Episode 12: How IBM and Graticule are linking up in rare diseases

Julie Krommenhoek, Vice President of Global Sales for IBM Watson Health, and Dan Housman, CTO from Graticule, discuss how the two companies are collaborating. Drawing on her extensive background in pharma research and real-world data, Julie shares how a partnering approach can solve the hard problems rare disease focused therapeutics companies face due to the limited volumes of patients they can track with RWD. The discussion explores a joint vision for creating and deploying algorithms to identify undiagnosed patients to support clinical trial recruitment or digital diagnostics to increase access to precision therapies. The podcast also explores how leveraging privacy preserving linking can extend claims data with other advanced patient data such as radiology, pathology, PROs, and genetic testing. These linked data sets can enable researchers even when analyzing small populations to generate necessary evidence to measure the value of new therapeutics on medical costs, outcomes, and patient experience in rare cohorts.

TRANSCRIPT

Dan Housman:

Hi this is Dan Housman from Graticule here with the Novel Cohorts Podcast. I’m here with Julie Krommenhoek from IBM Watson Health. I’m really excited about talking to her today because we’re announcing a partnership between Graticule and Watson Health. We’re here to talk about the kinds of things we can do together in rare diseases. Julie, welcome. To kick things off, give us a quick overview of what you’re doing at Watson Health and what you’re excited about in rare diseases.

Julie K:

Sure. Dan, thanks for having me. I’m Julie Krommenhoek and I’m with Watson Health. I lead Global Sales and my background has a long history in Real World Data. This is a really exciting partnership because we’re talking about some areas near and dear to my heart. We are working on not only rare diseases and how we tie them into clinical trials.  Most of my life was spent in researching clinical trials and then turning around and going to market with them. Today we’re going to talk about how we can take out the most prevalent mysteries in patient identification issues that exist, and how we’re going to link the data, extend the data, and most importantly, empower researchers to be better decision makers with more actionable data. As a researcher at heart, that’s what I set out doing years ago in rare diseases. This way we are extending this well past where we were before – which is trying to find data that is deep and meaningful – to now mapping that to the patient journey. With Graticule’s help, we’re able to make better decision making because we have the help that completes that journey, and answers the questions that all of us as researchers have been looking for – for years.  I’ll kick it off with that, Dan.

Dan Housman:

It’s great, and I’d say we work directly with Life Sciences customers in their rare disease groups.  They have been giving us all sorts of input into what are the hard things for them. I’d say among the big, hard problems we’re trying to solve with partnering with Watson Health is this patient identification problem. I’ll give the problem real quick and we can talk together about some of the solutions: if a patient does have a rare disease, they become a bit of a needle in a haystack, especially if they have a long process before they get diagnosed because a lot of these rare diseases look like a common disease. The people they’re going to see to get treated treat common diseases, so they get a diagnosis of a bad back when they have some kind of autoimmune disease. Or they get a diagnosis of a problem with their wrist when they have something that will eventually affect their heart. It’s really hard to come up with systems that can find those patients efficiently. If we can’t, then we can’t run a clinical trial or, at least, it becomes prohibitively expensive and slow. If we can’t, we can’t find the patients after the drugs are on the market, which means we have something that could help a group of people but we don’t know who they are.

We can talk a little bit, Julie, about areas where you see it’s exciting for us to work together to start building the solutions. We can take the Watson Health data and services to build out solutions in this patient identification space.

Julie K:

Absolutely. I love how you describe the fact that we have all these, and what you call and what we in the industry say is the data-to-study gap. What we were seeing is a couple major things happening when I was doing a rare disease back in neuropathic pain and we were talking about complex regional pain syndrome. We were doing small groups of studies, and it was all leading to neuropathic pain. We did have typical first responders saying they’re having neuropathic type of ailments, and we were mapping those ailments, but they were just in one slice of data. What you’re describing is that if we take those data slices, that doesn’t always meet all the radars for large pharma, because it’s more of your small size trials for rare diseases that are spurring the discovery. What I’m hearing us do today, which is completely different – and thanks for the FDA also accepting Real World Evidence now to support drug approvals in the setting of oncology and rare diseases – that’s a major breakthrough.  We’re seeing that COVID or not, we’re making some great entry points into those small size trials that you’re talking about. Dan, I think one area that the audience needs to hear about is how we, as a partnership between Watson Health and Graticule are solving some of those key crises between that data-to-study gap – to solve some of the rare diseases and those research challenges because they are small – and what we need to help in the future. Then we’ll get into the payer side and what we could do to help out on that, but let’s start with this.

