EUDAMED Is Coming Fast

Hey, welcome everyone. I'm Sean Smith. And this is Friday in Focus. Today, we're really lucky to have special guest Sam Patton. He's here from Utaprep, to talk about, the UtiMed deadline fast approaching at the end of the month. While we're waiting for everyone to join. Feel free to come on camera. Ask questions, participate in the call. I have some big news. We've created a voting system. So we selected 100 MedTech Leading Voices worth following in 2026, and now we've created a function where you can actually vote.

And so I'm going to show you that in a second. And then Jose is gonna come on, hopefully at the end of the call in the last few minutes, to, walk us through, all that. Next week, first in human clinical trials in Latin America. That ought to be super interesting. Then, on the 14th, there's a, MLV and ClareMed webinar on, identifying the barriers to real-world use. This is effectively a qualitative research, discussion about the importance of doing qualitative research and understanding, the actual blockers to adoption of your product. It ought to be, very enlightening. And then, May 15th, we have,

Charles Orlais, from TraceX is gonna be, Specifically addressing consultants, and how, they can, make money. With his platform. And, and get their work done for free. So then we have the International MedTech Safety Conference on June 2nd through 5th. Next week, I'm gonna be in RAPS, show some love to the, EU Breakthrough device pilot post, the article and link, and, right, I'm gonna be at RAPS next week. So next Friday, I'll be live from Lisbon, and…

we have the MLVX, so the thing that I wanted to show you was… you know. the, the voting. So, if you go to MLVX.ai, MLV slash MLV100, you can actually see, the 100 people. You can, vote for up to 3 of them, and… You can, add your own nominees. You can add yourself, you can add somebody else, and you just put their,

name, your name, their name, the LinkedIn URL, select the category, do your best, and then a word or two about why you're nominating them. And then hit vote. It's that easy. So, please, take a minute to do this, if you can, mlvx.ai slash mlv100, Rob, if you can put that in the… chat, or I will. That'd be great. And without further ado, I am going to, turn it over to Sam.

Awesome. And let him tell us about… as a small child, Sam, as a small child, did you dream of one day having the Utaprep platform?

Well, surprisingly, regulatory software for, Udipred submissions was not on my vision board, no. I kind of married into it, and I'll tell you a little bit more about that story, later, but yeah, yeah, it was, was not something that I had a deep yearning for as a 6-year-old growing up in Australia, that's for sure.

Where did you grow up in Australia?

I was on the East Coast, so I was born in, born in Melbourne, and lived on the sunny Gold Coast for, for 15 years before moving overseas. So, that's, I guess, kind of like your Miami Beach, I would imagine. It's the same style and everything like that, yeah, yeah.

Very nice. Miami is a lot of fun.

Excellent. Cool! Well, shall I, shall I flip up the slides then, so we can, we can get into this thing? Alright, let's do it.

Looking forward to it.

Cool. Alrighty. Well, look, thanks, Sean, for having me on. Thanks, everybody, for making the time, and, and joining us today. My name is Sam Patton, and as mentioned, I'm the founder of Udeprep, which is a software tool which has been developed. To help you get ready for Utimed, because Utimed is screaming around, it's coming very fast, the first deadline is in just 27 days from today. So essentially, what I want to take you through for the next 45, 50 minutes or so is why this work is quite a bit harder than most teams think, and what the current tools that are out there on the market today solve versus what they don't.

Little bit about me. So, I am a systems person, okay? So I'm an IT… IT background, nearly a decade of consulting across a bunch of different industries. I've been working with medical device manufacturers in various different capacities since 2017, and work in other regulated industries as well. Now, when I say that UtiMed is a software problem dressed up as a regulatory one, I'm saying it from both sides. Essentially, as part of coming through and developing the Utaprep tool, which I'm going to be showing a little bit about later. I reconstructed the entire schema for Unimed from scratch. using the European Commission's published documents, so every single field, every business rule, every enumeration code, every XML schema constraint, I've lived inside the data for months, and I know how every single piece of it connects.

I will say this, though. The regulatory judgment calls in this hour belongs to you, okay? I am not the regulatory affairs, specialist, that is, that is your domain. As I said, I am the, I'm the software guy, the systems infrastructure person that, hopefully is going to turn the expertise that you've got into valid submissions in a very quick turnaround time, because we absolutely need it. I'm going to share the numbers with you very shortly, but There's a lot of work and not a lot of hours to get all of the work done, and I'm sure Sean mentioned this to me before we got onto the call today. But,

the, there's been a bit of a wind-up, and people need to get things moving with this. So, just to set the table before we get into what this hour is, This is a working session. I'm going to take you through the submission preparation gap, which is really, I think, a quite underestimated, gap in terms of, of what, Device manufacturers are internally underestimating here, as far as also the work that the consultants in the room need to deal with. I also want to take you through the four-step process to do a UdiMed submission at the moment, and highlight where the work is actually hiding with this. I'll give you a short look at what it is that we're building, and we'll have some time for Q&A towards the end, but if you do have questions as we're going along.

