Survey Results: The Use of DoE in the Pharmaceutical Industry - Part II

With the ICH guidelines ICH Q8, Q11, Q14 and the revision of ICH Q2, the subject of developing processes and analysis methods has gained greater importance than in the past. One of the methods referred to as helpful here is Design of Experiments (DoE). A chapter of the European Pharmacopoeia also deals with this topic. Is DoE already being used in the pharmaceutical industry and in the manufacture of active ingredients? The ECA wanted to evaluate this through a survey. Read more results in Part II.

In Part I of this survey, we described the areas in which DoE is already being used and how often. In Part II, you will learn what purposes DoE will be used for in the future. What experiences have been made with DoE? And what problems still exist? Not every participant answered every single question; sometimes several answers were given by the same person, and rounding also contributes to the fact that the total sum of the answers does not always add up to 100%. Below are the results to the questions metioned above.

What DoE will be used for in the future?

Process understanding/characterization were mentioned by a large number of respondents (71%). Process/product/business optimization came in second (53%), followed closely by robustness testing (46%). In fourth place came use for method validation (42%). Somewhat behind was use in the area of submissions (12%).

Fig. 1: Purposes of DoE in future

The answers to the question "What is your experience with DoE?" are statistically less meaningful, as only a few respondents chose one of the options provided. The following were mentioned: We have selected too many factors for DoE (17%), 15% said they had no experience, and 13% admitted to having only used "one-factor-at-a-time" experiments so far. There were many individual responses (50%) regarding the use of DoE. Most participants viewed DoE positively, with some having used DoE for years or even decades. However, there were also some rather negative comments, such as "We have used small DoEs with little success" or "Challenging due to the large number of repetitions." Of particular interest was the statement by a DoE expert who views DoE as an effective tool when sufficient scientific background is available. However, he considered its use for "GMP brains" to be difficult due to the high procedural effort required in the GMP environment.

ECA also wanted to know whether there were any specific problems with DoE. 68% of participants answered "No." 32% answered "Yes," citing reasons such as:

• No experience with using DoE/no support (3x)
• Management does not support it
• Handling a large number of experiments
• DoE was discontinued because the necessary production line was not available
• There is no support for handling DoE data and documents
• DoE software is too expensive at the moment
• Resistance to using DoE in a GMP environment
• Some factors are difficult to use
• What type of DoE is necessary for approval documentation (format)?
• DoE is time-consuming

With regard to a DoE event, the ECA wanted to know what participants expected from a face-to-face event on this topic.

Twenty-five responses each mentioned a better understanding of how data is modeled and seeing case studies. Eighteen participants wanted hands-on training with statistical software. There was also one interesting individual response: "How DoE can be explained to people with no DoE background in an audit." The Minitab software is clearly the preferred choice for such an on-site DoE event. 

Conclusion 

DoE is already widely used in the pharmaceutical industry, and yet not as frequently as it could be. In some cases, the necessary experience and support are not available. There is skepticism regarding its use in the GMP environment with regard to data and documentation. 

To provide assistance here, the ECA is offering the on-site event "Introduction to Statistical Design of Experiments in the Pharmaceutical Industry" on 23-24 September 2025 in Berlin. Many exercises will be carried out directly by participants. Minitab will be used as the statistical software.