Statistical comparison of stability study designs

Studies are routinely conducted during development of a new drug to investigate the effect of strength, package, batch, storage condition, and storage time on stability of the dosage form. A typical study will have three batches, several strengths, several packages, and several storage conditions. Thus, if every batch by strength by package combination is tested for every storage condition, i.e., if a complete factorial design is used, a substantial expense is involved. To reduce expense, a matrix design is commonly used; e.g., only the smallest and largest bottles are tested. Fractional factorial-type designs for stability studies are proposed, and the statistical powers of various designs are compared.