In the rapidly evolving landscape of healthcare technology, different stakeholders are clashing over who can access what data and for what purpose. The potential impact on patient care looms large. Disruptions in data sharing could lead to fragmented care, where essential health information fails to reach the right hands at the right time. Healthcare executives must recognize the urgency of these issues and proactively engage in crafting solutions that ensure fair, secure, and comprehensive data exchange that benefits patients and clinicians alike.
On today’s episode of In Network's Designing for Health podcast, Nordic Chief Medical Officer Craig Joseph, MD, talks with Brendan Keeler, Interoperability and Data Liquidity Practice Lead at HTD Health. They discuss Brendan’s family background in medicine, and how he wound up choosing the path that he did, the current state of interoperability, and litigation around data disputes. They also talk about the impact those disputes are likely to have on patients, the role of the US government in healthcare standards and that compares with other countries around the world.
Listen here:
In Network's The Consulting Show podcast feature is available on all major podcasting platforms, including Apple Podcasts, Amazon Music, iHeart, Pandora, Spotify, and more. Search for 'In Network' and subscribe for updates on future episodes. Like what you hear? Leave a 5-star rating and write a review to help others find the podcast.
READ THE TRANSCRIPT
Show Notes:
[00:00] Intros
[01:51] Brendan’s background
[05:55] The challenges of interoperability
[09:48] E-prescribing vs prior authorization
[13:31] The government’s role in developing healthcare standards
[20:44] The current state of EHR vendor disputes
[32:13] How artificial intelligence can enhance data mapping and exchange
[39:00] The next thing that healthcare executives should be on the lookout for
[40:53] The future of prior authorization and interoperability.
[43:00] Brendan’s favorite well-designed things
[45:53] Outros
Transcript:
Dr. Craig Joseph: Brendan, welcome to the pod. How are you today?
Brendan Keeler: I'm doing great. It's a Friday. Thanks for having me.
Dr. Craig Joseph: And where do we find you today?
Brendan Keeler: I'm in Portland, Oregon. Somehow, in the craziness of conference fall and just, yeah, all the travel for work. But also, I have a new nephew, so a little bit of that. I’m going to see him in Denver.
Dr. Craig Joseph: So, you spend time in different parts of the country?
Brendan Keeler: Yeah, I bounce around. I was just in Nashville earlier this week. I was going to Denver, as mentioned. And my parents are from Philadelphia, so spread all out.
Dr. Craig Joseph: Okay. And it's probably a good idea for you to just stay, like, one step ahead of the law. That's what I'm hearing.
Brendan Keeler: Right. Yeah. They're after me for antitrust or information blocking.
Dr. Craig Joseph: Excellent. Well, we'll get into that because I don't even understand any of that stuff. So, let's talk about your past. My understanding is you are an interoperability guru, have you always wanted to be an interoperability guru? Was that your goal in kindergarten?
Brendan Keeler: I can’t claim that. That's a protected title. I haven't made it to interoperability sensei right now but working my way up to guru. But, no, I mean, I fell sort of ass backwards into it out of college. Went up to work for Epic just because I needed a job and thought it was, you know, the video game software company. It just sounded cool and yeah, I got to work up there in interoperability and integration. Because I protested against being a pure play engineer, and they said, okay, you can do something equally dorky connecting all the systems as part of implementations. Did that domestically, did that abroad in the Netherlands for a number of years. And then I've done that ever since and different roles and capacities across five or six companies now.
Dr. Craig Joseph: Tell us a little bit about what you're doing now.
Brendan Keeler: Sure, yeah so after the fun stint at Epic, I went to a company called Redox that helps applications connect to health systems to integrate, I then went to work for Jonathan Bush of Athenahealth fame, which was, a really fun ride at Zeus Health, working on health information exchange and then was head of product at Flexpa, which provided patient authorized access to claims data to other applications. I now work at HTD Health and HTD Health is a strategic consultancy and development agency focused. My particular practice is focused on interoperability, so I provide subject matter expertise, but executive partnership on any projects that we have related to health information exchange, all the interoperability topics we're talking about today, EHR integration or just general strategy. And as we have more and more and more connections between systems.
