Name: Komodo Target Smarter Spend Better Webinar (1080p) (1)
Uploaded: 2026-05-19T19:04:23.511Z
Duration: 28 min 39 s
Description: Komodo Target Smarter Spend Better Webinar (1080p) (1)

Transcript for "Komodo Target Smarter Spend Better Webinar (1080p) (1)": Welcome, everyone, to Target Smarter, Spend Better, and Unlocking ROI with Granular Insights by Komodo. I appreciate everyone taking a bit of time out of their day today to join us for this lunch and learn session and looking forward to a very productive and exciting opportunity to engage with, clients and members of life sciences. Let's, Stephanie, go to the next slide and review a few housekeeping items. And before we begin, just, for the group, your microphone is muted to reduce background noise, and should you encounter any technical issues, our dedicated support team is ready to assist. You can reach out to them through the chat, and they're happy to help resolve any issues. We do encourage active participation and work warmly welcome any questions and comments. So please feel free to use the q and a feature, and we'll be able to capture those questions, and we'll be addressing some of those at the end of the presentation today. And then be aware that the webinar is being recorded, and it will be distributed following this live event. Next slide. So for the group, my name is Chad Forsey. I'm a commercial segment strategy lead for our large biopharma customers. Been working in health care for the last fifteen years, mainly in commercial capacities, everything from carrying the bag out in the field to field marketing, commercial analytics, and business development. I have been with Komodo for the last four years, supporting commercial teams and life sciences to optimize their access to Komodo's products and solutions. Daniel Brocks, who will be joining me today, is our general manager of Komodo's analytics consulting team. Our analytics consulting team works with clients to turn our healthcare map into deeper insights, strategy, and action. Daniel has over twenty five years of experience bringing data driven analytics solutions to biopharma customers that span the enterprise across commercial, medical, clinical, and HUR. And Daniel's client portfolio has clients spanning pre commercial stage biotechs to really top five global pharmaceutical organizations and everything in between. And I'll turn it over to Daniel briefly. Thanks, Chad. You know, before we jump into the agenda, I really just wanna take a moment to acknowledge the team at Komodo that's really been responsible for building the Komodo drug projections, methodologies. There are so many people I could mention, but I'd especially like to acknowledge the inevitable doctor Paul Gurney along with his team, specifically Emily McGurney, Laura Chodes, and Hannah Lou. They've really built this capability and proven it across clients. And I also wanna thank our clients who have been working with us along on this journey. There's it's really been a true partnership, bringing Komodo projections to life to reduce the burden disease. So, just very appreciative for that, the shoulders, upon whom we sail. So thank you very much. Likewise. Next slide. While I review the agenda, our our tech team on the back end is gonna pop a poll question onto your screen. Please take a second, review and respond. So for our agenda today, we're gonna be reviewing today's challenges in a data driven decision making process, leveraging the Komodo data difference, simplifying your ability to target with precision, and then we'll end with a q and a session. The poll will be live for a couple more seconds. So as you have an opportunity, you can review the question and populate, it'll it'll populate your responses, and we'll review those briefly before diving in. So, give everyone another few seconds, to respond to the poll. And I'm sure you're all figuring this out, but it's a multiple choice. So please choose any any that apply. Any at all. Alright. So in terms of what are your most pressing challenges measuring product performance and market potential subnationally, please click all that's all that apply. And we will be able to address each of these throughout the presentation, but really popping to the top there is understanding data capture and adjusting for bias within a sample, and then understanding market share by indication or line of therapy. All real challenges, and looking forward to diving in and and sharing insights that will will really cover the gamut. Alright. Next slide, Stephanie, if you don't mind. So today's challenges in a data driven kind of decision making process. We can acknowledge that data is rapidly changing, and there are so many sources out there for you and teams to utilize. How do you focus on what will bring the most impact for your teams, your brands, and your organization? Precision medicine is growing rapidly and markets are becoming everly more complex, and we no longer live in this simplistic view of a market with Rx only insights. Medical benefit insights are becoming more vital for a total view of the patient journey, and you and your teams are facing increasing pressure to deliver more with less every single year. And as a result, the health care market and specifically the data market becoming more complex, traditional datasets struggle to capture medical benefit utilization, closed pharmacy networks, and competitive landscape shifts. They create blind spots for commercial teams. Inaccurate targeting can lead to directly misallocating resources, can lead to missed opportunities and high value prescribers, accounts, and territories. And finally, market trends evolve quickly. With new treatment guidelines or payer policy changes, most data insight and projected data models do not update frequently enough to capture these shifts in real time. This can lead to delayed response and cause potential