By: John Seaner, CMO, IPM.ai/Swoop
Challenges are an inevitability when building a drug portfolio in the rare disease space, and patient finding is no doubt chief among them. Clinical trial recruitment is highly capital intensive, and only more so for specialty and rare diseases. Patients are dispersed geographically, and most are unaware of their actual condition; going undiagnosed or receiving a misdiagnosis is a common outcome along the patient journey. According to a report on patient recruitment and retention by business research firm Roots Analysis, approximately 86% of registered clinical trials fail to enroll patients by the target deadline, 37% of clinical research sites do not meet their recruitment goals and more than 10% are unable to enroll any patients at all.1
John Garcia, former Vice President Global & US Business Operations at Alnylam Pharmaceuticals faced this exact predicament while working on ONPATTRO®, the only available therapy for hereditary ATTR (hATTR) amyloidosis, an amyloid disease with a diffuse symptomatology that resembles an array of other conditions. Adding to the confusion around this disorder is that electronic medical record (EMR) claims for ATTR appear very differently from its clinical presentation. Wanting to reconcile this consistently inaccurate data with the clinical symptomatology and gain a better understanding of the disease as well as the typical patient journey, Garcia sourced IPM.ai to make sense of the reoccurring signals within the claims, hone in on how the disease actually presents and uncover physicians likely caring for treatment appropriate patients. Key to this process was having the field reps spend time educating the health care providers (HCPs) about ATTR to help them accurately recognize the disease.
There was flexibility within Alnylam, in part, because no previous commercial brand existed. This afforded Garcia and his team the freedom to engage with IPM.ai – an organization they first became aware of through word-of-mouth – to solve the problem of patient finding. Internal expectations were set, and the AI/ML technology was positioned as a targeted approach to increasing HCP awareness of the disease.
The first step to uncovering likely ATTR candidates was for IPM.ai to score a universe of 300 million de-identified patients based on their resemblance to an ideal, true patient population. Alnylam questioned and refined patient markers throughout the process, taking genetic indicators into consideration for more accurate results. The resulting output benefited the company’s understanding of the disease and the patient’s journey over the long term.
While patient finding is a consistent challenge for rare disease indications, predictive platforms devoted to sourcing candidates and their physicians present a viable solution that is finally accessible for the first time. The unprecedented potential of AI/ML for uncovering lookalike patients coupled with an extensive real world data network of over 300 million de-identified patients, allows near real time insights to be shared and is pivotal for drug development. Garcia recommends using IPM.ai’s technology as early as possible to help remedy some of the inefficiencies that have become mainstays in the pharmaceutical industry and to speed up all commercialization endeavors. Going forward, Alnylam plans to amend their current predictive model in line with an expanded indication within the same therapeutic area and will continue to rely on IPM.ai as more developments are advanced through 2025.
- Roots Analysis. (2021, March). Patient Recruitment and Retention Services Market (2nd Edition) by Therapeutic Areas (Cardiovascular Diseases, Oncological Disorders, Infectious Diseases, CNS Disorders, Respiratory Disorders, Hematological Disorders and Others), Patient Recruitment Steps (Pre-screening and Screening), Trial Phases (Phase I, Phase II, Phase III and Phase IV), and Key Geographies (North America, Europe, Asia-Pacific, Latin America, MENA, and RoW) – Industry Trends and Global Forecasts, 2021–2030. https://www.rootsanalysis.com/reports/view_document/patient-recruitment-and-retention-services-market/245.html
Read on for the full interview between John Seaner, CMO of Swoop/IPM.ai and John Garcia, Vice President Global & US Business Operations at Alnylam Pharmaceuticals.
John Seaner: Let’s start out with you telling us a little bit about yourself and your role.
John Garcia: I've worked at Alnylam for about five years and, prior to that I was at a variety of different biotechs. In a previous life I worked in medical device and was a part of the Marine Corps. My more recent career has been focused on building out a commercial infrastructure, technology, capabilities, and teams across the globe for Alnylam, proceeding our three, and then hopefully more, launches coming in the near future.
John Seaner: That’s so great to hear. So, you're building out this team and a drug portfolio, and you obviously have some primary challenges, and IPM certainly helps and assists you with overcoming those. What was the primary challenge that led you to IPM when you realized there had to be a different approach to the way things were being done?
John Garcia: I think the primary challenge is one that many rare disease companies face, and one that we faced particularly early on, and we continue to. When we were going into an area that had never seen a drug before, and that therapeutic areas is what we call TTR, which is a form of amyloid disease. Within TTR, particularly in terms of the clinical symptomatology, it's rather diffuse and often looks like other things. There's also an obscurity when it comes to the data, which is derived from the clinical engagements with patients. What I mean by that is, while you can have a patient who shows up with TTR, how they show up via information within EMR claims or other sources, oftentimes looks very different than the clinical presentation.
