Name: Komodo National Drug Projections Webinar Recording
Uploaded: 2026-05-19T21:32:41.333Z
Duration: 40 min 25 s
Description: Komodo National Drug Projections Webinar Recording

Transcript for "Komodo National Drug Projections Webinar Recording": Alright. Let's go ahead and get started. Thank you all for joining us today on our webinar for national drug projections. We have a few housekeeping items to get us started. First if you have any technical issues throughout the webinar please do reach out to our support team so that we can resolve those quickly for you. If you have questions throughout the webinar please pop them into the Q and A feature at the bottom of the screen and we will do our best to address them at the end of the webinar. And finally, we are recording this webinar, so your microphone is muted to reduce that background noise. And also no worries if you have to hop early or to jump to another meeting, we will be sending out the recording later this week. Alright. As we are, sharing some future product direction and roadmap content in this presentation, we would like to remind you that this is for information purposes only. Please know that any future plans are subject to change at the discretion of Komodo Health. And with that, let's kick things off. It is my pleasure to introduce our presenter today. I have absolutely loved not only working with Emily McGurig, but also learning from her over the last couple of years, and you all are in for a real treat as she walks us through our latest offering. She has two and a half years of experience here at Komodo, but, and leads our commercial data strategy team that has over a decade of experience in partnering with life sciences organizations, and she excels in leveraging health care analytics and data sciences to help companies enhance patient care and outcomes. My name is Stephanie Idol. I've been with Komodo also a couple two and a half years and, have over twenty years of experience in the life sciences and med device industries, and I am one of the product marketing leads here at Komodo. And just real quick as before I hand it over to Emily, an overview of our agenda. We are really excited today to introduce Komodo drug projections and look forward to your engagement. Emily will be taking us through the details of the Komodo difference, walking us through a product demo, and then taking us through the vision, of the future. We will have time for q and a at the end, so please, as you think of questions, pop them into the q and a, and we'll do our best to respond to them at the end of the webinar. Without further ado, I'm gonna pass it over to Emily. Thanks so much, Stephanie. Thank you all for joining. It is my absolute pleasure to introduce Komodo drug projections to you, which we'll be launching this week. I wanted to get started by talking a little bit about why we embarked on this journey of, providing projected data in addition to the observed data that we already have within our health care map and offer to our customers. So it's our opinion here at Comodo that today's approach to projections are lacking. Current syndicated offerings on the market lack transparency, comprehensiveness, and foundation. And so what do I mean by that? On the transparency side, Black box methodologies provide, limited contextual information to help understand the outputs such as observed values or confidence intervals. On the comprehensiveness side, the evolving world of biosimilars and mixed benefit brands and regimens does require a combined medical and pharmacy benefit view, which is not on the market today without the use of expensive consultants. And then finally, lacking foundation. Outputs are currently refined by observation and experience rather than built from a standard scientific basis. And what this really leads is to us here at Comodo and our customers lacking trust in their projected data. We have a quote here from one of the, companies that we are currently working with in a projected offering. It's hard to get accurate projections for medical benefit and specialty products without spending a lot on consultants. A prescription audit misses most of what matters today. And so what we have embarked on starting today is offering syndicated offerings that account for all of these issues. So we have a poll question for you. And I said we were gonna put the poll up. There we go. So what is the one thing that sparks doubt about your projected data? Is it lack of transparency, lack of comprehensiveness, reliance on that gut feel and observation, or something else? Would love to hear your thoughts. And I'll give it just a minute. So it looks like the big winner so far is a lack of transparency followed by the lack of comprehensiveness. But it looks like the clear winner here is a lack of transparency. Give it just a few more seconds before we close the poll. So, yeah, as we see, most people are saying there's a lack of transparency, but also comprehensiveness and its reliance on the gut feeling and observation. We have few that have other. I'd be interested to kind of hear what other people think, maybe at another time. But from the transparency perspective, you now have Komodo National Drug Projections. We are working here at Komodo to bring back trust into projections. So we have over 10,000 drugs, with both medical and pharmacy benefit, with projected and observed values. So we're being completely transparent about our methodology as well as what we see within