out of the system actually that are not needing to see a doctor and kind of clunking up.

00:24:56: this is them keeping others that I'm more desperate need to see your specialist give them extra waiting times.

00:25:03: it's not uncommon that you're waiting for two or three months to actually see if specialists because And the system is so filled up with also patients that are not in that desperate need and dissecting, finding out what's really in desperate needs now to see a specialist.

00:25:17: Whereas they're other ones who don't need to see doctor or giving them signals stay at home.

00:25:22: everything is okay you don't have to see it.

00:25:24: It's also way too drive down system usage but also driving costs eventually.

00:25:30: Yeah!

00:25:30: You're nailing your goal.

00:25:32: So if I ask this question we come into an age of AI where we can be more permissive on a long tail of data signals because were more capable of sifting through the noise.

00:25:45: so you know if u asked a very pragmatic question.

00:25:48: What does it look like today to kind what do call unclogged system?

00:25:53: that usually means phone calls.

00:25:56: two patients check in on them care management calls, And that, you know if he did the math on that.

00:26:02: The majority of those calls don't result in any material change and behavior.

00:26:05: it might be a check-in.

00:26:06: hey how are you?

00:26:07: I'm good okay thanks a lot keep me posted.

00:26:10: But that may not necessarily uncover the data signal.

00:26:13: That actually might indicate that while they might answer the question In a certain way their underlying body Vital signs might indicate something else so that in the age Of ai back to the san patroda notion that there were emerging technologies that allow us you know, kind of move beyond a traditional left right divide.

00:26:33: that would mean there could be more signals collectible.

00:26:38: There can be wearables to digital tools where you plug in certain feelings and symptoms and this wider range of catchers men if u will for signals you tolerate in the age of i. any signal on its own may seem benign but when connected with other data points might reveal that underlying deterioration.

00:27:03: and i mean we're also working ourselves a lot with big pharma clients around the globe on their strategy defining use cases preparing there organizations for.

00:27:16: For the error of AI and capability build ups in your life.

00:27:20: as you are member of us national advisory committee I'm really intrigued.

00:27:25: buy your expert opinion because if you would separate then signal from noise where do you see the most meaningful near-term value creation through a?

00:27:35: i actually in health care?

00:27:37: is it more care management visit walk flow support optimization?

00:27:43: is that more in the area of data standardization and usage decision support patient engagement.

00:27:49: or is it completely elsewhere, giving us a bit of your sense.

00:27:53: Again not asking for the magic wand but where do you expect that to happen?

00:27:58: I'm grateful for it.

00:27:59: in fact we wrote a transition memo right at the end of the Biden administration and early in the Trump administration.

00:28:05: so i think that document is public.

00:28:08: But the transition document on our agenda for AI had a section on healthcare And we have two primary recommendations That I'll share with you.

00:28:16: now Let me start an obvious statement Getting the infrastructure right, getting the rules of the road write unlocks.

00:28:27: The creativity of the country or health care sector if you were to limit this in that way globally too invent to create a design kind magic bullet and air quotes we may be avoiding is particular answer.

00:28:44: so I shift towards two main gaps exist today.

00:28:48: One main gap is that the organization, that ultimately holds the incentive to improve quality and lower cost.

00:28:58: That could be the individual primary care doctor?

00:29:00: Mm-hmm.

00:29:01: That can be a regional accountable care organization.

00:29:04: it might be an integrated healthcare plan.

00:29:08: It might actually be a new service that has created akin to a concierge digital medical practice.

00:29:15: we don't know.

00:29:16: The options are limitless but someone will raise his or her hand and say, I will be the steward of this individual's health journey.

00:29:25: How can that individual seamlessly successfully access or aggregate all of the health information that is available about me?

00:29:36: And then organize various AI tools a decision support that would facilitate next best clinical action.

00:29:44: so one ensuring that we have an interoperable system pointing towards organization.

00:29:51: that's accountable.

