One important study objective is to estimate the shelf-life of the drug, which is used to determine the expiration
date that is printed on the package label of the drug. According to the Guideline for Stability Testing, ICHQ1 (which is
harmonized among US, EU and Japan, and can be viewed at http://www.ich.org/ich5q.html), the shelf-life is defined as
the time point where the 95% pointwise lower confidence limit for the regression line crosses the lowest acceptable limit
for drug content, which is usually 90% of the labelled amount. Typically, multiple batches of a drug are manufactured,
and it is desirable to obtain a single shelf-life for all batches. When a random sample of dosage units from each batch
is obtained, there is possibly a large variability of estimated shelf-life from batch to batch, especially when samples of
dosage units from batches are not large. Currently ICHQ1 suggests statistical testing for the homogeneity of degradation
rates (assuming equal intercepts) and pooling data from batches with degradation rates not declared different by the
tests into a subgroup.A single shelf-life is then calculated for each subgroup with the shortest computed shelf-life used
as the final shelf-life of the drug.