our health care map as well as other sources internally here at Comodo so that you can really understand the outputs that you'd be provided. And so these are kind of the main key features of our projected offering. First is again that transparent approach. So we're providing you both projected outputs with confidence intervals as well as observations to easily understand what you would be receiving from us. Second, we have comprehensive views. We have the ability to combine both medical and pharmacy benefit products as well as combine biosimilars or generic outputs into branded views. So an example of this is if you're in the, biosimilar market with Humira, Humira, AmgenVita, Ciltezo, those could all be aggregated up to one view within our template. Another example from the generic side is brand name Revlimid and generic lenalidomide. If you were interested in kind of seeing that in an aggregated view, we have that available for you. Third is a data driven foundation. Armando is grounded on facts from our data, and I'll go into further detail about what I mean by that later on in the presentation. And then finally, it's easy to use. All of our projected outputs will be available in MATLAB. With just a few clicks, you can have multiple views of your market at your fingertips. So let's go into the methodology a little bit. There are kind of three key elements to our projected outputs. The first is it it's grounded by our health care map. We've spent years selecting and incorporating multiple diverse sources to produce the most accurate insights. And second, we leverage those diverse source groups to their key advantages. So individual source group characteristics supply unique insights that contribute to the larger projected output hole. And then finally, we employ multiple statistical methods. We estimate the likelihood that we see an interaction with The US health care system, and we're updating those estimates as more evidence becomes available in our health care map. So these are kind of the three main ingredients that we use to project our offerings. So I'll go into each one of these in a little more detail now. So first, our health care map. Many of you are familiar with our health care map and our events based fabrics and utilize them for your analytics and different insights offerings. For projection specifically, I wanna highlight three things within our map that are really important. First is comprehensiveness. We have, on average, 85% true coverage for most products within the map. What I mean by true coverage is what's available within our health care map and we can provide to our customers, but also internal only sources such as our innovator license, which provides us a 100% visibility into CMS, Medicare, fee for service, advantage, and Medicaid claims. We also have other sources internally that we leverage to have that comprehensive view of products across the market. Second, it's representative. We have visibility across The entire US down to the zip three level at about a three quarter rate or more. So we have a lot of availability across the entire US, a representative mix of what's going on, not just pockets of geographies. And then finally, we have diversity. We combine many different types of data providers, each with unique views into The US healthcare system. And this is really a key element of our projected outputs. So to go into this in a little bit more detail, we have different source group characteristics which we isolate and gain advantages and insights from to have a larger whole of our projected output. First, being consistent visibility is not complete visibility. So you may observe a patient consistently filling a prescription at Walgreens. What what does that tell you about dispensing behavior at Onco three sixty? Does Walgreens contribute 50% of the volume in that area of The United States or that therapeutic area Or 5%? These are the questions that are key to being able to project accurately. Here at Komodo, we rely very heavily on our payer derived data. It provides us two key elements to being able to project this data. The first is a subset patient population with full patient journey visibility. This allows us to calculate accurate treatment rates at the individual drug level. The second is we have a minimal floor of visibility into every provider in The United States health care system, which enables a precise provider mix calculation. So even in the areas where there's specialty blocking, we have some visibility into those blocked drugs through our payer complete sources. We even have some visibility on the medical side and provider side into, you know, typically institutions that are very difficult to see such as Kaiser. So this is a very key element to the output. Second, we utilize different elements of our data to eliminate biases. So one example I have here is our provider complete data. We have a lot of advantages of our payer complete data, but it is definitely biased from a payer perspective. So isolating to provide our complete data enables us to provide, calculate a payer mix that's unbiased. And this is just one example of how we address bias within our model from multiple angles. And then finally, we use kind of our other data category to calculate a total population, which is the denominator for, on which we project up. So we utilize a mix of both internal and external sources to estimate the total US population. This includes the census data, Medicare public use