Umetrics Suite Blog

Using multivariate models in a regulated environment for continuous manufacturing of pharmaceuticals

January 14, 2020

Continuous manufacturing is one of the key trends within the pharmaceutical industry, both for the production of ‘classical’ drugs as well as large molecules. Companies are looking for ways to shift from traditional batch processing to a continuous method of operation. The main advantages associated with these processes are more room for modularity, automation and flexibility due to a smaller footprint, as well as more consistent quality of the drug product.

continuous-manufacturing-pharma

The FDA encourages manufacturers to adopt continuous manufacturing using advanced multivariate data models, because it improves process control, reduces variability and improves product quality and consistency.

The regulatory bodies such as the FDA see the promise of these processes and actively encourage companies to implement this manufacturing strategy. This is illustrated by the recently (2019) published draft guidance from the FDA¹ for the production of small molecule drugs using continuous processes, and their active discussion with the (bio)pharmaceutical industry through their Emerging Technology Program.

The FDA considers multivariate models to be surrogates for traditional release tests, for example, a stand in for dissolution. Christine Moore, as acting director of the FDA, described the use of real-time release testing (RTRT)² as:

“ the ability to evaluate and ensure the quality of in-process and/or final product based on process data” that “typically include a valid combination of measured material attributes and process controls.”

Moore asserted that real time release testing provides for an increased assurance of quality as well as increased manufacturing flexibility and efficiency. This means:

  • Shorter cycle time
  • Reduced inventory
  • Reduction in end product testing
  • Reduction in manufacturing cost

It also allows manufacturers to leverage enhanced process understanding to make corrective actions in real time.

The trend toward continuous manufacturing in pharmaceuticals

BCC Research's report on Continuous Manufacturing for Pharmaceuticals³ valued the market size for continuous manufacturing at $2.3 billion in 2018, and it is expected to grow at a CAGR of 8.8% through 2024 to reach $3.8 billion in 2024.

From a business perspective, continuous manufacturing promises to reduce costs and help products get to market faster. Better control means improved product quality and easier standardization, making it easier, for example, to manufacture biosimilars. It also becomes the foundation for a robust quality by design (QbD) strategy.

Pharma 4.0 analytics

This transition to more data-driven approaches and using continuous process manufacturing are sometimes referred to as pharma 4.0. The core of this shift lies in being able to use data analytics to create forecasting models for processes and process parameters. This means you will not only be able to understand what is happening in your processes as they are running, but also be able to predict the outcomes. Other data-driven benefits of pharma 4.0 include:

  • Model predictive control (MPC) – paving the way for more autonomous processes
  • Holistic overviews – giving you a view of your processes remotely, from anywhere
  • Predicting “soft” aspects – being able to predict quality in real-time based on online parameter estimations
  • Hybrid models – combining the mechanistic understanding of cell metabolism with data to create realistic models (or digital twins)
  • Increased understanding – data modeling will also improve the understanding of underlying kinetics of advanced biotransformations, such as protein separation and purification

In the end, we expect all of this together will be combined into a closed looped model of data based on predictive control strategies that get you closer to a fully automated process.

 

Using advanced data models in a regulated environment

It’s clear now we live in a world of data and the possibilities of data are endless. But what does it mean to use this in a regulated environment?

First of all, you need to make sure your data and your processes for collecting data (from research and development through prediction) are compliant. It’s essential, especially in highly regulated industries such as life science or pharma, that you follow all of the required steps to ensure the integrity of your data at every stage. This is more successful when you use tools and products designed to keep your data compliant.

  • Make sure your data is GAMP 5 compliant

GAMP 5 is a set of guidelines for manufacturers and users of automated systems in the pharmaceutical industry. These guidelines describe a set of principles and procedures to help ensure that your products have the required quality. They also cover all aspects of prediction and data management from the raw materials to facilities and equipment to the training and hygiene of your staff.

blog 30 data GAMP 5 compliant

  • Follow the ALCOA principle of the FDA

In addition to quality controls of GAMP 5, data integrity needs to be ensured as well, and this follows the ALCOA principle of the FDA. ALCOA is an acronym that stands for attributable, legible, contemporaneous, original, and accurate.

