RUGGEDNESS TESTS FOR A HIGH-PERFORMANCE LIQUID-CHROMATOGRAPHIC ASSAY - COMPARISON OF AN EVALUATION AT 2 AND 3 LEVELS BY USING 2-LEVEL PLACKETT-BURMAN DESIGNS

Ruggedness tests are performed on a high-performance liquid chromatogr aphic assay for a drug substance. For this purpose two-level Plackett- Burman designs are executed. The factors selected are tested at two an d three levels. Significant effects of factors are determined graphica lly and statistically. For the latter purpose experimental error is es timated from dummy factor effects. The use of other error estimates, a s for instance the variance from replicate nominal experiments, as an alternative for the dummies was also examined. The dummy factor effect s were found suitable to estimate the experimental error in a Plackett -Burman design. Appropriate replicate injections at nominal levels cou ld be an alternative. Testing at three levels gives more detailed info rmation about the behaviour of a response in function of the factors. It was also found that not only the effects on qualitative responses s hould be considered but those on quantitative ones as well in order to be able to make a full interpretation concerning the ruggedness of th e method.