Epic or open platforms? The choice facing European hospitals (Nils Hellrung)

[00:00:00] Dear listeners, welcome to Faces of Digital Health, a podcast about digital health and how healthcare systems around the world adopt technology with me, Tjasa Zajc. Most interoperability debates focus on exchange formats.

[00:00:16] In this discussion with Dr. Nils Hellrung, medical informatician and CEO of Vita Group, Nils focuses on the need for shaping hospital IT with governance and financing.

[00:00:31] Linking proprietary systems through standards, he says, will always lose the meaning of data, which is why he backs OpenEHR as a vendor-neutral clinical data core with FHIR standard used for exchange. Vita Group is providing the data platform for Catalonia and was just selected as a partner with a large healthcare group in France.

[00:00:56] The discussion that you are about to hear touches on the different market approaches to interoperability. We discussed challenges of markets such as Germany, where each federal state has its own authority to regulate and implement as they see fit.

[00:01:13] And we also talked about what does it mean to have a AI native data integration and why a 95% accuracy of AI is not good enough in healthcare where we're dealing with human lives. We also touched upon whether Europe will end up with monolithic US EHRs or open platforms. This episode is supported by Vita Group.

[00:01:43] Enjoy the show. And if you haven't yet, check out our newsletter, which you can find at fodh.substack.com. That's fodh.substack.com. Faces of Digital Health is present on LinkedIn and YouTube. So make sure to check out those platforms as well. And if you will enjoy the show, do leave a rating or a review wherever you listen to your podcast. Now let's dive in the discussion with Nils.

[00:02:23] Nils, hi, and thank you so much for joining me on Faces of Digital Health to discuss interoperability, open air, the European challenges and more.

[00:02:36] You are a medical informatician by background and actually did a PhD on architecture for transinstitutional health information systems, a big word, which basically refers to how information systems are structured when care crosses institutional boundaries. And this is a very relevant topic in the current European landscape. You also co-authored a textbook on healthcare information system.

[00:03:05] So I'm very curious to know, given that this textbook was published in 2011, how do you see the last 15 years of development? And especially when you write textbooks, theory can differ from practice.

[00:03:23] So how does the theory that you wrote about differ from practice you experience as a leader of a healthcare IT infrastructure company? Yeah, first of all, Tiasa, thank you for having me on your podcast. One good thing about healthcare, if you want, is that it moves very slowly. So a lot of things that you work on are still valid long time after that.

[00:03:48] For me, it was always interesting how, why do the processes look like they do? And also how do the IT infrastructures look like? And why are they looking like this? At this time, we were looking into hospitals much more than cross-institutional. I could always see, if I only look at the IT infrastructure or the software landscape within a hospital, I could always tell you, is this a university hospital? Is that a private-led hospital?

[00:04:16] Is that a public hospital? Why could I see that? Give you one example. So we were looking at one university hospital. They had three radiology institutes and they had three different radiology information systems, so PACS systems. And the reason for that was there were three professors and they didn't want to have the other professors look into their study material.

[00:04:45] So there was no trust between these professors and that's why they bought three different PACS systems. Which does not make any sense from a governance standpoint, but if you look behind it, there's always a history. If you look into private-led hospitals, they're normally much leaner, organized, much more straightforward. And of course, you can transfer that also to the, what we called at the time, trans-institutional case.

[00:05:11] So why is it so hard to bring together hospital information systems with information systems for physicians? It always depends on the reason. In Germany, for example, we have a very strict border between the hospital system and how it is financed and the ambulatory system. So the physicians and general practitioners and how they are financed.

[00:05:38] And since the billing administration is always the starting point, unfortunately, not the therapy, but the billing and administration, that's why the systems are also completely different. So a lot of things are still true today, I would say.

[00:05:53] Especially the people and these funny reasons why you can't, why collaboration, for example, doesn't happen, as you mentioned, where people are just very focused on their own careers, which is quite fascinating. Because when you design a technology, it's sometimes easy to think or be puzzled when it's not adopted because you don't understand these human elements.

[00:06:23] I thought it was interesting when you said that healthcare is moving slowly. Do you still have that feeling in the last few years with the rise of AI? How is that impacting the pace of change and especially the pace of expectations? Yeah, I would say there are, of course, fields where it's already in the everyday work, for example, again, in radiology, analyzing pictures and stuff like that.

