Blockchain Pharmaceuticals with Jaya Tripathi
Interviewer: Cameron Boozarjomehri
Welcome to the latest installment of the Knowledge-Driven Podcast. In this series, Software Systems Engineer Cameron Boozarjomehri interviews technical leaders at MITRE who have made knowledge sharing and collaboration an integral part of their practice.
Blockchain is everyone’s favorite buzz word. Whether it’s crypto currency or NFTs, the technology has been getting a lot of attention for how it could disrupt how we buy and interact. But MITRE’s own Jaya Tripathi sees a far more critical use for blockchain, tracking drugs. The Pharmaceuticals are a $400 Billion dollar industry, and yet every year large quantities of life saving drugs go missing from the global supply chain. Jaya is hoping to put an end to this loss by leveraging blockchains ability to track complex transactions across even more complex ecosystems. Listen in as she shares her vision of a brighter pharmaceutical future built on Blockchains.
Click below to listen to podcast:
Cameron Boozarjomehri (00:15):
Hello everyone. And welcome to MITRE’s Knowledge-Driven Podcast. A show where I, your host Cameron Boozarjomehri, interview brilliant minds across MITRE. Today, I’m joined by Jaya Tripathi, who will be talking to us about blockchain and pharmaceuticals. Jaya, would you like to introduce yourself?
Jaya Tripathi (00:31):
Hello everyone. I’m Jaya Tripathi, and I’m a principal in the data analytics department at MITRE. I’ve been at MITRE a little over 17 years, and it’s a pleasure to get this opportunity to speak here.
Cameron Boozarjomehri (00:46):
And so would you like to start with telling us a little more about the different kinds of work you do? Because, as I understand it, you wear many hats.
Jaya Tripathi (00:52):
Oh, yes. I have worked in several industries, telecommunications, oil and natural gas, the airline and then I was in the business school. I’m able to bring cross domain solutions, solutions that have worked, have been tried and tested in other industries and I’m able to bring them and apply them to the various FFRDCs at MITRE. Some of the work I’ve done at MITRE have involved policy simulations for opioid addiction treatment, policies, fraud, specifically looking at prescriber fraud schemes and the most recent work was applying Ethereum blockchain solutions to the pharmaceutical supply chain.
Cameron Boozarjomehri (01:37):
So for what we’re going to be discussing today, it actually sounds very synergistic from all these different, you have a strong understanding of the pharmaceutical industry but you also have a deep understanding of technologies like machine learning and blockchain. And as I understand it, your work is emerging of these to track something very important happening in the pharmaceutical industry.
Jaya Tripathi (01:57):
Yes. So some of the problems facing the pharmaceutical industry are counterfeit drugs, being able to track and trace a drug to its provenance. And as the recent public health emergency, the COVID pandemic, showed us, drug supply shortages, so managing supply and demand. In order to address all of these problems, we looked at current solutions and the gaps in them and we wanted to explore to see if blockchain was hype or a hope for this particular application. And so we compared it with the existing technology stack, we had a set of criteria that we wanted to go against, as well as the complicated statutes. And that’s what we set out to do, and we have a working prototype that we’ve developed.
Cameron Boozarjomehri (02:48):
That’s very cool. I think I’d actually want to step back for a second and start with… But I guess what was the inspiring moment for you that said this was the problem we wanted to tackle? At what point in your examinations of pharmaceutical industry did you say, “Okay, this is a real problem we need to understand.”?
Jaya Tripathi (03:05):
So I have been working on research projects on pharmaceutical drugs for almost a decade. So a lot of my research is motivated by events that have happened in my life. So that’s my starting point. Whether it would be a family member like my father who had prostate cancer or somebody in my family who had a skiing accident and then one thing would lead to another. And so that’s what motivates me. In this one was FDA modification to an act which is called the Drug Supply Chain Security Act, DSCSA, that requires the consistent traceability of pharmaceutical products at the package level.
Jaya Tripathi (03:46):
So, so far it’s been implemented at the lot level. So in the logistical hierarchy, lots are much larger than packages. And so this effect, not just drugs that are manufactured in the US, but drugs that enter of the US markets and the US, as you know, is one of the largest consumers of pharmaceuticals. And then COVID happened after that. So these two points were additional motivators and that’s when we realized that the solutions that we’re building can also be used in a predictive way, predicting shortages. So you know the history of where the drug was prior to you as a consumer going to a pharmacy and picking it up from the shelf.
