Accenture, Merck and Amazon Web Services (AWS) Collaborate on Cloud-Based Informatics Research Platform:
Accenture and Merck, collaborating with Amazon Web Services (AWS), has developed a cloud-based informatics research platform designed to improve productivity, efficiency and innovation in the early stages of drug development.
The platform will enable an ecosystem that accelerates innovation by creating open, industry-standard application programming interfaces for core research functions, allowing researchers to rapidly adopt new capabilities. Rx Data News is pleased to bring you an interview with two key players in the endeavor.
Managing Director, Life Sciences R&D, Accenture Philadelphia, PA
Joe Donahue is a Managing Director with Accenture’s Research and Development Services business where he leads Accenture’s global life sciences research practice. Mr. Donahue has more than twenty-five years of executive level entrepreneurial and Board experience in global life sciences and technology companies and private equity firms. He has degrees in Medicinal Chemistry and Computer Science from Villanova University in Villanova, PA.
Executive Director of Applied Technology Merck & Company
New York, NY
Hal Stern is Executive Director of Applied Technology at Merck & Company, where he focuses on building services and applications to define and shape the data-oriented and data-influenced adjacencies to Merck’s core markets. Hal was previously a VP at Juniper Networks and spent more than 20 years at Sun Microsystems. He holds a BSE degree from Princeton University and holds four issued and several filed patents in security, identity, user experience, and networking.
Rx Data News: Would you both provide a brief description of your background for our readers?
Joe Donahue: I am a managing director at Accenture’s Life Sciences Business. I’m a scientist and chemist by education and have been with Accenture for several years.
Hal Stern: I lead the Merck Research Labs IT Engineering Team, which is responsible for the design, development, deployment and data governance of all of our IT systems on the research side. This covers everything from very early discovery all the way through to clinical trials, pharmacovigilance and quality. I have been at Merck for about five years and spent twenty five years before that working in the technology industry.
Rx Data News: What about your backgrounds led you to see the need for a project of this kind?
Joe Donahue: My background as a scientist is really relevant, because I’ve lived this as a researcher in the laboratory. The scientific process is about generating data. It’s as much about informatics and understanding data today as it is about underlying science. It’s really hard to pull that data together. Scientist are working in an environment in which the systems they use are completely disconnected, the data is going into silos. In many large pharma and biotech organizations even basic chemical entities and molecular structures in biology aren’t integrated in a way that researchers can look at that data to understand if the molecule they’ve designed is actually achieving the biological activity they were hoping.
What we’ve seen is that this is still a very long process, it takes more than a decade for a new drug to be approved and costs north of $2.5 billion with a 90% failure rate. If we can cut down the amount of time that it is taking us to look at this data and also take advantage of all the new data that is being generated now, for example in genomics, gene editing, wearable devices are now generating data that is being fed into the research process, if we can create a better environment where folks don’t have to worry about how to integrate their data, how to access it or where it is and focus on the science, we really believe we can have a significant impact on a process that now takes over a decade. Even a small, incremental process potentially could have real, tangible results in the length of time it takes to get a new drug to a patient who needs it.
Hal Stern: I’m going to give you a radically different perspective of this from the other end of the spectrum. It’s an interesting question, what in our history makes us believe that we’re working on something important? I’ll start with Joe’s statement about the need to decouple data, to unhook data from the application that created and managed it and to separate it into multiple layers so that it is accessible to other applications that they may or may not have been involved with creating the original data.
We can actually assemble those things together. I’ll point to two things that I was involved with previously that come to mind when I think about our investment in this research life sciences platform. The first is the creation of the Java programming environment. Up to that point, we wrote programs that ran on servers and desktops, but this idea that we would be able to write programs that could pull from multiple network locations and could assemble them together into a different experience on your desktop had never been done before.
Obviously, we’ve seen Java go through multiple iterations, but the very first programming paradigm fundamentally changed the idea of where code ran and how you assembled together different facets of it. To me that is the solution part of all this, which is to say that we are going to move from purchasing applications that are purpose or function specific into thinking about assembling functions together to create time and value competitive workflows.