[00:06:52] But if we look now at the change of normal processes, I think it's still going very slow. And I would even say that the gap between what is possible and what is in reality gets even wider. And still, we are talking about humans. And most of the time, we are talking about humans that are not there because they're interested in tech stuff. But they are there for, first of all, their patients.

[00:07:22] And most of the time, a patient don't want to be a patient. They want to be healthy. We are talking about physician nurses and they didn't choose their job for AI or for tech, most of them at least, but to help people. So I still think, of course, the possibilities are getting more. But I still think it's quite slow, especially at the core of it. You know what I mean?

[00:07:50] So if you have some specific things where it goes very fast, but the core of it is very slow to adapt. Yeah, yeah, yeah, yeah. I guess the basic infrastructure that is, if it's a legacy infrastructure, you can just very easily switch.

[00:08:09] Not just because it's already embedded in everything that the hospitals do, but also because if you move to a different system, if you wanted to move to a different system, it's a clinical risk. Because you need to get everyone used to something new.

[00:08:29] And I'm sure that any U.S. audience that might be listening is always fascinated by how slowly we do things in Europe because things are moving much faster in the U.S. Since we're talking about... Sorry, just to go on to that point, that was not always the case. So when I started with medical informatics, Europe was far ahead. And that was a political initiative, actually, and a political will to go faster here.

[00:08:57] And as I said, it's not only a technical problem. You always have to see the combination of humans, processes, and then infrastructure and the whole governance. But there's no reason why we shouldn't be at the same pace or even faster in Europe. We will talk about this, I'm sure, later. Yeah, yeah. But it's not natural that the Americans must be faster here. Absolutely.

[00:09:25] Yeah, there's a whole range of reasons why they moved the way they did. And that's exactly what Europe is also trying to do with the European health data space. To just connect systems across borders, enable exchange of data. It's when we look at the numbers of when EHDS ideas started, they coincide, interestingly, with your own career as well. Because the idea for EHDS started shaping in April 10.

[00:09:54] And I think that's also when you started working for Vita Group. It took seven years till 2025 for EHDS to be formally adopted. So how do you see those seven years? How did the company that you work for now develop in that period?

[00:10:16] Did the fact that you knew that EHDS is coming, did it have any impact in terms of how you look at your own development and strategy? There are coincidentally some things that now very well fit into our strategy. But when we started or when I sold my own company to Vita Group and I started with Vita Group and had the chance to think a little bit bigger.

[00:10:41] My question was, okay, what is the biggest barrier to implement digital solutions into the health systems from my perspective? There are a lot of barriers, but what could I do? And for me, it was always data availability. So whenever I tried to implement a process between hospitals or whatever, it was always like, okay, how do we get the data? And then let's say 40% of the time of the projects, we were building interfaces that didn't work.

[00:11:10] And another 40% we were writing data protection concepts. And maybe 20% of the resources actually went into the problem. And when I started with Vita Group, I said, okay, this is the thing where I want to be active to care for data availability. And what it means in our case is that we extract data from existing systems and we standardize that data. That's very important.

[00:11:34] We standardize it in a way that we didn't invent, but we normalize it to international open standards. And then we make it available in real time for other applications, also for secondary use. So that's what we are doing since seven years. And of course, the HDS, it's more or less the same story. And I'm very happy that now they agreed on it. And it also slowly becomes relevant in the market.

[00:11:59] It's not relevant yet for that, I would say, the day-to-day usage or most of the hospitals are not at the border to another EU country. So the exchange of patients is not that relevant everywhere. But what is relevant is to think about that future data architecture in your hospital or your health regions. And that goes together quite well.

[00:12:24] If we just stop for a little bit with the kind of data normalization and curation that you mentioned, can you explain that a little bit? So is it actually like, how does that look like in practice? For example, you get a new client. Do you as a company actually take over that kind of data cleaning, if I call it like that? And is it to just various standards?

[00:12:53] Or which types of data curation do you do? Is it just open air or is it also other standards? Yeah, we are basically using two standards, FIRE and open air. We are using FIRE as a communication standard and also to have a lower entry barrier, to have some use cases quite fast. Because FIRE is just a communication standard that a lot of, let's say, at least modern applications are using.

[00:13:21] In open air, we are using to build up a health record that is very stable over time. So it's the clinic core of it. It has some advantages because it has a maximum data availability approach. That means everything you can think about the blood pressure can be recorded and stored in open air. And it's everywhere the same in Germany, like in Spain and France. It's an international thing.