Feel free to drop those into the chat. Before we do get into this, though, I always like to start with a question. So just in the, just in the chat here, too, I'd love to get a bit of an idea on where people are at. So, if you could just type, you know, 12345 into the chat here. And let me know, sort of, where your clients or your own team on UdiMed submission prep. The first one is that they're already submitting, seeing some of these coming through, great. Some in progress, some are scoped and not started. Some we have who are aware, who aren't scoped. I'm… I'm happy to report that I'm not seeing any, any fives of what's Utimed, because that would be a very worrying one to, to get, I think, at this point. But, it seems we've got…

quite a few who are underway, that's fantastic, so this won't be, this won't be completely, foreign to you, but, as far as the scoped and not started, cohort that we have here, I think you may see some interesting stuff as we, as we go through here as well, too. Alright, fantastic. Thank you so much for your feedback. So, the reason I ask, is that there's a really common position, at least from the conversations that I've been having, which is, you know, this, like, we'll start next month. And that was fine a few months ago, but it doesn't really work as a story now, particularly when you look at how much work is involved, which is what I want to go through here.

I do want to start by setting the stakes of this. The first hard deadline, as you may be aware, is the 28th of May, so that's 27 days from today. On the 28th of May, the first four UMED modules go from voluntary to mandatory under EU Regulation 2024-1860. So the first deadline, as I mentioned, it's the 28th. the… the modules that become mandatory is the actor module, so you do need to have yourself or your clients registered. As an actor, the UDID and the devices, notified bodies and certificates, and market surveillance. From the 28th of May, any new device that's placed on the EU market must be registered in Udamed before it's sold.

So that's the first deadline. Second deadline, which is coming up, is 6 months later, on the 28th of November. Now, this is the backfill deadline, if you want to think of it that way. So, every existing device that's being sold. still being sold, so Legacy MDD, AIMDD, and IVDD, as well as existing MDR and IVDR devices must be in UtiMed by that deadline. So the 6-month window between the deadlines if… for the consultants in the room, and actually, just… just so I can get a bit of a gauge, We have primarily consultants in the room here at the moment, just to… if you could just let me know in the chat who's sort of on the consultant side, or who's, you know, do we have device manufacturers here, or people that are working with,

Working inside of another device manufacturer. Okay, excellent. So we've got… the first one's coming through here. Okay, so we do have a mix. Okay, fantastic. No, that's really helpful, thank you. So, for the consultants in particular, there's a real opportunity within that 6-month, window, for billable engagement, and that's something that I do want to highlight as we're going through here. Now, I want to just share the live data, and this is, this is, this is… absolutely, blisteringly live. I, I pulled this data down, finished about 2 hours ago, so this is, this is right up-to-the-minute stuff.

As of today, Unimed has 28,865 manufacturers which are registered as actors. Of those, only 13,000 have actually loaded any devices, which is about 46%, as you can see in this one here. Which means that 16,500, give or take, manufacturers, roughly 57, 58% of Utimed's own registered population have a single registration number, so they've gotten into the system, but they haven't submitted any devices as of yet. And I do want that number to land, right? So, more than half the manufacturers who have… started the Utimed process have stopped before they've submitted a single device. So this is not like the haven't started, these are the ones that are… these are the ones that have stalled.

This is… there can be some reasons for that. It may just be that due to the fact that it's… only the actor registration module is mandatory from the 28th. Maybe we haven't gotten around to… to making the submissions yet. That being said. I know, at least anecdotally from conversations that I'm having, is that there are a number of there are a number of people who've began the process, so they've hit the work streams, you know, steps 1, 2, and 3, which I'm gonna take you through shortly, and they've hit the wall. So, these are… for the consultants in the room, if these are your clients, these are the ones who are not going to come back on their own. Deadline is 27 days, they are going to be looking for help.

And that's… this is just what we're looking at on screen here, is the ones that is, currently registered in… inside of UtiMed. This doesn't account for the population that hasn't registered at all. So let me dig into these numbers, real quick. So, the first is that this is inside of UniMed itself. So, 57.7% of the manufacturers who've started, they've registered it as an actor, so they've met their minimum requirements, for the 28th, but they've submitted zero devices. 662 of these also hold an active notified body certificate, right? So we know that we've spent a decent amount of capital figuring out that, you know, proving that their devices comply with MDR, paid for conformity, and these are people who, you know, like all of us, you can't afford to lose that investment.

The second one is, I think, perhaps more telling, is when we look at the cohort outside of Utimed. As a data problem, if you look at the registrations that are inside of Utimed, we can gauge, you know, where things are at at this present moment. What it doesn't, however, account for is the… the, those who are yet to register inside of UniMed. So what we're looking at here is, to give you a bit of an idea on the numbers that we have, we went through and looked at the FDA's list, so we filtered the manufacturers inside of Europe, UK, Switzerland, and this results in roughly 36,000 companies. When we cross-reference that data with the current registered Utimed data, we have approximately

14,000 European companies who are not yet registered as actors. When you compare these lists, so the companies that appear in the FDA list but not in Utimed's database, these are the ones that haven't registered yet. So there's an additional 23,000 minimum, that is verifiable that are not included in that cohort. I don't know whether you've been looking at this, in the trade press, or not. There's an estimate that's flying around which… which states that there's approximately 38,000 medical device companies in Europe. This is Europe… this is Europe only. So, if you look at… the, the European Commission's own numbers with this, over half of those are non-EU registered manufacturers.