files, trusted third party data such as the Kaiser Family Foundation, as well as Komodo's own Komodo patient insurance, which leverages enrollment files from our payer complete data as well as open data to help us understand age group as well as payer channel, for all the patients within our health care map. And so these are just examples of kind of the different key ingredients that we utilize to calculate our projected outputs. Finally, our our, outputs are grounded on a strong mathematical foundation. So we calculate those estimates that we were just talking about on the previous slide for each intersection of drug, payer channel, and age group. The model learns from our health care map over time as new information becomes available. We also account for bias and any type of trend breaks, And we do this by calculating these probability estimates for each data provider within the health care map, and we combine those estimates based on trustworthiness and bias level. And then we also do provider mix calculations on a weekly basis to ensure that there's consistency of a provider mix for a particular drug over time and adjusting accordingly if necessary. And so this was super important just a few months ago when the change health care cyber attack happened. The provider mix for many drugs changed dramatically, practically overnight within our map, and we were able to account for that within the model. So what are we offering in within our map lab template? We have multiple views and weekly insights for you to access. You'll be able to trend multiple drugs to understand the market, track market share over time, as well as visualize projected and observed insights together to understand the projection factors within the health care map for individual products. And we're also going to be record reporting across three key metrics. The first being total dispenses, which is a calculation of total unique patient days. New to molecule, which is our concept of new to brand enhanced, as well as total units. So estimated units for that product with the corresponding unit type, whether that be tabs, milligrams, milliliters, pens, etcetera. And again, all of these metrics are available for both medical and pharmacy benefit products and provide aggregation at the molecule level, which is either generics and brands combined or biosimilars combined into one simple view. So before I jump into the demo, I'll see if we have any questions. We do have a few questions in the, in the chat, that we'll probably walk through. How do you calculate 85% true coverage? That's a great question. We calculate this based on the information within our health care map as well as all the internal information that we have in house. So, again, our innovator license data, which is a 100% visibility into, the CMS data, as well as some of the historical sources that are no longer with our health care map. This will vary from product to product, but, again, on an average basis, particularly within the Medicare heavy space, we have 85% or more coverage. Does your source cover specialty pharmacies? We do cover specialty pharmacies. So we have, direct relationships with specialty pharmacies as well as visibility into those specialty pharmacies from our payer complete sources. Again, it's it's going to be a subset of patients, and that's kind of how why we leverage the total US population and balancing that patient population to the denominator to understand how much we have to scale up. But we do have visibility into the specialty products. This one's an interesting one. How do you translate these metrics into projected patient counts? How do we translate into projected patient counts? For the new patient starts specifically? It doesn't clarify, but maybe that person who asked that question can follow-up. Okay. Well, I could talk I saw I see someone else is also asking about NBRx and how we define it. So I'll talk a little bit about the new patient starts from that perspective. So we're looking at this, with a twelve month look back period, and we are biosimilar and generic agnostic. So, again, if someone starts on Humira and then switches to AMGEVITA, they will be counted as new to molecule for Humira. And so we're counting, from that perspective. Same with the brand name Revlimid to generic lenalidomide, we're counting across the molecule level to understand where that patient started within that product suite. I have one final question before we move on to your demo, Emily. Can you comment on how you project for brands promoted by private companies where you don't have visibility to their national sales? Yes. We actually, are working through that right now. One of the offerings that we'll be adding to the, product suite probably in the coming months is a gross sales metric offering. And so we're focusing on The US reported revenues, to help us understand kind of where our gross to nets fall between what they report and what what we project from a gross sales perspective and then leveraging that towards our, OUS product companies. Perfect. We'll leave the the rest until, after, the demonstration. Okay. Awesome. Let me know when you guys can see my screen. They can see. Awesome. Okay. So just to give you a quick view into how you could create your dashboard, when you come into the map lab environment, you have your create dashboard here at the top. We do have a template already set up for you here. It's our national drug projections offering. You have the option to