Transcript
Speaker: Welcome to Clinical Data Talks, a podcast brought to you by CRS-Cube. I'm your host, Sylvain Bertelot. Join me and industry experts as we discuss the latest trends impacting the world of clinical data.
Speaker: It's a real pleasure to welcome Chris Decker to the podcast today. He's the president and CEO of Cdisc and in his role, he helps shape the global standards that underpin how clinical trial data is collected, structured and shared across the industry.
Speaker: With decades of hands-on experience, both across clinical technology and data solutions, Chris brings a rare mix of technical depth and strategic strategic leadership.
Speaker: I'm really looking forward to today's discussion and to hearing Chris's perspective on the role of data in today's clinical research. Hi Chris, thank you for joining me. How are you doing?
Speaker: I'm doing great, Sylvain, and thank you for thank you for having me. I'm always, as my wife would tell you, I'm always ah i'm always willing to talk a lot. So I'm happy to be here and and and share some thoughts and just have a conversation. And by the way, I appreciate CRSQ being a member of CDIS. So thank you for your support as well.
Speaker: Well, i think the support is mutual and we really appreciate everything CDisk is doing. and about you loving talking, the flow will be mostly yours, so ah I'm sure you'll enjoy it.
Speaker: um So we'll get to talk about CDISC and what CDISC does. But before we go into that, I'm really curious about what has led you to being the president and and CEO of the of the organization today?
Speaker: A great question. And that's probably a bit of a long story, but I'll try not to spend the entire pa podcast talking about it. So I like to call myself a a recovering statistician.
Speaker: So i got I got several degrees in university and statistics. I got into the area of statistics and just really enjoyed the numbers and and the decisions and and the and the hypothesis that you could generate in that area.
Speaker: But then I started my career working in in pharmaceutical companies. So I've been in clinical research my entire career. and I started at what was at the time Burroughs Welcome, so I'm going back several companies in the early 90s as a statistician and a programmer.
Speaker: But what I quickly realized in that in that initial start was that I really didn't like statistics at all. I much more liked how do we innovate? How do we leverage data better?
Speaker: How do we make ourselves more efficient? So I remember generating three weeks into my job, generating tables and listings and thinking there's got to be a better way to do this than what we do today.
Speaker: and so early in my career, I was involved in every opportunity to leverage technology to optimize the work we were doing. And I realized really early on, even before C-DISC came onto the scene, that standardization of data was was critical. and I was involved in an early project before C-DISC when I was at Galacto at the time.
Speaker: which was looking at how do we bring 12 or 14 studies together and get and get hypotheses and decisions off of those studies and I ran a project at the time standardizing that information and I realized really quickly the challenge of that and why it was so important to do that so from a CEDIS perspective I got involved with CEDISC early on in the mid-2000s helping to lead and co-lead a pilot project with CEDISC we were looking at integrated submissions if you remember the regulation didn't come forward until 2014 so we were in those early days of just proving it out well fast forward 10 years i had the opportunity to join the cdisc board in 2018 it it really opened my eyes at the time i had used standards and i'd used the cdc standards i've been involved as a volunteer but it opened my eyes to both the challenges and the opportunities that a standards organization has
Speaker: to try to improve clinical research, to try to bring together the audience and the different perspectives. That's what I've found in this role is there's so many different views and so many different perspectives on the right way to approach this.
Speaker: So in 2023, Dave Evans, who was leading the organization, came to the board and said he wanted to retire. He was moving on to that next stage of his life. And he called me up and he said, you know, Chris, you should do this. You should take over this role.
Speaker: Because you really are thinking about how we can innovate, how we can sort of change and evolve the organization. And so I, I pondered a little bit and I went and asked my wife, I said, this is an interesting idea. And she said to me immediately, she said to me, you've been trying to fix the industry for 30 years. Why would you take this opportunity to do it from a place that maybe you can change things and influence a little bit?
Speaker: So as we all do, I listened wisely to, um, my better half and, um, decided to step into this role a few years ago. So beginning of last year and, uh, with the idea that, um, we were gonna try to innovate. We were gonna try to push things in a different direction.