If you were then to go and look at, say, who's registered in some of the other bigger databases, such as the FDA, etc, there's a lot. So the point… the point that I'm trying to… to tell you here is that this 38,000 number here, it's being thrown around, I don't believe really represents the full gravity of businesses. manufacturers which will be subject to these requirements. It's actually quite significant. the headline here is that the population that this is going to affect is materially larger than Europe only. I believe this could be, for the consultants in the room, the largest billable expansion event in EU medtech consulting in a decade.

This will be concentrated, the cohorts themselves don't fully overlap. But you've got tens of thousands of actors who haven't registered yet. Each of these actors is a manufacturer, or AR, or an importer with a portfolio of devices behind them. The unregistered actors are concentrated, that are not currently in, organized, and registered into UdiMed, may have smaller portfolios, but they're the least likely to have an in-house regulatory team that can hit the deadlines that are required here. These people will pay for help. They'll pay good money for help that's fast, and they'll pay a premium for…

Help that's fast and correct. I put it to you that the question isn't whether the work is coming. I think it's pretty clear at this stage that there is a phenomenal amount of work that's going to be coming as far as UNIMED submissions are concerned. The question is whether you can deliver fast enough to capture it. And that's sort of the through line for the rest of this hour here. So, when we look at… when we look at UdiMed, and what I believe to be sort of the core problem here, 60% of the data that is required for a UdiMed submission is generally not ready in the correct structure to do this.

Unimed submission is a four-step process. So the first step is that you do need to… Do document intake. So you have to find the relevant regulatory, data in the manufacturer's existing documents, so things like IFUs, labels, technical files, safety data sheets, etc. The second step is matching the codes. Now, every code in the European Commission's schema, right, the UTIMED, the framework that makes up UTIMED, is a list which is published and maintained. It's your job when you're making a UdiMed submission to ensure that the right data points that you have connect to the right code. Easy enough in practice, but in reality, it turns into quite a lot of heavy lifting, sorting through documents, etc.

The third is, of course, reviewing and auditing each match with the full provenance data. So that the submission that you make into Udamed survives an audit. The fourth, is generating the XML file, okay? Now, the XML is the… just… for those of you who are unfamiliar with the technical side of things, it's the… it's the technical file which you upload to Utimed if you're not doing it directly through, their online submission, tool. Now, The thing… I think… people are not talking about here is that… is that, and I'm sure some of you on the call here can agree to this as well, people are talking… you hear these things about, you know, a UdiMed automation, or a UdiMed, you know, automation tool, or an upload tool. And people generally think that it's… they're… they're solving

all of these steps. But realistically, a lot of the tools that are out there on the market at the moment, they're just solving the… this fourth step here, which is the upload. This presupposes that the documentation that you have, et cetera. is in a format which is ready to be submitted to Unimed. And for a lot of device manufacturers, that's simply not the case. Most of the products in this category are genuinely Step 4 products, so they automate the uploads. They handle what we… the machine-to-machine plumbing, right, which allows you to sort of push from whatever you're doing into the Utimed database.

That's absolutely fine, but this does, as I mentioned, assume that the data is already correct, structured, complete, validated. The… industry gap analyses that the literature tells us here is that approximately 60% of the data that Utimed requires is either missing or unstructured in manufacturers' internal systems, which means that even before you can run one of these Step 4 tools. You've got to get through steps 1 to 3, to do that. And steps 1 to 3 is where the actual time, the judgment, and the billable hours go. This is below the waterline, as I like to call it here. This is the work. And so this is why we built Uniprep to do that, but I'm gonna take you through what the steps are, and how you might go about doing this. Last poll for the moment,

I'm sorry, I'm one of these interactive speakers. Just to get a bit of a gauge here, because this does sort of impact the portfolio size that we're looking to do here, how many devices are you or your largest clients registering for UdiMed? So, 1 to 5, 6 to 25… We've got quite a few in the 1 to 5, okay… 26 to 100, over 200… Okay, fantastic, yeah, so some… some quite in the, in the larger cohort as well. 2 so far, three in the pipeline. Okay, fantastic. So we do have a range here as well. That's very useful. Thank you for the feedback.

So, let's jump into the first step in a Unimed submission. So… This is the document intake. Now, how this works is that, you've got a stack of documents, okay? You've got documents in different formats, you have documents in different languages, written by different people across different timelines for different purposes. Now, somewhere in these documents is the answers to the questions, is this device sterile, and what's the maximum number of reuses? And what are the contraindications of this? And a few hundred other questions which are… which are… mandated by the Udamed database. Now, your job in a submission is to find every one of these facts in the right language, with enough provenance that you can defend this answer.