name your, your template as well as add a description. And then you have the option to define your configuration at that brand level or molecule level. So, again, molecules are if you're interested in seeing biosimilars into one view or generic and brand options into one view versus brands that would those would be separate. But based on what you select, it'll allow you to key in a search here, and you can add your specific drugs and then generate your dashboard. But I have some dashboards already set up for us so we don't have to sit through the creation process. Today, I'm gonna focus on, Keytruda and Nobdivo. So within our template, we have different tabs to access different views of the data. I'll start here with our methodology. It cut it details the aggregation options that I just went through as well as the different definitions of variables and how they're calculated. And then we have a couple different views that you can access. So first, we have our trend report, which will allow you to see for the different products within your template that you selected. What are the projected outputs, for those brands over time? Just gonna get rid of our filters so we can see this a little bit more clearly. You have a toggle down here if you'd like to, you know, look at more recent data. But although we will be providing you back through, the beginning of the products life cycle if that's of interest to you. Then we have our market share template. It's just a little bit smaller so we can see it a little bit better. This is a 100% projected view of the data to be able to compare market share over time. And again, we have this ability to look at this from different viewpoints down here at the bottom. And I submit should mention that all of these are, exportable if you'd like these images to put into a a PowerPoint or a presentation. And finally, you have our projected versus observed tab. So this provides you on an individual brand level, or product level, depending on what you selected when you set up your template, to see the observed outputs within our, health care map as well as the projected outputs. And I'll make this a little bit bigger so it's easier to see. The confidence intervals are in here in this darker blue line as well. They're kind of make this a little bit bigger so it's easier to see. They're kind of shaded. There's one right there. So you have all of that kind of within your fingertips. And, again, you can look at this over time. And this is really helpful. I know a lot of our customers are interested in understanding some of the projection factors that they might want want to employ within our health care map, this is a view that could help you to accomplish that. Taking a look at some of the filter options, we have, the brand level and the molecule level. For this output, the molecule will just be the generic name. But if, you know, I had selected, for example, biosimilars, we would have adalimumab as one option, which would aggregate up all the different, biosimilars to Humira in one view. You have the option to change your, drug names as well as the metric that you're interested in seeing. So here I'll select units, for example. And so now we're looking at projected versus observed of Devo units over time. So that's our demo. Hope you enjoyed it. It's very quick and easy, obviously, as you can see to get set up and get started. And all of these are exportable if you wanna, you know, play around and and show different outputs, within a presentation. Very quick and easy to get those insights. So where is the future here? We have an evolving projection product suite. We recognize that national drug projections, while very difficult and, we're very excited about them, are probably just the starting point of where we'd like to take this, offering. And so just to give you a little bit of insight into our road map here, from the national drug projection standpoint, as I had mentioned a little earlier, we're going to be adding additional metrics and views over the coming months. So this will include gross sales projections as well as adjusted total dispenses. So this will account for patients that fill, more than a thirty day supply on the pharmacy side. Some sixty, ninety day supplies will account for things like that. We'll also be adding an output, that will help you to calculate the stability of our output of our projected outputs at over time. So I I think I mentioned these outputs will be generated on a weekly basis. And through this view, you'll be able to see how a point in time that calculation changed week over week so that you can start to understand where we start to stabilize particularly on the more recent, weeks and months that we'll be projecting out to. Then we'll be moving down into what we're calling complex events. The still will be a national view of drug projections, but we'll be providing the breakdown by key views that customers are interested in such as, by indication or by line of therapy or by the regimen. And then finally, the, kind of cherry on top to this suite is our subnational projection offering, which we are hoping to launch within the first half of the fiscal year twenty five, time frame. And this will allow us to get down to the individual territory provider or account level, depending on the the level of visibility that you're interested in seeing. So I hope you enjoyed today's presentation. I'm I'm sure we have tons of questions about the methodologies and the outputs, but