From Administrative Burden to Digital Advantage Streamlining Processes Under EU MDR & IVDR

Hello and welcome everyone. My name is Shawn Smith. Welcome to today's webinar from administrative burden to digital advantage. Streamlining processes under Umdr. A couple of housekeeping announcements before I turn it over to Basil to make the introductions, the recording and slides will be made available and emailed to everybody tomorrow. That's registered. I would like to ask that you use the Q. And a button for any questions that you would like to submit to the panelists speakers today? and that you use the chat for comments, or to request a certificate of attendance. If you would like a certificate of attendance, please put your name as it should appear on the certificate. And and very importantly, your email address, and it will be delivered within 48 h or so. We have to make them manually

with the housekeeping announcements out of the way. I'm going to go ahead and turn it over to Basil to make the introductions.

Thank you, Sean, for the introduction and lovely to see everyone online. I'm seeing a huge number of attendees joining us to my person. Vasil Akra is my name. I'm CEO, and owner of Akra team, a consultancy company and happy to co-host this event with 2 subject matter experts to help us actually understand the future of digitalization. And where can we go? And what can we do in the future. or what we can do today? Even so, to make sure that we are getting leaner and better, avoiding all the burdens that we are seeing on a regular basis. So with me, James Dewar, who is working for Scarlett and the CEO and co-founder of this notified body in the European Union. A very specialized, notified body with a focus on software as a medical device and AI solutions.

But you can learn more about Scarlett if you go to their website, visit their website, learn more about their services and abilities, and if you are still searching for a notified body, it's your time. Pick that name and make sure that you write an email. Next to him we have Michael Kanya. Michael is also CEO and co-owner of Medevo, a company which is focusing on digital solutions. So he faced the pain when he was in industry. And he told me one day, you know, after doing this experience of manual methods, I decided we can do it better, and I think he's doing a big difference with his new company with the co-founder, that they have to make sure that there is actually more digitalization and systematic approach in the way how things are delivered. So if you are interested in learning more about Medevo, visit their website or write Michael, he will be able to help you understand how they can

make a difference in your holistic process. So today, we're going to be talking about actually the current administrative burden and how the future digital solution might help you streamlining your processes. And I'm happy to have all of these speakers with us today, just to allow you to get like some insight on how things can be different, and before going into the discussion with all of them, I prepared the short presentation, just to give us, like some some taste about what is happening. And I would like to start with this just to make sure that everyone is reminded about the situation and getting some insight on how things can be done today, and what are the challenges, and how we can

improve them to make sure that we are generating a difference? Can you see my slides? So what we want to focus on? Like I said myself, I'm currently as a consultant, active with my team helping manufacturers, and I would like to share with you actually from our experience, but also with my previous experience, where I have been working for notified bodies, what are the challenges. And what are the bullet point which you need to consider to make sure that actually, the process getting leaner? Why, I'm saying this, because if you work for for industry. If you work for notified bodies, you recognize how much of these things are actually relying on excel sheets, word documents, and a lot of copy paste of information which lead to a lot of trouble. What do I mean with this trouble? Looking back to the numbers looking to the numbers

which are published on a regular basis by the study monitoring availability of medical devices. In the Union market we see key important numbers. The 1st numbers are, how many application were already filed in, and you see that they are a huge number of application, 27,000 application. And if you look to the number of issued certificate, as of today, you will see nearly a number of approximately 9,000 certificates. So one cert was achieved after 8 years of publication of the regulation. Why is it like that? Is it because the notified bodies or the manufacturer are too slow? No, it's actually connected to a lot of additional problems related in a lot of times with administration which requires a lot of documentation to be prepared twice triple and 4 times with a lot of similar information.

Also, it's connected to the complexity of the process, for all sides. Not just a notified body has complexity, but also the manufacturer has complexity in their documentation. I remember myself one day when I was working for a notified body, asking one single question, and all what I wanted to see, yes or a no, and the result was getting 1,300 pages for review. And this is where, actually, if you go into those pages, you will be actually stuck in reviewing and reviewing and reviewing. And then, instead of asking one question, you will get to 100 questions. All of these numbers are relevant for us. Why, we are looking to them. We are looking through them to see actually how slow the process is actually running today. And if we go deeper to the numbers. We see actually that the number is actually much

critical and connected to the process itself, because, looking to the application, when a manufacturer would apply till the signature, when a signature going to happen, you will see numbers which are going between actually one to 6 months, or even longer, where this initial application process and contractual agreement signing is taking too long. Why is it like that lot of excel sheets, a lot of information to be provided a lot of codes, a lot of clarification at the very beginning, which create a lot of complexity on top to this, even after the application. When you start the conformity assessment process. There is a lot of round of questions back and forth. Why? Because the documentation are inconsistent. They are not connected to each other.

The message is not provided in a clear way. Why is it like this? Because people are working are doing their best to deliver the documentation the right way. But if they are working actually in bigger team or involving internal external, the message might be broken, and this is also connected to the nature of this message, but also notified bodies on the other side, when they are working and doing the reviews, the reviews might lead to actually the same question coming from different members of the team, because there is no centralized system which is actually making sure you get the question just one time. So you get it from the clinical team. You get it from the technical team. What I'm trying to do today is not actually to finger point on either manufacturers or notified bodies. I'm trying to show you why we are sitting together today here to discuss how digitalization can be actually helping us for a better future. Today we see a lot of limitation, and there is a need

for a change. And this need for a change is actually reflected. But why the Commission has has doing actually, the past months is, is actually looking to quantitative data before we were always complaining. But no one had the visibility about what we are complaining about. Today. We have numbers in front of us and those numbers. To my point of view they are shocking. Why, they are shocking, because it's too lengthy, and it can't be that we spend actually up to 8 months or longer on an application where all of these back and forth communication are connected to the fact that we don't have a clear structure and the lean structure. And this is because we are still using methodologies where, if we look to our kids. They will love about us. We will say, How could you do this? This is ridiculous. You should do actually something in a different way. And this is where we can increase efficiency and make sure that the system is running much better.