00:29:52: the second piece is we need to have a feedback loop.

00:29:57: For our system.

00:29:58: now we haven't national weather service it constantly pinging cloud formations and other scientific measures help us then pool all of the raw material enable market economy to sort out what prediction should be for whether.

00:30:17: what's the health performance analogy here.

00:30:21: What is the safest and secure this way by which we can pool health information in such a manner that would reveal best practices, Wow, patients that got four or five chronic conditions when managed under this particular treatment regimen have much better outcomes than those who don't.

00:30:44: Being able to find that signal being able to disseminate that signal on the positive end around what you should be doing and on a negative end where do we see harm in avoidable spend?

00:30:56: Those were two pieces of infrastructure.

00:31:00: be done.

00:31:01: And I guess the third aspect of this, which we might get to a little bit later is you know This does create the room for new market entrant A new category that perhaps may not exist today as part Of our current economic model For health care delivery and so i've been Constantly evaluating.

00:31:20: how do We create conditions That would allow for breakthrough ideas?

00:31:23: Is it we don't yet Know?

00:31:25: at some point they May come what feels like a very difficult and long and frustrating challenge of how to balance the pharmaceutical sector with health insurance sector, with provider's sector.

00:31:39: And we're going back-and forth in ping pong match as who has leverage?

00:31:43: Who doesn't?

00:31:44: Who got advantage?

00:31:45: I

00:31:56: mean any AI needs work on data.

00:32:00: more than often.

00:32:01: real world healthcare data is messy incomplete sometimes biased.

00:32:08: when you look at That one actually flows through a platform like Arcadia.

00:32:14: How good is the underlying data that we're reading on to and how much does quality ceiling limit what AI can realistically deliver nowadays?

00:32:24: And hopefully, there's improvement coming I assume.

00:32:27: but what's the underlying dataset like?

00:32:30: if you could share it with us?

00:32:31: please

00:32:31: Let me take step back before answering this question.

00:32:35: What Is The Job To Be Done that a firm like Arcadia is hired to do and how does it relate to the data quality question?

00:32:43: you've asked this critical.

00:32:46: The emergence of accountable care organizations in the US economy starts from the premise, did the insurance sector must share data with the accountable care entity so you get longitudinal think about almost like a healthcare utilization map.

00:33:05: This patient has gone here for this reason, had this done to her and so forth that data combined with the clinical data held by The provider side in their electronic health records.

00:33:19: So this payer provider bridge Where today?

00:33:24: In fee-for service.

00:33:25: the two keep those datasets as far away from each other is physically possible.

00:33:33: there's contractual access to combining the datasets.

00:33:39: Before we get into an individual data streams quality, the aggregate story is that the typical Arcadia customer has much more data signal legally authorized to aggregate.

00:33:52: so we naturally have pipes that pull in more data than a typical IT firm serving and individual doctor or an individual plan independently.

00:34:03: And it is that unique customer base, that has given Arcadia a bit of a leg up on being able to deploy AI models.

00:34:11: because running an AI model on the patient's longitudinal record.

00:34:15: Is a far more effective and efficient approach than try to simulate or to synthetically generate that longitudinal record by holding only a fragment.

00:34:26: So that's a critical statement not about the company is unique, you know engineering ability but rather than nature of the job we are hired to solve.

00:34:35: The organizations rights to aggregate more data?

00:34:38: That is the gift and that's why you see Arcadia playing much more effective role in coming up with use cases that deploy AI that are really designed to hit that productivity goal we spoke of at the beginning, that's missing in healthcare because our customers are aligned financially towards that kind of aggregate population performance improvement objective.

00:35:01: Having said that... At the individual data element level this is where I mentioned earlier the technical standards issue We made The Judgment In The United States That electronic health records data models shall remain proprietary by the fragmented healthcare data ecosystem.

00:35:21: Company A has an electronic health record system, they've got some secret sauce as to how they organize their data tables and facilitate actions that is meant to stay private.