Dr. Craig Joseph: All right. So, you have convinced me that you know more about interoperability than I do. So let me ask you a question. So, you go right out of college with some sort of, computer science degree, I'm assuming, or something along those lines. And then, and you know nothing about healthcare, is that true at that time?
Brendan Keeler: My parents are both in healthcare, urologists but a long line of urologists. So, my father, my grandfather and my great grandfather were all urologists and then my mom is a registered nurse.
Dr. Craig Joseph: Oh, wow, so how did you end up not being a urologist?
Brendan Keeler: I worked in the lab, at CHOP, or not CHOP, one of the Philadelphia hospitals in high school and did not like sticking my hand and feeling the beating heart of what was it, a sheep they were dissecting or like, doing research on, you know, kind of, alien me to like, the pure play act of medicine, but deep appreciation for all those that are doing those things, and obviously for my parents.
Dr. Craig Joseph: So how do you learn about all the aspects of healthcare that you hadn't learned about? Is that just kind of an on-the-job thing that Epic offers, or is there a class or is it just kind of go learn as you go?
Brendan Keeler: Yeah, I mean, Epic does have like the three months of onboarding. But really, you know, part of their model is to bring smart or maybe not for me, not smart, but at least some smart people right out of college, throw them into the grinder and the ones that succeed and can work long hours, they can battle through tough implementations and pick up what's necessary, are the ones who make it pass a year or two years, and then, you know, it is trial by fire and learning through the industry. And so, I think about in a lot of ways as a second college or a second university, both in terms of learning. And also, I got to live in the Netherlands for four and a half years. I had a lot of fun too.
Dr. Craig Joseph: Yeah, that is not a horrible life. So, let me just ask you the question. I think most of us in healthcare wonder: why is this so difficult? Why is this so complicated? This is what I used to hear, when I worked for that software company. I can take my ATM card and go to any country and get money out, and it all seems to work, yet you can't send a list of diagnoses across the street.
Brendan Keeler: Yeah. I mean, interoperability is tough because it is a large superset of smaller workflows. And if you think about it in that light, like payment, the act of payment, which you're alluding to, building those rails is relatively easy as well. There's a forcing function and mutually aligning incentives to get it done and two, it's relatively simple transaction. People in the fintech and banking would dispute that, you know, the speed of payments, like fraud prevention, things like that. But at the end of the day, it's a debit and a credit. It's a movement of money. So, when you think about healthcare workflows, there's a number of different along different maturities that exist. And so, solving each of those at a ubiquitous national or international level is really hard and as you look at each of those, there's not always as clear an incentive to build the rails, to invest in those rails, rather than to rely on the existing ubiquitous method, which is phone call or even fax. Those are strong network effects because they work at scale. I can call the other hospital, I can fax the other hospital, I can fax the pharmacy. Whereas, you know, when you're building digital infrastructure to replace a workflow, when it only works for one out of ten or the hospital down the street, the efficacy isn't there and so you'll default to the thing that always works versus something that works some of the time. So that's one reason it's hard is that there's so many different workflows. There's e-prescribing, there's referrals. There's clinical data exchange with other providers. There's clinical data exchange with payers. There's interactions with payers for like click claims and eligibility. And there's hundreds of these different workflows that each are not the same transaction type. They're not the same data set. They are also different levels of solved or semi-solved or mature.
Dr. Craig Joseph: So, what's solved. What do you think is just like hey yeah, we've been there, done that. We've moved on.
Brendan Keeler: I don't want to call it solved, but ubiquitous. There's room for improvement as we've seen, with change healthcare in the spring but claims is relatively digitalized. You can send a digital claim. E-prescribing is relatively digitized. You can reliably send, you know, a prescription, typically over short scripts in a digital fashion to a pharmacy. And so, like while those are definitely you can look at them and zoom in on that, those parts of interoperability and say, we can make that better. And are people doing so in the grand scheme of things like providers would say, oh, well, I don't fax or a phone call, that stuff. Usually it's digitized, it's semi-solved, whereas, you know, things like referrals, things like prior authorization, things like claims, denials, like those are less solved workflows that are still heavily manual. They fall back to the ubiquitous solution, the solution of phone calls, a solution of fax and so, you can plot things along that spectrum and you can see we still have a lot of work to do, because there are still many problems that are not ubiquitous national infrastructure that can be relied on by providers, relied on by pharmacies, relied on by payers, operations teams.