The other thing it reminds me of, which occurred a couple years later, was the creation of Major League Baseball’s MLB.com. What happened there is that you had a sport that generates a tremendous amount of data, baseball has always been very rich in data, and what MLB.com did was to take all the data generated from the game and make it accessible in a new and novel way. They took data that historically you got from watching the game live or reading the box scores in the newspaper the next day and put it online and made it available for stakeholders to go build applications from it and what happened afterwards is they had the remarkable new set of ways for people to engage with the game of baseball.
I believe they fundamentally changed the way in which people interact with this national pastime because they took the data around it and sliced it up and made it accessible in new and interesting ways that didn’t involve watching a game live or watching it on TV or waiting for the paper the next day. Our new research platform does all of those things, it forces us to think about creating value through assembly and frees the data for other applications.
Rx Data News: Could you describe for our readers what exactly your new cloud-based informatics platform is and what it does?
Joe Donahue: We’ve announced a number of things, the first was a collaboration with AWS to co-create this platform on AWS architecture. It is important to point out that AWS is a vested partner in this so we’re not just using the AWS cloud environment, but that they are working on the project with us. We have access to their engineers, access to the latest technology on their platform and are working to optimize it for different types of data that we handle as part of drug discovery operations.
The second thing is a partnership ecosystem that includes technology providers, content providers and other services organizations that are providing capabilities and solutions to pharmaceutical researchers today. We’ve created effectively an environment to make it easy for them to plug and play and embed their capabilities into this platform. That benefits both those technology and content providers usually by providing an easier and more cost effective way for them to go to market. It opens up some really interesting opportunities from a business model perspective as well.
The third thing is the coalition of pharma companies that will eventually be involved. We’ve announced this initiative with Merck as our first client but it is the intent of all the parties involved, including Merck, that they will not be the only party to use the platform. Much like other platform initiatives that we’ve launched at Accenture, with this pre-competitive business model we expect other organizations to jump on board as well.
Hal Stern: I would characterize the platform as a research platform more than an informatics or analytics platform. While those are important, really our emphasis is, let’s understand some of the basic resource types that are used as part of the scientific method and then look at how we can accelerate the scientific method by encouraging a diverse set of partners, some from the industry and other more emergent players who have some really creative ideas about what to do with the data to improve the efficiency, speed and transparency so as to allow scientists to do science at scale.
We want to build a platform that will have a lot of the key data producers in it. It will also be an attractive place for new applications to be targeted. You want to have both, you want to have the best data and richest set of data resources available because that makes the applications interesting and you want to have the richest set of applications because that will attract more and more data providers and more and more core components to your platform.
Our initial challenge is to figure out how do you seed that. Apple created a phone operating system with a core set of applications in it by integrating mail and contacts and a camera and a phone, then along came Snapchat and Facebook which fundamentally changed the way we told stories with pictures. It is this ability to assemble on a common platform that consists of both core components and new applications that we would like to do for a much more data rich and computationally intense environment.
Rx Data News: What were the most difficult challenges you faced while developing the platform?
Joe Donahue: This represents a fundamental shift in terms of how pharma companies work with technology, data and applications. We’ve spoken with almost 50 pharma and biotech companies over the last two years to understand the challenges from a data handling perspective, from the utilization of new technologies such as artificial intelligence and machine learning, to how the companies collaborate with one another, in order to develop a strong understanding of their needs so that we can be confident the platform will address them.
The second part is to understand that this is a two-sided marketplace. It’s absolutely critical to the success of what we’re building and implementing that we build strong relationships with the technology vendors in the space that we’re talking about. We need to educate them about what we’re trying to do and to get them to buy into it.
As part of this, we’re creating standards for the capabilities that researchers need. We recently released some API’s for small molecule registration and published them for all of the vendors in the space that have applications or solutions for small molecule registration. There are a lot of drugs related to small molecules and managing that data is really critical.
At an AWS event at the end of November, three of those companies came up to us and showed us how they implemented their technology to these standards that we published and how their systems can plug and play in this environment. So getting the vendors involved and excited and helping them see the opportunities that exist has been one of these challenges. We’ve made great progress in a very short period of time, we’re actively engaged with more than 80 technology providers and more than 30 of them have signed up to be part of this partnership ecosystem that we’ve created which is the other half of this two-sided marketplace.