Now in reality, why all of these things are happening, because in reality notified bodies like manufacturers, like authorities, everyone has obligation, and the obligation are coming from the regulation for medical devices or for diabetes, and they expect notified bodies to develop a process. And if you look to the system itself, the processes are connected to how a notified body would do it. And we're going to learn more about how James, as a software expert AI expert, did it for his own notified body. But whatever James did for his own notified body. Scarlett and different notified body might do it in a different way, because the law will give us actually a frame and say, this is what you have to create.

and will not let us know to do it in a digital way or a manual way. They say you can offer service according to this frame how you do it. It's left to you. And this is where we need to think about it. To say, are we doing the right work at this moment, or can we do it in a better way for the future? Now, I'm going to give you some example to show you how complex it is. The examples are connected to some steps in the holistic conformity assessment application process where a manufacturer would apply for a notified body today. If you want to start the application process. The 1st thing that you have to do, you go into a Nando web page.

and you will find who are the notified bodies, and you will find their designation, scope and notification scope. So you will see which codes are acceptable. You have to do the search manually, one by one to see actually, which are the available notified bodies. There is now online on the website of the Commission some filters which you can apply, but they are actually not enabling you, if you have a portfolio of product, to get everything filtered. if you go on a manufacturer side, the manufacturer will go to the Mdcg. Mdr. Codes to find out 1st of all, which code would fit for me as well, and based on this, what is applicable for the manufacturer, and what is the notified body allowed to do? The application will run, depending on how the notified body build the process. Some notified body said, you apply via a form plus an excel sheet plus an additional form, and some of them they have a pre-application application.

and all of them they are still dealing with a lot of actually manual solution. Now some of them might be more advanced, some others less advanced. But what does it mean for us if we look to this, it means actually complexity, because you need to do multiple steps to achieve one simple target applying, saying, I want to get my products e-marked one day under the European law. And this is actually critical and complex, and lead to mistakes and delays which are unnecessary if we go further even in the holistic process we see that notified bodies are required to whatever they're going to be working on. They have to write reports.

I can tell you my story. When I was at a notified body. When we were doing reports under the directive, we were expected to have a summary of what has been done, and majority of the summary was connected to what the manufacturer delivered to us, so we had to copy the intended purpose. We had to copy some of the information of the manufacturer to write in a report and say what we have done, and why, we concluded is acceptable, and the report were short, they were actually 15 to 30 pages. Max, today with the medical device regulation or Ivdr. The reports are getting longer and longer. Why? Because the authorities are pushing this in that direction, and if you look to the reports, if you read the report of a notified body for quality management or for technical documentation assessment, you will find actually, the majority of this report is coming from your documentation.

whereas if they are copying this information, they will have also mistakes, because they copy information sometimes, and you might change it because they ask a question and they might copy it. And they don't change it anymore. And this might lead to actually incorrectness of the report, and then might lead to also questions afterwards if they got scrutinized by the authorities. But they want to do the work in a timely manner, but, as you know, no single manufacturer will deliver a technical documentation, and will get the green light immediately from a notified body, there is a back and forth. And this leads to a lot of actually

challenges in getting a report consistent from the beginning to the end, because they are still using traditional methods. And they are not actually using a connected digital system. So what I'm referring to is perhaps to have something digital where the manufacturer has a documentation notified body. Do the review, and then automatically, you get a report which is actually consistent. And referring to the right section of the manufacturer. And this is what we see from all sides. Manufacturers are facing challenges. Notified bodies are facing challenges. Also, surveillance authority are facing challenges to tell the story again and again in a way which is consistent. Going further to the last example that I want to do. You know, I don't know how many of you went through that process, you know some notified bodies. They have also, in the final decision step. It's required by law to have a final review

depends on the notified body. Some of them, they repeat review again for some of the activities, some they limit their review to whether the resources have done the job in the right way, and some others. They have a joint system from the beginning to the end. So there are a lot of methodology. I don't want to describe all of them. What I want to highlight and flag. A lot of these discussion, which happens at the end are connected to a story which is not clear enough. A notified body receive an application. They review the documentation aligned with that application, but during that review recognize they have to adopt an intended purpose, or whatever, so they expect that from the manufacturer and the manufacturer would do it, and at the end of the review process the final certification decision maker. They're going to be reviewing what has been applied for, what has been assessed and what is actually what is recommended for certification. And this point at the end.