market share loss and lack of an, understanding of your market as a whole. Next slide. So how do we understand these gaps at a granular level to enable our commercial teams to target the resources more effectively? How easily are you able to target a multi indication therapy? How long does it take for you to gain actual insights to make the right investments in your campaigns? Commercial teams are using multiple vendors to try to get a complete view of their market, especially to pull through granular insights, but integration barriers exist. Combining different data sources can be challenging and often limiting, and these limitations, they decrease visibility for marketing teams for commercial teams, making it very difficult to execute the right targeting strategy and the right investment decisions at the right time. And as we all know, data reliability challenges can lead to cautious decision making, ultimately slowing things down further when we all really need to move faster. So starting with the Komodo data difference, the evidentiary standard for real world data, we're gonna review Komodo's data, and then we'll follow that with the methodology around how we drive these granular insights with Komodo grub projections. Next slide. As a health tech company, establishing an evidentiary standard in life sciences requires more than just software, platform, services, and solutions. For our clients to realize a velocity advantage in bringing therapy to patients and reducing the burden of disease, we acknowledge that the products that we build at Komodo need to be powered by a best in class data product. Everything that we build at Komodo sits on top of our health care map. We're tracking over 330,000,000 patients across The US health care system, providing greater visibility to how patients are being treated. The health care providers that are part and parcel to their care, and the therapies and interventions that are critical to their patient journey. So what does this look like in terms of how we've built our data product? Next slide. In building in and investing in a data product over the last ten years, Komodo's taken a deliberate and unique sourcing strategy. This heterogeneous approach to source curation provides an advantage for building national and subnational projections. We pull data from a variety of sources that helps us address bias that we see in health care data. Bias like patient representation, like age, gender, race, and ethnicity, payer coverage, place of service visibility, medical or pharma pharmacy benefit visibility, and blocked specialty pharmacy drugs. We start by bringing in the switches and clearinghouse data. These are often low lag sources with broad coverage, and then we layer in provider complete data that matches up with the clearinghouse data. Not all, but a good chunk of this. And the provider complete data is informing us on payer derived insights. So bringing the compare complete data not only matches up with the clearinghouse data, but also creates a rich overlap of sources that provides optimal information to power a projections model at a product level. The payer information also helps inform us on provider, our provider complete data. And then finally, if we go to the end of the animation, we see this really rich overlap to generate projections with a Bayesian inference model with high confidence intervals because the integration and characterization work that we do on our data product. So in terms of addressing sources and the sources within the health care map and how we address different biases, let's talk about this in terms of coverage. Next slide. From a coverage perspective, you can see represented in these cartograms four different data sources and how much volume that we receive within the health care map at a zip three level. You might note things like source v and source x are pretty light in California, or sources y and z have strong volumes in California. If you were working with a vendor that was only relying on one of a few sources, you might be lacking visibility that's critical to how your teams are ultimately developing strategies, to bring your product and therapy to market. So our data sourcing strategy is designed to raise the visibility floor across the country so that we're not totally dark anywhere. How does this look when we put it all together? Next slide. When we pull all of this together, you can see that we have pretty great visibility everywhere, and we can use these visibility estimates as the bias from a coverage perspective in projecting up to census. And now that we've talked about data, the sourcing strategy, and some bias that might might exist within the data types and across different geographies, let's review the projections model that we've developed. And I'm gonna transition it over to Daniel. Great. Thanks, Jed. So as Jed said, we're gonna take a few minutes just to dive into the structure of the model itself and the intelligence that's been built into this offering. Next slide, please. So one way to think about it is that the the fundamental challenge of projections really is to develop trustworthy estimates of what we believe our visibility is at the national and subnational level. Said differently or maybe another way of saying it is, what's the probability that we actually observe an actual event that happened in the health care map? If we have reliable estimates of these probabilities, we can develop a projection, which we trust. But to address this fundamental question, Paul Gurney and his team have developed a very, very smart, hierarchical Bayesian inference model. So how does this work? So as you can see in the upper left hand side, the first thing we do with this model is we calculate visibility estimates at the intersection of drug, payer channel, and age group. We the the patient strata that's used fundamentally comes down