[00:13:48] When you change the versions of the FIRE standards from 3 to 4 to 5 to 6, you always have to change that data model also on the semantic level. So both of these standards have their intent. And we are using both for depending on the use case. And yes, we are doing the integration work, which is a lot of work with our customers or with our customers together.

[00:14:16] It always depends who our customers are. We have like small customers like a hospital with 150 beds. They don't have the data team. There we have to do it. And the other side of the road is Catalonia. Which has their own big data and data modeling team. And there we are not involved in data mapping or anything. They are doing this by themselves.

[00:14:45] We've got tons of questions related to what we discussed. And what I thought was interesting is when you mentioned that basically the FIRE profiles change with the versions. And in one of the previous episodes, when I talked to Herko Koemans from GDHP and from the Dutch Ministry of Health, and he, as someone super interested in interoperability, alluded to the fact that even if you use the same standards,

[00:15:15] they can differ. Like the FIRE profiles in one country can differ from FIRE profiles in another country. So to me, as someone who's not very deep in the whole interoperability field anymore, I just can't stop wondering if there are variations, what does that actually mean for connectivity? So I guess my question to you is, how do you as an expert in data standards and interoperability

[00:15:44] see this big vision of interconnected systems? Is it just an ideal that we can strive towards to or is it an achievable goal? On the European level, for example. I would say it's achievable for a core data set, but not for the vast majority of the data.

[00:16:07] I don't think that by connecting old systems or non-standardized systems that are not standardized at the core, so that started with proprietary data, connecting them over whatever kind of standard will lead to true interoperability. So just because you have, let's say, a phone line between two systems or two humans doesn't mean that I can understand Japanese at one second.

[00:16:35] I can understand some words that we agree on, but I cannot understand Japanese. And as long as the core data models of the systems are different, we will always have loss of semantic meanings when we exchange data, always. That's why we believe in open air, and that's why I think Finland, Catalonia,

[00:17:06] are great examples of how you can do that. Also, Karolinska Institute, for example, on a big scale of a university hospital, how you can achieve that. A lot of people argue, but you cannot expect that every vendor will now change their data model to open air. It's true, I don't expect that, but I expect that there will be enough vendors

[00:17:31] so that we have an ecosystem that is big enough to cover all aspects of medicine. Since you also mentioned different examples in different countries, one of the things that I'm very curious to hear about is how do you, as a German-based company, work on different markets? And the reason that's interesting is all the cultural and human elements that we discussed before.

[00:18:00] So, what's your experience with the work in Catalonia? How do you approach the cultural differences that need to be addressed when you are trying to implement in a foreign market? So, first of all, why does it work in Catalonia? And again, this has got nothing to do with technology. In the 70s or 80s, Catalonia had very bad health systems and they did a big reform.

[00:18:30] And one of the reforms was that there's only one system, one health system. In Germany, we still have around, I think, 110 or 120 public insurance companies. Basically, we have 120 different payer systems. Of course, they use the similar processes, but it's very complicated. And also, one advantage of this is that no matter if you are a GP or working in a hospital,

[00:18:58] you're part of the same system and you don't have any benefit of not working together. So, the physicians, they want to work together around a patient. And that's the precondition that things are getting adopted quite well here in Catalonia. I say here because I live in Barcelona for one year, which is very nice. Oh. That's the reason. And how do we approach it culturally? Yeah. Outside of Germany, I would say outside of the Dach region, outside of Germany, Austria and Switzerland,

[00:19:27] we are working together with partners very heavily. Here in Catalonia, we are working together with IBM Spain. We just entered the French market. We can talk about this in a minute, maybe. With a group of companies that they know the culture, they know the market. For us, we are not a very huge company. For us, it's impossible to enter every market with an office, with a team in place.