Dr. Craig Joseph: Why do you think that we're much further ahead for like e-prescribing than we are for prior authorization? Is it simply that was the priority at the time, or was there government incentive to solve one problem over another, or is it just too complicated?
Brendan Keeler: No, it's a mixture of aligned incentives and you know, artful network building by private actors and right tailwinds from regulatory, action. So, if you look at e-prescribing, certainly it's in the interests of providers and pharmacies to digitized that transaction. Like it's not so long ago in the 2000 that you were faxing or printing out scripts and people were bringing them in, people can lose them. And then there's not advance notice to the pharmacy that, you know, they need to start prepping the medication. And so, there's a lot of ways there's a lot of, you know, pain, there's a chance for fraud, right? Like certainly you can go get the script and, you know, scribble something and like maybe bring that with you to the pharmacy to get some controlled meds.
Brendan Keeler: And so, you know, there was some incentive alignment there. There were industry players, namely RxHub and SureScripts that provided services to supplement that, namely the medication history pool from the pharmacies or from the PBMs to start. But now the pharmacies as well, and also the active prescriptions they were building into 2000s. And then the government said you know, this is really important for us from a zoom in on interoperability, standpoint, to incentivize, to make sure we get done at national level. And so they backed the standard NCPDP and basically SureScripts became the de facto, solution to such an extent they were sued by the FTC not so long ago, for a monopoly, because, yeah, they grew ubiquitous, and a ubiquitous national solution like that, oftentimes converges on a monopoly, which can have antitrust implications if they behave anti-competitively. So, while antitrust and monopolies are not necessarily a good thing, ubiquitous infrastructure is, so we can consider ourselves pretty lucky because when I went over the Netherlands in 2014, I helped stand up the first e-prescribing. You know, they were starting to incentivize e-prescribing. So, with Epic, we did the first like, you know, e-prescribing connectivity over there. And it was just nascent. It was just so far behind. I was blown away. But it's because they focused on different subsets of interoperability to incentivize, to push forward. And so, like you look over there, digital image exchange, diagnostic images between hospitals, not ubiquitous yet, but like almost ubiquitous such that I can for Brendan see prior radiology images and things like that. And we don't have that at all here. And so like, you can just zoom in different countries and see how they've focused on different problem sets, different, you know, digitalization of manual workflows by deciding and prioritizing in a lot of ways, the government can act as the product manager writ large for the nation, for national infrastructure. And their bets don't always work out just like a product managers bets don't always work out, but they have unique influence and unique power to help stitch together. You know, what would otherwise be free market competition, which on its own can converge to national infrastructure, but is much harder, because of less incentive alignment.
Dr. Craig Joseph: So, kind of continuing along those lines, the government doesn't own the rails, they don't own the phone lines, right. For us old people to kind of understand, but they do own or often can point to, standards. Right. So, are there standards we hear about. Well, I don't know if we all hear about, you hear about other things like maybe kind of walk us through what are some of the high-level, standards and where do they come from? I don't think they come from the government. They come from consortia of just kind of organizations that are out there, right. How does that all work? Explain everything about standards in the next two minutes, Brendan please.
Brendan Keeler: I will try and do as much as I can. Standards can come in a lot of ways. In some countries they are government created and mandated. The government goes into the little innovation lab, and they come up with a way of encoding data or sending data over the pipe or, you know, transport standards for content standards. And then they release it to the world and mandate it. The US is not built like that, at least definitely not in healthcare, very rarely in other industries as well. Instead, they look for industry-adopted standards where groups of even competitors come together to say these are the agreed upon formats that we would use to transfer meds, or to send a continuity of care document, or in banking. You know, there's the FDX standard, which is for open banking for retrieval of banking data, via patient authorization or consumer authorization. And so, they look for industry standards that have been proven with some adoption and like pilots or small atomic networks. And then they say, okay, let's move this up the maturity curve and then let's either require it in different ways or, you know, incentivize it or recommend it different ways to promote it. So that, you know, it's just a different methodology just comes from the way we're structured as a country. Right? The government, the federal government doesn't have a ton of power allowed in a lot of ways compared to other countries, to mandate to require, things because of the, you know, the federal state interactions as well as, you know, our history as a union of 50 states, and so we end up with a number of different interesting, elaborate levers to get standards promoted, particularly the EHR certification program through the former ONC, Office of National Coordinator. But now the, ASTP, they renamed themselves. They did a rebrand over the summer. They have a voluntary, health technology certification program, totally voluntary. Nobody has to do it. No reason to do it. Except that their partners, the centers for Medicare and Medicaid Services, CMS, requires it for anyone that wants to, either hospitals that want to submit claims for Medicare or, physicians involved in the MIPs program. And so that double lever of like, here's a voluntary certification program. You don't have to do it. But then you kind of do have to do it to get this higher reimbursement, is the way today those things are primarily incentivized and promoted. Historically there was meaningful use, which had more active, like, here's $50,000, go use the EHR type incentives. That doesn't really exist today. We instead have a more elaborate series of levers and contraptions to incentivize.