for some notified body takes between 3 to 6 months, because there is an internal discussion. Did you assess actually all the devices. Did you miss one of them? Did you lost one of the intended purpose in your holistic assessment process? Because you don't have a system which is making sure what is applied for is always part of the assessment, and this leads to a lot of actually unnecessary challenges at the end, and the notified body might say at the end, oh, we're going to give you an approval. But for this intended purpose we have to redo the word, or for that device. We didn't consider it from the beginning, because it wasn't clear that we should confirm on that, or even much more critical. We don't believe that that device is actually a medical device, because at the beginning we thought it is. But during the assessment it seems to be different. And this actually

burden in the holistic conformity assessment process is connected to the manual steps that are done which can be improved if we look to a different solution for the future. I don't want to say that everyone is facing those challenges. I don't want to say that every manufacturer might be actually in such situation, or every notified body. I just want to flag. There is a need for an improvement. And we saw a lot of example which went wrong. And this is why we need to look to a system for the future which might be better, much more systematic. centralized, enabling actually from a digital solution both manufacture and notified bodies, to have a digital communication instead of an excel sheet communication. And what I say, digital. I'm not referring to Pdf, so I'm talking about a systematic approach which allows both sides to have an ability to review report in an automated manner without having all the burdens of manual methodologies. Now, I just

gave you a taste of what is actually the current situation, and before moving into the Q. And a. And I see James and Michael very interested in having a nice discussion on this. Let's let's go to a short poll and raise a question to all of you, and please contribute to this when when doing that, James, listening to me and listening to all what I said, what would you actually take as a key takeaway? Is it actually the reality? Or I'm telling something which is my own interpretation.

No, I definitely think it's the reality. I think that how you make this, how you turn this into like, into practice, into day to day. Life, I think, is the real question here. I think that the couple takeaways I would share on that is the 1st takeaway is, I think, when it comes to all of the data that manufacturers have to put together, whether it's in the Qms. Or it's in the technical file as part of a submission. I always feel like my mental model is always that there's what I kind of call high velocity and low velocity data. Right? I think there's information. And by low velocity data, I mean, information doesn't change necessarily very frequently. So that might be like an sop, a plan and something like that. And then there's high velocity data which might be a software requirement risk a harm. And that information tends to change very, very often. And one of my takeaways, I would say to kick off this conversation is that you have to know that for different speeds you need different tools, right?

And so I think the idea of using Google word, for example, as you say word, as a piece of software to be able to manage high velocity information like a risk or software requirement or testing requirement is not the right tool, and whereas obviously much more, you know, elaborate tools like, you know, some of the Sas providers have around Jira and confluence, and the list is quite long, is much more of a suitable tool in the context of high velocity information. And so I think one of the takeaways I'd share is that it's helpful to identify what information needs to be stored in what format? Because in sometimes low velocity, information like an sop. If it's just in a Pdf. As much as we talked about not liking Pdfs, it actually doesn't matter nearly as much as high velocity information in our experience. And I think the other thing I would say about another takeaway about digitalization is.

I think, that if we want it to happen in practice, and we do it at Scarlett, but we want it to be more broad in the notified body market. I think that we have to let notified bodies compete about it rather than it being something that goes through working groups that are centralized through the Commission and through Nbcg, and so on, because it's just going to take too long, and I don't necessarily think that manufacturers have the have the luxury of that time, and I think it's better that notified bodies find ways of kind of offering the best service possible, which really is our responsibility. And maybe the final thing I'll say as a takeaway on that is when it comes to digitization. One question which is still unanswered and we're working towards finding the sweet spot is

what level of granularity in the digitization is the right is the sweet spot. And what I mean by that specifically, is. every single manufacturer has their own ways of working, and they're going to generate data using different tools in different processes. Their work culture might also be different, and so on. And it's important that the digitization, the interface that they then have to the notified body, for submission has to be one, which kind of works for almost all of them right? And so what that means is that that has to be a very flexible interface right? And so, for example, the Eastar submission in the Us. It is a form of digitization, a form of, and one benefit of it is, it's found a level of granularity where it's able to work for almost everybody. Right now I'm sure that the FDA attorney would say, I still have issues with it because people submit information that's incomplete, and that leads to delays and so on. But you can see that it's

if you make it too granular, then I think that the manufacturers have to fundamentally change their Qms. For example, in order to submit a technical file. And so there's something interesting about finding that sweet spot. But I'll pose that.

So much, James, I think you mentioned multiple points which I will quote you about, which is actually which are key. The 1st point is. Actually, you are as a notified body service provider. And the best customer experience is actually a lean customer experience that they feel actually really catch up from the beginning and not facing challenges. And they feel actually, you are taking care of them. We're going to talk about that with Michael. But the other point which you just mentioned, which is also important. We need a harmonized methodology between notified bodies, because manufacturer are not always just working with one single notified body. They might have 2, 3, or 4, or sometimes, you know, even when I, when I in my company, I supported the manufacturer doing a harmonization with the notified bodies they had 17,

so they had to adopt to a lower number. So, having even 4 or 5 is a challenge if everyone has its own system. So we need to have systematic solution. With that, Michael. Looking to you, I know with your previous experience in industry, you created Medavia to help manufacture getting. And I remember in our discussion you said it's not just uploading a Pdf. But creating a system. You saw my numbers. I mentioned actually how long it takes to assess technical documentation to get the certificates. But also I mentioned how long it takes to apply even at the very beginning, how would actually a digital solution. Help on that to streamline the process. Do you have any ideas recommendation that you can share with us?

Yeah, you mean the the application process itself, or the really full tech file review.

Both of them. I want you to address both of them. So start with the application, but also talk about the technical documentation itself.

