Use Advanced Analytics To Usher In The Continuous Manufacturing Era

By Lisa J. Graham, Ph.D., P.E., Vice President of Analytics Engineering, Seeq

Regulatory support and innovations in technology have fueled the adoption of continuous manufacturing in pharma over the last several years, which exhibits a move toward operations excellence marked by rapid production within multi-use facilities, reduced scale-up risk, and greater quality expectations. The end-to-end, integrated, and intensified processing approach to drug production offers a wide range of benefits that help drive the industry’s goals to increase efficiency in drug development and manufacturing while lowering costs. With continuous manufacturing, you only have to run the process longer to generate the volumes required, significantly reducing the delays and risks associated with scale-up in batch manufacturing. Maintaining a steady state of production with continuous manufacturing also eliminates batch variability and improves product quality and availability. By facilitating scale-up while simultaneously ensuring high quality, continuous manufacturing opens the door to accelerating the development to commercial launch process.

However, the complexity of how quickly data must be analyzed is greater with continuous manufacturing. The start-and-stop steps of batch manufacturing allow more time for analysis and decision-making as opposed to continuous processing, where material rapidly moves from one unit operation to the next. As a result, analytics must be performed faster to quickly identify any issues and take action before valuable product is lost. In addition, continuous manufacturing consists of interrelated unit operations, where the control of a single unit operation affects what is happening in the next, requiring a greater level of understanding about how the various steps interact. Therefore, implementing continuous manufacturing requires a robust data collection and integration strategy across your entire organization.

Empower Data Capture And Sharing

While some companies still use the process of capturing data manually, others use SQL databases or have even moved toward other technologies, such as data historians. The result is critical information stored in disparate locations, preventing subject matter experts from easily connecting to the data and making decisions that are based on a holistic analysis of a process rather than only a portion of it. Working in siloes has long been an issue in the pharmaceutical industry that must be overcome. Breaking down walls and moving toward a future where collaboration acts as the engine for improved patient outcomes, though, requires partnership between situations, systems, and people.

For example, there are often situations where data scientists are asked to analyze data and provide a model for process engineers to implement. This essentially removes the engineer who has the context and understanding of the process that is a crucial component in ensuring an accurate model. It becomes even more problematic if that model is stored in an Excel spreadsheet to which only one person has access. Preventing this scenario and others like it requires an application that increases transparency across the process lifecycle, reinforcing decision support and not just execution. Journaling capabilities would also allow scientists to capture their thought process, leaving breadcrumbs of information for process engineers to click through to see how conclusions were drawn from that data or model and how the ideas and guidelines around it were established. A feature such as this would enable responsive decision-making as well as provides predictive analytics based on current-state data, so corrective action can be taken in time to either change, avoid, or plan for the expected negative outcomes.

As recently noted by McKinsey and Company, "Advanced data-analysis techniques are helping companies better understand and control the intricacies of their production processes. The result is better consistency, higher productivity, and superior quality. One major biopharmaceutical company used such techniques to tackle highly variable yields in vaccine production, leading to a major expansion in production capacity with no additional capital outlay."¹

Become An Analytics-Driven Organization

As humans, we are naturally skeptical of what we do not understand and sometimes have difficulty accepting change. This too applies to the pharmaceutical industry where the wrong decision can cost millions of dollars and, more importantly, negatively impact patients’ lives. That is why creating a culture that fosters the implementation of new approaches and technology may be the most challenging aspect of their adoption, regardless of the benefits they offer. Begin by acknowledging the pain points of your organization with those they affect and discuss potential solutions. Do you have the resources to successfully execute continuous manufacturing while capturing and integrating the data, so it can be effectively used for visualization, analysis, and process modeling? Have you considered how to influence your company culture into not just making, but also celebrating, these significant changes? With a thoughtful approach, you can help your team come to the conclusion on their own that using an application dedicated to process data analytics could be what they need to tap in-house expertise and improve their business and production results.

1. McKinsey and Company. (June 2017). The great re-make: Manufacturing for modern times. https://www.mckinsey.com/~/media/McKinsey/Business%20Functions/Operations/Our%20Insights/The%20great%20remake%20Manufacturing%20for%20modern%20times/The-great-remake-Manufacturing-for-modern-times-full-compenium-October-2017-final.ashx