Dr. Craig Joseph: Are there multiple ways to kind of skin the cat, so to speak, are there you know, we hear about APIs. I don't know, and there are other letters, combinations you know, is there one health system that might say, well, you know, we think that prior authorizations should use this format, and another, or maybe it's not the hospital system, it's the EHR vendor. Is there controversy there? Or is it pretty much. Yeah, we acknowledge that that's the standard. We're just not there yet, but we're moving in that direction.
Brendan Keeler: You know, it varies like, by standard, by use case. You see standards battles. Back in the day, there was a battle, not really well publicized between, HL7, one of the major standards bodies, and NCPDP over medication transfer, and they sort of divided and conquered and did a handshake where they passed the mantle of any prescription over to NCPDP, but then medication exchange within the hospital between systems to HL7. You certainly see controversies today, should we use FIHR the new standard or a more legacy standard called CDA for the exchange of data between providers for when, you know, I step out and get hit by a bus and then, you know, have to go over to OHSU and they need to pull my information from the variety of places I've lived across the US. Should that be one standard or the other. I think my personal opinion is that those are silly fights that people are just playing for their own interests of where they’ve placed their bets, what they invested into, content standards by and large, have like, minimal, advantages, disadvantages, just tools. And at the end of day, if I can get the meds, allergies and problems ubiquitously to any provider in the country, that is the best barometer for success. And you can always update to newer standards and over time, once you've built of ubiquitous digital infrastructure. So, I'd rather people fight for that latter point and push towards that latter point than get caught up in FHIR, or HL7, or CDA, they're all sending. It's like trying to argue whether English or Spanish is a better language or something like that, right? Like they convey meaning in different formats with different grammar and words. At the end of the day, they're conveying thought. And so, if you think about that parallel of speech and transmission over language, which we're getting pretty deep with metaphor here. But, you know, you see that it's not really the point, like the point is not English or Spanish. The point is can we talk.
Dr. Craig Joseph: This takes me back to kind of, the early 80s, the best time I think we can all agree.
Brendan Keeler: I was born in the 80s.
Dr. Craig Joseph: Okay, let's not get into details here, but, Brendan, there was, listen, this is very important. There was VHS versus Betamax, right? Two different standards about how videotapes were going to work and I'm not sure one was better than the other, but there were two different standards. The difference, though, I think, between there that that situation and the current situation is that there were companies that had money at stake, and that's not the case.
Dr. Craig Joseph: Right? We're not seeing arguments between Oracle Health and or maybe we are between Oracle Health and Epic in the United States, or a big EHR vendor or other vendors in Europe or is it simply just while we've invested in this and we'd like this to work, but we don't really, because we don't want to do a lot of extra work. Have I opened up a can of worms, Brendan?
