Jack O’Brien speaks with Michael Vande Vrede, head of digital products at EMD Digital, a Merck KGaA company. Vande Vrede discussed Syntropy’s collaboration with Evidium to address challenges involving low-quality data that are slowing research progress. 

Follow us: @mmmnews

To read more of the most timely, balanced and original reporting in medical marketing, subscribe here.

Note: The MM+M Podcast uses speech-recognition software to generate transcripts, which may contain errors. Please use the transcript as a tool but check the corresponding audio before quoting the podcast.

From HLTH 2023 in Las Vegas, Nevada. It’s the MM+M Podcast.

Hi and welcome. I am Jack O’Brien. I’m the digital editor. Mmm. We are here from the health conference in lovely Las Vegas. I’m Julie today by a very special guest. Hi, this is Michael vandeverede. I’m with centripee. I want to start off the conversation obviously talking about the latest news that came out of century with the Partnership. If you can give our audience a little bit of an insight into how that all came about and what it’s going to be. Yes exciting. So we announced a partnership with

AI company called avidium a few weeks ago and just a quick recap on what century is we’re a company that helps Health Systems make their data more useful for improving patient care and doing better research and I’ve sort of been long-term AI cynic in this space due to I think the regulatory nature of healthcare and so while we believe in the technology and it’s potential, we’ve always been sort of cautious and how we think about applying some of the AI to

Patient data specifically now that we’re in this sort of New Era with generative Ai and llms I think the stakes are even higher for us to do things in a very responsible and ethical and grounded way, which is kind of our philosophy and Mission a centropy and we found this company of idiom that has a really unique take on using clinical guidelines as sort of a referable knowledge-based data set to train their llms. And so it’s a really perfect Synergy of our ability to help integrate data contextualize that foreign organization and provide that clean data for them to then build. All of these llms on top of and the ultimate result of this hopefully will be higher quality output for

organizations, which is obviously the focus of any Healthcare Executive that I speak with is being able to take that and make it actionable. I want to go back and take a step back for a second though you talk about obviously the popularity that we’ve seen generator of AI it’s not like AI is anything new but certainly the chat GPT of it all has taken over really since around this time last year. What do you make of that kind of skyrocketed popularity not only in the healthcare industry, but kind of just in you know mainstream culture.

Yeah. I mean, I think it’s great we

Sort of democratized now access to super powerful technology. You don’t have to have a degree in machine learning or math to to access some of these powerful tools. So it’s really put into the fingertips of a lot of people in any industry really this potential power. I think in healthcare, we have a responsibility to not treat patient data, like click stream data and other Industries. And so I think we have to think about some of the the more existential questions around, you know, if we get this wrong for example and Healthcare, it can be the difference between life and death ultimately and so we have to do things in a more thoughtful way. So I’m like excited about the potential the technology, but I want to make sure everything we’re doing involves for now humans in the loop and doing things in a very grounded and responsible way. And so that’s sort of my take on where we are with llms and generative AI in general and as you look at like

some of the healthcare leaders and organizations that are experimenting the way because I feel like everyone’s dipping their toe in the water to some extent with gender of AI with machine learning our people doing in a responsible way are people kind of being more trepidation than they are.

Was saying like hey, let’s just go in we have this at our fingertips. Like you said we might as well try it.

Yeah, I think a lot of the promise right now is basically give us your dirty unstructured data. We’ll throw the llms on top of it. It will organize that and some interesting output will come out of that and that may be true but I think that the danger in some of it is we don’t know what bias necessarily is being at reduced. We don’t know what the hallucination risk will be from some of this output. So my sort of challenge for the industry and ourselves as an organization is like what if we started with cleaner more contextualized organized data as more of a referable data set, then we could really know and control more of the outputs of these llms. And so I think what I’m hearing across this, you know conference the last couple days is I think there’s a couple different perspectives. There’s the kind of move fast and Break Stuff approach which I think from people I’ve talked to most people are kind of saying that’s probably not the right approach here with healthcare because of the existential risk, I mentioned and so a little bit more thoughtfulness around how do we keep humans in the loop? How do we control the information flow and tie things back to source and Providence and making sure that