to this patient strata of payer channel and age group. And there's so much information contained in in that strata that we can use. The model then learns from the health care map over time as new information becomes available. Probability estimates are calculated for each data provider within the map. They're combined based on trustworthiness and bias levels. The provider mix has been calculated on a weekly basis to ensure consistency over time, and then adjust if necessary. This is really important. So if you if you think about, for instance, the CHC cyber attack last year, this market wide event that affect affected so many organizations. Because we can see the source level beta, we understand which customers were the most impacted by the attack, which allowed and and what the overall impact was. And so we can adjust our visibility estimates accordingly based upon all of that information. So the the model itself can also be tuned, as market events happen as well. As I mentioned, the projections are updated on a weekly basis, and they're updated across all the different sources of data. This model allows us to produce an output at both the national and the subnational level, which includes both the observed data. So the observed data, which you can calibrate back to or you can compare back to the, you know, the underlying data that you might be receiving from us today. But it also gives you projected drug events, both r x and m x. And because the model is probabilistic, it can also tell us what the confidence interval is around the projection, which we believe is super helpful just in terms of transparency, but also giving you the ability to contextualize the data and, that that we're we're sharing with you. If you go to the next slide, we also calibrate the model using multiple sources of truth. So there are public publicly available 10 queues. We've got a we're one of the CMS, innovators. So we can use the CMS DRDC data as a source of calibration. We also sometimes use our client's eight sixty seven data to to work within their specific projections. And all of these sources of data, there's, you know, there's, you know, publicly available epi data that we can use, kind of the list goes on and on. But if if we can find the source that we trust, that we can use to calibrate, we can improve, the model and its and its results. So how does the model do? So the model produces very accurate results. So let's let's take a look at an example. This chart that we're looking at, we're this is real data. And what we're doing is is we're comparing a client's eight sixty seven data. So the the volume that we see, in the, in the eight sixty seven data on the x axis. And we're comparing that to the projections on the y axis. Each dot, on the chart represents a single provider. The size of the dot is telling you some is is is in proportion to the volume of that account. The green bubbles indicate only the providers that we believe, a priority are are provider complete. So Chad was talking about our different source mix. These are the provider complete, customers. As you can see, those provider complete, the green bubbles, almost all, like, line up, you know, almost bang on perfect with the line. The deltas that we see, between the two, are are attributable largely to just timing effects. As you're probably aware, the eight sixty seven data is really telling us, what's happening as product is sort of moving into the back door of a of a of a location or into the refrigerator sometimes. Whereas projections is really about demand sales and estimates around when the product was administered or dispensed. So you can see, you know, when when we compare that particular segment of the business, we're getting, you know, very accurate, very reliable, results. Now what we wanna do is we wanna, look at a specific example. So in the case that we're gonna look at for the next few minutes, we're gonna look at an example to understand observed and projected data at the indication level. Before we dive into the indication of a data, though, just wanna talk a little bit about, visibility and how it varies by product, because this is a very important concept for projections. So in the overall Komodo Healthcare map, we have patient level event visibility to about 80% of all the Rx volume in The US and about 40% of all the MX volume. And we believe those are both market leading figures, in terms of access to data at the event level by for the patient. But that visibility can vary quite a bit by product. So for instance, in this chart, you're seeing drug a. Drug a has got, a pretty narrow specialty pharmacy network. We've got pretty good visibility to it. So we see about, let's call it, 60% of what's happening for that particular product. For drug b, because, they have a wider network and there's more blocking going on, our visibility is a bit lower, and we see about 30% of what's happening for that product. But notice how the visibility improves when we do the access that we have to the CMS VRDC data. Of course, that's a higher line data source. We don't see that data until a bit later, but by the by the time we have access to that data, we can actually see between 80 to 90% of the volume for these products. And that visibility at in that live data gives us a much better understanding of how our visibility is later in the time period that we're trying to project, and and that's information that helps us dial in the, the the projections. Let's let's see how this, looks by indication. If you go to the next slide, please. So in this example, we're looking at products in the myelofibrosis market. And we're focused on drug a, which is indicated for myelofibrosis with thrombocytopenia. So starting on the left, we can see that when you consider all the drugs that are used for myelofibrosis across all of their indications, the drug has about a three percent