Dan Housman:

One of the things that is fantastic about Watson Health is the MarketScan data set, which is pretty much the gold standard that health economists have been using forever. What it lacks often in rare diseases is a couple of things. One, there are certain extended data sets that would be really helpful beyond claims data, and sometimes we can expand out of MarketScan to Explorys to get to the EHR data. In working together, we have a nice partnership to enable a couple of things to extend beyond those two data sets. One, if we build a model, something that can predict that someone might have a rare disease, if we were to do a definitive test like a biopsy or a blood test, we can push those models into health systems that Graticule is working with and work like a CRO to then bring it closer to the point of care, which is really why they’re building out these patient identification solutions. They are most useful when they are at the point of care. We’re really enjoying the capacity to be able to take a big enough data set and rich enough data set to see what the gaps are and then move it forward into the health system environments. The other thing is, we can start to link this data with other useful scale data sets that are outside of Explorys. There will be more announcements to come but there’s already a rich set of partnerships that, given that we have a lot of expertise in what the problem is, we can already leverage within the Watson Health ecosystem to bring in extended data. Watson Health has linkage to lab data that can be brought in. We can bring in extended data in radiology and pathology. These are areas I think no one has thought to pursue until now. We are able to invent these new capabilities off of linked data as well as extending data that the customers we’re talking to in rare disease companies are getting really excited about.

Julie K:

Well not only excited Dan, you’re really answering what the synthetic control arm is bringing to market. We in the industry didn’t maybe title it as a synthetic control room or an external control arm, but we know that there are areas especially within rare diseases where it’s just not prudent or ethical to have a full trial. There are the expense and cost behind that, so replacing this from having actual patients with using real world data to our advantage by the multisource.  I think what you raised was the key contributor to this is having a diverse data set. Whether it’s imaging, whether it’s being able to mark where that disease is going and using those deep data sets, having diversity is where the future is in research. We know that the payers in the marketplace that are listening are seeing a future that’s going to increase, so they have to start looking at, as spend increases, health economic evidence increases, which is going to play a greater role in our partnership. Then the product evaluation process that lands in rare diseases and our orphan drug market requires us all to get really good evidence with a synthetic external control arm and start bringing that to market in a new way.  I know that you’ve mentioned burden of illness studies and things like that, which you’ve been personally working on that drive down the cost; they help with those treatment options so that pharma can actively manage new and many existing orphan disease types. I think this is an innovative way to think about new rare disease therapies as well.

Dan Housman:

It’s a good point, which is for the drug companies out there that are looking to bring a new product to launch, their drug works but it’s probably going to have a very high price point because these would be for small populations, and they need help from the Watson Health data, the right health economics arguments but also a very rich approach and an open approach to how we define burden of disease. So, on the one hand we have the strong traditional information that comes on healthcare utilization from things like MarketScan. On the other hand, we want to know what is the patient’s quality of life, where are their impairments? We can look inside some of the datasets like HPM, which is an absenteeism, presenteeism benefits data set within the Watson Health family that can give us this extended information about quality of life.  Or, we can link in data that comes from outside sources that can tell us about the quality of life. I think that’s going to help the groups trying to improve their dossiers for payers and their launch strategies and their communications to providers to communicate effectively because they have enough information on the patients that are in the small groups they want to treat.

Julie K:

Exactly. And then we’re also taking quality-of-life measurements, and we are creating our own inside Watson Health research; we are creating new historic landmarks in how we are bringing a sophisticated group to explain those quality-of-life measures.  Then we are showing that the measures can be available to all rare diseases.  So many different efforts are being taken, because we know that we already have raised the awareness of what needs to be done. With Graticule’s help, we are also implementing and then we are also looking into some genetic testing – genetic diagnostic tests – that are being used for rare and ultra-rare diseases. I know that you focus on a few rare diseases, but it expands beyond that.  There is a correlation that I see growing between that and quality of life and also a correlation with other types of clinical data that you’ve experimented with on identifying new target biomarkers and testing new panels for rare diseases. Do you want to talk a little bit about that?

Dan Housman:

Obviously we’re not going to make a ton of product roadmap commitments today. Among the many things we do for rare disease companies is merging and matching genetic testing data, which is now at large enough scale to be beneficial to identify rare disease patients. Here’s one of the problems – if you have a rare disease it doesn’t have an ICD 10 code. What you will find is you’re in some big broad category, and there are 12 different mutations, each of which is a different disease because only some can be treated with a certain drug. Because it doesn’t have an ICD 10 code, it’s stuck in the market. But by bringing the whole MarketScan data set or another data set into play with the definitive diagnosis that comes from genetic tests, which are now being done at scale from a small number of providers, we really can help a group get to that critical mass to say I can look at the 1,000 patients who have this disease, from this dataset, and know exactly which version of the disease they have in order to understand their drug and what the treatment impact will be. I think the higher-level point of looking into this is we’re hearing from customers the data they need, and we’re trying to be as responsive as possible to figuring out how to get it done. It’s just great to be working with your team at Watson Health, because you’re not opposed to innovation and creativity. You’re trying to do this through a “Smarter Health” approach, and it’s going to work because we’re willing to battle the unsolved problems and get to what people need.