Cameron Boozarjomehri (04:25):
And just to be clear, when we talk about the supply chain, we’re talking about everything from the raw materials through the manufacturing process to the moment it’s packaged into pills and sent to a wholesaler, which you could think is obviously not Costco, but if you think of a Costco where it’s they just have bulk supplies of, in this case, drugs, not, I don’t know, chicken or steaks or whatever, and they’re using that to then sell to other pharmacies that then go into what you would pick up. This is a really powerful problem. I can appreciate why you’d be drawn to this. I guess the next question is at what point were you guys thinking about these different problems and saying, “All right, yeah, let’s use a blockchain, let’s use this distributed ledger that will track all the drugs.”? What was that conversation like?
Jaya Tripathi (05:10):
So we looked at the existing technologies and what’s working and what isn’t, particular in terms of the statute that must be enacted upon by 2023. And we saw a lot of positives about blockchain particularly not needing to have those one to one trust established between all the players. So you have this internet of connections and these proprietary, siloed databases, forms of communication that are not necessarily standards. And so we thought blockchain was a viable one, particularly if there is no single central entity on whom the entire trust lays.
Cameron Boozarjomehri (05:52):
Actually if I can maybe give a little more background on this. So if you want to think of something like a t-shirt, maybe the t-shirt you’re wearing right now, it’s not just one thing that went into it. It’s not like someone picked all the cotton and then turned it into thread and then sewed it into fabric and then made your t-shirt and then maybe screen printed something on it. It’s someone got the materials then they sent it to someone else who wove it into yarn or thread, then that was sent to someone else. And every step along the way, that might be one unique entity interacting with that material and they’re doing one specific thing to it and they’re probably within their own business or company, you’re a thread manufacturer or you’re a fabrics manufacturer. And so what you’re saying is because you have each one of these places being its own siloed entity, it can be difficult to get all their systems to want to talk to each other.
Cameron Boozarjomehri (06:43):
I know we’ve done other blockchain episodes for the KDE, but I think one of the big things to understand about all distributed ledger, which is a blockchain technology, is that you can think of it as a piece of paper that each one of you, every time you record a piece of information onto it, ideally it’s updated for everyone. So if I’m saying I’ve made the thread that will go into this shirt, the next person in the supply chain will know, okay, Cameron has made the thread, and the person who’s supplied the materials to make that thread will know, okay, that’s been turned into thread so they know to send me more and I know to tell the next person to get ready to get it. But it also is really helpful because if I’m a regulator who wants to make sure your shirt is being manufactured properly and doesn’t have weird stuff in it, then I know each step along the way where a new thing happened to make that shirt. And this is critical because shirts go on our bodies, drugs go in our bodies, so it’s probably a little more important.
Jaya Tripathi (07:41):
That’s an excellent analogy. I love your example. In fact, I may borrow it.
Cameron Boozarjomehri (07:46):
Thank you so much. So the next question is because so many people know blockchain is this buzzword, I want to understand a little more about why it was being taken seriously. Because typically I’ll hear people say, “Blockchain is cool. We should do it.” And then that’s about the end of the conversation. Was there anything else that really went into motivating this kind of work?
Jaya Tripathi (08:09):
I begin with the problem, not so much with the technology. So what is the problem at hand? How impactful is it? MITRE does work in the public domain for the larger good, and so we typically like to tackle problems that affect a very, very large population and this particular problem affects everyone. And so I begin with the problem at hand and then I explore the technologies. So it’s not driven by the technology. And then I see what the gaps are, what it is I’m looking for in my solution, and then I explore all the choices. And right from the start we thought blockchain was a very viable one.
Jaya Tripathi (08:50):
And you mentioned about the auditor, the regulator being able to see everything, we found that we could leverage all the positives that blockchain brings, even though we employed a permission blockchain in which certain manufacturers may not be able to see what the other manufacturer has done. We thought this was a really, really viable solution so we explored it. One of the criticisms, as you know, of blockchain is scalability and performance. So that is what we are performing right now. We’ve done some preliminary tests. We know how many transactions per second are needed. In most cases for the pharma supply chain, you don’t need an immediate response. So that is one plus going for us. And we built this proof of concept based on all of these requirements.
Cameron Boozarjomehri (09:40):
To give a little context about what you mean by transactions, so going back to my point about having this piece of paper that tells you what everyone’s up to, it’s not super fast because everyone has to get together and agree this is what everyone is actually doing. It’s not just that I say it happens and then it goes on the ledger, a lot of the time there needs to be some sort of consensus, especially in something I imagine as distributed as a global supply chain.