Speaker: Yeah. So that's sort of a little bit of that background, the shorter version of it. Nice. Nice. And you've you've mentioned a few things that I'm sure we'll go back on, especially when you mentioned your initial experience at GSK.
Speaker: I was thinking, well, we we we still have the same problems or the same needs nowadays. so I'm sure we'll go back to that. But before we do, um and based maybe on your position at CDISC, could you tell us what role you see CDISC playing in the clinical trial industry?
Speaker: I think CDISC is an opportunity. We have an opportunity to partner with not only the industry, but with other groups to rethink the way that we're running our processes and the way that we're defining and using the standards.
Speaker: within our data life cycle. So we have a core foundation of standardized content. That's great. We have regulators who accept and understand those standards. That's very good. And we have significant expertise in implementing the standards and the content.
Speaker: But now I think we have to kind of shift the way we're thinking. And I think CEDIS can help drive this. We need industry's input. CEDIS can't do this by ourselves. But we have to break down the silos. We have to add semantics into our standards that really enable the connectivity from the beginning of of of the process, the study design, all the way through to the analysis.
Speaker: So today we look at standards in these silos, right? We have We have protocol, which isn't really a standard yet. Now we have, we have USDM and we have other things that we'll talk about that. I think later on the call or later on the podcast, but we have that and we have data collection standards. Then we have this, and again, we have this manual handoff. We have to break that down and we have to look across those. And I think where CDIS would like to help, and we need to do this in partnership with other organizations as well, as we think about real world data and other types of content.
Speaker: We need to break that down and really drive that transformation, that change in the standards from starting at your digital protocol and your study design all the way through. So that's that's where I think CEDIS can help to try to drive that that change that that that we have.
Speaker: It's not a small undertaking, but I think it's something that we can we can definitely help to do. No, it's definitely not a small undertaking, especially in our industry. So you shared a bit of history. Cdisc is 25 years old. I think we saw the first EDCs more than three decades ago, but we're still talking about digital transformation, although EDC, I'm pretty sure initially was coined as a way to embrace digital transformation.
Speaker: Do you know, ah do you have a feel for why we're still talking about digital transformation nowadays? Yeah, it's funny. my My short answer is we didn't know what it was 20 years ago or 25 years ago. Like when we when we use the phrase digital transformation, it's almost like the the Henry Ford quote, which I always love. I love my quote, so i apologize if I use too many.
Speaker: But, you know, he always said if he would have asked people what they wanted, they would have said a faster course. right, not a car. And I think we didn't really understand, but nor did i do I think we had necessarily the technology robustness that we have today.
Speaker: So I recently heard a quote at our conference this past, um was the fall or in the fall? There was a quote by a vendor that was presenting at the conference, an AI vendor, and they said, ai can't fix a broken process.
Speaker: and i'm bringing ai here not to talk about ai but just to reference the fact that we have processes that in my opinion are our legacy and broken is a harsh word but their processes these linear process that we use from beginning to end with these siloed groups i think is is is a challenge right and we have to break that down i've said it before but we we've allowed this linear manual and siloed process to dictate are our incremental changes. We made these tiny tweaks to our process, our changes, and we haven't fundamentally looked at the way that we do things.
Speaker: We thought EDC, to your point, decades ago, right we thought EDC was transformative in the late 90s. But let' let's be honest, right? In reality, All we did was move from ah from ah from paper to electronic paper.
Speaker: right We just moved it out of of paper with your pen and your pencil into, into and it really didn't fundamentally change the process. So patient and clinical research data to me is fundamentally complex as the relationships in a patient's data are not linear or two-dimensional.
Speaker: right there's relationships throughout a patient's data and and no offense to our edc vendors and and i think they would probably agree with us for the most part but that the way we collect the data and in terms of of edc we lose its context right we we collect this data we put it into forms and then we disconnect it and and we lose that context so by the time the data comes out to us downstream we've lost its meaning And so we almost have to try to put the meaning back together.
