Some people say digital manufacturing systems are too costly to implement and require too many resources to support going forward, according to Kevin Gordon, chief digital officer at Ori Biotech, who notes, however, that’s based on legacy information. New approaches are removing many of the barriers and helping companies recoup their investments faster. A 24-month return on investment is normal for companies replacing existing paper infrastructure with digital technologies. When starting fresh, the return on investment is even faster.
Traditional manufacturing systems are tightly linked. Here, data is pulled from databases by queries and put into other systems. “When one of those systems experiences a version update or major change control, all the coupled systems must be revalidated,” Gordon says.
The key difference is that “modern systems use a loosely-coupled publish and subscribe model using standard application programming interfaces. For example, the lab information management system (LIMS) can push data to the electronic batch record system (eBR), but only if the eBR subscribes to it. Consequently, elements (like the eBR) can be updated without revalidating all the other coupled components. As long as the information-handling method hasn’t changed, the rest can keep running,” Gordon continues.
Most small-molecule and antibody-drug manufacturers have already transitioned to digital manufacturing systems. Cell and gene therapies and other advanced therapeutics generally have not.
“Developers often believe the entirety of their processes are unique and would be difficult to digitize but, in reality, process commonality is about 80%,” points out Gordon, who advises exception management–“configuring to what’s different, but leveraging the fact that a large percentage of the process is universal. Use those commonalities as accelerators going forward.”
This approach de-risks the chemistry, manufacturing and controls (CMC) process during manufacturing scaleup from clinical to commercial quantities by making the technical transfer less susceptible to failure, and accelerates manufacturing going forward. Later, “the immutability of the data and the ability to generate reports directly from the system of record provides an audit-friendly platform that eliminates the risk of human error,” he explains.
Getting started requires data governance exercises. Gordon says they include “standardizing naming conventions and data to allow comparability and cross functionality that may not have existed before. Standardized data makes insights easier to gather, analyze and visualize, so learnings become clearer and more obvious. It comes down to creating the same version of truth throughout the organization.”