Consortium to Develop Knowledge Model for Biopharmaceutical Process Development

A new biotech consortium in the UK believes it can develop tools to improve biopharmaceutical process development.

A new biotech consortium in the UK believes it can develop tools to improve biopharmaceutical process development. Led by the consulting firm BioPharm Services Ltd (Chesham, UK, www.biopharmservices.com) and supported by the UK-based biopharmaceutical companies Avecia (www.avecia.com) and Cambridge Antibody Technology (www.cambridgeantibody.com, now part of AstraZeneca), the consortium was just awarded a $0.8 million Collaborative Research & Development Grant by the UK’s Department of Trade and Industry for a three-year project, expected to cost $2 million in total, to develop a knowledge-management model.

The idea, says Andrew Sinclair, managing director of BioPharm Services, grew out of manufacturing process simulations his firm was conducting to help biotech clients improve their manufacturing operations. Sinclair and his clients realized that they could increase the benefit of these simulations by applying them to process development, and by finding ways to profit from the data itself, even before using any simulation tools.

“Process information comes from development in isolated islands, and then the information has to be dragged together,” says Sinclair. “During the technology transfer process, staff end up replicating data or generating it again. We realized that if we could capture that data in a more structured way, we could make much better use of it.”

Bruce Williams, commercial engineering manager at Avecia Biotechnology, says his company has already seen the benefits of pulling together data electronically in its work with Sinclair’s group. “We got as much benefit from pulling the data together as we did from the simulation,” he says.

As an example, Williams cites an equipment cleaning study, in which they examined cleaning agent types, the consumption of cleaning agents and water, the time required, and choices between cleaning and sanitization processes. As a result of that work, Williams’ team was able to cut cleaning times and cleaning fluid usage in half.

The key difference between this new project and past work, Sinclair says, is that now they are working on an earlier stage in development. “We’re shifting the focus from manufacturing to process development,” he says. “We want to capture the data early, as it is being developed, so that we can look at manufacturability much sooner.”

Williams agrees that moving the focus earlier is useful. “It can be hard for someone in process development to understand the complexity and constraints of manufacturing operations,” he says. “So if we have a tool that allows someone in development to see the impact of process changes on manufacturing, and to see it in terms of cost of goods, that will be valuable.” He says the tool will also make it easier to re-use knowledge from other plants and processes. “It’s important to structure the way you go about design, so that you start from a fixed knowledge base rather than starting from scratch each time,” he says.

Sinclair says the consortium’s tools will differ from existing products because they will be designed specifically for process development and will focus on integrating data from a variety of sources. Williams agrees. “With existing tools, it’s hard to integrate data, so many companies are writing their own bits of programming code,” he says. “It can be very labor intensive to transfer data between different spreadsheets and databases.” The consortium plans to make the database compliant with ISA-88 and ISA-95 standards, so that it can integrate with existing process control and manufacturing execution systems.

Sinclair says the consortium hopes to have a beta product ready for testing in three years. He envisions that the product will include tools for simulation, costs of goods evaluations, and for creating process descriptions, but his primary focus is the database. “We can expand the tools later,” he says. “But it’s important, first, to develop a database that is expandable and validatable.”