Speaker: So when we talk about biomedical concepts at CDISC, by the time that data comes from EDC, um we again, we've lost that context. We've lost the meaning. and We have to kind of put it back together. We need to keep the meaning of the data at the very beginning and keep it together.
Speaker: You know, and as I mentioned at the beginning of this question, right? I also believe our technology wasn't robust enough 20 years ago, and it didn't force us to think about the data life cycle differently.
Speaker: As you know, in the last five years, right, explosion of AI. It's hard to believe ChatQPT didn't come on the onto the scene until it was 2022, I think, or 21. 21 or 22 is when it really came.
Speaker: That's only three years ago. yeah And look what's happened in that regard. So I think the explosion of technology and AI where we could avoid changing in the past, we can't avoid changing now because either we're going to change or technology is going to make us change.
Speaker: Right. It's going to force us. It's going to run us over and and and make us change the way we're thinking about it. So that's kind of the reason basically to make for your main long answer there, but is that 20 years ago, we didn't really know what digital transformation was. We weren't forced to change and now technology and the need to get insights out of our data much earlier in the process is going to require us to to change as we go forward. So we're at that point now where we're going to, it's going to happen because we're going to be forced.
Speaker: forced on them Yeah, I agree, fully agree. and So there are plenty of initiatives, well, or various initiatives, at least ongoing. ah I'm aware of the initiatives between CDISC and Transerate, for example. You've mentioned the USDM earlier on. ah Those initiatives have been looking at digitalizing protocols.
Speaker: Why is it so important to start with the protocol?
Speaker: You know, it's it's interesting. If you think about the CETA standards, they kind of got dropped to my earlier comment about EDC and how we lose that context and meaning, right? The CETA standards kind of got dropped in the middle of the process, if you think about it. We had to gen we had to generate ah tabulation data and then analysis data to submit for regulatory authorities.
Speaker: And we dropped it right in the middle. And so what happens is by not looking upstream, at the the protocol and that information, you've kind of broken the process from the beginning.
Speaker: And so we've made these small changes along downstream when we think about the data flow and we're trying to innovate, we're trying to leverage technologies and improve data quality. We're trying to automate, we're trying to show efficiency.
Speaker: But when we still live in a Word document that has to be interpreted by people, and manually turned into information downstream, it just kind of, to be honest, blows my mind that we're still there, right, in that world.
Speaker: And recently a study showed, it was last year actually, some a pharma company shared this with me, and i i'll I'll leave them anonymous, that 50% of a protocol is manually and incorrectly transcribed somewhere downstream.
Speaker: And that a piece of information in a protocol is manually copied on average 20 times. So the reality is that if if if we're treating a protocol like that as a Word document that we're trying to interpret and and and do things downstream, we're going to make errors and we're going to continue to have this challenge in this in this in this issue.
Speaker: And really where I think we need to start, I even wonder sometimes about the word protocol. right Protocol has its connotations of being a document. So should we start calling it something different like digital study design?
Speaker: But then someone reminded me that we have a tendency in our world when we innovate to still call things by legacy names to help people adopt things. yeah I'll use the word horsepower.
Speaker: Right? Why do we call an engine in a car? Because we wanted people to adopt it. We wanted people to accept it. And so I think as we move forward, our focus really needs to be on what are we trying to start What's the endpoint we're trying to achieve in a trial?
Speaker: And how does that endpoint drive everything else downstream? So the endpoints in the file must must define the definition of what we need to collect, what we need to analyze, and what we need to capture as we go through a a process. And I think the digital protocol can help put that framework around that to help us say, start at the end point, define the schedule of activities, and just have smarter trials.
Speaker: So that's why defining that information up front from a protocol perspective and in a digital way is so important because it can then drive everything else. And today, we don't do that.
Speaker: We manually read a document, we interpret it, and then we try to drive it. So it's little backwards. Yeah. and And we start with inefficiencies upfront.
Speaker: From the very beginning, you're spot on. Yeah, yeah no that makes sense. and And from what I see, i mean, there's so much we can do actually from a digital protocol that would really help at least the the study setup phase that would bring much more efficiencies than than what we can, I'm sure, think of at the moment.
