Equivalence tests for shelf life and average drug content in stability studies

Stability testing is a procedure frequently used in the pharmaceutical industry to estimate the shelf life of a drug. Hereby, a standard problem of interest is whether or not to pool a given number of batches to assign a single shelf life for the combined batches. In this paper, we propose two modified methods for the pooling process. One is based on the relative average drug contents of batches at a target shelf life, which basically involves specification of thresholds in relative terms, and the other is tests for the equivalence of shelf lives (using ratios of shelf lives) at a specified lower acceptance criteria of label claim. For both approaches, marginal and simultaneous confidence intervals are established. The methods will be illustrated using two different stability data. A simulation study is conducted to investigate the coverage probability of the proposed methods.