Brendan Keeler: Well, I mean, it's a fun and interesting time of year, I don't know if it's the tightening of funding across the digital health industry. I don't know if it's, sort of the cyclical nature of legislative and regulatory then judicial. But we're certainly seeing a push in judicial, in fights, in spats. And that could be Oracle versus Epic. They're not suing each other, but they are in this sort of turf war where, you know, the one of the leading marketing people put out a letter, sort of, I don’t know, if it's a technical term, bad mouthing Epic. And then Seema Verma, who is former head of CMS, but now works for Oracle, has been doing a lot of verbiage about Epic being a big blocker to interoperability. And Epic, who historically does not respond in kind, does not do marketing, has been doing some responses, has been pushing back. So that's a minor version of it, a more macro level. We see a dispute. We see, disputes, between apps and vendors over data access, overuse of the rails that are being built. And so, in particular, we have Particle Health, which is an on-ramp to these health information exchanges that serve for clinical data exchange between providers and other players. You know, as a dispute with Epic dating back to the spring, on whether rails that were built for treatment exchange, exchange of clinical data for treatment purpose of use so that when you're hit by a bus, Dr. Smith can see it, even though you know, you're 5000 miles away from where you live. Where that can be used for, by payers, by lawyers, by non-trading entities. And so, they had a dispute with that back and then that progressed over the summer. Then they actually sued Epic for antitrust, saying that Epic used its market power in ways that were anti-competitive. So that happened in September and is obviously going to continue for a long time. It's not the only thing we see, an application, real time medical systems, that provides analytics to skilled nursing facilities. They've sued PointClickCare, which is the largest skilled nursing facility here in the country that people just don't know about. They know about Epic and the inpatient ones. But like, this is the Epic of skilled nursing facilities, very, very large. They sued them for information blocking that, saying that PointClickCare was inappropriately preventing them with the permission of this skilled nursing facility, shared customers from accessing the health information needed to provide those analytics. And so big dramatic capital events that are happening in the industry now that, you know, the average provider doesn't really, see, but could drastically influence their day to day when all of a sudden their ability to see this outside data, the meds, allergies, problems of, you know, Brendan being hit by a bus goes away all of a sudden, like, that's one implication, or that they are radically empowered to choose the application of their choosing to do the sit alongside their EHR, in ways that deeply integrate beyond what they can do today. There's high upside and high downside to the capital, the drama that we're seeing. And I don't think a lot of providers know about it and maybe they shouldn't, but I think every hospital CEO should be intentionally and acutely watching these things and weighing in on these things because they are losing at the highest level. That's in fact, a doctor. But at the highest level right now, providers are in in an asymmetric relationship with payers where data is flowing freely to payers and they are getting no information, no care gaps, no claims data to build out their longitude picture and return. And so, if the CEO is at the C-suite of hospitals aren't paying attention and just pretend like it's not happening, then they're going to continue to have more and more of an asymmetric relationship where they are disadvantaged against payers rather than collaborating with payers, getting more symmetry and more, equitable data exchange.
Dr. Craig Joseph: Okay. Well that so that's interesting. I was going to say fights, arguments, lawsuits, debate.
Brendan Keeler: All of the above, all of the above.
Dr. Craig Joseph: It's not so much about how the information is exchanged. It's about the information itself. Like you had mentioned, there's one of many lawsuits that's happening because, or if I'm not sure if that's directly where the lawsuits coming from, but essentially, hey, you're using this information that you said was for treatment of the patient, but you're not treating any patients. And so, you shouldn't have. That's one of the assertions, you don't deserve this information because, you were not authorized to get it by the patient specifically and you're not treating the patient in general. Right? Did I simplify that correctly?
Brendan Keeler: Yeah. That's like that's the crux of the dispute and whether that's right or wrong that the infrastructure was built for treatment and that the, the way that the data was accessed hypothetically, was, you know, for treatment or claimed to be for treatment, but potentially was not. And like, I think people on the one side would say, well, it was for a very lightweight treatment. And then we received a HIPAA authorization, and we can do what we want with it. And then people on the other side would say, well, you stood up that treatment that single doc telehealth practice solely to funnel data to lawyers or to pharma or to payers so you can make money. And that doesn't feel right and that doesn't feel that feels asymmetric. And so, I don't see right or wrong. I just see an opportunity for us to think critically about serving payers, lawyers, individuals, these other use cases that are underserved, that the fact that they are underserved, unable to access data for, you know, in the case of payers care management, you just measure quality measures in the case of lawyers, mass tort or, you know, trials that require, health information, in the case of the patient controlling their own patient information and directing their care. These are all valid use cases, and they've chosen to use workarounds and asymmetric relationships, that are not that are opaque and actually, have a high propensity for abuse. They've chosen that because there's strong market demand for their use cases to be served. I think one of the positives of these fights is that has proven to be a catalyst for more movement on serving those use cases in the right way or in a way that is defined and agreed upon than we've seen in years past, than we’ve seen in my entire career, in just one year, in under one year. And so that’s the optimism. But I do think that we're not there yet. It could all fall apart. And the worst thing to happen would be for provider executives, people at the provider involved in this, to think that it's just going to go away because if they don't work towards the outcome of getting involved and helping push forward something that's symmetric and a shared value, where they get back care gaps and clinical data and those sorts of things, and they're going to be in an asymmetric relationship where preparers and other entities are pulling that data legally, but pulling that data and in such a way that they are completely disadvantaged as a business in a time when they have plenty of disadvantages as a business.