Speaker: um Now I want to to shift to 360i. So earlier um this year, at the time we record, Cdisk launched the 360i initiative.
Speaker: um and I understand that it's to create more connected and interoperable standards. Could you tell us a bit more about what you're aiming to achieve with three six ti Yeah, absolutely, Sylvain. And I'll give, again, just a little bit of context and background. So, you know, back in 2018, I think, when I joined the board, there was an initiative underway called the Blue Ribbon Commission.
Speaker: And the Blue Ribbon Commission at the time was to bring a bunch of industry leaders together and define sort of where we were with the standards and see this and where we needed to go. And a lot of feedback out of that was to make the standards um more interoperable, more something we can use in our process. Like I said in the very beginning, right, initially it wasn't necessarily designed with that in mind. So there was a lot of that feedback.
Speaker: So we started a project and in just in the beginning of end of 2019 called 360, which really was us exploring Some new areas we explore, we were starting to explore biomedical concepts and how do we define our standards more granularly than we do today.
Speaker: The protocol model, we would just started to work with Transcelerate and we were looking at how we could enable a digital protocol. Can we look at analysis results and can we standardize how we generate analysis within our industry. So this was more of a, I would say proof of concept exercise than it was full fledged, let's change.
Speaker: And so we proved out a lot of things. We found a lot of gaps. We didn't achieve everything we wanted to at the time, but what we proved was that there was value to the community and us striving in a different direction.
Speaker: So fast forward, when I came on board at the beginning of of last year, we had published an analysis results standard. We had partnered with Transcelerate to develop USDM, right, and the protocol model. And we had published over a thousand biomedical concepts to try to drive this forward.
Speaker: But we were still missing sort of those key components of how do we connect from end to end and how do we operationalize that so we can do this as an academic exercise, but how do we actually make it real? How do we make people so they can it to people so they can use it So our goal with 360i, and the i actually comes from implementation, is to change the way we develop and use the standards and to take our good content and add the additional semantics in there that are needed to make it interoperable.
Speaker: And so what we we did just in the first six months is we kind of showed how we can leverage the digital protocol to actually, and link it to our biomedical concepts to drive the auto generation of things like but defined xml the the eCRS,
Speaker: the that might be used for a variety of different data sources. um The ability to generate, auto-generate things like SDTM, right, leveraging these these different components.
Speaker: So what we also found in that exercise over the first six months is that we have missing metadata, and a lot of it's operational metadata. How do you drive this process as you're going through that linking of information?
Speaker: So what we want to be able to do at the end the day is be able to start to publish um i guess what we're calling is you know digital studies digital examples of studies that are taking our therapeutic area standards and actually digitizing those so one of our goals is to say can we take breast cancer for example and take our analog um you know pdf version of our our standard and actually digitize that information to find the endpoint and show it literally and publish it so that people can just pick it off the shelf and use it within their environment, within their their processes. Now, it does fundamentally change the processes within companies.
Speaker: yeah So that's one of the things that we, as we work with the community, how are they going to use this and how are they going consume it? So this year was about kind of proving out that we could do that. Now we have to actually start to operationalize. How can we work with with community to say, here's how you can use it in your in your world.
Speaker: So that's kind of where we're heading. It's it's fundamentally changing. And we've told the teams, you know, internal standards teams that we have, is that the days of publishing 400 page, you know, IGs, we're going to go away from that.
Speaker: And it's a change. It's a change for people because a lot of people have been, a lot of great parts of our community have developed exceptional content over the last 25 But we have to we have to move. We have to change.
Speaker: We have to evolve those standards. So that's that's our goal. And i get into more we get them more I get into more weeds, you know, not in a podcast. I spend hours on some of those things. ah valid and I like that because what your it's much more practical, that's what I'm hearing, and um less room for interpretation, which is what we've been struggling with for years or decades in the industry. When you read ah regulations, recommendations, there's always room for interpretation. and If you manage to remove this from what you're doing at CDISC and really share practical examples of how to implement standards, then it's much easier to implement it. Exactly.