I am hopeful that, we will be working together, to compare some of these outputs and, that you enjoy them. Thank you, Emily. That was fantastic. Before we head into q and a, we have one more poll question to pop up, and we'll give it a minute for them. This is around what your biggest challenge you have is in projected data today. I'll give it a few more, seconds here. A number of you asked questions regarding, the lowest level of projected data or being able to see the data by account, level. And if you you saw Emily's last slide there, that is coming through in our subnational, view, which is, coming probably in the first half of twenty twenty five. So, coming soon, but this first version of national drug projections is at the national level, just to be clear. Alright? And just just to give a breakdown of, what level of granularity we have in the template itself, it's, kind of aggregated at that brand or model to level. But we do have outputs down to the individual strength or HICS, PICS code, depending on the benefit type if if that is of interest to customers. Yeah. Great. And thank you for your answers on the poll questions. Looks like that that bottom one, that complex and time consuming to pull through, perhaps this very easy dashboard is is an answer to some of your challenges you have today. So let's get into some of the questions that you guys have had. There are quite a few to get through, so thank you and keep keep them coming, as people, as you think of them. So, this one was from from before. What is the granularity of the age breakdown for the national coverage? I can get you the age breakdowns. I don't know the groupings, off the top of my head. I mean, we go through you know, our map has all the way down to newborns up to, 89 and over. But I can get you the actual kind of groupings, if that's of interest. Alright. The, this is an interesting one. You mentioned that you use trustworthiness of different data sources. How do you calculate or measure that trustworthiness? Oh, it's an excellent question. Mhmm. I don't know if any of you have heard my, my boss, doctor Paul Gurney, speak about our health care map, but a a key component of our health care map is the deep characterization of the individual sources that contribute to the map. And so through painful hours of, you know, reviewing, comparing, understanding what we're getting from each individual data provider, we have internal knowledge of which sources that we find to be more trustworthy than others. And again, to that kind of group source characteristic aspect, where they have key advantages versus disadvantages to this model. So for example, again, that payer versus provider complete. We would never wanna use a payer complete dataset to understand payer mix. We we'd get pretty wrong answers. So that's kind of how we leverage the different data sources. Perfect. What about your coverage and projections into rare diseases? Yeah. I probably sound like a broken record now, but, we do leverage kind of that payer complete dataset where we have that minimal floor of visibility. But it will vary by therapeutic area and by drug class. And that's really where those confidence intervals come into play. Even when we get down to the individual account level, we do that on a, you know, project basis, from time to time. That becomes, like, really key to providing to our customers so that you can really understand how trustworthy the output is. So, for example, I had mentioned Kaiser Family Foundation earlier on. We're gonna obviously have wider confidence intervals, for that type of account compared to, you know, a Texas oncology or or something else. But we want to be transparent with you so that you can make the appropriate business decisions. And so we'll be doing that from the national output as well. So my, you know, I don't know every rare disease off the top of my head and what the outputs are, but, the confidence intervals and the widen it the the width of those confidence intervals will really help you understand kind of what our coverage metrics look like. Right. I have a very, couple interesting questions. First, are the numbers from the projections adjusted for lag? They are. Yes. We do adjust for lag. We adjust for lag as well as trend breaks, within the outputs. Right. And then the second this next one is the the projected trend just mirrors the raw data trend. So in terms of market share, how is this valuable? Well, so I guess you could you know, the one that I showed did have kind of that trend that way. I don't it's not every drug will have that kind of trend per se. Again, like, we're not we're not applying a standard projection factor, so each drug will look a little bit different, even within market baskets. Right? We we we could have, like, more visibility into one product than another within within a market basket itself, and so the projection factors might look a little bit different. So I don't know if that answers your question, but that's kind of how we're we're looking at this. Right. Oh, they keep popping up. This is great. All of these metrics are prescriptions and volume. How do you translate it into patient counts? What is the methodology used? I think maybe Yeah. So then the the numerator is kind of that patient closed patient population. Right? That's how we understand the treatment rates, but it also tells us us how many patients we see. And then that denominator is kind of that total US population. So that's kind of how