Hal Stern: When you talk about entertaining a shift in the way that technology works within the marketplace, clearly the new players like this because you are lowering the barriers of entry for them. For a company that has a great idea about how to apply machine learning to optimizing the drug discovery process, being on our platform benefits them because now they don’t need to worry about selling to each individual company, they don’t need to worry about how they can integrate with any number of data standards or foundational systems.
They can integrate with our core data system and off they go. I’ll go back to the notion of a two sided marketplace, there is a bit of rate factoring that happens here which is, what parts of the ecosystem are core service providers, they actually fit into the platform, and what parts of the ecosystem are actually applications that assemble that value together. Those are the more difficult, but also the more opportunistic conversations because it allows companies to deliver more value in more ways. It allows these software companies to reach a wider variety of customers than they would if they were a unit on a proprietary cloud environment. There’s a great set of opportunities there but it is also a challenge because I don’t want to be competing with them, I want to build a platform that encourage people to make software available to scientists.
We don’t want to be in a position where we’re competing with our partners, it’s never a good business model. So we’re trying to be very, very clear that this is not about competing with existing players, it’s not about competing with emerging players, it’s about trying to lower the barrier of entry to create a software ecosystem.
Rx Data News: Could you elaborate on what an open partner ecosystem is and why it is so integral to the platform?
Joe Donahue: We are creating standards for the capabilities that researchers need - things like accurate data capture and pulling data out for artificial intelligence and machine learning, small molecule registration and large molecule registration, and more. As an example, we recently released some standards for application programming interfaces for small molecule registration. These standards will provide any tech company that provides this as a solution to plug into the platform that we’re developing.
This will allow disparate systems to work together through standardized data and send that data into a single repository that researchers can access without the manual intervention that happens today with pharma companies. And we’re not just publishing APIs or micro-services for specialized capabilities, but also there are some core services we are providing on the platform such as getting data out. One of the biggest challenges for leveraging things such as artificial intelligence and machine learning is actually getting access to all of your data in a reasonable format so you can actually do something with it.
Hal Stern: I think we overuse the word open quite a bit. To me the point of open is we publish it and we let you use it for free. We’re not going to charge developers a tax to build to our platform, they can still make money off selling software. You can build to the Amazon APIs and Amazon doesn’t charge the developer, they encourage you to do it. We’re encouraging developers to not only work with our APIs but also to help evolve them. That is the point of the ecosystem. Our goal with building the first set of programming interfaces was to make them as simple as possible and then to eventually get more specialized as we can’t predict right now all the ways scientists will eventually use the platform so it will naturally evolve. We will figure out over time what the desire paths of our platform are and it will evolve over time.
Rx Data News: Do you see research platforms like this having uses for pharmaceutical companies outside of the R&D process?
Joe Donahue: From my perspective, absolutely. I think the pain is more pronounced in the research phase just because of the number of systems, the number of applications, the number of vendors, the diverse, heterogeneous data types that we are trying to manage, it’s a really hard problem, and that’s why we have placed this into focus. The length of time it takes to develop a new drug needs to be shortened and we’re hoping to do that through enabling efficiencies and innovation. But there is no reason that this approach, leveraging a platform to enable data integration, can’t be applied to other parts of R&D, development, clinical, etc.
Hal Stern: There is a lot of hard work that has to go into making a platform like this successful, from the architecture and engineering of the services, to building the ecosystem, to really understanding all its impacts on the market. So it is very much research and life sciences first. That being said, if you look at the patterns that emerge: How is data really shared? How do we do better data governance and better kinds of data governance? What are the patterns that we look at in terms of new kinds of applications that are assembled? It may become an attractive target for other parts of the business. They may say we have the exact same type of data problems. We’re starting simple and hopefully will discover a lot of really great use patterns and again, we’ll publish them, we’ll make them part of our ecosystem and I’m very interested to see where it goes.
If you found this content valuable, please consider subscribing to Rx Data News to receive monthly updates on the state of data analytics and machine learning in the pharmaceutical industry.