Giving it Your Best Shot: How Working on a COVID Vaccine Revealed Three Keys to Clinical Efficiency
While you may have heard that it took just two days to design some of the major COVID vaccines, the lesser-known story is how these companies scaled those benchtop discoveries to over a billion doses worldwide in months, not years.
How in the name of Dr. Anthony Fauci did this happen?
I spoke with Denis Kharlamov, EPAM’s Senior Delivery Manager who is leading projects for one of the major COVID vaccine developers, to hear more about EPAM's partnership in the journey of clinical optimization. He shared the three keys to process optimization for vaccine producers: (1) automation, (2) reproducibility and (3) traceability.
Automation
While many scientists and CTOs assert that the lab of the future will require some degree of automation, there’s less consensus on what needs to be automated and when to invest. To understand the best opportunities for automation, we employed our craft vs. stuff framework to assess which lab tasks classify as dutiful drudgery as opposed to delightful discovery. One area ripe for automation emerged: The not-so-glamorous tasks around inventory management.
“You need to know what you produced, when you produced it, quality results, in which container or vial it resides… and when it needs to be replenished,” says Kharlamov. “This kind of process should be automated.”
EPAM worked with one of the major COVID vaccine developers to design and develop their inventory management application, which uses lab IoT scanners on fridges and freezers to track the location of inventory without requiring extra documentation tasks for scientists and technicians. In addition to inventory management, Gaurav Rohatgi, EPAM Continuum’s Life Science Vertical Co-Lead, sees future applications for assembly line monitoring from the world of manufacturing entering the lab.
Craft vs. Stuff
From firefighting to litigation, we’ve learned that each occupation attracts people who resonate with core skills or craft required to excel in a given field.
Each job, however, requires a certain amount of drudgery or tedious minutiae we like to call “stuff.” Ethnography enables us to separate the craft from the stuff which in turn allows us to prioritize and empower a user’s craft while automating or removing the stuff that detracts from their experience.
“There's room to borrow from industry 4.0’s high-volume manufacturing, where there's a real sense that you have to instrument each critical step of the process,” he says. “With computer vision solutions like Drishti, you can actually create systems that can track if a particular protocol is being followed from a human perspective.”
Reproducibility
Our lab observations revealed that scientists often espouse the importance of documentation in theory but not in practice. Despite this theory-praxis gap, proper record-keeping can greatly improve the reproducibility crisis that many biolabs face.
Sridhar Iyengar, CEO of Elemental Machines, describes the opportunity created by this reproducibility crisis. “When you look at how modern software is written, you cannot write modern software without tools like debuggers… if you look at scaling up production from the lab to pilot production to mass production, it is just a physical process,” Iyengar says. “And there's a lot we can learn from the software industry for a virtual process, but what physical processes lack is the ability to capture the state of what you're working on and every single step of the process. For Elemental Machines, our ultimate goal is to create this wonderful platform for debugging physical processes.”
To improve documentation and quality assurance, EPAM supported a vaccine developer to create a custom electronic lab notebook to track hypotheses and outcomes continuously, with a portal for QA and legal teams to validate results.
Traceability
This validation requires both data integrity and traceability, which are some of the major drivers for digital transformation in the lab. “Many pharma companies are plagued by dark data—scientists within the organization or even within the same department may be unaware of similar or preexisting studies,” says EPAM’s Chief Scientist Chris Waller.
One cause of dark data is the inability to trace the development of initiatives through disjointed emails spread across several inboxes. Julie Gorenstein, Director of Data & Analytics at Takeda, elaborates on these challenges. “Generally speaking, everybody has exactly the same problem: There's a lot of data; how do you organize it to make it queryable and visualizable?”
Gorenstein says, “There's a lot of complexity on the process of alignment within the scientific community.” A simple example, she adds, can be found in our naming procedures here. “There are ontologies that exist to simplify complexity, but how do we trace the data from a raw value through an aggregated computed value? It might be one-to-one for a sample, or it could be one-to-two, or it could be to several samples, or it could be to several samples across many populations. So, how do you capture all this information? How do you trace through? That is a huge challenge.”
To address some of these traceability challenges, EPAM and the vaccine developer team implemented AWS cloud-based tools that leverage FAIR guiding principles for scientific data management.
The Result
One benefit of EPAM’s partnership was realized with early gains in the procurement application for this vaccine manufacturer, where the order throughput increased from 40 orders per month to over 400 orders per month, according to Kharlamov. This ten-fold increase in productivity requires simplified and optimized processes to facilitate scale-up.
Want to talk about clinical efficiency? Contact us.