If anybody audits the submission. And manually, this is the part where I see RA teams spending afternoons, days on this. Not because the work is necessarily hard, it's generally not. It's tedious, however, it's repetitive, and you can't outsource it without rechecking against the source documents yourself. So… some caveats, obviously, when it comes to doing this document intake, even if yourself or who you're doing the work for, are quite organized. There's 4 specific failure modes in this which aren't in the UdiMed training material. The first is your cross-document conflict, okay? So if you have an older label, an IFU was revised, and the safety data sheet says a third thing, which is the canonical answer for UTIMED. If you pick wrong, then you've created a discrepancy between the registered submission and your internal documentation, and an audit is going to surface later.

The second is source language drift. This is quite a big one, I think, particularly in Europe, given the number of languages which are as part of this, right? So, if you've got an IFU in Dutch. all of the European Commission's codes, so the business rules, which I'll share about in a moment, these are all only published in English. So the audit trail has to preserve the source language for provenance, and then prove that this matches across another language. Now, this is a non-trivial problem when you're doing it manually across 5 EU languages. I know for a fact that the current workflow for some RA consultants consists of having Google Translate open on one tab.

In order to try and sort of wade through the, the technical files there. Now, there is a better way to handle this, but most importantly, you need to be able to de-risk this part and make sure that you have a proper audit trail. And the third is partial documents, so some UDMED-required facts won't be in the documents at all. This gap is part of intake. It's not a separate phase. You will discover what's missing as you go through the submission. The fourth is ultimately the the multiplier on the other three. A typical product family has 20-plus documents in it. A typical portfolio, you know, as we've seen here in the chat, can have dozens of products, can have hundreds of products.

So you're not solving cross-document conflict once, you're solving it on every device, every product family, across every certification. So the work that's painful at one device becomes catastrophic at 50, because every failure mode I just described compounds with the portfolio size. So… how do you handle this? Well, how Udeprep handles intake is to… take your unstructured documents and turn them into structured items. So, essentially, we make the process of collecting all of your documents and the data archaeology, as I like to call it, is by taking these, is by taking your documents, so whether it's a PDF or image, Word docs, scanned, handwritten, it doesn't matter, essentially what we have is an engine where you can take all of that information, we extract it, we pull out the necessary regulatory facts with the source language preserved.

We also provide an English translation and a section detection as well, so that all of the relevant information for your UDMED submission gets auto-matched and mapped. to the correct fields. We also produce a gap analysis, and I can show you an example of what this looks like if there's interest, which shows you what's missing from your submission. So the point that I'm trying to make here is that step one ultimately needs to take you from your unstructured documents To structured, reviewable items, and that's the first cut into this 60% figure of mess that we talked about before. So… So that's step one, which is your document intake. Now, step two is code matching. Now, it should be said that UdiMed isn't a free text database, okay? So every fact that you've pulled out in step one has to be matched to an official code from one of the European Commission's published enumeration lists.

To be clear, these aren't small lists. If we look at the current breakdown of these, there's over 462 critical warning codes, there's approximately 30 storage and handling conditions. The EMDN codes as well, there's well over 8,000 of these which are arranged into the tree. And these are just the big ones, okay? So you still have risk class, legislation, market status, applicable directives. Every single drop-down in Utimed is a code list. Now, manually, the way I've seen teams handle this is that they have a spreadsheet of codes, open in one tab, and then you have the IFU open in another, and you're flipping backwards and forwards, looking for what the closest match is to these, these data that you've extracted.

Now… If you haven't looked at the way in which the data, is structured inside of Utimed. The EC publishes their validation logic in terms of what's called the business rules. Now, you need to be very careful here. The EC's business rules document for the UDI module, so what you can see here. This is hundreds of pages, they have narrative prose in there, so there's no structured data, nothing's machine-readable. You have text, right? You have, an ID, such as BRUDID114, You know, there is, there's hundreds and hundreds of these, these codes.

Now, every time the EC updates the platform, this document changes. Sometimes the rules change, sometimes the rule numbers change. Sometimes both. So, the point is that if you're going to validate your submissions against the EC's actual rules… and not just whether or not it will allow you to upload an XML file or not. Somebody has to read this PDF, somebody has to turn it into structured logic. It's not glamorous, and it's not in any product brochure that I've seen, but that is the difference between a submission that passes Udemed's validation, and one that doesn't. Now, let me jump over to the next one here, sorry…

So, I am also aware that there are more than a few, solutions whereby people are tracking this in terms of, inside of an Excel sheet. If you are doing this, if you are entertaining doing this internally. And there's some good reasons to think that. I do want to be honest about what this… what this buys you. So, as I mentioned, the EC's reference data drifts. The rules change, sometimes the rule numbers change, some get deprecated. The thing is that if your in-house business rules table isn't being maintained against those releases. you will have submissions that drift out of compliance silently. There's no error message inside of Unimed that says your logic is, you know, months late, or something like that.