Yeah. So I think, 1st topic is, we need to think data. We need to think about data and not documents. Because documents. If you have put information into a document. It's kind of dead. It's in there. You need to open a document. When you search for things you need to read it, you need to understand it. This is a cognitive burden for the manufacturer experts, but also for the notified bodies. And if you think in a structured data approach, then both areas, the application process and actually, the maintenance of the technical documentation is a different kind of level. Talking about the application you mentioned this kind of problems where it's like 6 months to get this kind of application process go through.

But imagine, if you have the structured approach where it's totally clear about your notified, your medical device, codes and everything. What you're trying to accomplish, and you are not able to ask a notified body that that is not applicable to you. So if this is inside of a digital process, you streamline the burden on the notified body side. To answer all this false requests from the manufacturers beforehand, and putting all the all the structured data into a application process on the other side notified body have everything in place. They need to assess if they want to take the job in assessing this technical file or this product or not and jumping over to the assessment itself.

Everything that is structured is basically served on a silver plate at the notified body site at the review. Imagine the clinical evaluation assessment report when we are looking for the simplest things like intended purpose or the device description. If it's already there, the people can do their jobs and simply assess what they are in front of what is in front of them, but not start the search where to find the intended purpose, and maybe I find 3 different kind of styles of the intended purpose, and so on. So the structured data approach is is central here.

Thank you, Michael, this is very helpful. So we are talking about data and not about documents and not about electronic documents, but talking actually about creating a system with all data information which allow us actually to connect the dots afterwards. James, looking to your background. This is what I learned that you are a software and AI expert as well. do you think? And this is where people now, everyone will put next to his website, AI driven. Do you think AI will be really the solution for the future? That manufacturer will say to the AI machine, please prepare my technical documentation for device A and the machine will do this, manage the documentation over the product lifecycle? And then, if so, what are the issues and limitation from a compliance point of view that you expect as an independent, impartial, notified body.

Yes, I think the answer. I think it's very likely that the way that that technology is going these days, and not just where it is today, but also the gradient of the curve at which it is improving means that I think that edge cases will disappear. And there's too much global.

Brainpower going towards making these products as incredible as they already, you know, even more incredible than they are today. And and so I think that. Yes, people will use Llm. Based products to generate and maintain compliance data related to Mdr. And we are very interested in finding the most effective ways of obviously providing our structure, or what we believe are the best way of submitting a technical documentation that is then generated, not completely by a human, but also by AI systems, you know, in a co-pilot based interface potentially with, you know, tools like Medievo, and so on. And and I think that I think the challenges there obviously are, you know, there's the kind of obvious challenges around hallucinations and so on. That manufacturers have to have to solve. But I think that it's, I think fundamentally, it's, you know.

the more specific the use cases are, the better it gets right. So I think that, like generally, these models. If you give it very broad, general prompts, and you ask it to generate a technical file. I guess it could do it tomorrow, but it'll be full of errors. Right? And I'm sure Mike will have a lot more direct experience with that part of the that side of the fence than I have, but I think that the challenge there is making sure that the use cases are as narrow as possible. And you can create many specific use cases. I think is what's going to generate far higher quality technical data as part of a submission than very broad ones, which I think are likely to

below quality.

Thank you so much, James, looking to what we just actually discussed, Michael. So far. what would be actually your your recommendation from a practical strategy for implementing a living quality management system, living documentation that reduce actually the administrative overhead, but still are actually meeting the requirement. What is from your view actually what should be considered from the beginning.

Yeah, this is a good one. Because spoiler if you decide today to to start with the digital documentation, it's not going to happen tomorrow or next week. So there are so many areas, and also, like James talked about the granularity. So is it on risk level, is it? On what kind of level do you tackle but the last 7 years, as we are doing this stuff, it was clear to us now that if you are at the very beginning of your journey, and you want to start this digitization journey in your company there are 2 good approaches how to structure that because you will not have a big bang implementation. It doesn't make sense. And the 1st approach is that you identify what is generating the most value for your company when you're looking at your processes.

This is sometimes a little bit harder. So that's why the second approach which I really like is, you are looking in your department, and you are asking you the question, What is my biggest pain? Where do I have my biggest pain? And this might lead you to a very valuable process, start to begin with. So either it's maybe the the way how you are managing your standards in your company. Maybe this is a pain or inconsistency between clinical Pms and your technical documentation. So you need to ask this question, and this might lead to that 1st touching point to start your journey, but it needs to be. And this is a very important recommendation. It needs to be iterative.

So you need to start, define a scope for a specific amount of time, maybe 3 months or so, and then learn, adapt and then move, move on. Do not plan for a 1 year or 2 year implementation of such a system or such a project, because it will cost you a lot of time without the learnings in the loop.

Thank you so much, Michael. This is very useful and helpful recommendation for everyone who's listening to us. It makes sense that you have really a stepwise approach. Don't jump and always measure the improvement that you're doing, and don't expect the system that will be delivering immediate results. It will take time till you see actually the output. James, is it reality today? Are we seeing? Are you seeing this as a notified bodies that manufacture are digitalizing? And where, or yourself, as well as a notified body, that you are applying digital method, and how they can actually support you, reducing the amount of resources that are like

an overkill for a company from an administrative point of view, because they are actually allowing you to use your resources more efficiently.

Yeah, I mean, on the side of our customers. Yeah, I mean, all of them are using digital solutions in order to submit Qms or technical file information to us. How many of them have used Llms to generate that information I don't know. Actually, I think that there might be a couple and a couple of use cases, but all of them are using, you know, a lot of the, you know, the kind of tools like Eqms, providers and confluence and jira, and all these kinds of different setups that are popular in the industry to be able to build and manage their compliance data. I think the challenge that's interesting in digitization is also this idea that there is this whole universe, that the manufacturer has this whole hub of all these different tools. Some of them, like a ketrix, will pull that information from different places and collate that information.

