What do you like about this case study?
- Analytical method-specific case study on applying Quality By Design
- Simple example that shows the value of DoE and the strengths of the Prediction Profiler and Design Space Profiler for determining the 'Method Operable Design Region' required by regulators.
- Covers most of the expected analytics/outputs required in an analytical method development process.
Who is it for?
- Analytical/QC/QA Scientists - Anyone involved with using chromatography or analytical measurement equipment, look out for terms such as 'HPLC, UPLC, GC, GC-MS, LC-MS'. They will value the ability to create experiments that are structured and allows them to gain more insight into their processes (and reach a 'ready to use' method quicker).
- Analytical Science Group Managers - Managers will be specifically involved in submitting their methods for approval to regulators according to their guidelines, they will value the ability of JMP to accelerate the QbD process and let them reach an approved method quicker.
How should I use it?
- The journal has lots of details covering the background of why each step is performed in QbD - the best focus is on the robustness profiler, the design space profiler and the graph builder (for intermediate precision).
What platforms can we use?
- Custom Design
- Prediction Profiler
- Design Space Profiler
- Graph Builder
- Fit Y by X
- Contour Profiler
Background
Analytical methods such as liquid or gas chromatography are key in industrial chemical/biochemical manufacturing and is used to assess raw material quality, productivity, consistency and identify potential drifts in the production process. This means that there is a lot of scrutiny on ensuring the methods are capable of consistently producing accurate and sensitive results. 99% of groups developing analytical methods for manufacturing are following the ICH guidelines - they state that a method must have proven specificity, accuracy, precision, linearity, and robustness, identifying the critical method parameters that affect the analytical tool, all of which falls under the 'Analytical Quality by Design' umbrella.
In the real world, analytical equipment like HPLC's will always vary in their ability to accurately control conditions; the temperature control of a column might vary significantly, there may be differences in the way different analysts prepare buffers etc. - the value of the AQBD process is that it allows them to understand how likely they are to have a failing method if these settings change and to focus on the settings that are the most critical to obtaining a suitable separation/quantification. The key suggestion from the ICH is that scientists should use DoE to facilitate this method development.
Most analytical scientists will not be confident in what DoE is and why to use it - analytical method development is very 'traditional' and always follows the same approach (change the column, then the temperature, then the flow rate etc. etc.) - JMP provides the benefits of generating and analyzing DoE's in a visual way that isn't stats-heavy and too 'scary' for a Scientist to want to attempt DoE with. Add on top the complexity of finding a design space that meets specification limits for multiple inputs and the use of the Design Space Profiler is a great alternative to classic Contour Plots and 'by eye' analysis.
The case study is based on the paper attached - it's an older paper but I like the simplicity of the challenge and I've altered some of the details to tell a more compelling story/streamlined demo. The key goals were:
- Find the HPLC settings that would meet the specification limits of 'Critical to Quality Attributes' - Custom Design.
- Find the setting ranges that would ensure robustness in the CTQA's - Contour Profiler or Design Space Profiler (better).
- Determine if the method was affected by changes in the equipment, day and analyst (intermediate precision) - Custom Design, Fit Y by X and/or Graph Builder.
Questions to ask:
- Do you develop analytical methods (HPLC, UPLC, LCMS, GCMS) for scale up/manufacturing/production?
- How do you determine what a 'good' analytical method is?
- Are you trying to find your 'Design Space' or 'Method Operable Design Region'? How do you do that?
- Do you follow the ICH guidelines?
- How do you determine the robustness of your method?
- What does a regulator expect to see in your methods?
- Do you have to prove Intermediate Precision for your quality/analytical/testing labs?