I know that's something that many people have likely faced at other rare disease companies, but what we were trying to think through was, how do we better understand the data that's related to the clinical symptomatology? Albeit it’s not necessarily accurate, it's consistently inaccurate and there are also particular signals which may be meaningful to our understanding of the disease and how we can help patients. Furthermore, in engaging with IPM, it was really about understanding what some of those clinical data symptoms might look like within the context of the information – how we connect those things together and extract meaning from them. In other words, how do we get to a point where we have a much higher likelihood that if we're engaging with a particular physician, that there might be a putative patient within that office that might have the disease. That was intended for us to drive incremental efficiency within our field sales teams, as they were in all our broader teams, because now we were spending our time to best educate physicians and staff and hopefully help those patients get recognized more quickly than they otherwise would.
John Seaner: Was there a breaking point that necessitated this change? Did you want to try something new because the old approaches weren’t working, or did you say, I have to change this up a bit?
John Garcia: I don't think it was necessarily a breaking point but having spent a lot of time in rare disease, it was, and it's always been, a challenge. It wasn't a new challenge for me. The solution, however, was evolving technology within IPM.ai and the broader space, which allowed for much easier access to information, particularly in the United States. Connecting that information using technologies allows us to understand the patient experience much better over a longer period of time. A lot of those things previously weren't possible because of many of the aspects of heavy computing in addition to ML or AI, not to suggest that we're necessarily doing those things in a pure way, but in some respects, we are – five years ago, those things required tons of money and computing power. All these things have become much more commoditized and egalitarian, where people have these skills and the tech to do it. Now we can do a whole bunch of things that otherwise we weren't able to.
So, it was a problem that we had experienced at several places and within several therapeutic areas. We knew that we had a very low understanding of the marketplace and that we needed to enrich our understanding through actual patients, obviously not identified patients, but actual patient information. That way we could say, okay, well, this is what we understand the disease to look like, but what does it actually look like for the patient both from a clinical symptomatology, as well as with respect to their medical data?
John Seaner: Got it. We have a lot of clients that tell us when they first started embarking on their machine learning and artificial intelligence journey that we had to do a lot of internal education to set expectations and kind of educate the organization on this way of doing things. Did you have to do any of that or was it pretty seamless?
John Garcia: Yeah, I think it's a little bit different for an Alnylam because there wasn't a previous commercial footprint. There wasn't a long-term and entrenched interest around X, Y, or Z activities and our teams. It was me and a couple of people sitting around the table going, okay, we have this problem, how do we solve it? Okay, IPM.ai can do this – let’s go. Once we started to do it, setting expectations became really important because one of the things that is easy to do is, when you start using terms like predictive analytics, cognitive computing, AI/ML and these kinds of catchphrases, most of the time the people saying them have very little understanding of what they actually mean. Many people might see them as a panacea, rather than just a piece of an overall approach to educating physicians and hopefully helping patients get diagnosed faster. I would say for anybody who's listening, if one is in a position where you've gained traction, it's hugely important that you let everybody know that it's not as though we're just going to find every patient and it's that easy because there are so many different, not only patient phenotypes but also how those show up within information. It's always a learning process and it takes time. There were times where people were kind of singing the glories of all of the things that we had built, and we hadn't even really gotten to proof-of-concept yet.
John Seaner: Did anything surprised you though during that time, did anything jump out at you that you didn’t expect to be derived from these projects and this whole notion of patient finding?
John Garcia: I think, in respect to how we've worked with IPM, I don't think there have been any surprises on that front. It's all been good. I think people aren’t aware of the aspects of data science and the complexity and difficulty of even getting a training cohort. If we had gone back, what we should have done is built that in within the clinical trial protocol but because the data were blinded, we didn't have that ability. If one is way out ahead of a phase III or a phase II/III, I would really think about it, particularly in a rare disease that doesn't have a therapeutic in the marketplace and figure out how one would key some of that patient information, obviously remaining blinded, into a place where one could do something with it. I think the most challenging aspect was figuring out how we identify the true positive and true negative patients and then use them in order to train the predictive power of the analytics that we were doing; that was tough.
Once you get into it, I think that the aspect of refining is important and taking the time to question some of your initial assumptions. Once you start to get patients on-drug and you have a commercial promotional team, you're going to learn a lot from those teams. So, listening, trying to internalize some of those things; if you have patient services, understanding how you might be able to use opted-in patients via patient services and some of these data. I think that the biggest part is not to become too enamored with what one has built initially, particularly when you're coming into a market that's kind of a blank slate because you're just going to learn so much and you have to pay attention and not take for granted what you've done before.