But eventually there's always a moment where that manufacturer wants to put it together into a release, a snapshot right where they're saying, Here's this kind of holistic view. It is my technical file. It is my Qms. And then to then send it over to the notified body. And from our perspective we need to know when there is a specific snapshot, a line in the sand if you will, where? Because obviously that's the moment that the manufacturer says, I am compliant to Mdr. And I am formally requesting that you do the conformity assessment right? And and so what that means is, there's still this need, even in digital world, to be able to draw that line and say, here's this holistic set of data or data and documents like where the combination may be and for somebody to assess. And

so definitely, our manufacturers are using those tools. And at times we've even asked. Some of our customers have asked to build integrations, to be able to provide data to us programmatically rather than manually to your second part of your question around, are we developing digitization tools? I mean, what I would say is that I think close to half the headcount of the company is product and engineering. And so these are people not in conformity assessment roles, but are individuals in the team designers, product managers, software engineers who are specifically focused on building tools, software tools to help really, really, in broadly, 3 areas, one is this area which is helping manufacturers submit information to us, so they might prepare it with tools like Madevo, and so on, and but eventually they have to give it to us, and we want to be able to give feedback such as you know.

your implication is incomplete, or this information is missing. This file is missing, and and things like that, so that we can again to your point around application Review. Make that as fast as possible. The second area where we're building tools obviously is we have, as you described in your presentation as a notified body. We have a lot of processes, right? And so we need to move, operate quickly internally. And so we have our own Qms as well. And so obviously, we build a lot of tools to be able to manage our own internal compliance, and those are all built in house and custom built, and then the final one, I would say, which is one that we're still only testing for now. But is this idea as well of being able to see where we can use software to help in assessments as well. So not just tooling to help with assessors. But are there even co-pilot based workflows that can help assessors when doing technical file reviews or in audits, and so on.

And those are just testing for now. But again, we think that there's quite a lot of promise in that area as well.

Sounds really promising. Yeah, go ahead.

Sorry, guys. I love the fact, James, that you mentioned that you have, like 50% engineers working on that kind of solutions. I mean when you are looking, maybe, on other notified bodies? What is the ratio between assessors and engineers that works on on streamlining processes and and works on tools. So I think this number is significant on your side.

Yeah, yeah, it is. It sounds like you're you're doing working on a big, different future hopefully, very soon and hopefully will influence actually, the the timeline.

And the costs connected to all of this, because this might be really a big differentiator in the future. If we have, like digital solution much more actually integrated. And I assume that you are doing all of this because you saw mistakes happening which are actually leading to delays in the holistic review process. Can you can you give us, like 1 2 examples of mistakes happening today between manufacturer notified bodies which lead to such a 50% investment in developments.

Yeah, I think that the 2 types of mistakes I would. Highlight is like at least these, probably to the audience, won't be a big surprise. But I think that the 2 are, I would say at times like a lack of detail, like a lack of depth of information on a specific piece of information. So, for example, clinical evaluation, activity, methodology is a common one where there's just not enough detail there, and and that can lead to delays because we have to ask for more detail, and I think the other one, which is again probably quite familiar, is sometimes like issues around traceability. And again. That, I think, is where structure and digitization is. Some of the most powerful places, is making sure that there is traceability when it's required, whether it's in the kind of software development lifecycle, or it's in risk management, obviously, which is very common. And I think that in those places issues with traceability

or lack of traceability is often still a pitfall. And so I think again, what can digital tools do there to help like we can do? We can help by providing some automated feedback, saying what you've given lacks detail and so on. But obviously, I think the real, the real place where that can be resolved really is on the side of the manufacturer. The tools that manufacturers use to be able to put together the technical documentations right? And so there, I think that it's. It's, you know. One thing that we're interested in, for example, is we write out and build quite a lot of enablement content. So this is for our customers to be able to understand how this conformity assessment work with Scarlett? What does Scarlett expect when it comes to certain

aspects of the technical file submission. But we're interested in making those more readily available to providers like Madevo, and so on, to be able to understand if they have customers who use scarlet. What's the best way to build your traceability, matrix, or all these kinds of tools, in order to make sure that when we receive them they look as similar as possible, so that we can optimize our assessment process on our side.

Which is very important. Going back to your initial comment, it would be great that we have one standard for everyone which is defined by all notified bodies together, making sure that companies like Madebo and others are making sure that they are connected. and then this will streamline the process and avoid actually the burden of actually the unnecessary challenges that we look to, because a digital solution will not fix the fact if I have good data or not so the data themselves. Whatever you're going to put in the system, your digital solution will not make you always 100% succeeding. If you have rubbish data from the beginning, it will be also leading to negative results.

So you need to make sure that you have the right data. But you are giving a structure which is reducing your burden and increasing actually the efficiency during the conformity assessment process. Michael, we hear about all of this. It sounds a bit actually tough, especially when we look into small or smallest manufacturer as they are defined in Europe, to be very tiny small manufacturer, which could be one or less than 10 people in one company, but sometimes just one single individual. What are the challenges, and that those companies might face in getting this implemented? And do you see, like a big investment that they have to do to make sure that they have such a system.

