Table 1 Example of challenges encountered using predictive stability studies

• Mistrust of models–lack of understanding–or too complex to model

• Lack of broad champion

• Lack of resources or different use of resources

• Lack of knowledge on approach

• Risk aversion

• Data quality is questioned if using low levels

• Product is too stable–difficult to develop predictive degradant model when change is small relative to noise

• If modeling is not successful for a product it may cast doubt on other products

• Challenges modeling dissolution