share. That share is very low, primarily because, many of the competitors in the market have a much wider much wider range of indications. So quite a bit of that 97% of the volume really isn't isn't addressable for this product. However, by leveraging patient level claims data, we can filter that data down to the drug volumes based on the indication, for every drug in the market. And when we filter the drug down, for patients who are diagnosed with myelofibrosis, we can see that the share doubles. It increases to that six percent. However, we're not just indicated for myelofibrosis, we're indicated for patients with myelofibrosis and thrombocytopenia. The shares, when we filter further, when we do that, you can see that the share goes to eight 18%. Now, of course, we could always do this kind of analytics, with the Komodo observed data, and many of you probably are. But now but, excuse me. We also know that there's inherent bias in the the source fix in the data. So now with projections, our clients can understand both at the national as well as the subnational level, how they're performing, where they're under and over performing the market, where there would be opportunities to, to to grow the brand or to inform any number of commercial or medical strategies. And, you know, just super helpful information, we think, for all kinds of planning purposes. If we go to the next slide, you can see what this looks like at the subnational level. So this table is showing, real data, same market, all the way down to the account level, and then it you're further double clicking into the HCTs. Let's just look at the just the first two rows. So you can see from this that we've got excellent visibility, to Texas Oncology. So if you're just looking at the last year, you can see that, the observed data is sixty four patients, and we're seeing and we're projecting seventy one. So the projection is taking us up about 11% because we have excellent visibility for that account. If you look at Florida Cancer, the second row in the file, there, you can see that we go from, twenty three observed patients to eighty nine projected patients. So, here, visibility is closer to twenty five percent, and the projection factor is is about 3.8, being a little precise. So you you can see that this helps you to contextualize the data in a way that might not have been possible before. This also helps you to make better resource allocation decisions. You know, now with the projection, we can see that, you know, Florida cancer is much more important than, you know, we might have believed from just the observed data. Now, of course, your field has always had that intelligence in the past and has, you know, probably always thought, like, you know, what's going on here? We know there's more patients in that place. So now you've got, you know, a a a model and a result that is more, attuned to to the ground reality, which we think increases trust and helps make better resource allocation decisions. The last thing I'll say is that when we do subnational projections, many of our clients need insights that go deeper than just the product level. So in this example, you can see that the emphasis is on the drug regimen. That's why the regimen is the first, drop down. But we have the ability to do this, subnational projections by indication, by line of therapy, by regimen, by patient age, etcetera. So the this level of sophistication can really help you dial in, your understanding of performance and potential for more complex therapies. Alright. You just go to the last slide. So we think those these insights will enable you to, you know, one, allocate resources for the best opportunities. You know, where does your therapy have the best success rates? Where can you allocate scarce resources to drive growth? It also helps to respond to dynamic changing markets. Did a new drug enter? Did one of the top payers change their coverage? Using these projections, you can see how the market is evolving and develop strategies to stay ahead. But, ultimately, what projections allows you to do really is to reach the right patients at the right time and with the right therapy to deliver the best outcome for the patient and ultimately reduce the burden of disease. And ultimately, that is why we built this offering and why we're so excited about it. So thank you so much for listening to that. I hope that was helpful. We're gonna use the remaining time for q and a. Yeah. Thank you, Daniel. And we do have a couple of questions that have come in. And if any questions, have yet, please use this opportunity. You can use the q and a, portion, or the q and a feature, and our team on the back end will make sure that we see them. The first question coming in from the audience was, so do you calculate coverage for each product and each HCP under each HCO? My understanding is the coverage will differ for each product. That's a great question. I think at a high level and and just for the group, if if if there's the need to follow-up and provide, deeper understanding on our sourcing strategy, what our visibility estimates are, or the specific methodology, feel free to follow-up with the team and email, and we're happy to schedule time, one on one to dive deeper into your question and how it's relatable to your your market and your market baskets. But a high level, looking at the combination of sources that we have within the health care map, we first start to focus on a highly confident closed sample within the population. From there, we establish what we see as observed treatment rates across the strata that Daniel had mentioned. So we're looking at, you know, age, pair type, etcetera, within that treated population of a highly confident sample. We then use third party data like Kaiser Family Foundation, epi data, our BRDC data, to get an estimated population size. And then the ratio between