[00:19:54] But we enable those partners to go into these markets. So, how many languages do you speak? How's your Spanish or Catalan going so far? That's a very bad question. So, I'm the only one in the family that didn't really succeed with speaking Spanish. I didn't have it at school. And I had 20 Spanish lessons. And if you start again learning the ABC,

[00:20:22] you're getting very small confidence. And I didn't really manage so far. Let's see what happens in the future. But, yeah, working from here, most of the time, home office, speaking English all day. Yeah. Yeah, yeah. You moved. So, that's already a good start on a daily basis to just getting to learn the language. If we stay with Catalonia a little bit longer,

[00:20:51] So, how long has that project been in place? How is it progressing? How realistic is it, in your view, for this consolidation of data that's happening on the Catalonia region level? How realistic is it that would also expand across Spain? So, Spain has several regions. So, it has a similar problem as Europe,

[00:21:20] where each region does their own thing, which then, again, brings new challenges when we want to achieve interoperability and cross-region exchange of data. Yeah, I think we started with the project in 22, if I'm not wrong. And after 18 months, the project was live, which tells a lot about the ambition and also the ability here to move forward quite quickly.

[00:21:49] We have data migrated backwards from 17 years. We have over 30 million individual electronic health records in the systems with billions of data sets. And there are now around 30 applications running every day in all hospitals of Catalonia and exchanging data with the core platform. So, I think that's a very successful project.

[00:22:17] The next phase is now to build on top of that what they call here the open health platform. That is, you can imagine, like an app store on top of our platform so that the hospitals can choose their own processes and applications that work together. So, how are the chances that this is being repeated? We see a lot of tenders at the moment but also coming up in Europe that actually have quite clearly

[00:22:46] Catalonia as a good example. In Spain, yes, it's very regional and they're not choosing all the same pattern. But they have a mechanism to make sure that on a, let's say, at some level they can exchange the data. And here the EHDS approach also plays a big role. In what sense does it play a big role? As you said, they have the same challenge. They have the same challenge to exchange data that has not been recorded in the same way

[00:23:15] but they need to find ways how to be interoperable at a certain level. That goes, but that comes with some disadvantages as we discussed before. In my opinion, it would have been better to implement the same strategy everywhere. But that's just a federal approach. I'm German. I can talk about this a lot. There's no chance in Germany to really centrally decide to do one thing

[00:23:43] and roll it out across all 16 parts of Germany. It's the same in Spain here, unfortunately. Yeah, there is the similarity between two markets. Since you mentioned Germany, just looking from the outside, if I look back 10 years, I remember a friend of mine also from the digital health space said in Germany, our national sport isn't football, it's data protection.

[00:24:13] And that kind of made it, made everyone very frustrated when you opened up a discussion about data exchange and data interoperability. But in the last few years, it seems that Germany does have large ambitions and is moving fast. Two years ago already, I think, the insurance data has been opened for secondary use, which is very advanced. So how do you see the development

[00:24:41] and just the advancement of interoperability in Germany? Is my kind of overview or view too optimistic or are things going in what direction are they going according to your observation? As a German, I always tend to be very pessimistic and very, let's say, critical of their own country, but probably you're right. There's very, very good approaches to some things. What really is, for me, it doesn't make any sense

[00:25:10] that what I said before, like, we have 16 federal states, but we have 17 federal data protection offices. Why? Because Bavaria has two. And so that's, you have to repeat the same process 16 times to roll something out in Germany if you do it on a federal level. And that, for me, does not make any sense at all.

[00:25:38] So that's just one of the innovation barriers, I would say. From a, let's say, cultural standpoint on the market, everybody agrees that we should be much faster and implement much more. But it takes time. And in Germany, that's why this comes a little bit bottom up. So a lot of hospitals are quite innovative because they have good CEOs with, let's say, strategic thinking. They want to implement their data infrastructure

[00:26:08] so that they are independent of the vendors and use the data for purposes like AI or automatization of processes. But it's not like we have this central, I would say, central capability of rolling out something. Yes, we have the electronic patient racket, but this project is 25 years old. And given that, I think it's a little bit sad where it stands. Okay. Yeah. It's when it comes to these types of things,

[00:26:37] it's all we sometimes need to remind ourselves that I guess in healthcare you need to look at 10 or 20 years kind of segments to assess what kind of progress has been made. So because on a daily basis, it can feel like things are just at a standstill. Speaking of different markets, you, so Vita Group recently announced a deployment that's going to happen in France. Can you tell me a little bit more about that? What does that mean for you?

[00:27:07] How much can you use the experience from Spain to also roll out in France? Yeah, I think it's a very interesting project because it's another approach again. It's not a public approach, but it's from companies. It's called the Alliance. It's an alliance of La Post-Santé and Cipage, which are two public companies in the healthcare market and also the University Hospital of Lyon. And they have their own IT company.