Inevitably, you'll just start producing submissions that perhaps pass your own internal checks. But we'll fail the EC's checks. This sounds like it's a pretty straightforward process, like, we just pull out the data that's on, on Utimed, and we will manage it through there, because it's easy to export XML from here. The point is that somebody has to keep this thing updated, distribute the new version, retain it forever. This is a… it's quite an ongoing contract that a lot of, a lot of… compliance specialists are not seeing, the, the real, deal with that. So, how we have solved that problem is that, when it comes to code matching and making sure that the codes match up with what's being published in the business rules,

Udeprep has a three-tier matching engine. The first thing we do is we take the documents that you've extracted in Part 1. We then go through and we're looking for keyword phrases against the enumeration list here. for multi-language responses, things like, for instance, keep out of sunlight, keep out of reach of children, do not reuse, etc, we solve these matches at this point, and that doesn't use any artificial intelligence. We have a proprietary matching engine that sorts that. The second part is, a constrained AI assist. So. with the extracted data that we've got, what we do is we compare that to the official list that Utimed publishes here. So, if there is a match in there, or we believe there to be an appropriate match,

within a certain threshold, and again, I can demo this in the actual software as well, we will assign it for your review with a… with a level of, with a level of confidence that we attach to that, so it's there for human review. We also then have an explicit no match, which is the Tier 3 matching. So, the European Commission has, these other codes, critical warning codes and storage handling codes. Which has free text, in there. Now, this is, this is where you can fill out your Unimed, submissions, right? So this can be. For those that aren't deterministic or value-bearing matches inside of your submission, this is where you can add the additional, context to the,

to your submission there as well. Now, the part that solves the problem from our previous slide, as far as keeping this up to date, every single day, the European Commission, we scrape their website, so we see if they've published a new version. of the Udamed, schema with that. If that happens, we catch it in their playground environment, so we ingest new business rules, and we update the engine with that. So this is done well in advance of the time that the production Udamed database is there, so you will never have to worry about it, being out of date, with that. We… we are ready to deploy, at the moment that the Utimed database is updated there as well.

The last point, and we're going to, we are going to get into this a little bit as well, shortly, but we have, export validation here, pre-export validation. If you have used any of the, any of the software tools that are doing things like the, the bulk uploads, etc. One of the things you'll note with that is that You may think that you've got everything correct, but if there's something wrong in an XML file, and if you're not a developer, this is quite a, quite a common thing to come up. Your submission into UniMed will fail, and then you'll have to go through and troubleshoot what it is. We solved that problem by simply just not allowing you to upload unless…

The… the information that you've put in your submission, perfectly validates against what is in in the EC's data rules. Now, step 3, for UdiMed submissions, this is your review and approval. So, even after you've gone through the technical documents, you've found the right sentence, you've matched it to the right code, you still have to be able to defend this. Udimed submissions are reviewed, You're all aware of this, and so if, you know, two years from now, you're in an audit, and somebody asks the question, why does this have critical warning code 135 on it. You need to be able to point at a specific sentence in a specific version of the IFU, or SDS, or something like that.

And sign off, chain with that. Now, how I've seen this defensibility play out, at the moment is that it's been, you know, quite primitive. It's email threads and spreadsheet comments and stuff like that. It doesn't scale, is what I'm… what I'm trying to highlight here. So it is really critical that you have an audit chain of command here. For those of you who have used, Utimed's UI, and I don't think I'm kidding when I've done… when I say this, it's a bit clunky, to say the least. The… there are 3 more failure modes that show up when it comes to the, the Utimed UI itself. Making sure that you have inter-team, consistency between RAs,

is really important here. You can have, you know, overlaps as far as codes, because there are some critical warning codes, storage and handling codes, etc, that are quite similar. It isn't the most intuitive UI. Yes, I think that's one way of putting it, a diplomatic way of putting it, but… Essentially, you do have these divergences where either case could be a reasonable suggestion for the correct code for that. But you do need to make sure that you understand why the decision was made for one than another. The second is, document revisions mid-engagement. Generally, we would assume that this doesn't happen, but if you aren't getting the most up-to-date data.

then you… the match that you made against the old IFU is now drifted from the source of truth. The third, as I've mentioned here, it can be quite fickle. It's not uncommon for Udamed, to throw you off, sort of, mid-submission, at which point you lose everything that you've done. Half-submitted submissions, they disappear, so… They've built a regulatory database, not a workflow tool, is, I think, the nicest way to put it here. How we've handled this is that, if you look on the screen here, essentially with Uda Prep, right, when you have done your document ingestion, we will show you, The critical warning code, so you can see this one here, read a label before use. We showcase that this has been matched with 95% confidence here. You'll see the source text here, so this is a Danish, product.

with an English translation to make it easy for you to scan and match the code here as well. So, as I mentioned before, Utamed codes are only published in English, so if you are doing translations, other localized products, etc. You do need a simple fashion to be able to match with that. What I think we do, which is… which is quite unique, is that we provide the reasoning for the matches as well. Again, we are not here to replace your job, we are here to make it easier, right? So… Essentially, this is the… this is the human, human approval here. We're just trying to… to make it easier, for the consultants, you know, to… and… and the device manufacturers here, that when you're doing these submissions.

it becomes a much more straightforward process to do that. We do include full tracking provenance with this as well, so if you do find yourself in an audit. say, 2 years down the track, 5 years down the track, something like that, and you do need to say, well, why was this code chosen instead of this other code? The reasoning, the extracted text, etc. That's all there. We maintain audit trails, for this exact reason, so you're able to point to that specific document that was uploaded. Here is why we chose it, and here's why. So… Oh, actually… Where are we at? 1040. What I'll do, I'll keep going through the submission here as well. If you do want to see a copy of the, if you do want to see what the UI looks like, in Udeprep as well, I can open that up, as we get into question time and sort of just give you a bit of a walkthrough as well.