Cameron Boozarjomehri (10:06):
So a very important step to this is making sure that those transactions are as legitimate and tamper evident, meaning that if someone tried to mess with one of it would be very clear that they tried to mess with one of these transactions so that anyone looking into the network would know that. And I think, to your point, because it is a global supply chain and ships are going back and forth across the ocean, carrying the supplies, it’s okay if there’s…. In this case, it could be like what milliseconds to minutes of delay. It’s not exactly days are going by without people getting their ledger updated.
Jaya Tripathi (10:41):
Correct. Minutes. We haven’t done a full stress test, but even in our preliminary one we were able to do a over 100 transactions per second with no special hardware. I also wanted to mention here in this context for our listeners that the term transactions is overloaded in this context. So what you were referring to earlier were the native blockchain transactions, something that writes and changes the state of the blockchain. And then we also have the transactions in the pharmaceutical supply chain, such as the change of ownership of a drug.
Cameron Boozarjomehri (11:16):
Actually that’s a really important point, because a transaction is supposed to represent that something happened, I turned thread into fabric, and then maybe it also represents that I’m giving this fabric to shirt manufacturer. But what you’re saying is that in the supply chain is literally I got the raw material, I turned it into a specific compound, I did the next thing. So every step along that way where I’m sending it from the person who got the raw material to the person who’s going to turn it into a compound is a transaction.
Jaya Tripathi (11:45):
So just wanted to clarify that the term there was overloaded.
Cameron Boozarjomehri (11:49):
I think that’s a very important distinction. I appreciate that. And it sounds like a big place where this is important is a lot of people don’t want to mess with the drug manufacturing until it’s manufactured. I don’t want to mess with insulin until I have insulin, but once I have insulin and I can sell it on a black market, obviously I’m going to start doing shady things and want to take it. Is that where the bulk of what the benefit of this was going to be comes in? It’s not necessarily that we’re going to just try to make a legitimate drug, it’s that once it’s made, it’s going to go through so many of these wholesalers that we mentioned earlier before it winds up in your hand, that there’s lots of places for it to fall off the back of a truck or for someone to just take a box without anyone noticing.
Jaya Tripathi (12:32):
So 90% of the drugs in the United States go through at least four different actors. So before they go from the manufacturer to ending up on the pharmacy shelf, they’ve gone through at least three hands and a lot of packaging and unpackaging and shipping and handling has happened. And so part of the statute is to be able to track all of this. And in fact, there’s a term called inference, which is you should be able to without just looking at the label or the serial numbers and everything else, the message that’s generated, being able to just know where that package came from. So a manufacturer would’ve sent the drugs, they’re homogenous, they’re all in the same kind of drugs, same manufacturer in a pallet. But once it reaches the wholesaler, they would be opened and repacked into smaller logistical units from the pallets into cases or even totes to meet the demands or the requests from the pharmacy.
Cameron Boozarjomehri (13:35):
And to be clear, the cases and totes, now we’re getting down to the point where is a tot what I would pick up from a pharmacy or what-
Jaya Tripathi (13:43):
Np, that’s a package.
Cameron Boozarjomehri (13:45):
So there’s cases, totes, packages, it sounds like there’s a variety of different sizes that we could be talking about.
Jaya Tripathi (13:53):
There’s a hierarchy. And then the fact that the smallest logistical unit is the package would have been opened and repackaged into different boxes, shipped through many destinations. That’s what makes it hard to simply just walk in and look. And at every stage there’s pieces information that the law requires should be generated. So being able to just look at that entire history from the package and know exactly where it came from is difficult because it’s not homogenous once it reaches the wholesaler.
Cameron Boozarjomehri (14:28):
Like I think we mentioned before, it sounds like just a lot of places where that tracking can break down because you have to account for so many different ways the drug can be packaged and sent off.
Jaya Tripathi (14:39):
Cameron Boozarjomehri (14:39):
So now that we talked about, I think, a lot of the problem and a little bit of what your approach was, I’m really curious to know what was the process like to actually build this? You have lots of actors, lots of supply chains, what was it like?
Jaya Tripathi (14:51):
Yeah, sure. I think at the outset, statute interpretation, so the DSCSA requirements were complex, several pages, and they were debated within the stakeholder community.
