ESTIMATING DRUG SHELF-LIFE WITH RANDOM BATCHES

The problem of assessing the shelf-lives of drug products in the market is considered. We focus on the situation where the drug characteristic of interest has a linear relationship over time and batch-to-batch variation is present. Two methods for assessing shelf-lives are proposed based on the weighted least squares method under a regression model with random coefficients. An application to some stability data from the pharmaceutical industry is presented.