00:35:31: however we have mandated that information be mapped To open source data model free of any intellectual property constraint And that mapping exercise is where we have a data quality gap.

00:35:47: In the extreme, every document, every piece of data should tag to an open source data model and made available downstream to organizations that are authorized access it.

00:36:00: however That step Of Data Quality doesn't Have yet A penalty for failure or An incentive To get a hundred percent right, so we continuously have to sort out what the data mapping looks like.

00:36:17: Ironically AI itself is playing a role.

00:36:21: if motivated to map We can now use AI To more effectively map unstructured data into a more formatted shareable dataset.

00:36:31: and in fact when In The last Biden administration I helped volunteer to help oncology clinics You know something like tumor stage is not structured in our electronic health records to map it from being an unstructured say piece of text.

00:36:48: To a more mapped data field on which you can run analytics.

00:36:51: and so yes we have any data quality problem in the narrow, which as we've got a map data.

00:36:58: but in aggregate if you have up opportunity work with organizations that can pool large amounts of data are gonna still see an AI advantage for good near term.

00:37:08: And both of us know health care is probably one the most, if not the most sensitive and personal matter in life.

00:37:16: In that regard a concept you have introduced I find especially interesting which is the idea of a Health Information Fiduciary Which is kind of trusted intermediary That helps patients use their data to get guidance that truly serves their interests.

00:37:33: What does that concept mean to you, and why might it become more important in a world where there is an eye-enabled healthcare to put forward?

00:37:45: When I mentioned earlier about Magic Bullet or products with great impact.

00:37:52: This category has been working towards for the past twenty years When you strip away all of the security and privacy restrictions in the American healthcare system, which I think carry globally.

00:38:07: There happens to be one common phenomenon The patient has the right To a copy Of their own data.

00:38:15: This is true i think In GDPR as well.

00:38:18: So this Is A global default setting.

00:38:21: The challenge Has been The lack of a business model For organizations to invest In helping those individuals use that data to more effectively and efficiently navigate the health care system.

00:38:37: That category is what I referred to as a Health Information Fiduciary, The good news in the interoperability world we find ourselves at earlier were at the tipping point from being twenty eighty to eighty-twenty In today's world.

00:38:53: you may have seen announcements from the Centers for Medicare Medicaid recently in that the large tech companies, OpenAI, Anthropic, Apple Google

00:39:02: etc.,

00:39:03: they've all basically made free of charge a service.

00:39:07: That would allow you to pool all your health records from any participating network and an age of AI more easily sift through information so it can have at fingertips human consumable understanding about what should do next.

00:39:26: stay healthy Now, this phenomenon is extraordinarily exciting but lacks a formal fiduciary concept.

00:39:39: It may be that the sponsor of this application is incentivized to get you to buy more prescription drugs and may default you towards recommendations making available rather than doing one that's in your best interest, which may end up costing the system less.

00:40:03: So you have a dynamic where if every single person had aggregated all of their records and run all of them or recent research through the decision support engine?

00:40:16: The great news for policymakers is at the healthcare system would be cheaper and better!

00:40:24: If you're a pharmaceutical company, if your insurance company and health system or provider network that would mean you get paid less.

00:40:33: So how much money are you going to invest?

00:40:36: To give people the tool That ultimately reduces your revenue?

00:40:40: It's just very challenging thing in our competitive private sector led market.

00:40:46: so I believe new category will emerge of organizations don't make on health care inflation.

00:40:54: They actually profit off of deflation, or they are incorporating these costs within a broader relationship that is profitable in other service

00:41:03: areas.".

00:41:04: That's what I'm hopeful for today.

00:41:07: and again if it's cheaper—if i make more money making the system better-I should take advantage of this movement towards consumerism And that's what I'm evangelizing wherever i go to try and recruit people To see that as a possible future.