Dr. Craig Joseph: Sure, oh yeah. You kind of mentioned that you see a risk that everything might fall apart. How real is that? And what is that risk? Actually, let's just do some doomscrolling here. What's the worst-case scenario? Will it be that I'm sitting in my hospital or my doctor's office, and I can't pull up the list of allergies from another hospital? Is that the worst case scenario?
Brendan Keeler: When confronted with fights? Fights provoke fear. When we feel fear, one natural reaction is to retreat, is to shelter, to lean out, to move away. And so, in that case, it's a low. I put it as a low, a low probability. Some values where you look at the probabilities and then the expected outcome and calculate the value. And in that five percent chance times these groups of providers retreating and saying, you know what, we can only trust other Epic providers. We Oracle groups can only trust other Oracle groups. Or we're just getting away from overall exchange, is so tremendously disruptive that patients would die, like someone would miss a drug allergy interaction. Somebody would miss a prior condition that today they're just doing because they have access to that data. It's popping up and saying, woah woah, they’re on the diabetes meds or whatever? And so, you know, I think in those terms that it's extremely important, not because it's a high probability, because the outcome times the probability is so catastrophic, it’s such a step back. It's a decade back for us in terms of health.
Dr. Craig Joseph: That’s how it used to be, right? I will share information with certain groups because I have a relationship with those groups, and I will make it easy. Actually, we're not saying you refuse to share. I will make it easy to share. Right. Anyone can call up and fax a release and get that information, but it comes in a way that's not easily consumable. Did I get that right? Is that the direction that you think?
Brendan Keeler: Yeah, I just think if everyone chooses to retreat. They say we can't save the exchange information. We're risking HIPAA breach or risking patient privacy. The fear, the uncertainty and doubt of this future causes them to retreat away from these networks. I think what would occur is that all of a sudden, you're lacking clinical data as a physician at the point of care that can be used to save lives and improve outcomes. And so, yeah, we have to work. We have to work to make that not happen at any cost.
Dr. Craig Joseph: All right. No more doomscrolling. Let's talk about, doom in a different way. AI. So, let me paint a picture, Brendan. Right now, forever, it's been very difficult to exchange data because I need to make sure that we're talking about the same thing. I need a list of medications, and my amoxicillin has to be the same as your amoxicillin. Otherwise, we're not really sure what we're talking about and then even when we know what we're talking about, my box three over here needs to go into your box four and if it doesn't go exactly right, that's why we need engineers, you know, developers such as yourself to make that happen. Listen, I don't need any of you, because I'll just have the AI do it. Is that true or am I extending it a little further than it would naturally go. I think it's probably not fully true, but where can AI help us with some of this data mapping or making kind of unnecessarily even know if you're talking about the same amoxicillin?
Brendan Keeler: Yeah, 100%, I mean you're talking about data quality. And so, like if we don't regress ten years in the past, that is a huge problem currently. Right. With clinical data exchange in particular. So, we're focusing on in the subset of interoperability exchange of clinical data for, you know, emergent treatment for continuity of care or even, you know, for to payers and others. But those use cases we end to that. That's all clinical data exchange. It's really hard because how do you codify the medications and the SIG data? How do you codify the diagnoses? How do you ensure structured in such a way that it, like you said, slots into the right places and all these different EHRs? That's part of standards job. But it's also, you know, fully utilizing the standards, fully ensuring that using the codes, same code systems, all these different small things that take a lot of work. AI will help, both on the demand side, AI is really hungry for well-structured data. And so, and AI is hungry for data, which increases the need for data exchange, in general and also help as a tool for mapping, for parsing, for structuring, unstructured data. And so, there's a ton of work being done now by different startups as well as established companies to instead of just like you are doctor, you go and see click the external documents tab and see a list of hundreds of summary documents. Sorting through that and pulling out the things that are relevant to you as a pulmonologist or relevant to you as a dermatologist, which is a subset of all those things. And so, AI is extremely effective even today. There are so many cool tools that are doing just that. These visit prep, where they've taken the hundreds of this cacophony of information exchanged over these networks and summarized exactly what you need for your specialty or even for your patient. I think that's one of the coolest developments of AI. It's definitely one of the most popular outside of like prior auth or like the admin stuff, which doesn't tick the box as much for me, even though there's good money to be out there, I think.
