What is my data telling me? Distinguish between process problem or measurement error using JMP to visualise the results
Christel Kronig, Senior Analytical Scientist, Dr Reddy's Laboratories
Analytical methods are crucial in supporting pharmaceutical development and manufacture. But should we always believe the data generated by these methods? How close is the testing result from the actual true value? Understanding of the measurement error is essential in order to make sound data driven decisions: is my product stable? Is there an issue with my process? Does my product meet specification?
To evaluate measurement error, data needs to be collected, and appropriate calculations such as bias, precision, uncertainty etc… can then be performed. Visualising the results graphically can be extremely powerful. Several examples of measurement error evaluation and visualisation will be presented using several JMP platforms such as Gauge R&R, ANOVA and control charts.
The accuracy profile validation methodology is a powerful way of understanding the total error for a given analytical method. An application of this methodology to compare the performance of 4 different analytical techniques for 3 separate attributes and 4 different products will be presented in one powerful graph generated using Graph Builder.