Cameron Boozarjomehri (15:05):
And just real fast, the DSCSA, that was the Drug Supply Chain Security Act, the one you mentioned from the FDA?
Jaya Tripathi (15:12):
Cameron Boozarjomehri (15:14):
I just wanted to check in with our acronyms.
Jaya Tripathi (15:16):
Yeah, sure. So interpreting the statute, so we worked with other organizations, a nonprofit that was also working with FDA, as well as some of the private stakeholders, the major manufacturers, and they had formed a consortium and they interpreted the statute for us because what I needed to build a system was very specific functions that could be implemented and where I could just put a check mark across. And we sought to implement a proof of concept that met almost all of them. And then on the technology side, once we had decided on working on blockchain solutions and selected Ethereum, we found there was a huge paradigm shift and we couldn’t leverage our previous scale with bitcoin blockchain, because the similarity between fungible cryptocurrency and the flow of drugs, that similarity is tenuous.
Cameron Boozarjomehri (16:16):
I feel like I need to jump in and unpack a lot of what you just said there. So I feel like a lot of people are familiar with bitcoin and it is a cryptocurrency. You think of it like dollars and cents moving around, which means that every dollar’s worth every other dollar, every penny’s worth every other penny. But what you’re talking about is Ethereum, you can also use it to share Ether, which is a lot of people think of it like bitcoins, and that they’re a cryptocurrency. You can use them. Every Ether is worth another ether, just like every bitcoin’s worth another bitcoin. But the problem is drugs are not Ethereum, drugs are not dollars, they are a unique, specific physical thing. Let’s say every tablet of ibuprofen is similar to every other tablet of the same amount of ibuprofen, but you might have different sized tablets, you might have different quantities in a specific box or whatever, and then you just have different drugs. So obviously the ability to use just this idea of cryptocurrency breaks down very quickly because it’s not currency, it’s each unique pill.
Jaya Tripathi (17:18):
Correct. It’s not a fungible, it’s an actual item. And also I think this is more complicated than bitcoin blockchain because the state is simpler there. It houses the list of balances each person has. So you have keys from the sender, the receiver and then you have the currency, but what we wanted and what Ethereum blockchain gave us was the ability to store all kinds of stuff. So here we have the logistical hierarchy to store. It’s not just the accounts, but we also have contract instances and there’s a lot of other associated data involved in it, storage and Ether like you mentioned. We found that the solution that we selected was able to tick off almost all the requirements that we had.
Cameron Boozarjomehri (18:09):
Actually, I think requirements is somewhere I’d like to end this because I think something people don’t always realize is okay, government says we want to do something better and then it just happens. And I feel like most people who watch government do anything realize that’s not usually what happens. And when a government organization says we want to improve supply chains and they make a law for it, a lot of the times that law is actually not very straightforward or even really considers how modern technology works. And so it sounds like, especially when you were working with these other organizations and nonprofits, that a big part of your job wasn’t necessarily building the cool new blockchain, it was actually just figuring out what the heck they were talking about in the first place so you could figure out what the requirements for what eventually might be a blockchain would be.
Jaya Tripathi (18:56):
Exactly. So they should be unambiguous in something that’s actionable and that’s measurable. Having that was the first step. Absolutely.
Cameron Boozarjomehri (19:06):
And was there specific things in that actionability and measurability that either were lacking or that you found to be surprisingly easy or difficult?
Jaya Tripathi (19:16):
One particular one that I wanted to talk about was the business requirement, which is in the statutes, that the implementation of roles and based on the roles who gets to see what part of the blockchain, as I mentioned before, this may sound like it’s going against the grain of what a public leisure stands for but we were able to leverage all the positives the blockchain brings and yet able to implement this by using a permission blockchain. Depending on whether you’re a manufacturer or a wholesaler, you may get to see the whole history of a product only if you was in the chain of that particular product. So, for example, if I’m manufacturer two and I’m looking at a product that I was never the ownership in the history of its ownership, I was never the owner, I don’t get to see that. But if I’m an FDA regulator, I should be able to see everything. Furthermore, even if you are in the chain, you don’t get to see what happened to it once you send it to the next person.