I think we need to cover the topic what you're trying to implement, basically. So imagine there is this, let's call it data model that all notified bodies are aligned. And you want to use this data model. So you still need software, a solution that is managing this kind of data. So you can exchange it with Scarlett or other notified bodies that are using that and right now there are no out of the box solutions for that it's simply not there. So one approach is that you might go along the way and say, Okay, I will build up an access database and stuff like that, and try to build it and try and error, and then move back and forth, and so on.

or you can look into the market what is maybe already existent in that area, and there are a few. So you might look when looking for your Pms processes. Then you might go look for met Board or Flynn AI. When you are going for technical documentation on content level, you might visit with Medevo and check that. Maybe that is a fit and this is a huge difference that we see. Maybe 10 years ago, 10 years ago, you need to start a large project, an assessment. And you need a project team. Today, you can kind of shop solutions. And these solutions can talk to each other. And I think this is the greatest opportunity for smaller manufacturers today, because you have today's technology and you have best practices solutions for the medtech industry. And this combination.

And speaking about today's technology and talking about software as a service solutions where you can simply subscribe to the service, and you can start using it. And if the service is tailored for the medtech industry you have out of the box processes, and you can start. And what we see the last 7 years is when we got approached by larger enterprises for our solutions. Then they have a lot of other solutions, a huge It landscape. And it's a kind of a process and a project to implement medevo into the landscape. But when you are a smaller manufacturer it's easier, and it's faster because you are not adapting than a lot of processes, because you simply use the best practice approach that the solutions already have, and then you are really faster, and it means you have a head start

against the larger corporations as a small manufacturer. And this is this is a pretty good, I would say notion of our time that we are in today.

Yeah. So if I take you well, big manufacturer have also challenges because of their expectation, structural expectation, number of devices, number of resources. But everyone can benefit from it in a different way. So small manufacturer or big manufacturer, they can benefit from digital solution, and the result of that benefit can be connected to actually reduction of capacities dealing with administrative burden internally. but also reducing the time in communication with authorities, notified bodies whatever, because they have better structured documentation. Looking to all of this, James, listening to what Michael was saying.

the product documentation are actually lifecycle documentation. So we have post-market. We have post-market evaluation, continuous like update. How? What is actually from your view? How can a manufacturer use this digital tools which are Michael mentioned, some of them to make sure that they're creation of documentation. Update of documentation is running smoother because everyone is complaining. Annual update manual update, even intermediate updates update on requests, unannounced audit announced, audit supplier audit. All of these things are leading to continuous review of documentation and update what is from your view, actually, how manufacturer can manage this in a better way.

I think that the best way for them to manage it is to use the best tools possible, right? Like, I think that nowadays a lot of these tools to Michael's point are, I think, the best tools are the ones that integrate with other tools. And so I think, because the reality is that I remember seeing a statistic around like the kind of average, let's say startup, for example, like small, you know. Smb, business or Smb business like, will you have on average, something like 50 or 70 pieces of software in their business? Right? And so, because businesses have so many different tools that they use. I think that it's quite critical that any kind of tool that you buy as part of your digitalization efforts or as part of maintaining your quality management. Technical file is effectively a tool that speaks to the other tools that you have right.

And I think that that's critical. In order to minimize that friction of to your point of. I've changed information somewhere, and it goes somewhere else, and so on and so forth. So I think that that's a critical element. And I think another critical element, again, is to my point at the beginning, about this high and low velocity point is, I think, your point, Basil around. Information that changes when you make changes, that changes often is falls in this kind of high velocity category right? And I think that's the information it is critical to remove the micro friction of that information being not updating automatically or not being notified at the very least, that you need to go and change that information. And so I think that it's important to invest in these tools that enable you to do that. I think that if in many ways like, even when I look at our own Qms, like, we build a lot of software to build automated notifications and things like that in order for us to know when

we have to do competence monitoring, or all these kinds of processes that we have. But there's certain things where we have certain policies. They just haven't changed since our designation audit. Right? And that's okay. And and so in that case, it doesn't need to be more than what it is right.

Perfect. This is this is great, because this is actually pushing us to the next question. So, Michael, you have a company which is dealing with technical documentation digitalization solution. Typical process. A manufacturer got the change. James was referring to a change or a standard change. And then you need to actually do a gap analysis. And you need to update your documentation. Why would the digital solution accelerate my process? What is what is the benefit connected to it. If you look to it from from creation of documentation and update of documentation in comparison to manual method, can you give us, like some insight from your expertise and knowledge in this field.

I like your example with the standards, because recently we had a discussion with a client of ours regards of standard updates. And there is a lot of impact analysis and stuff that you need to do manually when there is a new standard arising. So imagine if you have a new standard. It's not like, okay, I have a standard. And I need to implement it here. My technical documentation, the reality is, it's 1 standard update versus 100 documents. This is the reality. And it's this kind of administrative burden that the experts are facing right now. and especially this specific process. We have pretty cool Apis right now about the standard management.

because we have a manufacturer that are dealing with 20 files, and I don't know 100 or 150 documents that are affected. If this occurs and they time until it's then closest change might be between 3 and 4 weeks, because they need to have this sign of rounds and everything. And then the regulatory affairs. International team can start the impact assessment on actually international registrations. And when you are thinking back to our 1st discussion. The 1st questions that you had where we said structured data is something that is really leading us to this kind of benefits is right. Now, if you have this standard update, you will immediately see the impact out of the system because you are dealing with data structures. So there is this new standard. And you see this

150 documents flagged for you that are impacted. Your assessment is pretty fast, because you see this kind of 20 products are affected. You can run this implementation process and then simply click a few buttons to update all these documents and then create a new submission file, or how you call it James, the snapshot of the current technical documentation. To submit it, then to the authorities, or whatever you need to do. And especially this process is a speed up of around 90% of your time. And this 90% of your time is where Rn dealers, Ra experts can actually do value generating work and not copy and paste and

data search activities. And this is a really really strong use case that you mentioned. But it's applicable for everything. When you are having maybe a Pms trigger where you need to update your patient population. Then you have the impact on clinical evaluation reports on Pms and so on. This is the same methodology. You are not dealing with documents. You are dealing with data and let the machine do the boring stuff, and the experts are doing actually the value for your company.