Dr. Craig Joseph: So, let me just be clear. AI is not taking every interface engineer's job in the next year is what I'm hearing. But I can't trust you because you're one of those guys. I don't know. Why would I trust you with this question?
Brendan Keeler: I mean, I put this on the LinkedIn, I’m a LinkedIn poster now. I don't know what that means for me. I never imagined this happening. But I post online a lot. And one of the things I posted was the structuring of AI. So, you have the copilots, you have the agents and people are afraid of the agents. These autonomous, anything that's doing AI autonomously. People are gravitating a lot more towards the copilot to be the AI scribes. Right. They sit alongside the provider; they don't replace the provider. And so, that goes for developers to gravitating towards GitHub or Copilot and not these tools that would be that are submitting a bunch of PR. There's longer term upside to agents, to autonomous, you know, AI that does certain things, either entire roles or low-level work for a specific role. So, in the agent world, you can imagine, okay, the really hard mapping goes to the experienced person like Brendan. And then the low-level agents perform the junior work. Or you could say, the agents are so good at doing the QA task, but they're not good at doing the novel new file task. And so, agents have longer term upside in almost any industry category subsegment. But people are afraid of them because they remove roles. They should not be. The beauty of humanity is that with every new tool we get rid of some meaningless work, and we move to something that's higher value, right? We're not in the Industrial Revolution. Tons of roles were destroyed. Like, you know, we don't walk back. We don't even look back on. And so, I think the short term is there'll be tons of copilot that augment our ability to do the provider. Ambient Scribe's being a very popular and powerful one. I'm returning the position back to just talking to having the conversation, I think is just as a narrative, really impactful, especially as they do more than just transcribe and actually do all the order, placing all the diagnoses, all the keystrokes that you'd have to do as a physician. Wow. That's awesome. And short term, really powerful. But there's something really cool about unlocking us to work on the most important, highest leverage uses of our talent. And I think that's the promise of the agents, which I don't love, and where we gravitated on this, like, spy verbiage for these, you know, it's kind of nefarious, but, those agents, if they can allow us to work at top of license that are really cool, interesting problems, then we get to evolve to the next set of things that we're all supposed to work on and to make, to be doing 10x what we were doing before.
Dr. Craig Joseph: Yeah. All right. That's fair. That's less doomy to me. I thank you; I appreciate that.
Brendan Keeler: I'm pretty optimistic. So, like, you're going to hear me that flow through a lot of things. Well, I try to be realistic and paint the downside. Like, there will be jobs lost, right? But the net benefit is that it opens up entirely new categories that we didn't even imagine possible.
Dr. Craig Joseph: Sure, you had referenced earlier, that healthcare leaders, CEOs of healthcare systems really need to be on top of what's going on. What would you suggest they do if we have, a CEO or COO listening? What should they be doing? Should they be talking to their CIO, talking to other, executives?
Brendan Keeler: Yeah, I mean, there's a couple. I would talk to other executives. I would find your interop lead or whoever's in charge of this for your health system and say, what the hell is going on here? Do I have a status brief on this? How are we getting to a place where we are not strategically disadvantaged with our payer partners because of asymmetric relationships? That’s one question. Do we have HIPAA, breach risk because of this? And how are we getting to a place where we can move to use this as a catalyst to do better exchange with payers, that we get back care gaps in clinical data and the things we need to be successful for value-based care programs or whatever initiatives we have, or at least how do we stanch the bleeding of you know, losing leverage? You can go to your EHR vendor and push on them for the strategy there, in these turbulent times. I think that's a no brainer. And then, of course, they can get involved in care quality and common well and tough. And all these different initiatives across the industry. But I think start with the more tactical of that. Is your organization on top of it, and then is your EHR vendor on top of it, or like the must do's if you haven't done them already?