You know everything that we’re doing is done in this sort of responsible way. It’s interesting you bring that up because I’ve talked to a number of leaders who say that they’ve

seen, you know, kind of the lessons learned from the internet boom of the early 2000s and certainly the rise and I would say to a certain extent fall of social media of some of those, you know, move fast and break things the Facebook and Google approaches of it all and saying okay, we can’t do that with generative AI because it’s so much more powerful and maybe we don’t even know what potentially bring to the industry and like you said It ultimately goes to patient’s lives and they’re out. Yeah. Absolutely. Yeah,

I think as an industry, we’ve always liked technology adoption. Sometimes that’s very painful. If you think about like data integration for example, which is the area that we work on all the time. We’ve always talked to our customers and partners that there’s no magic Silver Bullet to making your data really useful and scalable. It takes a lot of investment and hard work. But if you do the work to integrate contextualize curate your data, you can get Downstream effects and there’s a compounding value over time of doing that but it’s not magic, you know, we’re kind of dancing on

Magic phase right now. Some of this stuff is really powerful. We have un we almost unlimited Computing potential we have data that’s growing exponential you combine these things together. And I mean, I think the potential future is awesome. But we just again I come back to this idea that we have to do this in a step-by-step fashion cautiously making sure that we don’t let the technology get ahead of you know, what we understand about it. Yeah

not having sort of hubris. If you will when you’re dealing with the sort of things when you look at the data conversation, I’ve had a number of conversations with leaders at this conference, but also, you know prior to coming here about the data aspect of people talk about data day to data it becomes almost kind of like, I don’t know I want to say white noise, but it’s like everyone’s going towards it when you look at that conversation, what’s maybe a misconception or misunderstanding about the data element that

you would want them to say be more proactive in this lens. So, you know, we work with a lot of Health Systems academic medical centers. And what we found is that the data collected in the course of patient care often isn’t sufficient from a clinical or scientific perspective to always answer all the downstream research questions, so,

So and that’s because when you’re collecting the data for the patient care, you’re not doing it for research. You’re doing it to help treat the patient. So how do we help those organizations make that data more of a clinical grade or research grade data asset that’s high quality there in the past pre-generative AI space. It was just hard work you had to go in you had to do all the data integration. You had to apply human and some machine curation to make that data useful, but that just took an investment of time and that was a hard investment because we believe that to get the highest quality data from these organizations. You have to have humans with contextual expert at those organizations in the loop. Like the clinicians themselves have to help sort of codify the data because they understand the context of how that unstructured clinical notes were written best. You can’t just Outsource it to somebody that has no contextual knowledge of that organization because the results not going to be high quality data. Now what we can do with some of this generative AI is we can augment a very human-centric problem because it’s never a good idea to take clinicians away from treating patients to go curate data. That’s a terrible value.

Opposition I would never you know, encourage any of our customers or Partners to make that trade. So how do you keep them in the loop without taking too much of their time away from the most important things that they’re doing which is taking care of patients. I think that’s where some of this AI space can come into play which is we can augment that human Centric problem still keep them in the loop in more of a supervisory role but not make them have to actually go and do all that that hard work themselves. It’s really interesting you talk about having the human

element be involved in it because I think a lot of people again this really kicked off last year, but I think over the past few months, there’s been a lot of conversation of like maybe we don’t fully understand the technology. It is very very powerful. But as long as we have some sort of human auditor if you will saying like from the clinical lens saying here’s the data and we’re not gonna act on it or let it just run wild without having somebody else being, you know there to double check it along the way so we’re not making mistakes that ultimately lead to you know, reduce patient outcomes or something even more dire.

Yeah, absolutely and I don’t think that we’re gonna head into a world even in the near future where machines are replacing humans. I am a huge proponent of the human creativity and

Our role as editors and curators of content and I think it’s going to become even more apparent as we move into this new era of misinformation hallucinations. Like how do we know? What’s true anymore? How will we know? What’s true? I think that’s where humans come in. We have to be the ones that are in control of this and are applying that sort of lens.