the closed observed or sorry, the closed population to the estimated population, we apply that ratio to the observed treated population to get us our projected treated population. So at a high level, that's kind of describing how we go about, you know, establishing that projected value. We then provide for the team a confidence interval, and confidence intervals exist across a few different dimensions. How much visibility do we have within the health care map around that product? Where is it being used? How is it being distributed? What types of sources do we see the product coming through? And how we tune the model for those dimensions ultimately affects that confidence interval. And so for us, taking that one step further, we apply that same logic at an HCO level, and we can apply this, as well as Daniel mentioned across indications and at different stages of, of line within within lines of therapy. So hopefully, that insight is helpful. I'm happy to schedule a follow-up if you'd like to dive in a bit deeper. Taking the next question, that we see, what biases exist in your coverage of stacked assets? So, again, I think, Daniel, this one comes back to the health care map. I'll take this one if you have any insights you wanna share on top of that. Our our map does have some geographic bias. As you saw in the, cartogram slides, no there's no, vertically integrated, you know, perfect source of truth within The US healthcare system. I think for those of us that have been working in life sciences, we all wish there was, but we do the best we can to address bias. I think for us, you know, the acknowledgment of what we do see allows us to build confidence with our clients so that you can establish more powerful analytic, and understanding, you know, what is in kind of your blind spots. One that we acknowledge is we're a bit light in Colorado. There's, limitations around California. But based on sources within the health care map that we do see, we're allowed we're able to make adjustments accordingly within the model. I think that, understanding this bias at a detailed level is, what has allowed our team to make the necessary adjustments. And the more we inform the map of sources of truth, the more we're able to make positive adjustments to the confidence interval and provide more accurate out outputs. We have more questions coming in. Happy to continue to take them. This one is, if the VRDC helps give you confidence in the total share for a model, however, VRDC data lags by one plus year, are you projecting VRDC data out to be able to provide projection estimates at a weekly recency? Daniel, we can tag team this one if you wanna start. Yeah. Yeah. I can start that. So, so the data lag in DRDC, depends on source. So for instance, we've got data today through, December 2024 for, the Medicare fee for service data, versus other, you know, Medicare Advantage, Medicaid. There there's lower latency, further back. So, but the idea is, you know, wherever we have that data, we can compare that to what we have in the map at that time, which gives us tremendous insight into visibility both at the national level as well as the subnational level, for that particular channel. So the the VRDC data is is super helpful for sure. Yeah. And I would say that that also helps characterize the populations. Right? And so that's what we're looking to do is characterize, you know, each given patient population based on, you know, what we're seeing from different source types. And so having that as a historical reference, again, kind of informing the model of bias, and then we make adjustments in real time as we see things come through those little lag sources like the open claims, the provider complete data, and then ultimately the pair complete data kind of rounding that out. But, hopefully, that is a sufficient answer. And, again, happy to follow-up on that, should, deeper dive be needed. Daniel, I think this is another good one for you. So, question came in that, our indication is going to be changing from third line to second line. So we need an understanding of how we're doing by line of therapy. Is this possible? Yeah. So that this is actually a really common use case for us. You know, Encommodo has been doing line of therapy analysis like this, really since our earliest days. So we've got a a really good understanding of how we develop business rules to tease out, you know, for each patient in their journey, you know, where are they at in their journey, what line of therapy are are they on. And we can use that same approach to developing those business rules to to tune this specific model. So that's that's how we're doing it today. And and fundamentally at its core, it's, you know, the kind of work that we've been doing since the amount of time for us. I think that's really a great point to end on, Daniel, as we look to wrap. I think the goal for us is we've, you know, brought this out of the lab as a commercially available product is to build trust and transparency with our clients. Not only being able to show you what do we see of your given market, but how do we go about solving this very complex problem. And as we acknowledge that health care is becoming more complex and more difficult to understand in some instances and help data even more volatile, a sophisticated approach is essential to drive more accurate and granular insights for your brand teams. And so we look forward to opportunities to make this relevant to you based on the patients that you're looking to engage, the drug therapies you're looking to bring to market, and ultimately your goals in reducing the disease burden. We hope that you'll trust Komodo to support you with our Velocity Advantage, and we look forward to follow ups accordingly. I appreciate everyone taking thirty minutes of your day. I'm looking forward to an opportunity to follow-up in the future. Thanks, Jed.