we project this to a patient level. Great. What testing do you do to measure the accuracy of your projection? Take the confidence intervals. Yeah. We are working on an entire white paper writing up kind of the methodology and the and the standards there. We employ a agent inference model. That's kind of the the model that we're specifically utilizing. So we can provide more information about that after today's call. Here's one from the chat that's not gonna do the QA, but the, how is Comodo accounting for products with specialty pharmacy blocks? Yeah. So, again, we we leverage our payer complete dataset. So even for drugs, like, the one that usually comes to mind for me is, Jakafi, where there's very heavy blocking. We do have a minimal floor of visibility within our payer complete data because it's not coming from the pharmacy the specialty pharmacies themselves. It's coming from individual payers. So, again, for those products, what I would expect is that our coverage is, probably lower than a retail product, and our confident confidence intervals may be wider. But we still have confidence in what we're projecting because we do have a minimal floor visibility compared to, some of our competitors that are really relying on those stable patients, but not necessarily the payer complete ones. Oh, I'm going in from every every direction. You guys are fantastic. Is MathLab subscription access a prerequisite for this projected data? Would it be a separate offering on top of Maplab and the health care map? And that's a great question. Yes. I if the Maplab subscription is not necessarily a prerequisite, it can it can be its own product. But if you're already using MATLAB and MathView, you can also add this as an addition to what you're already doing and what you're already using. For the subnational plans, I think this the answer is yes. But for does the does it include getting down to HTP level projections? Yes. It does. So, this piece of the puzzle, the national was kinda is kinda key to the next two phases of our of our evolving product suite. Everything will be a subset of this national ceiling that we've set. And, yes, we can get down to the individual HCP level. Again, once we're getting into those positions that are, you know, only treating, like, one patient or something like that, there will be some confidence intervals, etcetera. But, yes, we'll get down to the individual HCP level. Great. Another one from the q and a, section is is applying these projections to all HDPCS or ICDPCS on the road map or only drug codes? I know HCPCS codes are in there. I do not know about the ICD 10 procedure codes, so I will get back to you about that. Great thing to think about, as we think of future versions as well. So thank you. Yes. How does Komodo's lag compare with competitors? It's again, it's really drug dependent. I I I feel like it's a very unsatisfying answer, but it's true. It depends on kind of, like, the makeup of your of your product and market basket. On the pharmacy side, we have, plenty of data providers that are coming in daily. And then on the medical side, we have those that come in every week or two, up to monthly and quarterly is kind of the the the longest lag that we employ. So comparing to our competitors, we're we're, you know, what's the word? The same. Yeah. Comparable. Comparable. Comparable. There you go. For patients treated, how do you factor off label use or drug holiday? So I'm not sure. Yeah. That's an interesting question. So the complex events, the model too, as we're we call it internally, so being able to break down by indication and things like that, We are purposely going to have that be configurable for customers because we know that most customers see different, you know, see those worlds differently, like define line or define indication, you know, based on their unique business rules. So if we were to work together on a project like that, if we were able to get that through legal and compliance, we could maybe do some off label tagging. Great. Great question. And our final question is really about going more in-depth about the methodology, which I invite you to, reach out to your sales representative if you're currently in conversations with with Komodo. If not, we are more than happy to put you in touch with the right folks to have those deeper conversations about the methodology. And, Emily, if you wanna add anything, to that, please do. Yeah. We're working on a few white papers here in house to help detail that, in more depth. We're also working on a 10 reported 10 q to gross sales estimate white paper as well to really help ground, and contextualize the accuracy of these models. So look out for those in the coming weeks. And, yeah, we'd love to have more conversations with customers about kind of what we're doing and, share samples, get your feedback. This is definitely something that is going to be an iterative process. Another thing that my my boss always says is that here at Komodo, we're always looking towards a continuous improvement. So looking forward to those conversations. Alright. Yes. I see your question about recording. We will absolutely send a recording of this session, to you, within the within the next couple of days. So thank you so much for participating today. We really appreciate you joining us. We look forward to the journey. Yeah. Thank you, Stephanie. Thank you everyone for attending. Looking forward to the conversations. Have a great rest of your day. Thanks.