Speaker: Yep. So um we have to talk about AI. Of course. Everyone does, right? um But more seriously, because I've heard many things about AI, about standards, how they're meant to work better together. So from your point of view, how do you see standards and AI working together?
Speaker: um and do you think with standards, AI can be more useful? Well, I could answer this quickly and just say yes, and we can move on. But but let me expand a little bit.
Speaker: So, you know, it's interesting. i When I came into this role, I had been playing around with AI, you know, my own and things like that. But I hadn't i hadn't really jumped into it, per se.
Speaker: And so I was at a meeting early in my um tenure at CDIS last year in 2024. I think it was a DIA meeting. And we were in a future of standards session.
Speaker: And one of the moderators asked the question, they were trying to challenge the audience a little bit. They said, if we have AI, we don't need standards, right? You know, standards can, we we have AI, AI is going to solve that.
Speaker: And the reality is, you know, what I've kind of learned and in my own research is that not only do we need standards, we actually need more of a robust standards for AI to really effectively work.
Speaker: Good data leads to good AI. Bad data leads to bad AI. I mean, that's just, that's just a fact. And everybody is now experiencing that. And we're kind of over that. won't say we're over, we're kind of on that hype curve now where we're coming down off. If you look at the latest hype curve, it's like, we're coming down off when we realize what we can use AI for and maybe what it's not always as as good to use it for.
Speaker: and so as we think about standards and AI, we need and at C-Disc and in the community to leverage AI to better help us develop the standards. So this past summer, we ran an AI innovation challenge at C-Desk, and we had three fairly complex use cases.
Speaker: And it was really interesting to watch how the industry tackled these and in put forward solutions. And I think what we learned in that, and then in follow-up conversations that we had with several of these companies, especially ones that really focus on on AI, and I know you guys do a lot of work in this space as well,
Speaker: Um, when I asked them and they all said, oh yeah, we do okay with the data we have today. And we get, you know, 50% of the way there, 60% of the way. But if man, if if we could have structured underlying standards and in and just better structure and so to our content, we could do so much more. Like we could really explode in our ability to do this.
Speaker: So I think we really need to think about AI as a, an extraordinarily powerful enabler. Don't get me wrong. But it's an enabler, and you need to give it the information to make it do the things you really want it to do well.
Speaker: And so I think we need to use it within Cetus to help us develop the content more quickly. But then once we have the content, enabling the industry to use our content to do that. So I'll give protocols as a great example.
Speaker: right I've heard people say, we're just using AI, it's reading our own protocols and generating our stuff in 15 minutes. And then I said, oh, that's interesting. Tell me more. And they said, well, we're we're we're feeding it our well-structured protocols. you know We're not feeding it our are protocols that are not structured well.
Speaker: So, well, that's a standard. right youre're You're feeding it structured protocols. And so I think there's an example where we can use AI to help us generate from legacy protocols and generate a standard database of, of for example, USDM information, which is our protocol model.
Speaker: And then based on that that repository of information, be able to make much smarter decisions on the design of trials. So again, we can use AI to help us generate and then use AI to help us actually, once we have that content and that structured content,
Speaker: to be able to get a lot more out of it. So it's sort of that that development of standards and that use of standards where we need to be able to use it more and more. And at CDISC, we're developing, you know, Julie Smiley, who's our head of our data science group, is developing some strategy around how we're going to leverage AI at CDISC in the new year based on some of the work that we've done this year with various various vendors.
Speaker: So I think they go handin hand in hand. At the end of the they have to go handin hand in hand to really realize that the value on both sides. Amazing. i have one last question for you because you haven't shared enough. ah I want one last ah piece of advice for you. um What's the best piece of advice that you've received in your career that you consistently apply in your work?
Speaker: So Sylvain, you prepped me for this and I still i still like struggle to come up with one. um so I have two. because i couldn't I couldn't come up with one. um And it's funny, as you as you you know your career gets longer and longer and you get more seasoned, you have these moments, right, where you look back and and you think about what really drove you. And I think, so I'll do two. One is um a wrong choice is better than no choice.