Now… what I do want to say here, and I'd highlighted this before with our, sort of, you know, above the waterline, below the waterline, infographic, Utimed ultimately is not an upload problem, okay? This is a data correctness problem. Now, as a quick aside, before we get into Step 4 here, there are categories of tools that do machine-to-machine submissions to UDPREP, to UdiMed, rather. So these are direct upload, so no manual bulk upload step or anything like that. These are quite useful tools, but they don't solve the problem that most manufacturers are actually having here. These machine-to-machine tools automate the last 5%, okay? And that's the actual upload, so getting the data into UdiMed.

what this assumes, however, is that the… the data that you're uploading is already correct, it's structured, and it's schema valid. The problem isn't the upload itself, it's really all of the steps that come to that point, where you have something that's correct, to upload. So, when you hear UdiMed Automation, it's really worth asking which step of the 4 that it's automating. Utoprep currently operates in steps 1, 2, and 3, and we are moving on to step four next. Machine-to-machine tools live entirely in step four, and they stop there. These are not substitutes, you do still need to go through steps 1 through 3.

If your data is right, and the machine-to-machine tool is absolutely fine. Manual bulk uploading is fine too. Both of these work, but if your data is wrong, neither of these will help you. So, step four is, observationally, the one that I'm seeing a lot of consultants scoping for, and frankly, it's actually the easiest of the four, if the data, as I mentioned, going in is correct. Utimed, publishes their XSD schema, the XML, structured data schema there. Business rules are documented, validation is mechanical, okay? If you've done everything correctly from steps 1, 2, 3, then generating the file that needs to be uploaded into Udamed.

is… is deterministic. If it's wrong, you will be able to generate that XML file, but now you have to go back and figure out why, or which part of the previous step has broken it. if you've never seen an XML document, this is what a typical, file, XML file is that you are uploading into into Udamed, okay? Again, IT background, I can read this. I've had my head in this data for months and months and months. The problem is, though, however, the way that this is structured, as you sort of get further down into these, into these files and so forth. the codes themselves that they use for the internal naming structure doesn't actually match, a human-readable thing. My point being is that,

this is, generally speaking, non-human readable. And the problem with any submissions that you're doing inside of Udamed, if you get to the end of it and you're trying to go back and edit a submission that's already in there, because perhaps you made a typo, or you matched it to a different code and you'd like to update it. you have to download this file, you'll have to find some way to edit this file, to then go in and re-upload a brand new version of it. it does become a stack of rework with this, and to be clear, you know, UdiMed is the standard moving forward. This is an ongoing thing. This is not just a one and done.

Now, UdaPrep does support field-level editing on submitted records, okay? So, if you have gone through and created a whole submission through the Udeprep tool, you will be able to go through and make adjustments and so forth without having to redo the entire submission. The other way that we handle XML, with Udeprep is that we just lock the exports, as I mentioned before. So before you are able to produce a document to… to submit, to Utimed, we run a pre-export validation gate. We check the structure, we check that it… that it maps against the business rules. all the mandatory fields that are in there for your devices, class, and legislation, that they… they are complete. Cross-field consistency. All of this is deterministic, okay? So there's nothing… there's no AI in here at all.

This is red light, green light. If the fail… if there is a failure in it, then we will tell you exactly which business rule it contradicts, and then you can click on this, and you'll be able to jump back over and make that correction, or at least, highlight where you need to go to get the answers for this as well. So, step 4 is really the visible step, and we make this step easy by getting steps 1 through 3 with your submission right. One thing I do want to note on this as well, too, before we wrap up this category piece. So, Utimed's own validation is structural, and it's rule-based. It checks that the XML file is well-formed, and that the codes exist.

What it's not designed to catch, and this is a really important point, is a regulatorily wrong code, okay? So the code goes in, the… The submission accepts it, but then, you know, 6 months, 12 months, 24 months, whatever it is down the track. usually at a notified body review or something like that. Someone will read it and they'll ask, you know, why does this device have this code on it? This can be quite an expensive or, you know, quite an impactful mistake, because not just for that device in and to itself, but then if you've used this logic across a portfolio of devices. Now, everything in the catalog needs to be adjusted with that. This is really an asymmetrical cost that most teams aren't pricing in when they're deciding on how thorough to be in the first pass with this.