Very good point, because it's like, this is a typical problem which happens change, notification is submitted. Key documents are updated. The remaining supportive evidence are still referring to the old information questions, why is it like this? You are inconsistent with your documentation? And this is actually where I can get my process much more internal, audited. Before I submit the documentation, make sure that I'm addressing all impacted document, at least with a review and decision. Whether I'm going to change them or not, which makes actually me more consistent and more efficient in the activities before.

1.1 second buzzer, because I wanted to ask James a question. Right now, if you have the standard update, of course, the Gsprs needs to be updated and so on. But imagine now, if there is this approval on a data point level, so that the actually, documents are for maybe the vintage notified bodies. Let's call it that way where you have, then disapprovals. But if you are approving the data points, then all the review steps and and maybe wet signature steps are not necessary on the document level, so this will speed up the manufacturer process. But what about the notified body? James.

You mean about us receiving data rather than documents.

Yes.

Yeah, I mean, of course, I think that one of the things that's interesting to consider from the notified body perspective when it comes to digitization is one of the challenges that we have to solve, and we still continuously solving it is this idea that obviously, when you receive information, let's say, data from a manufacturer is like for us to be able to be more efficient in order for us to be able to optimize our processes, we need to know that what we're going to be receiving it's always going to be different because it's a different device, different risk class, different medical specialty. But if the information is in a similar, at least at some level of granularity, similar format, then we can optimize that process right? Because otherwise, if you open kind of Pandora's box, and it always looks completely different.

Then it makes it very much, much harder for us to be able to optimize internally. Right? You need to be able to kind of a B test right? And if you don't have a level of consistency. Now, the question is, what is the right level of consistency, right? And what level of detail is required. But I think it's quite crucial for us that we have that in order to be able to systematically be able to optimize our processes and say, Okay, it's very basic. But it's like, if we receive something we think about our process. For example, for an application review. And we think that it's like, okay, how can we make this application review faster, which is something we care about. And so then product and engineering work with some assessors.

We built some tools to make it faster, but then, if somebody else then submits for another application, and the way that they submit. It is completely different than what we just built can effectively be put in the bin. And so you can see why it's important from the refined body perspective in order to be able to optimize our processes, and so.

Think both sides not just notified. Bodies and manufacturer need to work together on the standard which is actually applicable for both sides, because this is how you can ensure that. Actually, your next digitalization steps are workable, because, if not, every time you need to adopt your process, it doesn't work. So and such a solution. We are not that far. Yet where the system can actually have that flexibility, because we need to train the system on what is right, what is wrong. This is how the machine will be actually more efficient over the time. And this can just happen by having actually a structure which is standardized for the future.

We had very nice and interesting conversations so far, but I'm pretty sure, Sean, we have questions from the audience, and I will hand it over to you to lead us through those questions, to see what is the audience also like keeping them awake at night? What are the problems that they see, and they want us to address for them.

Happy to launch into the questions. Before I do that I want to make a quick announcement. Our team is sponsoring quality accelerator program, and it was scheduled for June. But it has been moved to September 3rd and 4.th If you would like a copy of the agenda, please write agenda in the chat, and I will send it to you. Thanks now on to the questions. While you guys were talking I was reflecting on the fact that the FDA did its 1st review of a pharmaceutical submission last week completed that. And they're planning on rolling this out for medical devices in June. Any comments.

This is specifically the pilot around using AI within FDA within that department of FDA, right.

And they're planning on, rolling it out for devices as well.

I think it's great. I think that you know I think these tools like it's necessary that I think, like I say, in the kind of in this example, like, I guess, like we have like FDA, like with people like doing the reviews. I think that again, I think these tools are very powerful. And again, there's these 2 use cases where where we are interested in these tools for our internal processes. And again, Basil, in his presentation, showed some of the a snapshot of the many requirements that apply to notified bodies. But again, there's also this aim of being able to use these tools in assessments. And again, there are a lot of issues in assessments around retrieval of information around all these kinds of issues where this technology can be really powerful. And again, I think that it's necessary to be able to ensure that the best technology

safely gets to market as soon as possible. I think it's important that we use the best tools possible. And so I think it's important. Yes.

I heard some concerns about the data that is provided, then to the agency. If the data is then also incorporated into the learning process of the algorithms of the of the AI, and so on. I raised the the idea about bias options that are maybe there when fed with this data so that the AI is then maybe biased and leaning forwards to one decision making. But I think that this is a good approach, and we will. We will see the outcomes. And I'm really looking forward to the outcomes.