John Seaner: Any thoughts about any expanded use or using this earlier, maybe for true clinical study or even for the drug development?
John Garcia: Yeah, I think it's a no-brainer. I don't even know how many millions were spent on patient recruitment and I would imagine it was in hyper traditional ways. I believe those traditional ways are massively inefficient, particularly given the access to information that we have in the US; outside the US it's a much different question. I think there are lots of different approaches and use cases much earlier on, and it would've vastly helped us in respect to commercialization as well.
John Seaner: Anything upcoming that you can share that you're utilizing this data for or some new kind of angles for utilizing this data?
John Garcia: We're at a different stage as a company because we've really kind of built out a lot of these things for the products that we have in line right now and then we're going to be expanding in the same therapeutic area. So, in respect to the development of the predictive model, there will be some “unmet-ments” there, or I would call them adjustments, because, we'll likely have an expanded label and be able to do things that we don't do now. Then of course, as we kind of get through this latter part of 2022 - 2025, there'll be other things that will come into play later on.
John Seaner: Got it, wonderful. How did you find out about IPM.ai? What led you to it and how did that whole process work?
John Garcia: I'm pretty sure like so many things in life, it starts with somebody who knows somebody. As I recall, Marisa [Kelso] on my team knew Jonathan [Woodring] from another life and Jonathan had recently come to IPM.ai, so that's where that started. I'm sure if that connection hadn't have happened, I still would have been tracked down.
John Seaner: What more do you think we could do as an organization from a support perspective? I know we talked a lot about metrics and quantifying, but I wanted to kind of get your thoughts on that because now we have access to more resources, data and people.
John Garcia: Like so many things in biopharma, there are incumbents and there are new players, new entrants, and I think there's a number of incumbent areas that need quite a bit of work. I still think there's a large gap in terms of understanding how some of these digital and other interactions are actually working, or not. I think it’s easy if you're a digital marketer to report success on all digital activities, however, understanding what that multichannel mix looks like and its effectiveness is really important. I haven't seen a great solution around that yet.
I also think that when it comes to the integration of some of our activities with other parts of the company, whether it be clinical operations or developmental, I also think there's a lot more to do there. As you think about the broader ecosystem of initial therapeutic targets all the way through commercialization, the “customer base,” and I'll characterize that as the patient community, the family support, HCPs, the patients – it remains the same, that doesn't change. I'm not sure if there's a clear understanding of the marketplace over a much longer continuum and I think that that a lot of that starts with information. One would suppose that by the time we're halfway through a phase III, instead of running out and having to do a bunch of market research, most of that has already been done. One might qualify that as a role of a new product commercialization team, otherwise, but oftentimes they don't come in until later in the development, particularly when a company doesn't have a commercial team. So, how does IPM.ai help there? I think there's a real opportunity to facilitate that understanding earlier through the information you provide.
John Seaner: Do you do benchmarking? How quickly do you get patients on therapy, leveraging these types of solutions? How are you measuring efficiency in terms of the rep’s time and engaging with the right physician at the right time? What are your sales goals and are you getting there faster? Are those types of measurements being done or, or could that be more valuable? Could we help support that?
John Garcia: Potentially, I think we'd have to look at each of those use cases. I think that we look at many of those things across a variety of different areas. The forecast is an easy one because we know what our long-range forecast is, and whether we get there or not is obvious. In respect to rep efficiency, we look at a myriad of different areas, but again, I think it's one of those things where there's an opportunity to develop a much more holistic approach and an understanding not only of how the marketing and sales work, but also personal interaction.
John Seaner: We have a strong partnership and, and we'll continue to keep building and growing, thanks again.
John Garcia: Thank you so much, take care.
About the Author
Chief Marketing Officer
John leads a team charged with amplifying our brand, communicating the organization’s business value, and earning the loyalty of our clients. He has a 30 year track record of driving innovation, operational improvement, profitable revenue growth and sustainable competitive advantage for technology companies ranging from early-stage startups to global billion-dollar public organizations. John previously served as Chief Marketing Officer of 1010data, provider of analytical intelligence, consumer insights and data sharing solutions; Chief Marketing Officer of Signals Analytics, an early pioneer in the emerging Decision Science as a Service space; and Vice President of Global Growth Marketing at Medidata Solutions, the world’s leading clinical development lifecycle platform utilized by over 1,500 life sciences companies. Throughout his career his teams have earned notable awards and accolades, most recently for pioneering new approaches in account-based marketing, including the use of digital ethnography to uncover the customer decision journey, the utilization of data science to optimize demand activation and the application of signals intelligence to enhance customer empathy.