Julie K:

I hear you getting to our overall tagline which is “Smarter Health.”  That’s what Watson Health stands for – enabling partnerships like ours to reuse or repurpose data. We’re scaling this properly so that the clients you’re speaking with join our fight.  The sellers I have on my team at Watson Health are answering those key challenges from our key clients, and they’re identifying insights and how to have faster breakthroughs and how you improve the experiences that people are having once they receive a treatment. What I also really enjoy about this partnership is the ability to use this approach and technology to grow our relationships in our partnership. We already have deep relationships in the pharma and CRO markets, health systems, and also providers. What we’re able to do is bring all this together so we can deploy the right solutions and we’re leveraging them at the right time. What you’re helping us do is identify how those rare diseases in the broader populations can become more common diseases to solve. We’re okay with bringing in some of those extreme cases, because I think some of the smaller biotechs that you and I have talked about don’t get noticed.  We have to make it easier for not only drug approval but also getting to the right area where we can actually see those rare diseases get to launch and be a part of the life cycle so it becomes a bigger marketplace.

Dan Housman:

That’s a great point as well. You know rare disease is a strategy for a pharma, often to get into a broader disease category.

Julie K:

That’s exactly what it is.

Dan Housman:

So, if you want to figure out who’s going to respond to a drug, you look for the people with the most extreme phenotype. Maybe there’s only 1,000 of those people, but your drug will ideally, if it’s a good drug, work very well in that population. By working with your team and the data and services we get from Watson Health, there’s a pathway that we can offer to these groups that are fighting through their first indication. Whether it’s a small biotech or a large biotech with a rare disease strategy but has the data to provide the pathway to a more common disease. As an example, Prurigo Nodularis is an extremely rare, extremely strong itch disease, and we can identify all the things we need to get that drug into market. Itch is a huge broad market:  there’s dermatitis, atopic dermatitis, and you can walk across those various levels. Over the period of five or six years we can give them the correct route as they’re walking to find the root with the data and services we’re putting together. But it starts in a lot of these cases and I think it’s going to continue to start in these cases with rare diseases and real-world evidence, which is why we’re so excited about focusing on this seemingly narrow space.

Julie K:

Well, Dan, we’ve covered the data-to-study gap because that is one of our big callings that we really ask for help in creating a market and building channel expansion in that area. What I hear back, in case you don’t hear it all the time Dan, is what we hear from working with the team at Graticule is the partnership is both complementary and additive to our other businesses units.  So you’re able to pull in multiple different areas throughout our company to be able to go deeper within each area but also the intent is to create faster and scalable options for our customers. Watson Health sellers are saying that the pharma and CRO companies are responding, so what I really like is this is all encompassing. We’re able to meet the client where they’re at: starting with the population that they’re dealing with and then also bringing it through the patient journey, helping them with patient identification and bringing it all the way through, which is helping the entire health ecosystem. So, it’s definitely one that I can tell you our sellers have had a positive response, and it really does put a strong tile next to ”Smarter Health.”

Dan Housman:

And so I’ll leave with my vision and you can add a touch of your vision.  I think, especially in this space, success comes down to how we can take all of the piles of data, the piles of technology for doing machine learning and artificial intelligence and buzzword galore, and turn this data into something that really helps people at the other end.  The point of care, five years from now, will encompass a multitude of tools that equate to simple alerts for not only ensuring someone with a high A1c gets their diabetes treatment and counselling, but are now also directed at a larger population that will benefit because we’re finding the people with the rare disease, we’re finding the fast progressors and we’re impacting them by translating tools across from these big data sets into each health system that’s willing to adopt. So, I’m hoping we can get there. We’re still, I think, close to the starting line.

We look for companies like Watson Health to help accelerate our journey in this unique space, and have faith that by collaborating we will uncover gaps and expedite a faster turn in trial findings.  I’m not sure what you’re hoping is going to be the long-term outcome; I’d love to hear it.

Julie K:

Mine is finding the intersection of best-in-class technology and science to unlock rapid innovation. I don’t need to repeat what you said, that was perfect.   By applying different models  powered by RWD insights to our analytical workbench, this allows our teams to combine  in a “data exchange”.  This exchange can pull in a wide variety of our tech, algorithmic and AI services that deliver a uniquely centralized solution. The key is unifying not only the multitude of data sources available, but also capitalizing on partnerships like ours to advance the amount of rare disease medications brought to market. If Watson Health and Graticule can help handle one of the toughest areas, imagine what we can do with larger datasets where we don’t have to be as sticky with the deepness of the data and the richness of the data.  But I’m glad that you’re staying humble in your vision. You are right in that’s exactly where we have to start – rare diseases. What’s really nice is we get to focus on multiple areas of the ecosystem. You mentioned small biotechs that might be in their first expansion – we know how to do that. We could also be a part of the mid-market biotechs and pharmas, and we can also help the large pharma customers and CROs. The bigger biotech and pharma players are looking for specific sponsors that have needs across the board. This gives us the ability to be a solution provider for multiple markets.

 Dan Housman:

It’s been a pleasure speaking Julie. I’ll leave it with a call to action if anyone’s interested in some of the things we’re doing.  We’re happy to run workshops, Graticule and Watson Health together can offer expertise to groups that want to pursue one or more problems. The best way to connect Graticule is through our website, www.graticule.life, and go to our contact form. Obviously if you’re seeing this podcast you probably can find this already. We hope you enjoyed this discussion and that we get to hear from you sometime soon. Thanks again, Julie.

 Julie K:

Thanks, Dan. It’s been great, we expect great success and we’ve already seen it. Bye.