Data Analytics
Aug 06, 2019
| 5 min

How Digital Transformation Helped Amgen Implement Real-Time Process Control

In many manufacturing industries, variability in raw materials can lead to unexpected and undesirable changes in the final products. In regulated industries such as pharmaceuticals, this is especially problematic due to the need to maintain carefully controlled processes that stay within approved regulatory parameters for drug development and production. Embracing a total company-wide digital transformation enabled Amgen to align data across multiple systems to not only control, but also predict unacceptable deviations in time to make necessary adjustments. Read on to find out how they used data analytics to implement real-time process control.

This article is posted on our Science Snippets Blog.

Digital transformation is essential for successful real-time monitoring, which allows companies to incorporate data from raw materials to improve processes in real time.

Amgen Embraces Total Digital Transformation

For manufacturers, being able to address any problems arising from variability in raw materials usually means acting quickly to find the root cause. Otherwise, serious and significant setbacks to development and manufacturing can occur. In practice, this requires careful and ongoing monitoring and analysis of data coming from multiple sources: from raw material suppliers to production to patient responses. Process control requires coordination in how data is collected, shared, and analyzed starting with the raw material.

To accomplish this, Amgen created a Supplier Relationship Excellence (SRE) program that opened lines of communication with suppliers and created a feedback loop where operational performance data is shared. Key elements of the electronic data program included establishing digital exchange standards, using predictive models, and integrating artificial intelligence tools and technologies.

According to Dell Technologies Digital Transformation Index II Report, fewer than 28% of companies have adopted digital technologies that transform their business. Amgen is on the forefront for biopharma manufacturing companies implementing digital transformation for real-time process control.

Amgen’s successful digital transformation program for real-time process control included:

  • Standardization of eData for raw materials using a customized information system
  • The ability to identify, track, and control variation (ITCV) using multivariate analysis
  • Using mobile handheld technologies to capture data and enable real-time responses
  • Computational modeling for predictive analysis that mimics the process, equipment, and raw materials and identifies anything that might contribute to performance variability
  • Incorporation of artificial intelligence tools, such as Watson Natural Language Processing (NLP) for faster processing and search of large quantities of data

During a recent Umetrics User meeting, Cenk Undey, Executive Director of Process Development at Amgen presented a 10-year progressive overview of Amgen’s migration toward digital transformation and their experience using Real Time Multivariate Statistical Process Monitoring for biopharmaceutical manufacturing using SIMCA®-online.

Throughout the last decade, the company has continued to develop their use of data analytics to increase efficiencies and agility of processes and product development across all of their manufacturing plants. Real-time predictive monitoring and process control are key benefits of their digital transformation.

Read more about: The trending role of artificial intelligence in the pharmaceutical industry

Undey said the goal of digital transformation is to turn data into actionable insights. Advanced data analytics helped the company create processes to actually predicting problems before they occur.

Undey explained what success looks like at Amgen:

  • Saved batches. Contamination avoidance in bioreactors. Early leak detections and bioreactor vent heater failure early detection.
  • Yield increases. Titer increase due to induction timing, seeding and batch production optimization.
  • Efficiency gains. Faster non-conforming resolution due to ease of access to data and ability to compare historical batches. Downtime avoidance, quick trouble-shooting, preventive/condition-based maintenance.

Standardization of Data

For Amgen, one of the critical factors for successful implementation of real-time multivariate data analytics for a broad range of complex data was being able to consolidate and analyze an overwhelming amount of disparate data from different sources. They estimated that more that 500 million continuous data points are generated per product from manufacturing equipment used in therapeutic protein production. In research and development, one of the robotics-based drug candidate screenings generates an additional 200,000 data points per day.

In the past, some of this was collected in separate, unconnected Excel spreadsheets. Creating a standard file format and allowing seamless data exchange between suppliers and users, was a huge step forward for data integration.

“A significant amount of data is generated during the manufacturer of our products: More than 500 QC entries, over 2000 batch record entries, at least 500 million continuous data points.” – Cenk Undey, Amgen

Process Control Advances

Another major step forward was creating a predictive multivariate model that uses data readily available from certificates of analysis and prior to drug substance manufacturing to provide accurate expected impurity results from the pertinent step in less than an hour. Previously this step would take more than 80 hours, and required running a bench-scale model in the process development lab that negatively impacted production times. Using a data analytics process gives the team more time to work on important process improvement efforts.

Incorporating Mobile and AI Technologies

Amgen also incorporated data from other digital transformation technologies such as handheld Raman and NIR spectroscopy technology, which is used for raw material inspection. Amgen used this data to conduct mathematical modeling and statistical trending about raw material variability, creating a tool for predictive analysis, or what Amgen calls “preactive” (predictive in a proactive way) analysis.

The Amgen team also adopted advanced artificial intelligence (AI) tools, such as Watson Natural Language Processing (NLP), to achieve better insights. The system not only provides better understanding of raw material variability, it also creates feedback loops from patient experiences. For example, any information from patients that is made available in the company databases is connected to the Watson system, so the inspection team can perform text analytics using freeform sentences on the fly.

"Pre-active: To be predictive in a proactive way” - Amgen Process Development Team

What Does It Take for Digital Transformation to Work?

According to Undey, the digital transformation building blocks that lead to successful real-time predictive process monitoring, include:

  • The right hardware and software
  • Validation and supporting documents
  • On-demand models for relevant products & processes
  • Optimal alarm strategy for showing process deviations
  • A user experience that provides live updates and feedback
  • Clearly defined roles and responsibilities among the team and a commitment to follow processes, models and systems
  • Business processes for monitoring, responding to signals, and communicating findings
  • Users who are trained and empowered; and where desired behavior is rewarded

Find Out More

Find out more about how Amgen is using data analytics and digital transformation technologies to improve processes and predict problems before they occur.

Read the article "Amgen’s Digital Transformation: Linking raw material data from suppliers to patients”

Get the User Meeting Presentation

Get Amgen Presentation