00:41:28: So it is basically A new type of service entity, however we want to phrase It.

00:41:35: That has really rooted in trust Neutrality but also true interest for individuals Rather than For certain set Of product services solutions whatsoever Right?

00:41:47: Thats exactly right!

00:41:48: Don't know how easy your audience Is.

00:41:51: Professor Zach Kahane at Harvard, my guru.

00:41:55: He convened a bunch of stakeholders to benchmark existing off-the-shelf AI models frontier models against clinical case studies or scenarios akin to this fiduciary concept effectively asking should child be given very expensive growth hormone that may not have clinical efficacy but could kind drive revenue growth to one part of the ecosystem.

00:42:22: And conversely, what would be more conservative treatment for back pain?

00:42:26: When presenting a case against multiple frontier models you see a wide variation in default settings where AI because it's training history and post-training model whatever they've done that take into market has been effective is same case benchmarked against multiple models can have widely different recommendations And so with a fiduciary lens and some accountability, one might default towards models whose answers look more like evidence-based medicine wherever the answer may follow which might ultimately mean lower costs for given enterprise.

00:43:06: I think that tension is what we want to call out and do more research

00:43:11: on.

00:43:11: The billion dollar question remains whom?

00:43:14: What can you trust ultimately?

00:43:18: Thanks for sharing that and given your great set of experiences you have from decades at work in healthcare, I'm intrigued towards the end to ask if we would look ahead into five years.

00:43:31: One of the clearest signs you will be watching to tell whether healthcare is generally becoming more connected, more productive and even probably also more patient if not people-centered.

00:43:46: And not just merely digital.

00:43:48: on the surface

00:43:49: I think that number one metric to track Is this a share in American lives or global question?

00:43:57: Global lives that are in an accountable care relationship.

00:44:01: In five years and if we can flip from minority of the lives to majority, I think will naturally see the downstream effects greater AI use towards value based care more interoperability or consumer access and fiduciary services.

00:44:23: a lot really cool exciting promising things happen If we can find a way to move accountability from the minority, and if you understand this question what were really saying is.

00:44:38: The legacy fight of whether your in it tightly managed HMO product that limits your access to certain providers versus a wide open PPO product where he can go anywhere and see anyone that somehow this false divide of you can only have one of those two options.

00:44:58: Falls by the wayside.

00:45:00: if we can crack the code, then more and more of that PPO insurance product allows to see anyone but motivates you stay within a high quality accountable network?

00:45:11: That is unit I'm looking in five years to share with population.

00:45:18: that's accountable.

00:45:19: The more get number over the fifty percent line The better the demand signal will mean for overall health care technology.

00:45:27: The vendors that you know, the enterprises at service healthcare—the whole stack would default to value versus default fee-for-service as the motivator and that would just unleash this flywheel of innovation.

00:45:44: I believe ultimately make us proud —that we've retained a fifty-fifty mix but driven the value equation forward, making it more affordable.

00:45:58: Making it more accessible generally celebrating the innovative capacity of our healthcare economy just to do better and

00:46:05: better.".

00:46:06: And that's The Hopeful Spirit I carry—that we're going make it!

00:46:09: But that is one metric that is our North Star.

00:46:14: It's actually hard to find a better ending, so thanks Anish for sharing this.

00:46:19: it really was the true honor that you joined us here today.

00:46:23: Thank You For Having Me!

00:46:25: Hey and what a fascinating conversation we had bringing together policy technology healthcare delivery And also The bigger question of how We build systems That are Actually designed around people and outcomes rather than Stuff and actions that were sticking into system.

00:46:42: Thank you again for sharing your experience and more importantly, your optimism with us.

00:46:46: And to our listeners.

00:46:47: thank you for tuning into the Strategy & Insider podcast!

00:46:51: If you enjoyed this episode please don't forget to rate and follow the podcasts on the platform of your choice until next time and as always take care stay curious and Stay

00:47:07: Well!