Many people were surprised when in early November 2020, just 8 months after the world was largely bought to a standstill by the COVID19 pandemic, leading pharmaceutical companies Pfizer and BioNTech announced that they had developed a vaccine that they claimed was ‘90% effective’.
The surprise wasn’t borne of the fact that we had a vaccine at all – after all, vaccines have been responsible for the eradication of smallpox, among other diseases – but more the speed at which it had been developed.
This was a testament to not only human skill and determination, but also the technology, data and analytics that were used to develop it. In January 2020, it was announcement that the first drug designed entirely using AI had entered human clinical trials after less than a year of development, a huge milestone in the mission to unite the world of data analytics, technology and pharmaceuticals. But how has data analytics helped in the development of the COVID19 vaccines?
Design of Experiments (or DOE) is a tool that allows for a systematic approach to process development studies – ultimately reducing the number of experiments needed, and in the long run, also reducing the overall cost of experimentation.
A Quality by Design (QbD) approach to Design of Experiments (DOE) allows vaccine researchers to systematically determine the individual and interactive effects of various factors that influence the results of experiments.
For vaccine development, DOE can be broken down into three different investigational objectives; accelerating screening, supporting optimization, and ensuring robust characterization.
Accelerating screening means that vaccine developers can investigate many process parameters at the same time, enabling faster time to market, reduced costs for experimentation, and overall maximized knowledge. Process optimization means developers can use DOE to help determine factors, ranges and inputs needed to achieve specific process goals, like quality. DOE can also be used to analyse each unit operation’s design space, and then to calculate the extent of all the ICH Q8 guidelines. Bioprocess characterization ensures product stability, robustness and scalability, as well as staying in compliance with regulations.
With a vaccine developed and the need for hundreds of millions of doses to be produced in a short time frame, it's important to have an efficient and organized way to manage scale-up and technology transfer.
The expectation to scale-up to manufacturing in six to 12 months to address COVID-19 is an unprecedented speed. However, by using data analytics tools like MVDA (multivariate data analysis), the number of total batches needed to prove robustness can be less.
Being able to use data to improve the manufacturing process is also key. Using real-time analytics to monitor and control manufacturing has been a proven tool to ensure the robustness of the process as well as the product quality and consistency.
With multiple vaccines now approved for use, rolling them out to billions of people around the world as quickly as possible is a logistical challenge of unprecedented scale. This is perhaps the most traditional way that data analytics can help. The added issue for COVID19 vaccines is most of the ones available require two doses, with a different time between each dose.
Furthermore, the second dose needs to come from the same manufacturer as the first. With the two leading vaccines, one requires an ultra-cold storage (-100°F), has a shorter shelf-life, and needs more carefully transported than the other.
This will once again be the time to shine for data analytics, providing medical professionals and supply chain organisations vital modelling information and data repositories from which they can plan the rollout.