Dr. Craig Joseph: Okay, so, if I'm a CIO, I should be making sure to have these conversations with the boss.
Brendan Keeler: Yeah. Be ready with the report, the threat and the strategy.
Dr. Craig Joseph: Okay. Fair enough. One of my last questions will be about the future. So, if you could kind of magically take us 10 years from now or five to 10 years from now, like, assuming we get through the arguments we're having now, and some of those lawsuits get resolved one way or the other.
Dr. Craig Joseph: Where do you think, five to 10 years are we going to be night and day different, or it's going to be like, well, we'll probably have it figured out, but by then, med reconciliation is going to be complicated.
Brendan Keeler: Here's my projection is that health tech has been mostly bipartisan, in different ways. It's push forward in towards similar outcomes of patient access, focusing on subsets of interoperability to break down and make digital. I do think clinical data, these workflows will be nearly ubiquitous. And we will look back and say, why was this so weird and turbulent? In a way. But like this actually occurs across industries when building these sorts of networks. Bank of America, when they founded the predecessor to visa, man, it was on the ropes. And the leader of it, was just a legend sort of held it together and got it through that time. And then now we think visa and credit cards are ubiquitous in a way that we don't even question. But we should. It was at a time where Americans were not the Americans of today. They did not take on debt. They did not trust this flimsy little plastic card. And so, there were so many challenges back then that are akin to the turbulence of right now that we can get through and we will get through and our children will be like, why? Why was this such a problem? And so, like, I look forward to my daughter just being, like, not worrying about this. God, I want her to work in health tech, but, like, I get whatever she does, is just that her health is where it needs to be for the use cases that she needs.
Dr. Craig Joseph: Yeah, I know you make a great point about, Visa and Mastercard. Like, that's, most of us, we don't think about that anymore. So. Yeah. Awesome. Well, this has been great talking with you. We always like to end the conversation, talking about design and, anything in your life that is so well designed it brings you joy. So, what is it? I'm assuming there's one thing. Maybe two. Any thoughts?
Brendan Keeler: I have two, and one's really kind of basic. To the mocha masters, a coffee machine. I have dabbled. I'm not like, a coffee snob, but I'm just trying to figure out what is the right balance of what's good, what's fast and done like the caved glass thing. I forgot the name of that one I've done, I've lived in the Netherlands, where they have, like, entirely different sets of coffee makers. And the Milk Master is just, like, super-fast, you know, you prep it and it before, and you just click it and it makes an awesome, you know, full tank of it that stays hot for a long time. So, I really like that. I have a full list, my daughter is 20 months old, so I've gone through and now assessed and ranked all the different baby gear out there and so. I have a great travel stroller. Maybe I can send you a link or something that. The really expensive Uppa Baby type, one that people recommended. So, yeah, if anyone's like new parents looking at those things, I got a full set of recommendations and rankings. The SNOO is a wonderful device. And you know, the SNOO is this thing. Have you heard of this before?
Dr. Craig Joseph: I have not, we didn't have those things. We didn't have strollers when my children were small, we carried them by their hair. We didn't know any better, Brendan. We didn't know any better.
Brendan Keeler: The SNOO is a tech-enabled bassinet, which sounds super bougie and whatever, but the baby falls asleep like it's just crazy. She would cry, put her in the SNOO and it's suddenly she's out, and it has all these features and stuff. But one killer feature is that it's FDA approved, and it saves, it prevents SIDS and things like that. So, you're not only getting the nebulous benefits of, like, my baby sleeps like, oh, maybe it will sleep regular bassinet, but also preventing the off chance of SIDS and things like that. And so, I think it's awesome. It is expensive. You can rent one. Anyway, those are mine. I promised one, but I gave you like five.
Dr. Craig Joseph: Well, thank you so much, Brendan. This was great, we had just the right amount of you know, kind of, bad news, but, I think, overall, we're going in the right direction and really helps all of us. Helps patients, helps clinicians and hopefully helps the bottom line of the healthcare systems that we all depend on to keep us healthy. Thank you, sir. We appreciate it and good luck that daughter of yours and all the high-tech and low-tech gear that you have, to support her.
Brendan Keeler: Thanks so much. Thanks for having me on.