  • Attributable. First of all, your data must be stored so that it can be connected to the individual that produced it. Every piece of data entered into the record must be fully traceable with time and origination stamps.
  • Legible. Your data needs to become permanent, readable and understandable to anyone using the record. And this also applies to any meta data attached to the record.
  • Contemporaneous. Data must be fully documented at the time it’s generated.
  • Original. You need to have a copy of the original record or a certified copy, and the data record should include the first data entered as well as all successive entries following that.
  • Accurate. And finally, your data needs to be correct, complete, truthful and reliable.

Following FDA guidance

The FDA encourages manufacturers to adopt continuous manufacturing practices for two main reasons:

  1. It reduces manufacturing costs (which may reduce drug costs for consumers)
  2. It improves process control (which may improve product quality and consistency)

This helps in reducing variability and the risk of adverse reactions. From the consumer’s perspective, it may increase the speed at which drugs become available.

In their draft guidance, the FDA encourages the use of multivariate data analysis when creating prediction models. In Section III.B.2. on Process Monitoring and Control, the FDA encourages the use of models to ensure data provides value.

Continuous manufacturing relies on multivariatemodels that can predict process changes using real-time data input. These models summarize and transform data from univariate sensors into a multivariate trend graphs, making it easier to analyze.

Some of the ways multivariate models can be used to help improve process control for continuous manufacturing include:

  • early detection of equipment malfunction
  • insight into raw material impact
  • in-situ assessment of process stability
  • predictive monitoring

Read more: Multivariate vs univariate analysis in the pharma industry: analyzing complex data

Model Validation

An essential step in continuous manufacturing in regulated environments (or, in any manufacturing environment) is validating the accuracy of your model.

The model validation strategy ensures that you are continuously updating your production review methods. A successful prediction model must also consider how to handle batch data sets, and ensure they are used to confirm the reliability and robustness of the model. You also need to consider how to handle discrepancies between quantitative and qualitative models.

blog 30 considerations for model validation

When validating a model for use in a regulated industry, you’ll have an extra layer of questions that need to be addressed. Some key considerations of model quality will be:

  • Is the model complexity correct?
  • Is the model capable of differentiating between good and bad?
  • Can it avoid having false negative or positive?
  • Is the model capturing a sufficient amount of variation?
  • Is the prediction model robust enough? Precise enough?

And, in a regulated environment, you also have the administrative part to consider. This covers the user access. For example, what kinds of rights do the end users have? Does everyone have the appropriate level of control for handling the process? What does the system look like? Are any special SOPs required?

At the moment, there are no real guidelines from regulatory bodies for model validations, but there are best practices and examples from different parts of the industry, which can give a good insight into the line of thinking within the regulatory bodies.

Read more: How one company integrated data to implement real-time bioprocess monitoring

Want to know more?

Find out more about creating predictive models and validating them for continuous manufacturing.

Watch this recorded webinar: 

“The use of multivariate models in a regulated environment within continuous manufacturing for pharmaceutical products.”

Watch webinar

 

Sources:

1. Quality Considerations for Continuous Manufacturing , Guidance for Industry, DRAFT GUIDANCE , U.S. Department of Health and Human Services, Food and Drug Administration, Center for Drug Evaluation and Research (CDER), February 2019

2. Christine M. V. Moore , Regulatory Perspective on Real Time Release Testing (RTRT), AAPS Annual Meeting Washington, DC, 27 October 2011

3. BCC Research, Continuous Manufacturing in Pharmaceuticals: Implications for the Generics Market, global report, July 2019

 

 

Topics: Real Time Process Monitoring, Quality by Design (QbD), Pharmaceutical manufacturing, Continuous Process Monitoring

Lennart Eriksson

Written by Lennart Eriksson

Sr Lecturer and Principal Data Scientist at Sartorius Stedim Data Analytics

Leave a comment