Speaker: see So I think we have a tendency, don't know the phrase, deer in headlights or we have a tendency to get frozen and we don't, we don't act.
Speaker: Right. And I've always said in this industry, this industry is very risk adverse. As you know, this industry is difficult for us to make progress sometimes and change. yeah And so to me, if you have to make a choice and you have to move forward in something, even if you make a wrong choice and you have to correct it later, at least that's better than no choice.
Speaker: So as an industry, I feel like we have to, we have to, We know we're going to fail sometimes, and the industry does not like to fail. But we have to know we're going to fail sometimes, but we have to make choices, and we have to we have to continue to iterate and and try to find the right solution as we go forward.
Speaker: So that's one. The other one is is one that's been a challenge for me my whole career, which is listen to understand versus listen to respond. So a lot of times when we're in a conversation, we're immediately hearing the other person and saying, oh, I got an answer for that.
Speaker: I got a response to that. Instead of waiting and listening and really understanding what they're trying to convey. And as we as an industry start to collaborate more, because we're going to have to, right? We're going to have to work together more.
Speaker: We need to be able to come to the table and take time to understand the different perspectives, whether they be sponsors or regulatory or technology companies or a standards organization. and understand each other more if we're going to come to a decision on certain things as we move forward. And I think we have a tendency as different sectors and different individuals to kind of um dig in and not want to change. Right.
Speaker: And so I think that we need to do a better job. and um I've been me personally is tough for me is to listen to understand the other person and really try to understand where they're going before we kind of collectively have a response.
Speaker: So those are two things that always resonated with me oh that that, you know, as we push forward in what we're trying to accomplish in the industry. that I think we need to think about. So I had many more. I had to reduce it down the two.
Speaker: and we have thoughts no amazing I Amazing. I love the last one, especially. um And it's interesting because if you if you don't if you listen to respond in a way you're not ready to be challenged in your thinking, whereas if you should listen to understand, then naturally you your thinking will evolve.
Speaker: um and And I love that you picked this one as a piece of advice, because i imagine that in your role um and as an organization at CDISC, you need everyone to change a little and meet you in the middle.
Speaker: Otherwise, you can't you can't be effective in what you're trying to do. You know, absolutely. And you know, what people ask me after my first year or two, what's the biggest thing, right? What's the biggest thing you encountered right in this role? And yeah, after years and years of being in the consulting business and working with big pharma and dealing with stakeholders at at companies, right?
Speaker: When you dealt with a specific pharmaceutical or a big sponsor, you'd have three, maybe four stakeholders that you were trying to influence. And like you said, bring them together in a decision. In this world, in this role at CETA,
Speaker: the depth and breadth of those stakeholders is just tremendously bigger. I say it's actually much more harder to do. But also that that's exciting for me because it's an opportunity, it's a challenge to try to, as an organization, to your point, bring those people together and all their different views and see if we can come out the other side with agreement, right? Or at least another phrase to throw in here is I always like to disagree and commit.
Speaker: So you're always not going to agree all the time on everything. But the things you don't agree on, can we commit to a path forward? yeah right And so that's what my hope is as I work with the community is maybe you don't agree with everything you know that we're doing or we're trying to move forward, but can you can you commit to it?
Speaker: Can you say, hey, okay, I i see where you're going. And that's the listen to understand part. but is can we Can we come together and agree on a way forward and commit to it? That's what my old boss taught me was that disagree and commit, which I thought was always good. Trying to get consensus all the time doesn't work.
Speaker: No, no. yeah no So agree with your point. Well, thank you very much, Chris. It's been amazing talking to you. And I really love what what's happening in the industry and the role CDisk is playing in that.
Speaker: but I appreciate it. look forward to your continued, you know, you and and your team and continued support. And, um, I've always willing to, you know, share, cause I think we need to move this, move this forward. So I'm excited to where we're going and, um, we'll see where we get yeah the goal.
Speaker: Yep. ye And I'm sure the goal will move for the time today and appreciate contributing to to the discussion. No, thank you. And thanks everyone for listening and watching us.
Speaker: ah You can find more episodes of Clinical Data Talks on our website.