And this is the part that bites later with that. So, Again, I think just to round out the category piece here before we get to the offer, most of the tools that are currently out there on the market today, the tools that you've probably heard of, are Step 4 tools, okay? So these are for direct submission, bulk uploads, and the plumbing that's associated with that. You still have to do the work, through steps 1, 2, and 3. The other alternative is that in-house builds will solve the version of the problem that you can imagine to date, but if nobody is going through and making sure that the EC updates are there, this post-submission cost of maintenance is what's gonna get you with that.

Utiprep, as I said, we handle steps 1, 2, and 3, with Step 4 coming as the natural endpoint. We keep operating after the submission, because UtiMed isn't a thing that you do once. It's a thing that you maintain. We have a diverse, roadmap for features which are deploying, on Utoprep to make your life easy. The tool is live today, it's operational, so creating accounts, persistent submissions, you've got your portfolio inside of there. All of your audit logs, provenance, everything is inside of there. You can make an entire submission inside of Udeprep. download the XML file and have that ready to go in a validated format to make your submission into UdaMed easy. We have another series of features on the roadmap, which I'm quite excited about, which is going to be, you know.

differing against what's in the Utimed database, so as things change, you'll be able to be notified of the regulatory changes for submissions that you've already made. So we can flag, giving you an email that, is gonna notify you that you perhaps have to go back and check submissions that are inside of there. We're adding multi-SRN for consultancies, audit packets for, notified body visits, translation memory, monitoring against the public database as well. We're looking at doing some RIM platform integration, and M2M submissions, likely, further out, after a liability review on what that sort of looks like.

Now, a quick personal moment, if I may, before we get into the offer. So, I mentioned… a little bit myself… about myself, when we got started here. But I want to sort of share why… why this tool exists in the first place. A few months ago, I watched my wife, who works at a smaller medical device manufacturer. Now, she's not the PRRC, but she sat with the RA lead and two other people for an hour and a half while they were going through and doing a single UtiMed submission. So, 4 people. 90 minutes, right? And by the end of it, they weren't sure that they'd gotten it right. Now. what really stuck out to me is that you've got 4 smart people, and this is not the bottleneck. The process of doing this is the bottleneck.

They're sitting here doing the data shoveling, finding the right sentence, looking up the right codes, checking which version of the IFU is current, and they're doing this, you know, by hand, in real time, in a meeting. The judgment behind all of this is fine, but the work surrounding the judgment is awful. Now, this is the part that I really want to land before the offer, right? Utoprep doesn't replace anybody in this room, whether you're an RA consultant, whether you're a medical device manufacturer yourself. The regulatory call on everything that's matched as part of your submission is yours. You can approve it, you can edit it, reject it, it's your judgment on every item.

What we're doing is deleting this data shoveling step, that's eating your day, so you can get to the point where essentially going through and making the right calls is more or less a tick and flick, through the suggested codes, instead of having to go through all of these files and figure out which one matches to which, right? our matching engine will suggest what we believe to be the most appropriate code for this particular part of your document, and then we have yourself in the loop to review these suggestions. It's an architecture behind that, it's not a feature. Everything that comes, you know, down the track from that sits on top of these two commitments.

So, if this does sound interesting, and it does sound like something that you might be, might be liking to use for your business or for your clients, we do have a special offer for you to preprep, today. We're currently offering this at 399 euros as an unlimited tool until the deadline of the 28th of May here. What that means is that whether you've got 1 submission, 10 submissions, 100 submissions with that, you'll be able to use Utoprep, for the full extent of that, as much as you want between now and the 28th of May. We, after the 28th of May, if you would like to stay on as a… as a Utaprep customer, we'll have monthly subscription fees available depending on the amount of submissions that you need to do. The math, I think, here is quite simple.

For many of you here, you know, one to two billable hours will cover this. The rest is margin, right? You've got unlimited use of the tool for the next 27 days. A single unregistered manufacturer in your client list, if you're a consultant, will pay this back 10 times over on the first engagement. If you're in-house, this is the time that you're going to get back By spending on a single complex device. My bet here is simple, is that you can use this during the crunch. You'll see the leverage, and then after May has finished, you'll want to stay on, because you'll be slower without it. If it isn't earning its place by then, then you're welcome to cancel.

If you would like to do this, the offer is available now, at uniprep.com forward slash pricing. I'll just drop the URL here into the chat log, if you want to click on that. And we've also got a QR code, there to make it easy for you to sign up on there as well. Otherwise.

I can just say me, me, me in the chat, and that will work.

And that will work as well, absolutely, yeah. But that's the, that's the end of the, that's the end of the training that I have today. We've got some time to open up for, for questions. I'm also happy to, Happy to show you through what a submission looks like as well, depending on the timeline that you've got left for the session, Sean, because I know you've got some stuff that you need to get to as well.

So, Sam, this is Dan. I wanted to ask you, we know we talk about devices, but there is a second term that is now used for a lot of products, we call them combination drug device products. And this is extremely important, because if the device component is, let's say, the dominant one, I think that would… this process would be applicable to them. Am I right or wrong? And if I'm right, how does your process Provide the same level of, let's say, support for that type of, you know, registration.

Wow, what a hard question. Thanks, Daniel. Absolutely fine.

