For pharmaceutical manufacturers, a process deviation may not only mean a bad batch that affects a downstream process, it can also risk a regulatory violation that leads to fines or expensive market setback, or worse, it could endanger the health of the patient.
Even well-documented processes can go wrong. Statistical process control can help pharmaceutical manufacturers stay in compliance.
Your pharmaceutical production must follow a strict manufacturing process that adheres to the design space filed when the drug or product was approved. How do you ensure your product is manufactured in a controlled way with all parameters and correlation patterns within the approved ranges? If the manufacturing process deviates from this design space, the product may not fit the criteria to be considered safe for use by patients.
Continuous process manufacturing introduces control
Prior to 2011, you needed to show data that fulfilled specifications from the first three batches of your manufacturing process and the end product to have a validated process. But that changed with the 2011 guidance from the FDA that focused on continuous process verification, launching the modern era of quality by design and continuous process manufacturing.
The FDA realized that drugs with lower quality were being marketed that followed “validated” processes, but did not have a statistical basis for continuous process manufacturing. This has led to product recalls and complaints because of missing process understanding and process control. Since then, showing process knowledge and statistical validation has become an important aspect of regulatory approvals.
The FDA took this even further in 2019 by issuing a guidance for continuous process manufacturing.
Causes of process deviations
But even when processes have been validated, a lot of things can go wrong. Let’s take a look at some of the most common causes of pharmaceutical process deviations.
One of the most common causes of process deviation (especially in batch production) is contamination during a previous step or an ingredient impurity. The technicians might notice that cells are dying but not be sure what is killing them. Often this points to an impurity of some kind, such as contamination with bacteria, mycoplasma or viruses. But having the right data analysis processes can help you better identify the source of the contamination. Historical data can be analyzed using MVDA to determine what combination of impurities could potentially cause problems and create a standard of reference to help identify of these types of deviations early on.
2. Equipment failure
If a piece of equipment comes loose or a part is broken or malfunctions, it can cause problems in your process. Or, there could be a problem with one of the sensors. Maybe the database is compressing the data in a way that causes unusual readings (or worse, failing to make unusual readings obvious). That doesn’t necessarily mean the production is wrong, the data could be wrong, but even then you need to fix it. Because if the data is wrong, you cannot sell the product, since it's not compliant. The key is to be able to fix the problem before it affects the batch by applying predictive maintenance. For that, you must know that critical quality attributes were not affected and that you have caught the deviation in time. (Read more about batch process control).
3. Human error
If a process is not followed correctly, a stainless steel reactor wasn’t completely cleaned between batches, or chemicals were not properly flushed away, it could affect the next production run. Humans can, and do make mistakes, but don’t be too quick to blame all production problems on human error. Especially in today’s heavily automated world, human error is hopefully less and less common in pharma production. If you are seeing a large number of deviations from human error, it’s important to look at whether your process was designed properly, and if your operators have the right training.
“Regulators see human error as a last resort. Their expectation is that you can - and have - eliminated any possible process issues and confirmed that the individual had everything they needed and simply wasn’t focused.
Industry sees human error as a first-line response. We almost assume our processes, procedures and training are bulletproof, and the issue must have resulted from someone not paying appropriate attention to what they were doing at the time.”
- Joanna Gallant, in Pharmaceutical Online
4. Raw material problems
If there is a problem with the raw material that goes undiscovered, perhaps because it’s not properly measured or tested, you had a change in supplier or media, or it was contaminated during transport, you could start seeing unexpected deviations. These are often easy to spot because they create specific variations, but the sooner you discover them the better. They could impact the quality of production otherwise. (Read more about Reducing batch to batch variability)
5. Process interruption
If your process is interrupted for some reason, whether from equipment failure, power outage or another reason, your processes compromised. It’s important that you have the metrics and data available to be able to validate the critical process parameters to know if a short power interruption or other issue has caused an irreparable deviation to your process.
6. Process unknowns
There could be many unknowns in your production process or factors that impact each other in combination. Multiple unknown variables could be affecting each other and leading to deviations in your processes. In order to understand these better, you need to consider multivariate data analysis.
7. Limits of your design space
While it might not be practical (or cost-effective) to explore every possible variable when creating your design space (or the outer limits for your process control), having a very limited design space could lead to deviations that might otherwise be possible to correct falling outside the acceptable range. You also have to consider the accuracy limits for sensors, and allow a design space that is wide enough to contain such deviations.
Keeping your process in control
Whether your process deviation is caused by human error, contamination, equipment or sensor malfunction, or data that is out of range, knowing what to do to correct the problem, and whether your batch can be saved or process corrected in time to save the final product, depends on how quickly you detect it and whether you have the statistical data to back up your decisions. Having the right statistical tools for creating the optimal design space, establishing effective control parameters and continuously validating your process is important.
Our next article in this series will discuss “Finding effective ways to correct pharma process deviations.” Subscribe so you don’t miss it.
1. FDA/CDER/CBER/CVM, Guidance for Industry: Process Validation — General Principles and Practices, US Food and Drug Administration, Rockville, MD, Jan. 2011.
2. FDA/CDER/CBER/CVM, Quality Considerations for Continuous Manufacturing: Guidance for Industry, US Food and Drug Administration, Rockville, MD, Feb. 2019
3. Gallant, Joanna, “Human Error Is The Leading Cause Of GMP Deviations – Or Is It?," Pharmaceutical Online, May 1, 2014.