[00:27:37] And they're actually competitors in some ways, but they cover 70% of the French market. And so they decided, okay, if we want to be successful in the future, especially with innovations like AI, but also, I would say, in competition to big US vendors, we need to do something else. And we cannot go on competing, let's say, on a little small things, but we need to agree

[00:28:07] on the data layer, and then we can go on competing on the product level. And I'm very happy that they chose us, the health intelligence platform, as the basis for their data-centric approach for the French market. And here, it is like they are doing the marketing, and we are enabling them from, let's say, in the back how to sell, implement, and also operate the platform. How do you see the future development of healthcare and the

[00:28:36] use of healthcare data in Europe? It seems that a lot of discussions and also deployments that are happening now are really focused on making sure that this basic layer, the data layer, is done well, that the data is in a quality format. But as Rachel Dunscombe alluded to in her latest opinion piece, we made a lot of progress,

[00:29:05] or we are making a lot of progress in the interoperability sense of data, but now we need to start working also on the better use of focus on the basic layer could help Europe when it comes to AI development. I think it helps very much with implementing innovations at all and rolling it out. That's one of the typical problems I think

[00:29:35] IT companies have in Europe as a whole. We are still cultural difference because of the legislative difference because of the regulatory difference. But if we are looking at the data level, it doesn't make any difference if you are in Spain or in Belgium or in or in

[00:30:05] Germany. They all have the same problems. That's why here's a huge chance from my perspective and the EHDS will be a huge accelerator for the discussion here. So I think she's right. We are doing a lot of progress here. And yes, second part of the question, better use of data definitely. But then I'm coming back to what we discussed before. That also means that of course the processes and the culture

[00:30:34] within the hospitals is changing. If you suddenly have access to a huge amount of data and a huge amount of historic data and very detailed data about your patient, something else, then if you just have a fax like today or maybe just very short information about your patient. That's why we need different education as well here. And we also need

[00:31:04] AI. And I think one of the biggest chances is here for AI to summarize all that data and to present it in a meaningful way to the user because it's impossible for physicians or for nurses or for patients to read all the information that they have access now and also to understand all this is impossible. So we need AI as a translator between data and the user. Speaking of AI, how is AI impacting you?

[00:31:34] Hospitals, healthcare systems in general are under the demand of creating new roles to develop a workforce to reskill people to have new profiles that understand data management data governance AI development potentially do internal development based on the data especially since we are very protective of a third party can't easily work on

[00:32:08] manager so if you look at your company how is AI and agentic AI impacting your product development is it in any way helpful is it also helping you do all that kind of data structuring and cleaning we talked about at the beginning of this discussion yeah it's impact first if you started our internal review it's impacting our company a lot as I would say every software company is

[00:33:09] it's data which is an AI native product to help with the data integration from different systems and we are starting very enthusiastic and even I could now prompt into Claude or ChatGBT okay here this is the source data this is open air do a mapping and it looks very good we'll come up with the result convincing yeah very convincing

[00:33:39] but if you're an expert and you say did you also think of that you prompt this into ChatGBT said no but you are of course right I didn't think of that I will add this blah blah blah this convincing is very dangerous because we are talking about people's life I mean even 95% is not good enough you have to be very accurate about what you helps us to focus on the most complicated things and to make sure but at the same

[00:34:09] time it's like a pilot of a plane you cannot go into routine and say yeah just accept you're right you have to go in there and it still takes time but you mentioned that the new product that you have is AI

[00:34:39] native what does that actually mean what do you mean by that yeah it's a good question we can distinct between AI enabled and AI native and AI enabled means we just said we have a problem we throw it into it comes up with the result the AI native is more like it's not input output it's more like your colleague working beside you giving you suggestions telling you where to

[00:35:09] watch out and supporting your work process instead of just telling you the result and so we use a combination of AI based processes in the software but also deterministic because you cannot run for example if you then implement the process of really exchanging data from one system to the other and you run this on AI you will have problems now you need to be really

[00:35:39] sure that it's always the same that there are not hallucinations in it it's you can also always count on it then it's correct so it's a combination and AI native means it's more your colleague instead of your the one who really does the job from A to Z yeah and you internally build the connections and the guardrails to just make sure that there's a combination of the deterministic systems with the AI okay yes