Thanks, Daniel. Look, as it… as the Uniprep tool is designed today, we… we support the, basic UDIDIs, UDIDIs. MDRs and IVDR device categories inside of here. So… the… the way in which the software was put together was from the ground up. So we've essentially reverse-engineered the entire EC schema behind this. This specific, device category, so the combination devices, if it is something which you're currently submitting into Udamed, then Udeprep will support that. The matching engine is designed to work with it. It's the complete library database of how

The EC is currently doing this, so Utoprep will support you on that. That said, if you are using the tool, and you find that you would like, you know. some more feedback, or for it to be, you know, changed in some way to make those submissions easier for you as well. We are in an early launch of the tool at the moment. It's officially live and public as of this week. So we are constantly going through and wanting to get user feedback, et cetera, to refine the matching engine and the workflow that's in there for professionals such as yourself.

Thanks, Sam.

To sign up, I'm gonna interrupt for a moment, pass the mic over to Jose, because I do want to make sure that you all see the latest updates to MLVX before we drop. I'm gonna forward the questions in the QA to Sam. Sam, I'd like you to reach out to the people individually. Fantastic. You can stay on a couple minutes after, Jose, but let's switch over to Jose if he's with us. Are you with us, Jose?

Hey, Sean, yeah, how you doing? Yeah. Good to see everybody. Yeah, I'm just… Quick update, it hasn't been a little… It's been a little…

Close your camera.

Well, and I think Michelle…

up. Close your camera.

Kind of late, kind of delay closing my camera? Oh. Is that better?

Yes.

Yeah, it's working, yeah.

One second. Okay, alright. Yeah, so now, you know, the original list of the MLV100 was selected by Sean. What we've done is we've now allowed you to vote on them. Sean, I don't think you explained this at the beginning, what's the point of the voting for the MLV100, other than the pride of showing that you've got the most votes?

So my idea was that we would then create certain top 10 lists within the MLV100, basically, like, up-and-comers, you know, audience favorites, things like that. But we use, the data to kind of look at it differently. It's like, okay, these are the people that I picked. And these are, you know, and why I picked them. and why I follow them, but who do you follow, and what can we learn about them?

Okay, great. So you're gonna get some free exposure and free, free, kudos and marketing from Sean, who's got the largest list of medical device companies out there. So that's a pretty good reason to try to get your friends to come in here and vote for you. And then, of course, the other big thing is, if there's somebody who's not on the list, who you think should be on the list, who push, you know, somebody can nominate Sam over here. You know, clearly he's in the space, he's got good stuff to share, so put it on there, vote, nominate, and you'll be able to see, kind of, in real time. Obviously, this is its launch, so not a lot of votes, but a couple people have votes already. So that's the big thing, and then, you know, if you haven't heard of this before, what we're doing in this platform, it's a place for you to see who the experts in the field are. Many of them have been part of the MLV100, but some of them are not. You can sort of search by specialty, so, you know, if you look at cybersecurity.

which is our main area of expertise, you would find us there, and other people, you know, digital health, etc. And then, kind of a cool feature here is the Ask AI, where you can ask things like, hey, what regulatory code does electrosurgical device over-the-counter aesthetic fall under? And it's connected to the FDA, all of the FDA databases, or your mod database, your classification database, etc. And it'll show you that kind of information, it'll show you on the reference, like, where that came from, and if you ask follow-up questions where there's content from the experts on the platform that is relevant to the thing that you're asking about, then it would… it would bring up that kind of information, like, I don't know, if I look up at this.

this cybersecurity thing I asked about. you would see that I come up here, because there's a piece of content that it pulled from, from the database, and it said, like, here's some important information, and it would link to that, and then you can click on that profile, and you can see who the expert is, and you know, what other content they've got, you know, what their contact information is, what their LinkedIn, is, you know, CTAs, like, you can click on here and get, like, a link to our Cybersecurity quick start guide, etc. So, if you haven't tried it out, please check it out. There's a free tier, there's paid tiers, and if you're an expert, there's an ability to do all kinds of fun stuff on it. So, hope you check it out.

Yeah, and if you want to, edit your profile if you're listed, or, Add somebody, and if you… if you get added, then, that's all editable, you just need to claim your profile. Yeah. Right?

You can even, have it help you write LinkedIn posts and stuff. That's, anchored on regulatory guidances and your own… your own content. So, a lot of fun features coming.

Thank you. I saw Nellie voted. Thanks, Nellie. Thanks. Well, I appreciate everyone's time. It's, 12.01. Now's the time where you can unmute your mics and… scream at me, or whatever you want. And I want to thank Sam again for coming and sharing his platform with us. tell us who introduced us? It was… it was Martin and… and… oh…

And Dina, yeah.

Yeah, right, from…

From CreateRA, and Rayana, yeah.

Yeah, Rayana. Yeah. built a really cool platform, check that out, too. And, yeah, thanks, everyone, for your time. And.

Thanks so much for having me.

Visiting you next week for, what is it, First in Human Trials in Latin America. Yeah? Great. Thank you. Thank you again, Sam.

Thanks, thanks so much, have a great weekend.

Everything to you. Bye.