Cameron Boozarjomehri (20:14):
I was actually going to say, “I feel like what you’re saying is actually very understated.” I find in my own research, in my own work with blockchain and other distributed technologies, one of the biggest problems that industry has for wanting to adopt this, especially when you’re dealing with global supply chains, is transparency is antithetical to modern business. A lot of the times you’ll think of Facebook and Google or any large organization typically doesn’t want to share their secrets with other organizations. And a distributor ledger is basically saying, “We’re going to put as much out there as possible to get information to be as transparent as possible, which ends up making a lot of people hesitant to adopt the technology.” And it sounds like in your work, a permission blockchain, it’s not like a bitcoin or Ethereum in that you are not necessarily showing everyone every piece of information that happens.
Cameron Boozarjomehri (21:04):
You’re able to create these transactions even in the Ethereum network, that if I am the person who has all these raw materials and I send it to someone else, it’s not my business what happens next. If they’re the one who placed the order, then I’m just going to send it to them. Let’s go back to that shirt analogy, Target doesn’t need to know what I do with my shirt after I buy it. If I wear of the shirt, if I turn it into, I don’t know, a hat, if I turn it into a towel, that’s not really Target’s business and they shouldn’t need to know what I’m doing with it once I’ve bought it. And likewise, Target doesn’t really need to care or tell the person who made the shirt that sold it to Target what they’re doing with it and all the way down. So this is powerful in that you were saying you’ve built a system where everyone in the system can be held accountable but none of them have to give up the things that they’re so keen to keep secret.
Jaya Tripathi (21:59):
That’s right. So not only do drug companies want strict controls over who can see what to protect their competitive interests, as you mentioned, the law, the statute that I mentioned before, the FDA statute, also imposes these control constraints, so it’s actually requirement by law as well.
Cameron Boozarjomehri (22:16):
And then on top of that though you still have carved it out so that if a regulator needed to see into the network, they could still have that full transparency.
Jaya Tripathi (22:24):
That’s right. So in our demo, we actually demo different entities or actors in the supply chain logging in who have different roles and if they all want to see the same item, let’s say it’s item number 18, and so depending on who you are, whether you are the manufacturer of that drug who was involved in that particular package being sent, shipped in the pallets. So we show manufacturer one, manufacturer three, wholesaler two, regulator, and then we also show one of the blockchain consensus validator, the blockchain administrator, and you’ll be able to see what parts of the blockchain you can and cannot see. In some cases, you’ll see nothing at all.
Cameron Boozarjomehri (23:11):
That’s really powerful to be able to see all those different views. That’s amazing. I feel like it can be hard to wrap your head around just what a paradigm shift, what a game changer this could be for the drug supply chain, not just in the United States, but around the world.
Jaya Tripathi (23:30):
Sure. And we are used to, from our training in computer science, one particular way of thinking we’re used to being able to do all four crowd operations in traditional databases, or we are used to having a whole bunch of data objects available to us and thinking differently, but finding workarounds, alternatives of what was there and finding the solution, it was very satisfying-
Cameron Boozarjomehri (23:56):
Jaya Tripathi (23:57):
… When we were able to do that.
Cameron Boozarjomehri (24:01):
That’s great to hear. Is there anything else you’d like us to know before we go? Or is there anywhere we could go to learn more about either your work or the kinds of work that we should get excited about so that we can also contribute to this glorious blockchain future?
Jaya Tripathi (24:12):
Yeah. So two things I wanted to end with. So let me answer the question of where you can go. We look forward to sharing our findings in a white paper, which will be available through the MITRE public facing site, and also working with others who are in this field who may be listening and would like to collaborate with us. And then the other question is we want to extend this to supply chain and other domains. So we think the solution is very extendable. Like I said, the recent pandemic, when we had the supply chain shortages for things, even simple things like masks, and people had to wait for months to get a hold of masks, if we could be proactive and use our solution and apply predictive analytics methods, combine that with this, that’s where we want to go. We want to not look back. We want to look forward and be able to prevent some of these situations.
Cameron Boozarjomehri (25:09):
Before we go. A big thank you to MITRE and the Knowledge-Driven Enterprise for making this podcast possible. And a huge thank you to you, Jaya, for your incredible work and for taking the time to share it with us. This has been one of my favorite conversations. Thank you so much.
Jaya Tripathi (25:23):
Thanks, Cameron. Thanks for the opportunity.
Cameron Boozarjomehri is a Software Engineer and a member of MITRE’s Privacy Capability. His passion is exploring the applications and implications of emerging technologies and finding new ways to make those technologies accessible to the public.
© 2021 The MITRE Corporation. All rights reserved. Approved for public release. Distribution unlimited. Case number 21-3414
MITRE’s mission-driven team is dedicated to solving problems for a safer world. Learn more about MITRE.