I think if we are like scared and we don't try. We will never be able to to learn about the mistakes and the benefits connected to such a system. Therefore we need to try it. We need to test it. And it's still young. I would say it's comparable to the nineties when the PC. Started to get more more involved and developed. And this is where we are. But we need to make sure that the AI is getting an opportunity to learn with us. This is why we need to build the structure for it, to make sure that we are able to use it efficiently. and I would love sitting with James, smoking cigars, drinking our whiskey and the quality assessment processes are running via machine, and we say, Good luck go for it. Issue the next certificate. And we don't talk about how many application certification are happening because they are done automatically, and innovation are getting to market on time.

Well, that's a beautiful vision. I'd like to ask the other 2 how they think the role of regulatory consultants and notified body reviewers will change in light of the fast developments in AI and automation.

Do you want to go 1st for the consultants?

I have a clear vision on that. This is my personal view, but I think we saw a lot of consultants popping out in the Golden Mdr. Era, and then you had to buy templates and document templates and and sops and stuff like that, and then they got adapted to your needs. I think that in the era of high automated technical documentation on content basis with AI options. there will be a cleaning in the market where there will be a concentration on the experts where you will go to. If you have, maybe a device that is cardiovascular device, then you will have to go to the consultant, that is, with his high expertise on cardiovascular devices, because you don't need a consultant to explain to you how a clinical evaluation looks like, and how the process works. It's digitalized.

But you need, maybe the experts, then, to get your device safe to the market to do all the tasks and generate the data for it. So I think that this will be a consolidation on highly trained experts of specific areas, but not like generic consultants.

I think on the on the notify. Buddy. Reviewer perspective is what I would say is like. it just gives you so much leverage, right, because I think that the reality is, if you look at the profiles and the competence requirements for Mdr. For reviewers like the expertise that we have to have in the team is these people are world leading experts in our case, specifically on AI and healthcare, and so on. And all these medical specialties. And so what you want is you want these individuals to have as much leverage as possible to be able to use the skills that they've spent many, many years building, and I think that what these tools provide is an ability for the assessor to really spend much more of their time on the thing that is so valuable, the skills that they have that are so valuable to technical file review, but also fundamentally to the safety of the product. And so there are certain tasks. And Basil again, when he's talking about some of these

clinical evaluation assessment reports that have to be built by as an assessor. Again, like we build those, and we do use software to help us build them out faster. But still there's a huge, the number, the amount of artifacts that we generate as part of an assessment is absolutely ginormous. And so and so it's important that these tools help the assessor focus on the most critical questions.

I think you know, if I, if I take it from a consultant or notified body or manufacturer, point of view. AI shall not be actually misunderstood to say, it's taking our job. I think we will be more focusing. And in our job, more focus on the key element. And this is our expertise, a person coming to us to ask us about software expertise, a technical expertise, a clinical expertise. We'll get that value longer term as well. But what we should not spend time on is actually talking about structure baseline. How can I ensure consistency in documentation, doing review of documentation, doing update of documentation

because someone is not able to leverage the data because they don't have a system. So you need to take an excel sheet, make out of it a graph or whatever this should be automated. And this should be actually using machines done in a more efficient way. Because then, people, we are facing challenges to have the right resources. because there is no sufficient expertise available on the market, and with the limited expertise. And I was looking recently to a country where they looked into digitalization much more than any other European country. You know why they did it because they recognize with the limited population big country. They need to have more digital solution to be able to offer the expertise to everyone, and if they don't do it in such a way, people will not benefit in the same way. And this is why, if we look to the future system.

our consultancy expertise, notified body expertise, technical support, expertise can have a much bigger scale up because of the fact that we are using digital AI driven solution. And a lot of people will benefit from it in the same way.

Yes.

We only have a moment left, so I'd like to give you, gentlemen the opportunity to make any closing remarks and say goodbye to the audience. I remind everybody to please participate in the survey.

Thank you, Sean. So, James, please.

Closing remarks, I don't know. Yeah, I mean, obviously, like, you know, we had scarlet. We're notified body. But we're also fundamentally a technology company. And again, we've invested massively since the beginning in having in digitalization of all of the aspects, of all the aspects of conformity assessment. And we think it's critical to be able to do fast and safe assessments so far that I think that's also yielding fruit. You know, if I'm honest in our case, like we've done. All of our assessments have been done in less than 6 months from application accepted to certificate, and that includes like class 3 products. And so I think that like, that's evident that digitization

helps. And we have a whole bunch of other products coming through. And hopefully, we'll be on the market soon on similar timelines, and so I think it is here, and is here to stay.

Thank you so much. James. Michael.

Yeah, this is, thank you, James, for for that point, because, this is like, I'm seeing here a closing loop. We see this kind of acceleration and automation for manufacturers the last 7 years with our product. And hearing now that you are doing this 6 month cycle for your clients, class 3 devices. I can. Really, I'm really looking forward to have this connectivity from Madevo system to Scarlett, because then we will have this timeline boost on both sides. And this is something where I think is hope for the European process. And an example that it can be lean, and it can be really really fast if we do it right.

Thank you so much, Michael and James, and let me close it with that point and hand it over back to Sean. Digitalization helps. This is what James said. And this is what we need to keep in mind. We should use digitalization as a support and do not fight against digitalization, because without it we will not survive. Thank you so much, Sean, for hosting us and making sure that we have this nice conversation. Thank you for everyone who attended online. And if you feel we teach you something or we gave you something which is relevant. keep in touch. Follow us on, Linkedin. Write us an email. And we are happy to work with you on digitalization for the future. Because this is how we can make a difference.

Thank you all. Thank you.

All. Thank you. Bye, bye.