Well, they certainly couldn’t do this podcast. I think we want that on the records. The machine could never do this podcast you talked about some of the things you’ve been seeing around the conference. I’m curious what else in terms of conversations panels anything you’ve been paying attention to here at Health. Obviously. I mean, I’m looking out here just thousands of people thousands of booths like, you know, what’s caught your eye. So team last year

for me here was all around Health Equity. So that was just everywhere and this year. It’s obviously AI is here to stay for for a while. I think what I haven’t heard a lot of yet is how do we put patience in control of their information? There’s been a couple talks I’ve listened to where we’ve I’ve heard from people and I’m not gonna name names but they’ve basically said, you know, if patient data is de-identified that data,

Is no longer considered patient data anymore because it’s not attached directly to identifiable patient and we can use that for whatever we want. I think that’s that’s not correct. I think as technologists as health technologists, we have a responsibility to try to put the patient in control of their information and data as much as possible. I genuinely believe people will want to do the right thing. I think people will want to ultimately allow their data to be used for good research. But right now we’re not necessarily doing that transparently as an industry, especially in the US. It’s terrible. I mean when you go to get treated for something you sign, you know, some terms and conditions that allow the uses of your data for something Downstream that you have no idea of what you’re signing up for. I think there should be a responsible push as an industry all of us included to try to find ways where we put patients more and control of that allow them to permission the usage of their data for Downstream research. So that’s what I would like to see as

sort of the next, you know Horizon. It definitely ties in with that patient empowerment lens that I hear from a lot of different leaders, but also I imagine there’s a huge education element too where it’s like I think about you know, my uncle who you know.

He knows what I do as a living in terms of covering the healthcare industry, but let alone going to the doctor’s office you talk about signing all those paperwork and stuff and now it’s like, you know trying to inform him that can be a tall task. I imagine. It’s actually there’s a huge human factors piece here.

How do you educate people? How do you ensure that? They actually understand what they’re signing up for and that’s where the Health Equity thing comes into play. Also, you know, how do we ensure that? We can democratize and make accessible good research to everyone across the country regardless of where you live. Like if you’re living in a metropolitan area where you have access to like a leading Center, you might have that if you’re in a more rural area, you don’t you know, how do we ensure that? You know, we’re reaching out to those people as well. I think there’s a really human-centric, you know, patient-centric view that we can take care.

I’m really kind of curious. If we were to have this conversation at Health next year revisiting the partnership that you were talking about earlier. What would be the ideal World in terms of results or outcomes from that partnership? What would you like to be able to come back and say hey we did this and these are the metrics that we’ve been following. Yeah.

I would love to say that we’ve launched success.

Fully a couple projects with customers in the llm space where we’ve helped augment the humans and not replace them in the process, but we can do it in a fully traceable way where we can tie back to the actual knowledge underneath those llms and know that you know, we have controlled the inputs. We understand what the outputs are and everything’s done to sort of an open and transparent way. That would be like an ideal scenario. We have a couple exciting announcements that hopefully we’ll be able to talk about early next year, but that would be kind of a vision for me.

That’s how we do a teaser right there. That’s how you people wanting more Michael. I’ve really appreciate you being on the show. I have to ask you the question. I’ve been asking every other executive that’s kind of a hard left turn of everything. Obviously the health conference is the focus here in Vegas, but there’s so many other things in terms of concerts shopping casinos when you come to Vegas what catches your eye what’s the thing that you’re always looking forward to

I went to the Raiders Green Bay Packers game last night. We decide last minute to buy a few tickets for the team. The venue is incredibly impressive right? I’m a Packers fan. I’m from Wisconsin

originally, I’m sorry. Yeah.

I know. It’s it’s gonna be a rough few years. You can’t have three MVPs in a row.

Like but the stadium was amazing. Like it was beautifully designed the stage setup like the whole infrastructure the flow of people in and out of it. I was blown away. So that’s kind of a bright spot. I’m really excited about the sphere thing. Yeah. I tried to get tickets for you, too. They’re exorbitantly expensive this week, but I would love to go to a show there at some point in time that it looks amazing.

Yeah my fiance and I went to the Raiders Colts game last year which died in with the health conference phenomenal. I mean just you can tell why they’re gonna have the Super Bowl there. It’s two billion dollars very well.

Amazing. Yeah. Yeah,

but YouTube is on the horizon. That’s another one for whenever I come back here for this for sure. Absolutely. Well Michael really appreciate you being on the show sharing the Data Insights that you have and obviously best of luck on the partnership

really good to meet you.