[00:36:09] okay great because that's one of one of the bottlenecks today especially in our job setting up an open air or a fire platform on a green field it's quite easy to be honest and if you put some forms on top of the processes but you don't have any connections to the existing world that's quite easy the complexity comes if you want to connect it into the existing world and that's what we did over the last four

[00:36:39] five years a lot that's why we have a lot of experience how it works and correct and also for the hard edges cases that's where things really get interesting beyond just the proof of concept or the demo that you

[00:37:24] all products are built in a way that we've discussed some products are built on top of existing solutions and with agentic AI coming to the forefront in terms of governments there's the need to also figure out how do these different AI agents or different AI systems work together so I wonder how much potential new challenges are we creating

[00:37:54] in terms of interoperability because of different AI systems that are being used to which extent is there a problem and to which extent is it potentially not if the data layer is sound the most afraid I am of is that people think or maybe not think but they stop thinking because AI is doing everything and I don't want to go into hospital

[00:38:23] where the doctors are using AI and they are not trained on a real job and to think through the whole process by themselves anymore because I know how it feels for my everyday work I

[00:38:54] still need to think through and that's the thing that worries me the most that people get lazy because they think that AI is solving anything on the data layer there must be guard rates how these agents are talking to each other I think on the data layer once we get the data models right the exchange of data should not be based on AI at all it's just they are just pipelines ETL

[00:39:24] exchange pipelines that are deterministic so it's not good to throw AI at everything where you don't need it but for let's say sorting the data in the first place I think AI can help a lot yeah I guess there's still AI he basically talks how

[00:39:53] for the specialists I think yeah for the specialists AI isn't as useful because it doesn't really tell them something that they don't know but for younger doctors for junior doctors it can be super helpful and that also makes me personally worried to your point of really not being sure if the doctor is actually basing the decisions on experience or based on what AI

[00:40:23] has suggested so I guess yeah we still have tons of open questions around the impact that we're going to see in the end but speaking of speed and speaking of the potential and ambitions how do you see the European healthcare IT ecosystem in 2030 or in 2036 so 10 years from now in your view and understanding and knowledge

[00:40:53] of the field what do you think is the best case scenario and what's maybe a more realistic expectation of how things could turn out I I start with the worst case scenario again I'm German I'm sorry so the worst case scenario for me would be that and I see some risk at the moment despite the let's say geopolitical situation that also in the field of healthcare big US companies

[00:41:23] are taking over everything we've seen now the tender of charity with epic epic is getting gaining ground in Europe and what I consider ourselves to be part of a counter movement or another movement which from my perspective should be the European way based on open standards based on open source to really leverage the as we discussed before

[00:41:53] the innovativeness of the different European regions but bringing it that would be the ideal case so in a realistic scenario I would say we are stuck in between these two options so there will be one mainstream that are still thinking that monolith systems are the best way to go and others and I

[00:42:30] be able to influence what you are doing on a strategic level in your health region or in your hospital it sounds like an interesting question or balance between to which extent do we want to move together and look for the greater group and to which extent do you just want to make sure that you are taking care of your own needs and your own

[00:43:00] kind of benefit because monoliths in that sense do work great for that hospital that uses it it's just a very limited impact if you want to reach more broadly I'm not sure to be honest if they work that good maybe there are also differences of course between the product but the customers that we have today we have because the hospitals were just fed up with not being able to influence anything

[00:43:30] on the roadmap being completely depending always paying a lot of money for interfaces with not much data which doesn't have any good quality in the data we would not be here if the monolith would really be working in any case I can understand from some standpoint that you're saying as let's say medium sized hospitals okay capacity I don't have the expertise and I'm a little bit afraid of all this data work and AI that's too much for

[00:44:00] me that's why I buy something where I can say okay if it doesn't work I will call them what I need implementation is not done done yet or not finished yet Nils thank you so much for taking the time for this

[00:44:30] discussion today it's going to be interesting to see how things develop we are in the beginning of September and I believe you're also going to the open air conference which is happening in September what are you going to be talking about there lessons learned from AI based data integration so what we just talked about from the naive idea we can just throw everything into the

[00:45:00] AI then the problem is solved to where we in France thank you you've been listening to faces of digital health a proud member of the health podcast network

[00:45:29] if you enjoyed the show do leave a rating or review wherever you get your podcast subscribe to the show or follow us on LinkedIn additionally Finally, check out our newsletter. You can find it at fodh.substack.com. That's fodh.substack.com. Stay tuned.