Dear All,
I am looking for advice on the appropriate statistical approach in JMP to evaluate a formulation screening/stability study.
I have developed several formulations of an API solution containing different combinations of stabilizing components (A, B, C, D, E). The formulations have been filled into two different container types, so I would like to evaluate both the effect of the stabilizing components and the potential interaction with the container.
I currently have 2 months of stability data (t0, t1 m, t 2 m).
The main objective is to determine, statistically, which stabilizing component or combination of components provides the best API stability, and whether the performance depends on the container.
I would appreciate advice on how to structure and analyze the data in JMP, particularly regarding:
- How to define the stabilizing components as factors (including the presence/absence of each component and combinations).
- How to include container type as an additional factor.
- Whether the appropriate approach would be a factorial/DOE model, linear mixed model, ANOVA/GLM, or another approach.
- How to assess main effects and interactions, especially stabilizer × stabilizer and stabilizer × container interactions.
- How best to incorporate the 2-month stability measurements (e.g., initial vs. 1- and 2-month data) into the model.
- Which JMP platform/procedure would be most appropriate.
- How to determine statistically whether one formulation is superior to the others rather than simply comparing individual means.
- How to use the model to identify the best stabilizer or stabilizer combination, considering the stability responses simultaneously if several analytical attributes are being measured.
The stability responses include assay/API content and degradation products.
My primary goal is not simply to demonstrate statistical differences, but to use the available data to identify the most promising stabilizer combination for further formulation development.
